Restart problem using PIC method

Dear all,

I am using the PIC method with the version of 19.2.0(newest)
First, I calculate the case on my local computer for a while
Then, I transfer the case to the cluster and restart it, but there is an error occurred


What can I do about this, many thanks!

Kind regards

It could be several things.

  • Your restart file could be corrupt.
  • Your restart file may be empty (not written from previous simulation).
  • The version written by one machine isn’t compatible to the other.

It is recommend you not move restart files between machines for this reason.

Dear Dr. Musser
It does works and thanks for your suggestion.
However, I have got another problem with chemical reaction.
1, When I enable the species equation of gas (species_gas(0)= true) without any chemical equations, in my mind, the species equations should not be solved, but the residual of the gas species still get some values rather than zero


2, Based on the above settings, the X_g(1) always larger than 1 without any chemical reaction sets
3,
The calculation crashed after four timesteps without further detailed information and I have got no idea about how to improve

Appreciate a lot!

If you enable the species equations, then they are solved. If there are no chemical reactions or phase changes are defined, (i.e., usr_rates.f and usr_rates_des.f have not been modified), then you are solving a basic scalar transport equation.

By default, MFIX initializes the domain to the first species and sets all other species to zero. You can change this by modifying the gas composition in the initial conditions. In you case, all species except the first have a value of zero everywhere. The result is they have no-error (as there is no value to transport). However, the non-zero first species has some error given that it is non-zero, but the amount should be rather small. If it isn’t, it could indicate that your mesh quality is poor.

Check to see if there are errors in your LOG file. If you ran in parallel with MPI, there could be multiple LOG files. The route shown in the back trace, check_data_30.f assesses the value of field variables and stops the simulation if unphysical values are detected.

Many thanks!
I have refined the mesh but still the problem
image
The problem occurred only five timesteps without reactions. I guess the reason is not due to the unbalance of the mass fraction. Some other properties lead to this but I am not sure.
Besides, the value of the fraction is one, why it says it beyonds the bound

Kind regards

I have resolved this problem temporarily and looking forward to the results with chemical equations provided. Below is the explore.

  • I print the value of x_g and found that the value beyond one is really small
    image
  • I found the keywords in the check_data_30 subroutine
  • The limit value here is one so I change it to an appropriate value
    and the problem has gone