381
381
\section{Sediments}
383
\fluidity\ is capable of simulating an unlimited number of sediment concentration fields.
384
Each sediment field, with concentration, $c_{i}$, behaves as any other tracer field,
385
except that it is subject to a settling velocity, $u_{si}$. The equation of conservation
386
of suspended sediment mass thus takes the form:
383
\fluidity\ is capable of simulating an unlimited number of sediment concentration classes.
384
Each class has a separate grain size, density and settling velocity. The sediment behaves
385
as any other tracer field, except it is subject to a settling velocity:
388
387
\begin{equation}\label{eq:sediment_conc}
389
\ppt{c_i} + \nabla \cdot c_i ({\bf u} - \delta_{j3}u_{si}) = \nabla \cdot (\kaptens \nabla c_i)
392
Source and absorption terms have been removed from the above equation. These will only be
393
present on region boundaries.
395
Each sediment field represents a discrete sediment type with a specific diameter and
396
density. A distribution of sediment types can be achieved by using multiple sediment
399
{\bf Notes on model set up}
401
Each sediment field must have a sinking velocity. Note that this is not shown as a
402
required element in the options tree as it is inherited as a standard option for all
405
A sediment density, and sediment bedload field must also be defined. The sediment bedload
406
field stores a record of sediment that has exited the modelled region due to settling of
409
To use sediment, a linear equation of state must also be enabled
410
\option{\ldots/equation\_of\_state/fluids/linear}
412
\subsection{Hindered Sinking Velocity}
414
The effect of suspended sediment concentration on the fall velocity can be taken into
415
account by making the Sinking Velocity field diagnostic. The equation of Richardson and
416
Zaki [1954] is then used to calculate the hindered sinking velocity, $u_{si}$, based upon
417
the unhindered sinking velocity, $u_{s0}$, and the total concentration of sediment, $c$.
419
\begin{equation}\label{eq:hindered_sinking_velocity}
420
u_{si} = u_{s0}(1-c)^{2.39}
388
\ppt{S_m} + \nabla\cdot(\bmu S_m) = \nabla\cdot(\kaptens\nabla S_m) - \sigma S_m
391
The advection in the vertical direction is then modified with a downwards sinking
392
velocity $u_{\mathrm{sink}}$.
423
394
\subsection{Deposition and erosion}
425
A surface can be defined, the sea-bed, which is a sink for sediment. Once sediment fluxes
426
through this surface it is removed from the system and stored in a separate field: the
427
Bedload field. Each sediment class has its on bedload field.
429
Erosion of this bed can be modelled by applying the sediment\_reentrainment boundary
430
condition. There are several options for the re-entrainment algorithm that is used to
431
calculate the amount of sediment eroded from the bed.
434
1. Garcia's re-entrainment algorithm
436
Erosion occurs at a rate based upon the shear velocity of the flow at the bed, $u^*$, the
437
distribution of particle classes in the bed, and the particle Reynolds number,
438
$R_{p,i}$. The dimensionless entrainment rate for the i$^{th}$ sediment class, $E_i$, is
439
given by the following equation:
442
E_i = F_i \frac{AZ_i^5}{1-AZ_i^5/0.3}
446
Z_i = \lambda_m \frac{u^*}{u_{si}} R_{p,i}^{0.6} \left (\frac{d_i}{d_{50}} \right)^{0.2}
450
Where $F_i$ is the volume fraction of the relevant sediment class in the bed, $d_i$ is the
451
diameter of the sediment in the i$^{th}$ sediment class and $d_{50}$ is the diameter for
452
which 50\% of the sediment in the bed is finer. $A$ is a constant of value $1.3 \times
456
$u^*$ and $R_{p,i}$ are defined by the following equations:
459
u^* = \sqrt{\tau_b/\rho}
463
R_{p,i} = \sqrt{Rgd^{3}}/\nu
466
This is given dimension by multiplying by the sinking velocity, $u_{si}$, such that the
467
total entrainment flux is:
474
2. Generic re-entrainment algorithm
396
A surface can be defined, the sea-bed, which is a sink for sediment. Once sediment
397
fluxes through this surface it is removed from the system and stored in a separate
398
field: the SedimentFlux field. Each sediment class has an equivalent Flux field.
476
400
Erosion occurs when the bed-shear stress is greater than the critical shear stress. Each
477
sediment class has a separate shear stress, which can be input or calculated depending on
478
the options chosen. Erosion flux, $E_m$ is implemented as a Neumann boundary condition on
479
the bedload/erosion surface.
401
sediment class has a separate shear stress, which can be input or calculated depending
402
on the options chosen. Erosion flux, $E_m$ is implemented as a Neumann boundary condition
403
on the deposition/erosion surface.
481
405
\begin{equation}\label{eq:sediment_erosion_rate}
482
E_m = E_{0m}\left(1-\varphi\right)\frac{\tau_{sf} - \tau_{cm}}{\tau_{cm}}
406
E_m = E_{0m}\left(1-\varphi\right)\frac{\tau_{sf} - \tau_{cm}}{\tau_{cm}}
486
410
where $E_{0m}$ is the bed erodibility constant (kgm$^{-1}$s${-1}$) for sediment class $m$,
487
$\tau_{sf}$ is the bed-shear stress, $\varphi$ is the bed porosity (typically 0.3) and
488
$\tau_{cm}$ is the critical shear stress for sediment class $m$. The critical shear stress
489
can be input by the user or automatically calculated using:
411
$\tau_{sf}$ is the bed-shear stress, $\varphi$ is the bed porosity (typically 0.3)
412
and $\tau_{cm}$ is the critical shear stress
413
for sediment class $m$. The critical shear stress can be input by the user or
414
automatically calculated using:
491
416
\begin{equation}\label{eq:critical_shear_stress}
492
\tau_{cm} = 0.041\left(s-1\right)\rho gD
417
\tau_{cm} = 0.041\left(s-1\right)\rho gD
496
where s is the relative density of the sediment, i.e. $\frac{\rho_{S_{m}}}{\rho}$ and $D$
497
is the sediment diameter (mm). The SedimentDepositon field effectively mixes the deposited
498
sediment, so order of bedload is not preserved.
500
\subsection{Sediment concentration dependent viscosity}
502
The viscosity is also affected by the concentration of suspended sediment. This can be
503
taken account for by using the sediment\_concentration\_dependent\_viscosity algorithm on
504
a diagnostic viscosity field. If using a sub-grid scale parameterisation this must be
505
applied to the relevant background viscosity field.
507
The equation used is that suggested by Krieger and Dougherty, 1959, and more recently by
508
Sequeiros, 2009. Viscosity, $\nu$, is a function of the zero sediment concentration
509
viscosity, $\nu_0$, and the total sediment concentration, $c$, as follows.
511
\begin{equation}\label{eq:sediment_concentration_dependent_viscosity}
512
\nu = \nu_{0}(1-c/0.65)^{-1.625}
515
Note: a ZeroSedimentConcentrationViscosity tensor field is required.
421
where s is the relative density of the sediment, i.e. $\frac{\rho_{S_{m}}}{\rho}$ and $D$ is the sediment
422
diameter (mm). The SedimentFlux field effectively mixes the deposited sediment, so order of deposition
added
removed
manual/embedded_models.tex
