velcity residuals rising
 

Hi everyone.
I've got some troubles with a current calculation. My model is about a heating equipment, filled with oil , which is placed in a volume of air. It's a natural convection case. My problem is that during my calculation , velocity residuals are increasing after 1000 iterations. I had my case terminated by a divergence last week, which I Solved by checking my boudary conditions.

The main cause I suspect is the mesh size, which must be too important for the solver. I need your help to find other causes, in case my mesh refinment won't mange to make the solver converge

Regards...