Due to price changes, I have been asked to re-evaluate our existing general purpose CFD code in terms of capability, ease of use and cost, as compared to two other codes.

I am sure that many people will have performed this exercise, and as such I would be grateful to hear your impressions.

Our field of operation is general industrial CFD, which includes just about anything, and the three codes in question are Fluent, Star-CD and CFX.

Any and all input gratefully received....

(1). These are so-called general purpose CFD codes. Each one is a collections of modules and libraries. So, in principle, they are all good for general applications. (2). The easiest way to do is to try out the codes one by one for one year each. And by the end of the third year you should have all the information you are looking for. That is a practical approach. No one else can do that for you. (3). Actually, by the end of the third year, you will have used the current version of the program-A, and the current+1 version of the program-B, and the current+2 version of the program-C. At that point, you can go back to the program-A which will be at its fourth version. (4). In this way, the users and the developers can move forward. (5). Remember that there is no guarantee that a user can always assemble the model and the code together to obtain a solution. Therefore, the best place to start is the engineers who are going to use the codes. And in many cases, codes error and fix are handed down through the support engineer in person until the next version becomes available. This simply requires time and experience. (every vendor will put in as many new features as possible into the code along with some bugs. ) (6). If you are concerned about the cost, just ask your boss to pick one, any code is a good code if the user is an experienced engineer. good luck.

You should also look at CFD Research Corporation (CFDRC - a fairly large US company) software which is physics rich and mature. It compares very well and with the other software that you have mentioned. It is user freindly and the GUI is easy and intuitive to use. Their web site is www.cfdrc.com

Thank you for your detailed response.

I must admit, however, that it gave me a bit of a chuckle. Three years seems a bit far off when I have been given eight days! I know this is ludicrously short, hence my attempt to shorten the time by asking for inputs from others.

Another issue that must be factored into the cost is the change-over expense - learning curve etc. A substantial investment in skills in one code makes it expensive and impractical to switch on a whim.

One final note: If you know how to talk a supplier into giving a one year demo licence, I'd love to hear how!

The reason that CFDRC is not on my list is the latest figures I have seen places their prices way out of the ballpark. We do use their Fastran, but that is a specialized code for special purposes.

I am trying to verify this, however, as my figures could be wrong or outdated.

Thanks for your input.

There seems to be some misunderstanding about CFD-ACE+ price structure. We have some of the better prices among the high-end CFD tools. Please contact me directly (rmm@cfdrc.com), and I would provide you with the correct price information.

(1). There are always free demo codes available on the Internet. But then you are not going to get any support at all. (2). I have just attended a week long training on a popular commercial code. In the training, we couldn't get a sample case to run because of the code error. I am trying to get my test case to run this week, and I already have called support engineer three times through long distance phone call. (3).I had similar experience with other commercial codes. So, to select a code is a very simple decision, but it is just the begining of the real problem. (4).And sometimes, even the support engineer does not have the answer right away, and you have to send your case to them for analysis.

Brian,

By the time you cost in your time to learn a new code and the loss of productivity during that time I would be surprised if it made financial sense to change codes. That is unless the cost differences are huge, which generally for single licenses they are not. Obviously over the years the cost differential can mount up so perhaps a three year budgeting period makes more sense. I know it is often difficult to weigh internal man time, which is often 'free' to a department, against external costs but that is the only sensible way to argue the case.

Robert: I agree absolutely. Believe me, this is not a task I am enjoying. I have ten year's experience with a particular code, and as you say, the hidden costs of a switch can be horrendous. I also have to ask myself whether I want to throw that experience away...

It may take a few days to learn to use a code, but to become a "power user" takes months or years.

But unfortunately managers seem to get a rush of blood to the brain once in a while, and the CFD user gets asked to justify the cost of the code all over again. This has happened to me three times before, and frankly, I'm sick of it.

As to the differences mounting up, this is not even all that likely. This year developer A puts up his prices, next year developer B. Then you start the process again. The result is that you know a lot of codes, but are expert in none.

On another note, just to set the record straight, it seems I was wrong about the price of ACE+. The price is very similar to that of other codes.