Hi,

I am modeling a land speed record car, but due to the size and capacity of the machines that I am using I can only apply about 2million cells to my models. This gives my high continuity values. In order to reduced these I have reduced the Under-Relaxation values to (Pressure to 0.2 from 0.3 and Momentum 0.4 from 0.7). These do reduce the continuity values but how does it do this? What am I actually changing. I have looked on the Fluent manual but it is rather complex.

Any help would be greatly appriciated. Dan

URFs when low prevent large changes in variables from one iteration to the next. In this way divergence is resisted but can cause it to take far longer to reach a steady final solution. IMHO you need prism layers and a higher grid count as low URFs are not the answer as even if you do get it to converge the answer may not be correct (or accurate enough).

Typical analyses looking at drag, lift, separation and vortices for auto external aero use 15-25 million cells and that's for half a car. If looking at in yaw then double it. 2m cells just covers the front wing.

Under relaxation essentially takes a fraction (say 0.3 for pressure) of the value from the last iteration to prevent the solution changing too much from iteration to iteration. This means for example that for a relaxation factor of 0.5, with iterations 1 and 2 having data values 1 and 2 respectively, the solution after the second iteration would only actually go to 1.5 instead of 2.(A very simplified version of the process)

I've had alot of experience when I started out with difficult convergence and I've found the best policy is to try and have as good a quality of grid as possible, and preferably hexahedral. These are just the issues that need to be ironed out when modeling a complex geometry.

I try now not to change realxation but if you turn it down, you should do so at the tail end of the solution process - if it doesn't converge at the start then there is an underlying problem, normally grid quality, or boundary conditions.