|
Palabos
Version 1.1
$projectbrief
|
$projectbrief
|
$searchbox |
plb::Compute_C_processor< T, Descriptor > Class Template Reference
Compute the concentration C of the heavy phase. More...
#include <heLeeProcessor3D.h>
Public Member Functions | |
| Compute_C_processor (T M_) | |
| virtual void | processGenericBlocks (Box3D domain, std::vector< AtomicBlock3D * > blocks) |
| virtual Compute_C_processor< T, Descriptor > * | clone () const |
| virtual void | getTypeOfModification (std::vector< modif::ModifT > &modified) const |
Detailed Description
template<typename T, template< typename U > class Descriptor>
class plb::Compute_C_processor< T, Descriptor >
Compute the concentration C of the heavy phase.
GENERAL CONCEPTS ================ In this implementation of the He/Lee binary fluid, the macroscopic variables like velocity and concentration are stored in separate fields (unlike the Shan/Chen implementation in Palabos, in which this kind of variables are stored in external scalars). The reason is that the model makes a massive use of finite difference schemes to compute derivatives, and derivatives of derivaties (up to fourth derivative). Derivatives are non-local, and so, before each deriva- tive some inter-block communication is needed to fill the envelopes. This is most easily achieved by putting each macroscopic variable in a separate field which then spontaneously updates its envelope when it has been modified.
LITERATURE ==========
- T. Lee and C.-L. Lin, J Comp Phys 206 (2005), 16-47.
- T. Lee, Comp Math App 58 (2009), 987-994.
VARIABLES ========= The variables, each stored in a separate lattice or field, are: Block-Lattices:
- g: "The fluid" --> velocity u and pressure p1.
- f: "The concentration" --> Conc. C, density rho and potential mu. Scalar-Fields:
- C: Concentration of the heavy fluid.
- rho: Total density, summed over both components.
- mu, laplaceMu: Chemical potential and its Laplacian.
- p1: Flow pressure. Tensor-Fields:
- gradC, gradMu: Gradients of C and mu.
- u: Flow velocity.
Note that not all terms used in the He/Lee model are not represented as fields. For example, although the gradient of the pressure is used in the model, it is not stored in a tensor-field; instead, it is computed on-the-fly during the collision to save memory and to improve efficiency. This would not have been possible for the gradient of C, for example, because this quantity is used not only directly in the collision, but also for the computation of the velocity u.
DATA PROCESSORS FOR THE IMPLEMENTATION OF THE MODEL =================================================== The He/Lee algorithm is split over 4 data processors; the three first are needed to compute the macroscopic variables and their space deri- vatives from the populations f and g; the fourth implements the collision step.
- Compute_C_processor In: f, laplaceMu(t-1) Out: C
- Compute_gradC_rho_mu_processor In: C Out: gradC, rho, mu
- Compute_gradMu_laplaceMu_u_p1_processor In: g, C, rho, gradC, mu Out: gradMu, laplaceMu, u, p1
- HeLeeCollisionProcessor In: f, g, C, rho, gradC, mu, gradMu, laplaceMu, u, p1 Out: f, g
Note that all processors take the field f as their first argument, even if they don't need it, for a purely technical reason: in this way they can all be added as internal processors to f, to guarantee that they are executed automatically.
DATA PROCESSORS FOR SIMULATION SETUP ==================================== To create an initial condition, the user must assign a value to C, u, and p1. Then, derivative variables can be computed by invoking Compute_gradC_rho_mu_processor. A special data processor is available to compute the derivatives of mu without overwriting u and p1:
- Compute_gradMu_laplaceMu_processor In: mu Out: gradMu, laplaceMu Finally, the HeLeeCollisionProcessor can be used to initialize f and g to their equilibrium, by constructing this processor with an optional boolean argument initialize=true. In: f, laplaceMu(t-1) Out: C
Comment: laplaceMu is taken from the previous time step to preserve an explicit scheme.
Constructor & Destructor Documentation
| plb::Compute_C_processor< T, Descriptor >::Compute_C_processor | ( | T | M_ | ) | [inline] |
Member Function Documentation
| Compute_C_processor< T, Descriptor > * plb::Compute_C_processor< T, Descriptor >::clone | ( | ) | const [inline, virtual] |
Implements plb::BoxProcessingFunctional3D.
| void plb::Compute_C_processor< T, Descriptor >::getTypeOfModification | ( | std::vector< modif::ModifT > & | modified | ) | const [inline, virtual] |
Implements plb::BoxProcessingFunctional3D.
| void plb::Compute_C_processor< T, Descriptor >::processGenericBlocks | ( | Box3D | domain, | |
| std::vector< AtomicBlock3D * > | blocks | |||
| ) | [inline, virtual] |
Implements plb::BoxProcessingFunctional3D.
The documentation for this class was generated from the following files:
1.6.3