That's right. If you think about it, ANSYS is now a CFD company, not a finite element structural code company.

Not likely.

The CFX solver is at least 10 years ahead (when compared to Fluent's latest foray into pressure based coupled solvers which is in it's infancy) and 3 years ahead of Fluent in integrating CFX products into ANSYS.

CFX has been grown faster than Fluent in the past 3 years, taking market share away from them mainly....

Even if your right about the VPs, that doesn't mean the entire management structure is being replaced.

Here is the wording from directly from Ansys ....

"Upon the closing of the transaction, Daniel H. Blumenthal, a Managing Partner of Willis Stein & Partners, which is the controlling stockholder of Fluent, will join the ANSYS board of directors. Dr. Ferit Boysan will join ANSYS as Vice President and General Manager, reporting directly to ANSYS's President and CEO. Additionally, Dr. Bharatan Patel will continue to work closely with the combined company to provide his expertise and knowledge to the President and CEO and the Board of Directors of ANSYS under a separate multi-year consulting agreement."

It will be intersting to see how this plays out. Looking at the executive bios I would want the Fluent team runnng the business.

In 1996 (roughly), Ansys offered $50 millions for STAR-CD, which was rejected by the owners. According to its annual turnover and market position and Fluent's saling price, would it be reasonable to assume around $300 millions (just my opinion)?

In 1996 (roughly), Ansys offered $50 millions for STAR-CD, which was rejected by the owners. According to its annual turnover and market position and Fluent's saling price, would it be reasonable to assume around $300 millions (just my opinion)? (I posted in the wrong position above)

Hey, what do I know? You guys sound like insiders, I just read stuff on the web. And that says that fluents revenue is 3x that of CFX and almost equal to ansys as a whole. Buying market share only makes sense if you can keep the customers. I'm just saying its going to be a hard sell to decide to kill fluent.

But feel free to go on believing its all about which is the superior technology.

I agree, but my point is that neither code is going to go away anytime soon.

Guy

IPO = Initial Public Offering

Willis Stein and Partners (www.willisstein.com) owns Avid Thermal Technologies, of which Fluent is a subsidiary. Willis Stein is a holding company and typically keeps companies around for ~7 years before selling them off at a profit. Had ANSYS not agreed to buy Fluent, they would likely have been sold to another firm, or IPO'd on the stock market.

ANSYS' main competitor is ABAQUS, which was recently acquired by Dassault systems. Had ANSYS not agreed to acquire Fluent, Dassault certainly would have. Hence why this is such a strategic move for ANSYS.

Guy

Hi Neal,

there is a wonderful saying "people who are skilled with hammers see every problem as a nail". I guess it is true;-)

I would agree that some of the core techology in the CFX coupled pressure solver is great...but there is much more to a commercial success than the core...and a single solver strategy leaves gaps...

I think your quote was CFX against any solver anytime...do you really think that the coupled solver running transient at Cr=0.01 is going to win over Fluent's explicit solver?

How about a person who just wants to do a 2D case???

How about running Mach 10 against Fluent's density based solver?

Simply put the CFX solver may have some strengths but Fluent has less weaknesses!

Fluent has also demonstrated an ability over the years to get several solvers working together to solve real problems out there...that is where the money is!!

Does everybody out there really want all the latest stuff too.....no not always in business....

Take somebody who has been doing years of standard simulations with tried and true tet meshes and boundary layers in a cell centered code, with confidence on how to generate meshes for acceptable resolution, yplus, etc.

Then the new solver (like CFX) is vertex centered - giving a "volume coarsening" ratio of ?) and a different first cell off the wall height. On top of this the default wall function does a: if Yplus < 11.06 then Yplus = 11.06

...those are a lot of changes.....a company then has to totally re-evaluate their meshing strategy....a lot of work.

What if you want to use your "old" cell centered code along with it????

And I am not saying one is better than the other...!

I like to promote people thinking about things...not just spreading rhetoric.

Regards,

Bak_Flow

Stuart, you bring up a good point. A driving force behind this deal is ANSYS's desire to stay ahead of Dassault. They acquire market share and at the same time block their competitor. Now Dassault and MSC are left to fight over Adapco - nice position for Adapco to be in...!

I think $565M is a good price considering Fluent's revenue, growth, and (assumed) profit margin. Overall this is a good deal for ANSYS, they know what they're doing.

The only losers here are CFX folks (customers and those who develop/sell/manage the code), ANSYS is now forced to focus their attention on Fluent, the new big earner. Remember, they are publicly traded. Investment houses (who own 92% of ANSYS) will apply a lot of pressure to ANSYS management to justify the deal.

I meant Guy had a good point about Dassault (although maybe Stuart did too, I'm not sure...)

I wouldn't be so quick to assume that either code would be dropped. Fluent will certainly make up a larger portion of the CFD part of ANSYS, but they aren't stupid.

A sound approach would be to objectively benchmark the capabilities, something that was probably next to impossible to do before. There will certainly be gaps, and one would hope the new organization looks at the best way to close those gaps. In the meantime, they could each lend their expertise to the other's development activities, assuming personalities don't get in the way.

In the long term, I think the customer be the one who benefits the most.

Guy

Ansys bought Fluent because otherwise CATIA would. Ansys wants to make sure that a line CATIA-Abaqus-Fluent cannot be made (Abaqus was making a move towards Fluent).

Engineering packages (CATIA, Pro/E) are MUCH larger than FEM (Ansys, Abaqus), which are much larger than these tiny CFD communities.

The combination of a range of engineering packages (drawings, stress analysis, CFD) under 1 umbrella is very powerfull. Ansys wants to make sure they have the full line with top-of-the-market packages.

It had nothing to do with CFX versus Fluent, believe me. Although I must say: CFX ROCKS!!!!! ;-)

"do you really think that the coupled solver running transient at Cr=0.01 is going to win over Fluent's explicit solver?"

This statement implies that CFX is slow at running time accurate transient, which is totally false.

The CFX solver can run 1 coefficient loop per timestep at that sort of Courant number. So, then all it comes down to is a question of solver speed, i.e. which code is better optimised in terms of memory/CPU usage. In a fair comparison my bet is that the two codes are not that far off.

"How about a person who just wants to do a 2D case?"

This statement implies CFX can't do 2D, which is false.

2D cases are run in CFX by extruding to one element thick. This works fine.

"How about running Mach 10 against Fluent's density based solver?"

This statment implies that CFX cannot run high mach number flows, or cannot run them as well as a density based solver. This is completely false.

People run calculations like this in CFX 10 all the time. A density based solver has no clear advantage over pressure based for high Mach number. This is a falacy.

"Simply put the CFX solver may have some strengths but Fluent has less weaknesses!"

I guess this is in the eye of the beholder....

"Then the new solver (like CFX) is vertex centered - giving a "volume coarsening" ratio of ?) and a different first cell off the wall height."

This statment is ambiguous.

Yes, this is true, CFX is cell vertex based, not cell centered. I don't know what you are trying to say about volume coarsening ratio though. Control volumes in CFX are formed around the vertex. The volume coarsening will be similar to the cell based coarsenting.

"On top of this the default wall function does a: if Yplus < 11.06 then Yplus = 11.06"

What's wrong with this assumption?

If the scalable wall function is not for you then you should run SST. Standard wall functions do not converge to a mesh independent answer, hence the reason for the existence of a scalable w.f.

* "...those are a lot of changes.....a company then has to totally re-evaluate their meshing strategy....a lot of work."

Why do you have to revaluate meshing strategy? The meshing requirements are similar for cell-centered and cell-vertex. More cells in the b.l. is better... what's the difference?

* "I like to promote people thinking about things...not just spreading rhetoric."

Absolutely, so stop spreading rhetoric. Much of what you say makes no sense or is just false.

Hi there Hek

this thread is getting old so I can only reply to a couple of your comments:

"This statment implies that CFX cannot run high mach number flows, or cannot run them as well as a density based solver. This is completely false."

I agree that statement is false...(not sure I would use the terms completely false).....that is why I DID NOT SAY IT...

"this statement implies" that is putting words in my mouth...or keyboard...monitor..bus...etc ;-)......don't do that and then slag me about rhetoric...and telling me to stop.

But besides the argumentative tactics.....I have run both codes find Fluent's density based solver has sharper shock resolution than CFX....my opinion from my expereince that is all!

It appears as though you did not get my point...fluent has a number of different solvers and options...like 2D where it is by no means impossible to run in CFX-10...but the user has to play around to get what they really want!

Best Regards,

Bak_Flow