CFD Online Discussion Forums

CFD Online Discussion Forums (
-   NUMECA (
-   -   Where to find comparisons of CFD FEM codes (

J.T. March 6, 2003 22:32

Where to find comparisons of CFD FEM codes
Hi, All. There are so many CFD FEM comercial codes in the market. I need some advice of picking up a package for running Incompressible Turbulent Flows inside an eclosure. So far we have a budget of more than $80,000 annual license fee to lease or buy a software. We like to make sure it does the right jobs. Thanks!

Bruno Gouverneur March 10, 2003 11:55

Re: Where to find comparisons of CFD FEM codes

May I suggest you to describe more precisely the type of application that you would like to simulate and any other constraint that you are facing in your decision process ? If this is too confidential to be described on this public forum or if you would like to send me a geometry, do not hesitate to communicate directly by e-mail with me.

I am sure that we, at NUMECA, can offer you the most suitable solution to your problem, and I would be happy to demonstrate this to you.

Best regards,


A.S. May 24, 2003 12:03

Re: Where to find comparisons of CFD FEM codes

What kind of application are you looking for.

For chemical, you can consider StarCd and Fluent. For automotive, Star CD, for Incylinder flows : AVL Fire For Turbomachinery : CFX and Numeca For Aerospace : CFD-ACE (cfdrc's product), Cobalt CFD (, ACRI

The list is endless.

In my knowledge, normal licencing fees is around 20000 dollars per annum. Please keep your budget confidential.


James July 8, 2003 21:41

Re: Where to find comparisons of CFD FEM codes

Public comparisons of CFD codes are hard to find. In most cases, users will benchmark the codes for themselves.

The top three codes (in no particular order) are from CD-Adapco, (Star-CD), Aavid (Fluent), and Ansys (CFX). There are many other smaller codes, but there models tend to be more limited and they tend to concentrate or specific appliation/flow regimes.

CD-Adapco ( is primarily a consulting company. Star-CD has the advantage of having everything within a single solver and can handle a wide range of element types (if you have a grid generator that can create them). The solver is a pressure based, segregated, unstructured solver with User Fortran hooks into just about everything. It is cheap, but be prepared to spend a lot of time getting it to run. Fortunately, consulting is easy to come by at CD-Adapco, so if you can't get it to work, you can pay someone there to do it for you.

Aavid ( is a little more diverse, primarily concentrating on Thermal Technologies, although Fluent is the biggest moneymaker among it's ranks. Fluent has a lot of capability spread out over multiple solvers. A single user interface is used to set them up, but capability can be severely limited if you pick the wrong solver. The primary solver, Fluent UNS, is a pressure based, segregated, unstructured solver. Code modification is done through C subroutines. Convergence behavior tends to be good, but degrades with larger grid sizes. Better than Fluent's solver is there marketing machine, which is capable of convincing you that CFD can cure cancer (if they spend half of their marketing budget on R&D, this may be the case:).

Ansys ( is focused on engineering simulation with extensive multi-physics capablity. Up until recently, their only CFD offering was in the form of FLOTRAN, which was acquired a number of years ago and not further developed, although it is integrated with their existing FEA code. Ansys recently acquired CFX from AEA Technology, adding full-blown CFD to it's portfolio. There are three CFX codes: CFX-4, CFX-TASCflow and CFX-5. The first two are older, multiblock-structured codes and are no longer under development, although they will still sell them to you if you really want them.

CFX-5 is their flagship code. It has a coupled-multigrid, pressure based, unstructured solver, which is it's stongest selling point. The coupled-multigrid solution algorithm is faster, more efficient and scales linearly with grid size, although it does require more memory than other solvers. It also has a very good parallel implementation. Customization is done through an expression language, which allows you to simply enter equations where you need them instead of resorting to user subroutines (although user subroutines are available if required). It is a relatively new solver from an experienced CFD company and has lacked the necessary physical models to compete with Star-CD and Fluent for a number of years. However, there have been three major releases in the past 2 years, each adding significant capability and a lack of physical models is no longer the case. If I were a betting man, I would bet my life's savings on Ansys to dominate CFD in the coming years.

Ultimately you should choose the code that fits your budget and takes care of your needs. Chances are you do not have to look beyond these three, particularly given your budget.

Good luck, James

J.T. July 10, 2003 17:07

Re: Where to find comparisons of CFD FEM codes
Hi James,

Thank you for your kindly answers. I totally agree with your professional opinions. Now I found out most difficulties lay on the side of generating the HEX for very complicate geometries. I have tested the meshes generated by auto-tets-generator for the problems of unsteady turbulence flows using k-e model. The results are disappointing, sometimes it even diverged. HEX is the only hope for my simulations, Where to get the better HEX generator which can semi-automatically generate the HEX meshes and we can locally modify it (I realized that the fully auto Hex- generator only can be true in my dream.). So far I tried Harpoon, ANSYS ICEM, but not NUMEC HEX express. I don't know how good is NUMECA HEX Express. Thanks!

All times are GMT -4. The time now is 15:04.