I am facing similar issues when running fan simulations. I prefer running transient simulations using the k-e realizable turbulence model with scalable wall functions. I define fan boundary conditions as pressure - outlet (or pressure- inlet, pressure-outlet) while giving a rotational frame motion.

There might be problems with geometry (e.g. not proper boolean operation) or mesh, which seems unlikely. You could create only a cylinder as fan (or an axial fan with blades enclosed in a cylinder) in the room and define a rotational frame motion, if you want a steady-state simulation.

The convergence in the continuity equation can affect the numerical stability of the other equations. Please consider the below troubleshooting.

• is your simulation under the first-order discretization?
• reduce the under-relaxation factor.
• check your meshes, not including the ones with high skewness or aspect ratio.
• steady-state results can initialize the stable simulation when used as initial conditions.

Dear Bhat,

I am wondering why you modelled your fan as (as a simple cylinder with Inlet) while there is a chance to use a ready fan model in ANSYS?

Off course, like those errors are expected when you follow the method which you now do.

As for your current case, please try to benefit from the message that the software shows when it get diverged.

Kind regards