Velocity-pressure coupling
From CFD-Wiki
(One intermediate revision not shown) | |||
Line 8: | Line 8: | ||
==Formulation== | ==Formulation== | ||
we have at each cell descretised equation in this form, <br> | we have at each cell descretised equation in this form, <br> | ||
- | :<math> a_p \vec v_P = \sum\limits_{neighbours} {a_l } \vec v_l - \frac{{\nabla p}}{V} </math> ; <br> | + | :<math> a_p \vec v_P = \sum\limits_{neighbours} {a_l } \vec v_l - \frac{{\nabla p}}{V} </math> ; Where V = Volume of cell.<br> |
- | we have <br> | + | According to [[Rhie-Chow interpolation]], we have <br> |
:<math> \vec v_P = \frac{{\sum\limits_{neighbours} {a_l } \vec v_l }}{{a_p }} - \frac{{\nabla p}}{{a_p V}} </math> <br> | :<math> \vec v_P = \frac{{\sum\limits_{neighbours} {a_l } \vec v_l }}{{a_p }} - \frac{{\nabla p}}{{a_p V}} </math> <br> | ||
Line 15: | Line 15: | ||
:<math> \sum\limits_{faces} {\vec v_f \bullet \vec A} = 0 </math> <br> | :<math> \sum\limits_{faces} {\vec v_f \bullet \vec A} = 0 </math> <br> | ||
so we get: <br> | so we get: <br> | ||
- | :<math>\left[ {\frac{{\sum\limits_{neighbours} {a_l } \vec v_l }}{{a_p }}} \right]_{face} - \left[ {\frac{{\nabla p}}{{a_p V}}} \right]_{face} = 0 </math> <br> | + | :<math> \sum\limits_{faces} \left[ {\frac{{\sum\limits_{neighbours} {a_l } \vec v_l }}{{a_p }}} \right]_{face} - \sum\limits_{faces} \left[ {\frac{{\nabla p}}{{a_p V}}} \right]_{face} = 0 </math> <br> |
this gives us: <br> | this gives us: <br> | ||
- | :<math> \left[ {\frac{{\sum\limits_{neighbours} {a_l } \vec v_l }}{{a_p }}} \right]_{face} = \left[ {\frac{{\nabla p}}{{a_p V}}} \right]_{face} </math><br> | + | :<math> \sum\limits_{faces} \left[ {\frac{{\sum\limits_{neighbours} {a_l } \vec v_l }}{{a_p }}} \right]_{face} = \sum\limits_{faces} \left[ {\frac{{\nabla p}}{{a_p V}}} \right]_{face} </math><br> |
defining <math> H = \sum\limits_{neighbours} {a_l } \vec v_l </math> <br> | defining <math> H = \sum\limits_{neighbours} {a_l } \vec v_l </math> <br> | ||
- | :<math> \left[ {\frac{1}{{a_p }}H} \right]_{face} = \left[ {\frac{1}{{a_p }}\frac{{\nabla p}}{V}} \right]_{face} </math> <br> | + | :<math> \sum\limits_{faces} \left[ {\frac{1}{{a_p }}H} \right]_{face} = \sum\limits_{faces} \left[ {\frac{1}{{a_p }}\frac{{\nabla p}}{V}} \right]_{face} </math> <br> |
from this a pressure correction equation could be formed as: <br> | from this a pressure correction equation could be formed as: <br> | ||
- | :<math> \left[ {\frac{1}{{a_p }}H} \right]_{face} - \left[ {\frac{1}{{a_p }}\frac{{\nabla p^* }}{V}} \right]_{face} = \left[ {\frac{1}{{a_p }}\frac{{\nabla p^' }}{V}} \right]_{face} </math> <br> | + | :<math> \sum\limits_{faces} \left[ {\frac{1}{{a_p }}H} \right]_{face} - \sum\limits_{faces} \left[ {\frac{1}{{a_p }}\frac{{\nabla p^* }}{V}} \right]_{face} = \sum\limits_{faces} \left[ {\frac{1}{{a_p }}\frac{{\nabla p^' }}{V}} \right]_{face} </math> <br> |
- | This is a poisson equation. | + | This is a poisson equation. |
Here the gradients could be used from previous iteration. | Here the gradients could be used from previous iteration. |
Latest revision as of 05:50, 24 October 2005
If we consider the discretised form of the Navier-Stokes system, the form of the equations shows linear dependence of velocity on pressure and vice-versa. This inter-equation coupling is called velocity pressure coupling. A special treatment is required in order to velocity-pressure coupling. The methods such as:
- SIMPLE
- SIMPLER
- SIMPLEC
- PISO
provide an useful means of doing this for segregated solvers. However it is possible to solve the system of Navier-Stokes equations in coupled manner, taking care of inter equation coupling in a single matrix.
Contents |
Formulation
we have at each cell descretised equation in this form,
- ; Where V = Volume of cell.
According to Rhie-Chow interpolation, we have
For continuity :
so we get:
this gives us:
defining
from this a pressure correction equation could be formed as:
This is a poisson equation.
Here the gradients could be used from previous iteration.
SIMPLE
See SIMPLE algorithm
SIMPLER
SIMPLEC
PISO
See PISO algorithm
Return to: