Solving stokes flow use pde toolbox

Question

Dominique Gorissen 2023-7-13

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1995633-solving-stokes-flow-use-pde-toolbox

评论： Precise Simulation 2024-2-22

I would like the solve a stokes flow equation using the matlab PDE-toolbox.

and

Using an inital guess for the pressure field, the first equation can be solved using the toolbox. This will then give a solution for u that in general does not fullfill the second equation.

I would like to implement a pressure correction to update the pressure field using the second equation and then solve again until the second equation is fullfilled as well. Can someone help me out on this?

Many thanks!!

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Precise Simulation 2024-2-22

You can possibly try to using a potential flow solution with correction for the Kutta condition as described here

https://featool.com/model-showcase/04_Fluid_Dynamics_06_airfoil1

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Torsten 2023-7-13

2
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1995633-solving-stokes-flow-use-pde-toolbox#answer_1272838

编辑：Torsten 2023-7-13

The (simplified) Navier-Stokes equations need a special numerical treatment and cannot be solved by a general-purpose code like the PDE toolbox.

COMSOL with the fluid dynamics package and ANSYS are the commercial code options, OPENFOAM is a non-commercial alternative.

2 个评论
显示无隐藏无

Dominique Gorissen 2023-7-17

编辑：Dominique Gorissen 2023-7-17

Hi Toristen,

I understand you answer, however I do think there is a way to implement it. I should have explained my question more clearly. My idea for implementing is as follows:

make a Pressure array of size 1xNnodes. containing a guess for the pressure in each of the nodes. (this does not have to be the correct pressure just a first guess)
make a function that as input takes: (Pressure array, location.x, location.y, Mesh.nodes) and returns. This function will become the "f" coefficient
apply boundary conditions and parameters a=m=d=0 and c=viscosity
Solve
in each node compute using the "evaluateGradients" function. This should be zero, but since the first guess for the Pressure array is probably incorrect it will not be zero in all nodes. If <0 it means mass is flowing into the node and thus the pressure should be higher there compared to neigbouring nodes. If >0 it means mass is flowing out of the node and thus pressure is to high.
based on results of 4 update the values of the Pressure array and repeat until convergence.