Michael, I agree that if you just toss CFD software (no matter how easy it is) at a designer and walk away, you are asking for trouble. But if you take the time to do some training and pair them with a more experienced analyst, they should be able to turn out acceptable work. If, in your example, the experienced analyst sets the proper parameters (ie how far in front and in back of the grill to model, etc) initially, then why shouldn't the designer be able to run parametric studies with slightly different grills and different speed conditions? The analyst should be helping the next designer rather than just turning the crank on 20 different variations of the same analysis. Its a better use of his time.

How about something even simpler - trying to optimize pressure loss in the ducts that supply HVAC air to the passenger compartment? You don't even have to get the "right" answer. All you need to know is if one design is relatively better or worse than another. I don't think you need a PhD and 10 years of experience to solve flow in square pipe. The question that needs to be answered is, "Is the management who throws this new task on their designers going to give them the support they need, or do they just see this as a way to replace high priced engineers with cheaper labor?" Some companies will get it right and others will turn out garbage and then blame commercial CFD codes for their failures.

Please tell me which planet you live on where advertisements and sales people are 100% truthful about anything? Are you that naive? What you do expect them to write in adds: "Our software is so complicated that unless you send us your resume and 5 references about your experience we will not sell it to you."

None of the commercial vendors ever said their code is 100% perfect and solves every possible problem. Obviously there is still much research to be done in the field and different codes have different strengths and weaknesses. There is a big difference between saying that one code has a problem in a certain area and whining that everything is crap and nothing works and all salesmen are liars. In reality, you can get plenty of good data out and things are much easier than they were 15 years ago when I started.

Oh gee... I didn't realise that commercial CFD code never failed to solve a problem once. It's the stupid users who are responsible for all these failures. Thanks for pointing out the fact for me.

You are welcome. Unfortunately I never said that. In fact if you read all the posts I said more than once that its not perfect. There is a big difference between not perfect and totally useless. Its amazing how many people here can't seem to grasp that concept.