[sanbulaya]

I am simulating 2D flow geometry using fluent code. When I patch water with volume fraction of zero at initialization, the program is crushing after 246sec, which again takes so long(5days). Any one with an idea on what is wrong with my simulation. Please help!

[meangreen]

Why are you patching a volume fraction of 0? This may be your problem, but it would be wierd that your initial condition is getting the way after 256 s. Maybe if you provide more information on your system I can help. You generally dont need to patch in a value of 0 somewhere, you can maybe get around it by patching a real value somewhere else.

C

[sanbulaya]

Originally Posted by meangreen

Why are you patching a volume fraction of 0? This may be your problem, but it would be wierd that your initial condition is getting the way after 256 s. Maybe if you provide more information on your system I can help. You generally dont need to patch in a value of 0 somewhere, you can maybe get around it by patching a real value somewhere else.

C

Hullo Meangreen,
Thanks for your advice. I patched in a volume fraction of 1 for water in the domain and time step of 0.001. But again when I wanted to check the monitors, the program was not responding, meaning that it clashed or?
I have specified my boundaries as velocity inlet and space above the ogee spillway as airvent or pressure outlet and the downstream of the domain as presure outlet with backflow vol fraction for water as zero and Dh as 40m whilst inlet vol fract for water as zero and Dh as 33m. The airvent above has Dh=250m. So, how do I get along with this? I selected PISO and transient models and whatever goes with these models

[meangreen]

so a few things to try. I am still not sure what you are modeling, but just some things that may help.

1. if you try to open or close the residuals (or any fluent window) while it is iterating the whole program will crash. So if you ever want to see the plot of residuals or if you want to close an open plot first cancle the iteration then open/close then restart the iteration.
2. If you ever minimize the fluent window while it is working it will probably crash (not respond)

3. try using a very small fraction for the boundary conditions like 1e-5 for the water VOF in. WHen you use a zero, it may be skewing the matrix and fluent will have a hard time converging.

4. instead of making that one boundary condition an air-vent make it a pressure outlet and adjust the boundary conditions so not a lot of flow can come back in.

Let me know how this helps.
C

[sanbulaya]

Originally Posted by meangreen

so a few things to try. I am still not sure what you are modeling, but just some things that may help.

1. if you try to open or close the residuals (or any fluent window) while it is iterating the whole program will crash. So if you ever want to see the plot of residuals or if you want to close an open plot first cancle the iteration then open/close then restart the iteration.
2. If you ever minimize the fluent window while it is working it will probably crash (not respond)

3. try using a very small fraction for the boundary conditions like 1e-5 for the water VOF in. WHen you use a zero, it may be skewing the matrix and fluent will have a hard time converging.

4. instead of making that one boundary condition an air-vent make it a pressure outlet and adjust the boundary conditions so not a lot of flow can come back in.

Let me know how this helps.
C

Hullo Meangreen,
Thanks for your effort to try and assist, since I am very new with this kind of CFD Mathematical modellind of ogee spillway. As I mentioned earlier on, I am modelling a spillway which is a passage of water on top of the concrete dam (30m high and height of flood water is 3m, 4m and 1.5m respectively. The 2D model program is running well and I have to thank you again for your advice. The other problem I am facing is the production of graphs for pressure on the crest of the dam, mass flow rate in postprocessing. I have read the user manual but things are just not coming out. How do I go about it?
Thanking you in advance.

Sandamuh