Error in acusolve
Altair Employee
Dear all,
I have to do CFD Analysis (Rigid Body Dynamics) of a Wind turbine rotation, and I have just come across this error message when running an AcuSolve simulation. Attached herewith is my complete file for .Log and .inp
does anyone know about for this error? Thanks in advance
Best Regards,
Addham
<?xml version="1.0" encoding="UTF-8"?>
Answers
-
Somewhere around time step 9280, the convergence starts getting really poor - both residual ratio (a measure of how well the solution matches the equations) and solution ratio (a measure of how much the solution is changing). Some things to try:
1. I notice the mesh is very coarse - only 109,000 nodes - not much for a wind turbine2. I notice the max_stagger_iterations is set to 0 in Auto Solution Strategy. This will default to 2. You could try setting that to 4 or 6 instead.
3. It's possible the time increment of 0.0015 is too large to capture the rotation. A good rule of thumb is two to five degrees of rotation per time step. Depending on the resulting angular velocity, you may need to reduce the time increment.
It could be a combination of all three - but it appears the errors are accumulating, causing eventual failure.
0 -
Altair Forum User said:
Somewhere around time step 9280, the convergence starts getting really poor - both residual ratio (a measure of how well the solution matches the equations) and solution ratio (a measure of how much the solution is changing). Some things to try:
1. I notice the mesh is very coarse - only 109,000 nodes - not much for a wind turbine2. I notice the max_stagger_iterations is set to 0 in Auto Solution Strategy. This will default to 2. You could try setting that to 4 or 6 instead.
3. It's possible the time increment of 0.0015 is too large to capture the rotation. A good rule of thumb is two to five degrees of rotation per time step. Depending on the resulting angular velocity, you may need to reduce the time increment.
It could be a combination of all three - but it appears the errors are accumulating, causing eventual failure.
Regarding on above solution that your advice me to change I was did it:
1. the mesh element no increase from 109,000 to 115858 nodes
2. max stagger iterations I put it into 4
3. For this one, I did not confirm it. From what I know two to five degrees of rotation per time step is for sliding mesh method.FYI, the simulation was finished after a couple of days but my concern here the validation on that what I run in experimental and simulation analysis give the results a quite bigger gap. As an example, I want to know the rpm of the wind turbine. In the experiment, the rpm is 207 and the simulation the rpm is 92.
Btw, thanks in advance for your response acupro.0 -
I would guess you still don't have enough mesh for accurate representation of the flow and resulting forces. For a 3D model, you would probably need at least 4 to 5 Million nodes.
0 -
Thanks mate for feedback I will do it to enhance no of nodes. I will show to you the result later.
0