1
%-------------------------------------------------------------------------------
3
% This file is part of Code_Saturne, a general-purpose CFD tool.
5
% Copyright (C) 1998-2011 EDF S.A.
7
% This program is free software; you can redistribute it and/or modify it under
8
% the terms of the GNU General Public License as published by the Free Software
9
% Foundation; either version 2 of the License, or (at your option) any later
12
% This program is distributed in the hope that it will be useful, but WITHOUT
13
% ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
14
% FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
17
% You should have received a copy of the GNU General Public License along with
18
% this program; if not, write to the Free Software Foundation, Inc., 51 Franklin
19
% Street, Fifth Floor, Boston, MA 02110-1301, USA.
21
%-------------------------------------------------------------------------------
23
%===============================================
1
24
\programme{ Introduction}
2
{\huge sub-routines: co**, cp**, fu** ...}
25
%===============================================
4
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
5
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
27
%===============================================
6
28
\section{Use \& call}
7
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
8
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
10
From a CFD point of view combustion is a ({\small sometimes very})
11
complicated way to determine $\rho$.\\ Models needs few extra scalar
12
fields with regular transport equations, some of them with a
13
reactive or interfacial source term.\\ Modelling of combustion is able
14
to deal with gas phase combustion ({\small diffusion, premix, partial
15
premix}), and with solid or liquid fuels.\\ Combustion of condensed
16
fuels involves one-way interfacial flux due to phenomena in the
17
condensed phase ({\small evaporation or pyrolisis}) and reciprocal
18
ones ({\small heterogeneous combustion}). Many of the species injected
19
in the gas phase are afterwards involved in gas phase combsution.\\
20
That is the reason why many modules are similar for gas, coal and fuel
21
combustion modelling. Obviously, the thermodynamical description of
22
gas species is similar in every version as close as possible to the
23
JANAF rules.\\ All models are developped in both adiabatic and
24
unadiabatic ({\small permeatic: heat loss, eg. by radiation})
25
version, beyond the standard, the rule to call models is:
27
IPPMOD(index model) = -1 unused
29
IPPMOD(index model) = 0 simplest adiabatic version
31
IPPMOD(index model) = 1 simplest permeatic version
35
IPPMOD(index model) = 2.p Pļæ½ adiabatic version
37
IPPMOD(index model) = 2.p+1 Pļæ½ permeatic version
40
Every permeatic version involves the transport of enthalpy ({\small one more variable}).
42
%=================================
43
\subsection{Gas combustion modelling}
44
%=================================
46
Gas combustion is limited by disponibility ({\small in the same fluid
47
particle}) of both fuel and oxidant and by kinetic effects ({\small a
48
lot of chemical reactions involved can be described using an Arrhenius
49
law with high activation energy}). The mixing of mass ({\small atoms})
50
incoming with fuel and oxydant is described by a mixture fraction
51
({\small mass fraction of matter incoming with fuel}), this variable
52
is not affected by combustion. A progress variable is used to describe
53
the transformation of the mixture from fuel and oxydant to products
54
({\small carbon dioxyde and so on}).Combustion of gas is,
55
traditionnaly, splitted in premix and diffusion regimes.\\
57
In premix combustion process a first stage of mixing have been
58
realised ({\small without blast ...}) and the mixture is introduced in
59
the boiler ({\small or combustor can}). In common industrial
60
conditions the combustion is mainly limited by the mixing of fresh
61
gases ({\small inert}) and burnt ones resulting in the inflammation of
62
the first and their conversion to burnt ones; so an assumption of
63
chemistry much faster than mixing induces an intermittent regime. The
64
gas flow is constituted of totally fresh and totally burnt gases
65
({\small the flamme containing the gases during their transformation
66
is "extremely" thin}). With this previous assumptions,
67
Spalding \cite{1} established the "Eddy Break Up" model, which allows
68
a complete description with only one progress variable ({\small
69
mixture fraction is homogeneous}).\\
71
In diffusion flames the fuel and the oxydant are introduced by two
72
({\small at least}) inlets, in common industrial conditions, their
73
mixing is the main limitation and the mixture fraction is enough to
74
qualify a fluid particle, but in turbulent flows a {\em P}robability
75
{\em D}ensity {\em F}unction of the mixture fraction is needed to
76
qualify the thermodynamical state of the bulk. So both the mean and
77
the variance of the mixture fraction are needed ({\small two
80
Real world's chemistry is not so fast and, unfortunately, the mixing
81
can not be so homogeneous as wished. Then industrial combustion occurs
82
in partial premix regime. Partial premix occurs if mixing is not
83
finished ({\small at molecular level}) when the mixture is introduced,
84
or if air or fuel, are staggered, or if a diffusion flame is blown
85
off. For these situations, and specifically for lean premix gas
86
turbines Libby \& Williams \cite{2} developped a model allowing a
87
description of both mixing and chemical limitations. A collaboration
88
between the LCD Poitiers \cite{3} and EDF R\&D allows a simpler
89
version of their algorithm. Not only the mean and the variance of both
90
mixture fraction and progress variable are needed but also their
91
covariance ({\small five variables}).
94
%=================================
95
\subsection{Coal combustion modelling}
96
%=================================
98
Coal combustion is the main way to produce electricity in the world.
99
Coal is a natural product with a very complex composition. During the
100
industrial process of milling the raw coal is broken in tiny particles
101
of different sizes. After its introduction in the boiler, coal
102
particles undergoes drying, devolatilisation ({\small heating of coal
103
turn it in a mixture of char and gases}), heterogenous combustion
104
({\small of char in carbon monoxide}) leaving ash particles.\\ Each of
105
these phenomena are taken into account for some class of particles: a
106
class is caracterised by a coal ({\small it is useful to burn mixture
107
of coals with differents ranks or mixture of coal with biomasse ...})
108
and an initial diameter. For each class, \CS computes the number and
109
the mass of particles by unit mass of mixture.\\ The main assumption
110
is to solve only one velocity ({\small and pressure}) field: it means
111
the discrepancy of velocity between coal particles and gases is
112
assumed negligible.\\ Due to the radiation and heterogeneous
113
combustion, temperature can be different for gas and different size
114
particles: so the specific enthalpy of each particle class is
115
solved.\\ The description of coal pyrolysis proposed by Kobayashi \&
116
Ubhayakar \cite{4} is used, leading to two source terms for light and
117
heavy volatile matters ({\small the moderate temperature reaction
118
produces gases with low molecular mass, the high temperature reaction
119
produces heavier gases and less char}) represented by two passive
120
scalars: mixture fractions. The description of the heterogeneous
121
reaction ({\small which produce carbon monoxide}) produces a source
122
term for the carbon: a mixture fraction who can't be greater than the
123
results of stoechiometric oxidation of char by air ({\small carbon
124
can't be free in gas phase, it is always linked in an oxide}).\\ The
125
retained model for the gas phase combustion is the assumption of
126
diffusion flammelets surrounding each particle, so the previous
127
gaseous fuels are mixed in a local mean fuel and the mixing with air
128
is represented by a pdf between air and the mean local fuel
129
constructed with the variance of a passive scalar linked with air
130
({\small interfacial mass flux produce a source term for this
29
%===============================================
32
From a CFD point of view, combustion is a (sometimes very) complicated way to
33
determine $\rho$, the density.\\
34
Depending on the presence of a match or not, two solutions exist, known as
35
ignited and extinguished. From a numerical point of view, it means that these
36
two solutions have two attraction basin; the more representative the model, the
37
more difficult the stabilisation of the combustion ( may be difficult to
40
However, combustion models needs few extra fields of scalar with regular
41
transport equations, some of them with a reactive or interfacial source term. \\
43
This 2011 version of \CS focuses on stationnary industrial combustion processes
45
propagating fires are out of the present range (but in the short coming release). \\
47
In \CS~ modelling of combustion is able to deal with gas phase combustion
48
(diffusion, premix, partial premix), and with solid or liquid finely dispersed
49
fuels (fixed and fluidised beds are out of range).\\
50
Combustion of condensed fuels involves one-way interfacial flux due to phenomena
51
in the condensed phase (evaporation or pyrolisis) and reciprocal ones
52
(heterogeneous combustion and gasification). Many of the species injected in the
53
gas phase are afterwards involved in
54
gas phase combustion.\\
55
That is the reason why many modules are similar for gas, coal and fuel oil
56
combustion modelling. Obviously, the thermodynamical description of gas
57
species is similar in every version as close as possible to the JANAF rules.\\
58
All models are developped in both adiabatic and unadiabatic (permeatic : heat
59
loss, e.g. by radiation) version, beyond the standard (\fort{-1, 0, 1}), the
60
rule to call models in \fort{usppmo} is:
63
\fort{ippmod(index ; model)} &=& -1 ~~\quad \text{unused} \nonumber \\
64
\fort{ippmod(index ; model)} &=& ~~ 0 ~~~\quad \text{simplest adiabatic version} \nonumber \\
65
\fort{ippmod(index ; model)} &=& ~~ 1 ~~~\quad \text{simplest permeatic version} \label{Eqs_00A}\\
66
\text{and possibly:}\qquad\qquad\qquad\qquad\qquad\qquad\qquad \nonumber & &\\
67
\fort{ippmod(index ; model)} &=& 2.p \qquad \text{p} ~\text{adiabatic version} \nonumber \\
68
\fort{ippmod(index ; model)} &=& 2.p+1 ~ \text{p} ~\text{permeatic version} \nonumber
72
Every permeatic version involves the transport of enthalpy (one more variable).
74
%===============================================
75
\section{Gas combustion modelling}
76
%===============================================
78
Gas combustion is limited by disponibility (in the same fluid particle) of both
79
fuel and oxidizer and by kinetic effects (a lot of chemical reactions involved
80
can be described using an Arrhenius law with high activation energy). The mixing
81
of mass (atoms) incoming with fuel and oxydizer is described by a mixture
82
fraction (mass fraction of matter incoming with fuel), this
83
variable is not affected by combustion.\\
84
A progress variable is used to describe the transformation of the mixture from
85
fuel and oxydant to products (carbon dioxyde and so on).
86
Combustion of gas is, traditionnaly, splitted in premix and diffusion regimes.\\
87
In premixed combustion process a first stage of mixing has been realised
88
(without blast ...) and the mixture is introduced in the boiler (or combustor
89
can). In common industrial conditions the combustion is mainly limited by the
90
mixing of fresh gases (frozen) and burnt gases (exhausted) resulting in the
91
inflammation of the first and their conversion to burnt ones ; so an assumption
92
of chemistry much faster than mixing (characteristic time for chemistry much
93
smaller than characteristic time for turbulent mixing) induces an intermittent
94
regime. The gas flow is constituted of totally fresh and totally burnt gases
95
(the flamme containing the gases during their transformation is extremely
96
thin). With this previous assumptions, Spalding \cite{1} established the "Eddy
97
Break Up" model, which allows a complete description of the combustion process
98
with only one progress variable (mixture fraction is both constant
99
- in time - and homogeneous - in space).\\
100
In diffusion flames the fuel and the oxydant are introduced by, at least, two
101
inlets. In ordinary industrial conditions, their mixing is the main limitation
102
and the mixture fraction is enough to qualify a fluid particle, but in turbulent
103
flows a {\em P}robability {\em D}ensity {\em F}unction of the mixture fraction
104
is needed to qualify the thermodynamical state of the bulk. So, at least, both
105
the mean and the variance of the mixture fraction are needed (two variables) to
106
fit parameters of the pdf (the shape of whose is presumed).\\
107
Real world's chemistry is not so fast and, unfortunately, the mixing can not be
108
as homogeneous as wished. The main part of industrial combustion occurs in
109
partial premix regime. Partial premix occurs if mixing is not finished ( at
110
molecular level) when the mixture is introduced, or if air or fuel, are
111
staggered, or if a diffusion flame is blown off. For these situations, and
112
specifically for lean premix gas turbines \cite{2} developped a model allowing a
113
description of both mixing and chemical limitations. A collaboration between the
114
LCD Poitiers \cite{3} and EDF R\&D has produced a simpler version of their
115
algorithm. Not only the mean and the variance of both mixture fraction and
116
progress variable are needed but also their covariance (five variables).
136
%=================================
137
\subsection{Heavy Fuel Oil combustion modelling}
138
%=================================
140
Heavy fuel oil combustion have been hugely used to produced electrical
141
energy. Environmental regulation turning it more difficult and less
142
acceptable, a focus is needed on pollutant emission mainly sulphur
143
oxide and particles ({\small condensation of sulphuric acid can
144
aggregate soot}).\\ The description of fuel evaporation is done with
145
respect to its heaviness: after a minimum temperature is reached, the
146
gain of enthalpy is splitted between heating and evaporation. This way
147
the evaporation takes place on a range of temperature ({\small which
148
can be large}). The "total" evaporation is common for light ({\small
149
domestic}) oil but impossible for heavy ones: at high temperature,
150
during the last evaporation, a crakink reaction appears: so a
151
particle similar to char leaves. The heterogeneous oxydation of this
152
char particle is very similar to coal char ones.\\ Fuel injection is
153
described ({\small 2006 version}) with only one class of particles
154
({\small i.e. initial diameter}), the number, mass and specific
155
enthalpy of particles are calculated eveywhere. So three variables are
156
used to describe the condensed phase. In the same way as for coal,
157
only one velocity field is computed.\\ The model for gas combustion is
158
very similar to coal one but a special attention is paid to sulphur
159
({\small assumed to leave the particle as H2S during evaporation and
160
to be converted to SO2 during gas combustion}).
163
%==================================
164
%==================================
119
%===============================================
120
\section{Two-phase combustion modelling}
121
%===============================================
124
Coal combustion is the main way to produce electricity in the world. Heavy fuel
125
oil combustion have been hugely used to produce electrical energy.
126
Biomass is a promising fuel to be used alone or in blend.\\
127
Advanced combustion process may include exhaust gases recycling, pure oxygen or
128
steam injection, so the 2011 release of \CS ~takes in account
129
three oxidizers (tracked by three mixture fractions).\\
130
Coal is a natural product with a very complex composition. During the industrial
131
process of milling, the raw coal is broken in tiny particles of different
132
sizes. After its introduction in the boiler, coal particles undergoes drying,
133
devolatilisation (heating of coal turn it in a mixture of char and gases),
134
heterogenous combustion (of char by oxygen in carbon monoxide), gasification (of
136
dioxide or by water steam in carbon monoxide), leaving ash particles.\\
137
The description of fuel evaporation is done with respect to its heaviness :
138
after a minimum temperature is reached, the gain of enthalpy is splitted between
139
heating and evaporation. This way the evaporation takes place on a range of
140
temperature (which can be large). The total evaporation is common for light
141
(domestic) oil but impossible for heavy ones : at high temperature, during the
142
last evaporation, a craking reaction appears : so a particle similar to the char
143
is leaved. The heterogeneous oxydation of this char particle is very similar to
145
Each of these phenomena are taken into account for some classes of particles : a
146
solid class is caracterised by a coal (it is useful to burn mixture of coals
147
with differents ranks or mixture of coal with biomass ...) and an initial
148
diameter, for heavy fuel oil, liquid classes refer to initial diameter (neither
149
possibility of blending after injection nor cofiring with oil and coal). \CS~
150
computes the number, the mass and the enthalpy for each class of particles by
151
unit of mass of mixture; allowing the determination of local diameter and
152
temperature (for each class;
153
e.g. the finest will be be heated the fastest).\\
154
The main assumption is to solve only one velocity (and pressure) field : it
155
means that the discrepancy of velocity between coal particles and
156
gases is assumed to be negligible.\\
157
Due to the radiation, evaporation and heterogeneous combustion, temperature can
158
be different for gas and different size particles : so the specific enthalpy of
159
each particle class is solved.\\
160
The description of coal pyrolysis proposed by \cite{4} is used, leading to two
161
source terms for light and heavy volatile matters (the moderate temperature
162
reaction produces gases with low molecular mass, the high temperature reaction
163
produces heavier gases and less char) represented by two passive scalars :
164
mixture fractions. The description of the heterogeneous reaction (which produce
165
carbon monoxide) produces a source term for the carbon : the corresponding
166
mixture fraction is bounded far below one (the carbon can't be free, it is
167
always in carbon monoxide form, mixed with nitrogen or other).\\
168
The retained model for the gas phase combustion is the assumption of diffusion
169
flammelets surrounding particle (for a single paticvle or a cloud), this
170
diffusion flame establish itself between a mixing of the previous gaseous fuels
171
issued from fast phenomenon (pyrolysis or fuel evaporation) mixed in a local
172
mean fuel and the mixing of oxidizers, water vapor (issued from drying) and
173
carbon monoxide issued from slow phenomenon (heterogeneous oxydation and
174
gasification of char). The PDF diffusion approcah is used to describe the
175
conversion of hydrocarbon to carbon monoxide (hydrocarbon conversion is assumed
176
fast vs. mixing); the further conversion of carbon monoxide to carbon dioxyde
177
was (in previous release, still existing for fast first evaluation of carbon
178
dioxide usefull to initialize the kinetic model) ruled by mixing or is (now
179
recommended for better prediction of carbon monoxide at outlet and corrosion
180
risks) kineticaly ruled with respect to the mean mass fraction and temperature
181
(reach of equilibrium assumed slow vs. mixing). Special attention is paid to
182
pollutant formation (conversion of $H_{2}S$ to $SO_{2}$ involved in soot
183
agglomeration, NOx formation).
185
%================================================
165
186
\section{Bibliography}
166
%==================================
167
%==================================
187
%================================================
168
189
\begin{thebibliography}{4}
192
\bibitem{1}%Spalding_1971a
172
193
{\sc Spalding, D.B., {\em et al.}},\\
173
194
{\em Mixing and chemical reaction in steady confined turbulent turbulent flames},\\
174
195
13th Int.Symp. on Combustion , pp. 649-657, (1971).
197
\bibitem{2}%Libby_2000a
177
198
{\sc Libby, P.A. and Williams, F.A.},\\
178
199
{\em A presumed PDF analysis of lean partially premixed turbulent combustion},\\
179
200
Combust. Sci. Technol., 161, pp. 351-390, (2000)
202
\bibitem{3}%Ribert_2004a
182
203
{\sc Ribert, G.; Champion, M. and Plion, P.},\\
183
204
{\em Modeling turbulent reactive flows with variable equivalence ratio: application to the calculation of a reactive shear layer},\\
184
205
Combust. Sci. Technol., 176, pp. 907-923, (2004)
207
\bibitem{4}%Kobayashi_1976
187
208
{\sc Kobayashi, H. {\em et al.}},\\
188
209
16th Int.Symp. on Combustion , pp. 425-441, (1976).
194
211
\end{thebibliography}
196
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
197
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
198
\section{Discr\'etisation}
199
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
200
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
202
On se reportera aux sections relatives aux sous-programmes
203
\fort{cfmsvl} (masse volumique), \fort{cfqdmv}
204
(quantit\'e de mouvement) et \fort{cfener} (\'energie).
205
La documentation du sous-programme
206
\fort{cfxtcl} fournit des \'el\'ements relatifs aux