Is there such thing as a CFD "User"?
 

Hi all,

I just had a philosophical debate with a fellow grad student about this subject. I believe that a person cannot just "pick up" some CFD software and use it. Even if they generate a result that mimics an experimental result I would not trust the computation. I was just wondering what are your thoughts on this.

Thanks for your reply,

ryzd

I don't think that CFD is any different from the other tools that are used in industry. Anyone can use the tool if properly trained. There are many things that can cause an erroneous CFD solution, but that can be said for a structural tool like ANSYS.

Bottom line is that using a complex tool requires good understanding of not only the tool, but the underlying assumptions. For example: In structures, you need to be able to draw a free body diagram for complex loading scenarios. In CFD you need to understand the implications of the chosen grid and turbulence model.

I think that to learn, someone must use the tool. In college, you learn theory and then the tool. In the workforce you typically learn the tool with guidance from an expert, then you gradually learn the theory. Just a difference in approach.

I agree with Scott, but there's even more. Are you asking if someone can pick up a CFD tool and predict water flow in a pipe or are you asking if they can pick up a tool and solve a problem with multiple phases, complex chemistry, heat transfer, liquid films, droplets, moving boundaries with fluid-structure interaction added for good measure? I would hope almost anyone with a absolute minimum of training could do the first. Obviously they aren't going to do the second and somewhere in between is a line between what's reasonable and what isn't for someone with not a lot of background.

I would also add, that a good measure of suspicion is crucial for judging the results in any field, and specifically in CFD. Using a CFD tool without the needed knowledge for the specific problem at hand (the physics) and of the specific tool (numerics and the underlying models and assumptions) easily leads to a GIGO (Garbage In Garbage Out) situation, which occurs too often...

Let's demystify the question somewhat by replacing "CFD" with "paintbrush". Is there such a thing as a paintbrush user? Yes, certainly there is. Lot's of people are competent painters who can go down to the Home Depot, buy painting supplies, and do a great job on their house or a living room or a kid's project.

Is this same person going to be able to paint the next masterpiece that hangs in your favorite art museum? Probably not.

CFD is a tool like any other tool. In the hands of a competent, trained user it can produce good engineering results. You wouldn't give a toddler a brush and a bucket of paint and set them loose in your house. Nor should you give an untrained engineer a CFD code and pray for good results.

Now, I realize there are caveats so let's not argue the analogy. There are child prodigies who can outpaint most of us. And there are engineers who can pickup a CFD code and get good results right out of the box.

But when you turn to the high-end, one-off types of CFD problems, your CFD user probably isn't the one you want running the show. You want a CFD expert, a dedicated analyst who knows how to make the CFD code do what's needed and who can interpret the results.