"Who would even think of building a scale model for routine stress calculations these days?"

In the current Mechanical Engineering, there's a report on testing of a scale model of the pressure vessel that was damaged in the infamous Three Mile Island nuclear reactor accident. The tests were/are being done at Sandia National Labs, a US Department of Energy facility, but the interest and support are international.

I guess the key is that the baby was and is not routine!

John and Ken,

I suppose that I should clarify about the tools that I use. I work for CFD Research Corporation, and I use our commercial CFD-ACE+ code. On occation I do cheat for some of the more complex jobs by using new modules and features that are not yet available in the released version.

A couple more points before I shut up:

On the issue of commercial vs. in-house developed codes: I used to work for Electric Boat doing CFD work related to submarines using an in-house developed code. That code was (and still is) developed by a group of about 4 people. I have some idea of the improvements that have been made to that code in the 5 years since I left EB. When I compare that to the improvements that have been made to CFD-ACE+ in just the past 2 years, the differences are amazing. There is just no way that a handful of researchers/developers/programmers can keep up with the progress that can be made by more than 50 researchers/developers/programmers.

I'm going to go WAY out on a limb and give you my opinion on the best type of user of CFD in a "production" environment. What I'm referring to is a situation in which there has to be a product at the end of the process, not a code development effort, or a pure research environment. I would rather have a person who knows alot about the product and a little about CFD than the other way around. I think that I can teach someone CFD, that is rules for grid generation, and how the process works well enough to be proficient. More important, however, is that a user who know the product can recognize a bad result. Then that person can examine the process for errors, or get help from a CFD expert. A user that doesn't know the product can only judge the quality of the answer by looking at numerical data like convergence. I've sometimes had really good convergence to really bad solutions, and that doesn't help anyone.

Finally, John like chase perfect answers, but there are two things to recognize. 1. An answer doesn't have to be perfect to be useful. There is some range of accuracy that allows a design engineer to gain insight. To some extent, other disciplines have recognized this and built it into their design process. For example, the allowable level of stress in a pipe is typically 30-50% of the yield stress. This is in part because the ASME recognizes that there are errors inherent in the modelling process, and that in the field, shit happens, and there has to be margin.

2. The analysist also has to recognize that it is often not practical to get a perfect answer. I had to use a turbulence model, but there are no turbulence models that were developed specifically for vacuum relief valves. I can choose bewteen a model developed for a flat plate, and one that is supposed to be better for jets. The choice will have an effect on the answer.

Alton

Hello,

In my opinion it does not matter which code you use: a self-made code or a commercial code. It strictly depends on the engineer/scientist who uses the code and his/her knowledge of the fluid physics. Then, that engineer will know how to use this code (or not) to design the product.

I used a commercial CFD code to design a product. The CFD solution matched the experimental data. I was able to show with the CFD process that without drastic changes in the design the current product will not work. Well, I was laid-off and the company is going out of business.

So please do not say bad things about commercial codes, but rather work with commercial codes companies to improve and change their product to fit your requirements. Also, rely on your skills and team work to make sense out of CFD results.

I do not work for a commercial code company and do not have PhD.

Thanks, Tom

Again Alton, I can well see from where your coming on this important but not cfd life threatening issue. I think several links in recent times are centralising on fundamentals for academia sake and maybe its here that progress does not just spontaneously happen. In my book and experiences the gulf between research pure or applied and the just presented model[product] time lags. Your ability,training and knowledge to harness all that you could meant the difference between success or failure in the given time limits.

Your new or evolved CFD for the relief valve body design gave the real answers. Hope you got the credit you deserved.

Though what did bother me a little was the fact you said tried and tested flow design equations from old methods used by the manufacturer were not precise enough.Does that mean they were bad manufacturers[bad standards etc] or have more precise measuring instruments come along to question flow rate accuracy and possibly inline equipment.

Tom Very sorry that you have just lost your job but but either the company had fallen prey to lack of vision in being too late to incorporate CFD in whatever code form get the product right to win the market place or to at least compete or economics [stock market forces] have forced unwanted closures.

Don`t lose heart and keep pitching for another opportunity.

You miss the point.

FEM tends to be used upfront, with mechanical testing used to understand failures. With more immature (less trusted) technologies like CFD, it's usually the other way round - CFD is used to investigate odd behaviours spotted in physical tests.

My point is that CFD codes are generally run by CFD people, who need to understand how the codes work to use them properly (at all even?). This is not the case with more mature codes (maybe "spreadsheet" would have been a better "mature code" example - I haven't bothered writing one of those since about '87).

(1). I think, it is because the technology used is not validated. (2). The same is true for the instrument used in testing, if it is not calibrated, then the result is not good. (3). Wright brothers used the validated bicycle technology when designing the first airplane. He didn't use any FEM at all. (4). The same is true for the Ford model-T. I don't think Ford used FEM routinely to design model-T. They depended on the technology which was validated (reliable) in their time. (5). CFD can be very accurate and reliable when it is validated. (for that particular application)

The answer to your first question is yes.

the answer to the second is yes but I couldn't say that they'd necessarily work very well. Reviews I've gotten are mixed. Most commercial parallel codes prefer SMP machines I think.