Add another twist to this story ... Icepak is a joint venture of ICEM and Fluent.

CFX is too far behind Fluent and StarCD in terms of customers or industry focus to recover as the leading CFD code however targeting the design space with Ansys might be a good idea. New management will certainly also be good for CFX.

As you clearly stated that you have no experience with CFX, one would wonder your sudden very negative "concern" for how CFX will be run by Ansys. Does anyone else smell a rat?

Sniff, sniff.

Yes, I do

Steve mentioned that Fluent was by far the market leader in CFD sales, with Star being second and CFX third. How big is the gap between the different codes in terms of sales ? Bob

I thought the division was roughly 50% Fluent 25% Star 25% CFX

Star and cfx being as big nowadays. But these are very rough estimates.

I would say, that Fluent and Star are almost on the same level (my estimate: 35% Fluent 30%Star, 15% CFX, 20% others)

Star dominates the very important automobile sector. Here CFX does not have the necessary know-how, even if their solver is good.

This data is available at the Aavid and Ansys websites.... CFX had 19 M in sales (had operating loss this year). Fluent should be at least 70 M this year.

Steve2 (not Steve)

Steve,

Yes, ICEM does support many formats, I think I also pointed out the same thing when I said they have been careful not to align themselves in the past. But now, you never know what might happen.

I did not imply that ICEM will be solely focused on CFX. ICEM will continue to sell their products as they have been to keep that revenue (as you point out).

One thing I don't have to guess about is that ANSYS has no choice but to deliver some sort of joing CFX/ICEM product. This is because, being the cutthroat bussiness people they are, they will really want to get rid of the royalty stream that CFX is paying to MSC for using Patran in CFX-Build. ANSYS paying MSC (one of their competitors) money... I'm sure they are not happy with that arrangment.

So, I think what you will see is that ICEM will remain largely unchanged, but some sort of CFX product enhanced by geometry/meshing technology already available at ICEM will appear in the future.

Neale

Neale,

Your last message makes sense. My point before (perhaps too strongly) was that ICEM could not afford to align so closely with CFX that they ignore the majority of their users in the process and we agree. Its an interesting balancing act for Ansys all around it seems.

Steve

When looking at market share, how do you break down the sizes. Do you look at sales figures or profits ? The reason I ask is that both Star and Fluent have very large consultancy groups (as I understand it). If there figures were included in the profits then would that not bias the market shares ? Bob

This is both tedious and wrong. Star is almost non-existent in CP and strong only (really) in auto.

The market share was last year approx

40% Fluent 17% Star 16% CFX

27% others.

I would think its Star who are in the worst position w.r.t. Ansys buying CFX.

I believe consulting sales is much smaller than software sales for CFX and Fluent. I am sure the margins are much lower. I have heard that Fluent has had very good profitability and growth (15-20%) over the last several years. According to the AEA website CFX had an operating loss this year.

Sometime back we looked at CFX but felt they lacked industry expertise and where somewhat unstable as a company. Also CFX 5 was not well tested and lacked models found in Fluent.

Is the rat-smell back?

Hi Steve2,

I'm sorry, but I have read the ANSYS press release (http://www.ansys.com/corporate/news_...s/feb04_03.htm) and do not see where it states an operating loss this year, only revenues of $19 million for last year.

Secondly, CFX has a lot of experience in the CFD industry. CFX-4 and CFX-TASCflow, the legacy products of AEA and the former ASC, have long stood there ground in their respective fields and combined have a lot of capability.

CFX-5 may not have been sufficient for you a few years ago, but has seen intensive development in the last few years, combining the technology originally developed in these older codes. The upcoming CFX-5.6 release will completely replace CFX-4 and CFX-TASCflow for nearly all applications.

I can appreciate your opinion of CFX-5, particularly if you last looked at it two years ago. But it may be worth having another look at it now, particularly considering what may be coming in the future as a part of ANSYS. The techology is there, and any experience CFX may lack in the automotive sector can be gained through their new association with ANSYS ICEM-CFD.

Best regards, Robin

Hello Robin,

as I know that you work for CFX some annotations regarding the the future of CFX in the automotive sector:

The following points will be neccessary to gain market share:

1) The ability of grid-generation for complex geometries. CFX-Build canot do the job. Here ICEMCFD can perhaps help, as their mesher can better handle complex and bad CAD-data. But the ICEMCFD-programs have some severe problems regarding stability and the documentation is bad or non-existant. I have the feeling that star-cd with its unstructured/hybrid hex-mesher pro-am has the best meshing technology for complex geometries and CFX will have to fight hard to compete.

2) CFX must hire some experts for automotive CFD-applications. CFX has knowlege in turbomachinary and process engineering, but if a customer from the automotive sector has a question regarding CFD-simulation of an engine, he first has to explain to the CFX-support how an engine works.

3) CFX must offer interfaces to 1D-engine simulation software such as WAVE and GT-Power. Coupled with a 3D-CFD solver, these are the tools with which engines are designed. Without such a coupling no automotive company will decide to switch to CFX.

4) For in-cylinder calculations moving-mesh technology is essential.

I hope that ANSYS as the new owner will take steps in the right direction in order to establish CFX in the automotive sector.

Best regards, Larry

Hi Larry,

Thanks for the advice and will pass it along to the powers that be.

Best regards, Robin

I have copied the press release and link for you below but if you are an employee of CFX you should already know this. If AEA is wrong tell us how CFX did!

As for "CFD experience" we would expect that from any CFD company. We are more interested in the added value which comes from focusing CFD, related tools and staff on specific customer industries.

Promises of CFX5.6 is not going to cut it. Besides Fluent and Star are improving at the same or better rate and have more years of testing behind it.

http://www.aeat.com/corporate/news/2003/feb_04.htm

4 February 2003

--------------------------------------------------------------------------------

Sale of CFX AEA Technology agrees to sell its CFX software business for around £12.5 million

AEA Technology (AEAT) today announces that it has reached agreement for the sale of its specialist software business CFX to Ansys Inc for around £12.5 million in cash. Completion is subject to certain conditions and is expected by the end of February. The proceeds will be used to reduce the Group's net borrowings.

In the financial year ended 31 March 2002, CFX had sales of £13.0 million and made an operating loss of £1.9 million after exceptional items. Net assets as at 30 September 2002 were £2.5 million. The agreement to sell CFX is part of AEAT's disposal programme.

CFX specialises in developing and supplying computational fluid dynamics software and associated services worldwide. It was part of AEAT's Engineering Software division, the biggest part of which - Hyprotech - was sold last year for £67.5 million. Ansys Inc develops and globally markets engineering simulation software and technologies for engineers and designers in many industries.

For further information, please contact:

Mark Herbert/Catherine Lees, Bell Pottinger 07770 381 608 / 07836 298 811

I myself used STAR-CD (in combination with pro-am) and am now using CFX (also CFX5.6 already). For tet meshes ICEM and pro-am were horrible compared to CFX-build. The hybrid meshes generated by pro-am were also very bad and needed a lot of repairing (VERY time consuming). Putting some extra time in generating a good iges file saves a lot of time fixing it in your CFD package. I can only say that I'm very content with CFX 5, and I'm pleased with the next version (5.6).

Commercial CFD Marketplace, 2000

http://www.hp.co.kr/event/mcae/pdf/T1-1.pdf (1.5Mb), pp.8-11