[Abishek]

Is it possible to set the criteria for time-step advancement in an implicit transient FLUENT simulations as:

ABS(Y_current_iteration - Y_previous_iteration)/Y_previous_iteration < 0.0001 ?

where Y is some surface averaged parameter (not equal to 0), for example, average surface temperature.

The reason I would like to use this instead of the traditional residuals is, as follows:

I use the RPI model wall-boiling model and eulerian multiphase model in FLUENT for analysis a problem involving subcooled nucleate boiling. For some reason, the temperatures (say surface temperature) take several iterations (of the order of thousands for my 2D problem) after convergence of the residuals for steady state problems. Sometimes, as ANSYS itself suggests, the residuals (especially that of continuity) do not converge at all to any level, and the promblem is supposed to be deemed converged by monitoring certain other parameters, such as surface monitors and net mass flow rates, etc.