Has anyone seen this STAR-CCM+ code. It looks like the cfd software built as we expected it from a long time: with client-server and window tabs. You can also mesh your CAD geometry stuff into Solidworks and Pro-engineer and then start directly STAR-CCM+. It is plugged-in!!!

At last! This is how a cfd-code should be dude!

Have you seen the demo?

H.C.

NOPE DUDE There is much more to a CFD code than just some nice windows tabs

Funny that, I'm sure there's a direct correlation between the number of tabs and the resulting integrity of the solution

Instead of always announcing anything "new" CD should first release STAR 3.2 finally. Every three month they postpone the release date for the new version. It seems to me that they are not able to publish a new version and all the announcments are only good for showing possible customers that they are doing something so that they can still sell there 3 or 4 years old cfd code 3.15 which is absolutly out of date. Meanwhile I'm absolutley sure, that version 3.2 will not be release this year. I would bet on it.

Rather the integrity is inversely proportional to the # of tabs

...Or rather the cynicism increases with the ease of using a tool. Heaven forbid we move away namelist input, recreating geometry time-and-time again in different tools, ignoring real physical features of a component design, etc.

I agree there is a worring trend for designers to want to just "Do CFD" as if it takes as much skill as learning AutoCAD. To this end the packages that look user friendly get the sales and make $$$ for the companies even though the product is no doubt "abused" by the end user.. Garbage In - Garbage Out!

Well said - actually to obtain a good results from CFD, one needs to understand Fluid Mechanics as well as knowledge of numerics. While easy of use is important for users, it's important not to lose the sight of the core technology of CFD, which is the numerical discretisation of a set of governing equations.

Couldn't agree with you more. I understand fluid mechanics very well - what I do not understand is the mathematics that goes into creating geometry in CAD or a discretized mesh. Don't really want to know this. What I want is an easy way to work with an EXISTING 3D geometry (i.e. CAD model) and create a mesh, whether for CFD, FEA, or what other acronym you want to use. Also want an easy method to judge the quality of this mesh so it is appropriate to the physics I am trying to model. What I don't want to be forced to do is recreate a (simplified) geometry in a mesh generater after investing so much time in a 3D CAD model.

Just because something is more difficult to use doen't make it more correct. Difficulty = more correct is the message that so many CFD developers seem to perpetuate. What rational users are asking for is an easier way of doing repetitive and complex tasks, along with a method to evaluate the quality of the result. I do not feel this is losing site of the core technology of CFD or FEA.

I have tested some commercial packages and I most say that the ease of using them are bit different. (not talking about the mesh but pure solving and setting up the problem) Some times is more difficult = more correct ! In some cases it is very easy to set up the case, but with only very few options. I guess this is the user friendly part !! Don't get confused about different models, just use the simple default.

Commercial CFD softwares always said their products were very easy to use. When you really get into it, you will find lots of bugs after wasting lots of time figuring out what the problem really is and the support will always ask you to wait for the next version. Ha... It's totally like Microsoft product. How much time can you afford to wait for the crappy next version?

There's still a lot to catch up in the pre processing tool. Most commercial CFD softwares ignore the significance of this process, which waste users lots of time creating some crappy mesh and cause the error in the solution later.

The problem with the "garbage in - garbage out" philosophy is that people are inclined to believe that the inverse is also true, i.e. "good stuff in - good stuff out". That is not the case, it is entirely possible to have "good stuff in - garbage out"!

It is very strange that it took such a long time to get v3.2 released - probably due to the top management change and development resource allocation. CD has been developing 3 codes: v3.2, star-ccm (an upgrade version of star-cd), and star-ccm+ (done in North American by mostly former Fluent employees). And as I learnt, they also lost their 2 most experienced chief engineers in 2001.

Oh gee - I actually have to decide to 1 significant figure how many cells to use. Why doesn't the stupid mesher know what I want to do even when I don't.

Oh gee - I tried to put 150 layers in a gap 1 micron wide and 3 km long and now the stupid mesher gave me a mesh with aspect ratio problems.

Oh gee - why does the solver need a boundary layer? - it must be full of bugs.

Oh gee - there is so much high quality free software around, why would anybody spend money on a commercial code.

I just love people who whine. If commercial software is so bad then go out and write the perfect code yourself. No one stuck a gun to your head to force you to buy it.

I agree with Dudley. CFD requires knowledge of FD! If people aren't aware of what equations are being solved then they are bond for disaster when they try to interpret the answers to those equations!

Furthermore, the idea of simple, "push button", CFD packages for design engineers is ludicrous. CFD people don't buy design software for analysts that are "part time designers" so why should the reverse be true?

No CFD job is ever straight forward. For example consider the designer of a front grill for a car. He / she has just got a push button CFD package and wants to analyse the cooling ability of several different front grill designs. They throw the CAD data at the CFD package and let it compute flow rates. The CFD package inverts the volume and computes a mesh out to some bounding box. The designer now applies a inlet speed and outlet pressure just like the tutorial.

Assuming this works without any hassle, how is this any different to what an analysis would do?

My answer to this is that the experienced analyst would immediately realize that the grill is connected to a car and that part of that car would need to be included in the analysis. But how much of the car? Hence the need for engineering judgment and CFD experience!!

I wouldn't be too hard on the companies developing "CAD-integrated CFD", it is a logical development and shows much promise. On the Design vs. FD vs. CFD argument, any designer working in a field where fluid dynamics is important needs to understand his Fluid Dynamics. It really doesn't matter whether he uses "classical" analytical methods or CFD, these are just tools, which he should be able to use appropriately. Working on understanding the physics of fluid flow is what counts.

Michael, Thank you for your support, but I do agree very much with Charles' reply. One assumes that designers running CAD enabled CFD will not be solving eulerian two phase flow transients with multiple chemical reactions. They should be able to solve and understand problems with a limited range of physics and simple boundary conditions. I just get tired of people who claim to know what they are doing who blame every mistake they make on "lousy" commercial software. Nobody that I know was ever forced to buy it and the analysts where I work turn out quite a bit of high grade work with commercial software. It may not be perfect, but very little software (including open-software) is.

I guess the point I am trying to make is this - Are there actually any "simple" CFD problems out there in industry? The example I gave was flow over a grill - seems simple at first becuase it is just airflow but in reality the whole front of a car may need to be modelled to acuratley estimate the airflow through the fitted grill. How do people with no CFD experience know what is simple and what's not?

What would you bet? Your house, your car or lots of money? Anyway, you will loose.

If you go thru all the ads distributed by those commercial codes or even talk to the sales man, you may find the word like easy to use, user-friendly, etc.... It's the ad and the businessman that mislead the people and let everyone think that solving their problems on commerical softwares is as easy as playing the computer game. Who are responsible for all these craps? Hard to say.... Lots of user inputs are still needed to get a feasible solution out of the commercial code. I don't agree that the commercial code user can't make noise about how difficult to generate an accurate solution for their problem using their software. Indeed, improvements always come after the noise. It's the driving force to make the code better. MS Window has been out for more than a decade and people are still complaining about how crappy it is. In any case, it's stupid to shut the people mouth about how commercial code should work.

The true nature of the CFD is it's not that easy. Users with extensive knowledge about fluid mechanics and the physics of the problem are still required to run these commercial codes effectively. Unfortunately, not many distributors will tell their customers about this fact, fearing that it will scare their customers away. It's not like MS word where everyone can use it without knowing what the hell is going on behind the code. Thus, it's not surprising that people always complain about the code. Experience does matter in this field. Infact, none of the commercial code available on the market is perfect and nobody dare to say thier code is outstanding in every case. If someone does, he must be a liar.