| Author | Message | |||
     C. Reginald Jegathese (regi) New member Username: regi Post Number: 1 Registered: 11-2003 |
Hello!! There is an image in main page for Two Phase Porous Flow - example. Kindly give the problem (.pde) file name. Thanks. regi | |||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 64 Registered: 06-2003 |
The "two-phase porous flow" animation on our web page is a modification of the example problem "samples | time_dependent | fluids | black_oil.pde". The geometry of the example was modified to fit the web page. | |||
| C. Reginald Jegathese (regi) New member Username: regi Post Number: 2 Registered: 11-2003 |
I have modified black_oil.pde problem for only water seeping through porous sponge. I have enclosed the .pde I am trying to get pressure and velocity distribution. I am getting ERROR "Memory Protection fault" thanks for your help. regi./
| |||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 74 Registered: 06-2003 |
In the original "black_oil" formulation, the mobility M is guaranteed non-zero. In your modification, M goes to zero whenever S is zero. When this happens, the second PDE becomes 0=0, and the numerics crash. Put in a minimum mobility M=Max(S^2/muw, 0.1) or some reasonable number, so that the equation remains nonzero. Or you could use some other formula for mobility which remains nonzero. Also, because of your discontinuous initial condition on S, FlexPDE must cut the timestep to a very small number in the initial steps, until the front smooths out. You can avoid a lot of the time it spends retrying and cutting the timestep if you specify Time 0 to 120*day by 1e-6 (1e-6 is where it will end up after the retries) | |||
| C. Reginald Jegathese (regi) Junior Member Username: regi Post Number: 3 Registered: 11-2003 |
I am trying to analyse for shorter duration. (to simulate flow for an hour).. I am getting ERROR Floating-Point divide by Zero Thanks. regi./
| |||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 78 Registered: 06-2003 |
If you had been watching the output during the run, you would have noticed that long before the divide by zero, the solution was exponentiating at the outlet. This means there is trouble with the validity of the equations. I made several changes to the problem and got it to run seemingly correctly: 1) the original "Black_Oil" problem had an outflow boundary condition for S, which you deleted. I put it back in. 2) The nonlinear coefficients that depend on s^2 are unbounded if s exceeds 1. I put clamps on the value of S in the coefficients. 3) there seems to be a feedback problem at the discontinuous boundary condtion at the outlet. (See our documentation for diatribes against discontinuities). I put a fillet on the corner of the outlet. But most importantly, I am not convinced that by stripping out parts of a two-phase model you have achieved a valid 1-phase model. You should investigate the literature of the subject to determine what the correct equations are.
| |||
| C. Reginald Jegathese (regi) Member Username: regi Post Number: 4 Registered: 11-2003 |
Thanks for your corrections. I will check the equations for porous media for single-phase.. |
modified code of Samples\transient\black_oil.pde