different mesh size but same answer
 

same domain
but refinement mesh size 2 million ; 4 millions ; 6 million ; 8 million
still get same answer for coefficients
please help me:(

Without knowing more about your domain or simulation in general or what you consider the "same", it sounds like your domain is "mesh independent".

I'd agree with lcarasik.
Additional remark: have you heard of grid convergence index?
If not: have a look into the internet with your favorite search engine.

Cheers
Martin