|John C. Chien
||August 31, 1998 12:12
Re: How does FLUENT fair in absolute heat transfer analysis
I was a Fluent/UNS user and I had an opportunity to study the heat transfer problem related to the turbulent, separated flow in a turbine disk cavity. The main conclusion derived was that the heat transfer coefficient on the wall was sensitive to the mesh size ( triangular mesh) distribution on the wall. The problem can be minimized by using uniform mesh ( triangular mesh) on the wall, but in real applictions, it's not practical sometimes. Since the release of the " hybrid mesh" option with packed prism layers on the wall, I have not looked into this problem yet. In recent years, both AIAA and ASME have looked into the solution validation problem and have published some comparison results. Since the commercial codes are esentially black box to the users, it can always argue that users do not fully understand the code, or how to operate the code. So, it becomes the responsibility of the code developers to provide the benchmark test results to satisfy the need of the user community. Since the benchmark testing is always time concuming, sometimes it is not practical to get the quick answer to your question. You should present the problem to the code developer to see whether they can give you some benchmark test results. But in general, in terms of degree of difficulties, the sequence is something like: pressure field, velocity field, turbulence field, wall skin friction, wall heat transfer. So the heat transfer at the wall is among the most difficult quantities to predict. This raised the question of " can you obtain correct heat transfer coefficient at the wall by using a layer of irregular trangular mesh on the wall ? "