KISS upfront CFD

K-eep
I-t
S-imple
S-tupid

I seem to be having the same, repeated conversation with engineering leaders these days. They are often only seeing a modest return on their upfront CFD investment. Maybe they’ve cut the average # of prototypes from 10 to 8. Maybe the average design cycle has only dropped from 12 weeks to 10 weeks. In my experience, they should be seeing 2-4 times that improvement on both fronts. Why aren’t they?

There are a few typical culprits, but by far the biggest problem lies in the difficult path their engineers have chosen to climb the upfront CFD mountain. The engineers are starting with production ready models every time. The production CAD models contain thousands of parts when only 3 major, simple components have any real impact on the final solution. The engineers are spending time suppressing parts and fixing the mess as those suppressions cause delicate assemblies to disintegrate.

This is like attacking the shear face of the mountain with a couple ropes and cramp-ons when there is shiny new helicopter idling nearby.

Engineers are predispoded to running simulations on fully detailed models for some reason. I don’t completely understand this character quirk, but I’ve seen it thousands of times. It’s true. These guys will bang their heads against geometry cleanup for hours and days when they could easily create a fresh, simplified simulation model (including all 3 major components) in less than an hour.

If that simplified model is built with a few easy-to-edit features, the engineers could spend the next hours or days producing 10x-20x the number of simulations. They’ll spend their time producing rather than prepping. It’s your job to enforce KISS. The engineers are often unlikely to do it on their own. Point at the screen and ask, “Why do you have 3,000 rivets in your simulation model?”

Sometimes you gotta push your team (kicking and screaming) into the helicopter.

Advertisements

4 comments

  1. Jeff,
    Wouldn’t you say it’s a cultural thing? Technology is exciting, until it forces a change in habits or perspective. As an engineering manager, I have in the past asked my product engineers to use simulation to guide design, not just look for pre-test validation.
    I also have worked with a few software companies that could change the idea that CAD, not some other software drives design. Afterall who cares what the CAD entity is called as long as you can create the shape you want. Should we be a slave to CAD features as named and defined in a particular CAD system. There may soon be a better way.

  2. Jeff,
    Wouldn’t you say it’s a cultural thing? Technology is exciting, until it forces a change in habits or perspective. As an engineering manager, I have in the past asked my product engineers to use simulation to guide design, not just look for pre-test validation.
    I also have worked with a few software companies that could change the idea that CAD, not some other software drives design. Afterall who cares what the CAD entity is called as long as you can create the shape you want. Should we be a slave to CAD features as named and defined in a particular CAD system. There may soon be a better way.

  3. Jeff Waters · ·

    Yeah, I think it is a cultural thing… probably a human thing, in fact. The more I look into complex technologies, the more I realize that they are often like adding salt on the existing mindset of the user. If the user is predisposed to “doing it the hard way”, then the dish will taste bad. If they will adapt their underlying process and thinking a bit, the salt will make an exquisite dish out of the same ingredients.

    Using CAE for pre-test validation is usually worthless. But, that’s what many engineering teams are used to with standard development processes based on lab/prototype verification. Using CAE to guide design is where the money is…

  4. Jeff Waters · ·

    Yeah, I think it is a cultural thing… probably a human thing, in fact. The more I look into complex technologies, the more I realize that they are often like adding salt on the existing mindset of the user. If the user is predisposed to “doing it the hard way”, then the dish will taste bad. If they will adapt their underlying process and thinking a bit, the salt will make an exquisite dish out of the same ingredients.

    Using CAE for pre-test validation is usually worthless. But, that’s what many engineering teams are used to with standard development processes based on lab/prototype verification. Using CAE to guide design is where the money is…