~ubuntu-branches/ubuntu/precise/code-saturne/precise

We assume in this section that the user has at his disposal the calculation data file (calculation set up) or already prepared it following for instance the step-by-step guidance provided in \CS tutorial. The steps described below are intended to provide the user a way to run quickly on a workstation a calculation through the Graphical User Interface (GUI).

114

115

The first thing to do before running \CS is to add in the user \textasciitilde\texttt{/.profile}, \textasciitilde\texttt{/.bashrc} or similar file the path

116

leading to the chosen \CS version, or define an alias to the \texttt{code\_saturne} script. For example:\

117

\begin{center}

118

\texttt{export PATH=\$\{prefix\}/bin:\$PATH}.\

119

\end{center}

120

The second thing is to prepare the computation directories. For instance, the study directory \texttt{T\_JUNCTION}, containing a single calculation directory CASE1, will be created by typing the command:\

121

\begin{center}

122

\texttt{code\_saturne create -s T\_JUNCTION}\

123

\end{center}

124

The mesh files should be copied in the directory \texttt{MESH}, and the Fortran user files necessary for the calculation in the directory \texttt{CASE1/SRC}. Finally, the calculation data file \texttt{case\_name.xml} read by the GUI should be copied to the directory \texttt{CASE1/DATA}.

125

Once these steps completed, the user should go in the directory \texttt{CASE1/DATA} and type de command line \texttt{./SaturneGUI case\_name.xml} to load the calculation file into the interface. A window similar to fig.\ref{fig3_e1} will appear. Click on the heading ``Calculation management'', select the heading ``Prepare batch calculation'', see fig.\ref{fig43_e1}. After having chosen the number of processors, press ``start calculation'' to run the calculation.

126

127

\begin{figure}[!ht]

128

\begin{center}

129

\includegraphics[width=12cm]{gui_case_dir}

130

\caption{Identity and paths}

131

\label{fig3_e1}

132

\end{center}

133

\end{figure}

134

135

\begin{figure}[!ht]

136

\begin{center}

137

\includegraphics[width=12cm]{gui_prepare_execution}

138

\caption{Prepare execution}

139

\label{fig43_e1}

140

\end{center}

141

\end{figure}

142

143

If no problem arises, the simulation results can be found in the directory \texttt{CASE1/RESU} and be read directly by {\em ParaView} or {\em EnSight} in \texttt{CASE1/RESU/<YYYYMMDD-hhmm>/postprocessing}. Calculation history can be found in the file \texttt{<YYYYMMDD-hhmm>/listing}.

144

145

%==================================

146

\subsection{Troubleshooting}

147

%==================================

148

If the calculation does not run properly, the user is advised to check the

149

following points in\\

150

\texttt{CASE1/RESU/<YYYYMMDD-hhmm>}:

151

\begin{list}{$\bullet$}{}

152

\item if the calculation stops in the pre-processor, the user should check for error messages in the file \texttt{preprocessor*.log}.

153

\item if the problem is related to boundary conditions, the user should visualise the file \texttt{error.ensight} with {\em EnSight} or {\em Paraview},

154

\item if the calculation stops in the \CS core, the user should look for messages at the end of the files \texttt{listing} and \texttt{error*}. In addition, the user can track the following keywords in the listing. They are specific error signals:

155

\begin{list}{-}{}

156

\item \texttt{SIGFPE}: a floating point exception occurred. It happens when there is a division by 0, when the calculation did not converge, or when a real number reached a value over $10^{300}$. Depending on the architecture \CS is running

157

on, this type of execption may be caught or ignored.

158

\item \texttt{SIGSEGV}: a memory error such as a segmentation violation occurred. An array may have exceeded its allocated memory size and a memory location in use was overwritten.

159

\end{list}

160

\end{list}

120

161

%==================================

121

162

%==================================

122

163

\section{Practical information about \CS}

133

174

\label{prg_environementCS}

134

175

135

176

In order to use \CS, every user must add the following line (in their

136

\texttt{.profile}, or equivalent, depending on the environment):

177

\texttt{.profile}, \texttt{.bashrc}, or equivalent,

178

depending on the environment):

137

179

138

\hspace*{1cm}\texttt{export PATH=\{prefix\}/bin:\$PATH}\\

180

\hspace*{1cm}\texttt{export PATH=\$\{prefix\}/bin:\$PATH}\\

139

181

140

182

or define the following alias (in their \texttt{.bashrc}, or

141

183

equivalent, or \texttt{.alias} file, depending on the environment):

142

184

143

\hspace*{1cm}\texttt{alias cs='\{prefix\}/bin/cs'}\\

185

\hspace*{1cm}\texttt{alias cs='\$\{prefix\}/bin/cs'}\\

144

186

145

187

where \texttt{prefix} is the base directory where

146

188

\CS and its components have been installed\footnote{At EDF R\&D, \texttt{/home/saturne/Code\_Saturne/\verscs} is used}.

152

194

The standard architecture for the simulation studies is:

153

195

154

196

\noindent

155

A study directory containing:

197

An optional study directory containing:

156

198

\begin{list}{$\bullet$}{}

157

199

\item A directory \texttt{MESH} containing the mesh(es)

158

200

necessary for the study

171

213

\item A directory \texttt{SCRIPTS} for the launch script

172

214

\item A directory \texttt{RESU} for the results\\

173

215

To improve the calculation traceability, the files and directories

174

sent to \texttt{RESU} after a calculation are given a suffix

175

identifying the calculation start date and time by an eight-digit

176

number (two digits for each month, day, hour and minute; the result of a

177

calculation started at 14h03 on December 31$^{\text{st}}$ will therefore be indexed

178

12311403)

216

sent to \texttt{RESU} after a calculation are placed in a subdirectory

217

named after that run's ``id'', which is by default based on the run date

218

and time, using the format: \texttt{YYYYMMDD-hhmm}.

219

It is also possible to force a specific run id, using the \texttt{--id}

220

option to \texttt{code\_saturne run}.

179

221

\end{list}

180

222

181

223

\noindent

182

In the standard cases, \texttt{RESU} contains a directory

183

\texttt{CHR.ENSIGHT.mmddhhmm} with the post-processing files in {\em EnSight} format,

184

a directory \texttt{RESTART.mmddhhmm} for the calculation

185

restart files,

186

a directory \texttt{HIST.mmddhhmm} for the files of chronological

187

record of the results at specific locations (probes),\\

188

\texttt{listpre.mmddhhmm} and \texttt{listing.mmddhhmm} files reporting the

189

Preprocessor and the Kernel execution. For an easier follow-up of the modifications

190

in former calculations, the user-subroutines used in a calculation are stored in

191

a directory \texttt{SRC.mmddhhmm} in the directory \texttt{RESU}. The {\em Xml}

224

In the standard cases, \texttt{RESU/<run\_id>} contains a

225

\texttt{postprocessing} directory with the post-processing

226

(visualization) files, a \texttt{restart} directory for the calculation

227

restart files, a \texttt{monitoring} directory for the files of chronological

228

record of the results at specific locations (probes),\\

229

\texttt{preprocessor.log} and \texttt{listing} files reporting the

230

Preprocessor and the Kernel execution. For an tracing of

231

the modifications in prior calculations, the user-subroutines used in

232

a calculation are stored in a \texttt{src\_saturne} subdirectory. The {\em Xml}

192

233

Interface data file, thermo-chemical data files and launch script are also

193

copied into the directory \texttt{RESU} with the appropriate suffix (whatever

194

its name, the launch script appears in the directory \texttt{RESU} as

195

\texttt{runcase.mmddhhmm}). \texttt{compil.log.mmddhhmm} and

196

\texttt{summary.mmddhhmm} are respectively reports of the compilation phase and

197

general information on the calculation (which kind of machine, which user, which

198

version of the code, ...). Eventually, when the user subroutines produce

199

specific result files (extraction of 1D profiles for instance), a

200

directory \texttt{RES\_USERS.mmddhhmm} is created in the directory \texttt{RESU}

201

for these files\footnote{in order for the script to copy them properly, their

202

names have to be given in the variable \texttt{USER\_OUTPUT\_FILES}

203

of the launch script, see \S\ref{prg_runcase}}.

204

205

206

During calculations coupled with \syrthes (option specified in the launch

207

script of \CS or {\em via} the Interface) the same organisation is used for the

208

files related to \CS. For the files related

209

to \syrthes, the location of the upstream files is specified in the

210

\texttt{syrthes.env} file. Yet, the launch script is built presuming that the

211

following organisation is applied:

212

\begin{list}{$\bullet$}{}

213

\item a directory \texttt{SRC\_SYR} for the potential \syrthes user subroutines

214

\item a directory \texttt{DATA\_SYR} containing the configuration file \texttt{syrthes.env}

215

(location of files specific to \syrthes). The file defining the \syrthes

216

calculation options (\texttt{syrthes.data}) and the potential restart files can

217

also be placed in this directory.

218

\end{list}

219

220

The \syrthes result files (geometry file, chronological result files, calculation

221

restart files and the historic file) are placed in a sub-directory

222

\texttt{RESU\_SYR.mmddhhmm} of the \texttt{RESU} directory, where

223

\texttt{mmddhhmm} corresponds to the calculation identification suffix.

224

225

\sloppy

226

227

\noindent

228

The \syrthes execution report file is placed in the \texttt{RESU} directory (same as

229

for the \CS review) under the name \texttt{listsyr.mmddhhmm} and the compilation

230

report file is under the name \texttt{compil\_syrthes.log.mmddhhmm}.

231

For an easier follow-up of the modifications

232

in former calculations, the potential \syrthes user-subroutines used in a

233

calculation are stored in a directory \texttt{SRC\_SYR.mmddhhmm} in the

234

directory \texttt{RESU}.

235

236

\fussy

237

238

239

\begin{table}[!t]

234

copied into the results directory. \texttt{compil.log} and

235

\texttt{summary} are respectively reports of the compilation stage and

236

general information on the calculation (type of machine, user,

237

version of the code, ...).

238

239

\begin{table}[h!t]

240

\begin{tabular}{lll}

241

\multicolumn{3}{l}{Below are typical contents of a case directory CASE1 in a study STUDY} \\

242

\multicolumn{3}{l}{(\CS calculation coupled with \syrthes):}\\

243

242

\multicolumn{2}{l}{\texttt{STUDY/CASE1/DATA:}}&{\bf \CS data}\\

244

& \texttt{SaturneGUI} &Graphical User Interface launch script\\

245

& \texttt{study.xml} &Graphical User Interface parameter file\\

246

& \texttt{THCH} &example of thermochemical files

247

(used with the specific\\

248

& & physics modules for gas combustion,

249

pulverised coal \\

250

& & or electric arcs)\\

251

252

\multicolumn{2}{l}{\texttt{STUDY/CASE1/DATA\_SYR:}}&{\bf \syrthes data}\\

253

& \texttt{syrthes.data} &\syrthes data file \\

254

& \texttt{syrthes.env} &\syrthes configuration file\\

243

& \texttt{SaturneGUI} &Graphical User Interface launch script\\

244

& \texttt{study.xml} &Graphical User Interface parameter file\\

245

& \texttt{REFERENCE} &Example of user scripts and meteorological\\

246

& & or thermochemical date files (used with the\\

247

& & specific physics modules)\\

255

248

\multicolumn{2}{l}{\texttt{STUDY/CASE1/SRC:}}&{\bf \CS user subroutines }\\

256

& \texttt{REFERENCE} & examples of a user subroutines\\

257

& \texttt{usclim.f90} &user subroutines used for the present the calculation\\

249

& \texttt{REFERENCE} & Examples of a user subroutines\\

250

& \texttt{usclim.f90} & user subroutines used for the present calculation\\

258

251

& \texttt{usini1.f90} &\\

259

\multicolumn{2}{l}{\texttt{STUDY/CASE1/RESU:}}&{\bf results}\\

260

& \texttt{CHR.ENSIGHT.08211921} &directory containing the \CS

261

post-processing results\\

262

& &in the {\em EnSight} format for the

263

calculation 08211921\\

264

& &(contains both volume and boundary results;\\

265

& &the contents of the directory are user

266

modifiable)\\

267

& \texttt{SRC.08211921} &\CS user subroutines

268

used for the\\

269

& &calculation 08211921\\

270

& \texttt{SRC\_SYR.08211921} &\syrthes user subroutines

271

used in the calculation 08211921\\

272

& \texttt{HIST.08211921} &directory containing the chronological

273

records for \CS\\

274

& \texttt{RES\_USERS.08211921} &optional directory containing the

275

user results files\\

276

& \texttt{RESTART.08211921} &directory containing the \CS restart files \\

277

& \texttt{compile.log.08211921} &compilation report\\

278

& \texttt{study.xml.08211921} &Graphical User Interface parameter file

279

used for the\\

280

& &calculation 08211921\\

281

& \texttt{runcase.08211921} &launch script used for the calculation 08211921\\

282

& &(whatever the name given to the file in the

283

\texttt{SCRIPT} directory,\\

284

& & the file will be referred as

285

``\texttt{runcase.*}'' in the

286

\texttt{RESU} directory)\\

287

& \texttt{listpre.08211921} &execution report for the Preprocessor module of

288

\CS\\

289

& \texttt{listing.08211921} &execution report for the Kernel module of \CS\\

290

& \texttt{listsyr.08211921} &execution report for \syrthes\\

291

& \texttt{summary.08211921} &general information (machine, user, version, ...)\\

292

& \texttt{RESU\_SYR.08211921:} &{\bf \syrthes results (file names given in

293

the

294

\texttt{syrthes.env} file)}\\

295

& \hspace*{0.3cm}\texttt{geoms} &\syrthes \ solid geometry file\\

296

& \hspace*{0.3cm}\texttt{histos1}&\syrthes chronological records at

297

specified probes\\

298

& \hspace*{0.3cm}\texttt{resus1}&\syrthes calculation restart file (1

299

time step)\\

300

& \hspace*{0.3cm}\texttt{resusc1}&\syrthes chronological solid

301

post-processing file (may be transformed \\

302

& &into the {\em EnSight} format with the

303

{\em syrthes2ensight} utility)\\

252

\multicolumn{2}{l}{\texttt{STUDY/CASE1/RESU/20110509-1920:}}&{\bf results} for the

253

calculation 20110509-1920\\

254

& \texttt{postprocessing} &directory containing the \CS post-processing output\\

255

& &in the {\em EnSight} format (both volume and boundary);\\

256

& \texttt{src\_saturne} &copy of the \CS user subroutines used for the calculation\\

257

& \texttt{monitoring} &directory containing the chronological records for \CS\\

258

& \texttt{checkpoint} &directory containing the \CS restart files \\

259

& \texttt{compile.log} &compilation log\\

260

& \texttt{study.xml} &Graphical User Interface parameter file used for the\\

261

& &calculation\\

262

& \texttt{runcase} &copy of the launch script used for the calculation\\

263

& \texttt{preprocessor.log} &execution report for the \CS Preprocessor\\

264

& \texttt{listing} &execution report for the Kernel module of \CS\\

265

& \texttt{summary} &general information (machine, user, version, ...)\\

304

266

\multicolumn{2}{l}{\texttt{STUDY/CASE1/SCRIPTS:}}&{\bf launch script}\\

305

& \texttt{runcase} &launch script (compliant with all

306

architectures on which \\

307

& & \CS has been ported)\\

308

\end{tabular}

309

\end{table}

310

311

\clearpage

312

267

& \texttt{runcase} &launch script (which may contain batch

268

system keywords)\\

269

\end{tabular}

270

\end{table}

271

272

Note that the code may be run directly in the final \texttt{RESU/<run\_id>}

273

directory, or in a scratch directory (which may be recommended if the

274

compute environment includes different filesystems, some better suited

275

to data storage, others to intensive I/O). When running, the code

276

may use additional files or directories inside its execution directory, set

277

by the execution script, which include a \texttt{mesh\_input} file or directory,

278

as well as a \texttt{restart} directory (which is a link or copy of a previous

279

run's \texttt{checkpoint} directory), as well as a \texttt{run\_solver.sh}

280

script.

281

282

For coupled calculations, whether with \CS or \syrthes, each coupled

283

calculation domain is defined by its own directory (bearing the same name

284

as the domain), but results are placed in a \texttt{RESU\_COUPLING}

285

directory, with a subdirectory for each run, itself containing one

286

subdirectory per coupled domain. Coupled cases are not run through

287

the standard {\texttt{STUDY/CASE1/SCRIPTS/runcase} script or through

288

the {\texttt{code\_saturne run} command, but through a

289

{\texttt{STUDY/runcase\_coupling} script.

290

291

So in the coupled case, calculation results would not be placed in

292

\texttt{STUDY/CASE1/RESU/20110509-1920}, but in

293

\texttt{STUDY/RESU\_COUPLING/20110509-1920/CASE1}, with the \texttt{summary}

294

file being directly placed in \texttt{STUDY/RESU\_COUPLING/20110509-1920}

295

(as it references all coupled domains).

296

297

\begin{table}[h!t]

298

\begin{tabular}{lll}

299

\multicolumn{3}{l}{Below are typical additional contents with a coupled \syrthes

300

case SOLID1 in a study STUDY} \\

301

\multicolumn{2}{l}{\texttt{STUDY/runcase\_coupling}}&{coupled launch script}\\

302

\multicolumn{2}{l}{\texttt{STUDY/SOLID1/DATA:}}&{\bf \syrthes data}\\

303

& \texttt{syrthes.data} &\syrthes data file \\

304

& \texttt{syrthes.env} &\syrthes configuration file\\

305

\multicolumn{2}{l}{\texttt{STUDY/RESU\_COUPLING/20110509-1920/SOLID1:}}&{\bf results

306

(file names defined in \texttt{syrthes.env})}\\

307

& \texttt{src} &\syrthes user subroutines

308

used in the calculation\\

309

& \texttt{compile.log.08211921} &\syrthes compilation report\\

310

& \texttt{listsyr} &execution log\\

311

& \texttt{geoms} &\syrthes \ solid geometry file\\

312

& \texttt{histos1} &\syrthes chronological records at

313

specified probes\\

314

& \texttt{resus1} &\syrthes calculation restart file (1 time step)\\

315

& \texttt{resusc1} &\syrthes chronological solid

316

post-processing file\\

317

& &(may be transformed into the {\em EnSight}

318

format\\

319

& &with the {\em syrthes2ensight} utility)\\

320

\end{tabular}

321

\end{table}

313

322

314

323

%==================================

315

324

\subsubsection{\CS Kernel library files}

342

351

\texttt{elec} (electric module),

343

352

\texttt{fuel} (heavy fuel oil combustion module),

344

353

\texttt{lagr} (Lagrangian module,

345

\texttt{mati} (Matisse module),

346

354

\texttt{pprt} (general specific physics routines) and

347

355

\texttt{rayt} (radiative heat transfer).

348

356

The case preparer command \texttt{code\_saturne~create} copies all these files

349

357

in the user directory \texttt{SRC/REFERENCE} during the case preparation.

350

358

351

The ``include'' files are available in the directory

352

\texttt{include}, under similar subdirectories corresponding to each module:

353

\texttt{base}, \texttt{cfbl}, \texttt{cogz}, \texttt{cplv}, \texttt{elec},

354

\texttt{fuel}, \texttt{lagr}, \texttt{mati}, \texttt{pprt} and \texttt{rayt}.

355

356

359

The directory \texttt{bin} contains an example of the launch script, the

357

360

compilation parameter files and various utility programs.

358

361

359

362

%==================================

360

\subsection{Setting up and running of a calculation}

363

\subsection{Setting up and running a calculation}

361

364

%==================================

362

365

363

366

%==================================

377

380

\S\ref{prg_environementCS}).

378

381

379

382

\item Prepare the different directories using the \texttt{code\_saturne~create}

380

command (see \S\ref{prg_cscreate}) and, when needed, add the directories

381

\texttt{DATA\_SYR} and \texttt{SRC\_SYR} which are required to accomodate the

382

\syrthes files.

383

command (see \S\ref{prg_cscreate}).

383

384

\item Place the mesh(es) in the directory \texttt{MESH}. Make sure they are

385

\item It is recommendned to place the mesh(es) in the directory \texttt{MESH},

386

but they may be selected from other directories. Make sure they are

385

387

in a format compliant with \CS (see \S\ref{prg_maillages}). There can be

386

388

several meshes in case of mesh joining or coupling with

387

\syrthes\footnote{\syrthes uses meshes composed of 10-node tetrahedra (vertices

388

and centers of edges)}.

389

\syrthes\footnote{\syrthes 3 uses meshes composed of 10-node tetrahedra (vertices

390

and centers of edges, \syrthes 4 uses meshes composed of 4-node tetrahedra}.

389

391

390

392

\item Go to the directory \texttt{DATA} and launch the

391

Graphical User Interface using the command \texttt{./SaturneGUI} (see

392

\S\ref{prg_gui}).

393

Graphical User Interface using the command \texttt{./SaturneGUI}.

394

395

\item If not using the GUI, copy the

396

\texttt{DATA/REFERENCE/cs\_user\_scripts.py} file to \texttt{DATA} and

397

edit it, so that the correct run options and paths may be set. For advanced

398

uses, this file may also be used in conjunction with the GUI. Jus as with

399

user Fortran subroutines below, settings defined in this file have priority

400

over those defined in the GUI.

393

401

394

402

\item Place the necessary user subroutines in the directory \texttt{SRC} (see

395

403

\S\ref{prg_ssprgutilis}). When not using the Interface, some subroutines are

416

424

\end{list}

417

425

\end{list}

418

426

419

\item{\bf For the specific physics ``gas combustion'':}

427

\item{\bf For the ``gas combustion'' specific physics:}

420

428

421

429

(not accessible through the Graphical User Interface in version \verscs)

422

430

\begin{list}{}{}

448

456

\end{list}

449

457

\end{list}

450

458

451

\item{\bf For the specific physics ``coal combustion'':}

459

\item{\bf For the ``coal combustion'' specific physics:}

452

460

453

461

\begin{list}{}{}

454

462

\item {\em compulsory without Graphical User Interface:}

475

483

\end{list}

476

484

\end{list}

477

485

478

\item{\bf For the specific physics ``electric module''

486

\item{\bf For the ``electric module'' specific physics

479

487

(Joule effect and electric arcs):}

480

488

481

489

(not accessible through the Graphical User Interface in version \verscs)

507

515

\end{list}

508

516

\end{list}

509

517

510

\item{\bf For the specific physics ``heavy fuel oil combustion module'':}

518

\item{\bf For the ``heavy fuel oil combustion module'' specific physics:}

511

519

512

520

(not accessible through the Graphical User Interface in version \verscs)

513

521

\begin{list}{}{}

588

596

external data (input profiles, thermochemical data files, ...)

589

597

590

598

\item Prepare the launch script \texttt{runcase}, directly or through the

591

Graphical Interface (see \S\ref{prg_runcase})

599

Graphical Interface (see \S\ref{prg_runcase}), or prepare the

600

\texttt{DATA/cs\_user\_scripts.py} file.

592

601

593

602

\item Run the calculation and analyse the results

594

603

595

\item Purge the temporary files (in the directory \texttt{RUN} defined

596

in the launch script, see \S\ref{prg_runcase})

604

\item Purge the temporary files (in \texttt{RESU/<run\_id>} or

605

\texttt{<scratch>/<run\_id>} directory).

597

606

\end{list}

598

607

599

608

601

610

\subsubsection{Temporary execution directory}

602

611

%==================================

603

612

\label{prg_temporarydirectory}%

604

During a calculation, \CS uses a temporary directory for the compilation and

605

the execution, the result files being only copied at the end in the directory

606

\texttt{RESU}. This temporary directory is defined in the variable \texttt{RUN}

607

of the launch script. This variable is set top a default value in the non-user

608

section of the launch script, depending on the architecture:\\

609

\texttt{RUN=\$HOME/tmp\_Saturne/\$STUDY/\$CASE.mmddhhmm} for stand-alone

610

workstations or for the Chatou cluster\\

611

\texttt{RUN=\$SCRATCHDIR/tmp\_Saturne/\$STUDY/\$CASE.mmddhhmm} for

612

Platine at the CCRT\\

613

where \texttt{\$STUDY} and \texttt{\$CASE} are the names of the study and the

614

case. The usual suffix with the date and time is added so that successive

615

calculations will not get mixed-up.

616

617

This default value might not always be the optimal choice. Indeed, on a

618

stand-alone machine, it might be useful to take advantage of large sized local

619

disks on a machine when the \texttt{\$HOME} account is on an NFS disk.

620

621

For this matter, the variable \texttt{CS\_TMP\_PREFIX} of the launch script (see

622

\S\ref{prg_runcase}) allows the user to change this directory.

623

If the variable is empty, the default

624

\texttt{RUN} directory will be used. If it is not empty, the launch script will

625

set the \texttt{RUN} directory to

626

\texttt{\$CS\_TMP\_PREFIX/tmp\_Saturne/\$STUDY/\$CASE.mmddhhmm}.

613

During a calculation, \CS may use a temporary directory for the compilation and

614

the execution if such a ``scratch'' directory is defined. In that case, the

615

result files are only copied at the end in the directory

616

\texttt{RESU}. This is recommended if the compute environment includes different

617

filesystems, some better suited to data storage, others to intensive I/O.

618

If this is not the case, there is no point in running in a scratch directory

619

rather than the results directory, as this incurs additional file copies.

627

620

628

621

\noindent

629

{\em WARNING: in most cases, the temporary directories are not deleted

630

after a calculation. They will accumulate and may hinder the correct running of

631

the machine.\\

622

{\em WARNING: in case of an error, the temporary directories are not deleted

623

after a calculation, so that they may be used for debugging. They may then

624

accumulate and may hinder the correct operation of the machine.\\

632

625

\centerline{\bf It is therefore essential to remove them regularly.}}

633

626

634

627

637

630

%==================================

638

631

\label{prg_executionmodes}%

639

632

As explained before, \CS is composed of two main modules, the Preprocessor and the

640

Kernel, and an optional Partitioner. The Preprocessor reads the meshes and

641

performs the necessary joinings. The Partitioner optimizes domain decomposition

642

for parallel runs. The resulting data is transfered to the Kernel through specific

643

files, named \texttt{preprocessor\_output} and \texttt{domain\_number\_*}, where

633

Kernel, and an optional Partitioner. The Preprocessor reads the meshes.

634

The Partitioner optimizes domain decomposition for parallel runs.

635

The resulting data is transfered to the Kernel through specific

636

files, named \texttt{mesh\_input} and \texttt{domain\_number\_*}, where

644

637

\texttt{*} is the number of processors used. In a standard calculation, the files

645

638

are not copied from the temporary execution directory to the results directory,

646

639

as they have no interest for data analysis, and are considered ``internal''

647

640

files, whose format or contents is not guaranteed not to change between \CS versions.

648

641

649

Yet, the Preprocessor and Partitioner do not work in parallel and may require a

650

large amount of memory. Hence it is sometimes useful to run them

642

Yet, the Preprocessor does not run in parallel and may require a

643

large amount of memory. Similarly, the Partitioner may be run on a

644

reduced number of processors to obtain a better partition quality, so it

645

is sometimes useful to run the Preprocessor and Partitioner

651

646

separately, on a machine or in batch queues with extended memory, and to run the

652

647

proper parallel calculation on another machine or in another batch

653

648

queue. The launch scripts therefore allows specifically choosing

654

which modules to run, using the variables \texttt{EXEC\_PREPROCESS},

655

\texttt{EXEC\_PARTITION}, and \texttt{EXEC\_KERNEL} (see \S\ref{prg_runcase}).

649

which modules to run, either through the GUI or through \texttt{cs\_user\_scripts.py}:

656

650

657

651

\sloppy

658

652

659

\hspace*{0.5cm} If \texttt{EXEC\_PREPROCESS=no}, the Partitioner and Kernel will

660

copy or link the file defined by the \texttt{PREPROCESSOR\_OUTPUT\_IN}

661

variable to \texttt{preprocessor\_output}.

662

663

\hspace*{0.5cm} If \texttt{EXEC\_PARTITION=no}, the Kernel will search for a

664

corresponding \texttt{domain\_number\_*} file in the directory defined by

665

\texttt{PARTITION\_OUTPUT\_IN}.

666

667

\hspace*{0.5cm} If \texttt{EXEC\_KERNEL=no}, the Preprocessor and Partitioner

668

output will be saved respectively to a

669

\texttt{\$RESU/preprocessor\_output.\$SUFFIX} file

670

and to a \texttt{\$RESU/PARTITION\_OUTPUT.\$SUFFIX} directory.

671

In this case, the number of processors for which partitioning is required

672

is given by the \texttt{PARTITION\_LIST} variable of the launch script

673

(multiple partitionings may be done in one run).

653

\hspace*{0.5cm} If a {mesh\_input} file or directory is defined (which may be

654

either a {mesh\_input} from a previous Preprocessor run or a {mesh\_output}

655

from a previous solver run), the script will copy or link it to

656

the execution directory, and the Preprocessor will not be run again.

657

658

\hspace*{0.5cm} If \texttt{domain.exec\_partition = False}, the Partitioner

659

will not be run. A previous partitioning may be used by defining the

660

\texttt{domain.partition\_input} path.

661

662

\hspace*{0.5cm} If \texttt{domain.exec\_kernel = False}, the Kernel will not

663

be run. This is useful when only the mesh import and optionnaly partitioning

664

stages are required.

674

665

675

666

\fussy

676

667

668

It is encouraged to separate the mesh import and calculation runs, as

669

this speeds up calculations by not re-importing meshes for each run.

670

For some configurations, such as IBM Blue Gene machines with different

671

front-end an compute nodes, mesh import may be impossible on the

672

compute nodes (as the Preprocessor does not run in parallel, and may require

673

too much memory), so mesh import (and serial partitioning) should be

674

run separately on the front-end nodes, while later calculation stages

675

should be run on compute nodes.

676

677

Note also that depending on its configuration, the Partitioner may

678

be run either or both in serial or parallel mode. By default, serial

679

mode is currently chosen (due to limited feedback on partitioning

680

quality in parallel mode), but setting \texttt{domain.partition\_n\_procs}

681

to a value greater than 1 in the user scripts allows running

682

the Partitioner in parallel.

683

677

684

%==================================

678

685

\subsubsection{Interactive modification of the target time step}

679

686

%==================================

714

721

\texttt{DEBIT3} and \texttt{DEBIT4}.

715

722

716

723

An option \texttt{--nogui} is available for the use of \CS

717

without Graphic Interface (see \S\ref{prg_gui}). This option must

724

without Graphic Interface. This option must

718

725

be either the first or the last argument and appear only once.

719

726

720

727

In the directory \texttt{DATA}, the \texttt{code\_saturne~create} command

763

770

764

771

\item \hyperref[fmtdesc:des]{\simail (NOPO)}

765

772

\item \hyperref[fmtdesc:unv]{\ideas universal}

766

\item \hyperref[fmtdesc:igghexa]{NUMECA IGG/Hexa}

767

773

\item \hyperref[fmtdesc:med]{\med}

768

774

\item \hyperref[fmtdesc:cgns]{CGNS}

769

775

\item \hyperref[fmtdesc:ensight6]{\ensight 6}

828

834

Volume elements: & tetrahedra, prisms, hexahedra\\

829

835

\hline

830

836

Zone selection: & element face references and volume sub-domains\\

831

& (interpreted as numbered colors)\\

837

& (interpreted as numbered groups)\\

832

838

\hline

833

839

Compatibility: & all files of this type as long as the coordinate

834

840

system used is Cartesian and not cylindrical or

879

885

\hline

880

886

Volume elements: & tetrahedra, prisms, hexahedra\\

881

887

\hline

882

Zone selection: & colors (systematic) and named groups\\

888

Zone selection: & colors (always) and named groups\\

883

889

\hline

884

890

Compatibility: & \ideas (\emph{Master Series} 5 to 9, \emph{NX Series} 10 to 12)

885

891

at least\\

930

936

when considering definition of descriptions), so support for this format

931

937

is lightly tested.

932

938

933

Currently, geometric entity numbers are converted to ``color'' numbers.

934

This tends to lead to a large number of colors.

939

Currently, geometric entity numbers are converted to group numbers.

940

This tends to lead to a large number of groups.

935

941

936

942

\smallskip \noindent

937

943

\begin{tabular}[top]{|p{4.5cm}%

946

952

Volume elements: & polyhedra\\

947

953

\hline

948

954

Zone selection: & face and cell sets\\

949

& (interpreted as numbered colors)\\

955

& (interpreted as numbered groups)\\

950

956

\hline

951

957

Compatibility: & all files of this type ? (tested on purely polyhedral meshes)\\

952

958

\hline

965

971

of ADF modified for performance, but also works with a standard version

966

972

from CGNS.

967

973

968

Currently, geometric entity numbers are converted to ``color'' numbers,

974

Currently, geometric entity numbers are converted to numbered groups,

969

975

with the corresponding names printed to the \pcs log. Depending on whether

970

976

the names were generated automatically or set by the user, it would be

971

preferable to interpret such sets as named ``groups'' rather than numbered

972

``colors''.

977

preferable to use the original group names rather than base their

978

names on their numbers.

973

979

974

980

The CCMIO library is distributed freely by CD-Adapco upon demand.

975

981

986

992

Volume elements: & polyhedra\\

987

993

\hline

988

994

Zone selection: & named face and cell sets\\

989

& (interpreted as numbered colors, with names appearing in log)\\

995

& (interpreted as numbered groups, with names appearing in log)\\

990

996

\hline

991

997

Compatibility: & all files of this type ? (tested on purely polyhedral meshes)\\

992

998

\hline

1009

1015

mesh may also be separated into several \emph{parts} so as to identify

1010

1016

different zones. As \emph{part} names may contain up to 80 characters,

1011

1017

we do not transform them into groups (whose names could be unwieldy),

1012

so we simply interpret their number as a color.

1018

so we simply convert their numbers to group names.

1013

1019

1014

1020

Also note that files produced by \harpoon may contain badly oriented

1015

1021

prisms, so the \pcs orientation correction option

1016

1022

(\texttt{--reorient})may need to be used. Meshes built by this tool also

1017

1023

contain hanging nodes, with non-conforming elements sharing some vertices.

1018

The \texttt{--join --semi-conf} preprocessor option must thus be used.

1019

This option is not set automatically, as the user may prefer to specify

1020

which surfaces should be joined, and which ones should not (i.e. to

1021

conserve thin walls).

1024

Mesh joining must thus also be used, and is not activated automatically,

1025

as the user may prefer to specify which surfaces should be joined,

1026

and which ones should not (i.e. to conserve thin walls).

1022

1027

1023

1028

\smallskip \noindent

1024

1029

\begin{tabular}[top]{|p{4.5cm}%

1036

1041

\hline

1037

1042

Volume elements: & tetrahedra, pyramids, prisms, hexahedra\\

1038

1043

\hline

1039

Zone selection: & part numbers interpreted as color numbers\\

1044

Zone selection: & part numbers interpreted as numbered groups\\

1040

1045

\hline

1041

1046

Compatibility: & All files of this type\\

1042

1047

\hline

1043

1048

Documentation: & online documentation, also available at:

1044

\url{http://www.ensight.com/Downloads/}\\

1049

\href{http://www.ensight.com/downloads/cat\_view-5.html}

1050

{http://www.ensight.com/downloads/cat\_view-5.html}\\

1045

1051

\hline

1046

1052

\end{tabular}

1047

1053

1062

1068

arbitrary number of labels with each element, but files produced

1063

1069

by \gmsh use 2 labels, with the same meanings as with version 1.

1064

1070

1065

We chose to convert physical entity numbers to colors. It is possible

1071

We chose to convert physical entity numbers to groups. It is possible

1066

1072

to build a mesh using \gmsh without defining any physical entities

1067

(in which case all elements will have the same color, but the \gmsh

1073

(in which case all elements will belong to the same group, but the \gmsh

1068

1074

documentation clearly says that geometric entities are to be used

1069

1075

so as to group elementary entities having similar ``physical'' meanings.

1070

1076

1071

So as to obtain distinct colors with a mesh generated by \gmsh, it

1077

So as to obtain distinct groups with a mesh generated by \gmsh, it

1072

1078

is thus necessary for the user to define physical entities.

1073

1079

This requires an extra step, but allows for fine-grained control

1074

over the colors associated with the mesh, while using only elementary

1075

entities could lead to a high number of colors.

1080

over the groups associated with the mesh, while using only elementary

1081

entities could lead to a high number of groups.

1076

1082

1077

1083

\smallskip \noindent

1078

1084

\begin{tabular}[top]{|p{4.5cm}%

1086

1092

\hline

1087

1093

Volume elements: & tetrahedra, pyramids, prisms, hexahedra\\

1088

1094

\hline

1089

Zone selection: & physical entity numbers interpreted as color numbers\\

1095

Zone selection: & physical entity numbers interpreted as numbered groups\\

1090

1096

\hline

1091

1097

Compatibility: & all files of this type\\

1092

1098

\hline

1096

1102

\hline

1097

1103

\end{tabular}

1098

1104

1099

\subsubsubsection{IGG/Hexa (NUMECA)%

1100

\label{fmtdesc:igghexa}}

1101

1102

This format is quite peculiar in the sense that it defines a hierarchical mesh

1103

built exclusively of hexahedra, of quadrangular faces, and of edges, in which

1104

edges may be split in 2, faces in 2 or 4, and cells in 2, 4, or 8. Two

1105

neighboring elements are thus not always conforming. The \pcs only keeps the

1106

finest mesh level, and uses the hierarchical information so as to

1107

automatically build the appropriate conformal joining. CAD face

1108

information is transformed into color numbers.

1109

1110

\noindent

1111

\parbox[t]{.15\linewidth}{{\bf Remark}:}

1112

\parbox[t]{.85\linewidth}{

1113

The reader is based on the IGG/Hexa format as it was defined at the end of

1114

2001. Since that time, the mesher seems to have been renamed $HEXPRESS^{TM}$, and

1115

we do not have any recent documentation or test file. For this reason,

1116

support for this format is not directly included in the main \pcs executable,

1117

but requires conversion using the \texttt{igghexa\_to\_med} tool.

1118

}

1119

1120

\smallskip \noindent

1121

\begin{tabular}[top]{|p{4.5cm}%

1122

|>{\PreserveBackslash\raggedright\hspace{0pt}}p{10.5cm}|}

1123

\hline

1124

Default extension: & {\tt .hex}\\

1125

\hline

1126

File type: & portable binary by default

1127

(4-byte integers, 8-byte IEEE double-precision reals,

1128

big-endian), or text\\

1129

\hline

1130

Surface elements: & quadrangles\\

1131

\hline

1132

Volume elements: & hexahedra\\

1133

\hline

1134

Zone selection: & no volume selection, CAD surface numbers interpreted as

1135

colors\\

1136

\hline

1137

Compatibility: & unknown (at least files from late 2001 are readable)\\

1138

\hline

1139

Documentation: & elements provided by NUMECA (\href{http://www.numeca.be}

1140

{http://www.numeca.be})

1141

in 2001.\\

1142

\hline

1143

\end{tabular}

1144

1145

1105

%==================================

1146

1106

\subsubsection{Formats supported for input or output\label{cha:formats_inout}}

1147

1107

%==================================

1166

1126

do not necessarily support the whole format specification.

1167

1127

Especially, VTK does not support material types, and has only recently

1168

1128

added support for polyhedral elements in \ensightg files (interpreted

1169

as convex point sets, and not true polyhedra, but at least usable).

1129

as convex point sets in \paraview versions 2.4 to 2.8, and as true

1130

polyhedra starting with \paraview versions 2.10).

1170

1131

Also, both \ensightg (8.2 and above) and VTK allow for automatic distribution,

1171

1132

reducing the usefulness of pre-distributed meshes with per-processor files.

1172

1133

1176

1137

volume zones may only be considered to be part of the same mesh

1177

1138

if the file defines vertex IDs (which we consider to be

1178

1139

unique vertex labels). In this case, \emph{part} numbers

1179

are interpreted as colors. Without vertex IDs, only one part is read,

1180

and no colors are assigned.

1140

are interpreted as group names. Without vertex IDs, only one part is read,

1141

and no groups are assigned.

1181

1142

1182

1143

\smallskip \noindent

1183

1144

\begin{tabular}[top]{|p{4.5cm}%

1193

1154

Volume elements: & tetrahedra, pyramids, prisms, hexahedra, convex polyhedra\\

1194

1155

\hline

1195

1156

Zone selection: & possibility of defining element materials (not used), or

1196

interpret part number as color number if vertex IDs are

1157

interpret part number as group name if vertex IDs are

1197

1158

given\\

1198

1159

\hline

1199

1160

Compatibility: & files readable by \ensight 7.4 to 9.0, as well as tools

1201

1162

especially \paraview\ (\url{http://www.paraview.org})\\

1202

1163

\hline

1203

1164

Documentation: & online documentation, also available at:

1204

\url{http://www.ensight.com/Downloads/}\\

1165

\url{http://www.ensight.com/downloads/cat\_view-5.html}\\

1205

1166

\hline

1206

1167

\end{tabular}

1207

1168

1245

1206

\hline

1246

1207

Zone selection: & element families ({\it i.e.} colors and groups)\\

1247

1208

\hline

1248

Compatibility: & \med 2.3.5, \med 2.3.6, or \med 3.0.2 and above (only unstructured

1249

nodal connectivity is supported)\\

1209

Compatibility: & \med 2.3.5, \med 2.3.6, or \med 3.0.2 and above

1210

(only unstructured nodal connectivity is supported)\\

1250

1211

\hline

1251

1212

Documentation: & online documentation. Download link at \url{http://files.salome-platform.org/Salome/other/med-3.0.3.tar.gz}\\

1252

1213

\hline

1335

1296

selection, but the \pcs allows creation of groups associated to

1336

1297

zones or sections in the mesh using mesh selection sub-options\\

1337

1298

\hline

1338

Compatibility: & CGNS 2.5 or 3.1 on output\\

1299

Compatibility: & CGNS 2.5 or CGNS 3.1\\

1339

1300

\hline

1340

1301

Documentation: & See CGNS site: \url{http://www.cgns.org}\\

1341

1302

\hline

1342

1303

\end{tabular}

1343

1304

1344

1305

%==================================

1345

\subsubsection{Mesh meta-files}

1346

%==================================

1347

1348

The \pcs allows use of text files (preferably with a {\tt .mesh} extension)

1349

describing a set of meshes and their transformations, in place of (or

1350

combined with) ``true'' mesh files. These meta-files are

1351

described here:

1352

1353

Empty lines are ignored, and the \# character may be used to define

1354

comments (the part of a line following this character is ignored).

1355

1356

One may request the reading of as many meshes as one needs, using

1357

for each mesh a section such as:

1358

1359

\noindent

1360

\verb+read_mesh:+ {\it filename}

1361

1362

or:

1363

1364

\noindent

1365

\verb+read_mesh:+ {\it filename} {\tt <sub-options>}

1366

1367

If this section type appears more than once, the corresponding meshes

1368

are automatically appended. If needed, a mesh meta-file may itself

1369

declare another meta-file as a mesh file.

1370

Possible sub-options associated with a file may be separated by commas,

1371

semicolumns, or spaces, and are of the following form:

1372

1373

\noindent

1374

\begin{tabular}[top]{ p{2.5cm}%

1375

>{\PreserveBackslash\raggedright\hspace{0pt}}p{12.5cm} }

1376

{\tt format=} & {format name (identical to command-line options)}\\

1377

{\tt num=} & {mesh number (useful when a file contains multiple meshes)}\\

1378

{\tt grp\_cel=} & {{\tt <section} or {\tt zone>}, useful only for files in

1379

CGNS format}\\

1380

{\tt grp\_fac=} & {{\tt <section} or {\tt zone>}, useful only for files in

1381

CGNS format}\\

1382

\end{tabular}

1383

1384

\smallskip

1385

It is also possible to define a geometric transformation to apply to a mesh,

1386

using a homogeneous coordinates transformation matrix (3 lines, 4 columns,

1387

with the 3 first columns describing a rotation/scaling factor,

1388

and the last column describing a translation. The corresponding section

1389

is as follows (values may be spread over several lines):

1390

1391

\noindent

1392

\verb+transformation_matrix:+ $t_{11}$ $t_{12}$ $t_{13}$ $t_{14}$

1393

$t_{21}$ $t_{22}$ $t_{23}$ $t_{24}$ $t_{31}$ $t_{32}$ $t_{33}$ $t_{34}$\\

1394

1395

Note that the order in which multiple meshes are declared defines the

1396

order in which they are read and appended, but the geometric transformation

1397

is only applied at the end (this is a description file, not a command file).

1398

If multiple transformations are needed, a hierarchy of mesh meta-files

1399

may be used.

1400

1401

%==================================

1402

1306

\subsubsection{Meshing tools and associated formats}

1403

1307

%==================================

1404

1308

1420

1324

1421

1325

The use of files of the ``Common Solver'' type\footnote{File type specifically

1422

1326

developed for the early prototype versions of \CS (\texttt{tlc}) extension}

1423

is still possible but is deprecated. Such files are read directly from the

1424

Kernel, without the \pcs.

1425

The variable SOLCOM must

1426

be set to 1 in the launch scripts. Many potentialities of \CS are not

1327

is still in part possible but is deprecated. Such files are read directly

1328

from the Kernel, and this format is not handled by the launch

1329

scripts anymore. Many potentialities of \CS are not

1427

1330

usable with this file format (mesh joining with hanging nodes, periodicity,

1428

1331

parallel computing, ...).

1429

For all the other formats, the \pcs must be used (SOLCOM=0).

1430

1332

1431

1333

%==================================

1432

1334

\subsubsection{Meshing remarks}

1450

1352

\begin{list}{$\bullet$}{}

1451

1353

\item \texttt{--help}: gives a summary of the different command line options

1452

1354

1453

\item \texttt{-m mesh1 mesh2}: used to specify the names of the different meshes

1454

used. The launch script automatically calls the Preprocessor with the option

1455

\texttt{-m \$MESH}, where \texttt{MESH} is the variable where the user has

1456

specified the different meshes to be used.

1457

1458

\item \texttt{--join}: triggers the mesh joining functions. If nothing more is

1459

specified, every area of contact between two meshes will be joined together. The

1460

joining can be limited to certain selected faces. For instance, to join only

1461

the faces of colors 6 and 7, the full option will be \mbox{\texttt{--join --color

1462

6 7}}. These options are to be specified in the \texttt{COMMAND\_JOIN} variable in

1463

the launch script, to be automatically passed to the command line.

1464

1465

\item \texttt{--perio}: triggers periodic boundary conditions. Two types of

1466

periodic boundaries are possible: translation or rotation (see

1467

\S\ref{prg_paralperio} for additional details). For the translation, the basic

1468

option line is \texttt{--perio --trans tx ty tz}\\

1469

where \texttt{tx}, \texttt{ty} and \texttt{tz} are the coordinates of the

1470

translation vector. For the rotation, there are two possibilities. The

1471

transformation can be defined with a rotation angle (in degrees, between -180

1472

and 180), a direction and an

1473

invariant point\\

1474

\texttt{--perio --rota --angle a --dir dx dy dz --invpt px py pz}\\

1475

(with obvious notations), or by giving the rotation matrix and an invariant point\\

1476

\texttt{--perio --rota --matrix m11 m12 m13 m21 m22 m23 m31 m32 m33 --invpt px py

1477

pz}\\

1478

A rotation and a translation can be combined, by giving both \texttt{--rota} and

1479

\texttt{--trans} specifications. The translation will always be applied first,

1480

whatever the order in which the rotation and the translation have been given.\\

1481

The orientation of the transformations is not important since both the

1482

transformation and its inverse will be used to connect faces. Yet, when

1483

combining a translation and a rotation, the orientations given for both have to

1484

be consistent.\\

1485

It is possible (and usually recommended) to restrict the search for periodic

1486

connections between faces to a certain group of faces, by adding selection

1487

arguments like \texttt{--color}. It is also possible to specify up to 3

1488

independent periodicities, simply by repeating the \texttt{--perio} option. Below

1489

is given a example of the option line for a triple periodicity (the \verb+\+

1490

indicates the continuation of the command line):\\

1491

\texttt{--perio --trans -10.2 0 0 --color 2} \verb+\+\\

1492

\texttt{--perio --rota --angle 90 --dir 0 0 1 --invpt 0 0 0 --color 3 4} \verb+\+\\

1493

\texttt{--perio --trans 0 1 0 --rota --matrix 1 0 0 0 0 -1 0 1 0 --invpt 0 0 -0.2}\\

1494

This option is to be specified in the \texttt{COMMAND\_PERIO} variable in

1495

the launch script, to be automatically passed to the command line.

1355

\item \texttt{<mesh>}: the last argumen is used to specify the name of the mesh file.

1356

The launch script automatically calls the Preprocessor for every

1357

mesh in the \texttt{MESHES[]} list specified by th user.

1496

1358

1497

1359

\item \texttt{--reorient}: try to re-orient badly-oriented cells

1498

1360

if it is necessary to compensate for mesh-generation software

1508

1370

controlled by the launch script. The potential command line options are passed

1509

1371

through user modifiable variables at the beginning of the script. This way, the

1510

1372

user only has to fill these variables and doesn't need

1511

to search deep in the script for the Kernel command line. Yet, below is given

1512

the complete list of options, with the variables in which they should be

1513

specified in the script.

1373

to search deep in the script for the Kernel command line. For more advanced

1374

usage, the main options are described below:

1514

1375

1515

1376

\begin{list}{$\bullet$}{}

1516

1377

\item \texttt{--solcom}: this option indicates that the Kernel will read the

1517

1378

mesh directly, not using the Preprocessor output files. This is only possible

1518

1379

with ``Common Solver'' type of mesh files (see \S\ref{prg_maillages} for

1519

1380

restrictions).\\

1520

This option is triggered by the \texttt{SOLCOM} variable in the launch

1521

script. If \texttt{SOLCOM} is set to 1, the \texttt{-solcom} option is

1522

automatically added to the Kernel command line. The variable

1523

\texttt{ifoenv\index{ifoenv}} in the Fortran code will be set to 0 if the

1524

\texttt{--solcom} option has been used, otherwise it will be set to 1.

1381

This option is obsolete, and is not handled by the launch script anymore.

1382

1383

\item \texttt{--app-name}: specifies the application name. This is

1384

useful only in the case of code coupling, where the application name

1385

is used to distinguish between different code instances launched together.

1525

1386

1526

1387

\item \texttt{--mpi}: specifies that the calculation is running

1527

1388

with MPI communications. The number of processors used will be determined

1528

automatically by the Kernel. If necessary, the launch script automatically

1529

passes the \texttt{--mpi} option to the Kernel command line

1530

(see \ref{prg_runcase}).

1389

automatically by the Kernel. With most MPI implementations, the

1390

code will detect the presence of an MPI environment automatically, and

1391

this option is redundant. It is only kept for the rare case in which the

1392

MPI environment might not be detected.

1393

1394

\item \texttt{--mpi-io}: specifies that if MPI-IO should be used where

1395

available, and which mode should be used (\texttt{off} to disable,

1396

\texttt{eo} for explicit offsets, and \texttt{ip} for individual file

1397

pointers). MPI-IO is only available when running with MPI, and may improve

1398

performance only on parallel file systems. In other cases, it will incurr

1399

additional overhead.

1400

1401

\item \texttt{--preprocess}: triggers the preprocessing-only mode.

1402

The code may run without any Interface parameter file nor any user subroutine.

1403

Only the initial operations such as mesh joining and modification are

1404

executed.

1531

1405

1532

1406

\item \texttt{-q} or \texttt{--quality}: triggers the verification mode.

1533

The code runs without any Interface parameter file nor any user subroutine.

1534

The mesh is read and elementary tests are performed:\\

1407

The code may run without any Interface parameter file nor any user subroutine.

1408

This mode includes the preprocessing stages, and adds elementary tests:\\

1535

1409

\begin{list}{-}{}

1536

1410

\item the quality criteria of the mesh are calculated (non-orthogonality angles,

1537

1411

internal faces off-set, \ldots) and corresponding EnSight

1539

1413

\item test calculation of the gradient of $sin(x+2y+3z)$. The calculated

1540

1414

value is compared to the exact value, and an EnSight part for the

1541

1415

corresponding error is created. The gradient is calculated with each

1542

option IMRGRA from $0$ to $4$.\\

1416

option \texttt{imrgra} from $0$ to $4$.\\

1543

1417

\end{list}

1544

The command \texttt{-q} is to be placed in the \texttt{ARG\_CS\_VERIF}

1545

variable in the launch script to be added automatically to the Kernel

1546

command line.\\

1547

1548

\item \texttt{--cwf}: triggers the cutting of boundary and internal faces which

1549

have a warp angle larger than a certain limit\footnote{the warp angle is an

1550

indicator of the non-coplanarity of the different vertices of the face}. The

1551

concerned faces are divided into triangles. This option is to handle with care,

1552

since the division of the faces increases the non-orthogonalities of the mesh,

1553

but it is sometimes required (for the Lagrangian modeling, for instance, where

1554

non-plane faces lead to noticeable particle loss). By default, the faces are

1555

divided if their warp angle is larger than 0.01 degrees. This default value can

1556

be changed by adding an optional angle value to the command. For instance, to

1557

devide faces with a warp angle larger than 0.02 degrees, the full option will

1558

be \mbox{\texttt{-cwf 0.02}}.\\

1559

This option is to be specified in the \texttt{COMMAND\_CWF} variable in

1560

the launch script, to be automatically passed to the command line.

1561

1418

1562

1419

\item \texttt{--benchmark}: triggers the benchmark mode, for a timing

1563

1420

of elementary operations on the machine. A secondary option

1574

1431

\hspace*{0.5cm}\texttt{n=0}: output directed towards the standard output\\

1575

1432

\hspace*{0.5cm}\texttt{n=1}: output redirected towards a file \texttt{listing}

1576

1433

(default behaviour)\\

1577

This option can be specified in the \texttt{ARG\_CS\_OUTPUT} variable of the

1578

launch script.

1434

This option can be specified in the \texttt{domain.logging\_args} field

1435

of the user script.

1579

1436

1580

1437

\item \texttt{--logp n}: specifies the destination of the output for the

1581

1438

processors of rank 1 to $N-1$ in a calculation in parallel on $N$ processors

1588

1445

\hspace*{0.5cm}\texttt{n=1}: one file for the output of each processor. The

1589

1446

output of the processors of rank 1 to $N-1$ are directed to the files

1590

1447

\texttt{listing\_n0002} to \texttt{listing\_n$N$}.

1591

This option can be specified in the \texttt{ARG\_CS\_OUTPUT} variable of the

1592

launch script.

1448

This option can be specified in the \texttt{domain.logging\_args} field

1449

of the user script.

1593

1450

1594

1451

\item \texttt{-p xxx} or \texttt{--param xxx}: specifies the name of the GUI

1595

1452

parameter file to use for the calculation.\\

1596

The value of \texttt{xxx} is to be placed in the \texttt{PARAM} variable in the launch

1597

script (the file will be looked for in the directory \texttt{DATA}).

1598

The option \mbox{\texttt{-param \$PARAM}} is automatically added to the

1599

Kernel command line.

1453

The value of \texttt{xxx} is to be defined by the \texttt{--param} option

1454

of \texttt{code\_saturne run}, either directly or in the standard \texttt{runcase}

1455

script (the file will be searched for in the \texttt{data} directory, though

1456

an absolute path name may also be defined).

1600

1457

1601

1458

\item \texttt{-h} or \texttt{--help}: to display a summary of the different

1602

1459

command line options.

1603

1460

\end{list}

1604

1461

1605

1462

%==================================

1606

\subsection{Parameters in the launch script}

1463

\subsection{Launch scripts}

1607

1464

%==================================

1608

1465

\label{prg_runcase}%

1609

1466

1610

1467

The case preparer command \texttt{code\_saturne~create} places an example of launch script,

1611

\texttt{runcase}, in the \texttt{SCRIPTS} directory. This script is quite general

1612

and known to work on every architecture \CS has been tested on. The beginning

1613

if the script contains the definition of certain parameters (environment

1614

variables) necessary to set the calculation. The second part of the script

1615

contains the commands for the preparation and execution of the calculation. No

1616

user intervention should be necessary in this second part.\\

1617

The Graphical User Interface allows to fill in the major

1618

parameters of the script without having to edit the file.

1619

1620

Below is a list of the different variables and parameters that might be modified

1621

for a calculation, in their order of apparition in the script:

1622

\begin{list}{$\bullet$}{}

1623

\item LSF headers: definition of the headers for an LSF batch system, as can be

1624

found on the machines of the CCRT (Platine). The data expected are

1625

the number of processors reserved (\texttt{\#BSUB -n}), the CPU time limit

1626

(\texttt{\#BSUB -W}), the name of the standard output file (\texttt{\#BSUB -o}),

1627

the name of the standard error file (\texttt{\#BSUB -e}) and the name of

1628

the job (\texttt{\#BSUB -J}).

1629

1630

\item PBS headers: definition of the headers for a PBS batch system, as can be

1631

found on the machines of the Chatou cluster. The data expected are the number of

1632

nodes reserved (\texttt{nodes}), the number of processors per node

1633

(\texttt{ppn}), the total CPU time (\texttt{walltime}), the memory reserved

1634

(\texttt{mem}), and the name of the job (\texttt{\#PBS -N}).

1635

1636

\item Manchester headers: definition of the headers for the batch system

1637

specific to the cluster of the University of Manchester

1638

1639

\item \texttt{SOLCOM}: a value of 1 will pass the \texttt{-solcom} option to the

1640

Kernel (see \ref{prg_optappelnoy})

1641

\end{list}

1642

1643

\begin{list}{$\bullet$}{}

1644

\item \texttt{STUDY}: name of the study directory (automatically set by

1645

\texttt{code\_saturne~create}, see \S\ref{prg_architecture})

1646

1647

\item \texttt{CASE}: name of the case directory (automatically set by

1648

\texttt{code\_saturne~create}, see \S\ref{prg_architecture})

1649

1650

\item \texttt{PARAM}: name of the Interface parameter file, if necessary (see

1651

\ref{prg_optappelnoy})

1652

1653

\item \texttt{MESH}: name(s) of the mesh(es) used for the calculation (see

1654

\ref{prg_optappelecs} and \ref{prg_maillages}). The files will be looked for in

1655

the directory \texttt{MESHDIR} (see below).

1656

1657

\item \texttt{COMMAND\_JOIN}: Preprocessor command line option for mesh joining (see

1658

\ref{prg_optappelecs})

1659

1660

\item \texttt{COMMAND\_CWF}: Kernel command line option for the division of

1661

faces with too large a warp angle (see \ref{prg_optappelnoy})

1662

1663

\item \texttt{COMMAND\_PERIO}: Preprocessor command line option for the definition

1664

of periodic boundary conditions (see \ref{prg_optappelecs})

1665

1666

\item \texttt{THERMOCHEMISTRY\_DATA}: name of the thermochemical data file, if

1667

necessary (the file is looked for in the directory \texttt{DATA}, see

1668

\S\ref{prg_usppmo})

1669

1670

\item \texttt{NUMBER\_OF\_PROCESSORS}: number of processors (potentially virtual)

1671

to be used for the calculation.\\

1672

If the variable is left empty, the launch script

1673

will fill it automatically: on a batch system, \texttt{NUMBER\_OF\_PROCESSORS}

1674

will be equal to the number of processors reserved; in case of an

1675

interactive calculation, it will be set to 1.\\

1676

When using a batch system,

1677

\texttt{NUMBER\_OF\_PROCESSORS} should ideally be equal to the number of

1678

processors reserved, and can never be larger (one executable per

1679

processor). With an interactive calculation (like a Linux PC),

1680

\texttt{NUMBER\_OF\_PROCESSORS} can be larger than the total number of processors

1681

available, although it is not recommended (loss of efficiency because several

1682

executables share the same processor).\\

1683

In case of

1684

coupling with \syrthes, one processor is reserved for \syrthes, and the Kernel

1685

of \CS will therefore automatically be set to run on

1686

\texttt{NUMBER\_OF\_PROCESSORS-1} processors.

1687

1688

\item \texttt{PROCESSOR\_LIST}: list of nodes on which the calculation is to

1689

be run. On batch systems, this list is set automatically by the batch

1690

manager. For calculations on a stand-alone machine, the list is not used. Hence,

1691

except for very specific test (mainly for developing purposes), it is

1692

recommended to leave this variable empty.

1693

1694

\item \texttt{USER\_INPUT\_FILES}: list of the user data files to be

1695

copied in the temporary execution directory before the calculation (input

1696

profiles for instance). The files will

1697

be looked for in the directory \texttt{DATA}. The thermochemical data files,

1698

Interface parameter file and calculation restart files are specified in other

1699

variables and do not need to appear here. When using the vortex method for LES

1700

entry conditions, the corresponding data files have to be specified in

1701

\texttt{USER\_INPUT\_FILES} (see \S\ref{prg_usvort})

1702

1703

\item \texttt{USER\_OUTPUT\_FILES}: list of user result files to be

1704

copied in the directory \texttt{RESU} at the end of the calculation. A directory

1705

\texttt{RES\_USERS.mmddhhmm} will be created in the directory \texttt{RESU} and

1706

all the files will be stored in it. The files automatically created by the code

1707

(listings, post-processing, automatic chronological records\footnote{when using

1708

\texttt{ushist} for user-defined chronological records, the files created need

1709

to be specified in \texttt{USER\_OUTPUT\_FILES}},

1710

restart files) do not need to be specified in

1711

\texttt{USER\_OUTPUT\_FILES}.

1712

1713

\item \texttt{CS\_TMP\_PREFIX}: alternate temporary directory for the

1714

calculation (see \S\ref{prg_temporarydirectory})

1715

1716

\item \texttt{OPTIMIZATION}: optimisation level for compilation

1717

(LO, DBG, EF or PROF; see \S\ref{prg_library}). This optimisation

1718

level will be applied to all the modules of \CS (BASE, CFBL, COGZ, CPLV, ELEC,

1719

FUEL, LAGR, MATI, RAYT).

1720

Leaving the variable empty stands for ``standard''

1721

optimisation.

1722

1723

\item \texttt{CS\_LIB\_ADD}: additional commands for the link stage of the

1724

compilation. This can be especially useful if the user subroutines call routines

1725

provided by external libraries. To link with an external library ``foo'', the

1726

variable would be for instance\\

1727

\texttt{CS\_LIB\_ADD=``-L/opt/foo/lib -lfoo''}

1728

1729

\item \texttt{VALGRIND}: command to be placed before the \CS executable name on

1730

the execution command line ({\em i.e.} the launch script will execute the

1731

command \texttt{\$VALGRIND cs\_solver ...}). It is especially designed to use the

1732

valgrind debugging and profiling tool. The usual value to use valgrind is

1733

\texttt{VALGRIND=``valgrind --tool=memcheck''}

1734

1735

\item \texttt{ARG\_CS\_VERIF}: verification mode to be used for \CS (see

1736

\ref{prg_optappelnoy}). An empty variable implies standard calculation mode

1737

(IVERIF=0).

1738

1739

\item \texttt{ARG\_CS\_OUTPUT}: options for the redirection of the standard

1740

output (see \ref{prg_optappelnoy})

1741

1742

\item \texttt{ECHOCOMM}: level for the \texttt{--echo-comm} option of the Kernel

1743

command line (see \ref{prg_optappelnoy})

1744

1745

\item \texttt{ADAPTATION}: commands to trigger the automatic mesh

1746

adaptation with the software Homard. This option is still under development and

1747

restricted to developpers use.

1748

1749

\item \texttt{CASEDIR}: root directory of the calculation. This variable is

1750

automatically set by \texttt{code\_saturne~create} and should not be changed.

1751

1752

\item \texttt{DATA}: DATA directory of the case (see \ref{prg_architecture}).

1753

This variable is automatically set by \texttt{code\_saturne~create} and should not be

1754

changed.

1755

1756

\item \texttt{RESU}: RESU directory of the case (see \ref{prg_architecture}).

1757

This variable is automatically set by \texttt{code\_saturne~create} and should not be

1758

changed.

1759

1760

\item \texttt{SRC}: SRC directory of the case (see \ref{prg_architecture}).

1761

This variable is automatically set by \texttt{code\_saturne~create} and should not be

1762

changed.

1763

1764

\item \texttt{SCRIPTS}: SCRIPTS directory of the case (see

1765

\ref{prg_architecture}). This variable is automatically set by

1766

\texttt{code\_saturne~create} and should not be changed.

1767

1768

\item \texttt{RESTART\_IN}: directory containing the files for calculation

1769

restart.

1770

1771

\item \texttt{PREPROCESSOR\_OUTPUT\_IN}: \texttt{preprocessor\_ouput} file for a calculation in ``calculation'' mode (see \ref{prg_executionmodes})

1772

1773

\item \texttt{MESHDIR}: directory containing the mesh files (see

1774

\ref{prg_architecture}). This variable is automatically set by

1775

\texttt{code\_saturne~create} and should generally not be changed.

1776

1777

\item \texttt{DATA\_SYR}: directory for the \syrthes data. This directory has

1778

to be created by the user. It is advised to keep the location proposed

1779

in the launch script, which complies with the standard architecture

1780

of \CS (see

1781

\ref{prg_architecture}).

1782

1783

\item \texttt{SYRTHES\_ENV}: name of the environment file for \syrthes (usually

1784

\texttt{syrthes.env}, as proposed in the launch script).

1785

1786

\item \texttt{SRC\_SYR}: directory for the \syrthes user subroutines.

1787

This directory has to be created by the user. It is advised to keep the location

1788

proposed in the launch script, which complies with the standard architecture

1789

of \CS (see \ref{prg_architecture}).

1790

1791

\item \texttt{COUPLING\_MODE}: coupling mode between \CS and \syrthes 3.4, when

1792

such coupling is activated (see \texttt{COMMAND\_SYRTHES}). Two options are

1793

available:\\

1794

\hspace*{0.5cm}\texttt{MPI}: for a coupling based on MPI messages (requires a

1795

MPI library)\\

1796

\hspace*{0.5cm}\texttt{sockets}: for a coupling based on sockets

1797

1798

\item \texttt{EXEC\_PREPROCESSOR}: execution mode for \CS preprocessor (see \ref{prg_executionmodes})

1799

1800

\item \texttt{EXEC\_PARTITION}: execution mode for \CS partitionner (see \ref{prg_executionmodes})

1801

1802

\item \texttt{EXEC\_KERNEL}: execution mode for \CS kernel (see \ref{prg_executionmodes})

1803

1804

\end{list}

1805

1468

\texttt{runcase}, in the \texttt{SCRIPTS} directory. This script is quite minimalist and is known to work on every architecture \CS has been tested on.

1469

If a batch system is available, this script will contain options

1470

for batch submission.

1471

The script will then contain a line setting the proper \texttt{PYTHONPATH}

1472

variable for \CS to run.

1473

Finally, it simply contains the \texttt{code\_saturne run} command,

1474

possible with a \texttt{--param} option when a parameters file

1475

defined by the GUI is used. Other options recognized by

1476

\texttt{code\_saturne run} may be added.

1477

1478

In the case of a coupled calculation, this script also exists, and

1479

may be used for preprocessing stages, but an additional

1480

\texttt{runcase\_coupling} is added in the directory above the coupled case

1481

directories, and may be used to define the list of coupled cases,

1482

as well as global options, such as MPI options ot the temporary

1483

execution directory. An additional \texttt{runcase\_batch} file will

1484

contain batch submission options when a batch system is available

1485

(and is the file that should be submitted when using a batch system).

1486

1487

When not using the GUI, or if additional options need to be accessed,

1488

the \texttt{cs\_user\_scripts.py} file may be copied from

1489

the \texttt{DATA/REFERENCE} to the \texttt{DATA} and edited.

1490

This file contains several Python functions:

1491

1492

\begin{list}{$\bullet$}{}

1493

1494

\item \texttt{define\_domain\_parameter\_file} allows defining

1495

the choice of a parameters file produced by the GUI. This is

1496

generally not useful, as the parameters file may be directly

1497

defined in \texttt{runcase} or \texttt{runcase\_coupling}, or passed

1498

as an option to \texttt{code\_saturne run}, but could be useful

1499

when running more complex parametric scripts, and is provided for

1500

the sake of completeness.

1501

\item \texttt{define\_domain\_parameters} allows defining

1502

most paramters relative to case execution for the current

1503

domain, including advanced options not accessible

1504

through the GUI. This function is the most important one in the user

1505

scripts file, and contains descriptions of the various options.

1506

Note that in most examples, setting of options is preceded by

1507

a \texttt{if domain.param == None:} line, ensuring the settings

1508

are only active if no GUI-defined parameters file is present.

1509

This is used to prevent accidental override of parameters defined

1510

by the GUI: parameters defined through the user script have priority

1511

over the GUI parameters file, so if both are used, these tests

1512

may be removed for parameters which should be defined through

1513

user scripts.

1514

\item \texttt{define\_case\_parameters} allows defining

1515

most paramters relative to the global calculation, such as

1516

the number of processors or the execution directory.

1517

To avoid potentially conflicting definitions, this function is ignored

1518

for coupled calculations, where the corresponding parameters

1519

may be defined in the \texttt{runcase\_coupling} script.

1520

\item \texttt{define\_mpi\_environment} allows defining

1521

advanced MPI parameters or redefining MPI options if the automatic

1522

settings are incorrect, and ts use should only rarely be necessary.

1523

To avoid potentially conflicting definitions, this function is ignored

1524

for coupled calculations, where the corresponding parameters

1525

may be defined in the \texttt{runcase\_coupling} script.

1526

\end{list}

1806

1527

1807

1528

%==================================

1808

1529

\subsection{Graphical User Interface}

1809

1530

%==================================

1810

\label{prg_ihm}%

1531

\label{prg_gui}%

1811

1532

A Graphical User Interface is available with \CS.

1812

1533

This Interface creates or reads an XML file according to

1813

1534

a specific \CS syntax which is then interpreted by the code.

1856

1577

followings:

1857

1578

1858

1579

\begin{list}{-}{}

1859

\item Analysis environment: definition of the calculation directories

1860

(STUDY, CASE), mesh file(s), periodicity, coupling with \syrthes,

1861

stand-alone execution of the

1862

Preprocessor module (used by the Interface to get the colors of the boundary

1580

\item Identity and paths: definition of the calculation directories

1581

(STUDY, CASE, DATA, SRC, SCRIPTS, MESH).

1582

1583

\item Calculation environment: definition of the mesh file(s),

1584

stand-alone execution of the Preprocessor module

1585

(used by the Interface to get the groups of the boundary

1863

1586

faces).

1864

1587

1865

1588

\item Thermophysical models: physical model, ALE mobile mesh features,

1866

turbulence model, thermal model,

1867

initialisation of the variables.

1589

turbulence model, thermal model, coupling with \syrthes.

1590

1591

\item Additional scalars: definition, initialisation of the scalars,

1592

and physical characteristics.

1868

1593

1869

1594

\item Physical properties: reference pressure, fluid characteristics, gravity.

1595

It is also possible to write user laws for the density, the viscosity,

1596

the specific heat and the thermal conductivity in the interface through

1597

the use of a formulae interpreter.

1870

1598

1871

\item Additional scalars: definition, physical characteristics and

1872

initialisation of the scalars (apart from the temperature, which

1873

is treated separately in the Interface).

1599

\item Volume conditions: initialization of the variables, and definition of

1600

the zones where to apply head loss.

1874

1601

1875

1602

\item Boundary conditions: definition of the boundary conditions for

1876

1603

each variable. The colors of the boundary faces may be read

1877

directly from a ``listing'' file created by the Preprocessor. This

1878

file can be generated directly by the Interface under the heading\\

1879

``Analysis environment $\rightarrow$ Solution Domain $\rightarrow$

1880

Stand-alone running''.

1881

1882

\item Analysis control: parameters concerning the time averages, time step,

1604

directly from a ``listing'' file created by the Preprocessor.

1605

1606

\item Numerical parameters: number and type of time step, advanced parameters

1607

for the numerical solution of the equations.

1608

1609

\item Calculation control: parameters concerning the time averages, time step,

1883

1610

location of

1884

1611

the probes where some variables will be monitored over time,

1885

1612

definition of the frequency of the outputs in the calculation

1886

1613

listing and in the chronological records and of the EnSight outputs.

1887

1888

\item Numerical parameters: advanced parameters for the numerical solution of

1889

the equations

1614

The item {\itshape Profiles} allows to save, with a given frequency,

1615

1D profiles on an axis defined from two points provided by the user.

1890

1616

1891

1617

\item Calculation management: management of the calculation restarts,

1892

1618

updating of the launch script (temporary execution directory, parallel

1893

computing, user data or result files, ...), interactive launch of the

1894

calculation and user arrays definition.

1619

computing, user data or result files, ...) and interactive launch of the

1620

calculation.

1895

1621

1896

1622

\end{list}

1897

1623

1924

1650

out, thus saving the user the tedious task of uncommenting all the lines (and

1925

1651

the risk of skipping some of them).

1926

1652

1927

1928

%==================================

1929

\subsection{Face and cell mesh-defined properties and selection}

1930

%==================================

1931

\label{selection_criteria}

1932

1933

The mesh entities may be referenced by the user during the mesh

1934

creation. These references may then be used to mark out some mesh entities

1935

according to the need (specification of boundary conditions, pressure

1936

drop zones, ...). The references are generally of one of the two

1937

following types:

1938

\begin{list}{$\bullet$}{}

1939

\item color.

1940

A color is an integer possibly associated with boundary faces and

1941

volume elements by the mesh generator. Depending on the tool,

1942

this concept may have different names, which \CS interprets

1943

as colors. Most tools allow only one color per face or element.

1944

\begin{list}{-}{}

1945

\item I-deas uses a color number with a default of

1946

7 (green) for elements, be they volume elements or boundary

1947

``surface coating'' elements. Color 11 (red) is used for

1948

for vertices, but vertex properties are ignored by \CS.

1949

\item SIMAIL uses the equivalent notions of ``reference''

1950

for element faces, and ``subdomain'' for volume elements.

1951

By default, element faces are assigned no reference (0),

1952

and volume elements domain 1.

1953

\item Gmsh uses ``physical property'' numbers.

1954

\item EnSight has no similar notion, but if several parts

1955

are present in an EnSight 6 file, or several parts

1956

are present \emph{and} vertex ids are given in an

1957

Ensight Gold file, the part number is interpreted as

1958

a color number by the Preprocessor.

1959

\item The Comet Design (pro-STAR/STAR4) and NUMECA Hex file

1960

formats have a CAD section id that is interpreted

1961

as a color number. In the latter case, this notion

1962

only applies to faces, so volume elements are given

1963

color.

1964

\item The MED format allow integer ``attributes'', though

1965

many tools working with this format ignore those

1966

and only handle groups.

1967

\end{list}

1968

\item groups.

1969

Named ``groups'' of mesh entities may also be used with many

1970

mesh generators or formats. In some cases, a given cell or face may belong

1971

to multiple groups (as some tools allow new groups to be defined

1972

by boolean operations on existing groups).

1973

In \CS, every group is assigned a group number (base on alphabetical

1974

ordering of groups).

1975

\begin{list}{-}{}

1976

\item I-deas assigns a group number with each

1977

group, but by default, this number is just a counter.

1978

Only the group name is considered by \CS (so that elements

1979

belonging to two groups with identical names and different

1980

numbers are considered as belonging to the same group).

1981

\item CGNS allows both for named boundary conditions and mesh

1982

sections. If present, boundary condition names are

1983

interpreted as group names, and groups may also be defined

1984

based on element section or zone names using additional

1985

Preprocessor options (\texttt{-grp-cel} or

1986

\texttt{-grp-fac} followed by \texttt{section} or

1987

\texttt{zone}).

1988

\item Using the MED format, it is preferable to use ``groups''

1989

to colors, as many tools ignore the latter.

1990

\end{list}

1991

\end{list}

1992

1993

Selection criteria may be defined in a similar fashion whether

1994

using the GUI or in user subroutines.

1995

Typically, a selection criteria is simply a string containing

1996

the required color numbers or group names, possibly combined

1997

using boolean expressions. Simple geometric criteria are also

1998

possible.

1999

2000

A few examples are given below:

2001

2002

\verb+ENTRY+\\

2003

\verb+1 or 7+\\

2004

\verb+all[]+\\

2005

\verb+3.1 >= z >= -2 or not (15 or entry)+\\

2006

\verb+range[04, 13, attribute]+\\

2007

\verb+sphere[0, 0, 0, 2] and (not no_group[])+

2008

2009

Strings such as group names containing whitespace

2010

or having names similar to reserved operators may be protected

2011

using ``escape characters''.\footnote{Note that for defining a

2012

string in Fortran, double quotes are easier to use, as they do not

2013

conflict with Fortran's single quotes delimiting a string.

2014

In C, the converse is true. Also, in C, to define a string

2015

such as \texttt{{$\backslash$}plane}, the string

2016

\texttt{{$\backslash$}{$\backslash$}plane} must be

2017

used, as the first $\backslash$ character is used by the

2018

compiler itself. Using the GUI, either notation is easy.}

2019

More complex examples of strings whith protected strings are given here:

2020

2021

\verb+"First entry" or Wall\ or\ sym+\\

2022

\verb+entry or \plane or "noone's output"+

2023

2024

The following operators and syntaxes are allowed (fully capitalized

2025

versions of keywords are also allowed, but mixed capitals/lowercase

2026

versions are not):

2027

2028

\begin{tabular}[top]{p{6cm} l}

2029

\multicolumn{2}{l}{\bf escape characters }\\

2030

protect next character only: & \texttt{$\backslash$} \\

2031

protect string: & \texttt{{'}$string${'}} \quad \texttt{"$string$"}\\

2032

\end{tabular}

2033

2034

\begin{tabular}[top]{p{6cm} l}

2035

\multicolumn{2}{l}{\bf basic operators }\\

2036

priority: & \texttt{(} \quad \texttt{)} \\

2037

not: & \texttt{not} \quad \texttt{!} \quad \texttt{!=} \\

2038

and: & \texttt{and} \quad \texttt{\&} \quad \texttt{\&\&} \\

2039

or: & \texttt{or} \quad \texttt{|} \quad \texttt{||} \quad \texttt{,} \quad \texttt{;} \\

2040

xor: & \texttt{xor} \quad \texttt{\^} \\

2041

\end{tabular}

2042

2043

\begin{tabular}[top]{p{6cm} l}

2044

\multicolumn{2}{l}{\bf general functions }\\

2045

select all: & \texttt{all[]}\\

2046

entities having no group or color: & \texttt{no\_group[]} \\

2047

select a range of groups or colors: & \texttt{range[$first$, $last$]} \\

2048

& \texttt{range[$first$, $last$, group]} \\

2049

& \texttt{range[$first$, $last$, attribute]} \\

2050

\end{tabular}

2051

2052

For the range operator, $first$ and $last$ values are inclusive.

2053

For attribute (color) numbers, natural integer value ordering is used,

2054

while for group names, alphabetical ordering is used. Note also that in

2055

the bizarre (not recommended) case in which a mesh would contain for

2056

example both a color number 15 and a group named ``15'', using

2057

\texttt{range[15, 15, group]} or \texttt{range[15, 15, attribute]}

2058

could be used to distinguish the two.

2059

2060

Geometric functions are also available. The coordinates considered are

2061

those of the cell or face centers. Normals are of course

2062

usable only for face selections, not cell selections.

2063

2064

\begin{tabular}[top]{p{6cm} l}

2065

\multicolumn{2}{l}{\bf geometric functions }\\

2066

face normals: & \texttt{normal[$x$, $y$, $z$, $epsilon$]} \\

2067

& \texttt{normal[$x$, $y$, $z$, epsilon = $epsilon$]} \\

2068

plane, $ax + by + cz + d = 0$ form: & \texttt{plane[$a$, $b$, $c$, $d$, $epsilon$]} \\

2069

& \texttt{plane[$a$, $b$, $c$, $d$, epsilon = $epsilon$]} \\

2070

& \texttt{plane[$a$, $b$, $c$, $d$, inside]} \\

2071

& \texttt{plane[$a$, $b$, $c$, $d$, outside]} \\

2072

plane, normal + point in plane form: & \texttt{plane[$n_x$, $n_y$, $n_z$, $x$, $y$, $z$, $epsilon$]} \\

2073

& \texttt{plane[$n_x$, $n_y$, $n_z$, $x$, $y$, $z$, epsilon = $epsilon$]} \\

2074

& \texttt{plane[$n_x$, $n_y$, $n_z$, $x$, $y$, $z$, inside]} \\

2075

& \texttt{plane[$n_x$, $n_y$, $n_z$, $x$, $y$, $z$, outside]} \\

2076

box, extents form: & \texttt{box[$x_{min}$, $y_{min}$, $z_{min}$,

2077

$x_{max}$, $y_{max}$, $z_{max}$]} \\

2078

box, origin + axes form: & \texttt{box[$x_0$, $y_0$, $z_0$,}\\

2079

& \texttt{\qquad $dx_1$, $dy_1$, $dz_1$,

2080

$dx_2$, $dy_2$, $dz_2$,

2081

$dx_3$, $dy_3$, $dz_3$]} \\

2082

& \texttt{plane[$a$, $b$, $c$, $d$, epsilon = $epsilon$]} \\

2083

& \texttt{plane[$a$, $b$, $c$, $d$, inside]} \\

2084

& \texttt{plane[$a$, $b$, $c$, $d$, outside]} \\

2085

cylinder: & \texttt{cylinder[$x_0$, $y_0$, $z_0$, $x_1$, $y_1$, $z_1$, $radius$]} \\

2086

sphere: & \texttt{sphere[$x_c$, $y_c$, $z_c$, $radius$]} \\

2087

inequalities: & \texttt{>}, \texttt{<}, \texttt{>=}, \texttt{<=} associated

2088

with \texttt{x}, \texttt{y}, \texttt{z} or

2089

\texttt{X}, \texttt{Y}, \texttt{Z} keywords\\

2090

& and coordinate value; \\

2091

& \texttt{$x_{min}$ <= x < $x_{max}$} type syntax is allowed. \\

2092

\end{tabular}

2093

2094

In the current version of \CS, all selection criteria used

2095

are maintained in a list, so that re-interpreting a criterion already

2096

encountered (such as at the previous time step) is avoided.

2097

Lists of entities corresponding to a criteria containing no geometric

2098

functions are also saved in a compact manner, so re-using a previously

2099

used selection should be very fast. For criteria containing geometric

2100

functions, the full list of corresponding entities is not maintained,

2101

so each entity must be compared to the criterion at each time step.

2102

Heavy use of many selection criteria containing geometric functions

2103

may thus lead to reduced performance.

2104

2105

%==================================

2106

\section{Preprocessing}

2107

%==================================

2108

2109

The \pcs module of \CS reads the

2110

mesh file(s) (under any supported format) and translates the necessary

2111

information into a Kernel input file. Mesh joining and domain partitioning for parallel

2112

calculations are done at this stage. In case of periodic boundary

2113

conditions, the \pcs module also identifies the boundary faces that are

2114

related through periodicity and generates the corresponding connectivity.

2115

2116

The executable of the \pcs module is \texttt{cs\_preprocess},

2117

and the most useful options and sub-options are described briefly here.

2118

To obtain a complete and up-to-date list of options and environment

2119

variables, use the following command:

2120

\texttt{cs\_preprocess~-h} or \texttt{cs\_preprocess~--help}. Many options,

2121

such as this one, accept a short and a long version.

2122

2123

For the main options, the launch script \texttt{runcase} contains

2124

corresponding variables, that are used to define options for the

2125

\pcs. This way, the user only has to define these variables

2126

and does not detailed knowledge of the \pcs command line.

2127

2128

Nonetheless, it may be useful to call the \pcs manually

2129

in certain situations, especially for frequent verification when

2130

building a mesh, so its use is described here. Verification

2131

may also be done using the \texttt{code\_saturne~check\_mesh} command,

2132

which takes a subset of the same command-line arguments.

2133

2134

The \pcs is controlled using command-line arguments, some of these

2135

accepting sub-arguments. It is possible to complete the command-line

2136

with a file, using the same syntax as the command-line, but also

2137

allowing multiple lines and comments. A few environment variables

2138

allow an expert user to modify some behaviors or to obtain

2139

a trace of memory management.

2140

2141

%==================================

2142

\subsection{\pcs options and sub-options\label{sec:optpcs_ssopt}}

2143

%==================================

2144

2145

Main choices are done using command-line options. Sub-option arguments

2146

must directly follow the one activating the corresponding option, and

2147

the definition of sub-options ends with the first argument which does

2148

not match the same sub-option. Some options may be applied multiple

2149

times. For example, \texttt{-j} being the option which activates

2150

face joining, \texttt{--fraction} the sub-option modifying the

2151

default tolerance, and \texttt{--color} a face selection sub-option:

2152

2153

\texttt{cs\_preprocess -m fluid.msh -j --fraction 0.2 --color 2 --color 3 -j}

2154

2155

\noindent means that we apply a joining to boundary faces of

2156

color $2$ or $3$ of mesh \texttt{fluid.msh} with a tolerance of $0.2$,

2157

then that we apply a second joining to all remaining boundary faces

2158

(with default parameters), while:

2159

2160

\texttt{cs\_preprocess -m fluid.msh -j --fraction 0.2 -color 2 -j --color 3}

2161

2162

\noindent means that we apply a joining to boundary faces of

2163

color $2$ of mesh \texttt{fluid.msh} with a tolerance of $0.2$,

2164

then that we apply a second joining to boundary faces of color $3$

2165

(with default values for other parameters).

2166

2167

\subsubsection{Option files\label{sec:optpcs:input}}

2168

2169

It is possible to complete the command-line with a file, using the same

2170

syntax as the command-line, but allowing the addition of arbitrary line

2171

jumps and comments. On a given line, anything that follows a\texttt{\#}

2172

character until the end of the line is ignored in such a file.

2173

The use of such a file is activated by the argument \texttt{-i}.

2174

If for example a file \texttt{pp\_cmd} contains the following

2175

lines:

2176

2177

\noindent

2178

\begin{tabular}{l l}

2179

\texttt{--ensight} & \texttt{\# activate EnSight output}\\

2180

\texttt{--med} & \texttt{\# activate MED output}\\

2181

\end{tabular}

2182

2183

\noindent

2184

\begin{tabular}{l l}

2185

then command: & \texttt{cs\_preprocess -i pp\_cmd -m mesh.unv}\\

2186

is equivalent to: & \texttt{cs\_preprocess -m mesh.unv --ensight --med}\\

2187

\end{tabular}

2188

2189

The use of option files may circumvent command-line length limits

2190

(now rarely an issue), and allows easier manipulation and better

2191

readability of complex option combinations.

2192

2193

\subsubsection{Mesh selection\label{sec:optpcs:mesh}}

2194

2195

Any use of the preprocessor requires one or several meshes (except for

2196

\texttt{cs\_preprocess} and \texttt{cs\_preprocess~-h} which respectively

2197

print the version number and list of options).

2198

They are selected using the \texttt{--mesh} or \texttt{-m} option, followed

2199

by th list of meshes to read. The file format is usually automatically

2200

determined based on its extension (c.f. \ref{sec:formats_in}

2201

page~\pageref{sec:formats_in}) but this option also allows a \texttt{--format}

2202

sub-option, to force the format choice of selected files.

2203

2204

For formats allowing multiple meshes in a single file, the

2205

\texttt{--num} sub-option followed by a strictly positive integer allows

2206

selection of a specific mesh; by default, the first mesh is selected.

2207

2208

For meshes in CGNS format, we may in addition use the \texttt{--grp-cel}

2209

or \texttt{--grp-fac} sub-options, followed by the \texttt{section}

2210

or \texttt{zone} keywords, to define additional groups of cell or faces

2211

based on the organization of the mesh in sections or zones. The sub-options

2212

have no effect on meshes of other formats.

2213

2214

We may define as many mesh selections as we wish. Meshes are read and

2215

automatically compounded (but not joined) in the order of their appearance.

2216

For example:

2217

2218

\noindent

2219

\texttt{cs\_preprocess -m file.1 file.2 --format ideas -m branch.cgns --grp-cel section}

2220

2221

\noindent

2222

reads mesh files \texttt{file.1} and \texttt{file.2} as \ideas files,

2223

and appends the first mesh from the \texttt{branch.cgns} file,

2224

on which additional groups of cells corresponding to the CGNS section

2225

names to which cells belong.

2226

2227

\subsubsection{Post-processing output\label{sec:optpcs:post}}

2228

2229

By default, the \pcs does not generate any post-processor output.

2230

By adding \texttt{--ensight}, \texttt{--med}, or \texttt{--cgns} to

2231

the command-line arguments, the output of the mesh to the indicated format

2232

is provoked. It is possible to generate output to different formats

2233

simultaneously. Note that we will obtain additional surface meshes

2234

corresponding to the calculation domain boundary, to faces joined

2235

or modified by joinings or periodicity, etc. Note also that

2236

periodic faces are not considered to be part of the domain boundary.

2237

\footnote{Periodicity is interpreted as a ``geometric'' condition

2238

rather than a classical boundary condition.}

2239

2240

We consider 4 types of output: the volume mesh (output before possible

2241

joinings), the boundary mesh, informational meshes (such as faces

2242

stemming from or modified by a joining of periodicity, or

2243

selected interior faces), and meshes outlining to zones with errors.

2244

By default, all types of meshes are output. If one or several of

2245

the three filtering sub-options is used , only the meshes of the

2246

requested types will be output, as well as meshes corresponding to

2247

zones with errors (non-filterable). The filtering sub-options

2248

are: \texttt{--volume}, \texttt{--boundary}, and \texttt{--info}.

2249

2250

One may also optionally avoid outputting polygons or polyhedra

2251

by adding the \texttt{--discard-poly} sub-option.

2252

This allow working around bugs or limited polyhedron support

2253

in visualization tools or limitations of the CGNS 2 format.

2254

In this case, the post-processing output will be incomplete,

2255

but at least readable.

2256

2257

In the case of \emph{EnSight Gold} output, we may use the

2258

\texttt{--simple} sub-option to avoid splitting the output in

2259

several \emph{parts} corresponding to different attributes

2260

(colors and groups) borne by this mesh.

2261

2262

Still in the case of the \emph{EnSight Gold} output, we may

2263

force the output in text mode using the \texttt{--text} sub-option,

2264

or force binary output to ``big-endian'' variant using the

2265

\texttt{--big-endian} sub-option. By default, the output is native

2266

binary.

2267

2268

Finally, in the case of output in the \emph{MED} format, we may

2269

force the conversion of attributes (colors) to groups using

2270

the \texttt{--color-to-group} sub-option.

2271

2272

\subsubsection{Faces selection\label{sec:optpcs:select}}

2273

2274

Selection of faces occurs at several levels, especially for conforming

2275

joining.

2276

2277

It is also possible to select a set of interior faces bearing attributes

2278

(colors or groups) for verification purposes using the option

2279

\texttt{--int-face}. The \pcs prints the number of interior faces

2280

matching a selection criterion, and if in addition post-processing

2281

output is activated, a surface mesh corresponding to the selected faces

2282

will be generated.

2283

2284

\pcs selection criteria are more limited than those

2285

used by th kernel, and are generated using the following sub-options:

2286

2287

\noindent

2288

\begin{tabular}{l l}

2289

\texttt{--color} $c_1$ $c_2$ ... $c_n$~ &

2290

to select faces of the indicated color numbers\\

2291

\texttt{--group} $g_1$ $g_2$ ... $g_n$~ &

2292

to select faces of the groups named\\

2293

\texttt{--invsel} &

2294

to use the complement of the indicated colors and groups\\

2295

\end{tabular}

2296

2297

The same selection sub-options are used for conforming joining, periodicity,

2298

and interior faces verification. In the case of joining and periodicity,

2299

the selection is automatically restricted to boundary faces, while

2300

in the case of interior face verification, it is restricted to

2301

interior faces bearing attributes.

2302

2303

For example, to select all interior faces of mesh \texttt{example.des}

2304

which bear an attribute (color or group), but are not of color 2 and do not

2305

belong to group ``inlet'', with output of an \emph{Ensight Gold} part

2306

and with no kernel output (using \texttt{-sc}), we use the following

2307

command line (the \verb=\= character marks line continuation):

2308

2309

\noindent

2310

\begin{tabular}{l l l}

2311

\texttt{cs\_preprocess} & \texttt{-m example.des -sc --ensight } \verb=\= \\

2312

& \texttt{--int-face --group entree --color 2 --invsel}\\

2313

\end{tabular}

2314

2315

\subsubsection{Joining of non-conforming meshes\label{sec:optpcs:join}}

2316

2317

Conforming joining of possibly non-conforming meshes is one of the

2318

\pcs's main functions. To activate it, use the \texttt{--join} or

2319

\texttt{-j} command-line options, possibly followed by face selection

2320

criteria and sub-options controlling associated tolerance parameters.

2321

2322

For a simple mesh, it is rarely useful to specify face selection

2323

criteria, as joining is sufficiently automated to detect which faces

2324

may actually be joined. For a more complex mesh, or a mesh with thin

2325

walls which we want to avoid transforming into interior faces, it is

2326

recommended to filter boundary faces that may be joined by using

2327

face selection sub-arguments (see \S\ref{sec:formats_in}). This has the

2328

additional advantage of reducing the number of faces to test for

2329

in the intersection/overlap search, and thus reduced to time

2330

required by the joining algorithm.

2331

2332

One may also modify tolerance criteria using 2 sub-options:

2333

2334

\noindent

2335

\begin{tabular}[top]{p{3.0cm}%

2336

>{\PreserveBackslash\raggedright\hspace{0pt}}p{12.0cm}}

2337

\texttt{--fraction} $r$ &

2338

assigns value $r$ (where $0 < r < 0,49$) to the maximum

2339

intersection distance multiplier ($0,1$ by default). The maximum

2340

intersection distance for a given vertex is based on the length of

2341

the shortest incident edge, multiplied by $r$. The maximum intersection

2342

at a given point along an edge is interpolated from that at its

2343

vertices, as shown on the left of figure \ref{fig_join_tolerance}; \\

2344

\texttt{--plane} $c$ &

2345

assigns the minimum cosine between normals for two faces to be

2346

considered coplanar ($0,8$ by default, which corresponds to almost 37�);

2347

this parameter is used in the second stage of the algorithm, to

2348

reconstruct conforming faces, as shown on the right of figure

2349

\ref{fig_join_tolerance}.\\

2350

\end{tabular}

2351

2352

\begin{figure}[hp]

2353

\centerline{

2354

\includegraphics*[width=14cm]{join_tolerance}}

2355

\caption{Maximum intersection tolerance and faces normal angle

2356

\label{fig_join_tolerance}}

2357

\end{figure}

2358

2359

In practice, we are sometimes led to increase the maximum intersection

2360

distance multiplier to $0.2$ or even $0.3$ when joining curved surfaces,

2361

so that all intersection are detected. As this influences merging

2362

of vertices and thus simplification of reconstructed faces, but also

2363

deformation of ``lateral'' faces, it is recommended only to modify it

2364

if necessary. As for the \texttt{--plane} sub-option, its use has

2365

only been necessary on a few meshes up to now, and always in the

2366

sense of reducing the tolerance (i.e. bringing the minimum cosine closer

2367

to 1) so that face reconstruction does not try to generate faces

2368

from initial faces on different surfaces.

2369

2370

As an example, to join faces with color $2$ of the mesh \texttt{mesh.des}

2371

with an intersection distance multiplier of $0.15$, then joining all

2372

remaining joinable boundary faces with a default parameter of $0.1$,

2373

we may use the following command:

2374

2375

\texttt{cs\_preprocess -m mesh.des -j --fraction 0.15 --color 2 -j}

2376

2377

In cases where any two faces to join always have at least one common vertex

2378

(which is the case with meshes containing hanging nodes generated by

2379

tools such as \harpoon), we may use the \texttt{--semi-conf}

2380

sub-option, for a much faster edge intersection search than with

2381

the usual algorithm.

2382

2383

\subsubsection{Periodicity\label{sec:optpcs:period}}

2384

2385

Handling of periodicity is based on an extension of conforming joining,

2386

as shown on figure \ref{fig_join_periodic}. It is thus not necessary

2387

that periodic faces be conforming (though it usually leads to better

2388

mesh quality). All sub-options relative to conforming joining of

2389

non-conforming faces also apply to periodicity. Note also that

2390

once pre-processed, 2 periodic faces have the same orientation

2391

(possibly adjusted by periodicity of rotation).

2392

2393

\begin{figure}[hp]

2394

\centerline{

2395

\includegraphics*[height=8cm]{join_periodic}}

2396

\caption{Matching of periodic faces

2397

\label{fig_join_periodic}}

2398

\end{figure}

2399

2400

The sub-option for a translation-type periodicity is \texttt{--trans} followed

2401

by the three translation vector components $(tx, ty, tz)$.

2402

The sub-option for a rotation-type periodicity is \texttt{--rota},

2403

which must be completed either by:

2404

2405

\noindent

2406

\texttt{--angle} $\alpha$ \texttt{--dir} $dx$ $dy$ $dz$

2407

2408

\noindent

2409

where $\alpha$ is the rotation angle in degrees and $(dx, dy, dz)$ defines

2410

the rotation axis (using the trigonometric orientation), or by:

2411

2412

\noindent

2413

\texttt{--matrix} $m11$ $m12$ $m13$ $m21$ $m22$ $m23$ $m31$ $m32$ $m33$

2414

2415

\noindent

2416

where

2417

\begin{math}

2418

\left( \begin{smallmatrix}

2419

m_{11} & m_{12} & m_{13} \\ m_{21} & m_{22} & m_{23} \\ m_{31} & m_{32} & m_{33}

2420

\end{smallmatrix} \right)

2421

\end{math}

2422

defines the rotation matrix.

2423

2424

In either case, we may define an invariant point different from $(0, 0, 0)$

2425

by adding the rotation sub-option \texttt{--invpt} $px$ $py$ $pz$.

2426

A translation and a rotation may be combined: in this case, the composite

2427

transformation is always $R \circ T$, independently of the order in which

2428

the translation and rotation are defined.

2429

2430

The orientation of the transformations has no importance (but in the

2431

case of a composite transformation $R \circ T$, $R$ and $T$ must be

2432

consistent).

2433

As with joining, it is recommended to filter boundary faces to process

2434

using a selection criterion. As many periodicities may be built as desired,

2435

as long as boundary faces are present. One a periodicity is handled,

2436

faces having periodic matches do not appear as boundary faces, but as

2437

interior faces, and are thus not available anymore for other

2438

periodicities.

2439

2440

We finish with the example of a periodicity of rotation of $90$ degrees around

2441

an axis of direction $(0, 0, 1)$ passing through $(1, 0, 0)$,

2442

for faces of group ``perio\_1'' and with a joining tolerance of $20\%$ (the

2443

\verb=\= character marks line continuation):

2444

2445

\noindent

2446

\begin{tabular}{l l l}

2447

\texttt{--perio} & \texttt{--rota --angle 90 --dir 0 0 1 --invpt 1 0 0} \verb=\= \\

2448

& \texttt{--fraction 0.2 --group perio\_1}\\

2449

\end{tabular}

2450

2451

\subsubsection{Element orientation correction\label{sec:optpcs:orient}}

2452

2453

We may active the possible element orientation correction using the

2454

\texttt{--reorient} option.

2455

2456

Note that we cannot guarantee correction (or even detection) of a bad

2457

orientation in all cases.

2458

Not all local numbering possibilities of elements are tested,

2459

as we focus on ``usual'' numbering permutations. Moreover,

2460

the algorithms used may produce false positives or fail to find

2461

a correct renumbering in the case of highly non convex elements.

2462

In this case, nothing may be done short of modifying the mesh, as

2463

without a convexity hypothesis, it is not always possible to choose

2464

between two possible definitions starting from a point set.

2465

2466

With a post-processing option such as \texttt{--ensight},

2467

visualizable meshes of corrected elements as well as remaining

2468

badly oriented elements are generated.

2469

2470

\subsection{Environment variables\label{sec:envvpcs}}

2471

2472

Setting a few environment variables specific to the \pcs allows modifying

2473

its default behavior. In general, if a given behavior is modifiable

2474

through an environment variable rather than by a command-line option,

2475

it has little interest for a non-developer, or its modification is

2476

potentially hazardous. The environment variables used by the \pcs

2477

are described here:

2478

2479

\envvar{CS\_PREPROCESS\_MEM\_LOG}

2480

2481

Allows defining a file name in which memory allocation, reallocation,

2482

and freeing is logged.

2483

2484

\envvar{CS\_PREPROCESS\_MIN\_EDGE\_LEN}

2485

2486

Under the indicated length ($10^{-15}$ by default), an edge is considered

2487

to be degenerate and its vertices will be merged after the transformation

2488

to descending connectivity. Degenerate edges and faces will thus be

2489

removed. Hence, the post-processed element does not change, but the

2490

Kernel may handle a prism where the preprocessor input contained a

2491

hexahedron with two identical vertex couples (and thus a face of zero

2492

surface). If the \pcs does not print any information relative to this

2493

type of correction, it means that it has not been necessary. To completely

2494

deactivate this automatic correction, a negative value may be assigned

2495

to this environment variable.

2496

2497

\envvar{CS\_PREPROCESS\_IGNORE\_IDEAS\_COO\_SYS}

2498

2499

If this variable is defined and is a strictly positive integer, coordinate

2500

systems in \ideas universal format files will be ignored. The behavior

2501

of the \pcs will thus be the same as that of versions 1.0 and 1.1.

2502

Note that in any case, non Cartesian coordinate systems are not handled yet.

2503

2504

\envvar{CS\_PREPROCESS\_JOIN\_MAX\_SUB\_FACES}

2505

2506

Defines the number of new faces originating from an initial face ($100$

2507

by default) above which we consider that the joining reconstruction has

2508

probably entered an infinite loop and has thus failed.

2509

This criterion is the last to intervene in the detection of errors and

2510

only comes into play in pathological cases probably related to excessive

2511

deformation of faces to join (and thus a too large tolerance factor).

2512

It should not be necessary to increase it for a mesh really intended for

2513

a calculation, as $100$ faces correspond to a refinement ratio of $10$

2514

in each direction for joined cells, which is excessive from a numerical

2515

standpoint.

2516

2517

\subsubsection{System environment variables\label{sec:envpcs:sys}}

2518

2519

Some system environment variables may also modify the behavior of

2520

the \pcs. For example, if the \pcs was compiled with \med support

2521

on an architecture allowing shared (dynamic) libraries, the

2522

\texttt {LD\_PRELOAD} environment variable may be used to define a

2523

``prioritary'' path to load \med or HDF5 libraries, and thus experiment

2524

with another version of these libraries without recompiling the \pcs.

2525

To determine which shared libraries are used by an executable file, use

2526

the following command: \texttt{ldd~\{executable\_path\}}.

2527

2528

\subsection{Optional functionality\label{sec:pcs:lib_opt}}

2529

2530

Some functions of the \pcs are based on external libraries,

2531

which may not always be available. It is thus possible to configure

2532

and compile the \pcs so as not to use these libraries.

2533

When running the \pcs, the supported options are printed.

2534

The following optional libraries may be used:

2535

2536

\begin{itemize}

2537

2538

\item CGNS library. In its absence, \hyperref[fmtdesc:cgns]{CGNS}

2539

format support is deactivated.

2540

2541

\item \med-file library. In its absence, \hyperref[fmtdesc:med]{\med}

2542

format is simply deactivated.

2543

2544

\item Read compressed files using Zlib. With this option, it is

2545

possible to diretly read mesh files compressed with a

2546

\emph{gzip} type algorithm and bearing a \emph{.gz} extension.

2547

This is limited to formats not already based on an external

2548

library (i.e. it is not usable with CGNS or \med files),

2549

and has memory and CPU time overhead, but may be practical.

2550

Without this library, files must be uncompressed before use.

2551

2552

\end{itemize}

2553

2554

\subsection{General remarks}

2555

2556

Note that the \pcs is in general capable of reading all ``classical''

2557

element types present in mesh files (triangles, quadrangles, tetrahedra,

2558

pyramids, prisms, and hexahedra).

2559

Quadratic or cubic elements are converted upon reading into their

2560

linear counterparts. Vertices referenced by no element (isolated vertices

2561

or centers of higher-degree elements) are discarded. Meshes are read

2562

in the order defined by the user and are appended, vertex and element

2563

indices being incremented appropriately.

2564

\footnote{Possible entity labels are not maintained, as they would

2565

probably not be unique when appending multiple meshes.}

2566

2567

At this stage, volume elements are sorted by type, and the fluid domain

2568

post-processing output is generated if required. Conforming joining,

2569

which may transform cells into general polyhedra only occurs

2570

after this. This allows bypassing a limitation of most post-processing

2571

tools, which do not handle polyhedral elements.

2572

\emph{It is possible to require the simultaneous output with

2573

multiple formats.}

2574

2575

In general, colors or groups assigned to vertices are ignored.

2576

selections are thus based on faces or cells. with tools such

2577

as \simail, faces of volume elements may be referenced directly, while

2578

with \ideas or SALOME, a layer of surface elements bearing the required

2579

colors and groups must be added. Internally, the \pcs always considers

2580

that a layer of surface elements is added (i.e. when reading a \simail

2581

mesh, additional faces are generated to bear cell face colors.

2582

When building the descending connectivity, all faces with the

2583

same topology are merged: the initial presence of two layers of identical

2584

surface elements bearing different colors would thus lead to

2585

a calculation mesh with faces bearing two colors.

2586

2587

\subsection{Files passed to the Kernel\label{sec:pcs:mode_comm}}

2588

2589

Data passed to the Kernel by the \pcs is transmitted using a

2590

binary file, using ``big endian'' data representation, named

2591

\texttt{preprocessor\_output}.

2592

2593

When using the \pcs for mesh verification, data for the Kernel

2594

is usually not needed. In this case, the \texttt{-sc} option may be

2595

used to simulate output, without creation of a {\tt preprocessor\_output}

2596

file.

2597

2598

%==================================

2599

\section{Partitioning for parallel runs\label{sec:parall}}

2600

%==================================

2601

2602

Graph partitioning (using one of the optional \metis or

2603

\scotch libraries) is done using a secondary executable,\\

2604

\texttt{cs\_partition},

2605

which reads the file produced by the \pcs and builds one or

2606

several ``cell $\rightarrow$ domain'' distribution files, named

2607

{\tt domain\_number\_\it{p}} for a partitioning on $p$ sub-domains.

2608

2609

This separation leads to extra work for the Kernel, which must

2610

redistribute data read in the {\tt preprocessor\_output} file

2611

based on the associated partitioning, but avoids requiring

2612

re-running the \pcs whenever running on a different number of

2613

files. This useful mainly for large files (20 to 100 million

2614

cells, for which running the \pcs may require several hours and

2615

a machine with a very large memory capacity. Simple domain

2616

partitioning requires much less time, and in general slightly

2617

less memory than the \pcs.

2618

2619

Without partitioning (for example if neither \metis nor

2620

\scotch is available, or the partitioner has not been run

2621

for the required number of sub-domains), the Kernel will

2622

use a built-in partitioning using a space-filling curve

2623

(Z-curve) technique. This usually leads to partitionings

2624

of lower quality than with graph partitioning, but parallel

2625

performance remains reasonable.

2626

2627

\subsection{Options\label{sec:optcmd:parall}}

2628

2629

To list the partitioner's options, use the following command:

2630

{\tt cs\_partition~-h}

2631

2632

We provide the list of required partitionings an optionally additional

2633

options. For example, to simulate a partitioning for calculations on

2634

64 and 128 processes with no output, we may use the following command:

2635

2636

\texttt{cs\_partition 64 128 --no-write}

2637

2638

\subsubsection{Ignore periodicity\label{sec:optcmdpart:noperiod}}

2639

2640

By default, face periodicity relations are taken into account when building

2641

the ``cell $\rightarrow$ cell'' connectivity graph used for partitioning.

2642

This allows better partitioning optimization, but increases the probability

2643

of having groups of cells at opposite sides of the domain in a same

2644

sub-domain. This is not an issue for standard calculations, but may

2645

degrade performance of search algorithms based on bounding boxes.

2646

It is thus possible to ignore periodicity when partitioning a mesh

2647

using the \texttt{--no-perio} option.

2648

2649

Note that nothing guarantees that a graph partitioner will not place

2650

disjoint cells in the same sub-domain independently of this option,

2651

but this behavior is rare.

2652

2653

\subsubsection{Partitioner choice\label{sec:optpart:partlib}}

2654

2655

If the Partitioner has been configured with both \metis and

2656

\scotch libaries, using the \texttt{--metis} or \texttt{--scotch}

2657

option allows choosing between either library. By default, \emph{metis} is

2658

used if both choices are available.

2659

2660

\subsubsection{Simulation mode\label{sec:optpart:nowrite}}

2661

2662

Using the \texttt{--no-write} option, we can tell the partitioner

2663

not to output a {\tt domain\_number\_\it{p}} file.

2664

Partitioning is thus computed, but not saved.

2665

2666

\subsubsection{Environment variables\label{sec:optpart:envvar}}

2667

2668

\envvar{CS\_PARTITION\_MEM\_LOG}

2669

2670

Allows defining a file name in which memory allocations, reallocations,

2671

and frees will be logged.

2672

2673

%==================================

2674

\section{Main variables}

1653

%==================================

1654

\subsection{User subroutines}

1655

%==================================

1656

%==================================

1657

\label{prg_ssprgutilis}

1658

%==================================

1659

\subsubsection{Preliminary comments}

1660

%==================================

1661

1662

The user can run the calculations with or without an interface, with or

1663

without the user subroutines. Without interface, some user subroutines

1664

are needed. With interface, all the user subroutines are optional.

1665

1666

The parameters can be read in the interface and then in the user

1667

subroutines. In the case that a parameter is specified in the interface

1668

and in a user subroutine, it is the value in the user subroutine that

1669

is taken into acount. For this reason, all the examples of

1670

user subroutines are placed in the \texttt{REFERENCE} directory by the

1671

case setup \texttt{code\_saturne~create}.

1672

1673

%==================================

1674

\subsubsection{Main variables}

2675

1675

%==================================

2676

1676

2677

1677

This section presents a non-exhaustive list of the main variables which

2687

1687

integer array [ia], real array [ra].

2688

1688

2689

1689

%==================================

2690

\subsection{Array sizes}

1690

\subsubsubsection{Array sizes}

2691

1691

%==================================

2692

1692

\label{prg_dimensions}

2693

1693

2726

1726

2727

1727

\bigskip

2728

1728

2729

\variabsize{nphas}{Effective number of phases. \texttt{nphas} must be

2730

inferior or equal to \texttt{nphsmx}. In the current version, \texttt{nphas}

2731

is forced to 1 and should not be changed.}

2732

2733

\variabsize{nphsmx}{Maximum number of phases (default value:

2734

1)\footnote{the data structure of \CS is ready for a multiphase description,

2735

however no multiphase model has been implemented. Moreover, some options of the

2736

code are not compatible with \texttt{nphas} different from 1.}}

2737

2738

1729

\variabsize{nvar}{Number of solved variables (must be lower than

2739

1730

\texttt{nvrmax})}

2740

1731

2772

1763

2773

1764

\bigskip

2774

1765

2775

\variabsize{longia}{Size of the macro array of integer \texttt{ia}}

2776

2777

\variabsize{longra}{Size of the macro array of real \texttt{ra}}

2778

2779

1766

\variabsize{npromx}{Maximum number of physical properties. They will

2780

1767

be stored in the arrays \texttt{propce}, \texttt{propfa} or \texttt{propfb}}

2781

1768

2872

1859

standard use.

2873

1860

2874

1861

%==================================

2875

\subsection{Geometric variables}

1862

\subsubsubsection{Geometric variables}

2876

1863

%==================================

2877

1864

2878

1865

The main geometric variables are available in most of the

2879

subroutines and directly accessible through the following arrays.

1866

subroutines and directly accessible through the following arrays,

1867

defined in the \texttt{mesh} module (i.e. \texttt{use mesh}).

2880

1868

2881

1869

\variab{cdgfac}{cdgfac(ndim,nfac)}{ra}{Coordinates of the

2882

1870

centers of the internal faces}

2884

1872

\variab{cdgfbo}{cdgfbo(ndim,nfabor)}{ra}{Coordinates of the centers of the

2885

1873

boundary face}

2886

1874

2887

2888

1875

\variab{ifacel}{ifacel(2,nfac)}{ia}{Index-numbers of the two (only) neighboring

2889

1876

cells for each internal face}

2890

1877

2898

1885

\variab{ipnfbr}{ipnfbr(nfabor+1)}{ia}{Position of the first node of the each boundary

2899

1886

face in the array \texttt{nodfbr} (see note 3 in paragraph \ref{prg_dimensions}).}

2900

1887

2901

2902

1888

\variab{nodfac}{nodfac(lndfac)}{ia} {Index-numbers of the nodes of each

2903

1889

internal face (see note 3 in paragraph \ref{prg_dimensions}).}

2904

1890

2918

1904

2919

1905

\variab{xyznod}{xyznod(ndim,nnod)}{ra}{Coordinates of the mesh vertices}

2920

1906

2921

In addition, other geometric variables are accessible in sections of

2922

the unidimensional macro-arrays IA (for integers) and RA (for real numbers)

2923

which are passed as arguments

2924

in every subroutine (apart from a few ones of very low level). The

2925

index-number of the first element of these sections is stored in a ``common''

2926

(in the file \texttt{pointe.h}), passed to most of the routines. Hence, the

2927

surface of an internal face \texttt{ifac} is stored in

2928

\texttt{ra(isrfan+ifac-1)}.

2929

Or, the coordinate of vector $\vect{OF}$ (see below for

2930

definition) in the II$^{th}$ direction for face \texttt{ifac} is stored in

2931

\texttt{ra(idofij+(ifac-1)*ndim+ii-1)}\footnote{in Fortran, a multidimensional

2932

array \texttt{a(3,2)} is in fact a unidimensional array containing the elements

2933

\texttt{a(1,1)}, \texttt{a(2,1)}, \texttt{a(3,1)}, \texttt{a(1,2)},

2934

\texttt{a(2,2)}

2935

and \texttt{a(3,2)} in this order.}.

2936

1907

In addition, other geometric variables are useful for gradients

1908

reconstruction.

2937

1909

The main variables of this type are the following:

2938

1910

2939

\variab{idijpf}{idijpf}{i}{In \texttt{ra}, pointer to

2940

\texttt{dijpf(ndim,nfac)}, real array giving, for every internal face,

1911

\variab{dijpf}{dijpf(ndim,nfac)}{ra}{For every internal face,

2941

1912

the three components of the vector $\vect{I'J'}$, where I' and J' are

2942

1913

respectively the orthogonal projections of the neighboring cell centers I and J

2943

1914

on a straight line orthogonal to the face and passing through its center.}

2944

1915

2945

\variab{idiipb}{idiipb}{i}{In \texttt{ra}, pointer to

2946

\texttt{diipb(ndim,nfabor)}, real array giving, for every boundary

1916

\variab{diipb}{diipb(ndim,nfabor)}{ra}{For every boundary

2947

1917

face, the three components of the vector $\vect{II'}$. I' is the

2948

1918

orthogonal projection of I, center of the neighboring cell, on the

2949

1919

straight line perpendicular to the face and passign through its center}

2950

1920

2951

\variab{idist}{idist}{i}{In \texttt{ra}, pointer to \texttt{dist(nfac)}, real

2952

array giving, for every internal face, the scalar product between the

2953

vectors $\vect{IJ}$ and $\vect{n}$. I and J are respectively the centers

1921

\variab{idist}{dist(nfac)}{ra}{For every internal face,

1922

dot product of the vectors $\vect{IJ}$ and $\vect{n}$.

1923

I and J are respectively the centers

2954

1924

of the first and the second neighboring cell. The vector $\vect{n}$ is

2955

1925

the unit vector normal to the face and oriented from the first to the

2956

1926

second cell}

2957

1927

2958

\variab{idistb}{idistb}{i}{In \texttt{ra}, pointer to \texttt{distbr(nfabor)},

2959

real array giving, for every boundary face, the scalar product between

2960

the vectors $\vect{IF}$ and $\vect{n}$. I is the center of the

2961

neighboring cell. F is the face center. The vector $\vect{n}$ is the

2962

unit vector normal to the face and oriented to the exterior of the

2963

domain}

2964

2965

\variab{idofij}{idofij}{i}{In \texttt{ra}, pointer to \texttt{dofij(ndim,nfac)},

2966

real array giving, for every internal face, the components of the

2967

vector $\vect{OF}$. O is the intersection point between the face and

2968

the straight line joining the centers of the two neighboring cells. F

2969

is the face center}

2970

2971

\variab{iicelb}{iicelb}{i}{In \texttt{ia}, pointer to \texttt{icelbr(ncelbr)},

2972

integer array giving the list of cells having at least one boundary face}

2973

2974

\variab{ipond}{ipond}{i}{In \texttt{ra}, pointer to \texttt{pond(nfac)},

2975

real array giving

2976

$\displaystyle\frac{\vect{FJ}.\vect{n}}{\vect{IJ}.\vect{n}}$, for every

2977

internal face. With regard to the mesh quality, its ideal value is 0.5}

2978

2979

\variab{isrfan}{isrfan}{i}{In \texttt{ra}, pointer to \texttt{surfan(nfac)},

2980

real array giving the norm of the surface vector of the internal faces}

2981

2982

\variab{isrfbn}{isrfbn}{i}{In \texttt{ra}, pointer to \texttt{surfbn(nfabor)},

2983

real array giving the norm of the surface of the boundary faces}

1928

\variab{distbr}{distbr(nfabor)}{ra}{For every boundary face,

1929

dot product between the vectors $\vect{IF}$ and $\vect{n}$.

1930

I is the center of the neighboring cell. F is the face center.

1931

The vector $\vect{n}$ is the unit vector normal to the face and

1932

oriented to the exterior of the domain}

1933

1934

\variab{dofij}{dofij(ndim,nfac)}{ra}{For every internal

1935

face, the three components of the vector $\vect{OF}$. O is the intersection

1936

point between the face and the straight line joining the centers

1937

of the two neighboring cells. F is the face center}

1938

1939

\variab{icelbr}{icelbr(ncelbr)}{ia}{List of cells having at

1940

least one boundary face}

1941

1942

\variab{pond}{ipond(nfac)}{ra}{For every internal face,

1943

$\displaystyle\frac{\vect{FJ}.\vect{n}}{\vect{IJ}.\vect{n}}$.

1944

With regard to the mesh quality, its ideal value is 0.5}

1945

1946

\variab{surfan}{surfan(nfac)}{ra}{Norm of the surface vector of

1947

the internal faces}

1948

1949

\variab{surfbn}{surfbn(nfabor)}{ra}{Norm of the surface of the boundary faces}

2984

1950

2985

1951

%==================================

2986

\subsection{Physical variables}

1952

\subsubsubsection{Physical variables}

2987

1953

%==================================

2988

1954

2989

1955

The main physical variables are available in the majority of the

3014

1980

3015

1981

The indexes allowing to mark out the different variables (from 1 to

3016

1982

\texttt{nvar}) are integers available in a ``common'' file called

3017

\texttt{numvar.h}. Some solved variables (pressure, velocity, turbulence)

3018

depend on the considered phase, and the index which refers to it is then a

3019

array of size \texttt{nphsmx}, the maximum number of phases.

1983

\texttt{numvar.h}.

3020

1984

3021

For example, \texttt{ipr(iphas)} refers to the variable ``pressure'' of

3022

the phase \texttt{iphas} (with 1$\leqslant$iphas$\leqslant$nphas):

3023

the pressure of the phase \texttt{iphas} in the cell \texttt{iel}

3024

at the current time step is therefore \texttt{rtp(iel,ipr(iphas))}.

1985

For example, \texttt{ipr} refers to the variable ``pressure'':

1986

the pressure in the cell \texttt{iel} at the current time step is

1987

therefore \texttt{rtp(iel,ipr)}.

3025

1988

3026

1989

The list of integers referring to solved variables is given below. These

3027

1990

variable index-numbers are not only used for the \texttt{rtp} and

3028

1991

\texttt{rtpa} arrays, but also for some arrays of variable associated options

3029

(for instance, \texttt{blencv(ik(iphas))} is the percentage of second-order

3030

convective scheme for the turbulent energy of

3031

the phase \texttt{iphas} when a corresponding turbulent model is used).

1992

(for instance, \texttt{blencv(ik)} is the percentage of second-order

1993

convective scheme for the turbulent energy

1994

when a corresponding turbulent model is used).

3032

1995

3033

1996

\begin{list}{$\bullet$}{}

3034

\item \texttt{ipr(iphas)}\index{\texttt{ipr}}: pressure

3035

\footnote{\texttt{ipr(iphas)} corresponds to a

1997

\item \texttt{ipr}\index{\texttt{ipr}}: pressure

1998

\footnote{\texttt{ipr} corresponds to a

3036

1999

reduced pressure, from which the standard hydrostatic pressure has be

3037

2000

deduced. The total pressure is stored in the PROPCE array}.

3038

\item \texttt{iu(iphas)}\index{\texttt{iu}}: velocity along the X axis.

3039

\item \texttt{iv(iphas)}\index{\texttt{iv}}: velocity along the Y axis.

3040

\item \texttt{iw(iphas)}\index{\texttt{iw}}: velocity along the Z axis.

3041

\item \texttt{ik(iphas)}\index{\texttt{ik}}: turbulent energy, in $k-\varepsilon$,

2001

\item \texttt{iu}\index{\texttt{iu}}: velocity along the X axis.

2002

\item \texttt{iv}\index{\texttt{iv}}: velocity along the Y axis.

2003

\item \texttt{iw}\index{\texttt{iw}}: velocity along the Z axis.

2004

\item \texttt{ik}\index{\texttt{ik}}: turbulent energy, in $k-\varepsilon$,

3042

2005

$k-\omega$ modeling or v2f ($\varphi$-model) modeling.

3043

\item \texttt{ir11(iphas)}\index{\texttt{ir11}}: Reynolds stress R11, in

3044

$R_{ij}-\varepsilon$ or SSG modeling.

3045

\item \texttt{ir22(iphas)}\index{\texttt{ir22}}: Reynolds stress R22, in

3046

$R_{ij}-\varepsilon$ or SSG modeling.

3047

\item \texttt{ir33(iphas)}\index{\texttt{ir33}}: Reynolds stress R33, in

3048

$R_{ij}-\varepsilon$ modeling.

3049

\item \texttt{ir12(iphas)}\index{\texttt{ir12}}: Reynolds stress R12, in

3050

$R_{ij}-\varepsilon$ modeling.

3051

\item \texttt{ir13(iphas)}\index{\texttt{ir13}}: Reynolds stress R13, in

3052

$R_{ij}-\varepsilon$ modeling.

3053

\item \texttt{ir23(iphas)}\index{\texttt{ir23}}: Reynolds stress R23, in

3054

$R_{ij}-\varepsilon$ modeling.

3055

\item \texttt{iep(iphas)}\index{\texttt{iep}}: turbulent dissipation in $k-\varepsilon$,

2006

\item \texttt{ir11}\index{\texttt{ir11}}: Reynolds stress R11, in

2007

$R_{ij}-\varepsilon$ or SSG modeling.

2008

\item \texttt{ir22}\index{\texttt{ir22}}: Reynolds stress R22, in

2009

$R_{ij}-\varepsilon$ or SSG modeling.

2010

\item \texttt{ir33}\index{\texttt{ir33}}: Reynolds stress R33, in

2011

$R_{ij}-\varepsilon$ modeling.

2012

\item \texttt{ir12}\index{\texttt{ir12}}: Reynolds stress R12, in

2013

$R_{ij}-\varepsilon$ modeling.

2014

\item \texttt{ir13}\index{\texttt{ir13}}: Reynolds stress R13, in

2015

$R_{ij}-\varepsilon$ modeling.

2016

\item \texttt{ir23}\index{\texttt{ir23}}: Reynolds stress R23, in

2017

$R_{ij}-\varepsilon$ modeling.

2018

\item \texttt{iep}\index{\texttt{iep}}: turbulent dissipation in $k-\varepsilon$,

3056

2019

$R_{ij}-\varepsilon$ or v2f ($\varphi$-model) modeling.

3057

\item \texttt{iomg(iphas)}\index{\texttt{iomg}}: Specific dissipation rate $\omega$, in

2020

\item \texttt{iomg}\index{\texttt{iomg}}: Specific dissipation rate $\omega$, in

3058

2021

$k-\omega$ SST modeling.

3059

\item \texttt{iphi(iphas)}\index{\texttt{iphi}}: variable $\varphi=\overline{v^2}/k$ in v2f ($\varphi$-model).

3060

\item \texttt{ifb(iphas)}\index{\texttt{ifb}}: variable $\overline{f}$ in v2f ($\varphi$-model).

2022

\item \texttt{iphi}\index{\texttt{iphi}}: variable $\varphi=\overline{v^2}/k$ in v2f ($\varphi$-model).

2023

\item \texttt{ifb}\index{\texttt{ifb}}: variable $\overline{f}$ in v2f ($\varphi$-model).

3061

2024

\item \texttt{isca(j)}\index{\texttt{isca}}: scalar j(1$\leqslant$j$\leqslant$nscal).

3062

2025

\end{list}

3063

2026

3078

2041

``specific physics'' scalars, with

3079

2042

\texttt{nscal=nscaus+nscapp}. \texttt{nscal} must be inferior or

3080

2043

equal to \texttt{nscamx}.

3081

\item The phase related to the scalar \texttt{j} is

3082

\texttt{iphsca(j)}\index{\texttt{iphsca}}.

3083

2044

\item The \texttt{j}$^{\text{th}}$ user scalar is, in

3084

2045

the whole list of the \texttt{nscal} scalars, the scalar number

3085

2046

\texttt{j}. In the list of the \texttt{nvar} solved variables, it

3095

2056

\texttt{rtp(iel,isca(iscapp(j)))}.

3096

2057

3097

2058

\item The temperature (or the enthalpy) is the scalar number

3098

\texttt{iscalt(iphas)}\index{iscalt} in the list of the \texttt{nscal}

3099

scalars. It corresponds to the variable number \texttt{isca(iscalt(iphas))}

2059

\texttt{iscalt}\index{iscalt} in the list of the \texttt{nscal}

2060

scalars. It corresponds to the variable number \texttt{isca(iscalt)}

3100

2061

and its value in the cell \texttt{iel} is

3101

\texttt{rtp(iel,isca(iscalt(iphas)))}. if there is no thermal scalar,

3102

\texttt{iscalt(iphas)} is equal to -1.

2062

\texttt{rtp(iel,isca(iscalt))}. if there is no thermal scalar,

2063

\texttt{iscalt} is equal to -1.

3103

2064

\item A ``user'' scalar number \texttt{j} may represent the average of the

3104

2065

square of the fluctuations of a scalar \texttt{k} ({\em i.e.} the average

3105

2066

$\overline{\varphi^\prime\varphi^\prime}$ for a fluctuating scalar

3136

2097

3137

2098

\item All the properties (used or not) have a unique and distinct index-number,

3138

2099

given automatically by the code and stored in an integer or an integer array

3139

(its size may be the maximum number of phases, the maximum

3140

number of scalars or the maximum number of variables).

2100

(its size may be the maximum number of scalars

2101

or the maximum number of variables).

3141

2102

3142

2103

\item The indexes referring to the different properties stored in the

3143

2104

\texttt{prop{\bf xx}} arrays are given respectively by the following integer arrays:

3153

2114

3154

2115

\end{list}

3155

2116

3156

For instance, the index number corresponding to the density of the phase

3157

\texttt{iphas} is \texttt{irom(iphas)}.\\

3158

In the list of the properties defined at the cell center, the density of

3159

the phase \texttt{iphas} is therefore the \texttt{ipproc(irom(iphas))}$^{\text{th}}$

2117

For instance, the index number corresponding to the density is \texttt{irom}.\\

2118

In the list of the properties defined at the cell center, the density

2119

is therefore the \texttt{ipproc(irom)}$^{\text{th}}$

3160

2120

property: its value at the center of the cell \texttt{iel} is given by \\

3161

\texttt{prop{\bf ce}(iel,ippro{\bf c}(irom(iphas)))}.\\

2121

\texttt{prop{\bf ce}(iel,ippro{\bf c}(irom))}.\\

3162

2122

In the same way, in the list of the properties defined at the boundary

3163

faces, the density of the phase \texttt{iphas} is the

3164

\texttt{ipprob(irom(iphas)))}$^{\text{th}}$ property: its value at the boundary

2123

faces, the density is the

2124

\texttt{ipprob(irom))}$^{\text{th}}$ property: its value at the boundary

3165

2125

face is given by \\

3166

\texttt{prop{\bf fb}(iel,ippro{\bf b}(irom(iphas)))}

2126

\texttt{prop{\bf fb}(iel,ippro{\bf b}(irom))}

3167

2127

3168

2128

The list of properties accessible in the PROPxx arrays is given below (this does

3169

2129

not include the properties linked to the specific physics modules):

3170

2130

3171

\variab{irom}{irom(nphsmx)}{ia}{For each phase, property number

2131

\variab{irom}{irom}{ia}{Property number

3172

2132

corresponding to the density ({\em i.e.} $\rho$ in $kg.m^{-3}$)\\

3173

2133

stored at the cells and the boundary faces}

3174

2134

3175

\variab{iroma}{iroma(nphsmx)}{ia}{For each phase, property number

2135

\variab{iroma}{iroma}{ia}{Property number

3176

2136

corresponding to the density ({\em i.e.} $\rho$ in $kg.m^{-3}$) at the

3177

2137

previous time step, in the case of a second-order extrapolation in time\\

3178

2138

stored at the cells and the boundary faces}

3179

2139

3180

\variab{iviscl}{iviscl(nphsmx)}{ia}{For each phase, property number

2140

\variab{iviscl}{iviscl}{ia}{Property number

3181

2141

corresponding to the fluid molecular dynamic viscosity ({\em i.e.} $\mu$ in

3182

2142

$kg.m^{-1}.s^{-1}$)\\

3183

2143

stored at the cells}

3184

2144

3185

\variab{ivisla}{ivisla(nphsmx)}{ia}{For each phase, property number

2145

\variab{ivisla}{ivisla}{ia}{Property number

3186

2146

corresponding to the fluid molecular dynamic viscosity ({\em i.e.} $\mu$

3187

2147

in $kg.m^{-1}.s^{-1}$) at the previous time step, in the case of a

3188

2148

second-order extrapolation in time\\

3189

2149

stored at the cells}

3190

2150

3191

\variab{ivisct}{ivisct(nphsmx)}{ia}{For each phase, property number

2151

\variab{ivisct}{ivisct}{ia}{Property number

3192

2152

corresponding to the fluid turbulent dynamic viscosity ({\em i.e.}

3193

2153

$\mu_t$ in $kg.m^{-1}.s^{-1}$)\\

3194

2154

stored at the cells}

3195

2155

3196

\variab{ivista}{ivista(nphsmx)}{ia}{For each phase, property number

2156

\variab{ivista}{ivista}{ia}{Property number

3197

2157

corresponding to the fluid turbulent dynamic viscosity ({\em i.e.}

3198

2158

$\mu_t$ in $kg.m^{-1}.s^{-1}$) at the previous time step, in the case of a

3199

2159

second-order extrapolation in time\\

3200

2160

stored at the cells}

3201

2161

3202

\variab{icp}{icp(nphsmx)}{ia}{For each phase, property number

2162

\variab{icp}{icp}{ia}{Property number

3203

2163

corresponding to the specific heat, in case where it is variable

3204

2164

({\em i.e.} $C_p$ in $m^2.s^{-2}.K^{-1}$). See note below\\

3205

2165

stored at the cells}

3206

2166

3207

\variab{icpa}{icpa(nphsmx)}{ia}{For each phase, property number

2167

\variab{icpa}{icpa}{ia}{Property number

3208

2168

corresponding to the specific heat, in case where it is variable

3209

2169

({\em i.e.} $C_p$ in $m^2.s^{-2}.K^{-1}$), at the previous time step,

3210

2170

in the case of a second-order extrapolation in time. See note below\\

3211

2171

stored at the cells}

3212

2172

3213

\variab{itsnsa}{itsnsa(nphsmx)}{ia}{For each phase, in the case of a

2173

\variab{itsnsa}{itsnsa}{ia}{In the case of a

3214

2174

calculation run with a second-order discretisation in time with

3215

2175

extrapolation of the source terms, property number corresponding to the

3216

2176

source term of Navier-Stokes at the previous time step ($kg.m^{-1}.s^{-2}$)\\

3217

2177

stored at the cells}

3218

2178

3219

\variab{itstua}{itstua(nphsmx)}{ia}{For each phase, in the case of a

2179

\variab{itstua}{itstua}{ia}{In the case of a

3220

2180

calculation run with a second-order discretisation in time with

3221

2181

extrapolation of the source terms, property number corresponding to the

3222

2182

source terms of the turbulence at the previous time step\\

3223

2183

stored at the cells}

3224

2184

3225

\variab{itssca}{itssca(nphsmx)}{ia}{For each phase, in the case of a

2185

\variab{itssca}{itssca}{ia}{In the case of a

3226

2186

calculation run with a second-order discretisation in time with

3227

2187

extrapolation of the source terms, property number corresponding to the

3228

2188

source terms of the equations solved for the scalars at the previous

3229

2189

time step ($kg.m^{-1}.s^{-2}$)\\

3230

2190

stored at the cells}

3231

2191

3232

\variab{iestim}{iestim(nestmx,nphsmx)}{ia}{For each phase, property

2192

\variab{iestim}{iestim(nestmx)}{ia}{Property

3233

2193

number for the \texttt{nestmx} error estimators for Navier-Stokes. The estimators

3234

currently available are \texttt{iestim(iespre\index{iespre},iphas)},\\

3235

\texttt{iestim(iesder\index{iesder},iphas)},

3236

\texttt{iestim(iescor\index{iescor},iphas)},

3237

\texttt{iestim(iestot\index{iestot},iphas)}

2194

currently available are \texttt{iestim(iespre\index{iespre})},\\

2195

\texttt{iestim(iesder\index{iesder})},

2196

\texttt{iestim(iescor\index{iescor})},

2197

\texttt{iestim(iestot\index{iestot})}

3238

2198

stored at the cells}

3239

2199

3240

2200

\variab{ifluma}{ifluma(nvarmx)}{ia}{Property number corresponding to the

3241

2201

mass flow associated with each variable ({\em i.e.} for each face

3242

2202

of surface $S$, $\rho \vect{u} \,.\,\vect{S}$ in $kg.s^{-1}$). It

3243

2203

must be noticed that the mass flows are associated with the

3244

variables and not with the phases. This allows to have a distinct

3245

convective flow for each scalar.\\

2204

variables, which allows to have a distinct convective flow for each scalar.\\

3246

2205

stored at the internal faces and boundary faces}

3247

2206

3248

2207

\variab{ifluaa}{ifluaa(nvarmx)}{ia}{Property number corresponding to the

3263

2222

previous time step, in the case of a second-order extrapolation in time\\

3264

2223

stored at the cells}

3265

2224

3266

\variab{ismago}{ismago(nphsmx)}{ia}{For each phase, property number

2225

\variab{ismago}{ismago}{i}{Property number

3267

2226

corresponding to the variable $C$ of the dynamic model, {\em i.e}

3268

2227

so that $\mu_t=\rho C\overline{\Delta}^2\sqrt{2S_{ij}S_{ij}}$ (with the

3269

2228

notations of \cite{benhamadouche01}). $C$ corresponds to $C_s^2$ in the

3270

2229

classical model of Smagorinsky\\

3271

2230

stored at the cells}

3272

2231

3273

\variab{icour}{icour(nphsmx)}{ia}{For each phase, CFL number in each cell at the

2232

\variab{icour}{icour}{i}{CFL number in each cell at the

3274

2233

present time step\\

3275

2234

stored at the cells}

3276

2235

3277

\variab{ifour}{ifour(nphsmx)}{ia}{For each phase, Fourier number in each cell at

2236

\variab{ifour}{ifour}{i}{Fourier number in each cell at

3278

2237

the present time step\\

3279

2238

stored at the cells}

3280

2239

3281

\variab{iprtot}{iprtot(nphsmx)}{ia}{For each phase\footnote{Although the data

3282

structure of \CS allows multi-phase variables, the algorithm does not allow

3283

more than one pressure}, total pressure in each cell\\

2240

\variab{iprtot}{iprtot}{i}{Total pressure in each cell\\

3284

2241

stored at the cells}

3285

2242

3286

2243

\variab{ivisma}{ivisma(1 or 3)}{ia}{When the ALE method for deformable meshes is

3311

2268

\begin{list}{$\bullet$}{}

3312

2269

\item It is the case for the specific heat $C_p$.

3313

2270

\begin{list}{-}{}

3314

\item If $C_p$ is constant for the phase \texttt{iphas}, it can be specified in

3315

the interface or by indicating \texttt{icp(iphas)=0} in \texttt{usini1},

3316

and the property will be stored in the real number \texttt{cp0(iphas)}.

2271

\item If $C_p$ is constant, it can be specified in

2272

the interface or by indicating \texttt{icp=0} in \texttt{usini1},

2273

and the property will be stored in the real number \texttt{cp0}.

3317

2274

\item If $C_p$ is variable, it can be specified in the interface or by

3318

indicating \texttt{icp(iphas)=1} in \texttt{usini1}. The code will then

3319

modify this value to make \texttt{icp(iphas)} refer to the effective

3320

property number corresponding to the specific heat of the phase

3321

\texttt{iphas}, in a way which is transparent for the user. For each cell

2275

indicating \texttt{icp=1} in \texttt{usini1}. The code will then

2276

modify this value to make \texttt{icp} refer to the effective

2277

property number corresponding to the specific heat,

2278

in a way which is transparent for the user. For each cell

3322

2279

\texttt{iel}, the value of $C_p$ is then given in \texttt{usphyv} and

3323

stored in the array \texttt{propce(iel,ipproc(icp(iphas)))}.

2280

stored in the array \texttt{propce(iel,ipproc(icp))}.

3324

2281

\end{list}

3325

2282

\item It is the same for the diffusivity $K$ of each scalar \texttt{iscal}.

3326

2283

\begin{list}{-}{}

3337

2294

\end{list}

3338

2295

\end{list}

3339

2296

3340

3341

2297

\minititre{Note: cumulated duration associated with the averages

3342

2298

defined by the user}\label{prg_moyennes}

3343

2299

The cumulated duration associated with the calculation of a time averages

3381

2337

although they concern only the wall boundary faces.

3382

2338

3383

2339

3384

3385

%%%ICI%%%%%%%%%%%%%%%

3386

3387

2340

%==================================

3388

\subsection{Variables related to the numerical methods}

2341

\subsubsubsection{Variables related to the numerical methods}

3389

2342

%==================================

3390

2343

3391

2344

The main numerical variables and ``pointers''\footnote{As for the

3392

2345

geometrical variables, some variables may be accessed to directly in

3393

sections of the unidimensional macro-arrays IA (for the integers) and RA

3394

(for the real numbers) which are present as arguments in every

3395

subroutine (apart from a few ones of very low level). The number of the

3396

first position of these sections in IA and RA is indicated by an integer

3397

stored in a ``common''. These integers are called ``pointers''} are

2346

sections of the unidimensional macro-array \texttt{ra}

2347

(for the real numbers) which is present as an argument to many

2348

subroutines. The number of the first position of these sections in \texttt{ra}

2349

is indicated by an integer stored in a the \texttt{pointe} Fortran module.

2350

These integers are referred to as ``pointers''} are

3398

2351

displayed below.

3399

2352

3400

2353

\minititre{Boundary conditions}

3424

2377

element of the section allowing to mark out the ``wall'' (\texttt{itypfb=iparoi}

3425

2378

or \texttt{iparug})

3426

2379

or ``symmetry'' (\texttt{itypfb=isymet}) boundary faces in order to prevent the

3427

mass flow (these faces are impermeable). For instance, for the phase

3428

\texttt{iphas}, if the face \texttt{ifac} is a wall or symmetry face,

3429

\texttt{ia(iismph+ifac-1)=0} (with \texttt{iismph=iisymp+nfabor*(iphas-1))}.\\

2380

mass flow (these faces are impermeable). For instance,

2381

if the face \texttt{ifac} is a wall or symmetry face,

2382

\texttt{ia(iismph+ifac-1)=0} (with \texttt{iismph=iisymp+nfabor)}.\\

3430

2383

Otherwise \texttt{ia(iisymp+ifac-1)=1}. \\

3431

2384

In some subroutines, an array called \texttt{isympa(nfabor)}\index{\texttt{isympa}}

3432

2385

allows to simplify the coding with \texttt{isympa(ifac)=ia(iismph+ifac-1)}}

3433

2386

3434

\variab{iitrif}{iitrif}{i}{In \texttt{ia}, pointer to \texttt{itrifb}}

3435

3436

\variab{iitypf}{iitypf}{i}{In \texttt{ia}, pointer to \texttt{itypfb}}

3437

3438

\variab{itrifb}{itrifb(nfabor,nphas)}{ia}{Indirection array allowing to

2387

\variab{itrifb}{itrifb(nfabor)}{ia}{Indirection array allowing to

3439

2388

sort the boundary faces according to their boundary condition type \texttt{itypfb}}

3440

2389

3441

\variab{itypfb}{itypfb(nfabor,nphas)}{ia}{Boundary condition type at the

3442

boundary face \texttt{ifac} for the phase \texttt{iphas} (see user subroutine

3443

\texttt{usclim})}

2390

\variab{itypfb}{itypfb(nfabor)}{ia}{Boundary condition type at the

2391

boundary face \texttt{ifac} (see user subroutine \texttt{usclim})}

3444

2392

3445

\variab{iuetbo}{iuetbo}{i}{In \texttt{ra}, pointer to \texttt{uetbor}, used to

3446

store the friction velocity at the wall, in the case of a LES calculation with van

3447

Driest-wall damping}

2393

\variab{uetbor}{uetbor(nfabor)}{ra}{Friction velocity at the wall,

2394

in the case of a LES calculation with van Driest-wall damping}

3448

2395

3449

2396

\minititre{Distance to the wall}

3450

2397

3451

\variab{iifapa}{iifapa(nphsmx)}{ia}{For each phase, the pointer in \texttt{ia}

3452

which marks out the number of the wall face(type \texttt{itypfb=iparoi} or

3453

\texttt{iparug}) which is closest to the center of a given volume when necessary

2398

\variab{ifapat}{ifapat(ncelet)}{ra}{Number of the wall

2399

face(type \texttt{itypfb=iparoi} or \texttt{iparug}) which is closest

2400

to the center of a given volume when necessary

3454

2401

($R_{ij}-\varepsilon$ with wall echo, LES with van Driest-wall damping,

3455

2402

or SST $k-\omega$ turbulence model) and when \texttt{icdpar=2}.

3456

The number of the wall face (for the phase \texttt{iphas}) which is the closest to

3457

the center of the cell \texttt{iel} is therefore \texttt{ia(iifapa(iphas)+iel-1)}. This calculation method is not compatible with parallelism and periodicity}

2403

The number of the wall face which is the closest to

2404

the center of the cell \texttt{iel} is \texttt{ifapat(iel1)}.

2405

This calculation method is not compatible with parallelism and periodicity}

3458

2406

3459

\variab{idipar}{idipar}{i}{For each phase, pointer in \texttt{ra} to the section

3460

allowing to mark out the distance between the center of a given volume and the

3461

closest wall, when it is necessary ($R_{ij}-\varepsilon$ with wall echo,

2407

\variab{dispar}{dispar(ncelet)}{ra}{Distance between the center of

2408

a given volume and the closest wall, when it is necessary

2409

($R_{ij}-\varepsilon$ with wall echo,

3462

2410

LES with van Driest-wall damping, or SST $k-\omega$ turbulence model)

3463

2411

and when \texttt{icdpar=1}. The distance between the center of the cell

3464

\texttt{iel} and the closest wall is therefore \texttt{ra(idipar+iel-1)}}

2412

\texttt{iel} and the closest wall is \texttt{dispar(iel)}}

3465

2413

3466

\variab{iyppar}{iyppar}{i}{For each phase, pointer in \texttt{ra} to the section

3467

allowing to mark out the adimensional distance $y^+$ between a given

2414

\variab{yplpar}{yplpar}{ra}{Adimensional distance $y^+$ between a given

3468

2415

volume and the closest wall, when it is necessary (LES with van

3469

2416

Driest-wall damping) and when \texttt{icdpar=1}. The adimensional distance $y^+$

3470

2417

between the center of the cell \texttt{iel} and the closest wall is therefore

3471

\texttt{ra(iyppar+iel-1)}}

2418

\texttt{yplpar(iel1)}}

3472

2419

3473

2420

\minititre{Pressure drops}

3474

2421

3475

\variab{iicepd}{iicepd(nphsmx)}{ia}{For each phase \texttt{iphas}, pointer in

3476

\texttt{ia} to \\

3477

\texttt{icepdc(ncepdc(iphas))}, array allowing to mark out the index-numbers of

3478

the \\

3479

\texttt{ncepdc(iphas)} cells in which a pressure drop is imposed. \\

3480

The number of these cells is therefore given by

3481

\texttt{icepdc(ii)=ia(iicepd(iphas)+ii-1)}, with

3482

1$\leqslant$\texttt{ii}$\leqslant$\texttt{ncepdc(iphas)}. See the user subroutine

3483

\texttt{uskpdc}}

3484

3485

\variab{icepdc}{icepdc(ncepdc(iphas))}{ia}{Number of the \texttt{ncepdc(iphas)}

2422

\variab{icepdc}{icepdc(ncepdc)}{ia}{Number of the \texttt{ncepdc}

3486

2423

cells in which a pressure drop is imposed. See \texttt{iicepd} and the user

3487

2424

subroutine \texttt{uskpdc}}

3488

2425

3489

\variab{ickupd}{ickupd(nphsmx)}{ia}{For each phase \texttt{iphas}, pointer in

3490

\texttt{ra} to \\

3491

\texttt{ckupdc(ncepdc(iphas),6)}, array allowing to mark out the\\

3492

coefficients of the pressure drop tensor of the

3493

\texttt{ncepdc(iphas)} cells in which a pressure drop is imposed. See the user

3494

subroutine \texttt{uskpdc}}

3495

3496

\variab{ckupdc}{ckupdc(ncepdc(iphas),6)}{ra}{Value of the

3497

coefficients of the pressure drop tensor of the

3498

\texttt{ncepdc(iphas)} cells in which a pressure drop is imposed.

2426

\variab{ckupdc}{ckupdc(ncepdc,6)}{ra}{Value of the

2427

coefficients of the pressure drop tensor of the

2428

\texttt{ncepdc} cells in which a pressure drop is imposed.

3499

2429

See \texttt{ickpdc} and the user subroutine \texttt{uskpdc}}

3500

2430

3501

\variab{ncepdc}{ncepdc(nphsmx)}{ia}{For each phase, number of cells in

2431

\variab{ncepdc}{ncepdc}{ia}{Number of cells in

3502

2432

which a pressure drop is imposed. See the user subroutine \texttt{uskpdc}}

3503

2433

3504

2434

3505

2435

\minititre{Mass sources}

3506

2436

3507

\variab{iicesm}{iicesm(nphsmx)}{ia}{For each phase \texttt{iphas}, pointer in

3508

\texttt{ia} to \\

3509

\texttt{icetsm(ncetsm(iphas)}, array allowing to mark out the numbers of the \\

3510

\texttt{ncetsm(iphas)} cells in which a source of mass is imposed. The number of

3511

these cells is therefore given by \texttt{icetsm(ii)=ia(iicesm(iphas)+ii-1)}, with

3512

1$\leqslant$\texttt{ii}$\leqslant$\texttt{ncetsm(iphas)}. See the user subroutine

3513

\texttt{ustsma}}

3514

3515

\variab{iitpsm}{iitpsm(nphsmx)}{ia}{For each phase \texttt{iphas}, pointer in

3516

\texttt{ia} to \texttt{itypsm} (type of mass source for each variable).

3517

See \texttt{itypsm} and the user subroutine \texttt{ustsma}}

3518

3519

\variab{icetsm}{icetsm(ncetsm(iphas))}{ia}{Number of the \texttt{ncetsm(iphas)}

2437

\variab{icetsm}{icetsm(ncetsm)}{ia}{Number of the \texttt{ncetsm}

3520

2438

cells in which a mass source term is imposed. See \texttt{iicesm} and the user

3521

2439

subroutine \texttt{ustsma}}

3522

2440

3523

\variab{ismace}{ismace(nphsmx)}{ia}{For each phase \texttt{iphas}, pointer in

3524

\texttt{ra} to \texttt{smacel} (mass source term and if necessary injection value

3525

for every variable apart from pressure). See \texttt{smacel} and the user subroutine

3526

\texttt{ustsma}}

3527

3528

\variab{itypsm}{itypsm(ncetsm(iphas),nvar)}{ia}{Type of mass source term

2441

\variab{itypsm}{itypsm(ncetsm,nvar)}{ia}{Type of mass source term

3529

2442

for each variable (0 for an injection at ambient value, 1 for an

3530

2443

injection at imposed value). See the user subroutine \texttt{ustsma}}

3531

2444

3532

\variab{ncetsm}{ncetsm(nphsmx)}{ia}{For each phase, number of cells with

2445

\variab{ncetsm}{ncetsm}{i}{Number of cells with

3533

2446

mass sources. See the user subroutine \texttt{ustsma}}

3534

2447

3535

\variab{smacel}{smacel(ncetsm(iphas),nvar)}{ra}{Value of the mass source

2448

\variab{smacel}{smacel(ncetsm,nvar)}{ra}{Value of the mass source

3536

2449

term for pressure. For the other variables, eventual imposed injection

3537

2450

value. See the user subroutine \texttt{ustsma}}

3538

2451

3541

2454

\variab{nfpt1d}{nfpt1d}{i}{Number of boundary faces which are coupled

3542

2455

with a wall 1D thermal module. See the user subroutine \texttt{uspt1d}}

3543

2456

3544

\variab{iifpt1}{iifpt1}{i} {In \texttt{ia}, pointer to \texttt{ifpt1d(nfpt1d)},

3545

array allowing to mark out the numbers of the \texttt{nfpt1d} boundary faces

3546

which are coupled with a wall 1D thermal module. The number of these boundary

3547

faces is therefore given by \texttt{ifpt1d(ii)=ia(iifpt1+ii-1)}, with

2457

\variab{ifpt1d}{ifpt1d}{ia} {Array allowing to mark out the numbers of

2458

the \texttt{nfpt1d} boundary faces which are coupled with a wall 1D

2459

thermal module. The numbers of these boundary faces are

2460

given by \texttt{ifpt1d(ii)}, with

3548

2461

1$\leqslant$\texttt{ii}$\leqslant$\texttt{nfpt1d}.

3549

2462

See the user subroutine \texttt{uspt1d}}

3550

2463

3551

\variab{inppt1}{inppt1}{i}{In \texttt{ia}, pointer to \texttt{nppt1d(nfpt1d)},

3552

array giving the number of discretisation cells in the 1D wall for the

2464

\variab{nppt1d}{nppt1d}{ia}{Number of discretisation cells in the 1D wall for the

3553

2465

\texttt{nfpt1d} boundary faces which are coupled with a wall 1D thermal module. The

3554

number of cells for these boundary faces is therefore given by

3555

\texttt{nppt1d(ii)=ia(inppt1+ii-1)}, with

2466

number of cells for these boundary faces is given by

2467

\texttt{nppt1d(ii)}, with

3556

2468

1$\leqslant$\texttt{ii}$\leqslant$\texttt{nfpt1d}. See

3557

2469

the user subroutine \texttt{uspt1d}}

3558

2470

3559

\variab{ieppt1}{ieppt1}{i}{In \texttt{ia}, pointer to \texttt{eppt1d(nfpt1d)},

3560

array giving the thickness of the 1D wall for the \texttt{nfpt1d} boundary faces

3561

which are coupled with a wall 1D thermal module. The wall thickness for these

3562

boundary faces is therefore given by \texttt{eppt1d(ii)=ia(ieppt1+ii-1)}, with

2471

\variab{eppt1d}{eppt1d}{ia}{Thickness of the 1D wall for the

2472

\texttt{nfpt1d} boundary faces which are coupled with a wall 1D thermal

2473

module. The wall thickness for these boundary faces is therefore given

2474

by \texttt{eppt1d(ii)}, with

3563

2475

1$\leqslant$\texttt{ii}$\leqslant$\texttt{nfpt1d}. See the user subroutine

3564

2476

\texttt{uspt1d}}

3565

2477

3569

2481

3570

2482

\variab{ifmcel}{ifmcel(ncelet)}{ia}{Family number of the elements. See note 1}

3571

2483

3572

\variab{is2kw}{is2kw(nphsmx)}{ia}{For each phase \texttt{iphas}, pointer in

3573

\texttt{ra} to the section storing the square of the norm of the deformation

3574

rate tensor. In the cell \texttt{iel}, for the phase \texttt{iphas},

3575

$S^2=2S_{ij}S_{ij}$

3576

is therefore given by \texttt{ra(is2kw(iphas)+iel-1)}. This array is defined only

3577

when the phase \texttt{iphas} is treated with the SST $k-\omega$ turbulence model}

2484

\variab{s2kw}{s2kw(ncelet)}{ra}{Square of the norm of the deformation

2485

rate tensor. In the cell \texttt{iel}, $S^2=2S_{ij}S_{ij}$

2486

is given by \texttt{ra(is2kw+iel-1)}. This array is defined only

2487

with the SST $k-\omega$ turbulence model}

3578

2488

3579

\variab{idvukw}{idvukw(nphsmx)}{ia}{For each phase \texttt{iphas}, pointer in

3580

\texttt{ra} to the section storing the divergence of the velocity. In the

3581

cell \texttt{iel}, for the phase \texttt{iphas}, $div(\vect{u})$ is therefore

3582

given by \texttt{ra(idvukw(iphas)+iel-1)}. This array is defined only when the

3583

phase \texttt{iphas} is treated with the SST $k-\omega$ turbulence model

2489

\variab{divukw}{divukw}{ia}{Divergence of the velocity. In the

2490

cell \texttt{iel}, $div(\vect{u})$ is given by \texttt{divukw(iel1)}.

2491

This array is defined only with the SST $k-\omega$ turbulence model

3584

2492

(because in this case it may be calculated at the same time as $S^2$)}

3585

2493

3586

2494

\variab{ngrmmx}{ngrmmx}{i}{upper limit of the number of grid levels

3587

in the case of a multigrid solving (see \texttt{ngrmax})}

3588

3589

\variab{ia}{ia(longia)}{ia}{Integer work array}

3590

3591

\variab{ra}{ra(longra)}{ra}{Real work array}

3592

2495

when using the multigrid solver (see \texttt{ngrmax})}

2496

2497

\variab{ra}{ra(ifinra)}{ra}{Real work array}

3593

2498

3594

2499

\minititre{Note: boundary conditions}

3595

2500

The boundary conditions in \CS boil down to determine a value for the

3615

2520

\mbox{$\phi_f$=\texttt{coefa(ifac,iclvar)+coefb(ifac,iclvar)} $\phi_{I'}$}

3616

2521

with \texttt{iclvar=iclrtp(ivar,icoef)}

3617

2522

3618

3619

%==================================

3620

\subsection{User arrays}

3621

%==================================

3622

The code allows to define two user arrays, one integer array and one

3623

real array. The default size of these arrays is zero, and may be changed

3624

in \texttt{usini1}. The two arrays are then passed as arguments in every

3625

user subroutine of the code. For instance, a local variable calculated during

3626

the determination of the physical properties (user subroutine

3627

\texttt{usphyv}) may be stored in these arrays and sent to the

3628

post-processor at the end of the time step (user subroutine \texttt{usvpst}).

3629

3630

\variab{nituse}{nituse}{i}{Size of the user integer array}

3631

3632

\variab{nrtuse}{nrtuse}{i}{Size of the user real array}

3633

3634

\variab{ituser}{ituser(nituse)}{ia}{User integer array}

3635

3636

\variab{rtuser}{rtuser(nrtuse)}{ra}{User real array}

3637

3638

%==================================

3639

\subsection{Developer arrays}

3640

%==================================

3641

The code allows to define two developer arrays (similar to the user

3642

arrays \texttt{ituser} and \texttt{rtuser}), one integer array a one real array.

3643

The default size of these arrays is zero, and may be changed in

3644

\texttt{ustbus}. The two arrays are then passed as arguments in the rest

3645

of the code. They are designed to be used during the transitory

3646

development phases, in order to ease the tests (transfer of pieces of

3647

informations without consequence on the arguments of the subroutines).

3648

3649

\variab{nideve}{nideve}{i}{Size of the developer integer array}

3650

3651

\variab{nrdeve}{nrdeve}{i}{Size of the developer real array}

3652

3653

\variab{idevel}{idevel(nideve)}{ia}{Complementary integer array, used

3654

during development and test phases}

3655

3656

\variab{rdevel}{rdevel(nrdeve)}{ra}{Complementary real array, used

3657

during development and test phases}

3658

3659

%==================================

3660

\subsection{Parallelism and periodicity}

2523

%==================================

2524

\subsubsubsection{User arrays}

2525

%==================================

2526

Modules containing user arrays accessible from all user subroutines may

2527

be defined in the \texttt{user\_modules.f90} file. This file is

2528

compiled before any other Fortran user file, to ensure modules

2529

may be acessed in other user subroutines using the \texttt{use <module>}

2530

construct.

2531

2532

%==================================

2533

\subsubsubsection{Parallelism and periodicity}

3661

2534

%==================================

3662

2535

3663

2536

Parallelism is based on domain partitioning: each processor is assigned

3704

2577

\label{prg_paralperio}

3705

2578

{\bf Activation}

3706

2579

3707

Parallelism and periodicity are activated by means of the launch script

2580

Parallism is activated by means GUI or of the launch scripts

3708

2581

in the standard cases:

3709

2582

\begin{list}{$\bullet$}{}

3710

2583

3711

\item On clusters with PBS batch systems, the launching of a parallel run

3712

requires to complete the \texttt{PBS} batch cards located in the

3713

beginning of \texttt{runcase}, and particularly to set the number of

3714

physical nodes (\texttt{nodes}) and the number of physical

3715

processors per node (\texttt{ppn}) wanted. This can be done through

3716

the Graphical Interface or by editing the \texttt{runcase} file directly.

3717

The number of processors used for the calculation will then be set

3718

automatically to the number of processors reserved and the

3719

variable \texttt{NUMBER\_OF\_PROCESSORS} can be left empty

3720

(see also \S\ref{prg_runcase}).

3721

3722

\item On clusters with LSF batch systems (like the CCRT machines),

3723

the launching of a parallel run

3724

requires to complete the \texttt{LSF} batch cards located in the

3725

beginning of \texttt{runcase}, and particularly to set the

3726

number of processors (\texttt{\#BSUB -n}) wanted and the limit CPU

3727

time (\texttt{\#BSUB -W}). As for now, this can only be done

3728

by editing the \texttt{runcase} file directly.

3729

The number of processors used for the calculation will then be set

3730

automatically to the number of processors reserved and the

3731

variable \texttt{NUMBER\_OF\_PROCESSORS} can be left empty

3732

(see also \S\ref{prg_runcase}).

3733

3734

\item On clusters with other batch systems, \texttt{runcase} file may have to

3735

be modified manually. Please do not hesitate to contact the \CS support

2584

\item On clusters with batch systems, the launching of a parallel run

2585

requires to complete the batch cards located in the

2586

beginning of \texttt{runcase} or \texttt{runcase\_batch}

2587

script, and set the number of MPI processes, or the numbers

2588

of physical nodes and processors per node (\texttt{ppn}) wanted.

2589

This can be done through the Graphical Interface or by editing

2590

the \texttt{runcase} or \texttt{runcase\_batch} file directly.

2591

The number of processors defined here will override the number

2592

defined through the GUI in a non-batch environment

2593

(so that studies defined on one environment may be migrated

2594

to larger compute resources easily), but it may be overriden

2595

by the \item \texttt{define\_case\_parameters} function from

2596

the \texttt{cs\_user\_scripts.py} file, or by setting the

2597

\texttt{n\_procs\_weight}, \texttt{n\_procs\_min}, and

2598

\texttt{n\_procs\_max} parameters for the different domains

2599

defined in \texttt{runcase\_coupling}.

2600

2601

\item On clusters with unsupported batch systems,

2602

\texttt{runcase} file may have to be modified manually.

2603

Please do not hesitate to contact the \CS support

3736

2604

(saturne-support@edf.fr) so that these modifications can be added to

3737

2605

the standard launch script to make it more general.

3738

2606

3739

\item Although on batch systems the \texttt{NUMBER\_OF\_PROCESSORS} variable

3740

in the script (indicating the number of processors used for the

3741

calculation) is filled automatically to the number of processors

3742

reserved, the user can still choose to specify another value for it.

3743

This might only happen in very specific conditions and is not advised,

3744

as it will probably not be compatible with the batch system. Indeed,

3745

batch systems forbid to launch a calculation on more processors than the

3746

number of processors reserved, and some batch systems also forbid to

3747

launch a calculation on less processors than the number of processors

3748

reserved (automatic timeout on the idle processors that will stop the

3749

whole calculation).

3750

3751

3752

\item Periodicity is activated through the Graphical Interface or by completing

3753

the \texttt{COMMAND\_PERIO} of the launch script \texttt{runcase}.

3754

The transformation

3755

allowing to pass from a boundary to the other one must be defined

3756

(the direction does not matter) and the set of periodic faces

3757

should be (optional but strongly advised) marked out (for instance

3758

by means of a color).

3759

3760

\item Periodicity is compatible with parallelism.

3761

3762

\item Periodicity can also work when the periodic boundaries are meshed

3763

differently (periodicity of non-conforming faces), {\it apart} from

3764

the case of a 180 degree rotation periodicity with faces coupled

3765

on the rotation axis.

3766

3767

2607

\item A parallel calculation may be stopped in the same manner as a

3768

2608

sequential one using a \texttt{ficstp} file (see praragraph

3769

2609

\ref{prg_ficstp}).

3779

2619

\vspace{0.5cm}

3780

2620

{\bf User subroutines}

3781

2621

3782

The user can notice in a subroutine

2622

The user can check in a subroutine

3783

2623

\begin{list}{-}{}

3784

2624

\item that the presence of periodicity is tested with the variable

3785

2625

\texttt{iperio} (=1 if periodicity is activated);

3862

2702

(irangp.le.0) then ...}).

3863

2703

\end{list}

3864

2704

3865

3866

2705

{\bf Some notes about periodicity}

3867

2706

2707

Note that periodic faces are not part of the domain boundary:

2708

periodicity is interpreted as a ``geometric'' condition

2709

rather than a classical boundary condition.

2710

3868

2711

Some particular points should be reminded:

3869

2712

\begin{list}{-}{}

2713

\item Periodicity can also work when the periodic boundaries are meshed

2714

differently (periodicity of non-conforming faces), {\it apart} from

2715

the case of a 180 degree rotation periodicity with faces coupled

2716

on the rotation axis.

3870

2717

\item rotation periodicity is incompatible with

3871

2718

\begin{list}{-}{}

3872

2719

\item semi-transparent radiation,

3879

2726

\end{list}

3880

2727

3881

2728

%==================================

3882

\subsection{Geometry and particule arrays

2729

\subsubsubsection{Geometry and particule arrays

3883

2730

related to Lagrangian modeling}

3884

2731

%==================================

3885

2732

3934

2781

\minititre{Note: continuous eulerian phase number}

3935

2782

The current version of Lagrangian module is planned to work with only one

3936

2783

eulerian phase. This phase carries inclusions, and source terms of

3937

backward coupling are applied to it, if necessary. The number of this

3938

phase is stored in the variable \texttt{ilphas}\index{\texttt{ilphas}}.

3939

The standard value is \texttt{ilphas = 1}.

2784

backward coupling are applied to it, if necessary.

3940

2785

3941

2786

\minititre{Lagrangian arrays}

3942

2787

4105

2950

4106

2951

4107

2952

%==================================

4108

\subsection{Variables saved to allow calculation restarts}

2953

\subsubsubsection{Variables saved to allow calculation restarts}

4109

2954

%==================================

4110

2955

4111

The directory \texttt{RESTART*} contains:

2956

The directory \texttt{checkpoint} contains:

4112

2957

\begin{list}{-}{}

4113

\item \texttt{suiava}: main restart file,

4114

\item \texttt{suiavx}: auxiliary restart file (see \texttt{ileaux\index{ileaux}},

2958

\item \texttt{main}: main restart file,

2959

\item \texttt{auxiliary}: auxiliary restart file (see \texttt{ileaux\index{ileaux}},

4115

2960

\texttt{iecaux\index{iecaux}}),

4116

\item \texttt{rayava}: restart file for the radiation module,

4117

\item \texttt{lagava}: main restart file for the Lagrangian module,

4118

\item \texttt{lasava}: auxiliary restart file for the Lagrangian module (mainly

4119

for the statistics),

4120

\item \texttt{t1dava}: restart file for the 1D wall thermal module,

4121

\item \texttt{vorava}: restart file for the vortex method (see

2961

\item \texttt{radiative\_transfer}: restart file for the radiation module,

2962

\item \texttt{lagrangian}: main restart file for the Lagrangian module,

2963

\item \texttt{lagrangian\_stats}: auxiliary restart file for the Lagrangian module (mainly for the statistics),

2964

\item \texttt{1dwall\_module}: restart file for the 1D wall thermal module,

2965

\item \texttt{vortex}: restart file for the vortex method (see

4122

2966

\texttt{ivrtex\index{ivrtex}}).

4123

2967

\end{list}

4124

2968

4192

3036

file only when they are extrapolated in time, or with the Joule effect for the

4193

3037

specific heat).

4194

3038

4195

Apart from \texttt{vorava} which has a different structure and is

3039

Apart from \texttt{vortex} which has a different structure and is

4196

3040

always in text format, all the restart files are binary

4197

files. Nonetheless, they may be dumped by the \texttt{cs\_io\_dump}

4198

tool provided with the Preprocessor.

3041

files. Nonetheless, they may be dumped or compared using

3042

the \texttt{cs\_io\_dump} tool.

4199

3043

4200

3044

In the case of parallel calculations, it should be noted that all the processors

4201

3045

will write their restart data in the same files. Hence, for instance, there will

4202

always be one and only one \texttt{suiava} file, whatever the number of

3046

always be one and only one \texttt{main} file, whatever the number of

4203

3047

processors used. The data in the file are written according to the initial full

4204

domain index-numbers for the cells, faces and nodes. This allows in particular

4205

to continue with {\it p} processors a calculation begun with {\it n} processors,

4206

or to make the restart files independent of any vectorial renumbering that may

3048

domain ids for the cells, faces and nodes. This allows in particular

3049

to restart using {\it p} processors a calculation begun with {\it n} processors,

3050

or to make the restart files independent of any mesh renumbering that may

4207

3051

be carried out in each domain.

4208

3052

4209

{\bf On the other hand}, if the numbering of the initial full domain mesh is

4210

modified, the restart files will not be compatible. This may be the case if the

4211

mesh is composed of different elements that are pasted by the Preprocessor module

4212

and the order of the different elements has been changed in the Preprocessor command

4213

line between two calculations.

4214

4215

3053

{\em WARNING: if the mesh is composed of several files, the order

4216

3054

in which they appear in the launch script or in the Graphical Interface must not

4217

3055

be modified in case of a calculation restart\footnote{when uncertain, the user

4218

3056

can check the saved copy of the launch script in the \texttt{RESU} directory, or

4219

the head of the \texttt{listpre} file, which repeats the command line passed to

4220

the Preprocessor module}.}

3057

the head of the \texttt{preprocessor*.log} files, which repeat the

3058

command lines passed to the Preprocessor module}.}

4221

3059

4222

{\em NOTE: when meshes are pasted by the Preprocessor module with potential hanging

4223

nodes, two nodes closer than a certain (small) tolerance will be

4224

merged. Hence, due to numerical round-up errors, two different machines may

4225

yield different results. This might change the number of faces in the global

3060

{\em NOTE: when joining of faces or periodicity is used, two nodes closer

3061

than a certain (small) tolerance will be merged. Hence, due to numerical

3062

round-up errors, two different machines may yield different results.

3063

This might change the number of faces in the global

4226

3064

domain\footnote{the number of cells will not be modified, it is always the sum of the

4227

3065

number of cells of the different meshes} and make restart files

4228

3066

incompatible. Should that problem arise when making a calculation restart on a

4229

different architecture, the solution is to discard the \texttt{suiavx} file and

4230

use only the \texttt{suiava} file.}

4231

4232

4233

4234

4235

%==================================

4236

%==================================

4237

\section{User subroutines}

4238

%==================================

4239

%==================================

4240

\label{prg_ssprgutilis}

4241

%==================================

4242

\subsection{Preliminary comments}

4243

%==================================

4244

The user can run the calculations with or without an interface, with or

4245

without the user subroutine. Without interface, some user subroutines

4246

are needed. With interface, all the user subroutines are optional.

4247

4248

The parameters can be read in the interface and then in the user

4249

subroutine. In the case that a parameter is specified in the interface

4250

and in the user subroutine, it is the value in the user subroutine that

4251

is taken into acount. It is for that reason that all the examples of

4252

user subroutines are placed in the \texttt{REFERENCE} directory by the

4253

case preparer \texttt{code\_saturne~create}.

4254

4255

4256

4257

%==================================

4258

\subsection{Using selection criteria in user subroutines}

3067

different architecture, the solution is to ignore the \texttt{auxiliary}

3068

file and use only the \texttt{main} file, by setting \texttt{ileaux = 0}

3069

in \texttt{usini1.f90}}

3070

3071

%==================================

3072

\subsubsection{Using selection criteria in user subroutines}

4259

3073

%==================================

4260

3074

\label{fvm_selector}

4261

3075

4343

3157

\texttt{nomgrp} of the type \texttt{character*} and its lenght

4344

3158

\texttt{lngnom} of the type \texttt{integer}) may be used.

4345

3159

4346

4347

4348

%==================================

4349

\subsection{Initialisation of the main key words: \textmd{\texttt{usini1}}}

4350

%==================================

4351

4352

\noindent

4353

\textit{Subroutine only called during calculation initialisation.}

4354

4355

This subroutine is used to indicate the value of different calculation

4356

basic parameters: constant and uniform physical values, parameters of

4357

numerical schemes, input-output management ...\\

4358

4359

In the case of a calculation launched using the interface, it is only

3160

%==================================

3161

\subsection{Face and cell mesh-defined properties and selection}

3162

%==================================

3163

\label{selection_criteria}

3164

3165

The mesh entities may be referenced by the user during the mesh

3166

creation. These references may then be used to mark out some mesh entities

3167

according to the need (specification of boundary conditions, pressure

3168

drop zones, ...). The references are generally of one of the two

3169

following types:

3170

\begin{list}{$\bullet$}{}

3171

\item color.

3172

A color is an integer possibly associated with boundary faces and

3173

volume elements by the mesh generator. Depending on the tool,

3174

this concept may have different names, which \CS interprets

3175

as colors. Most tools allow only one color per face or element.

3176

\begin{list}{-}{}

3177

\item I-deas uses a color number with a default of

3178

7 (green) for elements, be they volume elements or boundary

3179

``surface coating'' elements. Color 11 (red) is used for

3180

for vertices, but vertex properties are ignored by \CS.

3181

\item SIMAIL uses the equivalent notions of ``reference''

3182

for element faces, and ``subdomain'' for volume elements.

3183

By default, element faces are assigned no reference (0),

3184

and volume elements domain 1.

3185

\item Gmsh uses ``physical property'' numbers.

3186

\item EnSight has no similar notion, but if several parts

3187

are present in an EnSight 6 file, or several parts

3188

are present \emph{and} vertex ids are given in an

3189

Ensight Gold file, the part number is interpreted as

3190

a color number by the Preprocessor.

3191

\item The Comet Design (pro-STAR/STAR4) and NUMECA Hex file

3192

formats have a CAD section id that is interpreted

3193

as a color number. In the latter case, this notion

3194

only applies to faces, so volume elements are given

3195

color.

3196

\item The MED format allow integer ``attributes'', though

3197

many tools working with this format ignore those

3198

and only handle groups.

3199

\end{list}

3200

\item groups.

3201

Named ``groups'' of mesh entities may also be used with many

3202

mesh generators or formats. In some cases, a given cell or face may belong

3203

to multiple groups (as some tools allow new groups to be defined

3204

by boolean operations on existing groups).

3205

In \CS, every group is assigned a group number (base on alphabetical

3206

ordering of groups).

3207

\begin{list}{-}{}

3208

\item I-deas assigns a group number with each

3209

group, but by default, this number is just a counter.

3210

Only the group name is considered by \CS (so that elements

3211

belonging to two groups with identical names and different

3212

numbers are considered as belonging to the same group).

3213

\item CGNS allows both for named boundary conditions and mesh

3214

sections. If present, boundary condition names are

3215

interpreted as group names, and groups may also be defined

3216

based on element section or zone names using additional

3217

Preprocessor options (\texttt{-grp-cel} or

3218

\texttt{-grp-fac} followed by \texttt{section} or

3219

\texttt{zone}).

3220

\item Using the MED format, it is preferable to use ``groups''

3221

to colors, as many tools ignore the latter.

3222

\end{list}

3223

\end{list}

3224

3225

Selection criteria may be defined in a similar fashion whether

3226

using the GUI or in user subroutines.

3227

Typically, a selection criteria is simply a string containing

3228

the required color numbers or group names, possibly combined

3229

using boolean expressions. Simple geometric criteria are also

3230

possible.

3231

3232

A few examples are given below:

3233

3234

\verb+ENTRY+\\

3235

\verb+1 or 7+\\

3236

\verb+all[]+\\

3237

\verb+3.1 >= z >= -2 or not (15 or entry)+\\

3238

\verb+range[04, 13, attribute]+\\

3239

\verb+sphere[0, 0, 0, 2] and (not no_group[])+

3240

3241

Strings such as group names containing whitespace

3242

or having names similar to reserved operators may be protected

3243

using ``escape characters''.\footnote{Note that for defining a

3244

string in Fortran, double quotes are easier to use, as they do not

3245

conflict with Fortran's single quotes delimiting a string.

3246

In C, the converse is true. Also, in C, to define a string

3247

such as \texttt{{$\backslash$}plane}, the string

3248

\texttt{{$\backslash$}{$\backslash$}plane} must be

3249

used, as the first $\backslash$ character is used by the

3250

compiler itself. Using the GUI, either notation is easy.}

3251

More complex examples of strings whith protected strings are given here:

3252

3253

\verb+"First entry" or Wall\ or\ sym+\\

3254

\verb+entry or \plane or "noone's output"+

3255

3256

The following operators and syntaxes are allowed (fully capitalized

3257

versions of keywords are also allowed, but mixed capitals/lowercase

3258

versions are not):

3259

3260

\begin{tabular}[top]{p{6cm} l}

3261

\multicolumn{2}{l}{\bf escape characters }\\

3262

protect next character only: & \texttt{$\backslash$} \\

3263

protect string: & \texttt{{'}$string${'}} \quad \texttt{"$string$"}\\

3264

\end{tabular}

3265

3266

\begin{tabular}[top]{p{6cm} l}

3267

\multicolumn{2}{l}{\bf basic operators }\\

3268

priority: & \texttt{(} \quad \texttt{)} \\

3269

not: & \texttt{not} \quad \texttt{!} \quad \texttt{!=} \\

3270

and: & \texttt{and} \quad \texttt{\&} \quad \texttt{\&\&} \\

3271

or: & \texttt{or} \quad \texttt{|} \quad \texttt{||} \quad \texttt{,} \quad \texttt{;} \\

3272

xor: & \texttt{xor} \quad \texttt{\^} \\

3273

\end{tabular}

3274

3275

\begin{tabular}[top]{p{6cm} l}

3276

\multicolumn{2}{l}{\bf general functions }\\

3277

select all: & \texttt{all[]}\\

3278

entities having no group or color: & \texttt{no\_group[]} \\

3279

select a range of groups or colors: & \texttt{range[$first$, $last$]} \\

3280

& \texttt{range[$first$, $last$, group]} \\

3281

& \texttt{range[$first$, $last$, attribute]} \\

3282

\end{tabular}

3283

3284

For the range operator, $first$ and $last$ values are inclusive.

3285

For attribute (color) numbers, natural integer value ordering is used,

3286

while for group names, alphabetical ordering is used. Note also that in

3287

the bizarre (not recommended) case in which a mesh would contain for

3288

example both a color number 15 and a group named ``15'', using

3289

\texttt{range[15, 15, group]} or \texttt{range[15, 15, attribute]}

3290

could be used to distinguish the two.

3291

3292

Geometric functions are also available. The coordinates considered are

3293

those of the cell or face centers. Normals are of course

3294

usable only for face selections, not cell selections.

3295

3296

\begin{tabular}[top]{p{6cm} l}

3297

\multicolumn{2}{l}{\bf geometric functions }\\

3298

face normals: & \texttt{normal[$x$, $y$, $z$, $epsilon$]} \\

3299

& \texttt{normal[$x$, $y$, $z$, epsilon = $epsilon$]} \\

3300

plane, $ax + by + cz + d = 0$ form: & \texttt{plane[$a$, $b$, $c$, $d$, $epsilon$]} \\

3301

& \texttt{plane[$a$, $b$, $c$, $d$, epsilon = $epsilon$]} \\

3302

& \texttt{plane[$a$, $b$, $c$, $d$, inside]} \\

3303

& \texttt{plane[$a$, $b$, $c$, $d$, outside]} \\

3304

plane, normal + point in plane form: & \texttt{plane[$n_x$, $n_y$, $n_z$, $x$, $y$, $z$, $epsilon$]} \\

3305

& \texttt{plane[$n_x$, $n_y$, $n_z$, $x$, $y$, $z$, epsilon = $epsilon$]} \\

3306

& \texttt{plane[$n_x$, $n_y$, $n_z$, $x$, $y$, $z$, inside]} \\

3307

& \texttt{plane[$n_x$, $n_y$, $n_z$, $x$, $y$, $z$, outside]} \\

3308

box, extents form: & \texttt{box[$x_{min}$, $y_{min}$, $z_{min}$,

3309

$x_{max}$, $y_{max}$, $z_{max}$]} \\

3310

box, origin + axes form: & \texttt{box[$x_0$, $y_0$, $z_0$,}\\

3311

& \texttt{\qquad $dx_1$, $dy_1$, $dz_1$,

3312

$dx_2$, $dy_2$, $dz_2$,

3313

$dx_3$, $dy_3$, $dz_3$]} \\

3314

& \texttt{plane[$a$, $b$, $c$, $d$, epsilon = $epsilon$]} \\

3315

& \texttt{plane[$a$, $b$, $c$, $d$, inside]} \\

3316

& \texttt{plane[$a$, $b$, $c$, $d$, outside]} \\

3317

cylinder: & \texttt{cylinder[$x_0$, $y_0$, $z_0$, $x_1$, $y_1$, $z_1$, $radius$]} \\

3318

sphere: & \texttt{sphere[$x_c$, $y_c$, $z_c$, $radius$]} \\

3319

inequalities: & \texttt{>}, \texttt{<}, \texttt{>=}, \texttt{<=} associated

3320

with \texttt{x}, \texttt{y}, \texttt{z} or

3321

\texttt{X}, \texttt{Y}, \texttt{Z} keywords\\

3322

& and coordinate value; \\

3323

& \texttt{$x_{min}$ <= x < $x_{max}$} type syntax is allowed. \\

3324

\end{tabular}

3325

3326

In the current version of \CS, all selection criteria used

3327

are maintained in a list, so that re-interpreting a criterion already

3328

encountered (such as at the previous time step) is avoided.

3329

Lists of entities corresponding to a criteria containing no geometric

3330

functions are also saved in a compact manner, so re-using a previously

3331

used selection should be very fast. For criteria containing geometric

3332

functions, the full list of corresponding entities is not maintained,

3333

so each entity must be compared to the criterion at each time step.

3334

Heavy use of many selection criteria containing geometric functions

3335

may thus lead to reduced performance.

3336

3337

%==================================

3338

\section{Importing and Preprocessing Meshes}

3339

%==================================

3340

3341

Importing and preprocessing meshes is done both by the

3342

\pcs module, which is used to import meshes, and using

3343

preprocessing functions of the code Kernel.

3344

3345

The \pcs module of \CS reads the

3346

mesh file(s) (under any supported format) and translates the necessary

3347

information into a Kernel input file.

3348

3349

When multiple meshes are used, the \pcs is called once per mesh,

3350

and each resulting output is added in a \texttt{mesh\_input}

3351

directory (instead of a single \texttt{mesh\_input} file).

3352

3353

The executable of the \pcs module is \texttt{cs\_preprocess},

3354

and the most useful options and sub-options are described briefly here.

3355

To obtain a complete and up-to-date list of options and environment

3356

variables, use the following command:

3357

\texttt{cs\_preprocess~-h} or \texttt{cs\_preprocess~--help}. Many options,

3358

such as this one, accept a short and a long version.

3359

3360

For the main options, the launch script \texttt{runcase} contains

3361

corresponding variables, that are used to define options for the

3362

\pcs. This way, the user only has to define these variables

3363

and does not detailed knowledge of the \pcs command line.

3364

3365

Nonetheless, it may be useful to call the \pcs manually

3366

in certain situations, especially for frequent verification when

3367

building a mesh, so its use is described here. Verification

3368

may also be done using the GUI or the mesh quality check mode

3369

of the general run script.

3370

3371

The \pcs is controlled using command-line arguments.

3372

A few environment variables allow an expert user to modify

3373

some behaviors or to obtain a trace of memory management.

3374

3375

%==================================

3376

\subsection{\pcs options\label{sec:optpcs}}

3377

%==================================

3378

3379

Main choices are done using command-line options. For example:

3380

3381

\texttt{cs\_preprocess --num 2 fluid.med}

3382

3383

\noindent means that we read the second mesh defined in the

3384

\texttt{fluid.med} file, while:

3385

3386

\texttt{cs\_preprocess --no-write --post-volume fluid.med fluid.msh}

3387

3388

\noindent means that we read file \texttt{fluid.msh}, and

3389

do not produce a \texttt{mesh\_input} file, but do output

3390

a \texttt{fluid.med} file (effectively converting a Gmsh file to

3391

a MED file).

3392

3393

\subsubsection{Mesh selection\label{sec:optpcs:mesh}}

3394

3395

Any use of the preprocessor requires one mesh file (except for

3396

\texttt{cs\_preprocess} and \texttt{cs\_preprocess~-h} which respectively

3397

print the version number and list of options).

3398

This file is selected as the last argument to \texttt{cs\_preprocess},

3399

and its format is usually automatically determined based on its

3400

extension (c.f. \ref{sec:formats_in} page~\pageref{sec:formats_in})

3401

but a \texttt{--format} option allows forcing the format choice of

3402

the selected file.

3403

3404

For formats allowing multiple meshes in a single file, the

3405

\texttt{--num} option followed by a strictly positive integer allows

3406

selection of a specific mesh; by default, the first mesh is selected.

3407

3408

For meshes in CGNS format, we may in addition use the \texttt{--grp-cel}

3409

or \texttt{--grp-fac} options, followed by the \texttt{section}

3410

or \texttt{zone} keywords, to define additional groups of cell or faces

3411

based on the organization of the mesh in sections or zones. The sub-options

3412

have no effect on meshes of other formats.

3413

3414

\subsubsection{Post-processing output\label{sec:optpcs:post}}

3415

3416

By default, the \pcs does not generate any post-processor output.

3417

By adding \texttt{--post-volume [format]},

3418

with the optional \texttt{format} argument being one of \texttt{ensight},

3419

\texttt{med}, or \texttt{cgns} to the command-line arguments,

3420

the output of the volume mesh to the default or indicated format

3421

is provoked.

3422

3423

In case of errors, output of error visualization output is always

3424

produced, and by adding \texttt{--post-error [format]},

3425

the format of that output may be selected (from one of \texttt{ensight},

3426

\texttt{med}, or \texttt{cgns}, assuming MED and CGNS are

3427

available),

3428

3429

\subsubsection{Element orientation correction\label{sec:optpcs:orient}}

3430

3431

We may activate the possible element orientation correction using the

3432

\texttt{--reorient} option.

3433

3434

Note that we cannot guarantee correction (or even detection) of a bad

3435

orientation in all cases.

3436

Not all local numbering possibilities of elements are tested,

3437

as we focus on ``usual'' numbering permutations. Moreover,

3438

the algorithms used may produce false positives or fail to find

3439

a correct renumbering in the case of highly non convex elements.

3440

In this case, nothing may be done short of modifying the mesh, as

3441

without a convexity hypothesis, it is not always possible to choose

3442

between two possible definitions starting from a point set.

3443

3444

With a post-processing option such as \texttt{--post-error}

3445

or, \texttt{--post-volume},

3446

visualizable meshes of corrected elements as well as remaining

3447

badly oriented elements are generated.

3448

3449

\subsection{Environment variables\label{sec:envvpcs}}

3450

3451

Setting a few environment variables specific to the \pcs allows modifying

3452

its default behavior. In general, if a given behavior is modifiable

3453

through an environment variable rather than by a command-line option,

3454

it has little interest for a non-developer, or its modification is

3455

potentially hazardous. The environment variables used by the \pcs

3456

are described here:

3457

3458

\envvar{CS\_PREPROCESS\_MEM\_LOG}

3459

3460

Allows defining a file name in which memory allocation, reallocation,

3461

and freeing is logged.

3462

3463

\envvar{CS\_PREPROCESS\_MIN\_EDGE\_LEN}

3464

3465

Under the indicated length ($10^{-15}$ by default), an edge is considered

3466

to be degenerate and its vertices will be merged after the transformation

3467

to descending connectivity. Degenerate edges and faces will thus be

3468

removed. Hence, the post-processed element does not change, but the

3469

Kernel may handle a prism where the preprocessor input contained a

3470

hexahedron with two identical vertex couples (and thus a face of zero

3471

surface). If the \pcs does not print any information relative to this

3472

type of correction, it means that it has not been necessary. To completely

3473

deactivate this automatic correction, a negative value may be assigned

3474

to this environment variable.

3475

3476

\envvar{CS\_PREPROCESS\_IGNORE\_IDEAS\_COO\_SYS}

3477

3478

If this variable is defined and is a strictly positive integer, coordinate

3479

systems in \ideas universal format files will be ignored. The behavior

3480

of the \pcs will thus be the same as that of versions 1.0 and 1.1.

3481

Note that in any case, non Cartesian coordinate systems are not handled yet.

3482

3483

\subsubsection{System environment variables\label{sec:envpcs:sys}}

3484

3485

Some system environment variables may also modify the behavior of

3486

the \pcs. For example, if the \pcs was compiled with \med support

3487

on an architecture allowing shared (dynamic) libraries, the

3488

\texttt {LD\_PRELOAD} environment variable may be used to define a

3489

``prioritary'' path to load \med or HDF5 libraries, and thus experiment

3490

with another version of these libraries without recompiling the \pcs.

3491

To determine which shared libraries are used by an executable file, use

3492

the following command: \texttt{ldd~\{executable\_path\}}.

3493

3494

\subsection{Optional functionality\label{sec:pcs:lib_opt}}

3495

3496

Some functions of the \pcs are based on external libraries,

3497

which may not always be available. It is thus possible to configure

3498

and compile the \pcs so as not to use these libraries.

3499

When running the \pcs, the supported options are printed.

3500

The following optional libraries may be used:

3501

3502

\begin{itemize}

3503

3504

\item CGNS library. In its absence, \hyperref[fmtdesc:cgns]{CGNS}

3505

format support is deactivated.

3506

3507

\item \med-file library. In its absence, \hyperref[fmtdesc:med]{\med}

3508

format is simply deactivated.

3509

3510

\item Read compressed files using Zlib. With this option, it is

3511

possible to diretly read mesh files compressed with a

3512

\emph{gzip} type algorithm and bearing a \emph{.gz} extension.

3513

This is limited to formats not already based on an external

3514

library (i.e. it is not usable with CGNS or \med files),

3515

and has memory and CPU time overhead, but may be practical.

3516

Without this library, files must be uncompressed before use.

3517

3518

\end{itemize}

3519

3520

\subsection{General remarks}

3521

3522

Note that the \pcs is in general capable of reading all ``classical''

3523

element types present in mesh files (triangles, quadrangles, tetrahedra,

3524

pyramids, prisms, and hexahedra).

3525

Quadratic or cubic elements are converted upon reading into their

3526

linear counterparts. Vertices referenced by no element (isolated vertices

3527

or centers of higher-degree elements) are discarded. Meshes are read

3528

in the order defined by the user and are appended, vertex and element

3529

indices being incremented appropriately.

3530

\footnote{Possible entity labels are not maintained, as they would

3531

probably not be unique when appending multiple meshes.}

3532

3533

At this stage, volume elements are sorted by type, and the fluid domain

3534

post-processing output is generated if required.

3535

3536

In general, groups assigned to vertices are ignored.

3537

selections are thus based on faces or cells. with tools such

3538

as \simail, faces of volume elements may be referenced directly, while

3539

with \ideas or SALOME, a layer of surface elements bearing the required

3540

colors and groups must be added. Internally, the \pcs always considers

3541

that a layer of surface elements is added (i.e. when reading a \simail

3542

mesh, additional faces are generated to bear cell face colors.

3543

When building the $faces \rightarrow cells$ connectivity, all faces with the

3544

same topology are merged: the initial presence of two layers of identical

3545

surface elements belonging to different groups would thus lead to

3546

a calculation mesh with faces belonging to two groups.

3547

3548

\subsection{Files passed to the Kernel\label{sec:pcs:mode_comm}}

3549

3550

Data passed to the Kernel by the \pcs is transmitted using a

3551

binary file, using ``big endian'' data representation, named

3552

\texttt{mesh\_input} (or contained in a directory of that name).

3553

3554

When using the \pcs for mesh verification, data for the Kernel

3555

is not always needed. In this case, the \texttt{--no-write} option may be

3556

avoid creating a \pcs output file.

3557

3558

%==================================

3559

\subsection{Mesh preprocessing%

3560

\label{sec:prepro}}

3561

%==================================

3562

3563

\subsubsection{Joining of non-conforming meshes}\label{sec:optpcs:join}

3564

3565

Conforming joining of possibly non-conforming meshes may be done by the

3566

solver, and defined either using the Graphical User Interface (GUI) or the

3567

\texttt{cs\_user\_join} user function. In the GUI, the user needs to

3568

add entries in the ``Face joining'' section of the ``Meshes'' tab in the item

3569

``Calculation environment $\rightarrow$ Meshes selection''.

3570

The user may specify faces to be joined, and can also modify basic joining

3571

parameters, see fig. \ref{fig:joining}.

3572

3573

\begin{figure}[!h]

3574

\begin{center}

3575

\includegraphics[width=15cm]{gui_mesh_join}

3576

\caption{Conformal or non-conformal joining}

3577

\label{fig:joining}

3578

\end{center}

3579

\end{figure}

3580

3581

For a simple mesh, it is rarely useful to specify strict face selection

3582

criteria, as joining is sufficiently automated to detect which faces

3583

may actually be joined. For a more complex mesh, or a mesh with thin

3584

walls which we want to avoid transforming into interior faces, it is

3585

recommended to filter boundary faces that may be joined by using

3586

face selection criteria. This has the

3587

additional advantage of reducing the number of faces to test for

3588

in the intersection/overlap search, and thus reduced to time

3589

required by the joining algorithm.

3590

3591

One may also modify tolerance criteria using 2 options:

3592

3593

\noindent

3594

\begin{tabular}[top]{p{3.0cm}%

3595

>{\PreserveBackslash\raggedright\hspace{0pt}}p{12.0cm}}

3596

\texttt{fraction} $r$ &

3597

assigns value $r$ (where $0 < r < 0,49$) to the maximum

3598

intersection distance multiplier ($0,1$ by default). The maximum

3599

intersection distance for a given vertex is based on the length of

3600

the shortest incident edge, multiplied by $r$. The maximum intersection

3601

at a given point along an edge is interpolated from that at its

3602

vertices, as shown on the left of figure \ref{fig_join_tolerance}; \\

3603

\texttt{lane} $c$ &

3604

assigns the maximum angle between normals for two faces to be

3605

considered coplanar ($25�$ by default);

3606

this parameter is used in the second stage of the algorithm, to

3607

reconstruct conforming faces, as shown on the right of figure

3608

\ref{fig_join_tolerance}.\\

3609

\end{tabular}

3610

3611

\begin{figure}[!hp]

3612

\centerline{

3613

\includegraphics*[width=14cm]{join_tolerance}}

3614

\caption{Maximum intersection tolerance and faces normal angle

3615

\label{fig_join_tolerance}}

3616

\end{figure}

3617

3618

In practice, we are sometimes led to increase the maximum intersection

3619

distance multiplier to $0.2$ or even $0.3$ when joining curved surfaces,

3620

so that all intersection are detected. As this influences merging

3621

of vertices and thus simplification of reconstructed faces, but also

3622

deformation of ``lateral'' faces, it is recommended only to modify it

3623

if necessary. As for the \texttt{plane} parameter, its use has

3624

only been necessary on a few meshes up to now, and always in the

3625

sense of reducing the tolerance so that face reconstruction does not

3626

try to generate faces from initial faces on different surfaces.

3627

3628

\subsubsection{Periodicity\label{sec:optpcs:period}}

3629

3630

Handling of periodicity is based on an extension of conforming joining,

3631

as shown on figure \ref{fig_join_periodic}. It is thus not necessary

3632

for the periodic faces to be conforming (though it usually leads to better

3633

mesh quality). All options relative to conforming joining of

3634

non-conforming faces also apply to periodicity. Note also that

3635

once pre-processed, 2 periodic faces have the same orientation

3636

(possibly adjusted by periodicity of rotation).

3637

3638

This operation can also be performed by the solver and specified

3639

either using the GUI or the \texttt{cs\_user\_periodicity} function.

3640

3641

\begin{figure}[!hp]

3642

\centerline{

3643

\includegraphics*[height=8cm]{join_periodic}}

3644

\caption{Matching of periodic faces

3645

\label{fig_join_periodic}}

3646

\end{figure}

3647

3648

As with joining, it is recommended to filter boundary faces to process

3649

using a selection criterion. As many periodicities may be built as desired,

3650

as long as boundary faces are present. One a periodicity is handled,

3651

faces having periodic matches do not appear as boundary faces, but as

3652

interior faces, and are thus not available anymore for other

3653

periodicities.

3654

3655

%==================================

3656

\subsubsection{Parameters for conforming or non-conforming mesh joinings}

3657

%==================================

3658

3659

The setting of these parameters is done in the user subroutine \texttt{cs\_user\_join} (called once). The user can specify the parameters used for the joining of different meshes. Below is given the list of the standard parameters which can me modified:

3660

\begin{list}{-}{}

3661

\item \texttt{fract}: the initial tolerance radius associated to each vertex is equal to the lenght of the shortest incident edge, multiplied by this fraction,

3662

\item \texttt{plane}: when subdividing faces, 2 faces are considered coplanar and may be joined if the angle between their unit normals (in degree) does not exceed this parameter,

3663

\item \texttt{iwarnj}: the associated verbosity level (debug level if over 3).

3664

\end{list}

3665

In the call of the function \texttt{cs\_join\_add}, a selection criteria for

3666

mesh faces to be joined is specified.

3667

The list of advanced modifiable parameters is given below:

3668

\begin{list}{-}{}

3669

\item \texttt{mtf}: a merge tolerance factor, used to locally modify the tolerance associated to each vertex before the merge step. Depending on its value four scenarii are possible:

3670

\begin{list}{$\rightarrow$}{}

3671

\item if $mtf=0$, no vertex merge

3672

\item if $mtf<1$, the vertex merge is more strict. It may increase the number of tolerance reduction and therfore define smaller subset of vertices to merge together but it can drive to loose intersections.

3673

\item if $mtf=1$, no change occurs

3674

\item if $mtf>1$, the vertex merge is less strict. The subset of vertices able to merge is greater.

3675

\end{list}

3676

\item \texttt{pmf}: a pre-merge factor. This parameter is used to define a limit under which two vertices are merged before the merge step,

3677

\item \texttt{tcm}: a tolerance computation mode. If its value is:

3678

\begin{list}{$\rightarrow$}{}

3679

\item 1 (default), the tolerance is the minimal edge length related to a vertex, multiplied by a fraction.

3680

\item 2, the tolerance is computed like for 1 with, in addition, the multiplication by a coefficient equal to the maximum between $sin(e1)$ and $sin(e2)$; where $e1$ and $e2$ are two edges sharing the same vertex V for which we want to compute the tolerance.

3681

\item 11, it is the same as 1 but taking into account only the selected faces.

3682

\item 12, it is the same as 2 but taking into account only the selected faces.

3683

\end{list}

3684

\item \texttt{icm}: the intersection computation mode. If its value is:

3685

\begin{list}{$\rightarrow$}{}

3686

\item 1 (default), the original algorithm is used. Care should taken to clip the intersection on an extremity.

3687

\item 2, a new intersection algorithm is used. Caution should be used to avoid to clip the intersection on an extremity.

3688

\end{list}

3689

\item \texttt{maxbrk}: defines the maximum number of equivalence breaks which is enabled during the merge step,

3690

\item \texttt{maxsf}: defines the maximum number of sub-faces used when splitting a selected face

3691

\end{list}

3692

3693

The followings are advanced parameters used in the search algorithm for face intersections between selected faces (octree structure). They are useful in case of memory limitation:

3694

\begin{list}{-}{}

3695

\item \texttt{tml}: the tree maximum level is the deepest level reachable during the tree building,

3696

\item \texttt{tmb}: the tree maximum boxes is the maximum number of bounding boxes (BB) which can be linked to a leaf of the tree (not necessary true for the deepest level),

3697

\item \texttt{tmr}: the tree maximum ratio. The building of the tree structure stops when the number of bounding boxes is superior to the product of \texttt{tmr} with the number of faces to locate. This is an efficient parameter to reduce memory consumption.

3698

\end{list}

3699

The call to the subroutine '\texttt{setajp}' returns the value of these parameters.

3700

3701

%==================================

3702

\subsubsection{Parameters for the periodicity}

3703

%==================================

3704

3705

Periodicities can be set directly in the Graphical User Interface (GUI) or using the user subroutine \texttt{cs\_user\_periodicity} (called when once during the calculation initialisation). In the GUI, the user can choose between 3 types of periodicities: translation, rotation, or mixed (see fig. \ref{fig:periodicities}).

3706

Then specific parameters must be set.

3707

3708

\begin{figure}[!h]

3709

\begin{center}

3710

\includegraphics[width=14cm]{gui_mesh_periodicity}

3711

\caption{Periodicity}

3712

\label{fig:periodicities}

3713

\end{center}

3714

\end{figure}

3715

3716

\texttt{cs\_user\_periodicity} can be used instead of the GUI, it allows also the user to specify the parameters used to set periodicities and gives access to more advanced parameters. Below is given the list of the main parameters which can me modified:

3717

\begin{list}{-}{}

3718

\item \texttt{fract}: the initial tolerance radius associated to each vertex is equal to the lenght of the shortest incident edge, multiplied by this fraction,

3719

\item \texttt{plane}. When subdividing faces, 2 faces are considered as coplanar and may be joined if the angle between their unit normals (in degree) does not exceed this parameter,

3720

\item \texttt{iwarnj}: the associated verbosity level (debug level if over 3).

3721

\end{list}

3722

The second part of the subroutine is used to define the periodic transformations. The user provides in the subroutine '\texttt{defpro}' the reference of the mesh the transformation applies to, as well as:

3723

\begin{list}{-}{}

3724

\item the translation vector, if a periodicity of translation is used,

3725

\item the axis, the angle of rotation, and an invariant point if a periodicity of rotation is used,

3726

\item an homogeneous matrix if a general transformation is used.

3727

\end{list}

3728

In addition, the user can modify advanced parameters in case problems occur during the joining step, or to get a better mesh quality:

3729

\begin{list}{-}{}

3730

\item \texttt{mtf}: a merge tolerance factor, used to locally modify the tolerance associated to each vertex before the merge step. Depending on its value four scenarii are possible:

3731

\begin{list}{$\rightarrow$}{}

3732

\item if $mtf=0$, there is no vertex merge.

3733

\item if $mtf<1$, the vertex merge is more strict. It may increase the number of tolerance reduction and therfore define smaller subset of vertices to merge together, but it can drive to loose intersections.

3734

\item if $mtf=1$, no changes occur.

3735

\item if $mtf>1$, the vertex merge is less strict. The subset of vertices able to merge is greater.

3736

\end{list}

3737

\item \texttt{pmf}: a pre-merge factor. This parameter is used to define a limit under which two vertices are merged before the merge step,

3738

\item \texttt{tcm}: a tolerance computation mode. If its value is:

3739

\begin{list}{$\rightarrow$}{}

3740

\item 1 (default), the tolerance is the minimal edge length related to a vertex, multiplied by a fraction.

3741

\item 2, the tolerance is computed like for 1 with, in addition, the multiplication by a coefficient equal to the maximum between $sin(e1)$ and $sin(e2)$, where $e1$ and $e2$ are two edges sharing the same vertex V for which we want to compute the tolerance.

3742

\item 11, it is the same as 1 but taking into account only the selected faces.

3743

\item 12, it is the same as 2 but taking into account only the selected faces.

3744

\end{list}

3745

\item \texttt{icm}: the intersection computation mode. If its value is:

3746

\begin{list}{$\rightarrow$}{}

3747

\item 1 (default), the original algorithm is used. Care should taken to clip the intersection on an extremity.

3748

\item 2, a new intersection algorithm is used. Caution should be used to avoid to clip the intersection on an extremity.

3749

\end{list}

3750

\item \texttt{maxbrk}: defines the maximum number of equivalence breaks which are enabled during the merge step,

3751

\item \texttt{maxsf}: defines the maximum number of sub-faces used when splitting a selected face

3752

\end{list}

3753

3754

The following are advanced parameters used in the search algorithm for face intersections between selected faces (octree structure). There are useful in case of memory limitation:

3755

\begin{list}{-}{}

3756

\item \texttt{tml}: the tree maximum level is the deepest level reachable during the tree building

3757

\item \texttt{tmb}: the tree maximum boxes is the maximum number of bounding boxes (BB) which can be linked to a leaf of the tree (not necessary true for the deepest level)

3758

\item \texttt{tmr}: the tree maximum ratio. The building of the tree structure stops when the number of bounding boxes is superior than the product of \texttt{tmr} with the number of faces to locate. This is an efficient parameter to reduce memory consumption.

3759

\end{list}

3760

The call to the routine '\texttt{setapp}' returns the value of these parameters.

3761

3762

%==================================

3763

\subsubsection{Modification of the mesh geometry}

3764

%==================================

3765

\noindent

3766

\textit{Subroutines called only during the calculation initialisation.}

3767

3768

The user subroutine \texttt{cs\_user\_mesh\_input} allows a detailed

3769

selection of imported meshes read, reading files multiple times,

3770

applying geometric transformations, and renaming groups.

3771

3772

The user subroutine \texttt{cs\_user\_mesh\_modify} may be used

3773

for advanced modification of the main \texttt{cs\_mesh\_t}} structure.

3774

3775

{\em WARNING: Caution must be exercised when using this subroutine

3776

along with periodicity. Indeed, the periodicity parameters are not

3777

updated accordingly, meaning that the periodicity may be unadapted

3778

after one changes the mesh vertex coordinates. It is particularly

3779

true when one rescales the mesh. Rescaling should thus be done

3780

in a separate run, before defining periodicity.}

3781

3782

The user subroutine \texttt{cs\_user\_mesh\_thinwall} allows

3783

insertion of thin walls in the calculation mesh.

3784

Faces on each side of a thin wall will share the same vertices,

3785

so postprocessing of the main volume mesh may not show the

3786

inserted walls, thouhg they will appear in the main boundary

3787

output meshmay be used to modify

3788

``manually'' the mesh vertices coordinates, \textit{i.e.}

3789

the array \mbox{\texttt{mesh->vtx\_coord[]}}.

3790

3791

%==================================

3792

\section{Partitioning for parallel runs\label{sec:parall}}

3793

%==================================

3794

3795

Graph partitioning (using one of the optional \metis or

3796

\scotch libraries) is done using a secondary executable,\\

3797

\texttt{cs\_partition},

3798

which reads the file produced by the \pcs and builds one or

3799

several ``cell $\rightarrow$ domain'' distribution files, named

3800

{\tt domain\_number\_\it{p}} for a partitioning on $p$ sub-domains.

3801

3802

This separation leads to extra work for the Kernel, which must

3803

redistribute data read in {\tt mesh\_input}

3804

based on the associated partitioning, but avoids requiring

3805

re-running the \pcs whenever running on a different number of

3806

processors.

3807

3808

Without partitioning (for example if neither \metis nor

3809

\scotch is available, or the partitioner has not been run

3810

for the required number of sub-domains), the Kernel will

3811

use a built-in partitioning using a space-filling curve

3812

(Z-curve) technique. This usually leads to partitionings

3813

of lower quality than with graph partitioning, but parallel

3814

performance remains reasonable.

3815

3816

\subsection{Options\label{sec:optcmd:parall}}

3817

3818

To list the partitioner's options, use the following command:

3819

{\tt cs\_partition~-h}

3820

3821

We provide the list of required partitionings an optionally additional

3822

options. For example, to simulate a partitioning for calculations on

3823

64 and 128 processes with no output, we may use the following command:

3824

3825

\texttt{cs\_partition 64 128 --no-write}

3826

3827

\subsubsection{Ignore periodicity\label{sec:optcmdpart:noperiod}}

3828

3829

By default, face periodicity relations are taken into account when building

3830

the ``cell $\rightarrow$ cell'' connectivity graph used for partitioning.

3831

This allows better partitioning optimization, but increases the probability

3832

of having groups of cells at opposite sides of the domain in a same

3833

sub-domain. This is not an issue for standard calculations, but may

3834

degrade performance of search algorithms based on bounding boxes.

3835

It is thus possible to ignore periodicity when partitioning a mesh

3836

using the \texttt{--no-perio} option.

3837

3838

Note that nothing guarantees that a graph partitioner will not place

3839

disjoint cells in the same sub-domain independently of this option,

3840

but this behavior is rare.

3841

3842

\subsubsection{Partitioner choice\label{sec:optpart:partlib}}

3843

3844

If the Partitioner has been configured with both \metis and

3845

\scotch libaries, using the \texttt{--metis} or \texttt{--scotch}

3846

option allows choosing between either library. By default, \emph{metis} is

3847

used if both choices are available.

3848

3849

\subsubsection{Simulation mode\label{sec:optpart:nowrite}}

3850

3851

Using the \texttt{--no-write} option, we can tell the partitioner

3852

not to output a {\tt domain\_number\_\it{p}} file.

3853

Partitioning is thus computed, but not saved.

3854

3855

\subsubsection{Environment variables\label{sec:optpart:envvar}}

3856

3857

\envvar{CS\_PARTITION\_MEM\_LOG}

3858

3859

Allows defining a file name in which memory allocations, reallocations,

3860

and frees will be logged.

3861

3862

3863

%==================================

3864

\section{Basic modelling setup}

3865

%==================================

3866

3867

%==================================

3868

\subsection{Initialisation of the main parameters}

3869

%==================================

3870

3871

This operation is done in the Graphical User Interface (GUI) or by using the user subroutine \texttt{usini1}.

3872

In the GUI, the initialisation is performed by filling the parameters displayed in Figs. \ref{fig:6_GUI} to \ref{fig:28_GUI}. If the option 'Mobile mesh' is activated in Fig. \ref{fig:6_GUI}, please see Section \ref{sec:ALE} for more details. In fig. \ref{fig:11_GUI}, the equivalent initialisations occur in the subroutine \texttt{usiniv} when the GUI is not used.

3873

The headings filled for the initialisation are the followings:

3874

\begin{list}{-}{}

3875

\item Thermophysical model options: ALE mobile mesh,

3876

turbulence model, thermal model, see figs. \ref{fig:6_GUI} to \ref{fig:8_GUI}.

3877

\item Additional scalars: definition, initialisation of the scalars, and physical characteristics, see figs. \ref{fig:11_GUI} and \ref{fig:12_GUI}. In fig. \ref{fig:12_GUI}, the initial values are given in the subroutine \texttt{usiniv} if the GUI is not used, see Section \ref{sec:usiniv}.

3878

\item Physical properties: reference pressure, fluid characteristics, gravity, see figs. \ref{fig:13_GUI} to \ref{fig:15_GUI}. If non-constant values are used for the fluid properties, and if the GUI is not used, see Section \ref{sec:usphyv}.

3879

\item Numerical parameters: number and type of time steps, and advanced parameters for the numerical solution of the equations, see figs. \ref{fig:21_GUI} to \ref{fig:23_GUI}.

3880

\item Calculation control: parameters related to the time averages, the time step, the

3881

locations of the probes where some variables will be monitored over time

3882

(if the GUI is not used, this information is specified in Section

3883

\ref{sec:usiniv}), the definition of the frequency of the outputs in the calculation

3884

listing, the postprocessing output writer frequency and format options, and

3885

the postprocessing output meshes and variables selection, see

3886

figs. \ref{fig:24_GUI}, \ref{gui_output_log}, \ref{gui_output_writers},

3887

and \ref{gui_output_meshes}. The item ``Profiles'' allows to save, with a

3888

frequency defined by the user, 1D profiles on an axis defined by two

3889

points, see fig. \ref{fig:26_GUI}.

3890

\end{list}

3891

3892

\begin{figure}[!ht]

3893

\begin{center}

3894

\includegraphics[width=13cm]{gui_mobile_mesh}

3895

\caption{Mobile mesh option}

3896

\label{fig:6_GUI}

3897

\end{center}

3898

\end{figure}

3899

3900

\begin{figure}[!ht]

3901

\begin{center}

3902

\includegraphics[width=13cm]{gui_turbulence_models}

3903

\caption{Turbulence model selection}

3904

\label{fig:7_GUI}

3905

\end{center}

3906

\end{figure}

3907

3908

\begin{figure}[!ht]

3909

\begin{center}

3910

\includegraphics[width=13cm]{gui_thermal_scalar}

3911

\caption{Thermal scalar selection}

3912

\label{fig:8_GUI}

3913

\end{center}

3914

\end{figure}

3915

3916

\begin{figure}[!ht]

3917

\begin{center}

3918

\includegraphics[width=13cm]{gui_user_scal_def_init}

3919

\caption{Definition and initialisation of the scalars}

3920

\label{fig:11_GUI}

3921

\end{center}

3922

\end{figure}

3923

3924

\begin{figure}[!ht]

3925

\begin{center}

3926

\includegraphics[width=13cm]{gui_user_scal_phys_props}

3927

\caption{Associated physical properties of the scalars}

3928

\label{fig:12_GUI}

3929

\end{center}

3930

\end{figure}

3931

3932

\begin{figure}[!ht]

3933

\begin{center}

3934

\includegraphics[width=13cm]{gui_phys_prop_ref_pressure}

3935

\caption{Setting of the reference pressure}

3936

\label{fig:13_GUI}

3937

\end{center}

3938

\end{figure}

3939

3940

\begin{figure}[!ht]

3941

\begin{center}

3942

\includegraphics[width=13cm]{gui_fluid_props}

3943

\caption{Fluid properties}

3944

\label{fig:14_GUI}

3945

\end{center}

3946

\end{figure}

3947

3948

\begin{figure}[!ht]

3949

\begin{center}

3950

\includegraphics[width=13cm]{gui_gravity_hyd_pressure}

3951

\caption{Settings of the gravity and of the hydrostatic pressure}

3952

\label{fig:15_GUI}

3953

\end{center}

3954

\end{figure}

3955

3956

\begin{figure}[!ht]

3957

\begin{center}

3958

\includegraphics[width=13cm]{gui_time_step}

3959

\caption{Time step settings}

3960

\label{fig:21_GUI}

3961

\end{center}

3962

\end{figure}

3963

3964

\begin{figure}[!ht]

3965

\begin{center}

3966

\includegraphics[width=13cm]{gui_numerical_parameters}

3967

\caption{Numerical parameters for the main variables resolution}

3968

\label{fig:22_GUI}

3969

\end{center}

3970

\end{figure}

3971

3972

\begin{figure}[!ht]

3973

\begin{center}

3974

\includegraphics[width=13cm]{gui_global_res_parameters}

3975

\caption{Global resolution parameters}

3976

\label{fig:23_GUI}

3977

\end{center}

3978

\end{figure}

3979

3980

\begin{figure}[!ht]

3981

\begin{center}

3982

\includegraphics[width=13cm]{gui_time_averages}

3983

\caption{Management of time averaged variables}

3984

\label{fig:24_GUI}

3985

\end{center}

3986

\end{figure}

3987

3988

\begin{figure}[!ht]

3989

\begin{center}

3990

\includegraphics[width=13cm]{gui_output_log}

3991

\caption{Parameters of chronological logging options}

3992

\label{gui_output_log}

3993

\end{center}

3994

\end{figure}

3995

3996

\begin{figure}[!ht]

3997

\begin{center}

3998

\includegraphics[width=13cm]{gui_output_writers}

3999

\caption{Management of postprocessing writers}

4000

\label{gui_output_writers}

4001

\end{center}

4002

\end{figure}

4003

4004

\begin{figure}[!ht]

4005

\begin{center}

4006

\includegraphics[width=13cm]{gui_output_meshes}

4007

\caption{Management of postprocessing meshes}

4008

\label{gui_output_meshes}

4009

\end{center}

4010

\end{figure}

4011

4012

\begin{figure}[!ht]

4013

\begin{center}

4014

\includegraphics[width=13cm]{gui_output_1d_profiles}

4015

\caption{Management of 1D profiles of the solution}

4016

\label{fig:26_GUI}

4017

\end{center}

4018

\end{figure}

4019

4020

In the case of a calculation launched using the interface, the subroutine \texttt{usini1} is only

4360

4021

used to modify high-level parameters which can not be managed by the

4361

4022

interface. In the case of a code utilisation without interface, this

4362

subroutine is compulsory and all the headings must be completed.

4023

subroutine is compulsory and all the headings must be completed. \texttt{usini1} is used to indicate the value of different calculation

4024

basic parameters: constant and uniform physical values, parameters of

4025

numerical schemes, input-output management...\\

4026

It is called only during the calculation initialisation.

4363

4027

4364

4028

For more details about the different parameters, please refer to the key

4365

4029

word list (\S\ref{prg_motscles}).

4366

4030

4367

\texttt{usini1.f90} is in fact a gouping of 6 sperate subroutines: \texttt{usipph},

4031

\texttt{usini1} is in fact constituted of 6 seperate subroutines: \texttt{usipph},

4368

4032

\texttt{usinsc}, \texttt{usipsc}, \texttt{usipgl},

4369

\texttt{usipsu}and \texttt{usipes}. Each one controls the management of various

4370

specific parameters. The key words that dont' feature in the supplied example

4371

can be provided by the user in \texttt{SRC/REFERENCE/base}; in this case,

4372

understanding of the comments is needed to add the key words in the appropriate

4373

subroutine (the most widely used is \texttt{iphas}, it will assure that the value

4033

\texttt{usipsu} and \texttt{usipes}. Each one controls various

4034

specific parameters. The key words which are not featured in the supplied example

4035

can be provided by the user in \texttt{SRC/REFERENCE/base}; in this case,

4036

understanding of the comments is required to add the key words in the appropriate

4037

subroutine, it will ensure that the value

4374

4038

has been well defined). The modifiable parameters in each of the subroutines of

4375

\texttt{usini1.f90} are:

4039

\texttt{usini1} are:

4376

4040

4377

4041

\begin{list}{$\bullet$}{}

4378

4042

\item \texttt{usipph}: \texttt{iturb} and \texttt{icp} (don't modify these

4382

4046

\item \texttt{usipsc}: \texttt{iscavr} and \texttt{ivisls} (don't modify these

4383

4047

parameters anywhere else)

4384

4048

\item \texttt{usipgl}: \texttt{idtvar}, \texttt{ipucou}, \texttt{iphydr} and the

4385

parameters related to the error estimators(don't modify these parameters

4049

parameters related to the error estimators (don't modify these parameters

4386

4050

anywhere else).

4387

4051

\item \texttt{usipsu}: physical parameters of the calculation (thermal scalar, physical

4388

properties,...), numeric parameters (time steps, number of iterations, ...),

4052

properties,...), numerical parameters (time steps, number of iterations, ...),

4389

4053

definition of the time averages.

4390

\item \texttt{usipes}: post processing output parameters (periodicity, variable names,

4054

\item \texttt{usipes}: post-processing output parameters (periodicity, variable names,

4391

4055

probe positions,...)

4392

4056

\end{list}

4393

4057

4394

For more details of the different parameters, see the list of key words (\S\ref{prg_motscles}).

4395

The names of the key words can also be seen in the helps sections of the interface.

4058

For more details on the different parameters, see the list of key words (\S\ref{prg_motscles}).

4059

The names of the key words can also be seen in the help sections of the interface.

4396

4060

4397

4061

\minititre{Notes}

4398

$\bullet\ $ Determined in the list of \texttt{nscaus} user scalars, representing

4399

the mean square fluctuations of another whilst informing the \texttt{iscavr}

4400

array (warning, this was not the case in version 1.0). For

4401

the other scalars, \texttt{iscavr} does not need to be completed (by default,

4062

$\bullet\ $ The table \texttt{iscavr} is filled with the user scalars which represent the mean square fluctuations of another scalar amongst the list of the \texttt{nscaus} scalars (warning, this was not the case in version 1.0). For the other scalars, \texttt{iscavr} does not need to be completed (by default,

4402

4063

\texttt{iscavr(ii)}$\leqslant$0). For instance, if the scalar \texttt{jj}

4403

4064

represents the average of the square of the fluctuations of the scalar \texttt{kk},

4404

4065

the user must indicate \texttt{iscavr(jj)=kk}

4406

4067

4407

4068

\noindent

4408

4069

$\bullet\ $ When using the interface, only the

4409

supplementary parameters (which can not be defined in the interface)

4070

additional parameters (which can not be defined in the interface)

4410

4071

should appear in \texttt{usini1}. To spare the user the necessity to

4411

delete the other parameters appearing as examples in the subroutine, the

4412

utility program \texttt{code\_saturne~create} comments automatically all the

4413

example lines of \texttt{usini1} with a code \texttt{!ex}. The user

4414

needs then only to uncomment the lines which are useful in his

4072

delete the other parameters given as examples in the subroutine, the

4073

setup program \texttt{code\_saturne~create} comments automatically all the

4074

example lines of \texttt{usini1} with a code ``\texttt{!ex}''. The user

4075

needs then only to remove comments at the lines which are useful for his

4415

4076

case. This function of

4416

4077

\texttt{code\_saturne~create} can be deactivated with

4417

the option \texttt{--nogui} (useful if the user knows that he will not

4078

the option ``\texttt{--nogui}'' (useful if the user knows that he will not

4418

4079

use the interface).

4419

4080

4420

%==================================

4421

\subsection{Management of boundary conditions: \textmd{\texttt{usclim}}}

4081

\subsection{Selection of mesh inputs: \textmd{\texttt{cs\_user\_mesh\_input}}}

4422

4082

%==================================

4423

4083

4424

4084

\noindent

4425

\textit{Subroutine called every time step.}

4426

4427

It is the second compulsory subroutine for every calculation launched

4428

without interface(except in the specific physics case where the

4085

\textit{Subroutine called only during the calculation initialisation.}

4086

4087

This C function may be used to select which mesh input files

4088

are read, and apply optional coordinate transformations or group renumberings

4089

to them. By default, the input read is a file or directory named

4090

\texttt{mesh\_input}, but if this function is used, any file may be selected,

4091

and the same file may be read multiple times (applying a different

4092

coordinate transformation each time).

4093

All inputs read through this function are automatically concatenated, and

4094

may be later joined using the mesh joining options.

4095

4096

Geometric transformations are defined using a homogeneous coordinates

4097

transformation matrix. Such a matrix has 3 lines and 4 columns, with the

4098

3 first columns describing a rotation/scaling factor, and the last column

4099

describing a translation. A 4th line is implicit, containing zeroes

4100

off-diagonal, and 1 on the diagonal. The advantages of this representation

4101

is that any rotation/translation/scaling combination may be expressed

4102

by matrix multiplication, while simple rotations or translations

4103

may still be defined easily.

4104

4105

%==================================

4106

%==================================

4107

\subsection{Non-default variables initialisation} \label{sec:usiniv}

4108

%==================================

4109

4110

The non-default variables initialisation is performed in the subroutine \texttt{usiniv} (called only during the calculation initialisation).\\ At the calculation beginning, the variables are initialised

4111

automatically by the code. Velocities and scalars are set to 0 (or \texttt{scamax} or \texttt{scamin} if 0 is outside the acceptable

4112

scalar variation range), and the turbulent variables are estimated from

4113

\texttt{uref} and \texttt{almax}. \\

4114

For the $k$ in $k-\varepsilon$, $R_{ij}-\varepsilon$, v2f or $k-\omega$

4115

model:\\

4116

{\texttt{rtp(iel,ikiph) = 1.5*(0.02*uref)**2}

4117

(in $R_{ij}-\varepsilon$, $R_{ij}=\frac{2}{3}k\delta_{ij}$)\\

4118

For the $\varepsilon$ in $k-\varepsilon$, $R_{ij}-\varepsilon$ or v2f model:\\

4119

\texttt{rtp(iel,ieiph) = rtp(iel,ikiph)**1.5*cmu/almax}\\

4120

For $\omega$ in the $k-\omega$ model:\\

4121

\texttt{rtp(iel,iomgip) = rtp(iel,ikiph)**0.5/almax}\\

4122

For $\varphi$ and $\overline{f}$ in the v2f model:\\

4123

\texttt{rtp(iel,iphiph) = 2/3}\\

4124

\texttt{rtp(iel,ifbiph) = 0}

4125

4126

The subroutine \texttt{usiniv} allows if necessary to initialise certain

4127

variables to values closer to their estimated final values, in order to

4128

obtain a faster convergence.

4129

4130

This subroutine allows also to make a non-standard initialisation of

4131

physical parameters (density, viscosity, ...), to impose a local

4132

value of the time step, or to modify some parameters (time step,

4133

variable specific heat, ...) in the case of a calculation restart.

4134

4135

\minititre{Note: value of the time step}

4136

\begin{list}{-}{}

4137

\item For calculations with constant and uniform time step

4138

(\texttt{idtvar}=0), the value of the time step is \texttt{dtref},

4139

given in the parametric file of the interface or \texttt{usini1}.

4140

\item For calculations with a non-constant time step

4141

(\texttt{idtvar}=1 or 2) which is not a calculation restart,

4142

the value of \texttt{dtref} given in the parametric file of the interface

4143

or in \texttt{usini1} is used to initialise the time step.

4144

\item For calculations with a non-constant time step

4145

(\texttt{idtvar}=1 or 2) which is a restart of a

4146

calculation whose time step type was different (for instance, restart

4147

using a variable time step of a calculation run using a constant time

4148

step), the value of \texttt{dtref}, given in the parametric file of the

4149

interface or in \texttt{usini1}, is used to initialise the time step.

4150

\item For calculations with non-constant time step

4151

(\texttt{idtvar}=1 or 2) which is a restart of a

4152

calculation whose time step type was the same (for instance, restart with

4153

\texttt{idtvar}=1 of a calculation run with \texttt{idtvar}=1), the time

4154

step is read from the restart file and the value of \texttt{dtref} given

4155

in the parametric file of the interface, or in \texttt{usini1}, is not used.

4156

\end{list}

4157

It follows, that for a calculation with a non-constant time step (\texttt{idtvar}=1

4158

or 2) which is a restart of a calculation in which

4159

\texttt{idtvar} had the same value, \texttt{dtref} does not allow to modify the

4160

time step. The user subroutine \texttt{usiniv} allows to modify the array

4161

\texttt{dt} which contains the value of the time step read from the restart file

4162

(array whose size is \texttt{ncelet}, defined at the cell centers whatever the

4163

chosen time step type is).

4164

4165

{\em WARNING: to initialise the variables in the framework of a

4166

specific physics module} (\texttt{nscapp.gt.0}),

4167

one of the subroutines

4168

\texttt{usebui}, \texttt{usd3pi}, \texttt{uslwci} or \texttt{uscpiv}

4169

should be used (depending on the activated module) instead of \texttt{usiniv}.

4170

4171

%==================================

4172

\subsection{Manage boundary conditions}

4173

%==================================

4174

The boundary conditions can be specified in the Graphical User Interface (GUI) under the heading ``Boundary conditions'' or in the user subroutine \texttt{usclim} called every time step. With the GUI, each region and the type of boundary condition associated to it are defined in fig. \ref{fig:19_GUI}. Then, the parameters of the boundary condition are specified in fig. \ref{fig:20_GUI}. The colors of the boundary faces may be read directly from a ``listing'' file created by the Preprocessor. This file can be generated directly by the interface under the heading ``Definition of boundary regions $\rightarrow$ Add from Preprocessor listing $\rightarrow$ import groups and references from Preprocessor listing'', see fig. \ref{fig:19_GUI}.

4175

4176

\begin{figure}[!ht]

4177

\begin{center}

4178

\includegraphics[width=13cm]{gui_bc_regions}

4179

\caption{Definition of the boundary conditions}

4180

\label{fig:19_GUI}

4181

\end{center}

4182

\end{figure}

4183

4184

\begin{figure}[!ht]

4185

\begin{center}

4186

\includegraphics[width=13cm]{gui_bc_parameters}

4187

\caption{Parameters of the boundary conditions}

4188

\label{fig:20_GUI}

4189

\end{center}

4190

\end{figure}

4191

4192

\texttt{usclim} is the second compulsory subroutine for every calculation launched

4193

without interface (except in the case of specific physics where the

4429

4194

corresponding boundary condition user subroutine must be used) \\

4430

When the interface is used, \texttt{usclim} is used to define complex

4195

When the subroutine is used, \texttt{usclim} is used to define complex

4431

4196

boundary conditions (input profiles, conditions varying in time, ...)

4432

which could not be specified by means of the interface, and only these

4197

which could not be specified by the means of the interface, and only these

4433

4198

need to be defined. In the case of a calculation launched without the

4434

4199

interface, all the boundary conditions must appear in \texttt{usclim}.\\

4435

4200

\texttt{usclim} is essentially constituted of loops on boundary

4436

4201

face subsets. Several sequences

4437

4202

of \verb+call getfbr+ \verb+('criterion', nlelt, lstelt)+ (cf.

4438

\S\ref{fvm_selector}) allow to differentiate

4439

the boundary faces according to their group(s), their

4203

\S\ref{fvm_selector}) allow to select

4204

the boundary faces with respect to their group(s), their

4440

4205

color(s) or geometric criteria. If needed, geometric and

4441

physical variables are also available to the user, these allow him

4442

to differentiate the boundary faces using other criteria.

4206

physical variables are also available to the user. These allow him

4207

to select the boundary faces using other criteria.

4443

4208

4444

4209

For more details about the treatment of boundary conditions, the user

4445

4210

may refer to the theoretical and computer documentation \cite{theory} of

4446

the subroutine \texttt{condli} (for the wall conditions, see

4447

\texttt{clptur}) (to access to this document on a workstation, use

4211

the subroutine \texttt{condli} (for wall conditions, see

4212

\texttt{clptur}) (to access this document on a workstation, use

4448

4213

\mbox{\texttt{code\_saturne~info --guide theory}}).

4449

4214

4450

From the user point of view, the boundary conditions are totally

4451

determined by three arrays\footnote{except with Lagrangian}:

4452

\texttt{itypfb(nfabor,nphas)}\index{\texttt{itypfb}},

4215

From the user point of view, the boundary conditions are fully

4216

defined by three arrays\footnote{except with Lagrangian}:

4217

\texttt{itypfb(nfabor)}\index{\texttt{itypfb}},

4453

4218

\texttt{icodcl(nfabor,nvar)}\index{\texttt{icodcl}} and

4454

4219

\texttt{rcodcl(nfabor,nvar,3)}\index{\texttt{rcodcl}}.

4455

4220

\begin{list}{-}{}

4456

\item \texttt{itypfb(ifac,iphas)} defines the type of the face \texttt{ifac}

4457

(input, wall, ...) for the phase \texttt{iphas}.

4221

\item \texttt{itypfb(ifac)} defines the type of the face \texttt{ifac}

4222

(input, wall, ...).

4458

4223

\item \texttt{icodcl(ifac,ivar)} defines the type of boundary

4459

condition for the variable \texttt{ivar} at the face \texttt{ifac}

4224

condition for the variable \texttt{ivar} on the face \texttt{ifac}

4460

4225

(Dirichlet, flux ...).

4461

4226

\item \texttt{rcodcl(ifac,ivar,.)} gives the numerical values associated with the

4462

4227

type of boundary condition (value of the Dirichlet, of the flux ...).

4463

4228

\end{list}

4464

4229

4465

4230

In the case of standard boundary conditions (see

4466

\S\ref{prg_clstandard}), it is enough to complete \texttt{itypfb(ifac,iphas)} and

4467

some boxes of the array \texttt{rcodcl}, the array \texttt{icodcl} and most of the

4468

boxes of \texttt{rcodcl} are completed automatically. For non-standard boundary

4231

\S\ref{prg_clstandard}), it is sufficient to complete \texttt{itypfb(ifac)} and

4232

parts of the array \texttt{rcodcl}; the array \texttt{icodcl} and most of \texttt{rcodcl} are filled automatically. For non-standard boundary

4469

4233

conditions (see \S\ref{prg_clnonstandard}), the arrays \texttt{icodcl} and

4470

\texttt{rcodcl} must be totally completed.

4234

\texttt{rcodcl} must be fully completed.

4471

4235

4472

4236

%==================================

4473

4237

\subsubsection{Coding of standard boundary conditions}

4474

4238

%==================================

4475

4239

\label{prg_clstandard}%

4476

The standard values taken by the indicator \texttt{itypfb} are:

4240

The standard key words used by the indicator \texttt{itypfb} are:

4477

4241

\texttt{ientre\index{ientre}}, \texttt{iparoi\index{iparoi}},

4478

4242

\texttt{iparug\index{iparug}}, \texttt{isymet\index{isymet}},

4479

4243

\texttt{isolib\index{isolib}} and \texttt{iindef\index{iindef}}.

4484

4248

\begin{list}{$\rightarrow$}{}

4485

4249

\item Zero-flux condition for pressure and Dirichlet condition for all

4486

4250

other variables. The value of the Dirichlet must be given in

4487

\texttt{rcodcl(ifac,ivar,1)} for every value of \texttt{ivar}, apart from

4488

\texttt{ivar=ipr(iphas)}. The other values of \texttt{rcodcl} and

4489

\texttt{icodcl} are completed automatically.

4251

\texttt{rcodcl(ifac,ivar,1)} for every value of \texttt{ivar}, except for

4252

\texttt{ivar=ipr}. The other values of \texttt{rcodcl} and

4253

\texttt{icodcl} are filled automatically.

4490

4254

\end{list}

4491

4255

4492

4256

\item If \texttt{itypfb=iparoi}: smooth solid wall face, impermeable and with friction.

4493

4257

4494

4258

\begin{list}{$\rightarrow$}{}

4495

\item the eventual moving velocity of the wall tangent to the face is

4496

given by \texttt{rcodcl(ifac,ivar,1)} (\texttt{ivar} being

4497

\texttt{iu(iphas)}, \texttt{iv(iphas)} or \texttt{iw(iphas))}. The initial

4498

value of \texttt{rcodcl(ifac,ivar,1)} is zero for

4499

the three velocity components (and therefore needs to be specified

4500

only in the case of the existence of a slipping velocity). \\

4501

{\em WARNING: the wall moving velocity must be in the boundary face

4502

plane. By security, the code uses only the projection of this

4259

\item the potential sliding wall velocity of the face is

4260

found in \texttt{rcodcl(ifac,ivar,1)} (\texttt{ivar} being

4261

\texttt{iu}, \texttt{iv} or \texttt{iw)}. The initial

4262

values of \texttt{rcodcl(ifac,ivar,1)} are zero for

4263

the three velocity components (and therefore are to be specified

4264

only if the velocity is not equal to zero). \\

4265

{\em WARNING: the wall sliding velocity must belong to the boundary face

4266

plane. For safety, the code only uses the projection of this

4503

4267

velocity on the face. As a consequence, if the velocity specified

4504

by the user is not in the face plane, the wall moving velocity really

4268

by the user does not belong to the face plane, the wall sliding velocity really

4505

4269

taken into account will be different.}

4506

4270

4507

\item Concerning the scalars, two kinds of boundary conditions can be

4271

\item For scalars, two kinds of boundary conditions can be

4508

4272

defined:

4509

4273

\begin{list}{$\rightsquigarrow$}{}

4510

4274

\item Imposed value at the wall. The user must write\\

4512

4276

\hspace*{1cm}\texttt{rcodcl(ifac,ivar,1)}=imposed value\\

4513

4277

\item Imposed flux at the wall. The user must write\\

4514

4278

\hspace*{1cm}\texttt{icodcl(ifac,ivar)}=3\\

4515

\hspace*{1cm}\texttt{rcodcl(ifac,ivar,3)}=flux imposed value (for the flux

4516

definition according to the variable, the user may refer to the

4517

case \texttt{icodcl}=3 of the paragraph \ref{prg_clnonstandard}).

4518

\item If the user does not complete these arrays, the default condition

4279

\hspace*{1cm}\texttt{rcodcl(ifac,ivar,3)}=flux imposed value (depending on the variable, the user may refer to the case \texttt{icodcl}=3 of Section \ref{prg_clnonstandard} for the flux definition).

4280

\item If the user does not fill these arrays, the default condition

4519

4281

is zero flux.

4520

4282

\end{list}

4521

4283

\end{list}

4525

4287

\begin{list}{$\rightarrow$}{}

4526

4288

\item the eventual moving velocity of the wall tangent to the face is

4527

4289

given by \texttt{rcodcl(ifac,ivar,1)} (\texttt{ivar} being

4528

\texttt{iu(iphas)}, \texttt{iv(iphas)} or \texttt{iw(iphas)}). The initial

4290

\texttt{iu}, \texttt{iv} or \texttt{iw}). The initial

4529

4291

value of \texttt{rcodcl(ifac,ivar,1)} is zero for

4530

4292

the three velocity components (and therefore needs to be specified

4531

4293

only in the case of the existence of a slipping velocity). \\

4534

4296

velocity on the face. As a consequence, if the velocity specified

4535

4297

by the user is not in the face plane, the wall moving velocity really

4536

4298

taken into account will be different.}

4537

\item The dynamic roughness must be specified in \texttt{rcdocl(ifac,iu(iphas),3)}.

4538

The values of \texttt{rcdocl(ifac,iv(iphas),3)} and

4539

\texttt{rcdocl(ifac,iw(iphas),3)} are not used.

4299

\item The dynamic roughness must be specified in \texttt{rcdocl(ifac,iu,3)}.

4300

The values of \texttt{rcdocl(ifac,iv,3)} and

4301

\texttt{rcdocl(ifac,iw,3)} are not used.

4540

4302

\item For scalars, two kinds of boundary conditions can be defined:

4541

4303

\begin{list}{$\rightsquigarrow$}{}

4542

4304

\item Imposed value at the wall. The user must write\\

4552

4314

is zero flux.

4553

4315

\end{list}

4554

4316

\end{list}

4555

4556

4557

4558

4317

\item If \texttt{itypfb=isymet}: symmetry face (or wall without friction)

4559

4318

\begin{list}{$\rightarrow$}{}

4560

\item Nothing to write in \texttt{icodcl} and \texttt{rcodcl}.

4319

\item Nothing to be writen in \texttt{icodcl} and \texttt{rcodcl}.

4561

4320

\end{list}

4562

4321

4563

4322

\item If \texttt{itypfb=isolib}: free outlet face (or more precisely free

4564

4323

inlet/outlet with forced pressure)

4565

4324

\begin{list}{$\rightarrow$}{}

4566

4325

\item The pressure is always treated with a Dirichlet condition, calculated

4567

in order to have $\displaystyle \frac{d}{dn}\left(\frac{dP}{d\tau}\right)=0$.

4568

The pressure is given the value $P_0$ at the first face \texttt{isolib} met.

4569

The pressure drop is always linked to just one face, even if there are

4326

with the constraint $\displaystyle \frac{d}{dn}\left(\frac{dP}{d\tau}\right)=0$.

4327

The pressure is set to $P_0$ at the first \texttt{isolib} face met.

4328

The pressure calibration is always done on a single face, even if there are

4570

4329

several outlets.

4571

4330

\item If the mass flow is coming in, the ``infinite'' velocity is retained

4572

and Dirichlet condition for the scalars and the turbulent quantities is used

4331

and a Dirichlet condition for the scalars and the turbulent quantities is used

4573

4332

(or zero-flux condition if no Dirichlet value has been specified).

4574

\item If the mass flow is going out, zero-flux condition for the velocity,

4333

\item If the mass flow is going out, zero-flux condition are set for the velocity,

4575

4334

the scalars and the turbulent quantities.

4576

\item Nothing to write in \texttt{icodcl} or \texttt{rcodcl} for pressure or

4577

velocity. An optional Dirichlet condition can be specified for the scalars

4335

\item Nothing is written in \texttt{icodcl} or \texttt{rcodcl} for the pressure or

4336

the velocity. An optional Dirichlet condition can be specified for the scalars

4578

4337

and turbulent quantities.

4579

4338

\end{list}

4580

4339

4581

\item If \texttt{itypfb=iindef}: non-defined type face (non-standard case)

4340

\item If \texttt{itypfb=iindef}: undefined type face (non-standard case)

4582

4341

\begin{list}{$\rightarrow$}{}

4583

\item The coding is done by completing every array \texttt{rcodcl} and

4342

\item Coding is done in a non-standard way by filling both arrays \texttt{rcodcl} and

4584

4343

\texttt{icodcl} (see \S\ref{prg_clnonstandard}).

4585

4344

\end{list}

4586

4345

\end{list}

4587

4346

4588

4347

\minititre{Notes}

4589

$\bullet\ $ Whatever the value of the indicator \texttt{itypfb(ifac,iphas)}, if

4348

$\bullet\ $ Whatever is the value of the indicator \texttt{itypfb(ifac)}, if

4590

4349

the array \texttt{icodcl(ifac,ivar)} is modified by the user ({\em i.e.} filled

4591

in by a value different from zero), the code will not use the default

4592

conditions for the variable \texttt{ivar} at the face \texttt{ifac}, but will

4593

take into account the values of \texttt{icodcl} and \texttt{rcodcl} given by the

4594

user (these arrays must then be totally completed, like in the non-standard case). \\

4595

For instance, for a symmetry face at which the scalar 1 is given a

4350

with a non-zero value), the code will not use the default

4351

conditions for the variable \texttt{ivar} at the face \texttt{ifac}. It will

4352

take into account only the values of \texttt{icodcl} and \texttt{rcodcl} provided by the

4353

user (these arrays must then be fully completed, like in the non-standard case). \\

4354

For instance, for a normal symmetry face where scalar 1 is associated with a

4596

4355

Dirichlet condition equal to 23.8 (with an infinite exchange

4597

4356

coefficient):\\

4598

\hspace*{2cm}\texttt{itypfb(ifac,iphas)=isymet}\\

4357

\hspace*{2cm}\texttt{itypfb(ifac)=isymet}\\

4599

4358

\hspace*{2cm}\texttt{icodcl(ifac,isca(1))=1}\\

4600

4359

\hspace*{2cm}\texttt{rcodcl(ifac,isca(1),1)=23.8}\\

4601

(\texttt{rcodcl(ifac,isca(1),2)=rinfin} is the default value, so it is

4602

not necessary to specify it)\\

4603

The boundary conditions for the other variables are still automatically

4360

(\texttt{rcodcl(ifac,isca(1),2)=rinfin} is the default value, therfore it is

4361

not necessary to specify a value)\\

4362

The boundary conditions for the other variables are remain automatically

4604

4363

defined.

4605

4364

4606

4365

\noindent

4607

$\bullet\ $The user may define new types of wall faces. He only needs to

4608

choose a value $N$ and to specify completely the boundary conditions

4609

corresponding to this new wall face type (see

4610

\S\ref{prg_clnonstandard}). He must then specify

4611

\texttt{itypfb(ifac,iphas)}=$N$. The value of $N$ must be between 1 and

4366

$\bullet\ $The user can define new types of boundary faces. He only needs to

4367

choose a value $N$ and to fully specify the boundary conditions (see

4368

\S\ref{prg_clnonstandard}). He must specify

4369

\texttt{itypfb(ifac)}=$N$ where $N$ range is 1 to

4612

4370

\texttt{ntypmx\index{ntypmx}} (maximum number of boundary face types), and of

4613

4371

course different from the values \texttt{ientre}, \texttt{iparoi},

4614

\texttt{iparug}, \texttt{isymet}, \texttt{isolib} and \texttt{iindef} (the value

4615

of these variables is given in the file \texttt{paramx.h}). This allows to

4616

easily isolate some boundary faces, in order to calculate balances.

4372

\texttt{iparug}, \texttt{isymet}, \texttt{isolib} and \texttt{iindef} (the values

4373

of these variables are given in the file \texttt{paramx.h}). This allows to

4374

easily isolate some boundary faces, in order for instance to calculate balances.

4617

4375

4618

4376

%==================================

4619

4377

\subsubsection{Coding of non-standard boundary conditions}

4620

4378

%==================================

4621

4379

\label{prg_clnonstandard}%

4622

In the case of a face not corresponding to a standard type, the user

4623

must complete all of the arrays \texttt{itypfb}, \texttt{icodcl} and

4624

\texttt{rcodcl}. \texttt{itypfb(ifac,iphas)} is then worth \texttt{iindef}

4625

or another value defined by the user (see note in the end of paragraph

4380

In the case a face does not correspond to a standard type, the user

4381

must fill completely the arrays \texttt{itypfb}, \texttt{icodcl} and

4382

\texttt{rcodcl}. \texttt{itypfb(ifac)} is then equal to \texttt{iindef}

4383

or another value defined by the user (see note at the end of Section

4626

4384

\ref{prg_clstandard}). The arrays \texttt{icodcl} and \texttt{rcodcl}

4627

must be completed as follows:

4385

must be filled as follows:

4628

4386

4629

4387

\begin{list}{$\bullet$}{}

4630

4388

\item If \texttt{icodcl(ifac,ivar)}=1: Dirichlet condition at the face

4636

4394

4637

4395

\item \texttt{rcodcl(ifac,ivar,2)} is the value of the exchange coefficient

4638

4396

between the outside and the fluid for the variable \texttt{ivar}. An

4639

infinite value (\texttt{rcodcl(ifac,ivar,2)=rinfin}) indicates a

4640

perfect transfer between the outside and the fluid (default case).

4397

infinite value (\texttt{rcodcl(ifac,ivar,2)=rinfin}) indicates an

4398

ideal transfer between the outside and the fluid (default case).

4641

4399

4642

4400

\item \texttt{rcodcl(ifac,ivar,3)} is not used.

4643

4401

4644

\item \texttt{rcodcl(ifac,ivar,1)} is expressed in the unit of the variable

4402

\item \texttt{rcodcl(ifac,ivar,1)} has the units of the variable

4645

4403

\texttt{ivar}, {\em i.e.}:

4646

4404

\begin{list}{$\rightsquigarrow$}{}

4647

4405

\item $m/s$ for the velocity

4656

4414

4657

4415

\item $J.kg^{-1}$ for the enthalpy

4658

4416

4659

\item \degresC$^2$ for the temperature fluctuations

4417

\item \degresC$^2$ for temperature fluctuations

4660

4418

4661

\item $J^2.kg^{-2}$ for the enthalpy fluctuations

4419

\item $J^2.kg^{-2}$ for enthalpy fluctuations

4662

4420

\end{list}

4663

4421

4664

\item \texttt{rcodcl(ifac,ivar,2)} is expressed in the following unit (defined so

4665

that by multiplying the exchange coefficient and the variable, the

4666

obtained flux has the same unit as the flux defined below for

4422

\item \texttt{rcodcl(ifac,ivar,2)} has the following units (defined in such way that when multiplying the exchange coefficient by the variable, the

4423

given flux has the same units as the flux defined below when

4667

4424

\texttt{icodcl=3)}:

4668

4425

4669

4426

\begin{list}{$\rightsquigarrow$}{}

4690

4447

wall. This flux is negative if it is a source for the fluid. It corresponds to:

4691

4448

\begin{list}{$\rightsquigarrow$}{}

4692

4449

\item

4693

$\displaystyle -C_p(\frac{\lambda_T}{C_p}+\frac{\mu_t}{\sigma_T})\grad T\cdot\vect{n}$ in the case of a temperature (in $W/m^2$).

4450

$\displaystyle -C_p(\frac{\lambda_T}{C_p}+\frac{\mu_t}{\sigma_T})\grad T\cdot\vect{n}$ for a temperature (in $W/m^2$).

4694

4451

4695

4452

$\displaystyle -(\lambda_h+\frac{\mu_t}{\sigma_h})\grad h\cdot\vect{n}$

4696

in the case of an enthalpy (in $W/m^2$).

4697

4698

$\displaystyle -(\lambda_\varphi+\frac{\mu_t}{\sigma_\varphi})\grad\varphi\cdot\vect{n}$ in the case of another scalar $\varphi$ (in $kg.m^{-2}.s^{-1}.[\varphi]$, where $[\varphi]$ is the unit of $\varphi$).

4699

4700

\item $-\Delta t\ \grad P\cdot\vect{n}$ in the case of the pressure (in $kg.m^{-2}.s^{-1}$).

4701

4702

\item $-(\mu+\mu_t)\grad U_i\cdot\vect{n}$ in the case of a velocity component (in $kg.m^{-1}.s^{-2}$).

4703

4704

\item $-\mu\grad R_{ij}\cdot\vect{n}$ in the case of a $R_{ij}$ tensor component (in $W/m^2$).

4453

for an enthalpy (in $W/m^2$).

4454

4455

4456

4457

\item $-\Delta t\ \grad P\cdot\vect{n}$ for the pressure (in $kg.m^{-2}.s^{-1}$).

4458

4459

\item $-(\mu+\mu_t)\grad U_i\cdot\vect{n}$ for a velocity component (in $kg.m^{-1}.s^{-2}$).

4460

4461

\item $-\mu\grad R_{ij}\cdot\vect{n}$ for a $R_{ij}$ tensor component (in $W/m^2$).

4705

4462

\end{list}

4706

4463

4707

4464

\end{list}

4708

4465

4709

4466

\item If \texttt{icodcl(ifac,ivar)}=4: symmetry condition, for the symmetry

4710

4467

faces or wall faces without friction. This condition can only be

4711

used for the velocity components ($\vect{U}\cdot\vect{n}=0$) and

4712

the $R_{ij}$ tensor components (for the other variables, a zero-flux

4713

condition type is generally used).\\

4468

used for velocity components ($\vect{U}\cdot\vect{n}=0$) and

4469

the $R_{ij}$ tensor components (for other variables, a zero-flux

4470

condition type is usually used).\\

4714

4471

4715

\item If \texttt{icodcl(ifac,ivar)}=5: friction condition, for the smooth-wall faces

4472

\item If \texttt{icodcl(ifac,ivar)}=5: friction condition, for wall faces

4716

4473

with friction. This condition can not be applied to the pressure.

4717

4474

\begin{list}{$\rightsquigarrow$}{}

4718

4475

\item For the velocity and (if necessary) the turbulent variables, the

4719

4476

values at the wall are calculated from theoretical profiles. In

4720

the case of a moving wall, the three components of the slipping

4721

velocity are given by (\texttt{rcodcl(ifac,iu(iphas),1)},

4722

\texttt{rcodcl(ifac,iv(iphas),1)}, and \texttt{rcodcl(ifac,iw(iphas),1)}).\\

4723

{\em WARNING: the wall moving velocity must be in the boundary face

4724

plane. By security, the code uses only the projection of this

4725

velocity on the face. As a consequence, if the velocity specified

4726

by the user is not in the face plane, the wall moving velocity really

4477

the case of a sliding wall, the three components of the sliding

4478

velocity are given by (\texttt{rcodcl(ifac,iu,1)},

4479

\texttt{rcodcl(ifac,iv,1)}, and \texttt{rcodcl(ifac,iw,1)}).\\

4480

{\em WARNING: the wall sliding velocity must belong to the boundary face

4481

plane. For safety, the code uses only the projection of this

4482

velocity on the face. Therefore, if the velocity vector specified

4483

by the user does not belong to the face plane, the wall sliding velocity really

4727

4484

taken into account will be different.}

4728

4485

4729

\item For the other scalars, the condition \texttt{icodcl}=5 is similar to

4486

\item For other scalars, the condition \texttt{icodcl}=5 is similar to

4730

4487

\texttt{icodcl=1}, but with a wall exchange coefficient calculated from a

4731

theoretical law. The values of \texttt{rcodcl(ifac,ivar,1)} and

4732

\texttt{rcodcl(ifac,ivar,2)} must therefore be specified: see \cite{theory}.

4488

theoretical law. Therefore, the values of \texttt{rcodcl(ifac,ivar,1)} and

4489

\texttt{rcodcl(ifac,ivar,2)} must be specified: see \cite{theory}.

4733

4490

\end{list}

4734

4491

4735

4492

\item If \texttt{icodcl(ifac,ivar)}=6: friction condition, for the rough-wall faces

4736

with friction. This condition can not be applied to the pressure.

4493

with friction. This condition can not be used with the pressure.

4737

4494

\begin{list}{$\rightsquigarrow$}{}

4738

4495

\item For the velocity and (if necessary) the turbulent variables, the

4739

4496

values at the wall are calculated from theoretical profiles. In

4740

the case of a moving wall, the three components of the slipping

4741

velocity are given by (\texttt{rcodcl(ifac,iu(iphas),1)},

4742

\texttt{rcodcl(ifac,iv(iphas),1)}, and \texttt{rcodcl(ifac,iw(iphas),1)}).\\

4743

{\em WARNING: the wall moving velocity must be in the boundary face

4744

plane. By security, the code uses only the projection of this

4745

velocity on the face. As a consequence, if the velocity specified

4746

by the user is not in the face plane, the wall moving velocity really

4497

the case of a sliding wall, the three components of the sliding

4498

velocity are given by (\texttt{rcodcl(ifac,iu,1)},

4499

\texttt{rcodcl(ifac,iv,1)}, and \texttt{rcodcl(ifac,iw,1)}).\\

4500

{\em WARNING: the wall sliding velocity must belong to the boundary face

4501

plane. For safety, the code uses only the projection of this

4502

velocity on the face. Therefore, if the velocity vector specified

4503

by the user does not belong to the face plane, the wall sliding velocity really

4747

4504

taken into account will be different.}\\

4748

The dynamic roughness height is given by \texttt{rcodcl(ifac,iu(iphas),3)} only.

4505

The dynamic roughness height is given by \texttt{rcodcl(ifac,iu,3)} only.

4749

4506

4750

4507

\item For the other scalars, the condition \texttt{icodcl}=6 is similar to

4751

4508

\texttt{icodcl}=1, but with a wall exchange coefficient calculated from a

4755

4512

\end{list}

4756

4513

4757

4514

\item If \texttt{icodcl(ifac,ivar)}=9: free outlet condition for the

4758

velocity. This condition can only be applied to the velocity

4515

velocity. This condition is only applicable to velocity

4759

4516

components.\\

4760

If the mass flow at the face is going out, this condition is equivalent

4517

If the mass flow at the face negative, this condition is equivalent

4761

4518

to a zero-flux condition.\\

4762

If the mass flow at the face is coming in, the value zero is imposed to

4763

the velocity at the face (but not to the mass flow).\\

4519

If the mass flow at the face is positive, the velocity at the face is set to zero (but not to the mass flow).\\

4764

4520

\texttt{rcodcl} is not used.

4765

4521

4766

4522

\end{list}

4778

4534

4779

4535

The code checks the main compatibilities between the boundary

4780

4536

conditions. In particular, the following rules must be respected: \\

4781

$\bullet\ $On each face, the three components of the velocity

4782

must belong to the same type. The same must be true for the different

4783

components of the $R_{ij}$ tensor.\\

4537

$\bullet\ $On each face, the boundary conditions of the three velocity components must belong to the same type. The same is true for the components of the $R_{ij}$ tensor.\\

4784

4538

$\bullet\ $If the boundary conditions for the velocity belong to the

4785

``slipping'' type (\texttt{icodcl}=4), the conditions for $R_{ij}$ must belong to

4539

``sliding'' type (\texttt{icodcl}=4), the conditions for $R_{ij}$ must belong to

4786

4540

the ``symmetry'' type (\texttt{icodcl}=4), and vice versa.\\

4787

4541

$\bullet\ $If the boundary conditions for the velocity belong to the

4788

``friction'' type (\texttt{icodcl}=5 or 6), the conditions for the turbulent variables

4542

``friction'' type (\texttt{icodcl}=5 or 6), the boundary conditions for the turbulent variables

4789

4543

must belong to the ``friction'' type, too.\\

4790

$\bullet\ $If the boundary condition for a scalar belongs to the

4791

``friction'' type, the boundary condition for the velocity must belong to

4544

$\bullet\ $If the boundary condition of a scalar belongs to the

4545

``friction'' type, the boundary condition of the velocity must belong to

4792

4546

the ``friction'' type, too.

4793

4547

4794

In case of error, if post-processing output is activated (which is the default),

4795

a special error output to the same mesh format occurs, so as to help

4548

In case of mistakes, if the post-processing output is activated (which is the default setting),

4549

a special error output, similar to the mesh format, is produced in order to help

4796

4550

correcting boundary condition definitions.

4797

4551

4798

4552

%==================================

4800

4554

%==================================

4801

4555

4802

4556

In the code, it may be necessary to have access to all the boundary

4803

faces of a given type. In order to ease this kind of search, an array of

4804

sorted faces is automatically completed (and updated at every time step)

4805

for each phase \texttt{iphas}: \texttt{itrifb(nfabor,iphas)\index{itrifb}}.\\

4806

\texttt{ifac=itrifb(i,iphas)} is the number of the i$^{\text{th}}$ face of type

4557

faces of a given type. To ease this kind of search, an array made of

4558

sorted faces is automatically filled (and updated at each time step):

4559

\texttt{itrifb(nfabor)\index{itrifb}}.\\

4560

\texttt{ifac=itrifb(i)} is the number of the i$^{\text{th}}$ face of type

4807

4561

1.\\

4808

\texttt{ifac=itrifb(i+n,iphas)} is the number of the i$^{\text{th}}$ face de type

4562

\texttt{ifac=itrifb(i+n)} is the number of the i$^{\text{th}}$ face of type

4809

4563

2, if there are $n$ faces of type 1.\\

4810

4564

... etc.

4811

4565

4812

4566

Two auxiliary arrays of size \texttt{ntypmx} are also defined.\\

4813

\texttt{idebty(ityp,iphas)\index{idebty}} is the number of the first box

4814

corresponding to the

4815

faces of type \texttt{ityp} in the array \texttt{itrifb}.\\

4816

\texttt{ifinty(ityp,iphas)\index{ifinty}} is the number of the last box

4817

corresponding to the

4818

faces of type \texttt{ityp} in the array \texttt{itrifb}.

4819

4820

Therefore, a number \texttt{ifac0} between \texttt{idebty(ityp,iphas)} and

4821

\texttt{ifinty(ityp,iphas)} corresponds to each face of type

4822

\texttt{ityp=itypfb(ifac,iphas)}, so that \texttt{ifac=itrifb(ifac0,iphas)}.

4823

4824

If there is no face of type \texttt{ityp}, the code imposes \\

4825

\texttt{ifinty(ityp,iphas)=idebty(ityp,iphas)-1},\\

4826

which allows to bypass, for all the missing \texttt{ityp}, the loops like \\

4827

\texttt{do ii=idebty(ityp,iphas),ifinty(ityp,iphas)}.

4828

4829

The values of all these indicators are displayed in the beginning of the

4567

\texttt{idebty(ityp)\index{idebty}} is the index

4568

corresponding to the first

4569

face of type \texttt{ityp} in the array \texttt{itrifb}.\\

4570

\texttt{ifinty(ityp)\index{ifinty}} is the index

4571

corresponding to the last

4572

face of type \texttt{ityp} in the array \texttt{itrifb}.

4573

4574

Therefore, a value \texttt{ifac0} found between \texttt{idebty(ityp)} and

4575

\texttt{ifinty(ityp)} is associated to each face \texttt{ifac} of type

4576

\texttt{ityp=itypfb(ifac)}, so that \texttt{ifac=itrifb(ifac0)}.

4577

4578

If there is no face of type \texttt{ityp}, the code set \\

4579

\texttt{ifinty(ityp)=idebty(ityp)-1},\\

4580

which enables to bypass, for all the missing \texttt{ityp}, the loops such as \\

4581

\texttt{do ii=idebty(ityp),ifinty(ityp)}.

4582

4583

The values of all these indicators are displayed at the beginning of the

4830

4584

code execution listing.

4831

4585

4832

4586

%=============================================================

4833

\subsection[Management of the boundary conditions with LES: \texttt{usvort}]

4834

{Management of the boundary conditions with LES: \textmd{\texttt{usvort}}}

4587

\subsubsection[Boundary conditions with LES]

4588

{Boundary conditions with LES}

4835

4589

%===============================================================

4836

4590

\label{prg_usvort}%

4837

This subroutine allows to generate the non-stationary inlet boundary

4591

The subroutine \texttt{usvort} allows to generate the non-stationary inlet boundary

4838

4592

conditions for the LES by the vortex method. The method is based on

4839

4593

the generation of vortices in the 2D inlet plane with help from

4840

4594

the pre-defined functions. The fluctuation normal to the inlet plane

4909

4663

The 4 edges are numbered relative to the directions \texttt{dir1} and

4910

4664

\texttt{dir2} as shown in figure \ref{Fig_vortex}:

4911

4665

4912

\begin{figure}[hp]

4666

\begin{figure}[!ht]

4913

4667

\centerline{

4914

4668

\includegraphics*[width=8cm]{vortex}}

4915

4669

\caption{Numbering of the edges of a rectangular inlet(\texttt{icas}=1)

5006

4760

\texttt{phidat}.

5007

4761

\end{list}

5008

4762

5009

5010

%==================================

5011

\subsection{Management of the variable physical properties:

5012

\textmd{\texttt{usphyv}}}

5013

%==================================

5014

5015

\noindent

5016

\textit{Subroutine called every time step.}

5017

5018

If necessary, all the variation laws related to the fluid physical

5019

parameters (density, viscosity, thermal diffusivity, ...) are written in this

5020

subroutine.

4763

%==================================

4764

\subsection{Manage the variable physical properties}

4765

%==================================

4766

%==================================

4767

\subsubsection{Basic variable physical properties}\label{sec:usphyv}

4768

%==================================

4769

When the fluid properties are not constant, the user is offered the choice to define the variation laws in the Graphical User Interface (GUI) or in the subroutine \texttt{usphyv} which is called at each time step. In the GUI, in the item ``Fluid properties'' under the heading ``Physical properties'', the variation laws are defined for the fluid density, viscosity, specific heat and thermal conductivity through the use of a formula editor, see figs. \ref{fig:V-60_GUI} and \ref{fig:UL1_GUI}.

4770

4771

\begin{figure}[!ht]

4772

\begin{center}

4773

\includegraphics[width=13cm]{gui_fluid_props}

4774

\caption{Physical properties - Fluid properties}

4775

\label{fig:V-60_GUI}

4776

\end{center}

4777

\end{figure}

4778

4779

\begin{figure}[!ht]

4780

\begin{center}

4781

\includegraphics[width=9cm]{gui_density_law}

4782

\caption{Definition of a user law for the density}

4783

\label{fig:UL1_GUI}

4784

\end{center}

4785

\end{figure}

4786

4787

If necessary, all the variation laws related to the fluid physical properties are written in the subroutine \texttt{usphyv}.

5021

4788

5022

4789

The validity of the variation laws must be checked, particularly when

5023

4790

non-linear laws are defined (for instance, a third-degree polynomial law

5024

4791

may produce negative density values).

5025

4792

5026

{\bf \underline{WARNING}}\label{prg_propvar}

4793

{\bf \underline{WARNING}}\label{prg_propvar}%

5027

4794

\begin{list}{$\bullet$}{}

5028

\item If one wishes to impose a density or viscosity variable in

5029

\texttt{usphyv}, it can be done either in the interface or in

5030

\texttt{usini1}(\texttt{irovar(iphas)}=1, \texttt{ivivar(iphas)}=1).

4795

\item If one wishes to impose a variable density or variable viscosity in

4796

\texttt{usphyv}, it must be flagged either in the interface or in

4797

\texttt{usini1}(\texttt{irovar}=1, \texttt{ivivar}=1).

5031

4798

\item In order to impose a physical property ($\rho$, $\mu$,

5032

4799

$\lambda$, $C_p$)\footnote{except for some specific physics} a reference

5033

value must be inputted to the interface or in \texttt{usini1}(in

5034

particular for $\rho$, the pressure will contain 1 part as $\rho_0 gz$)

5035

\item By default, the $C_p$ coefficient of the phase \texttt{iphas} and the

5036

diffusivity of the scalars \texttt{iscal} ($\lambda/C_p$ for the

4800

value should be provided in the interface or in \texttt{usini1} (in

4801

particular for $\rho$, the pressure will be function of $\rho_0 gz$)

4802

\item By default, the $C_p$ coefficient and the

4803

diffusivity for the scalars \texttt{iscal} ($\lambda/C_p$ for the

5037

4804

temperature) are considered as constant in time and uniform in

5038

space, with the values \texttt{cp0(iphas)} and \texttt{visls0(iscal)}

4805

space, with the values \texttt{cp0} and \texttt{visls0(iscal)}

5039

4806

specified in the interface or in \texttt{usini1}.\\

5040

To give a variable value to $C_p$, the user must specify it in the

5041

interface or give the value 1 to \texttt{icp(iphas)} in \texttt{usini1},

5042

and complete for each cell \texttt{iel} the array

4807

To assign a variable value to $C_p$, the user \textbf{must}t specify it in the

4808

interface or assign the value 1 to \texttt{icp} in \texttt{usini1},

4809

and fill for each cell \texttt{iel} the array

5043

4810

\texttt{propce(iel,ipccp)} in \texttt{usphyv}. Completing the array

5044

\texttt{propce(iel,ipccp)} while \texttt{icp(iphas)=0} induces array

4811

\texttt{propce(iel,ipccp)} while \texttt{icp=0} induces array

5045

4812

overwriting problems and produces wrong results.

5046

4813

5047

4814

\item In the same way, to have variable diffusivities for the scalars

5048

\texttt{iscal}, the user must specify it in the interface or give the value

4815

\texttt{iscal}, the user \textbf{must} specify it in the interface or give the value

5049

4816

1 to \texttt{ivisls(iscal)} in \texttt{usini1}, and complete for each cell

5050

4817

\texttt{iel} the array \texttt{propce(iel,ipcvsl)} in \texttt{usphyv}.

5051

4818

Completing \texttt{propce(iel,ipcvsl)} while \texttt{ivisls(iscal)}=0

5052

4819

induces memory overwriting problems and produces wrong results.\\

5053

4820

5054

{\em Example}: If the scalars 1 and 3 have a constant and uniform

5055

viscosity, and if the scalars 2 and 4 have a variable viscosity,

5056

the following values must be imposed in \texttt{usini1}: \\

4821

{\em Example}: If scalars 1 and 3 have a constant and uniform

4822

diffusivity, and if scalars 2 and 4 have a variable diffusivity,

4823

the following values must be set in \texttt{usini1}: \\

5057

4824

\texttt{ivisls(1)}=0, \texttt{ivisls(2)}=1, \texttt{ivisls(3)}=0

5058

4825

and \texttt{ivisls(4)}=1. \\

5059

4826

The indicators \texttt{ivisls(2)} and \texttt{ivisls(4)} are then

5060

modified automatically by the code in order to reflect the rank

4827

modified automatically by the code in order to return the rank

5061

4828

corresponding to the diffusivity of each scalar in the list of physical

5062

properties\footnote{they are no longer worth 1 but stay positive

5063

so that \texttt{ivisls}$>$0 is synonymous with variable property}.

4829

properties\footnote{they are no longer equal to 1 but stay positive

4830

so that \texttt{ivisls}$>$0 is synonymous with variable diffusivity}.

5064

4831

The arrays \mbox{\texttt{propce(iel,ipcvsl)}} in \texttt{usphyv} must

5065

4832

then be completed with \texttt{ipcvsl=ipproc(ivisls(2))} and

5066

4833

\texttt{ipcvsl=ipproc(ivisls(4))}. \\

5075

4842

is variable.

5076

4843

\end{list}

5077

4844

5078

%==================================

5079

\subsection{Non-default variables initialisation:

5080

\textmd{\texttt{usiniv}}}

5081

%==================================

5082

5083

\noindent

5084

\textit{Subroutine only called during calculation initialisation.}

5085

5086

At the calculation beginning, the variables are initialised

5087

automatically by the code. Velocities and scalars are set to the value

5088

0 (or \texttt{scamax} or \texttt{scamin} if 0 is outside the acceptable

5089

scalar variation range), and the turbulent variables are estimated from

5090

\texttt{uref} and \texttt{almax}. \\

5091

For $k$ in $k-\varepsilon$, $R_{ij}-\varepsilon$, v2f or $k-\omega$

5092

model:\\

5093

{\texttt{rtp(iel,ikiph) = 1.5*(0.02*uref(iphas))**2}

5094

(in $R_{ij}-\varepsilon$, $R_{ij}=\frac{2}{3}k\delta_{ij}$)\\

5095

For $\varepsilon$ in $k-\varepsilon$, $R_{ij}-\varepsilon$ or v2f model:\\

5096

\texttt{rtp(iel,ieiph) = rtp(iel,ikiph)**1.5*cmu/almax(iphas)}\\

5097

For $\omega$ in $k-\omega$ model:\\

5098

\texttt{rtp(iel,iomgip) = rtp(iel,ikiph)**0.5/almax(iphas)}\\

5099

For $\varphi$ and $\overline{f}$ in v2f model:\\

5100

\texttt{rtp(iel,iphiph) = 2/3}\\

5101

\texttt{rtp(iel,ifbiph) = 0}

5102

5103

The subroutine \texttt{usiniv} allows if necessary to initialise some

5104

variables at values closer to their estimated final values, in order to

5105

obtain a faster convergence.

5106

5107

This subroutine allows also to make non-standard initialisation of

5108

physical parameters (density, viscosity, ...), to impose a local

5109

value of the time step, or to modify some parameters (time step,

5110

variable specific heat, ...) in the case of a calculation restart.

5111

5112

\minititre{Note: value of the time step}

5113

\begin{list}{-}{}

5114

\item In the case of a calculation with constant and uniform time step

5115

(\texttt{idtvar}=0), the value of the time step is \texttt{dtref},

5116

given in the parametric file of the interface or \texttt{usini1}, the

5117

calculation being whether a restart (\texttt{isuite}=1) or not (

5118

\texttt{isuite}=0).

5119

\item In the case of a calculation with non-constant time step

5120

(\texttt{idtvar}=1 or 2) which is not a calculation restart

5121

(\texttt{isuite}=0), the

5122

value of \texttt{dtref} given in the parametric file of the interface or in

5123

\texttt{usini1} is used to initialise the time step.

5124

\item In the case of a calculation with non-constant time step

5125

(\texttt{idtvar}=1 or 2) which is a restart (\texttt{isuite}=1) of a

5126

calculation whose time step type was different (for instance, restart

5127

using a variable time step of a calculation run using a constant time

5128

step), the value of \texttt{dtref} given in the parametric file of the

5129

interface or in \texttt{usini1} is used to initialise the time step.

5130

\item In the case of a calculation with non-constant time step

5131

(\texttt{idtvar}=1 or 2) which is a restart (\texttt{isuite}=1) of a

5132

calculation whose time step type was the same (for instance, restart with

5133

\texttt{idtvar}=1 of a calculation run with \texttt{idtvar}=1), the time

5134

step is read from the restart file and the value of \texttt{dtref} given

5135

in the parametric file of the interface or in \texttt{usini1} is not used.

5136

\end{list}

5137

It follows that for a calculation with non-constant time step (\texttt{idtvar}=1

5138

or 2) which is a restart (\texttt{isuite}=1) of a calculation in which

5139

\texttt{idtvar} had the same value, \texttt{dtref} does not allow to modify the

5140

time step. The user subroutine \texttt{usiniv} allows to modify the array

5141

\texttt{dt} which contains the value of the time step read from the restart file

5142

(array whose size is \texttt{ncelet}, defined at the cell centers whatever the

5143

chosen time step type).

5144

5145

{\em WARNING: to initialise the variables in the framework of a

5146

specific physics module} (\texttt{nscapp.gt.0}) {\em one of the subroutines

5147

\texttt{usebui}, \texttt{usd3pi}, \texttt{uslwci} or \texttt{uscpiv}

5148

should be used instead of \texttt{usiniv} (depending on the activated module).}

5149

5150

%==================================

5151

\subsection{Non-standard management of the chronological record files:

5152

\textmd{\texttt{ushist}}}

5153

%==================================

5154

\label{prg_ushist}

5155

5156

\noindent

5157

\textit{Subroutine called every time step}

5158

5159

The interface and the subroutine \texttt{usini1} allow to manage the

5160

``automatic'' chronological record files in an autonomous way:

5161

position of the probes, printing frequency and concerned variables. The

5162

results are written in a different file for each variable. These files

5163

are written in {\em xmgrace or {\em gnuplot}} format and contain the profiles corresponding to

5164

every probe. This type of output format may not be well adapted if, for

5165

instance, the number of probes is too high. The subroutine

5166

\texttt{ushist} allows then to personalise the output format of the

5167

chronological record files. The version given as example in the

5168

directory works as follows:

5169

5170

\begin{list}{-}{}

5171

\item Positionning of the probes (only at the first passage): the index

5172

\texttt{ii} varies between 1 and the number of probes. The coordinates

5173

\texttt{xx}, \texttt{yy} and \texttt{zz} of each probe are given.

5174

The subroutine \texttt{findpt}

5175

gives then the number \texttt{icapt(ii)\index{icapt}} of the cell center

5176

which is the closest to the defined probe.

5177

5178

\item Opening of the output files (only at the first pass): in the

5179

version given as example, the program opens a different file for

5180

all the \texttt{nvar} variables. \texttt{ficush(j)} contains the name of the

5181

J\raisebox{1ex}{\small th} file and \texttt{impush(j)} its unit number

5182

(\texttt{impush} is initialised by default so that the user has at his

5183

disposal specific unit numbers and does not run the risk to overwrite an

5184

already open file).

5185

5186

\item Writing to the files: in the version given as example, the program

5187

writes the time step number, the physical time step (based on the

5188

standard time step in the case of a variable time step) and the

5189

value of the selected variable at the different probes.

5190

5191

\item Closing of the files (only at the last time step).

5192

5193

\end{list}

5194

5195

{\em WARNING: The use of {\em\texttt{ushist}} neither erases nor replaces the

5196

parameters given in the interface or in {\em\texttt{usini1}}. Therefore, in

5197

the case of the use of {\em\texttt{ushist}}, and to avoid the creation

5198

of useless files, the user should set {\em \texttt{ncapt}=0} in the interface or

5199

in {\em \texttt{usini1}} to deactivate the automatic production of

5200

chronological records}.\\

5201

In addition, {\em \texttt{ushist}} generates supplementary result

5202

files. The user shoud remember to add in the launch script the necessary

5203

command to copy them in the directory \texttt{RESU} at the end of the

5204

calculation. The interface allows the specification of the name of the copied

5205

user results files. For the calculations without interface, the variable must

5206

be inputted in USER\_OUTPUT\_FILES in the launch script.

5207

5208

%==================================

5209

\subsection{User source terms in Navier-Stokes: \textmd{\texttt{ustsns}}}

5210

%==================================

5211

5212

\noindent

5213

\textit{Subroutine called every time step}

5214

5215

This subroutine is used to add user source terms to the Navier-Stokes

5216

equations. For each phase \texttt{iphas}, it is called three times every time

5217

step, once for each velocity component (\texttt{ivar} is successively worth

5218

\texttt{iu(iphas)}, \texttt{iv(iphas)} and \texttt{iw(iphas)}).

5219

At each passage, the user must complete if necessary the arrays \texttt{crvimp}

5220

and \texttt{crvexp} expressing respectively the implicit and explicit part of

5221

the source term. If no other source terms apart from \texttt{ivar=iu(iphas)} for

5222

example, are required, \texttt{crvimp} and \texttt{crvexp} must be read over and

5223

their 2 other components, \texttt{ivar=iv(ihpas)} and \texttt{ivar=iw(iphas)}

5224

must be cancelled.

4845

4846

%==================================

4847

\subsubsection{Modification of the turbulent viscosity}

4848

%==================================

4849

4850

The subroutine \texttt{usvist} is used to modify the calculation of the turbulent

4851

viscosity, {\em i.e.} $\mu_t$ in $kg.m^{-1}.s^{-1}$

4852

(this piece of information, at the mesh cell centers, is conveyed by the

4853

variable \texttt{propce(iel,ipcvst)}, with

4854

\texttt{ipcvst = ipproc(ivisct)}). The

4855

subroutine is called at the beginning of every time step, after the

4856

calculation of the physical parameters of the flow and of the

4857

``conventional'' value of $\mu_t$ corresponding to the chosen turbulence

4858

model (indicator \texttt{iturb}).\\

4859

{\em WARNING: The calculation of the turbulent viscosity being a

4860

particularly sensible stage, a wrong use of {\em\texttt{usvist}} may

4861

seriously distort the results.}

4862

4863

%==================================

4864

\subsubsection{Modification of the variable $C$ of the dynamic LES model}

4865

%==================================

4866

4867

\noindent

4868

\textit{Subroutine called every time step in the case of LES with the

4869

dynamic model.}

4870

4871

The subroutine \texttt{ussmag} is used to modify the calculation of the variable $C$ of

4872

the LES sub-grid scale dynamic model.

4873

4874

Let us first remind that the LES approach introduces the notion of

4875

filtering between large eddies and small motions. The solved variables

4876

are said to be filtered in an ``implicit'' way. Sub-grid scale models

4877

(``dynamic'' models) introduce in addition an explicit filtering.

4878

4879

The notations used for the definition of the variable $C$ used in the

4880

dynamic models of \CS are specified below. These notations are the ones

4881

assumed in the document \cite{benhamadouche01}, to which the user may

4882

refer for more details.

4883

4884

The value of $a$ filtered by the explicit filter (of width

4885

$\widetilde{\overline{\Delta}}$) is called $\widetilde{a}$ and the value

4886

of $a$ filtered by the implicit filter (of width $\overline{\Delta}$) is

4887

called $\overline{a}$.

4888

We define:

4889

\begin{equation}

4890

\begin{array}{ll}

4891

\overline{S}_{ij}=\frac{1}{2}(\frac{\partial \overline{u}_i}{\partial x_j}

4892

+\frac{\partial \overline{u}_j}{\partial x_i}) &

4893

||\overline{S}||=\sqrt{2 \overline{S}_{ij}\overline{S}_{ij}}\\

4894

\alpha_{ij}=-2\widetilde{\overline{\Delta}}^2

4895

||\widetilde{\overline{S}}||

4896

\widetilde{\overline{S}}_{ij}&

4897

\beta_{ij}=-2\overline{\Delta}^2

4898

||\overline{S}||

4899

\overline{S}_{ij}\\

4900

L_{ij}=\widetilde{\overline{u}_i\overline{u}_j}-

4901

\widetilde{\overline{u}}_i\widetilde{\overline{u}}_j&

4902

M_{ij}=\alpha_{ij}-\widetilde{\beta}_{ij}\\

4903

\end{array}

4904

\end{equation}

4905

4906

4907

In the framework of LES, the total viscosity (molecular + sub-grid) in

4908

$kg.m^{-1}.s^{-1}$ may be written in \CS:

4909

\begin{equation}

4910

\begin{array}{llll}

4911

\mu_{\text{total}}&=&\mu+\mu_{\text{sub-grid}} &

4912

\text{\ \ if\ \ }\mu_{\text{sub-grid}}>0\\

4913

&=&\mu &

4914

\text{\ \ otherwise }\\

4915

\text{with\ }\mu_{\text{sub-grid}}&=&\rho C \overline{\Delta}^2 ||\overline{S}||

4916

\end{array}

4917

\end{equation}

4918

4919

$\overline{\Delta}$ is the width of the implicit filter, defined at the

4920

cell $\Omega_i$ by \\

4921

$\overline{\Delta}=XLESFL*(ALES*|\Omega_i|)^{BLES}$

4922

\index{xlesfl}\index{ales}\index{bles}.

4923

4924

In the case of the Smagorinsky model (\texttt{iturb=40}), $C$ is a

4925

constant which is worth $C_s^2$. $C_s^2$ is the so-called Smagorinsky

4926

constant and is stored the variable \texttt{$csmago$\index{csmago}}.

4927

4928

In the case of the dynamic model (\texttt{iturb=41}), $C$ is variable in

4929

time and in space. It is determined by

4930

$\displaystyle C=\frac{M_{ij}L{ij}}{M_{kl}M_{kl}}$.

4931

4932

In practice, in order to increase the stability, the code does not use the

4933

value of $C$ obtained in each cell, but an average with the values

4934

obtained in the neighboring cells (this average uses the extended

4935

neighborhood and corresponds to the explicit filter). By default, the

4936

value calculated by the code is

4937

\begin{displaymath}

4938

C=\frac{\widetilde{M_{ij}L{ij}}}{\widetilde{M_{kl}M_{kl}}}

4939

\end{displaymath}

4940

4941

The subroutine \texttt{ussmag} allows to modify this value. It is for

4942

example possible to calculate the local average after having calculated the

4943

ratio

4944

\begin{displaymath}

4945

C=\widetilde{\left[\frac{M_{ij}L{ij}}{M_{kl}M_{kl}}\right]}

4946

\end{displaymath}

4947

4948

{\em WARNING: The subroutine {\em\texttt{ussmag}} can be activated only when

4949

the dynamic model is used.}

4950

4951

%==================================

4952

\subsection{User source terms}

4953

%==================================

5225

4954

5226

4955

Let us assume that the user source terms modify the equation of a

5227

4956

variable $\varphi$ in the following way:

5228

4957

\begin{displaymath}

5229

4958

\rho\frac{\partial \varphi}{\partial t}+\ldots = \ldots + S_{impl}\times\varphi+S_{expl}

5230

4959

\end{displaymath}

5231

$\varphi$ is here a velocity component, but the examples are also valid

5232

for a turbulent variable ($k$, $\varepsilon$, $R_{ij}$, $\omega$,

4960

The example is valid a velocity component, for a turbulent variable ($k$, $\varepsilon$, $R_{ij}$, $\omega$,

5233

4961

$\varphi$ or $\overline{f}$) and for a scalar (or for the average of the

5234

4962

square of the fluctuations of a scalar), because the syntax of the

5235

4963

subroutines \texttt{ustske}, \texttt{ustsri}, \texttt{ustsv2},

5236

4964

\texttt{ustskw} and \texttt{ustssc} is similar.

5237

4965

5238

In finite volume formulation, the solved system is then modified as

4966

In the finite volume formulation, the solved system is then modified as

5239

4967

follows:

5240

4968

\begin{displaymath}

5241

4969

\left(\frac{\rho_i\Omega_i}{\Delta t_i}-\Omega_iS_{impl,i}\right)

5246

4974

$\text{\texttt{crvimp}}_i=\Omega_iS_{impl,i}$\\

5247

4975

$\text{\texttt{crvexp}}_i=\Omega_iS_{expl,i}$

5248

4976

5249

In practice, it is essential that the term

4977

In practice, it is essential for the term

5250

4978

$\displaystyle \left(\frac{\rho_i\Omega_i}{\Delta

5251

t_i}-\Omega_iS_{impl,i}\right)$ is positive. To ensure this property,

4979

t_i}-\Omega_iS_{impl,i}\right)$ to be positive. To ensure this property,

5252

4980

the equation really taken into account by the code is the following:

5253

4981

\begin{displaymath}

5254

4982

\left(\frac{\rho_i\Omega_i}{\Delta t_i}-

5257

4985

+\ldots = \ldots + \Omega_iS_{impl,i}\varphi_i^{(n)} + \Omega_iS_{expl,i}

5258

4986

\end{displaymath}

5259

4987

To make the ``implicitation'' effective, the source term decomposition

5260

between implicit and explicit parts will be done by the user who must

5261

make sure $\text{\texttt{crvimp}}_i=\Omega_iS_{impl,i}$ is always negative

5262

(otherwise the solved equation remains right, but there is no

4988

between the implicit and explicit parts will be done by the user who must

4989

ensure that $\text{\texttt{crvimp}}_i=\Omega_iS_{impl,i}$ is always negative

4990

(otherwise the solved equation remains right, but there will not be

5263

4991

``implicitation'').

5264

4992

5265

{\em WARNING: When the second-order in time with extrapolation of the

4993

{\em WARNING: When the second-order in time is used along with the extrapolation of the

5266

4994

source terms\footnote{indicator \texttt{isno2t} for the velocity,

5267

\texttt{ISTO2T} for the turbulence and \texttt{isso2t} for the scalars}

5268

is activated, it is no longer possible to test the sign of $S_{impl,i}$,

4995

\texttt{ISTO2T} for the turbulence and \texttt{isso2t} for the scalars}, it is no longer possible to test the sign of $S_{impl,i}$,

5269

4996

because of coherence reasons (for more details, the user may refer to

5270

4997

the theoretical and computer documentation \cite{theory} of the

5271

4998

subroutine \texttt{preduv}). The user must therefore make sure it is

5272

always positive.}

4999

always positive (or take the risk to affect the calculation stability).}

5273

5000

5274

5001

\minititre{Particular case of a linearised source term}

5275

5002

5280

5007

\begin{displaymath}

5281

5008

\rho\frac{\partial\varphi}{\partial t}=F(\varphi)

5282

5009

\end{displaymath}

5283

5284

5010

We want to make it implicit using the following method:

5285

5011

\begin{eqnarray*}

5286

5012

\frac{\rho_i\Omega_i}{\Delta t}\left(\varphi_i^{(n+1)}-\varphi_i^{(n)}\right) & = &

5303

5029

$\text{\texttt{crvimp}}_i=-2K\Omega_i\varphi_i^{(n)}$\\

5304

5030

$\text{\texttt{crvexp}}_i=K\Omega_i[\varphi_i^{(n)}]^2$

5305

5031

5306

5307

5308

%==================================

5309

\subsection{User source terms for $k$ and $\varepsilon$:

5310

\textmd{\texttt{ustske}}}

5032

%==================================

5033

\subsubsection{In Navier-Stokes}

5034

%==================================

5035

5036

The subroutine \texttt{ustsns} is used to add user source terms to the Navier-Stokes

5037

equations (at each time step). It is called three times every time

5038

step, once for each velocity component (\texttt{ivar} is successively worth

5039

\texttt{iu}, \texttt{iv} and \texttt{iw}).

5040

At each passage, the user must complete if necessary the arrays \texttt{crvimp}

5041

and \texttt{crvexp} expressing respectively the implicit and explicit part of

5042

the source term. If no other source terms apart from \texttt{ivar=iu} for

5043

example, are required, \texttt{crvimp} and \texttt{crvexp} must be read over and

5044

their 2 other components, \texttt{ivar=iv(ihpas)} and \texttt{ivar=iw}

5045

must be cancelled.

5046

5047

%==================================

5048

\subsubsection{For $k$ and $\varepsilon$}

5311

5049

%==================================

5312

5050

5313

5051

\noindent

5314

5052

\textit{Subroutine called every time step, in $k-\varepsilon$ and

5315

5053

in v2f.}

5316

5054

5317

This subroutine is used to add source terms to the transport equations

5055

The subroutine \texttt{ustske} is used to add source terms to the transport equations

5318

5056

related to the turbulent kinetics energy $k$ and to the turbulent

5319

dissipation $\varepsilon$ (for each phase \texttt{iphas}).

5320

This subroutine is called every time step, once for each phase (the

5057

dissipation $\varepsilon$.

5058

This subroutine is called every time step (the

5321

5059

treatment of the two variables $k$ and $\varepsilon$ is made

5322

5060

simultaneously). The user is expected to provide the arrays \texttt{crkimp} and

5323

\texttt{crkexp} for $k$ and \texttt{creimp} and \texttt{creexp} for

5061

\texttt{crkexp} for $k$, and \texttt{creimp} and \texttt{creexp} for

5324

5062

$\varepsilon$. These arrays are similar to the arrays \texttt{crvimp} and

5325

5063

\texttt{crvexp} given for the velocity in the user subroutine \texttt{ustsns}.

5326

5064

The way of making implicit the resulting source terms is the same as the one

5327

5065

presented in \texttt{ustsns}. For $\varphi$ and $\bar{f}$

5328

in v2f, see \texttt{ustsv2}, \S\ref{prg_ustsv2}.

5329

5330

5066

in the v2f model, see \texttt{ustsv2}, \S\ref{prg_ustsv2}.

5331

5067

5332

5068

%==================================

5333

\subsection{User source terms for $R_{ij}$ and $\varepsilon$: \textmd{\texttt{ustsri}}}

5069

\subsubsection{For $R_{ij}$ and $\varepsilon$}

5334

5070

%==================================

5335

5071

5336

5072

\noindent

5337

5073

\textit{Subroutine called every time step, in $R_{ij}-\varepsilon$.}

5338

5074

5339

This subroutine is used to add source terms to the transport equations

5075

The subroutine \texttt{ustsri} is used to add source terms to the transport equations

5340

5076

related to the Reynolds stress variables $R_{ij}$ and to the turbulent

5341

dissipation $\varepsilon$ (for each phase \texttt{iphas}).

5342

This subroutine is called 7 times every time step and for each phase

5077

dissipation $\varepsilon$.

5078

This subroutine is called 7 times every time step

5343

5079

(once for each Reynolds stress component and once for the

5344

5080

dissipation). The user must provide the arrays \texttt{crvimp} and

5345

5081

\texttt{crvexp} for the variable \texttt{ivar} (referring successively to

5346

\texttt{ir11(iphas)}, \texttt{ir22(iphas)}, \texttt{ir33(iphas)},

5347

\texttt{ir12(iphas)}, \texttt{ir13(iphas)}, \texttt{ir23(iphas)} and

5348

\texttt{iep(iphas)}). These arrays are similar to the arrays \texttt{crvimp}

5082

\texttt{ir11}, \texttt{ir22}, \texttt{ir33},

5083

\texttt{ir12}, \texttt{ir13}, \texttt{ir23} and

5084

\texttt{iep}). These arrays are similar to the arrays \texttt{crvimp}

5349

5085

and \texttt{crvexp} given for the velocity in the user subroutine

5350

5086

\texttt{ustsns}. The method for impliciting the resulting source terms is the

5351

5087

same as that presented in \texttt{ustsns}.

5352

5088

5353

5089

%==================================

5354

\subsection{User source terms for $\varphi$ and $\overline{f}$:

5355

\textmd{\texttt{ustsv2}}}

5090

\subsubsection{For $\varphi$ and $\overline{f}$}

5356

5091

%==================================

5357

5092

\label{prg_ustsv2}

5093

5358

5094

\noindent

5359

5095

\textit{Subroutine called every time step, in v2f.}

5360

5096

5361

This subroutine is used to add source terms to the transport equations

5097

The subroutine \texttt{ustsv2} is used to add source terms to the transport equations

5362

5098

related to the variables $\varphi$ and $\overline{f}$ of the v2f

5363

$\varphi$-model (for each phase \texttt{iphas}). This subroutine is called twice

5364

every time step and for each phase (once for $\varphi$ and once for

5365

$\overline{f}$). The user is expected to provide the arrays \texttt{crvimp} and

5366

\texttt{crvexp} for \texttt{ivar} referring successively to \texttt{iphi(iphas)}

5367

and \texttt{ifb(iphas)}. Concerning $\varphi$, these arrays are similar to the arrays

5099

$\varphi$-model. This subroutine is called twice

5100

every time step (once for $\varphi$ and once for $\overline{f}$).

5101

The user is expected to provide the arrays \texttt{crvimp} and

5102

\texttt{crvexp} for \texttt{ivar} referring successively to \texttt{iphi}

5103

and \texttt{ifb}. Concerning $\varphi$, these arrays are similar to the arrays

5368

5104

\texttt{crvimp} and \texttt{crvexp} given for the velocity in the user subroutine

5369

5105

\texttt{ustsns}. Concerning $\overline{f}$, the equation is slightly

5370

5106

different:

5371

5107

\begin{displaymath}

5372

5108

L^2 div(\grad(\overline{f})) = \overline{f} + \ldots + S_{impl}\times\overline{f}+S_{expl}

5373

5109

\end{displaymath}

5374

In finite volume formulation, the solved system is written:

5110

In the finite volume formulation, the solved system is written as:

5375

5111

\begin{displaymath}

5376

5112

\int_{\partial\Omega_i}\grad(\overline{f})^{(n+1)}dS=\frac{1}{L_i^2}\left(

5377

5113

\Omega_i\overline{f}^{(n+1)}_i + \ldots + \Omega_iS_{impl,i}\overline{f}_i^{(n+1)} +

5385

5121

one presented in \texttt{ustsns}.

5386

5122

5387

5123

%==================================

5388

\subsection{User source terms for $k$ and $\omega$: \textmd{\texttt{ustskw}}}

5124

\subsubsection{For $k$ and $\omega$}

5389

5125

%==================================

5390

5126

5391

5127

\noindent

5392

5128

\textit{Subroutine called every time step, in $k-\omega$.}

5393

5129

5394

This subroutine is used to add source terms to the transport equations

5130

The subroutine \texttt{ustskw} is used to add source terms to the transport equations

5395

5131

related to the turbulent kinetics energy $k$ and to the specific

5396

dissipation rate $\omega$ (for each phase \texttt{iphas}). This subroutine is

5397

called every time step, once for each phase (the treatment of the two

5398

variables $k$ and $\omega$ is made simultaneously).The user is expected

5132

dissipation rate $\omega$. This subroutine is

5133

called every time step (the treatment of the two

5134

variables $k$ and $\omega$ is made simultaneously). The user is expected

5399

5135

to provide the arrays \texttt{crkimp} and \texttt{crkexp} for the variable

5400

$k$ the arrays \texttt{crwimp} and \texttt{crwexp} for the variable $\omega$.

5136

$k$, and the arrays \texttt{crwimp} and \texttt{crwexp} for the variable $\omega$.

5401

5137

These arrays are similar to the arrays \texttt{crvimp} and \texttt{crvexp}

5402

5138

given for the velocity in the user subroutine \texttt{ustsns}. The way of

5403

impliciting the resulting source terms is the same as the one presented in

5139

making implicit the resulting source terms is the same as the one presented in

5404

5140

\texttt{ustsns}.

5405

5141

5406

5142

%==================================

5407

\subsection{User source terms for the user scalars: \textmd{\texttt{ustssc}}}

5143

\subsubsection{For user scalars}

5408

5144

%==================================

5409

5145

5410

5146

\noindent

5411

5147

\textit{Subroutine called every time step.}

5412

5148

5413

This subroutine is used to add source terms to the transport equations

5149

The subroutine \texttt{ustssc} is used to add source terms to the transport equations

5414

5150

related to the user scalars (passive or not, average of the square of

5415

5151

the fluctuations of a scalar, ...). In the same way as

5416

5152

\texttt{ustsns}, this subroutine is called every time step, once for

5417

5153

each user scalar. The user needs to provide the arrays \texttt{crvimp}

5418

5154

and \texttt{crvexp} related to each scalar. \texttt{cvimp} and \texttt{crvexp}

5419

5155

must be set to 0 for the scalars on which it is not wished for the user source

5420

term term to be applied (the arrays are initially at 0 at each inlet in the subroutine.)

5421

5422

%==================================

5423

\subsection{Management of the pressure drops: \textmd{\texttt{uskpdc}}}

5424

%==================================

5425

5426

\noindent

5427

\textit{Subroutine called every time step.}

5428

5429

This subroutine is called three times every time step and for each phase

5430

\texttt{iphas}.

5431

5432

The tensor representing the pressure drops is supposed to be symmetric

5156

term to be applied (the arrays are initially set to 0 at each inlet in the subroutine.)

5157

5158

%==================================

5159

\subsection{Pressure drops (head losses)}

5160

%==================================

5161

5162

Pressure drops can be defined in the Gaphical User Interface (GUI) or in the subroutine \texttt{uskpdc} (called three times every time step). In the GUI, under the heading ``Volume conditions'', the item ``Volume regions definition'' allows to define areas where pressure drops occur, see an example in fig \ref{fig:hl1}. The item ``Head losses'' allows to specify the head loss coefficients, see fig \ref{fig:hl2}. The tensor representing the pressure drops is supposed to be symmetric

5433

5163

and positive.

5434

5164

5165

\begin{figure}[!ht]

5166

\begin{center}

5167

\begin{tabular}{c}

5168

\includegraphics[width=14cm]{gui_volume_regions} \\

5169

5170

\includegraphics[width=9cm]{gui_head_loss_regions}

5171

\end{tabular}

5172

\caption{Creation of head losses region}

5173

\label{fig:hl1}

5174

\end{center}

5175

\end{figure}

5176

5177

\begin{figure}[!ht]

5178

\begin{center}

5179

\includegraphics[width=9cm]{gui_head_loss_coeffs}

5180

\caption{Head losses coefficients}

5181

\label{fig:hl2}

5182

\end{center}

5183

\end{figure}

5184

5185

If necessary, the pressure drops are written in the

5186

subroutine \texttt{uskpdc}.

5187

5435

5188

\begin{list}{$\bullet$}{}

5436

5189

\item During the first call, all the cells are checked to know the

5437

number of cells in which a pressure drop is present for the phase

5438

\texttt{iphas}. This number is called \texttt{ncepdp\index{ncepdp}} in

5190

number of cells in which a pressure drop is present.

5191

This number is called \texttt{ncepdp\index{ncepdp}} in

5439

5192

\texttt{uskpdc} (and corresponds to

5440

\texttt{ncepdc(iphas)\index{ncepdc}}). It is used to lay out the arrays

5193

\texttt{ncepdc\index{ncepdc}}). It is used to lay out the arrays

5441

5194

related to the pressure drops. If there is no pressure drop,

5442

5195

\texttt{ncepdp} must be equal to zero (it is the default value, and the

5443

5196

rest of the subroutine is then useless).

5445

5198

\item During the second call, all the cells are checked again to

5446

5199

complete the array \texttt{icepdp\index{icepdp}} whose size is

5447

5200

\texttt{ncepdp}. \mbox{\texttt{icepdc(ielpdc)}} is the number of the

5448

\texttt{ielpdc}\raisebox{1ex}{\small th} cell containing pressure drops (for

5449

the current phase).

5201

\texttt{ielpdc}\raisebox{1ex}{\small th} cell containing pressure drops.

5450

5202

5451

5203

\item During the third call, all the cells containing pressure drops

5452

(for the current phase) are checked in order to complete the array

5204

are checked in order to complete the array

5453

5205

containing the components of the tensor of pressure drops

5454

5206

\mbox{\texttt{ckupdc(ncepdp,6)}\index{ckupdc}}. This array is so that

5455

5207

the equation related to the velocity may be written:

5468

5220

zones or values may be treated.

5469

5221

5470

5222

%==================================

5471

\subsection{Management of the mass sources: \textmd{\texttt{ustsma}}}

5223

\subsection{Management of the mass sources}

5472

5224

%==================================

5473

5225

5474

\noindent

5475

\textit{Subroutine called every time step.}

5476

5477

This subroutine is used to add a density source term in some cells of

5478

the domain. The mass conservation equation is then modified as follows:

5226

The subroutine \texttt{ustsma} is used to add a density source term in some cells of

5227

the domain (called at each time step). The mass conservation equation is then modified as follows:

5479

5228

\begin{displaymath}

5480

5229

\frac{\partial \rho}{\partial t} + div(\rho\vect{u})=\Gamma

5481

5230

\end{displaymath}

5510

5259

5511

5260

\bigskip

5512

5261

5513

This subroutine is called three times every time step (for each phase).

5262

This subroutine is called three times every time step.

5514

5263

5515

5264

\begin{list}{$\bullet$}{}

5516

5265

\item During the first call, all the cells are checked to know the

5517

number of cells containing a mass source term for the current

5518

phase \texttt{iphas}. This number is called \texttt{ncesmp\index{ncesmp}} in

5266

number of cells containing a mass source term.

5267

This number is called \texttt{ncesmp\index{ncesmp}} in

5519

5268

\texttt{ustsma} (and corresponds to

5520

\texttt{ncetsm(iphas)\index{ncetsm}}). It is used to lay out the arrays

5269

\texttt{ncetsm\index{ncetsm}}). It is used to lay out the arrays

5521

5270

related to the mass sources. If there is no mass source,

5522

5271

\texttt{ncesmp} must be equal to zero (it is the default value, and the

5523

5272

rest of the subroutine is then useless).

5525

5274

\item During the second call, all the cells are checked again to

5526

5275

complete the array \texttt{icetsm\index{icetsm}} whose dimension is

5527

5276

\texttt{ncesmp}. \mbox{\texttt{icetsm(ieltsm)}} is the number of the

5528

\texttt{ieltsm}\raisebox{1ex}{\small th} cell containing a mass source (for

5529

the current phase).

5277

\texttt{ieltsm}\raisebox{1ex}{\small th} cell containing a mass source.

5530

5278

5531

5279

\item During the third call, all the cells containing mass sources are

5532

5280

checked in order to complete the arrays

5537

5285

containing a mass source.\\

5538

5286

\hspace*{1cm}\texttt{itypsm}=0: $\varphi_i=\varphi^{(n+1)}$ condition\\

5539

5287

\hspace*{1cm}\texttt{itypsm}=1: imposed $\varphi_i$ condition\\

5540

\hspace*{1cm}\texttt{itypsm} is not used for \texttt{ivar=ipr(iphas)}\\

5541

- \texttt{(ieltsm,ipr(iphas))} is the value of the mass source term $\Gamma$, in

5288

\hspace*{1cm}\texttt{itypsm} is not used for \texttt{ivar=ipr}\\

5289

- \texttt{(ieltsm,ipr)} is the value of the mass source term $\Gamma$, in

5542

5290

$kg.m^{-3}.s^{-1}$.\\

5543

5291

- \texttt{smacel(ieltsm,ivar)}, for \texttt{ivar} different from

5544

\texttt{ipr(iphas)}, is the value

5292

\texttt{ipr}, is the value

5545

5293

of $\varphi_i$ for the variable \texttt{ivar} in the

5546

5294

\texttt{ielstm}\raisebox{1ex}{\small th} cell containing a mass source.\\

5547

5295

5548

5296

\minititre{Notes}

5549

5297

$\bullet$ If \texttt{itypsm(ieltsm,ivar)=0}, \texttt{smacel(ieltsm,ivar)}

5550

5298

is not used.\\

5551

$\bullet$ If $\Gamma$=\texttt{smacel(ieltsm,ipr(iphas))}$<$0, mass is removed from

5299

$\bullet$ If $\Gamma$=\texttt{smacel(ieltsm,ipr)}$<$0, mass is removed from

5552

5300

the system, and \CS considers automatically a

5553

5301

$\varphi_i=\varphi^{(n+1)}$ condition, whatever the values given

5554

5302

to \texttt{itypsm(ieltsm,ivar)} and \texttt{smacel(ieltsm,ivar)}

5563

5311

For the variance, do not take into account the scalar $\varphi_i$ in the environment

5564

5312

where $\varphi\ne\varphi_i$ generates a variance source.

5565

5313

5566

%========================================

5567

\subsection{Thermal module in a 1D wall}

5568

%========================================

5569

5570

\noindent

5571

\textit{subroutine called at every time step}

5572

5573

This subroutine takes into account the affected thermal inertia by a wall.

5574

Some boundary faces are treated as a solid wall with a given thickness, on

5575

which the code resolves an undimensional equation for the heat conduction.

5576

The coupling between the 1D module and the fluid works in a similar way to

5577

the coupling with the \syrthes. In construction, the user is not able to

5578

account for the heat transfer between different parts of the wall. A physical

5579

analysis of each problem, case by case is required to evaluate the relevance

5580

of its usage by way of a report of the simple conditions (temperature, zero-flux

5581

) or a coupling with \syrthes.\\

5582

5583

The use of this code requires that theres is only 1 phase (\texttt{nphas=1})

5584

and that the thermal scalar is defined as (\texttt{iscalt}$>0$).

5585

5586

{\em WARNING: The 1D thermal module is developped assuming the thermal scalar

5587

as a temperature. If the thermal scalar is an enthalpy, the code calls the

5588

subroutine \texttt{usthht} for each transfer of information between the fluid

5589

and the wall in order to convert the enthalpy to temperature and vice-versa.

5590

This function has not been tested and is firmly discouraged. If the thermal

5591

variable is the total (compressible) energy, the thermal module will not work.}

5592

5593

\bigskip

5594

5595

This procedure is called twice,on initialisation and again at each time step.

5596

5597

\begin{list}{$\bullet$}{}

5598

\item The 1st call (initialisation) all the boundary faces that will be treated

5599

as a coupled wall are marked out. This figure is written noted as

5600

\texttt{nfkpt1d}. It applies dimension to the arrays in the thermal module.

5601

\texttt{nfkpt1d} will be at 0 if there are no coupled faces (it is in fact the

5602

default value, the remainer of the subroutine is not used in this case).

5603

The parameter \texttt{isuit1} also need to be defined, this indicates if the

5604

temperature of the wall must be initialised or written in the file (stored in

5605

the variable \texttt{filmt1}).

5606

\item The 2nd call (initialisation) again concern the wall faces, it completes

5607

the \texttt{ifpt1d} array of dimension \texttt{nfpt1d}.

5608

\mbox{\texttt{ifpt1d(ifbt1d)}} is the number

5609

\texttt{ifbt1d}\raisebox{1ex}{\small th} boundary faces coupled with the thermal module

5610

of a 1D wall. The directional parameters are then completed for a pseudo

5611

wall associated to each face

5612

\begin{list}{-}{}

5613

\item \texttt{nppt1d(nfpt1d)\index{nppt1d}}: number of cells in the 1D mesh associated

5614

to the pseudo wall.

5615

\item \texttt{eppt1d(nfpt1d)\index{eppt1d}}: thickness of the pseudo wall.

5616

\item \texttt{rgpt1d(nfpt1d)\index{rgpt1d}}: geometery of the pseudo wall mesh (refined

5617

as a fluid if \texttt{rgt1d} is smaller than 1)

5618

\item \texttt{tppt1d(nfpt1d)\index{tppt1d}}: initialisation temperature of the wall

5619

(uniform in thickness). In the course of the calculation, the array stores the

5620

temperature of the solid at the fluid/solid interface.

5621

\end{list}

5622

5623

Other than for re-reading a file (\texttt{ficmt1}), \texttt{tppt1d} is not used.

5624

\texttt{nppt1d}, \texttt{ifpt1d}, \texttt{rgpt1d} and \texttt{eppt1d} are

5625

compared to data from the follow-up file and they must be identical.

5626

5627

{\em WARNING: The test in \texttt{ifpt1d} implicilty assumes that the array is completed

5628

in ascending order (i.e \texttt{ifpt1d(ii)}$>$\texttt{ifpt1d(jj)} if ii$>$jj.

5629

This will be the case if the coupled faces are defined starting from the unique loop on the

5630

boundary faces (as in the example). If this is not the case, contact the development

5631

team to short circuit the test.}

5632

5633

\item The 3rd call (at each time step) is for the confirmation that all the arrays

5634

involving physical parameter and external boundary conditions have been completed.

5635

\begin{list}{-}{}

5636

\item \texttt{iclt1d(nfpt1d)\index{iclt1d}}:Typical boundary condition at the external

5637

(pseudo) wall: Dirichlet condition (\texttt{iclt1d}=1) or flux condition (\texttt{iclt1d}=3)

5638

\item \texttt{tept1d(nfpt1d)\index{tept1d}}: External temperature of the pseudo wall in the

5639

Dirichlet case.

5640

\item \texttt{hept1d(nfpt1d)\index{hept1d}}: External coefficient of transfer in the pseudo

5641

wall under Dirichlet conditions(en $W.m^{-2}.K^.$).

5642

\item \texttt{fept1d(nfpt1d)\index{nfpt1d}}: External heat flux in the pseudo wall under

5643

the flux conditions(en $W.m^{-2}$,negative value for energy entering the wall)

5644

\item \texttt{xlmt1d(nfpt1d)\index{xlmt1d}}: Conductivity$\lambda$ of the wall uniform

5645

in thickness, (in $W.m^{-1}.K^{-1}$).

5646

\item \texttt{rcpt1d(nfpt1d)\index{rcpt1d}}: Volumetric heat capacity $\rho C_p$ of the

5647

wall uniform in thickness in $J.m^{-3}.K^{-1}$)

5648

\item \texttt{dtpt1d(nfpt1d)\index{dtpt1d}}: Physical time step ascociated with the solved

5649

1D equation of the pseudo wall(which can be different from the time step in the

5650

calculation)

5651

\end{list}

5652

5653

\end{list}

5654

5655

The 3rd call, done at each time step, allows the imposition of boundary conditions

5656

and physical values in time.

5657

5658

%==================================

5659

\subsection{ Initialization of the options of the variables related

5660

to the ale module: \textmd{\texttt{usalin}} and \textmd{\texttt{usstr1}} }

5661

%==================================

5662

\label{prg_usalin}%

5663

\noindent

5664

\textit{Subroutine called at the start.}

5665

5666

5667

\minititre{Subroutine \texttt{usalin}}

5668

This subroutine completes \texttt{usini1}.

5669

5670

\texttt{usalin} allows to set option for the ale module, and in

5671

particular to active the ale module

5672

5673

\minititre{Subroutine \texttt{usstr1}}

5674

5675

\texttt{usstr1} allows to specify for the structure module the

5676

following pieces of information:

5677

\begin{list}{-}{}

5678

\item number of structure (\texttt{nbstru}).

5679

\item initial value of deplacement, velocity and acceleration

5680

(\texttt{xstr0}, \texttt{xstreq} and \texttt{vstr0}).

5681

\end{list}

5682

5683

Below is a list of the different variables that might be modified:

5684

5685

\begin{list}{$\bullet$}{}

5686

5687

\item{\texttt{nbstru}} \\

5688

{the number of structures}

5689

5690

\item{\texttt{idfstr(i)}} \\

5691

{index of the structure, where I is the index of the face}

5692

5693

\item{\texttt{xstr0(i,k)}} \\

5694

{initial position of a structure, where \texttt{i} is the dimension of space

5695

and \texttt{k} the index of the structure}

5696

5697

\item{\texttt{xstreq(i,k)}} \\

5698

{position of balance of a structure, where \texttt{i} is the dimension of space

5699

and \texttt{k} the index of the structure}

5700

5701

\item{\texttt{vstr0(i,k)}} \\

5702

{initial velicity of a structure, where \texttt{i} is the dimension of space

5703

and \texttt{k} the index of the structure }

5704

\end{list}

5705

5706

%==================================

5707

\subsection{Management of the boundary conditions of velocity mesh related to the ale module: \textmd{\texttt{usalcl}}}

5708

%==================================

5709

5710

5711

\noindent

5712

\textit{Subroutine called every time step.}

5713

5714

\minititre{Subroutine \texttt{usalcl}}

5715

The use of \texttt{usalcl} is obligatory to run a calculation using

5716

the ale module just as it is in \texttt{usini1}. The way of using it

5717

is the same as the way of using \texttt{usclim} in the framework of

5718

standard calculations, that is to say a loop on the boundary faces

5719

marked out by their colour (or more generally by a property of their

5720

family), where the type of boundary condition of velocity mesh for

5721

each variable are defined.

5722

5723

The main numerical variables are described below.

5724

5725

\variab{ialtyb}{ialtyb(nfabor)}{ia}{In the ale module, the user

5726

defines the velocity mesh from the colour of the boundary faces, or

5727

more generally from their properties (colours, groups, ...), from the

5728

boundary conditions defined in \texttt{usclim}, or even from their

5729

coordinates. To do so, the array \texttt{ialtyb(nfabor)} gives for each face

5730

\texttt{ifac} the velocity mesh boundary condition types marked out by the key

5731

words \texttt{ivimpo\index{ivimpo}}, \texttt{igliss\index{igliss}},

5732

\texttt{ibfixe\index{ibfixe}}

5733

5734

\begin{list}{$\bullet$}{}

5735

5736

\item If \texttt{ialtyb=ivimpo}: imposed velocity.

5737

5738

\begin{list}{$\rightarrow$}{}

5739

\item In the case where all the nodes of a face have a imposed displacement,

5740

it is not necessary to fill the tables with boundary conditions

5741

velocity mesh for this face, they will be erased. In the other case,

5742

the value of the Dirichlet must be given in \texttt{rcodcl(ifac,ivar,1)} for

5743

every value of \texttt{ivar} (\texttt{iuma}, \texttt{ivma} and \texttt{iwma})

5744

The other boxes of \texttt{rcodcl} and \texttt{icodcl} are completed automatically.

5745

5746

The tangential velocity mesh is taken like a tape speed under the

5747

boundary conditions of wall for the fluid, except if wall velocity

5748

was specified by the user in the interface or usclim (in which case

5749

it is this speed which is considered).

5750

\end{list}

5751

5752

\item if \texttt{ialtyb(nfac) = ibfixe}: fixed wall

5753

\begin{list}{$\rightarrow$}{}

5754

\item the velocity is null.

5755

\end{list}

5756

5757

\item if \texttt{ialtyb(nfac) = igliss}: sliding wall

5758

\begin{list}{$\rightarrow$}{}

5759

\item the tangential velocity is not used.

5760

\end{list}

5761

5762

\end{list}

5763

5764

}

5765

5766

5767

%==================================

5768

\subsection{Management of the structure property: \textmd{\texttt{usstr2}}}

5769

%==================================

5770

5771

\noindent

5772

\textit{Subroutine called every time step.}

5773

5774

The use of \texttt{usstr2}

5775

is obligatory to run a calculation using the ale module with a structure module.

5776

5777

For each structure, the system that will be solved is:

5778

5779

\begin{equation}

5780

M.x^{''}+C.x^{''}+K.(x-x_{0} = 0

5781

\end{equation}

5782

5783

where

5784

5785

\begin{list}{-}{}

5786

\item $M$ is the mass stucture (\texttt{xmstru}).

5787

\item $C$ is the dumping coefficient of the stucture (\texttt{xcstru}).

5788

\item $K$ is the spring constant or force constant of the stucture (\texttt{xkstru}).

5789

\item $x_{0}$ is the initial position

5790

\end{list}

5791

5792

Below is a list of the different variables that might be modified:

5793

5794

\begin{list}{$\bullet$}{}

5795

5796

\item{\texttt{xmstru(i,j,k})} \\

5797

{the mass stucture of the structure, where \texttt{i},\texttt{j} is

5798

the array of mass structure and \texttt{k} the index of the structure.}

5799

5800

\item{\texttt{xcstru(i,j,k})}\\

5801

{dumping coefficient of the stucture, where \texttt{i},\texttt{j} is the array of

5802

dumping coefficient and \texttt{k} the index of the structure.}

5803

5804

\item{\texttt{xkstru(i,j,k)}}\\

5805

{spring constant of the stucture, where \texttt{i},\texttt{j} is the array of spring

5806

constant and \texttt{k} the index of the structure.}

5807

5808

\item{\texttt{forstr(i,k)}}\\

5809

{force vector of the stucture, where \texttt{i} is the force vector and

5810

\texttt{k} the index of the structure.}

5811

\end{list}

5812

5813

5814

%==================================

5815

\subsection{Modification of the turbulent viscosity: \textmd{\texttt{usvist}}}

5816

%==================================

5817

\noindent

5818

\textit{Subroutine called every time step.}

5819

5820

This subroutine is used to modify the calculation of the turbulent

5821

viscosity of the phase \texttt{iphas}, {\em i.e.} $\mu_t$ in $kg.m^{-1}.s^{-1}$

5822

(this piece of information, at the mesh cell centers, is conveyed by the

5823

variable \texttt{propce(iel,ipcvst)}, with

5824

\texttt{ipcvst = ipproc(ivisct(iphas))}). The

5825

subroutine is called at the beginning of every time step, after the

5826

calculation of the physical parameters of the flow and of the

5827

``conventional'' value of $\mu_t$ corresponding to the chosen turbulence

5828

model (indicator \texttt{iturb(iphas)}).\\

5829

{\em WARNING: The calculation of the turbulent viscosity being a

5830

particularly sensible stage, a wrong use of {\em\texttt{usvist}} may

5831

seriously distort the results.}

5832

5833

%==================================

5834

%==================================

5835

\subsection{Modification of the variable $C$ of the dynamic LES model:

5836

\textmd{\texttt{ussmag}}}

5837

%==================================

5838

5839

\noindent

5840

\textit{Subroutine called every time step in the case of LES with the

5841

dynamic model.}

5842

5843

This subroutine is used to modify the calculation of the variable $C$ of

5844

the LES sub-grid scale dynamic model.

5845

5846

Let us first remind that the LES approach introduces the notion of

5847

filtering between large eddies and small motions. The solved variables

5848

are said to be filtered in an ``implicit'' way. Sub-grid scale models

5849

(``dynamic'' models) introduce in addition an explicit filtering.

5850

5851

The notations used for the definition of the variable $C$ used in the

5852

dynamic models of \CS are specified below. These notations are the ones

5853

assumed in the document \cite{benhamadouche01}, to which the user may

5854

refer for more details.

5855

5856

The value of $a$ filtered by the explicit filter (of width

5857

$\widetilde{\overline{\Delta}}$) is called $\widetilde{a}$ and the value

5858

of $a$ filtered by the implicit filter (of width $\overline{\Delta}$) is

5859

called $\overline{a}$.

5860

We define:

5861

\begin{equation}

5862

\begin{array}{ll}

5863

\overline{S}_{ij}=\frac{1}{2}(\frac{\partial \overline{u}_i}{\partial x_j}

5864

+\frac{\partial \overline{u}_j}{\partial x_i}) &

5865

||\overline{S}||=\sqrt{2 \overline{S}_{ij}\overline{S}_{ij}}\\

5866

\alpha_{ij}=-2\widetilde{\overline{\Delta}}^2

5867

||\widetilde{\overline{S}}||

5868

\widetilde{\overline{S}}_{ij}&

5869

\beta_{ij}=-2\overline{\Delta}^2

5870

||\overline{S}||

5871

\overline{S}_{ij}\\

5872

L_{ij}=\widetilde{\overline{u}_i\overline{u}_j}-

5873

\widetilde{\overline{u}}_i\widetilde{\overline{u}}_j&

5874

M_{ij}=\alpha_{ij}-\widetilde{\beta}_{ij}\\

5875

\end{array}

5876

\end{equation}

5877

5878

5879

In the framework of LES, the total viscosity (molecular + sub-grid) in

5880

$kg.m^{-1}.s^{-1}$ may be written in \CS:

5881

\begin{equation}

5882

\begin{array}{llll}

5883

\mu_{\text{total}}&=&\mu+\mu_{\text{sub-grid}} &

5884

\text{\ \ if\ \ }\mu_{\text{sub-grid}}>0\\

5885

&=&\mu &

5886

\text{\ \ otherwise }\\

5887

\text{with\ }\mu_{\text{sub-grid}}&=&\rho C \overline{\Delta}^2 ||\overline{S}||

5888

\end{array}

5889

\end{equation}

5890

5891

$\overline{\Delta}$ is the width of the implicit filter, defined at the

5892

cell $\Omega_i$ by \\

5893

$\overline{\Delta}=XLESFL(IPHAS)*(ALES(IPHAS)*|\Omega_i|)^{BLES(IPHAS)}$

5894

\index{xlesfl}\index{ales}\index{bles}.

5895

5896

In the case of the Smagorinsky model (\texttt{iturb(iphas)=40}), $C$ is a

5897

constant which is worth $C_s^2$. $C_s^2$ is the so-called Smagorinsky

5898

constant and is stored the variable \texttt{$csmago$\index{csmago}}.

5899

5900

In the case of the dynamic model (\texttt{iturb(iphas)=41}), $C$ is variable in

5901

time and in space. It is determined by

5902

$\displaystyle C=\frac{M_{ij}L{ij}}{M_{kl}M_{kl}}$.

5903

5904

In practice, in order to increase the stability, the code does not use the

5905

value of $C$ obtained in each cell, but an average with the values

5906

obtained in the neighboring cells (this average uses the extended

5907

neighborhood and corresponds to the explicit filter). By default, the

5908

value calculated by the code is

5909

\begin{displaymath}

5910

C=\frac{\widetilde{M_{ij}L{ij}}}{\widetilde{M_{kl}M_{kl}}}

5911

\end{displaymath}

5912

5913

The subroutine \texttt{ussmag} allows to modify this value. It is for

5914

example possible to calculate the local average after having calculated the

5915

ratio

5916

\begin{displaymath}

5917

C=\widetilde{\left[\frac{M_{ij}L{ij}}{M_{kl}M_{kl}}\right]}

5918

\end{displaymath}

5919

5920

{\em WARNING: The subroutine {\em\texttt{ussmag}} can be activated only when

5921

the dynamic model is used.}

5922

5923

%==================================

5924

\subsection{Temperature-enthalpy and enthalpy-temperature conversions: \textmd{\texttt{usthht}}}

5925

%==================================

5926

5927

\noindent

5928

\textit{Subroutine optionally called.}

5929

5930

This subroutine is used to encapsulate a simple enthalpy-temperature

5931

conversion law and its inverse.

5932

This subroutine is called in \texttt{usray4}, user subroutine from the

5933

radiation module.

5934

5935

%==================================

5936

\subsection{Modification of the mesh geometry: \textmd{\texttt{usmodg}}}

5937

%==================================

5938

5939

\noindent

5940

\textit{Subroutine called only during the calculation initialisation.}

5941

5942

This subroutine may be used to modify ``manually'' the mesh vertices

5943

coordinates, \textit{i.e.} the array:

5944

\begin{list}{$\bullet$}{}

5945

\item \mbox{\texttt{xyznod(3,nnod)}} (vertex coordinates)

5946

\end{list}

5947

5948

{\em WARNING: Caution must be exercised when using this subroutine

5949

along with periodicity. Indeed, the periodicity parameters are not

5950

updated accordingly, meaning that the periodicity may be unadapted

5951

after one changes the mesh vertex coordinates. It is particularly

5952

true when one rescales the mesh.}\\

5953

5954

%==================================

5955

\subsection{Management of the post-processing intermediate outputs: \textmd{\texttt{usnpst}}}

5956

%==================================

5957

5958

\noindent

5959

\textit{Subroutine called every time step(even if the user hasn't moved it to the SRC directroy).}

5960

5961

This subroutine is used to determine when post-processing outputs will be

5962

generated. By default, it tests if the current time step number (\texttt{ntcabs}) is a

5314

%==================================

5315

\section{Results analysis}

5316

%==================================

5317

5318

%==================================

5319

\subsection{Management of the post-processing intermediate outputs}

5320

%==================================

5321

5322

The subroutine \texttt{usnpst} is used to specify when post-processing outputs will be

5323

generated (it is called at each time step even if the user hasn't moved it to the directory SRC). By default, it tests if the current time step number (\texttt{ntcabs}) is a

5963

5324

multiple of the chosen output frequency (\texttt{ntchr}). If it is the case, the

5964

5325

indicator \texttt{iipost} turns to 1, which triggers the writing of an

5965

5326

intermediate output. If the frequency is given a negative value, the

5966

test is not done.

5327

test is not performed.

5967

5328

5968

5329

For instance, a user who wants to generate post-processing outputs (also

5969

5330

called ``chronological outputs'') at

5992

5353

5993

5354

\newpage

5994

5355

%==================================

5995

\subsection{Definition of post-processing and mesh zones:

5996

\textmd{\texttt{usdpst}}}

5356

\subsection{Definition of post-processing and mesh zones}

5997

5357

%==================================

5998

5358

5999

\noindent

6000

\textit{Subroutine called at the calculation beginning..}

6001

6002

This subroutine allows for the definition of surface or volume

6003

sections, in the form of lists of \texttt{nlfac} internal faces

6004

(\texttt{lstfac}) and \texttt{nlfab} boundary faces (\texttt{lstfab}),

5359

The functions defined in \texttt{cs\_user\_postprocess.c}, namely

5360

\texttt{cs\_user\_postprocess\_writers}, \texttt{cs\_user\_postprocess\_meshes},

5361

and \texttt{cs\_user\_postprocess\_activate} allow for

5362

the definition of postprocessing output formats and frequency, and

5363

for the definition of surface or volume sections, in the form

5364

of lists of \texttt{nlfac} internal faces (\texttt{lstfac}) and

5365

\texttt{nlfab} boundary faces (\texttt{lstfab}),

6005

5366

or of \texttt{nlcel} cells (\texttt{lstcel}), in order to generate

6006

5367

chronological outputs in {\em EnSight}, {\em MED} or {\em CGNS} format.

6007

5368

6008

One or several ``writers'' can be associated with each visualization

6009

mesh, or ``part'' created. The arguments of the function \texttt{pstcwr}

6010

defining a ``writer'' are as follows:

5369

One or several writers can be associated with each post-processing

5370

mesh, or ``part'' created. The arguments of the function

5371

\texttt{cs\_post\_define\_writer} are as follows:

6011

5372

6012

5373

\begin{list}{$\bullet$}{}

6013

\item \texttt{nomcas}: basic name of the associated case.\\ {\em

5374

\item \texttt{writer\_id}: id the the associated writer.\\

5375

negative ids are reserved (-1 for the main output),

5376

but the matching writer's options may be redifined by

5377

calls to this function.

5378

\item \texttt{case\_name}: basic name of the associated case.\\ {\em

6014

5379

WARNING}: depending on the chosen format, this name may

6015

5380

be shortened (maximum number of characters: 32 for {\em MED},

6016

5381

19 for {\em EnSight}) or modified automatically (whitespaces or

6017

5382

forbidden characters will be replaced by '\_')

6018

\item \texttt{nomrep}: name of the output directory

6019

\item \texttt{nomfmt}: choice of the output format:

5383

\item \texttt{dir\_name}: name of the output directory

5384

\item \texttt{fmt\_name}: choice of the output format:

6020

5385

\begin{list}{$\rightarrow$}{}

6021

5386

\item {\em EnSight Gold} ({\em EnSight} also accepted)

6022

5387

\item {\em MED\_fichier} ({\em MED} also accepted)

6023

5388

\item {\em CGNS}

6024

\item {\em text} (mesh output, no variables output, for debug only).

6025

5389

\end{list}

6026

5390

The options are not case-sensitive, so {\em ensight} or {\em cgns} are valid, too.

6027

\item \texttt{optfmt}: character string containing a list of

5391

\item \texttt{fmt\_opts}: character string containing a list of

6028

5392

options related to the format, separated by commas; for the

6029

5393

{\em EnSight Gold} format, these options are:

6030

5394

\begin{list}{$\rightarrow$}{}

6046

5410

variable as a series of independent variables (a variable is recognised as a

6047

5411

tensor if its dimension is 6 or 9); not implemented yet.

6048

5412

\end{list}

6049

\item \texttt{indmod}: indicates if the post-processing (i.e. visualization) meshes

6050

(or ``parts'') are:

5413

\item \texttt{time\_dep}: indicates if the post-processing

5414

(i.e. visualization) meshes (or ``parts'') are:

6051

5415

\begin{list}{$\rightarrow$}{}

6052

\item 0 fixed (usual case)

6053

\item 1 deformable (the vertex positions may vary over time)

6054

\item 2 modifiable: (the lists of cells or faces

6055

defining these ``parts'' can be changed over time)

5416

\item \texttt{FVM\_WRITER\_FIXED\_MESH} fixed (usual case)

5417

\item \texttt{FVM\_WRITER\_TRANSIENT\_COORDS} deformable

5418

(the vertex positions may vary over time)

5419

\item \texttt{FVM\_WRITER\_TRANSIENT\_CONNECT} modifiable:

5420

(the lists of cells or faces

5421

defining these meshes can be changed over time)

6056

5422

\end{list}

6057

\item \texttt{ntchrl}: default output frequency associated with

6058

this ``writer'' (the output may be forced or prevented at

6059

every time step using the subroutine \texttt{usnpst})

5423

\item \texttt{output\_at\_end}: force output at calculation end

5424

if not 0

5425

\item \texttt{frequency\_n}: default output frequency in time steps

5426

associated with this writer, or $< 0$ (the output may be forced

5427

or prevented at any time step using the function

5428

\texttt{cs\_user\_postprocess\_activate})

5429

\item \texttt{frequency\_t}: default output frequency in seconds

5430

associated with this writer, or $< 0$ (has priority over

5431

\texttt{frequency\_n}, and the output may be forced or prevented

5432

at any time step using the function

5433

\texttt{cs\_user\_postprocess\_activate})

6060

5434

\end{list}

6061

5435

6062

5436

In order to allow the user to add an output format to

6063

5437

the main output format, or to add a mesh to the default

6064

output, the lists of standard and user meshes and ``writers'' are not

5438

output, the lists of standard and user meshes and writers are not

6065

5439

separated. Negative numbers are reserved for the non-user items. For

6066

5440

instance,the mesh numbers -1 and -2 correspond respectively to the global

6067

mesh and to boundary faces, generated by default, and the ``writer'' -1

5441

mesh and to boundary faces, generated by default, and the writer -1

6068

5442

corresponds to the usual post-processing case defined {\em via}

6069

5443

\texttt{usini1} or {\em via} the interface.

6070

5444

6071

5445

The user chooses the numbers corresponding to the post-processing

6072

meshes and ``writers'' he wants to create. These numbers must be positive

5446

meshes and writers he wants to create. These numbers must be positive

6073

5447

integers. It is possible to assocate a user mesh with the standard

6074

5448

post-processing case (-1), or to ask for outputs regarding the boundary

6075

faces (-2) associated with a user ``writer''.

5449

faces (-2) associated with a user writer.

6076

5450

6077

5451

For safety, the output frequency and the possibility to modify the

6078

post-processing meshes are associated with the ``writers'' rather than

6079

with with the ``parts''. This logic avoids unwanted generation of

5452

post-processing meshes are associated with the writers rather than

5453

with the meshes. This logic avoids unwanted generation of

6080

5454

inconstitent post-processing outputs. For instance EnSight would not

6081

5455

be able to read a case in which one field is output to a given part

6082

5456

every 10 time steps while another field is output to the same part

6083

5457

every 200 time steps.

6084

5458

6085

The possibility to modify a mesh over time is limited by the more restrictive

6086

``writer'' which is associated with it. For instance, if the ``writer''

6087

1 allows the modification of the mesh topology (argument \texttt{indmod

6088

= 2} in the call to \texttt{pstcwr}) and the ``writer'' 2 allows no

6089

modification (\texttt{indmod = 0}), a user post-processing mesh

6090

associated with the ``writers'' 1 and 2 will not be modifiable, but a

6091

mesh associated only with the ``writer'' 1 will be modifiable. The

5459

The possibility to modify a mesh over time is limited by the most restrictive

5460

writer which is associated with. For instance, if writer

5461

1 allows the modification of the mesh topology (argument \texttt{time\_dep

5462

= FVM\_WRITER\_TRANSIENT\_CONNECT} in the call to \texttt{cs\_post\_define\_writer}) and writer 2 allows no

5463

modification (\texttt{time\_dep = FVM\_WRITER\_FIXED\_MESH}),

5464

a user post-processing mesh

5465

associated with the writers 1 and 2 will not be modifiable, but a

5466

mesh associated only with the writer 1 will be modifiable. The

6092

5467

modification is done by means of the user subroutine \texttt{usmpst},

6093

which is called only for the currently modifiable meshes.

5468

which is called once per time step and per modifiable mesh.

6094

5469

6095

5470

It is possible to output variables which are normally automatically

6096

5471

output on the main volume or boundary meshes to a user mesh which is a subset

6097

of one of these by assigning the corresponding category to the user

6098

mesh. By default, a meshe's category is identical to its number, so

6099

the category associated with the main volume output is -1, and that

6100

associated with the main boundary output is -2. A category may be assigned

6101

using the \texttt{pstcat} subroutine.

5472

of one of these by setting the \texttt{auto\_variables} argument of

5473

one of the \texttt{cs\_post\_define\_...\_mesh} to \texttt{true}.

6102

5474

6103

5475

It is also possible to define an alias of a post-processing mesh. An

6104

alias shares all the attributes of a ``part'' (without duplication),

5476

alias shares all the attributes of its parent mesh (without duplication),

6105

5477

except its number. This may be used to output different variables on a

6106

same ``part'' with 2 different writers: the choice of output variables

6107

is based on the ``part'', so if $P_a$ is associated with writer $W_a$,

5478

same mesh with 2 different writers: the choice of output variables

5479

is based on the mesh, so if $P_a$ is associated with writer $W_a$,

6108

5480

all that is needed is to define an alias $P_b$ to $P_a$ and associate

6109

5481

it with writer $W_b$ to allow a different output variable selection with

6110

5482

each writer. An alias may be created using the \texttt{pstalm} subroutine.

6111

5483

6112

Modification of a part or it's alias over time is always limited by the

6113

most restrictive "writer" to which it's meshes have been asscoiated (parts of

6114

the structures being shared in memory). It is possible to define as many

6115

alias' as are required for a "part", but an alias cannot be defined for another alias.

6116

6117

It is not possible to mix cells and faces in the same ``part'' (most of

6118

the post-processing tools being perturbed by such a case)\footnote{in thr

6119

future, it will probably be possible to automatically add faces bearing

6120

group or attribute characteristics to a cell mesh, but those faces will

6121

only be written for formats supporting this (such as MED 2.3),

6122

and will only bear attributes, not variable fields}. If the user

6123

defines lists of faces and cells simultaneously, only the higher dimension

6124

entities (the cells) will be taken into account.

6125

6126

For a better understanding, the user may refer to the example given in

6127

{\texttt{usdpst}}. We can note that the whitespaces in the beginning or in

6128

the end of the character strings given as arguments of the functions

6129

called are suppressed automatically.

6130

6131

The variables to post-process on the defined ``parts'' will be specified

6132

in the subroutine \texttt{usvpst}. ``

6133

6134

{\em WARNING In the parallel case, some ``parts'' may not contain any

5484

Modification of a postprocessing mesh or it's alias over time is always

5485

limited by the most restrictive "writer" to which it's meshes have been

5486

asscoiated (parts of the structures being shared in memory). It is

5487

possible to define as many aliases as are required for a true mesh,

5488

but an alias cannot be defined for another alias.

5489

5490

It is not possible to mix cells and faces in the same mesh (most of

5491

the post-processing tools being perturbed by such a case)\footnote{actually,

5492

faces adjacent to selected cells and belonging to face or cell groups

5493

may be selected when the \texttt{add\_groups} of

5494

\texttt{cs\_post\_define\_...\_mesh} is set to \texttt{true},

5495

so as to maintain group information, but those faces will

5496

only be written for formats supporting this (such as MED),

5497

and will only bear groups, not variable fields}.

5498

5499

For a better understanding, the user may refer to the examples given in

5500

\texttt{cs\_user\_postprocess\_meshes}. We can note that the whitespaces

5501

in the beginning or in the end of the character strings given as arguments

5502

of the functions called are suppressed automatically.

5503

5504

The additional variables to post-process on the defined meshes

5505

will be specified in the subroutine \texttt{usvpst}. ``

5506

5507

{\em WARNING In the parallel case, some meshes may not contain any

6135

5508

local elements on a given processor. This is not a problem at all, as

6136

long as the ``part'' is defined for all processors (empty or not).

6137

It would in fact not be a good idea at all to define a ``part'' only

6138

if it contains local elements, global operations on the ``part'' would

6139

become impossible, leading to probable deadlocks or crashes.}

5509

long as the mesh is defined for all processors (empty or not).

5510

It would in fact not be a good idea at all to define a postprocessing

5511

mesh only if it contains local elements, global operations on that

5512

mesh would become impossible, leading to probable deadlocks or crashes.}

6140

5513

6141

5514

%==================================

6142

\subsection{Modification of the mesh zones to post-process:

6143

\textmd{\texttt{usmpst}}}

5515

\subsection{Modification of the mesh zones to post-process}

6144

5516

%==================================

6145

5517

6146

5518

\noindent

6147

5519

\textit{Subroutine called only for each modifiable ``part'', at every

6148

5520

active time step of an associated ``writer''.}

6149

5521

6150

For the user ``parts'' defined {\em via} the user subroutine

6151

\texttt{usdpst} and associated only with ``writers'' allowing the ``part''

6152

modification over time ({\em i.e.} created with the

6153

parameter \texttt{indmod} = 2), this subroutine is used to modify the

6154

lists of cells, internal and boundary faces defining this ``part'' (or

6155

post-processing mesh).

5522

The subroutine \texttt{usmpst} is used to modify the

5523

lists of cells, internal and boundary faces which define a ``user part'' (or

5524

post-processing mesh) defined through the user subroutine

5525

\texttt{usdpst} and associated only with ``writers''; allowing ``part''

5526

modifications over time ({\em i.e.} created with the

5527

parameter \texttt{indmod} = 2).

6156

5528

6157

5529

At first, the corresponding lists contain the previously defined

6158

5530

values. If these lists are modified for a given post-processing mesh, the

6187

5559

velocity isosurface.

6188

5560

6189

5561

%==================================

6190

\subsection{Definition of the variables to post-process:

6191

\textmd{\texttt{usvpst}}}

5562

\subsection{Definition of the variables to post-process}

6192

5563

%==================================

6193

5564

6194

\noindent

6195

\textit{Subroutine called for each ``part'', at every active time step of an

6196

associated ``writer'' (see \texttt{usdpst}).}

6197

6198

5565

For the parts defined in \texttt{usdpst}, the subroutine \texttt{usvpst}

6199

is used to specify the variables to post-process.

5566

is used to specify the variables to post-process (called for each ``part'', at every active time step of an

5567

associated ``writer'', see \texttt{usdpst}).

6200

5568

6201

5569

The output of a given variable is generated by means of a call to

6202

5570

\texttt{psteva}, whose arguments are:

6247

5615

variable at the cells. If \texttt{ivarpr} $= 1$, this

6248

5616

argument will be replaced by the position of the beginning

6249

5617

of the array on which the variable in defined, for instance

6250

\texttt{rtp(1, iu(1))} for the velocity of the phase 1.

5618

\texttt{rtp(1, iu(1))} for the velocity.

6251

5619

\item \texttt{trafac}: equivalent of \texttt{tracel} for the

6252

5620

internal faces.

6253

5621

\item \texttt{trafbr}: equivalent of \texttt{tracel} for the

6266

5634

associate it with a different ``writer'' in the subroutine \texttt{usdpst}.}

6267

5635

6268

5636

%==================================

6269

\subsection{Modification of the variables at the end of a time step: \textmd{\texttt{usproj}}}

5637

\subsection{Modification of the variables at the end of a time step}

6270

5638

%==================================

6271

5639

6272

\noindent

6273

\textit{Subroutine called every time step.}

6274

6275

This subroutine is called at the end of every time step. It is used to

5640

The subroutine \texttt{usproj} is called at the end of every time step. It is used to

6276

5641

print of modify any variable at the end of every time step.

6277

5642

6278

5643

Several examples are given:

6283

5648

\item Modification of the temperature in a given area starting from a

6284

5649

given time

6285

5650

6286

\item Extraction of a 1D profile

5651

\item Extraction of a 1D profile, see fig. \ref{fig:26_GUI}

6287

5652

6288

5653

\item Printing of a moment

6289

5654

6307

5672

the framework of periodic or parallel calculations (in particular

6308

5673

when we want to calculate the gradient of a variable or to have

6309

5674

access to values at the cells neighboring a face).

6310

\item Finally it must not be forgotten that the solving with

5675

\item Finally it must not be forgotten that the resolution with

6311

5676

temperature as a solved variable is questionable when the specific

6312

5677

heat is not constant.

6313

5678

\end{list}

6314

5679

6315

5680

%==================================

6316

\subsection{Radiative thermal transfers in semi-transparent gray media}

6317

%==================================

6318

6319

%==================================

6320

\subsubsection{Initialisation of the radiation main key words: \textmd{\texttt{usray1}}}

6321

%==================================

6322

6323

\noindent

6324

\textit{Subroutine called only during calculation initialisation.}

6325

6326

This subroutine is one of the two which must be completed by the user for all

6327

calculations including radiative thermal transfers. This subroutine is

6328

composed of three headings. The first one is dedicated to the activation

6329

of the radiation module, only in the case of classic physics. \\

6330

{\em WARNING: when a calculation is run using a specific physics module,

6331

this first heading must not be completed. The radiation module is then

6332

activated or not according to the parameter file related to the considered

6333

specific physics.} \\

6334

6335

\noindent

6336

In the second heading the basic parameters of the radiation module are indicated.\\

6337

Finally, the third heading deals with the selection of the

6338

post-processing graphic outputs. The variables to treat are splitted

6339

into two categories: the volumetric variables and those related to the

6340

boundary faces.\\

6341

6342

\noindent

6343

For more details about the different parameters, the user may refer to the

6344

key word list (\S\ref{prg_motscles}).

6345

6346

6347

%==================================

6348

\subsubsection{Management of the radiation boundary conditions:

6349

\textmd{\texttt{usray2}}}

6350

%==================================

6351

6352

\noindent

6353

\textit{Subroutine called every time step.}

6354

6355

This is the second subroutine is necessary for every calculation

6356

including radiative thermal transfers. It is used to give all the

6357

necessary parameters concerning, in the one case, the wall temperature

6358

calculation, and in the other, the coupling between the termal

6359

scalar (temperature or enthalpy) and the radiation module at the

6360

calculation domain boundaries. It must be noted that the boundary conditions

6361

concerning the thermal scalar which may have been defined in the

6362

subroutine \texttt{usclim} will be modified by the radiation module

6363

according to the data given in \texttt{usray2} (cf. \S\ref{fvm_selector}).\\

6364

A zone number must be given to each boundary face \footnote{this must be less

6365

than the maximum allowable by the code, \texttt{nozrdm}. This is fixed at 2000

6366

in \texttt{radiat.h} and cannot be modified.}and, specifically for

6367

the walls, a boundary condition type and an initialisation temperature

6368

(in Kelvin). The initialisation temperature is only used to make the

6369

solving implicit at the first time step. The zone number allows to assign

6370

an arbitrary integer to a set of boundary faces having the same

6371

radiation boundary condition type. This gathering is used by the

6372

calculation, and in the listing to print some physical values (mean

6373

temperature, net radiative flux ...). An independent graphic output in

6374

{\em EnSight} format is associated with each zone and allows the display on

6375

the boundary faces of the variables selected in the third heading of the

6376

subroutine \texttt{usray1}.\\

6377

A boundary condition type stored in the array ISOTHP is associated with

6378

each boundary face. There are five different types:

6379

6380

\begin{list}{$\bullet$}{}

6381

6382

\item \texttt{\textbf{itpimp}}: wall face with imposed temperature,

6383

6384

\item \texttt{\textbf{ipgrno}}: for a gray or black wall face, calculation of the

6385

temperature by means of a flux balance,

6386

6387

\item \texttt{\textbf{iprefl}}: for a reflecting wall face, calculation of the

6388

temperature by means of a flux balance.

6389

This is fixed at 2000 in \texttt{radiat.h} and cannot be modified.

6390

6391

\item \texttt{\textbf{ifgrno}}: gray or black wall face to which a conduction

6392

flux is imposed,

6393

6394

\item \texttt{\textbf{ifrefl}}: reflecting wall face to which a conduction

6395

flux is imposed, which is equivalent to impose this flux directly

6396

to the fluid.

6397

6398

\end{list}

6399

6400

\noindent

6401

Depending on the selected boundary condition type at every wall face,

6402

the code needs to be given some supplementary pieces of information:

6403

6404

\begin{list}{$\bullet$}{}

6405

6406

\item \texttt{\textbf{itpimp}}: the array \texttt{tintp} must be completed

6407

with the imposed temperature value and the array \texttt{epsp} must

6408

be completed with the emissivity value (strictly positive).

6409

6410

\item \texttt{\textbf{ipgrno}}: must be given: an initialisation temperature in

6411

the array \texttt{tintp}, the wall emissivity (strictly positive, in

6412

\texttt{epsp}), thickness (in \texttt{epap}), thermal conductivity

6413

(in \texttt{xlamp}) and an external temperature (in \texttt{textp})

6414

in order to calculate a conduction flux across the wall.

6415

6416

\item \texttt{\textbf{iprefl}}: must be given: an initialisation temperature (in

6417

\texttt{tintp}), the wall thickness (in \texttt{epap}) and thermal conductivity (in

6418

\texttt{xlamp}) and an external temperature (in \texttt{textp}).

6419

6420

\item \texttt{\textbf{ifgrno}}: must be given: an initialisation temperature (in

6421

\texttt{tintp}), the wall emissivity (in \texttt{epsp}) and the conduction

6422

flux (in $W/m^2$ whatever the thermal scalar, enthalpy or temperature) in

6423

the array \texttt{rcodcl}. The value of \texttt{rcodcl} is positive when the

6424

conduction flux is directed from the inside of the fluid domain to the

6425

outside (for instance, when the fluid heats the walls). If the

6426

conduction flux is null, the wall is adiabatic.

6427

6428

\item \texttt{\textbf{ifrefl}}: must be given: an initialisation temperature (in

6429

\texttt{tintp}) and the conduction flux (in $W/m^2$ whatever the thermal

6430

scalar) in the array \texttt{rcodcl}. The value of \texttt{rcodcl} is

6431

positive when the conduction flux is directed from the inside of the

6432

fluid domain to the outside (for instance, when the fluid heats the

6433

walls). If the conduction flux is null, the wall is adiabatic. The flux

6434

received by \texttt{rcodcl} is directly imposed as boundary condition for

6435

the fluid.

6436

6437

\end{list}

6438

6439

\noindent

6440

{\em WARNING: it is obligatory to set a zone number to every boundary

6441

face, even those which are not wall faces. These zones will be used during the

6442

printing in the listing. It is recommended to gather together the

6443

boundary faces of the same type, in order to ease the reading of the

6444

listing.}\\

6445

6446

%==================================

6447

\subsubsection{Absorption coefficient of the medium, boundary conditions

6448

for the luminance and calcualtion of the net radiative flux:

6449

\textmd{\texttt{usray3}}}

6450

%==================================

6451

6452

\noindent

6453

\textit{Subroutine called every time step.}

6454

6455

This subroutine is composed of three parts. In the first one, the user

6456

must provide the absorption coefficient of the medium in the array CK,

6457

for each cell of the fluid mesh. By default, the absorption coefficient

6458

of the medium is 0, which corresponds to a transparent medium.\\

6459

6460

{\em WARNING: when a specific physics is activated, it is forbidden to

6461

give a value to the absorption coefficient in this subroutine. In this

6462

case, it is calculated automatically, or given by the user {\em via} a

6463

thermo-chemical parameter file (dp\_C3P or dp\_C3PSJ for gas combustion,

6464

and dp\_FCP for pulverised coal combustion).}\\

6465

6466

\noindent

6467

The two following parts of this subroutine concern a more advanced use

6468

of the radiation module. It is about imposing boundary conditions to the

6469

equation of radiative transfer and net radiative flux calculation, in

6470

coherence with the luminance at the boundary faces, when the user wants

6471

to give it a particular value. In most cases, the given examples do not

6472

need to be modified.

6473

6474

%==================================

6475

\subsubsection{Encapsulation of the temperature-enthalpy conversion:

6476

\textmd{\texttt{usray4}}}

6477

%==================================

6478

6479

\noindent

6480

\textit{Subroutine called every time step.}

6481

6482

This subroutine is used to call the subroutine \texttt{usthht}. The user

6483

can implement his own conversion formulas into it. \\

6484

This subroutine is useless when the thermal scalar is the temperature.\\

6485

\noindent

6486

6487

{\em WARNING: when a specific physics is activated, it is forbidden to use this

6488

subroutine. In this case, {\em \texttt{usray4}} is replaced by {\em

6489

\texttt{ppray4}}, which is not a user subroutine.}\\

6490

6491

6492

\noindent

6493

The value of the argument \texttt{mode} allows to know in which direction the

6494

conversion will be made:

6495

\begin{list}{$\bullet$}{}

6496

6497

\item \texttt{\textbf{mode = 1}}: the fluid enthalpy in the cell must be

6498

converted into temperature (in Kelvin),

6499

6500

\item \texttt{\textbf{mode = -1}}: the wall temperature (\texttt{text}

6501

or \texttt{tparoi}, in Kelvin) must be converted into enthalpy.

6502

6503

\end{list}

6504

6505

{\em WARNING: the value of \texttt{mode} is passed as argument and must not be

6506

modified by the user.}\\

6507

6508

6509

6510

%==================================

6511

\subsection{Utilisation of a specific physics: \textmd{\texttt{usppmo}}}

5681

\subsection{Non-standard management of the chronological record files}

5682

%==================================

5683

\label{prg_ushist}

5684

5685

The interface and the subroutine \texttt{usini1} allow to manage the

5686

``automatic'' chronological record files in an autonomous way:

5687

position of the probes, printing frequency and related variables. The

5688

results are written in a different file for each variable. These files

5689

are written in {\em xmgrace or {\em gnuplot}} format and contain the profiles corresponding to

5690

every probe. This type of output format may not be well adapted if, for

5691

instance, the number of probes is too high. The subroutine

5692

\texttt{ushist}, called at each time step, allows then to personalise the output format of the

5693

chronological record files. The version given as example in the

5694

directory works as follows:

5695

5696

\begin{list}{-}{}

5697

\item Positionning of the probes (only at the first passage): the index

5698

\texttt{ii} varies between 1 and the number of probes. The coordinates

5699

\texttt{xx}, \texttt{yy} and \texttt{zz} of each probe are given.

5700

The subroutine \texttt{findpt}

5701

gives then the number \texttt{icapt(ii)\index{icapt}} of the cell center

5702

which is the closest to the defined probe.

5703

5704

\item Opening of the output files (only at the first pass): in the

5705

version given as example, the program opens a different file for

5706

all the \texttt{nvar} variables. \texttt{ficush(j)} contains the name of the

5707

J\raisebox{1ex}{\small th} file and \texttt{impush(j)} its unit number

5708

(\texttt{impush} is initialised by default so that the user has at his

5709

disposal specific unit numbers and does not run the risk to overwrite an

5710

already open file).

5711

5712

\item Writing to the files: in the version given as example, the program

5713

writes the time step number, the physical time step (based on the

5714

standard time step in the case of a variable time step) and the

5715

value of the selected variable at the different probes.

5716

5717

\item Closing of the files (only at the last time step).

5718

5719

\end{list}

5720

5721

{\em WARNING: The use of {\em\texttt{ushist}} neither erases nor replaces the

5722

parameters given in the interface or in {\em\texttt{usini1}}. Therefore, in

5723

the case of the use of {\em\texttt{ushist}}, and to avoid the creation

5724

of useless files, the user should set {\em \texttt{ncapt}=0} in the interface or

5725

in {\em \texttt{usini1}} to deactivate the automatic production of

5726

chronological records}.\\

5727

In addition, {\em \texttt{ushist}} generates supplementary result

5728

files.

5729

5730

5731

%==================================

5732

\section{Advanced modelling setup}

5733

%==================================

5734

5735

%==================================

5736

\subsection{Use of a specific physics}

6512

5737

%==================================

6513

5738

\label{prg_usppmo}%

6514

\noindent

6515

\textit{Subroutine called only during calculation initialisation.}

6516

6517

This is one of the three subroutines which must be obligatory completed

6518

by the user in order to use a specific physics module.

5739

Specific physics such as dispersed phase, atmospheric flows and coal combustion models can be added by the user from the interface, or by using the subroutine \texttt{usppmo} (called only during the calculation initialisation). With the interface, when a specific physics is activated in fig. \ref{fig:5_GUI}, additional items or headings may appear (see for instance Sections \ref{sec:Ini-lag} and \ref{sec:Ini-coal}).

5740

5741

\begin{figure}[!ht]

5742

\begin{center}

5743

\includegraphics[width=14cm]{gui_thermo_phys_models}

5744

\caption{Thermophysical models selection}

5745

\label{fig:5_GUI}

5746

\end{center}

5747

\end{figure}

5748

5749

When the interface is not used, \texttt{usppmo} is one of the three subroutines which must be obligatory completed by the user in order to use a specific physics module. Also, some specific physics modules can not yet be activated through the interface such as the modules listed below which were not quoted at the beginning of this section.

6519

5750

At the moment, \CS allows to use two ``pulverised coal'' modules

6520

(Lagrangian coupling or not), two ``gas combustion'' modules, two

6521

``electric'' modules and one ``compressible'' module. To activate one of

5751

(with Lagrangian coupling or not), two ``gas combustion'' modules, two

5752

``electrical'' modules, a ``compressible'' module, an ``cooling towers'' module and an ``atmospheric'' module. To activate one of

6522

5753

these modules, the user needs to complete one (and only one) of the

6523

5754

indicators \texttt{ippmod(i.....)\index{ippmod}} in the subroutine

6524

5755

\texttt{usppmo}. By default, all the indicators \texttt{ippmod(i.....)} are

6553

5784

\item \texttt{ippmod(icolwc)}=2 three peak model with adiabiatic conditions.

6554

5785

\item \texttt{ippmod(icolwc)}=3 three peak model with permeatic conditions.

6555

5786

\item \texttt{ippmod(icolwc)}=4 four peak model with adiabiatic conditions.

6556

\item \texttt{ippmod(icolwc)}=5 four peak model with permeatic condintions.

5787

\item \texttt{ippmod(icolwc)}=5 four peak model with permeatic conditions.

6557

5788

\item \texttt{ippmod(icolwc)}=-1 module not activated.

6558

5789

\end{list}

6559

5790

\item Multi-coals and multi-classes pulverised coal combustion:

6606

5837

\item \texttt{ippmod(icompf)} = 0 module activated

6607

5838

\item \texttt{ippmod(icompf)} =-1 module not activated

6608

5839

\end{list}

6609

5840

\item atmospheric flow module: indicator {\bf \tt

5841

ippmod(iatmos\index{icompf})}

5842

\begin{list}{$\rightarrow$}{}

5843

\item \texttt{ippmod(iatmos)} =-1 module not activated

5844

\item \texttt{ippmod(iatmos)} = 0 standard modelling

5845

\item \texttt{ippmod(iatmos)} = 1 dry atmosphere

5846

\item \texttt{ippmod(iatmos)} = 2 humid atmosphere (NOT functional)

5847

\end{list}

5848

\item cooling towers module: indicator {\bf \tt

5849

ippmod(iaeros\index{icompf})}

5850

\begin{list}{$\rightarrow$}{}

5851

\item \texttt{ippmod(iaeros\index)} =-1 module not activated

5852

\item \texttt{ippmod(iaeros\index)} = 0 no model (NOT functional)

5853

\item \texttt{ippmod(iaeros\index)} = 1 Poppe's model

5854

\item \texttt{ippmod(iaeros\index)} = 2 Merkel's model

5855

\end{list}

6610

5856

\end{list}

6611

5857

6612

5858

{\em WARNING: Only one specific physics module can be activated at the

6727

5973

\tiny{

6728

5974

\begin{tabular}{|c|c|c|c|} \hline

6729

5975

Lines & Examples of values & Variables & Observations \\ \hline

6730

1 & THERMOCHIMIE & & Comment line \\ \hline

5976

1 & THERMOCHEMISTRY & & Comment line \\ \hline

6731

5977

2 & 8 & \texttt{ncoel\index{ncoel}} & Number of current species \\ \hline

6732

5978

3 & 8 & \texttt{npo\index{npo}} & Number of points for the \\

6733

5979

& & & enthalpy-temperature tabulation \\ \hline

6734

4 & ESPECES COURANTES & & Comment line \\ \hline

5980

4 & CURRENT SPECIES & & Comment line \\ \hline

6735

5981

5 & CH4 C2H4 CO O2 CO2 H2O N2 C(S) & \texttt{nomcoel\index{nomcoel}}(\texttt{ncoel}) & List of the \\

6736

5982

& & & current species \\ \hline

6737

5983

6 & 300. & \texttt{tmin\index{tmin}} & Temperature inferior limit (Kelvin) \\

6738

5984

& & & for the enthalpy-temperature tabulation \\ \hline

6739

5985

7 & 2400. & \texttt{tmax\index{tmax}} & Temperature superior limit (Kelvin) \\

6740

5986

& & & for the enthalpy-temperature tabulation \\ \hline

6741

8 & 4 & \texttt{nato\index{nato}} & Number of elemental species \\ \hline

5987

8 & 4 & \texttt{nato\index{nato}} & Number of elementary species \\ \hline

6742

5988

9 & .012 1 2 1 0 1 0 0 1 & & Molar mass of the elemental species \\

6743

5989

10 & .001 4 4 0 0 0 2 0 0 &\texttt{wmolat\index{wmolat}}(\texttt{nato}), & (first column) \\

6744

5990

11 & .016 0 0 1 2 2 1 0 0 &\texttt{atcoel\index{atcoel}}(\texttt{ncoel,nato})& and composition of the current species \\

6745

5991

12 & .014 0 0 0 0 0 0 2 0 & & as a function of the elemental species \\ \hline

6746

13 & RAYONNEMENT & & Comment line \\ \hline

5992

13 & RADIATION & & Comment line \\ \hline

6747

5993

14 & 0.1 & \texttt{ckabs1\index{ckabs1}} & Constant absorption coefficient \\

6748

5994

& & & for the gas mixture \\ \hline

6749

15 & CARACTERISTIQUES CHARBONS & & Comment line \\ \hline

5995

15 & COAL CHARACTERISTICS & & Comment line \\ \hline

6750

5996

16 & 2 & \texttt{ncharb\index{ncharb}} & Number of coal types \\ \hline

6751

5997

17 & 1 1 & \texttt{nclpch\index{nclpch}}(\texttt{ncharb}) & Number of classes for each coal \\

6752

5998

& & & (each column corresponding to \\

6773

6019

28 & 0. 0. & \texttt{ock\index{ock}}(\texttt{ncharb}) & Composition in O (mass.\%, dry) of the coke \\

6774

6020

& & & for each coal \\ \hline

6775

6021

29 & 0. 0. & \texttt{pcick\index{pcick}}(\texttt{ncharb})& PCI of the dry coke ($Jkg^{-1}$) for each coal \\ \hline

6776

30 & Cendres & & Comment line \\ \hline

6022

30 & Ashes & & Comment line \\ \hline

6777

6023

31 & 6.3 6.3 & \texttt{xashch\index{xashch}}(\texttt{ncharb})& Ash mass fraction (mass.-\%, dry) in each coal \\ \hline

6778

6024

32 & 0. 0. & \texttt{h0ashc\index{h0ashc}}(\texttt{ncharb}) & Ash formation enthalpy ($Jkg^{-1}$) \\

6779

6025

& & & for each coal \\ \hline

6780

6026

33 & 0. 0. & \texttt{cpashc\index{cpashc}}(\texttt{ncharb}) & CP of the ashes ($Jkg^{-1}K^{-1}$) for each coal \\ \hline

6781

34 & D\'evolatilisation (Kobayashi) & & Comment line \\ \hline

6782

35 & 1 0.37 0 0.37 & \texttt{iy1ch\index{iy1ch}}(\texttt{ncharb}), & For each coal, pairs (\texttt{iy1ch}, \texttt{y1ch}). \\

6027

34 & 0. 0. & \texttt{xwatch\index{cpashc}}(\texttt{ncharb}) & humidity rate of the ashes (mass.-\%) for each coal \\ \hline

6028

35 & Devolatilisation (Kobayashi) & & Comment line \\ \hline

6029

36 & 1 0.37 0 0.37 & \texttt{iy1ch\index{iy1ch}}(\texttt{ncharb}), & For each coal, pairs (\texttt{iy1ch}, \texttt{y1ch}). \\

6783

6030

& & \texttt{y1ch\index{y1ch}}(\texttt{ncharb}) & The real \texttt{y1ch} is the adimensional stoich. coefficient\\

6784

6031

& & & If the integer \texttt{iy1ch} is worth 1, \\

6785

6032

& & & the provided value of \texttt{y1ch} is adopted and \\

6789

6036

& & & the provided value of \texttt{y1ch} is ignored: \\

6790

6037

& & & \texttt{y1ch} is calculated automatically (the light \\

6791

6038

& & & volatiles are then composed of {$CH_{4}$}, {$CO$}). \\ \hline

6792

36 & 1 0.74 1 0.74 & \texttt{iy2ch\index{iy2ch}}(\texttt{ncharb}), & For each coal, pairs (\texttt{iy2ch}, \texttt{y2ch}). \\

6039

37 & 1 0.74 1 0.74 & \texttt{iy2ch\index{iy2ch}}(\texttt{ncharb}), & For each coal, pairs (\texttt{iy2ch}, \texttt{y2ch}). \\

6793

6040

& & \texttt{y2ch\index{y2ch}}(\texttt{ncharb}) & The real \texttt{y2ch} is the adimensional stoich. coefficient\\

6794

6041

& & & If the integer \texttt{iy2ch} is worth 1, \\

6795

6042

& & & the provided value of \texttt{y2ch} is adopted and \\

6799

6046

& & & the provided value of \texttt{y2ch} is ignored: \\

6800

6047

& & & \texttt{y2ch} is calculated automatically (the heavy \\

6801

6048

& & & volatiles are then composed of {$C_{2}H_{4}$}, {$CO$}).\\ \hline

6802

37 & 370000. 410000. & \texttt{a1ch\index{a1ch}}(\texttt{ncharb})& Devolatilisation pre-exponential factor A1 ($s^{-1}$)\\

6803

& & & for each coal (light volatile matters) \\ \hline

6804

38 & 1.3E13 1.52E13 & \texttt{a2ch\index{a2ch}}(\texttt{ncharb})& Devolatilisation pre-exponential factor A2 ($s^{-1}$)\\

6805

& & & for each coal (heavy volatile matters) \\ \hline

6806

39 & 74000. 80000. & \texttt{e1ch\index{e1ch}}(\texttt{ncharb})& Devolatilisation activation energy E1 ($Jmol^{-1}$) \\

6807

& & & for each coal (light volatile matters) \\ \hline

6808

40 & 250000. 310000. & \texttt{e2ch\index{e2ch}}(\texttt{ncharb})& Energie d'activation E2 ($Jmol^{-1}$) de d\'evolatilisation\\

6809

& & & for each coal (heavy volatile matters) \\ \hline

6810

41 & Combustion h\'et\'erog\`ene & & Ligne de commentaire \\ \hline

6811

42 & 17.88 17.88 & \texttt{ahetch\index{ahetch}}(\texttt{ncharb}) & Char burnout pre-exponential constant \\

6812

& & & ($kgm^{-2}s^{-1}atm^{-1}$) for each coal \\ \hline

6813

43 & 16.55 16.55 & \texttt{ehetch\index{ehetch}}(\texttt{ncharb}) & Char burnout activation energy ($kcalmol{-1}$) \\

6814

& & & for each coal \\ \hline

6815

44 & 1 1 & \texttt{iochet\index{iochet}}(\texttt{ncharb}) & Char burnout reaction order for each coal \\

6816

& & & 0.5 if \texttt{iochet} = 0 and 1 if \texttt{iochet} = 1 \\ \hline

6049

38 & 370000. 410000. & \texttt{a1ch\index{a1ch}}(\texttt{ncharb})& Devolatilisation pre-exponential factor A1 ($s^{-1}$)\\

6050

& & & for each coal (light volatile matters) \\ \hline

6051

39 & 1.3E13 1.52E13 & \texttt{a2ch\index{a2ch}}(\texttt{ncharb})& Devolatilisation pre-exponential factor A2 ($s^{-1}$)\\

6052

& & & for each coal (heavy volatile matters) \\ \hline

6053

40 & 74000. 80000. & \texttt{e1ch\index{e1ch}}(\texttt{ncharb})& Devolatilisation activation energy E1 ($Jmol^{-1}$) \\

6054

& & & for each coal (light volatile matters) \\ \hline

6055

41 & 250000. 310000. & \texttt{e2ch\index{e2ch}}(\texttt{ncharb})& Activation energy E2 ($Jmol^{-1}$) of devolatilisation\\

6056

& & & for each coal (heavy volatile matters) \\ \hline

6057

42 & heterogeneous combustion $O_2$ & & Comment lign \\ \hline

6058

43 & 17.88 17.88 & \texttt{ahetch\index{ahetch}}(\texttt{ncharb}) & Char burnout pre-exponential constant \\

6059

& & & ($kgm^{-2}s^{-1}atm^{-1}$) for each coal \\ \hline

6060

44 & 16.55 16.55 & \texttt{ehetch\index{ehetch}}(\texttt{ncharb}) & Char burnout activation energy ($kcalmol^{-1}$) \\

6061

& & & for each coal \\ \hline

6062

45 & 1 1 & \texttt{iochet\index{iochet}}(\texttt{ncharb}) & Char burnout reaction order for each coal \\

6063

& & & 0.5 if \texttt{iochet} = 0 and 1 if \texttt{iochet} = 1 \\ \hline

6064

46 & heterogeneous combustion $CO_2$ & & Comment lign \\ \hline

6065

47 & 1.788 1.788 & \texttt{ahetch\index{ahetch}}(\texttt{ncharb}) & Char burnout pre-exponential constant \\

6066

& & & ($kgm^{-2}s^{-1}atm^{-1}$) for each coal \\ \hline

6067

48 & 1.655 1.655 & \texttt{ehetch\index{ehetch}}(\texttt{ncharb}) & Char burnout activation energy ($kcalmol^{-1}$) \\

6068

& & & for each coal \\ \hline

6069

49 & 1 1 & \texttt{iochet\index{iochet}}(\texttt{ncharb}) & Char burnout reaction order for each coal \\

6070

& & & 0.5 if \texttt{iochet} = 0 and 1 if \texttt{iochet} = 1 \\ \hline

6071

50 & OXYDIZERS CHARACTERISTICS & & Comment lign \\ \hline

6072

51 & 3 & \texttt{noxyd\index{noxyd}} & Number of oxydizers \\

6073

& & & (mixtures of $O_2,N_2,H_2O,CO_2$) \\ \hline

6074

52 & 1. 0. 1. & \texttt{oxyo2\index{oxyo2}}(\texttt{noxyd}) & Composition in $O_2$ of each oxydizer ($moles$) \\

6075

53 & 0. 0. 1. & \texttt{oxyn2\index{oxyn2}}(\texttt{noxyd}) & Composition in $N_2$ of each oxydizer ($moles$) \\

6076

54 & 0. 0. 1. & \texttt{oxyh2o\index{oxyh2o}}(\texttt{noxyd}) & Composition in $H_2O$ of each oxydizer ($moles$) \\

6077

55 & 2.39 1. 1. & \texttt{oxyco2\index{oxyco2}}(\texttt{noxyd}) & Composition in $CO_2$ of each oxydizer ($moles$) \\ \hline

6817

6078

\end{tabular}

6818

6079

}

6819

6080

\caption{Example of file for the pulverised coal combustion:

6821

6082

\end{center}

6822

6083

\end{table}

6823

6084

6824

6825

6085

\item Example of file for the electric arc: \texttt{dp\_ELE} (see

6826

6086

array \ref{tab_dpELE}).

6827

6087

6829

6089

\begin{center}

6830

6090

\small{

6831

6091

\begin{tabular}{|c|l|c|c|} \hline

6832

Li nes & Examples of values & Variables & Observations \\ \hline

6092

Lines & Examples of values & Variables & Observations \\ \hline

6833

6093

1 &\# Fichier ASCII format libre ... & & Free comment \\ \hline

6834

6094

2 &\# Les lignes de commentaires ... & & Free comment \\ \hline

6835

6095

3 &\# ... & & Free comment \\ \hline

6843

6103

& & & (\texttt{npo} $\leqslant$ \texttt{npot} set in \texttt{ppthch.h}) \\

6844

6104

& & & So there will be \texttt{ngazg} blocks of \texttt{npo} lines each \\ \hline

6845

6105

9 &\# ... & & Free comment \\ \hline

6846

10 &\# Proprietes ... & & Free comment \\ \hline

6847

11 &\# ~~~~T~~~~~~~~~~~H ... & & Free comment \\ \hline

6848

12 &\# Temperature Enthalpie ... & & Free comment \\ \hline

6849

13 &\# ... & & Free comment \\ \hline

6850

14 &\# ~~~~~K~~~~~~~~~J/kg ... & & Free comment \\ \hline

6106

14 & 0 & \texttt{ixkabe\index{ixkabe}} & Radiation options for \texttt{xkabe\index{xkabe}} \\ \hline

6851

6107

15 &\# ... & & Free comment \\ \hline

6852

16 & ~~~300.~~~~~~14000. ... & & Tabulation in line of the physical properties \\

6108

16 &\# Proprietes ... & & Free comment \\ \hline

6109

17 &\# ~~~~T~~~~~~~~~~~H ... & & Free comment \\ \hline

6110

18 &\# Temperature Enthalpie ... & & Free comment \\ \hline

6111

19 &\# ... & & Free comment \\ \hline

6112

20 &\# ~~~~~K~~~~~~~~~J/kg ... & & Free comment \\ \hline

6113

21 &\# ... & & Free comment \\ \hline

6114

22 & ~~~300.~~~~~~14000. ... & & Tabulation in line of the physical properties \\

6853

6115

& & & as a function of the temperature in Kelvin \\

6854

6116

& & & for each of the \texttt{ngazg} species \\

6855

6117

& & \texttt{h} & Enthalpy in J/kg \\

6869

6131

\end{list}

6870

6132

6871

6133

\clearpage

6872

%==================================

6873

\subsection{Management of the boundary conditions related to pulverised

6874

coal and gas combustion: \textmd{\texttt{usebuc, usd3pc, uslwcc,

6875

uscpcl and uscplc}} }

6876

%==================================

6877

6878

\noindent

6879

\textit{Subroutines called every time step.}

6880

6881

In this paragraph, ``specific physics'' refers to gas combustion or

6134

6135

6136

%==================================

6137

\subsection{Pulverised

6138

coal and gas combustion module}

6139

%==================================

6140

%==================================

6141

\subsubsubsection{Initialisation of the variables}\label{sec:Ini-coal}

6142

%==================================

6143

For coal combustion, it is possible to initialise the specific variables in the Graphical User Interface (GUI) or in the subroutines \texttt{usebui, usd3pi, uslwci} and \texttt{uscpiv}. In the GUI, when a coal combustion physics is selected in the item ``Calculation features'' under the heading ``Thermophysical models'', an additional item appears:``Pulverized coal combustion''. In this item the user can define coal types, its composition, the oxydant and reactions parameters, see figs. \ref{fig:Ini-coal1} to \ref{fig:Ini-coal5}.

6144

6145

\begin{figure}[!ht]

6146

\begin{center}

6147

\includegraphics[width=12cm]{gui_coal_classes}

6148

\caption{Thermophysical models - Pulverized coal combustion, coal classes}

6149

\label{fig:Ini-coal1}

6150

\end{center}

6151

\end{figure}

6152

6153

\begin{figure}[!ht]

6154

\begin{center}

6155

\includegraphics[width=11cm]{gui_coal_coke}

6156

\caption{Pulverized coal combustion, coke}

6157

\label{fig:Ini-coal2}

6158

\end{center}

6159

\end{figure}

6160

6161

\begin{figure}[!ht]

6162

\begin{center}

6163

\includegraphics[width=11cm]{gui_coal_composition}

6164

\caption{Pulverized coal combustion, coal composition}

6165

\label{fig:Ini-coal3}

6166

\end{center}

6167

\end{figure}

6168

6169

\begin{figure}[!ht]

6170

\begin{center}

6171

\includegraphics[width=11cm]{gui_coal_reaction}

6172

\caption{Pulverized coal combustion, reaction parameters}

6173

\label{fig:Ini-coal4}

6174

\end{center}

6175

\end{figure}

6176

6177

\begin{figure}[!ht]

6178

\begin{center}

6179

\includegraphics[width=11cm]{gui_coal_oxydant}

6180

\caption{Pulverized coal combustion, oxydant}

6181

\label{fig:Ini-coal5}

6182

\end{center}

6183

\end{figure}

6184

6185

If the user deals with gas combustion or if he (or she) does not want to use the GUI for coal combustion, the subroutines \texttt{usebui, usd3pi, uslwci} and \texttt{uscpiv} are used (only during the calculation initialisation).\\

6186

In this section, ``specific physics'' will refer to gas combustion or

6882

6187

to pulverised coal combustion.

6883

6188

6884

As are \texttt{usini1} and \texttt{usppmo}, the use of \texttt{usebuc},

6885

\texttt {usd3pc}, \texttt{uslwcc}, \texttt{uscpcl} or \texttt{uscplc} is

6886

obligatory to run a calculation concerning a specific physics

6887

modeling. The way of using them is the same as the way of using

6888

\texttt{usclim} in the framework of standard calculations, that is to

6889

say several loops on the boundary faces lists (cf. \S\ref{fvm_selector})

6189

These subroutines allow the user to initialise some variables specific

6190

to the specific physics activated {\em via} \texttt{usppmo}. As usual,

6191

the user may have access to several geometric variables to discriminate

6192

between different initialisation zones if needed.

6193

6194

{\em WARNING: in the case of a specific physics modeling, all the

6195

variables will be initialised here, even the potential user scalars: {\em

6196

\texttt{usiniv}} is no longer used.}

6197

6198

6199

\begin{list}{$\bullet$}{}

6200

\item in the case of the EBU pre-mixed flame module, the user can

6201

initialise in every cell \texttt{iel}: the mixing rate

6202

\texttt{rtp\index{rtp}(iel,isca(ifm))} in variable richness, the

6203

fresh gas mass fraction \\

6204

\texttt{rtp(iel,isca(iygfm\index{iygfm}))}

6205

and the mixture enthalpy \texttt{rtp(iel,isca(ihm\index{ihm}))} in

6206

permeatic conditions

6207

6208

\item in the case of the rapid complete chemistry diffusion flame

6209

module, the user can initialise in every cell \texttt{iel}: the

6210

mixing rate \texttt{rtp(iel,isca(ifm\index{ifm}))}, its variance

6211

\texttt{rtp(iel,isca(ifp2m\index{ifp2m}))} and the mixture mass

6212

enthalpy \texttt{rtp(iel,isca(ihm))} in permeatic conditions

6213

6214

\item in the case of the pulverised coal combustion module, the

6215

user can initialise in every cell \texttt{iel}:

6216

\begin{list}{$\rightarrow$}{}

6217

\item the transport variables related to the solid phase

6218

\begin{list}{}{}

6219

\item \texttt{rtp(iel,isca(ixch\index{ixch}(icla)))} the reactive coal mass fraction related to the class \texttt{icla} (\texttt{icla} from 1 to \texttt{nclacp} which is the total number of classes, {\em i.e.} for all the coal type)

6220

\item \texttt{rtp(iel,isca(ixck(\index{ixck}icla)))} the coke mass fraction related to the class \texttt{icla}

6221

\item \texttt{rtp(iel,isca(inp\index{inp}(icla)))} the number of particles related to class \texttt{icla} per kg of air-coal mixture

6222

\item \texttt{rtp(iel,isca(ih2\index{ih2}(icla)))} the mass enthalpy related to the class \texttt{icla} in permeatic conditions

6223

\end{list}

6224

\item \texttt{rtp(iel,isca(ihm))} the mixture enthalpy

6225

\item the transport variables related to the gas phase

6226

\begin{list}{}{}

6227

\item

6228

\texttt{rtp(iel,isca(if1m\index{if1m}(icha)))} the mean value of the tracer 1 representing the light volatile matters released by the coal \texttt{icha}

6229

\item

6230

\texttt{rtp(iel,isca(if2m\index{if2m}(icha)))} the mean value of the tracer 2 representing the heavy volatile matters released by the coal \texttt{icha}

6231

\item \texttt{rtp(iel,isca(if3m\index{if3m}))}

6232

the mean value of the tracer 3

6233

representing the carbon released

6234

as CO during coke burnout

6235

\item \texttt{rtp(iel,isca(if4p2m\index{if4p2m}))} the variance associated with the tracer 4 representing the air (the mean value of this tracer is not transported, it can be deduced directly from the three others)

6236

\item \texttt{rtp(iel,isca(ifp3m\index{ifp3m}))} the variance associated with the tracer 3

6237

\end{list}

6238

\end{list}

6239

\end{list}

6240

6241

%==================================

6242

\subsubsection{Boundary conditions}\label{sec:coal-cl}

6243

%==================================

6244

In this section, ``specific physics'' refers to gas combustion or

6245

to pulverised coal combustion.\\

6246

For coal combustion, it is possible to manage the boundary conditions in the Graphical User Interface (GUI). When the coal combustion physics is selected in the heading ``Thermophysical models'', specific boundary conditions are activated for inlets, see fig. \ref{fig:cond_lim-coal}. The user fills for each type of coal previously defined (see Section \ref{sec:Ini-coal}) the initial temperature and initial composition of the inlet flow, as well as the mass flow rate.

6247

6248

\begin{figure}[!ht]

6249

\begin{center}

6250

\includegraphics[width=13cm]{gui_coal_bc}

6251

\caption{Boundary conditions for the combustion of coal}

6252

\label{fig:cond_lim-coal}

6253

\end{center}

6254

\end{figure}

6255

6256

For gas combustion or if the GUI is not used for coal combustion, the use of \texttt{usebuc} (called at every time step),

6257

\texttt {usd3pc}, \texttt{uslwcc}, \texttt{uscpcl} or \texttt{uscplc} is as

6258

mandatory as \texttt{usini1} and \texttt{usppmo} to run a calculation involving specific physics. The way of using them is the same as using

6259

\texttt{usclim} in the framework of standard calculations, that is, run several loops on the boundary faces lists (cf. \S\ref{fvm_selector})

6890

6260

marked out by their colors, groups, or geometrical criterion, where

6891

6261

the type of face, the type of boundary condition for each variable and

6892

6262

eventually the value of each variable are defined.

6932

6302

velocity vector direction

6933

6303

by giving the components of

6934

6304

a direction vector in

6935

\texttt{rcodcl\index{rcodcl}(ifac,iu\index{iu}(iphas))}, \texttt{rcodcl(ifac,iv\index{iv}(iphas))} and \texttt{rcodcl(ifac,iw\index{iw}(iphas))}

6305

\texttt{rcodcl\index{rcodcl}(ifac,iu\index{iu})}, \texttt{rcodcl(ifac,iv\index{iv})} and \texttt{rcodcl(ifac,iw\index{iw})}

6936

6306

\end{list}

6937

6307

6938

6308

{\em WARNING:

6949

6319

\texttt{iqimp(izone)} the value 0 and set

6950

6320

the three velocity components (in

6951

6321

$m.s^{-1}$) in

6952

\texttt{rcodcl(ifac,iu(iphas))},

6953

\texttt{rcodcl(ifac,iv(iphas))} and

6954

\texttt{rcodcl(ifac,iw(iphas))}

6322

\texttt{rcodcl(ifac,iu)},

6323

\texttt{rcodcl(ifac,iv)} and

6324

\texttt{rcodcl(ifac,iw)}

6955

6325

\end{list}

6956

6326

\item finally he specifies for each gas inlet type

6957

6327

the mixing rate \texttt{fment\index{fment}(izone)} and

6994

6364

0). If the mass flow is imposed, the user

6995

6365

must set the air mass flow value

6996

6366

\texttt{qimpat\index{qimpat}(izone)}, its direction in

6997

\texttt{rcodcl(ifac,iu(iphas))}, \texttt{rcodcl(ifac,iv(iphas))}

6998

and \\ \texttt{rcodcl(ifac,iw(iphas))} and the incoming

6367

\texttt{rcodcl(ifac,iu)}, \texttt{rcodcl(ifac,iv)}

6368

and \\ \texttt{rcodcl(ifac,iw)} and the incoming

6999

6369

air temperature \texttt{timpat\index{timpat}(izone)} in

7000

6370

Kelvin. If the velocity is imposed, he has to

7001

set \texttt{rcodcl(ifac,iu(iphas))}, \\

7002

\texttt{rcodcl(ifac,iv(iphas))} and \texttt{rcodcl(ifac,iw(iphas))}.

6371

set \texttt{rcodcl(ifac,iu)}, \\

6372

\texttt{rcodcl(ifac,iv)} and \texttt{rcodcl(ifac,iw)}.

7003

6373

7004

6374

\item if the inlet belongs to the ``primary air and

7005

6375

pluverised coal'' \texttt{type (ientcp(izone) = 1)} the

7016

6386

\end{list}

7017

6387

7018

6388

%==================================

7019

\subsection{Initialisation of the variables related to pulverised

7020

coal and gas combustion: \textmd{\texttt{usebui, usd3pi, uslwci and uscpiv}}}

7021

%==================================

7022

7023

\noindent

7024

\textit{Subroutines called only during the calculation initialisation.}

7025

7026

In this paragraph, ``specific physics'' refers to gas combustion or

7027

to pulverised coal combustion.

7028

7029

These subroutines allow the user to initialise some variables specific

7030

to the specific physics activated {\em via} \texttt{usppmo}. As usual,

7031

the user may have access to several geometric variables to discriminate

7032

between different initialisation zones if needed.

7033

7034

{\em WARNING: in the case of a specific physics modeling, all the

7035

variables will be initialised here, even the eventual user scalars: {\em

7036

\texttt{usiniv}} is no longer used.}

7037

7038

7039

\begin{list}{$\bullet$}{}

7040

\item in the case of the EBU pre-mixed flame module, the user can

7041

initialise in every cell \texttt{iel}: the mixing rate

7042

\texttt{rtp\index{rtp}(iel,isca(ifm))} in variable richness, the

7043

fresh gas mass fraction \\

7044

\texttt{rtp(iel,isca(iygfm\index{iygfm}))}

7045

and the mixture enthalpy \texttt{rtp(iel,isca(ihm\index{ihm}))} in

7046

permeatic conditions

7047

7048

\item in the case of the rapid complete chemistry diffusion flame

7049

module, the user can initialise in every cell \texttt{iel}: the

7050

mixing rate \texttt{rtp(iel,isca(ifm\index{ifm}))}, its variance

7051

\texttt{rtp(iel,isca(ifp2m\index{ifp2m}))} and the mixture mass

7052

enthalpy \texttt{rtp(iel,isca(ihm))} in permeatic conditions

7053

7054

\item in the case of the pulverised coal combustion module, the

7055

user can initialise in every cell \texttt{iel}:

7056

\begin{list}{$\rightarrow$}{}

7057

\item the transport variables related to the solid phase

7058

\begin{list}{}{}

7059

7060

\item \texttt{rtp(iel,isca(ixck(\index{ixck}icla)))} the coke mass fraction related to the class \texttt{icla}

7061

\item \texttt{rtp(iel,isca(inp\index{inp}(icla)))} the number of particles related to class \texttt{icla} per kg of air-coal mixture

7062

\item \texttt{rtp(iel,isca(ih2\index{ih2}(icla)))} the mass enthalpy related to the class \texttt{icla} in permeatic conditions

7063

\end{list}

7064

\item \texttt{rtp(iel,isca(ihm))} the mixture enthalpy

7065

\item the transport variables related to the gas phase

7066

\begin{list}{}{}

7067

\item

7068

\texttt{rtp(iel,isca(if1m\index{if1m}(icha)))} the mean value of the tracer 1 representing the light volatile matters released by the coal \texttt{icha}

7069

\item

7070

\texttt{rtp(iel,isca(if2m\index{if2m}(icha)))} the mean value of the tracer 2 representing the heavy volatile matters released by the coal \texttt{icha}

7071

\item \texttt{rtp(iel,isca(if3m\index{if3m}))}

7072

the mean value of the tracer 3

7073

representing the carbon released

7074

as CO during coke burnout

7075

7076

\item \texttt{rtp(iel,isca(ifp3m\index{ifp3m}))} the variance associated with the tracer 3

7077

\end{list}

7078

\end{list}

7079

\end{list}

7080

7081

7082

%==================================

7083

\subsection{Initialisation of the options of the variables related to

7084

pulverised coal and gas combustion: \textmd{\texttt{usebu1, usd3p1,

7085

uslwc1, uscpi1 and uscpl1}}}

7086

%==================================

7087

7088

\noindent

7089

\textit{Subroutines called at calculation beginning.}

7090

7091

In this paragraph, ``specific physics'' refers to gas combustion or

6389

\subsubsection{Initialisation of the options of the variables}

6390

%==================================

6391

In the case of coal combustion, time averages, chronological records and listings follow-ups can be set in the Graphical User Interface (GUI) or in the subroutines \texttt{usebu1, usd3p1, uslwc1, uscpi1} and \texttt{uscpl1}. In the GUI, under the heading ``Calculation control'', additional variables appear in the list in the items ``Time averages'' and ``Profiles'', as well as in the item Volume solution control'', see figs. \ref{fig:t_average-coal} and \ref{fig:V_control-coal}.

6392

6393

\begin{figure}[!ht]

6394

\begin{center}

6395

\includegraphics[width=12cm]{gui_coal_time_average}

6396

\caption{Calculation control - Time averages}

6397

\label{fig:t_average-coal}

6398

\end{center}

6399

\end{figure}

6400

6401

\begin{figure}[!ht]

6402

\begin{center}

6403

\includegraphics[width=13cm]{gui_coal_solution_control}

6404

\caption{Calculation control - Volume solution control}

6405

\label{fig:V_control-coal}

6406

\end{center}

6407

\end{figure}

6408

6409

In this section, ``specific physics'' refers to gas combustion or

7092

6410

pulverised coal combustion.

7093

6411

7094

These 3 subroutines are used to complete \texttt{usini1} for the

7095

considered specific physics.

6412

For gas combustion or if the GUI is not used for coal combustion, the 3 subroutines \texttt{usebu1, usd3p1, uslwc1, uscpi1} and \texttt{uscpl1} can be used to complete \texttt{usini1} for the

6413

considered specific physics. These subroutines are called at the calculation start.

7096

6414

They allow to:

7097

6415

\begin{list}{$\bullet$}{}

7098

6416

\item generate, for the variables which are specific to the activated

7140

6458

\end{list}

7141

6459

\item rapid complete chemistry diffusion flame modeling:

7142

6460

\begin{list}{}{}

7143

\item everything is identical to the ``EBU'' case, except from

6461

\item everything is identical to the ``EBU'' case, except

7144

6462

the fresh gas mass fraction which is replaced by the

7145

6463

variance of the mixing rate \texttt{ivar=isca(ifp2m\index{ifp2m})}

7146

6464

\end{list}

7147

\item pulverised coal modeling with 3 comustables:

6465

\item pulverised coal modeling with 3 combustibles:

7148

6466

\begin{list}{}{}

7149

6467

\item {\em variables shared by the two phases}:

7150

6468

\begin{list}{-}{}

7251

6569

7252

6570

\item set the value of the constant \texttt{cebu\index{cebu}} of the Eddy Break

7253

6571

Up model (only in \texttt{usebu1}. By default \texttt{cebu}=2.5)

7254

7255

\end{list}

7256

7257

7258

7259

%==================================

7260

\subsection{Management of Boundary Conditions of the electric arc: \texttt{uselcl}}

7261

%==================================

7262

7263

7264

\noindent

7265

\textit {sub routine called at each time step.}

7266

7267

As in the \texttt{usini1} and \texttt{usppmo}, the use of \texttt{usecl}

7268

is required to run an electric calculation. The main use is the same as

7269

occurs in \texttt{usclim} for the standard \CS calculations, for which

7270

different loops on the boundary faces is defined. Each faces list is

7271

built with the use of selection criteria (cf. \S\ref{fvm_selector}),

7272

and is referenced by their group(s), their color(s) or geometrical

7273

criterions. The face type, the boundary conditions for each variable,

7274

and finally the value of each variable or imposed flow are fixed.

7275

7276

{\em WARNING:for the electric module, , the boundary conditions of all

7277

the variables are defined here,

7278

even those of the eventual user scalars: {\em \texttt{usclim}} is not

7279

used at all.}

7280

7281

For the electric module, each boundary face is associated with a number

7282

\texttt{izone} \footnote{\texttt{izone} must be less than the maximum

7283

value allowed by the code, \texttt{nozzppm}. This is fixed at 2000 in \texttt

7284

{ppvar.h} and cannot be modified.}(the color \texttt{icoul} for example) in

7285

order to group together all the boundary faces of the same type. In the report

7286

\texttt{usclim}, the main change from the users point of view concerns the

7287

specification of the boundary conditions of the potential, which isn't

7288

implied by default. The Dirichlet and Neumann conditions must be imposed

7289

explicitly using \texttt{icodcl} and \texttt{rcodcl} (as would be done for

7290

the classical scalar).

7291

7292

Whats more, if one wishes to slow down the power dissipation(Joule module

7293

effect) or the current (electric arc module) from the imposed values

7294

\texttt{(puismp\index{puismp}} and \texttt{couimp\index{couimp}} respectively),

7295

they can be changed by the potential scalar as shown below:

7296

7297

\begin{list}{-}{}

7298

\item For the electric arc, the imposed potential difference can be a fixed variable:

7299

for example, the cathode can be fixed at 0 and the potential at the anode

7300

contains the variable \texttt{dpot}. This variable is initialised in \texttt{useli1}

7301

by an estimated potential difference. If \texttt{ielcor}=1 (see

7302

\texttt{useli1}), \texttt{dpot} is updated automatically during the

7303

calculation to obtain the required current.

7304

\item For the Joule module effect, \texttt{dpot} is again used with the same

7305

signification as in the electric arc module. If \texttt{dpot} is not wanted

7306

in the setting of the boundary condtions, the variable \texttt{coejou} can be

7307

used. \texttt{coejou} is the coefficient by which the potential difference is

7308

multiplied to obtain the desired power dissipation. By default this begins at

7309

1 and is updated automatically. If \texttt{ielcor}=1 (see \texttt

7310

{useli1}), multiply the imposed potentials in \texttt{uselcl} by \texttt{coejou}

7311

at each time step to achieive the desired power dissipation.

7312

\end{list}

7313

7314

{\em WARNING: In alternative current, attention should be paid to the values of potential

7315

imposed at the limits: the variable named "real potential" represents an affective

7316

value if the current is in single phase, and a "real part" if not.}

7317

\begin{list}{-}{}

7318

\item For the Joule studies, a complex potential is someitmes needed

7319

(\texttt{ippmod(ieljou)}=2): this is the case in particular where the current

7320

has 3 phases. To have access to the phase of the potential, and not just its

7321

amplitude, the 2 variables must be deleted: in \CS, there are 2 arrays

7322

specified for this role, the real part and the imaginary

7323

part of the potential. For use in the code, these variables are named

7324

``real potential'' and ``imaginary potential''. For an alternative

7325

sinusoidal potential $Pp$, the maximum value is noted as $Pp_\text{max}$,

7326

the phase is noted as $\phi$, the real potential

7327

and the imaginary potential are respecively $Pp_\text{max}\,cos\phi$ and

7328

$Pp_\text{max}\,sin\phi$.

7329

\item For the Joule studies in which one does not have access to the phases, the real

7330

potential (imaginary part =0) will suffice (\texttt{ippmod(ieljou)=1}): this is obviously the case with

7331

continous current, but also with single phase alternative current. In \CS

7332

there is only 1 varialbe for the potential, called "real potential". Pay attention to

7333

the fact that in alternate current, the "real potential" represents a effective value

7334

of potential , $\frac{1}{\sqrt{2}}\,Pp_\text{max}$ (in continous current there is no

7335

such ambiguity).

7336

\end{list}

7337

7338

7339

%==================================

7340

\subsection{Initialisation of the variables in the electric module}

7341

%==================================

7342

7343

\noindent

7344

\textit{subroutine called only at the initialisation of the calculation}

7345

7346

This subroutine allows the user to initialise some of the specific physics variables

7347

prompted via \texttt{usppmo}

7348

. The user has access, as usual, to many geometric variables so that the zones can

7349

be differentiated if needed.

7350

7351

{\em WARNING: For the specific physics, it is here that all varialbes are initialsed:

7352

\texttt{usiniv} is not used}

7353

7354

This subroutine works like \texttt{usiniv}. The values of potential and its

7355

constiuents are initialised if required.

7356

7357

It should be noted that the enthalpy is important.

7358

7359

\begin{list}{-}{}

7360

\item For the electric arc module, the enthalpy value is taken from the temperature

7361

of reference \texttt{t0(iphas)} (given is \texttt{usini1}) from the temperature-enthalpy tables

7362

supplied in the data file \texttt{dp\_ELE.} The user must not intervene here.

7363

\item For the Joule effect module, the value of enthalpy must be specified by the user

7364

. An example is given of how to obtain the enthalpy from the temperature of reference

7365

T0 (IPHAS)(given in \texttt{usini1}), the the temperature-enthalpy low must be

7366

supplied. A code is suggested in the sub routine \texttt{usthht}(which is there for

7367

the determination of physical properties).

7368

\end{list}

7369

%====================================

7370

\subsection { Initialisation of the variable options in the electric module}

7371

%==================================

7372

\label{prg_useli1}%

7373

\noindent

7374

\textit{subroutine called at each time step}

7375

7376

This subroutine is completed in \texttt{usini1} for the specific physics. It allows:

7377

\begin{list}{$\bullet$}{}

7378

\item Activates the variables in the specific physics module, the chronolgical

7379

outputs (\texttt{ichrvr(ipp)} indicators), the listings (\texttt{ilisvr(ipp)}

7380

indicators) and the

7381

historical exits at the probes defined in \texttt{usini1} (\texttt{ihisvr(ipp)}

7382

indicators).

7383

The functions are the same as in \texttt{usini1} and the script frequency of the

7384

exits are fixed using \texttt{usini1.} The indicators \texttt{ipp} are for the value \texttt{ipp=ipppro} (\texttt{ipproc(ivar)}, with \texttt{ivar}, the number of specific physics varibles. With the main variables

7385

which concern the user (velocity, pressure, etc), the user must always use \texttt{usini1} if the history,the listings or the chronological files are required.

7386

The variables which the user can activate are marked out. The number of variables in

7387

the calculation is given in \texttt{ivar} (defined by

7388

\texttt{propce(iel,ipproc(iprop)} for cell \texttt{iel}):

7389

7390

\begin{list}{$\rightarrow$}{}

7391

\item Electric Arc Module:

7392

\begin{list}{-}{}

7393

\item Calculation variables \texttt{rtp(iel,ivar)}

7394

\begin{list}{\texttt{ivar} = }{}

7395

\item \texttt{isca(ihm\index{ihm})} enthalpy

7396

\item \texttt{isca(ipotr\index{ipotr})} real potentiel

7397

\item \texttt{isca(ipotva(i)\index{ipotva})} solved components of the potential vector.

7398

\item \texttt{isca(iycoel(iesp)\index{iycoel})} the mass fraction of \texttt{ngazg} composites if there are more than 1

7399

\end{list}

7400

\item Properties \texttt{propce(iel,ipproc(iprop))}

7401

\begin{list}{\texttt{iprop} = }{}

7402

\item \texttt{itemp\index{itemp}} temperature

7403

\item \texttt{iefjou\index{iefjou}} power dissipation by the Joule effect.

7404

\item \texttt{ilapla(i)\index{ilapla(i)}} components of the laplace forces.

7405

\end{list}

7406

\end{list}

7407

\item Joule Module effect~:

7408

\begin{list}{-}{}

7409

\item Calculation variables \texttt{rtp(iel,ivar)}

7410

\begin{list}{\texttt{ivar} = }{}

7411

\item \texttt{isca(ihm\index{ihm})} enthalpy

7412

\item \texttt{isca(ipotr\index{ipotr})} real potential

7413

\item \texttt{isca(ipoti\index{ipoti})} imaginary potential if its to be taken into account

7414

\item \texttt{isca(iycoel(iesp)\index{iycoel}}) the mass fraction of \texttt{ngazg} composites if there are more than 1

7415

\end{list}

7416

\item Properties \texttt{propce(iel,ipproc(iprop))}

7417

\begin{list}{\texttt{iprop} = }{}

7418

\item \texttt{itemp\index{itemp}} temperature

7419

\item \texttt{iefjou\index{iefjou}} volumic power dissipation by Joule effect.

7420

\end{list}

7421

\end{list}

7422

\end{list}

7423

7424

\item to give the coefficient of relaxation of the density \texttt{srrom}:\\

7425

$\rho^{n+1}=\texttt{srrom}*\rho^{n}+(1-\texttt{srrom})\rho^{n}$\\

7426

(for the electric arc, the sub-relaxation is taken into account during the 2nd time step; for the Joule effect the sub relaxation is not accounted for unless the user specifies in \texttt{uselph}

7427

7428

\item indicates if the data will be fixed in the power dissipation or in the current, done in \texttt{ielcor}.

7429

\item target current fixed as \texttt{couimp} (electric arc module) or the power dissipation \texttt{puism} (Joule module effect).

7430

\item Fix the initial value of potential difference \texttt{dpot},

7431

the for the calculations with a single fixed parameter as \texttt{couimp} or \texttt{puism}.

7432

7433

\end{list}

7434

7435

%==================================

7436

\subsection{Management of variable physical properties in the electric module}

7437

%==================================

7438

7439

\noindent

7440

\textit{Subroutine called at each time step}

7441

7442

All the laws of the variation of physical data of the fluid are written (where neccessary)

7443

in this subroutine... The subroutine replaces \texttt{usphyyv} and a similar component.

7444

7445

{\em WARNING: For the electric module, it is here that all the physical variables are defined

7446

(including the relative cells and the eventuel user scalars):\texttt{usepelph} is not used.}

7447

7448

The user should ensure that the defined variation laws are valid for the whole range of

7449

variables. Particular attention should be taken with the non-linear laws (for example, a

7450

3rd degree polynomial law giving negative values of density)

7451

7452

{\em WARNING: with the electric module, all the physical propertie are assumed as variables

7453

and so are stored in the \texttt{propce} array. \texttt{cp0}, \texttt{viscls0}, \texttt{viscl0} are not used}

7454

7455

For the Joule effect, the user is required to supply the physical properties in the sub-

7456

routine. Examples are given which are to be adapted by the user. If the temperature is

7457

to be determined to calculate the physical properties, the solved variable, enthalpy must

7458

be deduced. The preffered temperature-enthalpy law can be selected in the subruotine

7459

\texttt{usthht} (an example of the interpolation is given from the law table. This

7460

subroutine can be re-used for the initialisation of the variables(\texttt{useliv}))

7461

For the elecrtic arc module, the physical properties are intepolated from the data file

7462

\texttt{dp\_ELE} supplied by the user. Modification is not generally necessary.

7463

7464

7465

7466

%==================================

7467

\subsection[Management of the {\em EnSight} output in the electric module: \texttt{uselen}]

7468

{Management of the post-processing output in the electric module: \textmd{\texttt{uselen}}}

7469

%==================================

7470

7471

\noindent

7472

\textit{Subroutine called at each chronological output}

7473

7474

This subroutine allows the addition on $n$ variables in the preprocessing output and

7475

works like the subroutine \texttt{usvpst} (with the electric module, it is however also

7476

possible to use \texttt{usvpst}.

7477

7478

The algebraic variables related to the electric module are provided by default provided

7479

that they are not explicitely contained in the \texttt{propce} array:

7480

\begin{list}{-}{}

7481

\item gradient of real potential in $V m^{-1}$ ($\grad Pot_R = -\vect{E}$)

7482

\item density of real current in $A m^{-2}$ ($\vect{j}=\sigma \vect{E}$)

7483

\end{list}

7484

specifically for the Joule module effect with \texttt{ippmod(ieljou)}=2~:

7485

\begin{list}{-}{}

7486

\item gradient of imaginary potential in $V m^{-1}$

7487

\item density of real current in $A m^{-2}$

7488

\end{list}

7489

specifically for the electric arc module with \texttt{ippmod(ielarc)}=2~:

7490

\begin{list}{-}{}

7491

\item magnetic field in $T$ (\vect{B} = \vect{rot}\,\vect{A})

7492

\end{list}

7493

7494

If it is convenient for the user, there is no need to add this subroutine into the

7495

SRC directory: the post-processing will be done automatically (at the same frequency

7496

(\texttt{NTCHR}) as the other calculation variables)

7497

7498

7499

%==================================

7500

\subsection{Compressible module}

7501

%==================================

7502

7503

When the compressible module\footnote{For more details concerning the

7504

compressible version, the user may refer to the document ``Implantation

7505

d'un algorithme compressible dans \CS'', Rapport EDF 2003,

7506

HI-83/03/016/A, P. Mathon, F. Archambeau et J.-M. H\'erard.} is

7507

activated, it is recommended to:

7508

\begin{list}{-}{}

7509

\item use the option ``time step variable in time and uniform in

7510

space'' (\texttt{idtvar}=1) with a maximum Courant number of 0.4

7511

(\texttt{coumax}=0.4): these choices must be written in \texttt{usini1}

7512

\item keep the convective numerical schemes proposed by default.

7513

\end{list}

7514

7515

%==================================

7516

\subsubsection{ Initialisation of the options of the variables related

7517

to the compressible module: \textmd{\texttt{uscfx1}} and \textmd{\texttt{uscfx2}}}

7518

%==================================

7519

\label{prg_uscfx12}%

7520

\noindent

7521

\textit{Subroutine called every time step.}

7522

7523

These subroutines complete \texttt{usini1}.

7524

7525

\texttt{uscfx1} allows to set non standard calculation options related to the

7526

compressible module, and in particular to fill in the key word \texttt{icfgrp}

7527

allowing to take into account the hydrostatic equilibrium in the

7528

boundary conditions.

7529

7530

\texttt{uscfx2} allows to specify for the molecular thermal conductivity and

7531

the volumetric viscosity the following pieces of information:

7532

\begin{list}{-}{}

7533

\item variable or not (\texttt{iviscv})

7534

\item reference value (\texttt{viscv0})

7535

\end{list}

7536

7537

%==================================

7538

\subsubsection{Management of the boundary conditions related to the

7539

compressible module: \textmd{\texttt{uscfcl}}}

7540

%==================================

7541

7542

7543

\noindent

7544

\textit{Subroutine called every time step.}

7545

7546

The use of \texttt{uscfcl}

7547

is obligatory to run a calculation using the compressible module just

7548

as it is in both \texttt{usini1} and \texttt{usppmo} . The

7549

way of using it is the same as the way of using

7550

\texttt{usclim} in the framework of standard calculations, that is to

7551

say several loops on the boundary faces lists (cf. \S\ref{fvm_selector})

7552

marked out by their colors, groups, or geometrical criterion, where

7553

the type of face, the type of boundary condition for each variable and

7554

eventually the value of each variable are defined.

7555

7556

{\em WARNING: in the case of a calculation using the compressible

7557

module, the boundary conditions of all the variables are defined here,

7558

even those of the eventual user scalars: {\em \texttt{usclim}} is not

7559

used at all.}

7560

7561

In the compressible module, the different available boundary conditions

7562

are the followings:

7563

7564

\begin{list}{-}{}

7565

\item inlet/outlet for which everything is known

7566

\item supersonic outlet

7567

\item subsonic inlet

7568

\item subsonic wall

7569

\item wall

7570

\item symmetry

7571

\end{list}

7572

7573

7574

%==================================

7575

\subsubsection{Ininitialisation of the variables related to the

7576

compressible module: \textmd{\texttt{uscfxi}}}

7577

%==================================

7578

7579

\noindent

7580

\textit{Subroutine called only during calculation initialisation.}

7581

7582

This subroutine is used to initialise some variables specific to the

7583

specific physics activated {\em via} \texttt{usppmo}. As usual,

7584

the user may have access to several geometric variables to discriminate

7585

between different initialisation zones if needed.

7586

7587

{\em WARNING: in the case of a specific physics modeling, all the

7588

variables are initialised here: {\em \texttt{usiniv}} is not used at

7589

all.}

7590

7591

This subroutine works like \texttt{usiniv} for velocity,

7592

turbulence and passive scalars. Concerning pressure, density,

7593

temperature and specific total energy, only 2 variables out of the 4 are

7594

independant. The user may also initialise the variable pair he wants

7595

(apart from temperature-energy) and the two other variables will be

7596

calculated automatically by giving the right value to the variable

7597

\texttt{iccfth} used for the call to \texttt{uscfth}.

7598

7599

%==================================

7600

\subsubsection{Compressible module thermodynamics: \textmd{\texttt{uscfth}}}

7601

%==================================

7602

7603

\noindent

7604

\textit{This subroutine is called several times every time step (boundary conditions, physical properties, solving of the energy equation, ...).}

7605

7606

This subroutine is used to set the thermodynamics parameters. By

7607

default, the perfect gas laws are implemented. If the user needs to use

7608

other laws (perfect gas with variable Gamma, Van der Waals), he must

7609

modify this subroutine.

7610

7611

%==================================

7612

\subsubsection{Management of the variable physical properties in the

7613

compressible module: \textmd{\texttt{uscfpv}}}

7614

%==================================

7615

7616

\noindent

7617

\textit{Subroutine called every time step.}

7618

7619

If necessary, all the variation laws of the fluid physical properties

7620

(viscosity, specific heat, ...) are described here. This subroutine

7621

replaces and is similar to \texttt{usphyv}.

7622

7623

The user should make sure that the defined variation laws are valid for

7624

the whole variation range of the variables.

7625

7626

%==================================

7627

\subsection{Lagrangian modeling of multiphasic flows with dipersed inclusions}

7628

%==================================

7629

7630

7631

%==================================

7632

\subsubsection{Initialisation of the main key words in the Lagrangian

7633

modeling: \textmd{\texttt{uslag1}}}

7634

%==================================

7635

7636

\noindent

7637

\textit{Subroutine called only during calculation initialisation.}

7638

7639

\noindent

7640

This is one of the two subroutines which must be completed in

7641

the case of a calculation modeling a Lagrangian multiphasic flow. This

6572

\end{list}

6573

6574

%==================================

6575

\subsection{Heavy fuel oil combustion module}

6576

%==================================

6577

%==================================

6578

\subsubsection{Initialisation of transported variables}

6579

%==================================

6580

To initialise or modify (in case of a continuation) values of transported variables and of the time step, the subroutine \texttt{usfuiv} is used. It is similar to \texttt{usiniv}. It is called at the beginning of every computation (new or continuation) before the time loop.

6581

6582

Physical properties are stored in \texttt{propce} (cell center), \texttt{propfa} (inner face) and \texttt{propfb}. For instance, \texttt{propce(iel, ipproc(irom ))} is \texttt{rom(iel)}, the mean density (in $kg.m^{-3}$), and \texttt{propfa(ifac,ipprof(ifluma(ivar)))} is \texttt{flumas(IFAC,IVAR)}, the convective flux of the variable \texttt{ivar}.\\

6583

Physical properties (\texttt{rom, viscl, cp, ...}) are computed in \texttt{ppphyv} and are not to be modified here.

6584

6585

All cells can be identified by using the subroutine '\texttt{getcel}'. All boundary faces may be identified using the '\texttt{getfbr}' subroutine. All internal faces may be identified using the '\texttt{getfac}' subroutine. Details of the syntax of these three subroutines are given in \texttt{usfuiv}.

6586

6587

In \texttt{usfuiv} the user initialise quantities related to the turbulent model chosen, and to gaseous species and droplets compositions. Exemples are provided in the subroutine.

6588

6589

%==================================

6590

\subsubsection{Boundary conditions}

6591

%==================================

6592

Boundary conditions are defined on a per-face basis in \texttt{usfucl}. Boundary faces may be identified using the '\texttt{getfbr}' subroutine. \texttt{usfucl} is very similar to \texttt{uscpcl}, see Section \ref{sec:coal-cl}. Boundary conditions may be assigned in two ways:

6593

\begin{list}{.}{}

6594

\item for ``standard'' boundary conditions (inlet, free outlet, wall, symmetry): a code is defined in the array \texttt{itypfb} (of dimensions equal to the number of boundary faces). This code will then be used by a non-user subroutine to assign the conditions.

6595

\item for ``non-standard'' conditions: see details given in \texttt{usfucl}.

6596

\end{list}

6597

6598

%==================================

6599

\subsubsection{Initialisation of the options of the variables}

6600

%==================================

6601

6602

The presence of a fuel combustion module variable in the listing, {\itshape histo} files, and the output frequency are set in the subroutine \texttt{usfui1}. If the vectors below are not allocated, default values will be used:

6603

\begin{list}{-}{}

6604

\item \texttt{ichrvr}: chronological output (1:yes / 0:no)

6605

\item \texttt{ilisvr}: listing output (1:yes / 0:no)

6606

\item \texttt{ihisvr}: {\itshape histo} output (number of probes and probe numbers), if $= -1$, every probes defined in \texttt{usini1} will be found in the {\itshape histo} files

6607

\end{list}

6608

6609

Calculation options such as a the relaxation parameter the for density (recommended when starting a combustion computation but forbidden for unstationnary computations) can also be set, as well as physical constants like the the laminar viscosity for the enthalpy.

6610

6611

%==================================

6612

\subsection{Radiative thermal transfers in semi-transparent gray media}

6613

%==================================

6614

%==================================

6615

\subsubsection{Initialisation of the radiation main parameters}

6616

%==================================

6617

6618

The main radiation parameters can be initialise in the Graphical User Interface (GUI) or in the user subroutine \texttt{usray1}. In the GUI, under the heading ``Thermophysical models'', when one of the two thermal radiative transfers models is selected, see fig. \ref{fig:0_ray}, additional items appear. The user is asked to choose the number of directions for angular discretisation, to define the absorption coefficient and select if the radiative calculation are restarted or not, see figs. \ref{fig:1_ray} and \ref{fig:3_ray}. When ``Advanced options'' is selected for both models figs. \ref{fig:2_ray} or \ref{fig:4_ray} appear, the user must fill the resolution frequency and verbosity levels. In addition, the activation of the radiative transfer leads to the creation of an item ``Surface solution control'' under the heading ``Calculation control'', see fig. \ref{fig:5_ray}, where radiative transfer variables can be selected to appear in the output listing.

6619

6620

\begin{figure}[ht]

6621

\begin{center}

6622

\includegraphics[width=13cm]{gui_rad_transf_models}

6623

\caption{Radiative transfers models}

6624

\label{fig:0_ray}

6625

\end{center}

6626

\end{figure}

6627

6628

\begin{figure}[ht]

6629

\begin{center}

6630

\includegraphics[width=10cm]{gui_rad_transf_do_params}

6631

\caption{Radiative transfers - parameters of the DO method}

6632

\label{fig:1_ray}

6633

\end{center}

6634

\end{figure}

6635

6636

\begin{figure}[ht]

6637

\begin{center}

6638

\includegraphics[width=7cm]{gui_rad_transf_do_advanced}

6639

\caption{Radiative transfers - advanced parameters of the DO method}

6640

\label{fig:2_ray}

6641

\end{center}

6642

\end{figure}

6643

6644

\begin{figure}[ht]

6645

\begin{center}

6646

\includegraphics[width=10cm]{gui_rad_transf_p1_params}

6647

\caption{Radiative transfers - parameters of the P-1 model}

6648

\label{fig:3_ray}

6649

\end{center}

6650

\end{figure}

6651

6652

\begin{figure}[ht]

6653

\begin{center}

6654

\includegraphics[width=7cm]{gui_rad_transf_p1_advanced}

6655

\caption{Radiative transfers - advanced parameters of th P-1 model}

6656

\label{fig:4_ray}

6657

\end{center}

6658

\end{figure}

6659

6660

\begin{figure}[ht]

6661

\begin{center}

6662

\includegraphics[width=6cm]{gui_rad_transf_post_output}

6663

\caption{Calculation control - Radiative transfers postprocessign output}

6664

\label{fig:5_ray}

6665

\end{center}

6666

\end{figure}

6667

6668

If the GUI is not used, \texttt{usray1} is one of the two subroutine which must be completed by the user for all

6669

calculations including radiative thermal transfers. It is called only during the calculation initialisation. It is composed of three headings. The first one is dedicated to the activation

6670

of the radiation module, only in the case of classic physics. \\

6671

{\em WARNING: when a calculation is ran using a specific physics module,

6672

this first heading must not be completed. The radiation module is then

6673

activated or not, according to the parameter file related to the considered

6674

specific physics.} \\

6675

6676

\noindent

6677

In the second heading the basic parameters of the radiation module are indicated.\\

6678

Finally, the third heading deals with the selection of the

6679

post-processing graphic outputs. The variables to treat are splitted

6680

into two categories: the volumetric variables and those related to the

6681

boundary faces.\\

6682

6683

\noindent

6684

For more details about the different parameters, the user may refer to the

6685

key word list (\S\ref{prg_motscles}).

6686

6687

6688

%==================================

6689

\subsubsection{Radiative transfers boundary conditions}

6690

%==================================

6691

These informations can be filled by the user through the Graphical User Interface (GUI) or by using the subroutine \texttt{usray2} (called every time step). If the interface is used, when one of the ``Radiative transfers'' options is selected in fig. \ref{fig:0_ray}, it activates specific boundary conditions each time a ``Wall'' is defined, see fig. \ref{fig:6_ray}. The user can then choose between 3 cases. The parameters the user must specify are displayed for one of them in fig. \ref{fig:7_ray}.

6692

6693

\begin{figure}[ht]

6694

\begin{center}

6695

\includegraphics[width=11cm]{gui_rad_transf_wall_model}

6696

\caption{Boundary conditions - choice of wall thermal radiative transfers}

6697

\label{fig:6_ray}

6698

\end{center}

6699

\end{figure}

6700

6701

\begin{figure}[ht]

6702

\begin{center}

6703

\includegraphics[width=11cm]{gui_rad_transf_wall_params}

6704

\caption{Boundary conditions - example of wall thermal radiative transfer}

6705

\label{fig:7_ray}

6706

\end{center}

6707

\end{figure}

6708

6709

When the GUI is not used, \texttt{usray2} is the second subroutine necessary for every calculation which includes radiative thermal transfers. It is used to give all the

6710

necessary parameters concerning, in the one case, the wall temperature

6711

calculation, and in the other, the coupling between the thermal

6712

scalar (temperature or enthalpy), and the radiation module at the

6713

calculation domain boundaries. It must be noted that the boundary conditions

6714

concerning the thermal scalar which may have been defined in the

6715

subroutine \texttt{usclim} will be modified by the radiation module

6716

according to the data given in \texttt{usray2} (cf. \S\ref{fvm_selector}).\\

6717

A zone number must be given to each boundary face \footnote{this must be less

6718

than the maximum allowable by the code, \texttt{nozrdm}. This is fixed at 2000

6719

in \texttt{radiat.h} and cannot be modified.}and, specifically for

6720

the walls, a boundary condition type and an initialisation temperature

6721

(in Kelvin). The initialisation temperature is only used to make the

6722

solving implicit at the first time step. The zone number allows to assign

6723

an arbitrary integer to a set of boundary faces having the same

6724

radiation boundary condition type. This gathering is used by the

6725

calculation, and in the listing to print some physical values (mean

6726

temperature, net radiative flux ...). An independent graphic output in

6727

{\em EnSight} format is associated with each zone and allows the display on

6728

the boundary faces of the variables selected in the third heading of the

6729

subroutine \texttt{usray1}.\\

6730

A boundary condition type stored in the array ISOTHP is associated with

6731

each boundary face. There are five different types:

6732

6733

\begin{list}{$\bullet$}{}

6734

6735

\item \texttt{\textbf{itpimp}}: wall face with imposed temperature,

6736

6737

\item \texttt{\textbf{ipgrno}}: for a gray or black wall face, calculation of the

6738

temperature by means of a flux balance,

6739

6740

\item \texttt{\textbf{iprefl}}: for a reflecting wall face, calculation of the

6741

temperature by means of a flux balance.

6742

This is fixed at 2000 in \texttt{radiat.h} and cannot be modified.

6743

6744

\item \texttt{\textbf{ifgrno}}: gray or black wall face to which a conduction

6745

flux is imposed,

6746

6747

\item \texttt{\textbf{ifrefl}}: reflecting wall face to which a conduction

6748

flux is imposed, which is equivalent to impose this flux directly

6749

to the fluid.

6750

6751

\end{list}

6752

6753

\noindent

6754

Depending on the selected boundary condition type at every wall face,

6755

the code needs to be given some supplementary pieces of information:

6756

6757

\begin{list}{$\bullet$}{}

6758

6759

\item \texttt{\textbf{itpimp}}: the array \texttt{tintp} must be completed

6760

with the imposed temperature value and the array \texttt{epsp} must

6761

be completed with the emissivity value (strictly positive).

6762

6763

\item \texttt{\textbf{ipgrno}}: must be given: an initialisation temperature in

6764

the array \texttt{tintp}, the wall emissivity (strictly positive, in

6765

\texttt{epsp}), thickness (in \texttt{epap}), thermal conductivity

6766

(in \texttt{xlamp}) and an external temperature (in \texttt{textp})

6767

in order to calculate a conduction flux across the wall.

6768

6769

\item \texttt{\textbf{iprefl}}: must be given: an initialisation temperature (in

6770

\texttt{tintp}), the wall thickness (in \texttt{epap}) and thermal conductivity (in

6771

\texttt{xlamp}) and an external temperature (in \texttt{textp}).

6772

6773

\item \texttt{\textbf{ifgrno}}: must be given: an initialisation temperature (in

6774

\texttt{tintp}), the wall emissivity (in \texttt{epsp}) and the conduction

6775

flux (in $W/m^2$ whatever the thermal scalar, enthalpy or temperature) in

6776

the array \texttt{rcodcl}. The value of \texttt{rcodcl} is positive when the

6777

conduction flux is directed from the inside of the fluid domain to the

6778

outside (for instance, when the fluid heats the walls). If the

6779

conduction flux is null, the wall is adiabatic.

6780

6781

\item \texttt{\textbf{ifrefl}}: must be given: an initialisation temperature (in

6782

\texttt{tintp}) and the conduction flux (in $W/m^2$ whatever the thermal

6783

scalar) in the array \texttt{rcodcl}. The value of \texttt{rcodcl} is

6784

positive when the conduction flux is directed from the inside of the

6785

fluid domain to the outside (for instance, when the fluid heats the

6786

walls). If the conduction flux is null, the wall is adiabatic. The flux

6787

received by \texttt{rcodcl} is directly imposed as boundary condition for

6788

the fluid.

6789

6790

\end{list}

6791

6792

\noindent

6793

{\em WARNING: it is mandatory to set a zone number to every boundary

6794

face, even those which are not wall faces. These zones will be used during the

6795

printing in the listing. It is recommended to gather together the

6796

boundary faces of the same type, in order to ease the reading of the

6797

listing.}\\

6798

6799

%==================================

6800

\subsubsection{Absorption coefficient of the medium, boundary conditions

6801

for the luminance and calculation of the net radiative flux}

6802

%==================================

6803

When the absorption coefficient is not constant, the subroutine \texttt{usray3} is called instead at each time step. It is composed of three parts. In the first one, the user

6804

must provide the absorption coefficient of the medium in the array CK,

6805

for each cell of the fluid mesh. By default, the absorption coefficient

6806

of the medium is 0, which corresponds to a transparent medium.\\

6807

6808

{\em WARNING: when a specific physics is activated, it is forbidden to

6809

give a value to the absorption coefficient in this subroutine. In this

6810

case, it is calculated automatically, or given by the user {\em via} a

6811

thermo-chemical parameter file (dp\_C3P or dp\_C3PSJ for gas combustion,

6812

and dp\_FCP for pulverised coal combustion).}\\

6813

6814

\noindent

6815

The two following parts of this subroutine concern a more advanced use

6816

of the radiation module. It is about imposing boundary conditions to the

6817

equation of radiative transfer and net radiative flux calculation, in

6818

coherence with the luminance at the boundary faces, when the user wants

6819

to give it a particular value. In most cases, the given examples do not

6820

need to be modified.

6821

6822

%==================================

6823

\subsubsection{Encapsulation of the temperature-enthalpy conversion}

6824

%==================================

6825

6826

\noindent

6827

\textit{Subroutine called every time step.}

6828

6829

The user subroutine \texttt{usray4} is used to call the user subroutine \texttt{usthht}. \texttt{usthht} is used to encapsulate a simple enthalpy-temperature

6830

conversion law and its inverse. The user

6831

can implement his own conversion formulas into it. \\

6832

This subroutine is useless when the thermal scalar is the temperature.\\

6833

\noindent

6834

6835

{\em WARNING: when a specific physics is activated, it is forbidden to use this

6836

subroutine. In this case, {\em \texttt{usray4}} is replaced by {\em

6837

\texttt{ppray4}}, which is not a user subroutine.}\\

6838

6839

\noindent

6840

The value of the argument \texttt{mode} allows to know in which direction the

6841

conversion will be made:

6842

\begin{list}{$\bullet$}{}

6843

6844

\item \texttt{\textbf{mode = 1}}: the fluid enthalpy in the cell must be

6845

converted into temperature (in Kelvin),

6846

6847

\item \texttt{\textbf{mode = -1}}: the wall temperature (\texttt{text}

6848

or \texttt{tparoi}, in Kelvin) must be converted into enthalpy.

6849

6850

\end{list}

6851

6852

{\em WARNING: the value of \texttt{mode} is passed as argument and must not be

6853

modified by the user.}\\

6854

6855

%==================================

6856

\subsubsection{Input of radiative transfer parameters}

6857

%==================================

6858

6859

\noindent

6860

\textit{The routine \texttt{usray5} is called twice. The first time is for boundary

6861

conditions. The second time is for the net radiation flux computation}

6862

6863

In this subroutine, during the first call (\texttt{iappel=1}), the boundary conditions

6864

are filled:

6865

\begin{list}{-}{}

6866

\item the radiative intensity must be set in the array \texttt{cofrua} when the discrete

6867

ordinates model is used; an example is given in \texttt{usray5} for an isotropic radiation

6868

field on a gray wall. Proposed boundary conditions for the intensity in \texttt{usray5} are:

6869

symmetry, inlet/oulet, and wall boundary,

6870

\item the entering intensity for free boundaries is set to zero in \texttt{cofrua} (if the

6871

user has more information, he can improve it),

6872

\item arrays \texttt{cofrua} and \texttt{cofrub} must be filled when the P-1 model is

6873

used. The boundary conditions proposed are the same as with the discret ordinates model.

6874

\end{list}

6875

During the second call (\texttt{iappel=2}), the density of the net radiation flux must be

6876

calculated consistently with the boundary conditions of the intensity considering that the

6877

density of net flux is the balance between the radiative emiting part of a boundary face

6878

(and not the reflecting one) and the radiative absorbing part. The provided example is

6879

consistent with the example of the intensity boundary conditions given when the discret

6880

ordinates model is used.

6881

6882

6883

%==================================

6884

\subsection{Conjugate heat transfers}

6885

%==================================

6886

%========================================

6887

\subsubsection{Thermal module in a 1D wall}

6888

%========================================

6889

6890

\noindent

6891

\textit{subroutine called at every time step}

6892

6893

This subroutine takes into account the affected thermal inertia by a wall.

6894

Some boundary faces are treated as a solid wall with a given thickness, on

6895

which the code resolves an undimensional equation for the heat conduction.

6896

The coupling between the 1D module and the fluid works in a similar way to

6897

the coupling with the \syrthes. In construction, the user is not able to

6898

account for the heat transfer between different parts of the wall. A physical

6899

analysis of each problem, case by case is required to evaluate the relevance

6900

of its usage by way of a report of the simple conditions (temperature, zero-flux

6901

) or a coupling with \syrthes.\\

6902

6903

The use of this code requires that the thermal scalar is

6904

defined as (\texttt{iscalt}$>0$).

6905

6906

{\em WARNING: The 1D thermal module is developped assuming the thermal scalar

6907

as a temperature. If the thermal scalar is an enthalpy, the code calls the

6908

subroutine \texttt{usthht} for each transfer of information between the fluid

6909

and the wall in order to convert the enthalpy to temperature and vice-versa.

6910

This function has not been tested and is firmly discouraged. If the thermal

6911

variable is the total (compressible) energy, the thermal module will not work.}

6912

6913

\bigskip

6914

6915

This procedure is called twice,on initialisation and again at each time step.

6916

6917

\begin{list}{$\bullet$}{}

6918

\item The 1st call (initialisation) all the boundary faces that will be treated

6919

as a coupled wall are marked out. This figure is written noted as

6920

\texttt{nfkpt1d}. It applies dimension to the arrays in the thermal module.

6921

\texttt{nfkpt1d} will be at 0 if there are no coupled faces (it is in fact the

6922

default value, the remainer of the subroutine is not used in this case).

6923

The parameter \texttt{isuit1} also need to be defined, this indicates if the

6924

temperature of the wall must be initialised or written in the file (stored in

6925

the variable \texttt{filmt1}).

6926

\item The 2nd call (initialisation) again concern the wall faces, it completes

6927

the \texttt{ifpt1d} array of dimension \texttt{nfpt1d}.

6928

\mbox{\texttt{ifpt1d(ifbt1d)}} is the number

6929

\texttt{ifbt1d}\raisebox{1ex}{\small th} boundary faces coupled with the thermal module

6930

of a 1D wall. The directional parameters are then completed for a pseudo

6931

wall associated to each face

6932

\begin{list}{-}{}

6933

\item \texttt{nppt1d(nfpt1d)\index{nppt1d}}: number of cells in the 1D mesh associated

6934

to the pseudo wall.

6935

\item \texttt{eppt1d(nfpt1d)\index{eppt1d}}: thickness of the pseudo wall.

6936

\item \texttt{rgpt1d(nfpt1d)\index{rgpt1d}}: geometery of the pseudo wall mesh (refined

6937

as a fluid if \texttt{rgt1d} is smaller than 1)

6938

\item \texttt{tppt1d(nfpt1d)\index{tppt1d}}: initialisation temperature of the wall

6939

(uniform in thickness). In the course of the calculation, the array stores the

6940

temperature of the solid at the fluid/solid interface.

6941

\end{list}

6942

6943

Other than for re-reading a file (\texttt{ficmt1}), \texttt{tppt1d} is not used.

6944

\texttt{nppt1d}, \texttt{ifpt1d}, \texttt{rgpt1d} and \texttt{eppt1d} are

6945

compared to data from the follow-up file and they must be identical.

6946

6947

{\em WARNING: The test in \texttt{ifpt1d} implicilty assumes that the array is completed

6948

in ascending order (i.e \texttt{ifpt1d(ii)}$>$\texttt{ifpt1d(jj)} if ii$>$jj.

6949

This will be the case if the coupled faces are defined starting from the unique loop on the

6950

boundary faces (as in the example). If this is not the case, contact the development

6951

team to short circuit the test.}

6952

6953

\item The 3rd call (at each time step) is for the confirmation that all the arrays

6954

involving physical parameter and external boundary conditions have been completed.

6955

\begin{list}{-}{}

6956

\item \texttt{iclt1d(nfpt1d)\index{iclt1d}}:Typical boundary condition at the external

6957

(pseudo) wall: Dirichlet condition (\texttt{iclt1d}=1) or flux condition (\texttt{iclt1d}=3)

6958

\item \texttt{tept1d(nfpt1d)\index{tept1d}}: External temperature of the pseudo wall in the

6959

Dirichlet case.

6960

\item \texttt{hept1d(nfpt1d)\index{hept1d}}: External coefficient of transfer in the pseudo

6961

wall under Dirichlet conditions (in $W.m^{-2}.K^.$).

6962

\item \texttt{fept1d(nfpt1d)\index{nfpt1d}}: External heat flux in the pseudo wall under

6963

the flux conditions(in $W.m^{-2}$, negative value for energy entering the wall).

6964

\item \texttt{xlmt1d(nfpt1d)\index{xlmt1d}}: Conductivity$\lambda$ of the wall uniform

6965

in thickness (in $W.m^{-1}.K^{-1}$).

6966

\item \texttt{rcpt1d(nfpt1d)\index{rcpt1d}}: Volumetric heat capacity $\rho C_p$ of the

6967

wall uniform in thickness (in $J.m^{-3}.K^{-1}$).

6968

\item \texttt{dtpt1d(nfpt1d)\index{dtpt1d}}: Physical time step ascociated with the solved

6969

1D equation of the pseudo wall(which can be different from the time step in the

6970

calculation).

6971

\end{list}

6972

6973

\end{list}

6974

6975

The $3^{rd}$ call, done at each time step, allows to impose boundary conditions

6976

and physical values in time.

6977

6978

%==================================

6979

\subsubsection{Fluid-Thermal coupling with \syrthes}

6980

%==================================

6981

When the user wishes to couple \CS with \syrthes to include heat transfers, it can be

6982

done in the Graphical User Interface (GUI) or in the user subroutine

6983

\texttt{cs\_syrthes\_coupling}.

6984

In the GUI, to set such a coupling, a thermal scalar must be

6985

selected first in the item ``Thermal scalar'' under the heading ``Thermophysical models''.

6986

Then the item ``Conjugate heat transfer'' will appear, see fig. \ref{fig:syrthes}.

6987

The zones where the coupling occurs must be defined and a projection axis can be

6988

specified in case of 2D coupling.

6989

6990

\begin{figure}[ht]

6991

\begin{center}

6992

\includegraphics[width=8cm]{gui_syrthes_coupling}

6993

\caption{Thermophysical models - coupling with \syrthes}

6994

\label{fig:syrthes}

6995

\end{center}

6996

\end{figure}

6997

6998

If the subroutine \texttt{ussyrc} is used, the user must specify the arguments

6999

passed to the subroutine '\texttt{defsyr}'. These arguments are:

7000

\begin{list}{-}{}

7001

\item \texttt{numsyr} is the matching \syrthes application \texttt{id} number, or $-1$,

7002

\item \texttt{namsyr} is the matching \syrthes application name,

7003

\item \texttt{cprjsy}: ' ' if the user wishes to use a 3D standard coupling,

7004

or specify '$x$', '$y$', or '$z$' as the projection axis if a 2D coupling with \syrthes is used,

7005

\item \texttt{critsu} is the surface selection criteria,

7006

\item \texttt{critvl} is the volume selection criteria (only with \syrthes 4),

7007

\item \texttt{iwarns} is the verbosity level.

7008

\end{list}

7009

Examples are provided in '\texttt{ussyrc}'.

7010

7011

7012

%==================================

7013

\subsection{Lagrangian modeling of multiphase flows with dipersed inclusions}

7014

%==================================

7015

7016

%==================================

7017

\subsubsection{Initialisation of the Lagrangian

7018

modeling parameters}\label{sec:Ini-lag}

7019

%==================================

7020

7021

The initialisation of the Lagrangian module parameters can be performed in

7022

the Graphical User Interface (GUI) or in the user subroutine \texttt{uslag1}

7023

(called only during the calculation initialisation). In the GUI, the selection

7024

of the Lagrangian module in the item ``Calculation features'' under the heading

7025

``Thermophysical models'' activates the heading ``Particle and droplets

7026

tracking''. The initialisation is performed in the three items included in

7027

this heading. In ``Global settings'', the user defines the Eulerian/Lagrangian

7028

multi-phase treatment, the main parameters, the specific physics associated with

7029

the particles and numerical adavanced options, see figs. \ref {fig:Ini-Lag1}

7030

to \ref {fig:Ini-Lag3}. In the item ``Statistics'', names are associated to

7031

volume and boundary statistical variables for listing and post-processing,

7032

see fig. \ref {fig:Ini-Lag4}. In the item ``Output'', the user defines the

7033

output frequency, post-precessing options for particles and selects the variables

7034

that will appear in the listing, see.fig. \ref {fig:Ini-Lag5}.

7035

7036

\begin{figure}[ht]

7037

\begin{center}

7038

\includegraphics[width=12cm]{gui_lagr_global_settings}

7039

\caption{Lagrangian module - global settings}

7040

\label{fig:Ini-Lag1}

7041

\end{center}

7042

\end{figure}

7043

7044

\begin{figure}[ht]

7045

\begin{center}

7046

\includegraphics[width=10cm]{gui_lagr_global_settings_coal}

7047

\caption{Lagrangian module - global settings, specific physics}

7048

\label{fig:Ini-Lag2}

7049

\end{center}

7050

\end{figure}

7051

7052

\begin{figure}[ht]

7053

\begin{center}

7054

\includegraphics[width=8cm]{gui_lagr_global_advanced}

7055

\caption{Lagrangian module - global settings, advanced numerical options}

7056

\label{fig:Ini-Lag3}

7057

\end{center}

7058

\end{figure}

7059

7060

\begin{figure}[ht]

7061

\begin{center}

7062

\includegraphics[width=11cm]{gui_lagr_statistics}

7063

\caption{Lagrangian module - statistics}

7064

\label{fig:Ini-Lag4}

7065

\end{center}

7066

\end{figure}

7067

7068

\begin{figure}[ht]

7069

\begin{center}

7070

\includegraphics[width=11cm]{gui_lagr_output}

7071

\caption{Lagrangian module - output}

7072

\label{fig:Ini-Lag5}

7073

\end{center}

7074

\end{figure}

7075

7076

\noindent

7077

When the GUI is not used, \texttt{uslag1} is one of the two subroutines which must be completed in

7078

the case of a calculation using a Lagrangian multiphase flow model. This

7642

7079

subroutine gathers in different headings all the key word which are

7643

7080

necessary to configure the Lagrangian module. The different headings

7644

7081

refer to:

7687

7124

modeling, the user must make sure that the volumetric statistics are

7688

7125

directly used for the calculation of the locally undisturbed fluid flow

7689

7126

field. \\

7690

7691

7127

\noindent

7692

7128

When the thermal evolution of the particles is activated, the associated

7693

7129

particulate scalars are always the inclusion temperature and the locally

7714

7150

7715

7151

%==================================

7716

7152

\subsubsection{Management of the boundary conditions related to the

7717

particles: \textmd{\texttt{uslag2}} and \textmd{\texttt{uslain}}}

7153

particles}

7718

7154

%==================================

7155

The boundary conditions related to particles can be defined in the

7156

Graphical User Interface (GUI) or in the subroutines \texttt{uslag2} and

7157

\texttt{uslain}. In the GUI, the selection of the Lagrangian module in the

7158

item ``Calculation features'' under the heading ``Thermophysical models''

7159

activates the item ``Particle boundary conditions'' under the heading

7160

``Boundary conditions''. Different options are available depending on the

7161

type of standard boundary conditions selected (wall, inlet/outlet, etc...),

7162

see fig. \ref{fig:CL-Lag}.

7163

7164

\begin{figure}[ht]

7165

\begin{center}

7166

\includegraphics[width=16cm]{gui_lagr_bc}

7167

\caption{Lagrangian module - boundary conditions}

7168

\label{fig:CL-Lag}

7169

\end{center}

7170

\end{figure}

7719

7171

7720

7172

\noindent

7721

In the framework of the multiphasic lagrangrian modeling, the management

7173

In the framework of the multiphase lagrangrian modeling, the management

7722

7174

of the boundary conditions concerns the particle behaviour when there

7723

7175

is an interaction between its trajectory and a boundary face. These

7724

7176

boundary conditions may be imposed independently of those concerning the

7739

7191

7740

7192

\noindent

7741

7193

It is the second indispensable subroutine for every calculation using the

7742

Lagrangian module. The main numerical variables and ``pointers'' are

7743

described below.

7194

Lagrangian module. The main numerical variables are described below.

7744

7195

7745

7196

\variab{ifrlag}{ifrlag(nfabor)}{ia}{In the Lagrangian module, the user

7746

7197

defines \texttt{nfrlag\index{nfrlag}} boundary zones from the color of the

7824

7275

7825

7276

\end{list}

7826

7277

}

7827

7828

7829

7278

\variablist{iuslag}{iuslag(nclagm, nflagm, ndlaim)}{ia}{Some pieces of

7830

7279

information must be given for each particle class associated with an

7831

7280

injection zone. The first part consists in integers contained in the

7868

7317

7869

7318

\end{list}

7870

7319

}

7871

7872

7320

\variablist{ruslag}{ruslag(nclagm, nflagm, ndlagm)}{ra}{Some pieces of

7873

7321

information must be given for each particle class associated with an

7874

7322

injection zone. The second and last part consists in real numbers

7944

7392

7945

7393

\end{list}

7946

7394

}

7947

7948

7395

\variab{iusvis}{iusvis(nflagm)}{ia}{In order to display the variables at

7949

7396

the boundaries defined in the subroutine \texttt{uslag1}, this array

7950

7397

allows to select the boundary zones on which a display is wanted. To do

7962

7409

\textit{Subroutine called every time step.}

7963

7410

7964

7411

\noindent

7965

It is not obligatory to intervene in this subroutine.

7412

It is not mandatory to intervene in this subroutine.

7966

7413

7967

7414

\noindent

7968

7415

\texttt{uslain} is used to complete \texttt{uslag2} when the particles

7980

7427

the calculation domain at the current iteration are marked out by an index

7981

7428

varying between \texttt{nbpart}+1 and \texttt{nbpnew}.

7982

7429

7983

7984

%==================================

7985

\subsubsection{Treatment of the particle/boundary interaction:

7986

\textmd{\texttt{uslabo}}}

7987

%==================================

7988

7989

\noindent

7990

\textit{Subroutine called at every particle/boundary interaction.}

7991

7992

\noindent

7993

It is not obligatory to intervene in this subroutine, but it is required

7994

in four different cases.

7430

\clearpage

7431

7432

%==================================

7433

\subsubsection{Treatment of the particle/boundary interaction}

7434

%==================================

7435

7436

\noindent

7437

The subroutine \texttt{uslabo} is not mandatory but is required in four different

7438

cases. It is called for each particle/boundary interaction.

7995

7439

7996

7440

\noindent

7997

7441

Firstly, an intervention is required when \texttt{jbord}* type boundary

8077

7521

second-order scheme is used elsewhere.

8078

7522

8079

7523

%==================================

8080

\subsubsection{Option for particle cloning/merging: \textmd{\texttt{uslaru}}}

7524

\subsubsection{Option for particle cloning/merging}

8081

7525

%==================================

8082

7526

8083

7527

\noindent

8084

7528

\textit{Subroutine called every Lagrangian iteration.}

8085

7529

8086

7530

\noindent

8087

An intervention in this subroutine is required if the particle

7531

An intervention in the subroutine \texttt{uslaru} is required when the particle

8088

7532

cloning/merging option is activated {\em via} the key word \texttt{iroule}. The

8089

importance function \texttt{croule} must then be completed. \\

7533

important function '\texttt{croule}' must then be completed. \\

8090

7534

The aim of this technique is to reduce the number of particles to treat in

8091

7535

the whole flow and to refine the description of the particle cloud only

8092

where the user wants to get volumetric statistics more accurate than in the

7536

where the user wants to get more accurate volumetric statistics than in the

8093

7537

rest of the calculation domain. \\

8094

7538

The values given to the importance function are strictly positive real

8095

7539

numbers allowing to classify the zones according to their

8117

7561

8118

7562

%==================================

8119

7563

\subsubsection{Manipulation of particulate variables at the end of an

8120

iteration and user volumetric statistics: \textmd{\texttt{uslast}} and

8121

\textmd{\texttt{uslaen}}}

7564

iteration and user volumetric statistics}

8122

7565

%==================================

8123

7566

8124

7567

\noindent

8181

7624

\end{list}

8182

7625

8183

7626

%==================================

8184

\subsubsection{User stochastic differential equations:

8185

\textmd{\texttt{uslaed}}}

7627

\subsubsection{User stochastic differential equations}

8186

7628

%==================================

8187

7629

8188

7630

\noindent

8189

\textit{Subroutine called every Lagrangian sub-step.}

8190

8191

\noindent

8192

An intervention in this subroutine is required if supplementary user

7631

An intervention in the subroutine \texttt{uslaed} is required if supplementary user

8193

7632

variables are added to the particle state vector (arrays \texttt{ettp}

8194

and \texttt{ettpa}).

7633

and \texttt{ettpa}). This subroutine is called at each Lagrangian sub-step.

8195

7634

8196

7635

\noindent

8197

7636

The integration of the stochastic differential equations associated with

8237

7676

8238

7677

8239

7678

%==================================

8240

\subsubsection{Particle relaxation time: \textmd{\texttt{uslatp}}}

7679

\subsubsection{Particle relaxation time}

8241

7680

%==================================

8242

7681

8243

7682

\noindent

8244

\textit{Subroutine called every Lagrangian sub-step.}

8245

8246

\noindent

8247

An intervention in this subroutine is not obligatory.

8248

8249

\noindent

8250

In this subroutine, the particle relaxation time may be modified

7683

An intervention in this subroutine is not mandatory.

7684

7685

\noindent

7686

The particle relaxation time may be modified in the subroutine \texttt{uslatp}

8251

7687

according to the chosen formulation of the drag coefficient. \\

8252

7688

The particle relaxation time, modified or not by the user, is available

8253

7689

in the array \texttt{taup}.

8254

7690

8255

7691

%==================================

8256

\subsubsection{Particle thermal characteristic time: \textmd{\texttt{uslatc}}}

7692

\subsubsection{Particle thermal characteristic time}

8257

7693

%==================================

8258

7694

8259

7695

\noindent

8260

\textit{Subroutine called every Lagrangian sub-step.}

8261

8262

\noindent

8263

An intervention in this subroutine is not obligatory.

8264

8265

\noindent

8266

In this subroutine, the particle thermal characteristic time may be

8267

modified according to the chosen correlation for the calculation of the

8268

Nusselt number. \\

7696

An intervention in this subroutine is not mandatory.

7697

7698

\noindent

7699

The particle thermal characteristic time may be

7700

modified in the subroutine \texttt{uslatc} according to the chosen correlation

7701

for the calculation of the

7702

Nusselt number. This subroutine is called ar each Lagrangian sub-step. \\

8269

7703

The thermal characteristic time, modified or not by the user, is

8270

7704

available in the zone \texttt{tempct(nbpmax,1)} of the array \texttt{tempct}.

7705

7706

%==================================

7707

\subsection{Compressible module}

7708

%==================================

7709

7710

When the compressible module\footnote{For more details concerning the

7711

compressible version, the user may refer to the document ``Implantation

7712

d'un algorithme compressible dans \CS'', Rapport EDF 2003,

7713

HI-83/03/016/A, P. Mathon, F. Archambeau et J.-M. H\'erard.} is

7714

activated, it is recommended to:

7715

\begin{list}{-}{}

7716

\item use the option ``time step variable in time and uniform in

7717

space'' (\texttt{idtvar}=1) with a maximum Courant number of 0.4

7718

(\texttt{coumax}=0.4): these choices must be written in \texttt{usini1}

7719

\item keep the convective numerical schemes proposed by default.

7720

\end{list}

7721

7722

%==================================

7723

\subsubsection{ Initialisation of the options of the variables}

7724

%==================================

7725

\label{prg_uscfx12}%

7726

\noindent

7727

\textit{Subroutines called at each time step.}

7728

7729

The subroutines \texttt{uscfx1} and \texttt{uscfx2} complete \texttt{usini1}.

7730

7731

\texttt{uscfx1} allows to set non standard calculation options related to the

7732

compressible module, and in particular to fill in the key word \texttt{icfgrp}

7733

allowing to take into account the hydrostatic equilibrium in the

7734

boundary conditions.

7735

7736

\texttt{uscfx2} allows to specify for the molecular thermal conductivity and

7737

the volumetric viscosity the following pieces of information:

7738

\begin{list}{-}{}

7739

\item variable or not (\texttt{iviscv})

7740

\item reference value (\texttt{viscv0})

7741

\end{list}

7742

7743

%==================================

7744

\subsubsection{Management of the boundary conditions}

7745

%==================================

7746

7747

\noindent

7748

\textit{Subroutine called every time step.}

7749

7750

The use of \texttt{uscfcl}

7751

is compulsory when running a calculation that uses the compressible module, just

7752

as it is in both \texttt{usini1} and \texttt{usppmo}. The

7753

way of using it is the same as the way of using

7754

\texttt{usclim} in the framework of standard calculations, that is to

7755

say several loops on the boundary faces lists (cf. \S\ref{fvm_selector})

7756

marked out by their colors, groups, or geometrical criterion, where

7757

the type of face, the type of boundary condition for each variable and

7758

eventually the value of each variable are defined.

7759

7760

{\em WARNING: in the case of a calculation using the compressible

7761

module, the boundary conditions of all the variables are defined here,

7762

even those of the eventual user scalars: {\em \texttt{usclim}} is not

7763

used at all.}

7764

7765

In the compressible module, the different available boundary conditions

7766

are the followings:

7767

7768

\begin{list}{-}{}

7769

\item inlet/outlet for which everything is known

7770

\item supersonic outlet

7771

\item subsonic inlet

7772

\item subsonic wall

7773

\item wall

7774

\item symmetry

7775

\end{list}

7776

7777

%==================================

7778

\subsubsection{Initialisation of the variables}

7779

%==================================

7780

7781

The subroutine \texttt{uscfxi}, called during the calculation initialisation,

7782

is used to initialise some variables specific to the

7783

specific physics activated {\em via} \texttt{usppmo}. As usual,

7784

the user may have access to several geometric variables to discriminate

7785

between different initialisation zones if needed.

7786

7787

{\em WARNING: in the case of a specific physics modeling, all the

7788

variables are initialised here: {\em \texttt{usiniv}} is not used at

7789

all.}

7790

7791

This subroutine works like \texttt{usiniv} for velocity,

7792

turbulence and passive scalars. Concerning pressure, density,

7793

temperature and specific total energy, only 2 variables out of the 4 are

7794

independant. The user may also initialise the variable pair he wants

7795

(apart from temperature-energy) and the two other variables will be

7796

calculated automatically by giving the right value to the variable

7797

\texttt{iccfth} used for the call to \texttt{uscfth}.

7798

7799

%==================================

7800

\subsubsection{Thermodynamics}

7801

%==================================

7802

7803

\noindent

7804

\textit{The subroutine \texttt{uscfth} is called several times at each time step

7805

(boundary conditions, physical properties, solving of the energy equation, ...).}

7806

7807

This subroutine is used to set the thermodynamics parameters. By

7808

default, the perfect gas laws are implemented. If the user needs to use

7809

other laws (perfect gas with variable Gamma, Van der Waals), he (or she) must

7810

modify this subroutine.

7811

7812

%==================================

7813

\subsubsection{Management of variable physical properties}

7814

%==================================

7815

7816

If necessary, all the variation laws of the fluid physical properties

7817

(viscosity, specific heat, ...) can be described in the subroutine \texttt{uscfpv}

7818

which is then called at each time step. This subroutine replaces and is similar to \texttt{usphyv}.

7819

7820

The user should make sure that the defined variation laws are valid for

7821

the whole variation range of the variables.

7822

7823

%==================================

7824

\subsection{Management of the electric arc module}

7825

%==================================

7826

%==================================

7827

\subsubsection{Initialisation of the variables}

7828

%==================================

7829

7830

\noindent

7831

\textit{subroutine called only at the initialisation of the calculation}

7832

7833

The subroutine \texttt{useliv} allows the user to initialise some of the specific

7834

physics variables

7835

prompted via \texttt{usppmo}. It is called only during the initialisation of

7836

the calculation. The user has access, as usual, to many geometric variables so

7837

that the zones can be treated separately if needed.

7838

7839

{\em WARNING: For the specific physics, it is here that all variables are initialised:

7840

\texttt{usiniv} is not used}

7841

7842

This subroutine works like \texttt{usiniv}. The values of potential and its

7843

constituents are initialised if required.

7844

7845

It should be noted that the enthalpy is relevant.

7846

7847

\begin{list}{-}{}

7848

\item For the electric arc module, the enthalpy value is taken from the temperature

7849

of reference \texttt{t0} (given in \texttt{usini1}) from the temperature-enthalpy

7850

tables

7851

supplied in the data file \texttt{dp\_ELE.} The user must not intervene here.

7852

\item For the Joule effect module, the value of enthalpy must be specified by the user

7853

. An example is given of how to obtain the enthalpy from the temperature of reference

7854

\texttt{t0}(given in \texttt{usini1}), the temperature-enthalpy law must be

7855

supplied. A code is suggested in the sub routine \texttt{usthht}(which is there for

7856

the determination of physical properties).

7857

\end{list}

7858

7859

%==================================

7860

\subsubsection{Variable physical properties}

7861

%==================================

7862

7863

All the laws of the variation of physical data of the fluid are written (when neccessary)

7864

in the subroutine \texttt{uselph}... The subroutine replaces \texttt{usphyyv} and works

7865

in a similar manner. It is called at each time step.

7866

7867

{\em WARNING: For the electric module, it is here that all the physical variables are defined

7868

(including the relative cells and the eventuel user scalars):}\texttt{usepelph} {\em {is not used.}}

7869

7870

The user should ensure that the defined variation laws are valid for the whole range of

7871

variables. Particular care should be taken with non-linear laws (for example, a

7872

$3^{rd}$ degree polynomial law giving negative values of density)

7873

7874

{\em WARNING: in the electric module, all the physical propertie are considered as variables

7875

and are therefore stored in the \texttt{propce} array. \texttt{cp0}, \texttt{viscls0}, \texttt{viscl0}

7876

are not used}

7877

7878

For the Joule effect, the user is required to supply the physical properties in the sub-

7879

routine. Examples are given which are to be adapted by the user. If the temperature is

7880

to be determined to calculate the physical properties, the solved variable, enthalpy must

7881

be deduced. The prefered temperature-enthalpy law can be selected in the subruotine

7882

\texttt{usthht} (an example of the interpolation is given from the law table. This

7883

subroutine can be re-used for the initialisation of the variables(\texttt{useliv}))

7884

For the electric arc module, the physical properties are interpolated from the data file

7885

\texttt{dp\_ELE} supplied by the user. Modifications are generally not necessary.

7886

7887

%==================================

7888

\subsubsection{Boundary Conditions}

7889

%==================================

7890

7891

\minititre{Subroutine \texttt{uselcl}}

7892

\noindent

7893

\textit {subroutine called at each time step.}

7894

7895

As much as \texttt{usini1} and \texttt{usppmo}, the use of \texttt{usecl}

7896

is required to run an electric calculation. The main use is the same as

7897

occurs in \texttt{usclim} for the standard \CS calculations, for which

7898

different loops on the boundary faces is defined. Each faces list is

7899

built with the use of selection criteria (cf. \S\ref{fvm_selector}),

7900

and is referenced by their group(s), their color(s) or geometrical

7901

criterions. The face type, the boundary conditions for each variable,

7902

and finally the value of each variable or imposed flow are fixed.

7903

7904

{\em WARNING:for the electric module, the boundary conditions of all

7905

the variables are defined here,

7906

even for those of the eventual user scalars: {\em \texttt{usclim}} is not

7907

used at all.}

7908

7909

For the electric module, each boundary face is associated with a number

7910

\texttt{izone} \footnote{\texttt{izone} must be less than the maximum

7911

value allowed by the code, \texttt{nozzppm}. This is fixed at 2000 in \texttt

7912

{ppvar.h} and cannot be modified.}(the color \texttt{icoul} for example) in

7913

order to group together all the boundary faces of the same type. In the report

7914

\texttt{usclim}, the main change from the users point of view concerns the

7915

specification of the boundary conditions of the potential, which isn't

7916

implied by default. The Dirichlet and Neumann conditions must be imposed

7917

explicitly using \texttt{icodcl} and \texttt{rcodcl} (as would be done for

7918

the classical scalar).

7919

7920

Whats more, if one wishes to slow down the power dissipation(Joule

7921

effect module) or the current (electric arc module) from the imposed values

7922

\texttt{(puismp\index{puismp}} and \texttt{couimp\index{couimp}} respectively),

7923

they can be changed by the potential scalar as shown below:

7924

7925

\begin{list}{-}{}

7926

\item For the electric arc, the imposed potential difference can be a fixed variable:

7927

for example, the cathode can be fixed at 0 and the potential at the anode

7928

contains the variable \texttt{dpot}. This variable is initialised in \texttt{useli1}

7929

by an estimated potential difference. If \texttt{ielcor}=1 (see

7930

\texttt{useli1}), \texttt{dpot} is updated automatically during the

7931

calculation to obtain the required current.

7932

\item For the Joule module effect, \texttt{dpot} is again used with the same

7933

signification as in the electric arc module. If \texttt{dpot} is not wanted

7934

in the setting of the boundary conditions, the variable \texttt{coejou} can be

7935

used. \texttt{coejou} is the coefficient by which the potential difference is

7936

multiplied to obtain the desired power dissipation. By default this begins at

7937

1 and is updated automatically. If \texttt{ielcor}=1 (see \texttt

7938

{useli1}), multiply the imposed potentials in \texttt{uselcl} by \texttt{coejou}

7939

at each time step to achieve the desired power dissipation.

7940

\end{list}

7941

7942

{\em WARNING: In alternative current, attention should be paid to the values of potential

7943

imposed at the limits: the variable named "real potential" represents an affective

7944

value if the current is in single phase, and a "real part" if not.}

7945

\begin{list}{-}{}

7946

\item For the Joule studies, a complex potential is sometimes needed

7947

(\texttt{ippmod(ieljou)}=2): this is the case in particular where the current

7948

has 3 phases. To have access to the phase of the potential, and not just to its

7949

amplitude, the 2 variables must be deleted: in \CS, there are 2 arrays

7950

specified for this role, the real part and the imaginary

7951

part of the potential. For use in the code, these variables are named

7952

``real potential'' and ``imaginary potential''. For an alternative

7953

sinusoidal potential $Pp$, the maximum value is noted as $Pp_\text{max}$,

7954

the phase is noted as $\phi$, the real potential

7955

and the imaginary potential are respecively $Pp_\text{max}\,cos\phi$ and

7956

$Pp_\text{max}\,sin\phi$.

7957

\item For the Joule studies in which one does not have access to the phases, the real

7958

potential (imaginary part =0) will suffice (\texttt{ippmod(ieljou)=1}): this is

7959

obviously the case with

7960

continous current, but also with single phase alternative current. In \CS

7961

there is only 1 varialbe for the potential, called "real potential". Pay attention to

7962

the fact that in alternate current, the "real potential" represents a effective value

7963

of potential , $\frac{1}{\sqrt{2}}\,Pp_\text{max}$ (in continous current there is no

7964

such ambiguity).

7965

\end{list}

7966

7967

\minititre{Subroutine \texttt{usetcl}}

7968

\noindent

7969

\textit{Subroutine called every time step.}

7970

7971

This subroutine is compulsory when the electrical module is used. It

7972

manages the boundary conditions for variables unknown by \texttt{usclim}.

7973

It calculates:

7974

\begin{list}{$\bullet$}{}

7975

\item the intensity at each electrode

7976

\item the voltage on each termin of transformers. To achieve it, the intensity,

7977

the rvoltage at each termin, the Rvoltage, and the total intensity of the

7978

transformer are calculated.

7979

\end{list}

7980

7981

Finally, a test is performed to check if the offset is zero or if a boundary

7982

face is in contact with the ground.

7983

7984

%====================================

7985

\subsubsection {Initialisation of the variable options}

7986

%==================================

7987

\label{prg_useli1}%

7988

7989

The subroutine \texttt{useli1} is completed in \texttt{usini1} for the specific

7990

physics. It is called at each time step. It allows:

7991

\begin{list}{$\bullet$}{}

7992

\item to activate the variables in the specific physics module, the chronological

7993

outputs (\texttt{ichrvr(ipp)} indicators), the listings (\texttt{ilisvr(ipp)}

7994

indicators) and the

7995

historical exits at the probes defined in \texttt{usini1} (\texttt{ihisvr(ipp)}

7996

indicators).

7997

The functions are the same as in \texttt{usini1} and the script frequency of the

7998

exits are fixed using \texttt{usini1.} The indicators \texttt{ipp} are for the

7999

value \texttt{ipp=ipppro} (\texttt{ipproc(ivar)}, with \texttt{ivar}, the number

8000

of specific physics variables. With the main variables

8001

which concern the user (velocity, pressure, etc), the user must always use

8002

\texttt{usini1} if the history, the listings, or the chronological files are required.

8003

The variables which the user can activate are marked out. The number of variables in

8004

the calculation is given in \texttt{ivar} (defined by

8005

\texttt{propce(iel,ipproc(iprop)} for cell \texttt{iel}):

8006

8007

\begin{list}{$\rightarrow$}{}

8008

\item Electric Arc Module:

8009

\begin{list}{-}{}

8010

\item Calculation variables \texttt{rtp(iel,ivar)}

8011

\begin{list}{\texttt{ivar} = }{}

8012

\item \texttt{isca(ihm\index{ihm})} enthalpy

8013

\item \texttt{isca(ipotr\index{ipotr})} real potentiel

8014

\item \texttt{isca(ipotva(i)\index{ipotva})} solved components of the potential vector.

8015

\item \texttt{isca(iycoel(iesp)\index{iycoel})} the mass fraction of \texttt{ngazg}

8016

composites if there are more than 1

8017

\end{list}

8018

\item Properties \texttt{propce(iel,ipproc(iprop))}

8019

\begin{list}{\texttt{iprop} = }{}

8020

\item \texttt{itemp\index{itemp}} temperature

8021

\item \texttt{iefjou\index{iefjou}} power dissipation by the Joule effect.

8022

\item \texttt{ilapla(i)\index{ilapla(i)}} components of the laplace forces.

8023

\end{list}

8024

\end{list}

8025

\item Joule Module effect~:

8026

\begin{list}{-}{}

8027

\item Calculation variables \texttt{rtp(iel,ivar)}

8028

\begin{list}{\texttt{ivar} = }{}

8029

\item \texttt{isca(ihm\index{ihm})} enthalpy

8030

\item \texttt{isca(ipotr\index{ipotr})} real potential

8031

\item \texttt{isca(ipoti\index{ipoti})} imaginary potential if its to be taken into account

8032

\item \texttt{isca(iycoel(iesp)\index{iycoel}}) the mass fraction of \texttt{ngazg}

8033

composites if there are more than 1

8034

\end{list}

8035

\item Properties \texttt{propce(iel,ipproc(iprop))}

8036

\begin{list}{\texttt{iprop} = }{}

8037

\item \texttt{itemp\index{itemp}} temperature

8038

\item \texttt{iefjou\index{iefjou}} volumic power dissipation by Joule effect.

8039

\end{list}

8040

\end{list}

8041

\end{list}

8042

8043

\item to give the coefficient of relaxation of the density \texttt{srrom}:\\

8044

$\rho^{n+1}=\texttt{srrom}*\rho^{n}+(1-\texttt{srrom})\rho^{n}$\\

8045

(for the electric arc, the sub-relaxation is taken into account during the 2nd time

8046

step; for the Joule effect the sub relaxation is not accounted for unless the user

8047

specifies in \texttt{uselph}

8048

8049

\item indicates if the data will be fixed in the power dissipation or

8050

in the current, done in \texttt{ielcor}.

8051

\item target current fixed as \texttt{couimp} (electric arc module)

8052

or the power dissipation \texttt{puism} (Joule module effect).

8053

\item Fix the initial value of potential difference \texttt{dpot},

8054

the for the calculations with a single fixed parameter as \texttt{couimp}

8055

or \texttt{puism}.

8056

8057

\end{list}

8058

8059

%==================================

8060

\subsubsection[{\em EnSight} output]

8061

{Post-processing output}

8062

%==================================

8063

8064

The subroutine \texttt{uselen} allows the addition on $n$ variables in the

8065

preprocessing output and

8066

works like the subroutine \texttt{usvpst} (with the electric module, it is however also

8067

possible to use \texttt{usvpst}. It is called at each chronological output

8068

8069

The algebraic variables related to the electric module are provided by default provided

8070

that they are not explicitely contained in the \texttt{propce} array:

8071

\begin{list}{-}{}

8072

\item gradient of real potential in $V m^{-1}$ ($\grad Pot_R = -\vect{E}$)

8073

\item density of real current in $A m^{-2}$ ($\vect{j}=\sigma \vect{E}$)

8074

\end{list}

8075

specifically for the Joule module effect with \texttt{ippmod(ieljou)}=2~:

8076

\begin{list}{-}{}

8077

\item gradient of imaginary potential in $V m^{-1}$

8078

\item density of real current in $A m^{-2}$

8079

\end{list}

8080

specifically for the electric arc module with \texttt{ippmod(ielarc)}=2~:

8081

\begin{list}{-}{}

8082

\item magnetic field in $T$ (\vect{B} = \vect{rot}\,\vect{A})

8083

\end{list}

8084

8085

If it is convenient for the user, there is no need to add this subroutine into the

8086

SRC directory: the post-processing will be done automatically (at the same frequency

8087

(\texttt{NTCHR}) as the other calculation variables)

8088

8089

8090

%==================================

8091

\subsection{\CS-\CS coupling}

8092

%==================================

8093

8094

\noindent

8095

\textit{Subroutine called once during the calculation initialisation.}

8096

8097

This user subroutine \texttt{ussatc} is used to couple \CS with itself.

8098

It is used for turbomachine applications for instance, the first \CS managing

8099

the fluid around the rotor and the other the fluid around the stator.

8100

In the case of a coupling between two \CS instances, the \texttt{numsat}

8101

and \texttt{namsat} arguments of the subroutine '\texttt{defsat}' are ignored.

8102

In case of multiple couplings, a coupling will be matched with available \CS

8103

instances prioritarily based on the \texttt{namsat} (\CS instance name) argument,

8104

then on the \texttt{numsat} (\CS instance application number) argument.\\

8105

If \texttt{namsat} is empty, matching will be based on \texttt{numsat} only.\\

8106

The arguments of '\texttt{defsat}' are:

8107

\begin{list}{-}{}

8108

\item \texttt{numsat}: the matching \CS application \texttt{id}, or $-1$,

8109

\item \texttt{namsat}: the matching \CS application name,

8110

\item \texttt{crtcsu}: the cell selection criteria for support,

8111

\item \texttt{crtfsu}: the boundary face selection criteria for support (not functional),

8112

\item \texttt{crtccp}: the cell selection criteria for coupled cells,

8113

\item \texttt{crtfcp}: the boundary face selection criteria for coupled faces,

8114

\item \texttt{iwarns}: the verbosity level.

8115

\end{list}

8116

8117

8118

%==================================

8119

\subsection{Fluid-Structure external coupling}

8120

%==================================

8121

8122

\noindent

8123

\textit{???Subroutine called only once or at each iteration???.}

8124

8125

The subroutine \texttt{usaste} belongs to the module dedicated to external

8126

Fluid-Structure coupling with \textit{Code\_Aster}. Here one defines the boundary

8127

faces coupled with \textit{Code\_Aster} and the fluid forces components which are

8128

given to structural calculation. When using external coupling with \textit{Code\_Aster},

8129

structure number necessarily needs to be negative; the references of coupled faces being

8130

i.e. -1, -2, etc...

8131

The subroutine performs the following operations:

8132

\begin{list}{-}{}

8133

\item '\texttt{getfbr}' is called to get a list of elements matching a

8134

geometrical criterion or reference number then a colour (negative value) is associated

8135

to these elements.

8136

\item the value passed to \texttt{asddlf}, for user-chosen component, for every negative

8137

colour, defines the movement imposed to the external structure.

8138

\item the user specify with the value of \texttt{isyncp} if \CS and \textit{Code\_Aster}

8139

use synchronised chronological output or not.

8140

\end{list}

8141

8142

%==================================

8143

\subsection{ALE module}

8144

%==================================

8145

%==================================

8146

\subsubsection{Initialisation of the options}

8147

%==================================

8148

\label{prg_usalin}%

8149

This initialisation can be performed in the Graphical User Interface (GUI)

8150

or in the subroutines \texttt{usalin} and \texttt{usstr1}. First of all,

8151

in the GUI when the ``Mobile mesh'' is selected in the ``Thermophysical models''

8152

heading, additional options are displayed. The user must choose a type of mesh

8153

viscosity and how to describe its spatial distribution, see fig. \ref{fig:Ini-ale}.

8154

8155

\begin{figure}[ht]

8156

\begin{center}

8157

\includegraphics[width=12cm]{gui_ale_mei}

8158

\caption{Thermophysical models - mobile mesh (ALE method)}

8159

\label{fig:Ini-ale}

8160

\end{center}

8161

\end{figure}

8162

8163

The following paragraphs are relevant if the GUI is not used.

8164

8165

\minititre{Subroutine \texttt{usalin}}

8166

\noindent

8167

\textit{Subroutine called at the start.}

8168

This subroutine completes \texttt{usini1}.

8169

8170

\texttt{usalin} allows to set option for the ale module, and in

8171

particular to active the ale module

8172

8173

\minititre{Subroutine \texttt{usstr1}}

8174

8175

\texttt{usstr1} allows to specify for the structure module the

8176

following pieces of information:

8177

\begin{list}{-}{}

8178

\item number of structure (\texttt{nbstru}).

8179

\item initial value of deplacement, velocity and acceleration

8180

(\texttt{xstr0}, \texttt{xstreq} and \texttt{vstr0}).

8181

\end{list}

8182

8183

Below is a list of the different variables that might be modified:

8184

8185

\begin{list}{$\bullet$}{}

8186

8187

\item{\texttt{nbstru}} \\

8188

{the number of structures}

8189

8190

\item{\texttt{idfstr(i)}} \\

8191

{index of the structure, where I is the index of the face}

8192

8193

\item{\texttt{xstr0(i,k)}} \\

8194

{initial position of a structure, where \texttt{i} is the dimension of space

8195

and \texttt{k} the index of the structure}

8196

8197

\item{\texttt{xstreq(i,k)}} \\

8198

{position of balance of a structure, where \texttt{i} is the dimension of space

8199

and \texttt{k} the index of the structure}

8200

8201

\item{\texttt{vstr0(i,k)}} \\

8202

{initial velicity of a structure, where \texttt{i} is the dimension of space

8203

and \texttt{k} the index of the structure }

8204

\end{list}

8205

8206

%==================================

8207

\subsubsection{Boundary conditions of velocity mesh}

8208

%==================================

8209

The boundary conditions can be managed with the Graphical User Interface (GUI)

8210

or with the subroutine \texttt{usalcl} (called at each time step). In the GUI,

8211

when the item ``Mobile mesh'' is activated the item ``Fluid structure interaction''

8212

appears under the heading ``Boundary conditions''. Two types of Fluid-structure

8213

coupling are offered. The first one is internal, using a simplified structure

8214

model and the second is external with \textit{Code\_Aster}, see figs.

8215

\ref{fig:CL-ale1} and \ref{fig:CL-ale2}.

8216

8217

\begin{figure}[ht]

8218

\begin{center}

8219

\includegraphics[width=12cm]{gui_ale_internal}

8220

\caption{Boundary conditions - internal coupling}

8221

\label{fig:CL-ale1}

8222

\end{center}

8223

\end{figure}

8224

8225

\begin{figure}[ht]

8226

\begin{center}

8227

\includegraphics[width=12cm]{gui_ale_external}

8228

\caption{Boundary conditions - external coupling}

8229

\label{fig:CL-ale2}

8230

\end{center}

8231

\end{figure}

8232

8233

\minititre{Subroutine \texttt{usalcl}}

8234

When the GUI is not used, the use of \texttt{usalcl} is mandatory to run

8235

a calculation using

8236

the ale module just as it is in \texttt{usini1}. The way of using it

8237

is the same as the way of using \texttt{usclim} in the framework of

8238

standard calculations, that is to say a loop on the boundary faces

8239

marked out by their colour (or more generally by a property of their

8240

family), where the type of boundary condition of velocity mesh for

8241

each variable are defined.

8242

8243

The main numerical variables are described below.

8244

8245

\variab{ialtyb}{ialtyb(nfabor)}{ia}{In the ale module, the user

8246

defines the velocity mesh from the colour of the boundary faces, or

8247

more generally from their properties (colours, groups, ...), from the

8248

boundary conditions defined in \texttt{usclim}, or even from their

8249

coordinates. To do so, the array \texttt{ialtyb(nfabor)} gives for each face

8250

\texttt{ifac} the velocity mesh boundary condition types marked out by the key

8251

words \texttt{ivimpo\index{ivimpo}}, \texttt{igliss\index{igliss}},

8252

\texttt{ibfixe\index{ibfixe}}

8253

8254

\begin{list}{$\bullet$}{}

8255

8256

\item If \texttt{ialtyb=ivimpo}: imposed velocity.

8257

8258

\begin{list}{$\rightarrow$}{}

8259

\item In the case where all the nodes of a face have a imposed displacement,

8260

it is not necessary to fill the tables with boundary conditions

8261

velocity mesh for this face, they will be erased. In the other case,

8262

the value of the Dirichlet must be given in \texttt{rcodcl(ifac,ivar,1)} for

8263

every value of \texttt{ivar} (\texttt{iuma}, \texttt{ivma} and \texttt{iwma})

8264

The other boxes of \texttt{rcodcl} and \texttt{icodcl} are completed automatically.

8265

8266

The tangential velocity mesh is taken like a tape speed under the

8267

boundary conditions of wall for the fluid, except if wall velocity

8268

was specified by the user in the interface or usclim (in which case

8269

it is this speed which is considered).

8270

\end{list}

8271

8272

\item if \texttt{ialtyb(nfac) = ibfixe}: fixed wall

8273

\begin{list}{$\rightarrow$}{}

8274

\item the velocity is null.

8275

\end{list}

8276

8277

\item if \texttt{ialtyb(nfac) = igliss}: sliding wall

8278

\begin{list}{$\rightarrow$}{}

8279

\item the tangential velocity is not used.

8280

\end{list}

8281

8282

\end{list}

8283

}

8284

8285

%==================================

8286

\subsubsection{Modification of the viscosity}

8287

%==================================

8288

8289

The user subroutine \texttt{usvima} is used along the ALE (Arbitrary Lagrangian Eulerian

8290

Method) module, it fills mesh viscosity arrays. It is called at each time step.

8291

The user can modify mesh viscosity values to prevent cells and nodes from huge

8292

displacements in awkward areas, such as boundary layer for example.

8293

If \texttt{iortvm} = 0, the mesh viscosity modelling is considered as isotropic and

8294

therefore only the \texttt{viscmx} array needs to be filled.

8295

If \texttt{iortvm} = 1, mesh viscosity modeling is orthotropic therefore all arrays

8296

\texttt{viscmy}, \texttt{viscmx}, and \texttt{viscmz} need to be filled.

8297

Note that \texttt{viscmx}, \texttt{viscmy} and \texttt{viscmz} arrays are initialized

8298

at the first time step with the value 1.

8299

8300

%==================================

8301

\subsubsection{Fluid - Structure internal coupling}\label{sec:ALE}

8302

%==================================

8303

8304

In the subroutine \texttt{usstru} the user provides the parameters of two other subroutines.

8305

\texttt{usstr1} is called at the beginning of the calculation. It is used to define

8306

and initialise the internal structures where Fluid-Structure coupling occurs.

8307

For each boundary face \texttt{ifac}, \texttt{idfstr(ifac)} is the number of the structure

8308

the face belongs to (if \texttt{idfstr(ifac)} = 0, the face \texttt{ifac} doesn't belong

8309

to any structure). When using internal coupling, structure number necessarily needs to be

8310

positive. The number of "internal" structures is automatically defined with the maximum

8311

value of the \texttt{idfstr} table, meaning that internal structure numbers must be defined

8312

sequentially with positive values, beginning with integer value '1'.

8313

8314

For each internal structure one can define here:

8315

\begin{list}{-}{}

8316

\item an initial velocity \texttt{vstr0}

8317

\item an initial displacement \texttt{xstr0} (i.e. \texttt{xstr0} is the value of the

8318

displacement \texttt{xstr} compared to the initial mesh at time t = 0)

8319

\item a displacement compared to equilibrium \texttt{xstreq} (i.e. \texttt{xstreq}

8320

is the initial displacement of the internal structure compared to its position at

8321

equilibrium; at each time step t and for a displacement \text{xstr}(t), the associated

8322

internal structure will undergo a force $-k*(\text{}(t)+XSTREQ)$ due to the spring).

8323

\end{list}

8324

\text{xstr0} and \text{vstr0} are initialised with the value 0.

8325

When starting a calculation using ALE, or re-starting a calculation with ALE, based

8326

on a first calculation without ALE, an initial iteration 0 is automatically performed

8327

in order to take initial arrays \text{xstr0}, \text{vstr0} and \text{xstreq} into

8328

account. In any other case, add the following expression '\text{italin=1}' in subroutine

8329

\text{usalin}, so that the code can deal with the arrays \text{xstr0}, \text{vstr0} and \text{xstreq}.

8330

8331

When \texttt{ihistr} is set to 1, the code writes in the output the history of the

8332

displacement, of the structural velocity, of the structural acceleration anf of the

8333

fluid force. The value of structural history output step is the same as the one for

8334

standard variables \text{nthist}.

8335

8336

The second subroutine, \text{usstr2}, is called at each iteration. One defines in this

8337

subroutine structural parameters (considered as potentially time dependent): i.e.,

8338

mass m \text{xmstru}, friction coefficients c \text{xcstru}, and stiffness k \text{xkstru}.

8339

\text{forstr} array gives fluid stresses acting on each internal structure. Moreover it's

8340

possible to take external forces (gravity for example ) into account, too.

8341

\begin{list}{.}{}

8342

\item \text{xstr} array indicates the displacement of the structure compared to it

8343

s position in initial mesh,

8344

\item \text{xstr0} array gives the displacement of the structures in initial mesh

8345

compared to structural equilibrium,

8346

\item \text{vstr} array stands for structural velocity.

8347

\end{list}

8348

\text{xstr}, \text{xstr0} and \text{vstr} are \text{DATA} tables that can be used to

8349

define arrays Mass, Friction and Stiffness. Those are not to be modidfied.

8350

8351

The 3D structural equation that is solved is the following one :

8352

\begin{equation}\label{eq:FluidStruct}

8353

\displaystyle

8354

\tens{m}.\partial_{tt}\vect{x}+\tens{c}.\partial_{t}\vect{x}+\tens{k}.\left(\vect{x}+\vect{x_0}\right)=\vect{f},

8355

\end{equation}

8356

where $x$ stands for the structural displacement compared to initial mesh postition

8357

\text{xstr}, $x_0$ represents

8358

the displacement of the structure in initial mesh compared to equilibrium.

8359

Note that $\tens{m}$,$\tens{c}$, and $\tens{k}$ are 3\text{x}3 matrices.

8360

Equation \eqref{eq:FluidStruct} is solved using a Newmark HHT algorithm.

8361

Note that the time step used to solve this equation, \text{dtstr}, can be

8362

different from the one of fluid calculations. The user is free to define \text{dtstr}

8363

array. At the beginning of the calculation \text{dtstr} is initialised to the value of

8364

\text{dtcel} (fluid time step).

8365

8366

%==================================

8367

\subsection{Management of the structure property}

8368

%==================================

8369

8370

The use of \texttt{usstr2} is mandatory to run a calculation using the ale

8371

module with a structure module. It is called at each time step.

8372

8373

For each structure, the system that will be solved is:

8374

8375

\begin{equation}

8376

M.x^{''}+C.x^{''}+K.(x-x_{0} = 0

8377

\end{equation}

8378

8379

where

8380

8381

\begin{list}{-}{}

8382

\item $M$ is the mass stucture (\texttt{xmstru}).

8383

\item $C$ is the dumping coefficient of the stucture (\texttt{xcstru}).

8384

\item $K$ is the spring constant or force constant of the stucture (\texttt{xkstru}).

8385

\item $x_{0}$ is the initial position

8386

\end{list}

8387

8388

Below is a list of the different variables that might be modified:

8389

8390

\begin{list}{$\bullet$}{}

8391

8392

\item{\texttt{xmstru(i,j,k})} \\

8393

{the mass stucture of the structure, where \texttt{i},\texttt{j} is

8394

the array of mass structure and \texttt{k} the index of the structure.}

8395

8396

\item{\texttt{xcstru(i,j,k})}\\

8397

{dumping coefficient of the stucture, where \texttt{i},\texttt{j} is the array of

8398

dumping coefficient and \texttt{k} the index of the structure.}

8399

8400

\item{\texttt{xkstru(i,j,k)}}\\

8401

{spring constant of the stucture, where \texttt{i},\texttt{j} is the array of spring

8402

constant and \texttt{k} the index of the structure.}

8403

8404

\item{\texttt{forstr(i,k)}}\\

8405

{force vector of the stucture, where \texttt{i} is the force vector and

8406

\texttt{k} the index of the structure.}

8407

\end{list}

8408

8409

%==================================

8410

\subsection{Management of the Atmospheric module}

8411

%==================================

8412

%==================================

8413

\subsubsection{Initialisation of the variables}

8414

%==================================

8415

8416

The initialisation can be done in the Graphical User Interface (GUI)

8417

or in the subroutine \texttt{usativ} (called only during the calculation

8418

initialisation). Under the heading ``Thermophysical models'', when in the

8419

item ``Calculation features'' one of the atmospheric flow model is selected,

8420

it activates an item under the same heading:``Atmospheric flows'' where the

8421

path leading to a file containing meteorological data must be specified, see

8422

fig. \ref {fig:Ini-atmo}. In addition is the atmospheric flow model chosen is

8423

the ``dry atmosphere'', an option appear the item ``Time step'' under the

8424

heading ``Numerical parameters'' plus an additional variable ``PotTemp''

8425

in the table of the ``Equation parameters'' item.

8426

8427

\begin{figure}[ht]

8428

\begin{center}

8429

\includegraphics[width=12cm]{gui_atmo_read}

8430

\caption{Thermophysical models - atmospheric flows}

8431

\label{fig:Ini-atmo}

8432

\end{center}

8433

\end{figure}

8434

8435

When the GUI is not used, \texttt{usativ} allows to initialise or modify

8436

(in case of a restarted calculation) the calculation variables and the

8437

values of the time step. It plays a similar role as \texttt{usiniv} for

8438

the additional variables introduced with the air-cooling module. The

8439

quantities that can be initialised here in user-selected zones are:

8440

\begin{list}{-}{}

8441

\item the air velocity with the array \texttt{rtp(iel,iu)} (with

8442

\texttt{iv} and \texttt{iw} for the other components),

8443

\item the air temperature with the array \texttt{rtp(iel,isca(ihumid))},

8444

\item turbulent quantities depending on the turbulent model selected.

8445

\end{list}

8446

The example provided in the user file performs the initialisation of the

8447

variables from meteorological profiles using the interpolation routine \texttt{intprf}.

8448

8449

%==================================

8450

\subsubsection{Non standard options}

8451

%==================================

8452

The subroutine \texttt{usati1} initialises non-standard parameters for

8453

atmospherical calculations. These parameters are for instance:

8454

\begin{list}{-}{}

8455

\item \texttt{imeteo},

8456

\item \texttt{irovar} for each phase,

8457

\item \texttt{ivivar} for each phase.

8458

\end{list}

8459

8460

%==================================

8461

\subsubsection{Management of the boundary conditions}

8462

%==================================

8463

8464

The user subroutine \texttt{usatcl} allows to define the boundary conditions of the variables

8465

unknown by \texttt{usclim}. It is called at each time step. Boundary conditions are

8466

applied to mesh faces selected using the subroutine '\texttt{getfbr}' for instance. For

8467

each type of boundary condition, these faces are grouped as physical zones characterised

8468

by an arbitrary number \texttt{izone} chosen by the user. If a boundary condition is

8469

retrieved from a meteorological profile, the variable \texttt{iprofm(izone)} of the zone

8470

must be set to 1.\\

8471

Examples are provided in \texttt{usatcl}.

8472

8473

%==================================

8474

\subsection{Cooling tower modelling}

8475

%==================================

8476

8477

%==================================

8478

\subsubsection{Parameters}

8479

%==================================

8480

8481

\noindent

8482

\textit{Subroutine called only during calculation initialisation? OR AT EACH ITERATION?.}

8483

8484

The subroutine \texttt{uscti1} contains calculation parameters such as:

8485

\begin{list}{-}{}

8486

\item temperature parameters,

8487

\item the number of exchange zones at various locations,

8488

\item the air properties.

8489

\end{list}

8490

8491

%==================================

8492

\subsubsection{Initialisation of the variables}

8493

%==================================

8494

8495

The subroutine \texttt{usctiv} allows to initialise or modify

8496

(in case of a restarted calculation) the calculation variables and the

8497

values of the time step. It is called only during the calculation

8498

initialisation. It plays a similar role as \texttt{usiniv} for the

8499

additional variables introduced with the air-cooling module. The

8500

quantities that can be initialised here in user-selected zones are:

8501

\begin{list}{-}{}

8502

\item the air temperature by filling the array \texttt{rtp(iel,isca(ihumid))},

8503

\item the air humidity by filling the array \texttt{rtp(iel,isca(itemp4))},

8504

\item the air velocity by filling the array \texttt{rtp(iel,iu)} (with

8505

\texttt{iv} and \texttt{iw} for the other components),

8506

\end{list}

8507

where \texttt{iel} can be an element found in a list returned by the routine

8508

'\texttt{getcel}'.

8509

8510

%==================================

8511

\subsubsection{Definition of the exchange zones}

8512

%==================================

8513

8514

The subroutine \texttt{usctdz} is used to define the exchange zones of a cooling

8515

tower. The user provides the following parameters:

8516

\begin{list}{-}{}

8517

\item \texttt{imzech}: its value is related to the model used:

8518

\begin{list}{$\rightarrow$}{}

8519

\item 0: no model is used,

8520

\item 1: Merkel model is used,

8521

\item 2: Poppe model is used,

8522

\end{list}

8523

\item 10 exchange zone parameters.

8524

\end{list}

8525

These arguments are passed to the subroutine '\texttt{defct}' along with a

8526

geometrical selection criterion.

8527

8528

%==================================

8529

\subsubsection{Management of the boundary conditions}

8530

%==================================

8531

8532

The subroutine \texttt{usctcl}, called at each time step, allows to define

8533

the boundary conditions of the variables unknown by \texttt{usclim}. Boundary

8534

conditions are applied to mesh faces selected using the subroutine \texttt{getfbr}

8535

for instance. For each type of boundary condition, these faces are grouped as

8536

physical zones characterised by an arbitrary number \texttt{izone} chosen by the

8537

user. The list of boundary conditions offered in this module is given below:

8538

\begin{list}{-}{}

8539

\item Dirichlet,

8540

\item flux density (velocities, pressure, scalar),

8541

\item sliding wall (velocity),

8542

\item friction (velocity),

8543

\item roughness (velocity),

8544

\item free inlet/outlet (velocity),

8545

\item symmetry.

8546

\end{list}

Older »