natural convection

Juliet

hello,everybody

i am solving a problem involved the natural convection and eletro-chemical reaction, i make the sorsca.f,sorent.f,and fluinj.f for some layer, also the density is set to idea f(t) or isobaric. and Gy=9.81, PISO algorithm,steady state. in addition, i add about same length ambient air mesh as actual model. but when calluate it, always divergence and unstable, and enthalpy residuel is INF,or very large. why? i would appreciate if you can give me some explain or advice. if you anybody once have done some work like this or have some example about this, whoud you please sent me that? by the way, for natural convection, how much additional mesh should be? thanks a lot

Juliet