Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Failed to find consistent initial values

Please login with a confirmed email address before reporting spam

I am trying to simulate incompressible Navier-Stokes equation.

The setting looks fine and it looks like there is nothing to skip. But I got the following error.

Failed to find consistent initial values.
Last time step is not converged.

Would you tell me what cause this error usually?

P.S.

This is a transient case and inlet sets as flow-rate and outlet is a zero-traction. The inlet flow-rate is defined used functions based on *.txt.


6 Replies Last Post Oct 6, 2010, 2:45 a.m. EDT

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 9, 2010, 2:30 a.m. EDT
The inlet flowrate is too steep for the start I would assume. The workaround is to start from very low and then ramp it up to your desired velocity, or run a steady-state simulation with your initial target velocity and use it as initial condition.
The inlet flowrate is too steep for the start I would assume. The workaround is to start from very low and then ramp it up to your desired velocity, or run a steady-state simulation with your initial target velocity and use it as initial condition.

Rinaldo Gonzalez Galdamez

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 9, 2010, 2:01 p.m. EDT
Hello,

I am having a similar problem with my model. The difference is that it is a multiphase flow. I am using the level-set method. All the subdomain settings and boundary conditions seem to be fine. I initialize the level-set function as it is recommended in the user guide and use that initial value for the transient solution. I keep getting the same error: " Failed to find consistent initial values. Last time step not converged". Could it be that the inlet pressure is too big? Could it be the solver? the mesh?

Thanks in advance.
Hello, I am having a similar problem with my model. The difference is that it is a multiphase flow. I am using the level-set method. All the subdomain settings and boundary conditions seem to be fine. I initialize the level-set function as it is recommended in the user guide and use that initial value for the transient solution. I keep getting the same error: " Failed to find consistent initial values. Last time step not converged". Could it be that the inlet pressure is too big? Could it be the solver? the mesh? Thanks in advance.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 9, 2010, 9:53 p.m. EDT
I am not sure exactly what I done but I am trying to explain what happend.

The result looks fine and maybe I solved this problem. However I am not sure the procedure.

What I done for this problem is that I set up stationary solver 1 and transient solver1 under study 1 and

right click the study 1 and click show basic configuration for the study 1. And I changed the method from iterative to direct.

And I run and get the result without any error. And the result looks fine.

I think this may be the procedure to setup for initial value for transient solver.

But unitl now I am not sure that this setting make the stationary solver run with the initial value from transient case at t=0.

Is there anyone can confirm this?

Thanks in advance.
I am not sure exactly what I done but I am trying to explain what happend. The result looks fine and maybe I solved this problem. However I am not sure the procedure. What I done for this problem is that I set up stationary solver 1 and transient solver1 under study 1 and right click the study 1 and click show basic configuration for the study 1. And I changed the method from iterative to direct. And I run and get the result without any error. And the result looks fine. I think this may be the procedure to setup for initial value for transient solver. But unitl now I am not sure that this setting make the stationary solver run with the initial value from transient case at t=0. Is there anyone can confirm this? Thanks in advance.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Aug 10, 2010, 2:16 a.m. EDT
You use usually the steady-state case as an initial value for the transient case (I think you meant the other way around?).

So first setup your st.st solver, initial values set to 0. Then your transient solver has the initial values set to solution, and the sol is from the st st solver. You can also have a "store solution" on the st.st solver to keep the results and refer to it later.

You use usually the steady-state case as an initial value for the transient case (I think you meant the other way around?). So first setup your st.st solver, initial values set to 0. Then your transient solver has the initial values set to solution, and the sol is from the st st solver. You can also have a "store solution" on the st.st solver to keep the results and refer to it later.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Oct 5, 2010, 11:28 p.m. EDT
Hi,

I'm heading down the same road, but am stuck on something even simpler. In the attached, I solve first for purely stationary "p0" and then want to use that as the initial value for subsequent transient value "p".

If you look at "Line Graph 1", you can see that the limit solution for large time is correct, but it has picked up some seemingly random function for t=0.

Any suggestions appreciated,

Thanks, John
p.s. Duh. Apparently all it took was me posting here. This is working now. I changed the selections for the initial values in dependent variables 2. Initial values now set to "initial expresion" .... second attached file is the good one.
Hi, I'm heading down the same road, but am stuck on something even simpler. In the attached, I solve first for purely stationary "p0" and then want to use that as the initial value for subsequent transient value "p". If you look at "Line Graph 1", you can see that the limit solution for large time is correct, but it has picked up some seemingly random function for t=0. Any suggestions appreciated, Thanks, John p.s. Duh. Apparently all it took was me posting here. This is working now. I changed the selections for the initial values in dependent variables 2. Initial values now set to "initial expresion" .... second attached file is the good one.


Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Oct 6, 2010, 2:45 a.m. EDT
Hi

have you niticed that you have two different pressure values for Pressure 2 in the two Darcy dl1 and dl2. I would expect them to be equal no ?

--
Good luck
Ivar
Hi have you niticed that you have two different pressure values for Pressure 2 in the two Darcy dl1 and dl2. I would expect them to be equal no ? -- Good luck Ivar

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.