We are getting ready to purchace CFX-5 or Fluent 6 for us on centrifugal pumps, pd pumps and their housing only. Does anyone have any comments for or against either program to help me deside?

Thanks

Have each of them solve a typical problem that you want to model and go with the 'most accurate'. That's what we did. Good luck.

What code did you go for, and why?

It was over 6 years ago, and we went with CFX-TASCflow. We had a particular turbomachinery problem we wanted to model and they were the only close participant at the time (we gave out a sample problem to a number of commercial cfd vendors to solve, if they would, with the understanding that we would purchase the code with the most 'accurate' solution). Their people are very knowledgeable about our type of problem as well. I can't recall, but I believe it was before Fluent had multiple frame of reference capability (could be wrong here). Needless to say, have had nothing but great support from them...

In any case, my old graduate advisor from college, first gave me the idea of the test case. It takes an investment of time on their part, and to a certain degree, they must be confident in their product to attempt a sample problem. Some companies gave up after looking at the geometry - something better to see, before you buy a code...

I have used both. They both have avantages and disadvantages, so probably your best bet is to have both companies do a demo case on a typical problem, see what the results look like, and also have them demonstrate how to set it up.

From the "number of models" point of view Fluent 6 is probably the clear winner there. However, that's where it ends. From the overall stability, robustness and performance point of view CFX-5 wins hands down any day. This is mainly due to the coupled solver, however the post processing and mesher in CFX-5 are also superior to Fluent's offerings. One thing with Fluent is that you generally have more "ways" through the software than in CFX-5, simply because it has more stuff, but complex is not necessarily better, and I'm never impressed when I have to spend time working around a problem in Fluent because some particular feature doesn't work. You do this more in Fluent than CFX-5.

If rotating frame problems are what you need then it's even more in favour of CFX-5 because the MFR/GGI features come from CFX-TASCflow which has the best implementation of these sort of features. However the other features in CFX-5 are also similarly well implemented.

Neale

I disagree, looking at both I would say Fluent 6 is probably 3 years ahead of CFX 5. The MRF capability is superb and the technical support guys are very knowledgeable and prompt. CFX is ok but has some catching up to do, Fluent is a much better all round proposition. I work in the auto industry and to honest I don't know of anyone using CFX.

As far as their solvers are concerned the only thing that Fluent is ahead on is the sheer number of models. So, yes, I agree that Fluent is ahead in the model implementation department at the moment (maybe 3 years is a bit of an overstatement though, I'd say a year at most). I think that one also needs to asses the usefulness of what is there. Fluent has sort of taken a shotgun approach to model development wheras the models put into CFX-5 are more targeted.

The other area Fluent has exceled in is a fairly usable interface, although CFX-5 is better these days, not to mention that it has a better mesher and a more flexible post processor.

Fluent's coupled solver has a farily sparse feature matrix and it doesn't work well at all. I always reverted to using the segregated solver. In this respect Fluent is 5-10 years behind CFX-5 and CFX-TASCflow. CFX-TASCflow had conservative MFR models long before Fluent ever did, and they are much more robust. These are the same models now available in CFX-5. CFX has tended to do well in the chemical process, turbomachinery and rotating equipment markets, and not so much in the automotive industry. But, times are changing....

In the end I just use the tool that is best for the job. These days I find that running CFX-5 is faster and much less painful.

Neale

The opinion I respect the most is the one that comes from industry. Which product is a company prepared to put it's money behind rather than it's mouth (year upon year), in the auto and most other industries it's overwhelmingly Fluent over CFX and I've seen nothing to indicate that changing, lets face it if CFX was so much better and had such a great future then why would AEA technology be putting it up for sale? Further more why has no one spapped up this 'wonderful' technology, I may be wrong but I don't think anyone has been rushing to buy them?

Corus is selling their alumunium department. Don't ask why because that department is currently the only one that makes profit. After the sale, they want to spend the money in the steel department (I start talking like to good old mr. Chien).

So, can you tell me which arguments the AEA board has for selling CFX? Making profit seems to be important but it also seems to be important to stick to your core business. Apparently CFX is not the core business for AEA.

If a company has implemented all their models into FLUENT, it would be idiot to buy CFX and re-implement all models. So, when you say FLUENT is better because it has a larger market share, you are ignoring history. When Neale says CFX is better (assuming he has used both), he refers to the current state of the art.

Astrid

I think market share is a very good indicator, history has shown us that industry uses the best tools for the job - remember when CHAM was the market leader, I actually think CFX was a good deal stronger three or four years ago. Lets not pretend CFX5 is an up and coming state of the art product, it's just better than CFX4 and still lags behind most of the competition (not just Fluent).

CFX-5 VS Fluent 6?

Fluent

The person who said 'I dont know anybody using cfx for turbomachinary' is plain lying. If you read his letter, it is easy to see he either works for Fluent or is closely affiliated.

Why are aea selling cfx? to raise capital.

Why hasn't it been sold yet? It hasn't actually been on the market yet. aeat made an statement of intention to sell it, FY 2002/2003.

The people who have used both, I mean cfx5.5 vs Fluent 6, not earlier versions are choosing cfx5.5 becquse its easier to get a good solution more quickly.

I advise you to try both and not just listen to hype.

In fact, why are you listening to me? If you want the true answer by an unbiased person, i.e. youself, then try them both with an open mind. You might just be surprised.

Gaelle,

Try reading the email properly it says 'automotive industry' and as you well know if you have any grasp of the North American market that CFX has a tiny presence in auto compared to Fluent and Star. I agree with your comment though that the best way to really find out is to try for yourself. Both products can solve the problem fairly easily, there's also the issues of long term corporate stability and direction of the vendor, technical support, training and whether or not you are using software that your customer is comfortable with. Again in my opinion Fluent wins as the best solver and all round solution - at the end of the day it's just my opinion.