I'm currently looking at several options for purchasing a general purpose CFD code. I know that all commercial CFD codes are limited in their accuracy and require a high degree of knowledge to be applied correctly.

In particular, it would seem that turbulence modelling using variations of the standard k-epsilon model are the basics for any of the commercial codes, and this will always be inherently limited.

One of the codes under consideration is the Flotran module available as a package with the Ansys mechanical/structural solver. It has the obvious advantage of also including a respected and widely used FEA mechanical solver together with the CFD software. However, I believe that the Flotran component is rather more primitive/limited than the standalone codes, eg. Fluent.

I am wondering if anyone has - experience with using Flotran, and comments on its performance, especially if any experimental validation has been performed; - general comments/opinions/information on the accuracy and adequacy of Flotran (compared with specific CFD solvers such as Fluent) for various flow problems.

My main applications are likely to be electronics packaging, heat transfer, natural convection in sealed and vented enclosures, and occasional external flow problems such as flow over aircraft stores.

I have already looked at the "CFD Benchmarks" ASME paper, which was useful but, as stated within it, the findings are only as good as the understanding of the problem by those who provided the model solutions ... and the CFD codes have already grown considerably since that paper was written.

Any assistance would be greatly appreciated. Also, any responses which could be emailed directly to me would be most helpful.

Regards,

Simon Langsford Vision Abell p/l

I can not tell you which code to use. It depends on :1) the problem you are trying to solve, 2) the hardware you have, 3) the experience level of the engineer, 4) the turn-around time of your design process, 5) the amount of money you are willing to spend. The standard procedure is:1) define your problem first, conditions, geometry , solution you are looking for, 2) send it to the software developer to see whether they have sample cases for you, 3) ask them to solve your case problem ( free or at a small cost ) for you. 4) determine whether your system can handle it, 5) determine whether you like it or not. CFD codes are not CAD programs or word processor programs, you really have to do some home work first. By the way, if it's possible, I would run different codes to make sure the solution is reliable for my need. ( you may want to have more than one CFD code in your inventory . This is always a good idea .)

Hi, Simon, try to contact Flotherm, which has been applied by many companies and universities to Electronics Packaging design. People working in Flotherm is very helpful and friendly.

Michael

IcePak 2.2 is a fully unstructured (tetra-hexa) solver for electronic cooling applications. IcePak uses a FV algorithm (same as Fluent 5) which has proven to be stable and computationally economical. As a Fluent product, Icepak enjoys the benefits of leading edge CFD physical models and algorithms.

Automatic cartesian and/or structured grids generation presents serious difficulties for simulations in very complex geometries, specially if the design envelope is being pushed to the limit.

Please check:

www.icepak.com