Some times inadequate initial guesses or inadequate boundary conditions may lead to this error. In some cases i changed the B.C. or initial guesses and this problem removed. I think this error show that the process of numerical solution can not find a converged solution, and the error is propagated in the domain.

If you are simulating for an unsteady case, e.g, by means of VOF technique, it is usually dependent to the time step size and transient algorithm. Try to decrease the time step to half and always use the Iterative method instead of Non-Iterative method. Hope it works!

Thank you - Is the AMG solver is by default in Fluent, can't I use FAS solver ?

My application is mainly related to Transient and high pressure raise in a short span of time (something similar to Arc Blast)

Therefore, you should certainly use the small time steps in order of 10^(-6). Despite, try to change the default settings for AMG sover, for example use .001 value for pressure instead of default value. In addition, use the PRESTO algorithm for pressure coupling instead of simple. It is very useful for capturing the pressure correctly.

The AMG is the FLUENT default, but I think you can change it if you want but I don't suggest you.

I'm agree with Adel Hashiehbaf, as your case is strongly time dependent use very small time steps, after some iterations maybe you can increase the size of time steps.

thank you to both of you, I am doing in that way and look forward for the positive results.

Can I use density based solver or pressure based solver...? Which is most preferable ?

Dear Adel & Mohammad....I have tried with reduced time steps(10^-6) and PRESTO algorithm...But it is not at all converging...Is this due to some meshing errors...or Do I need to scale down my model and do iterations..? The AMG solver error is gone by reduced time steps but it is not converging, please help mewith your thoughts