I'm doing a 3D transient simulation of a compressor (with non iterative time advancement) I get an error after 18599 iterations: "divergence in AMG solver: temperature". What is quite strange is the moment at which it happens: 18599 because at the 18600th, it normally needs to write my .cas and .dat file (every 200 timesteps). I don't know if it is casuality but in all the cases: can someone help me on the problem?

please send me cause of divergence and how we can solve this problem.

One cause is negative cell volumes. Try grid->check. Look for a negative volume, if so there's your problem and you must re-mesh with much better quality elements.

Carlos.