CAD for FEA in the GOOD old days

(see Part I: CAD for FEA in the BAD old days)

I spent my first year in the CAE industry as an Application Engineer. An AE in most CAE software companies does a lot of pre-sales work. Prospects often want to “see your stuff work on our stuff” before making an investment. This was an exciting time for me. I was definitely drinking from the fire hose. My experience in a ton of industries and products exploded.

Change was in the air, however. FEA companies were successfully targeting multi-tasking, upfront Engineers with easier-to-use and less capable simulation tools like COSMOS and Mechanica. At the same time, the CAD companies were trying to move their products beyond the Design department and hoping to have a license on every Engineer’s desk.

It seemed natural to think that an easier to use FEA toolset for generalist engineers featuring a direct linkage to their CAD tool would take over the world. The PhD specialists could use their complex pre-processors, but the regular Joe-Sixpack engineers would prefer to work with geometry in CAD.

For a half dozen years or so, that theory worked out.

There was a rapid increase in total number of 3D CAD seats that eventually plateaued to about 1 million users worldwide. Much of that came from companies buying 3D CAD for the first time, and a bit of an expansion into the multi-tasking Engineer base. The term “Design Engineer” (as opposed to “Designer” or “Engineer”) began to show up. Some of these early adopters of upfront CAE and everyman CAD usage had success. But the trend did not take off as quickly as I had anticipated.

After an initial shakeup, job descriptions and functions settled back into older patterns through the subtle force of practicality. BSME Engineers were still spending a very small percentage of the workday using CAD. Complex tools like Pro/Engineer, CATIA, and Unigraphics required a massive amount of time to and effort to master. You definitely needed 40 hours/week of stick time to maintain proficiency. So, the vast majority of companies still had a Design department staffed with professional Designers for detail design and documentation control.

Getting these designers to run upfront CAE was largely a non-starter. They just didn’t have the stress, strain, and mechanics instruction to properly setup the FEA models or interpret the results.

I call this period “The good old days” because at least stronger integrations between CAD and CAE had been developed. More multi-tasking Engineers were engaging in CAE earlier in the product development process. Where these engineers were unlikely to learn a complicated pre-processor like their PhD counterparts, they could wrap their heads around the simpler upfront CAE tools. They often needed to work closely with a designer to simplify manufacturing ready CAD models. But if you could get a robust, simulation-ready CAD model on the screen, you could do some truly amazing what-if studies by simply adjusting a few dimensions.

Getting to that simulation-ready CAD model was no easy task, though. Most of the engineers couldn’t or wouldn’t create bare-bones CAD models from scratch. And, the designers were more inclined to clean up manufacturing-ready models instead of starting over. So, while these companies enjoyed a nice ROI from their upfront CAE investment, the bigger potential return never materialized.

Read the final installment of this 3 part series: CAD for FEA today

email_digest
  • http://www.efieldsolutions.com/ Harald Hermansson

    Hi,

    I’m eagerly waiting for part three, but dare make a comment already now:

    One of the most important objections to starting over with a simulation model from scratch instead of repairing an existing production CAD model is the risk that you do some mistake and don’t end up with exactly the same geometry in the end, even if this should be a much simpler path to follow. And even if there is no such mistake made, there will be people suggesting the models are not equivalent.

    I’ve been working with CAE in different physical domains (FEA, CFD, CEM) for more than 20 years, and knows that there is much value in a simulation ready production CAD model, and that it can have a very long life (as long as the complete product life cycle actually). As it is now, people working with transforming raw CAD models into simulation ready models in these fields very often use different tools and methodology even within the same company, and does the same work three times over. It would be interesting to read your comment on that in the next part!

    Best regards,

    Harald Hermansson

  • http://www.efieldsolutions.com Harald Hermansson

    Hi,

    I’m eagerly waiting for part three, but dare make a comment already now:

    One of the most important objections to starting over with a simulation model from scratch instead of repairing an existing production CAD model is the risk that you do some mistake and don’t end up with exactly the same geometry in the end, even if this should be a much simpler path to follow. And even if there is no such mistake made, there will be people suggesting the models are not equivalent.

    I’ve been working with CAE in different physical domains (FEA, CFD, CEM) for more than 20 years, and knows that there is much value in a simulation ready production CAD model, and that it can have a very long life (as long as the complete product life cycle actually). As it is now, people working with transforming raw CAD models into simulation ready models in these fields very often use different tools and methodology even within the same company, and does the same work three times over. It would be interesting to read your comment on that in the next part!

    Best regards,

    Harald Hermansson

  • http://lifeupfront.com/ Jeff Waters

    Hello Harald,
    I’ll have to consider your comments as I write the next piece! There is a spectrum of CAE usage from “early, rough trend” to “dead-on, .1% accuarcy against test data.” I tend to personally lean towards the “early, rough trend” end of that scale in terms of overall value.

    In those scenarios, it’s less likely that subtle differences between a simulation-ready and mfg-ready CAD model will matter.

    But when you get more to the other end of the scale where you are more using CAE as a validation tool, you are absolutely correct.

    Coincidentally, the modern solution for handling CAD in CAE will appeal to engineers at both ends of that spectrum.

  • http://lifeupfront.com Jeff Waters

    Hello Harald,
    I’ll have to consider your comments as I write the next piece! There is a spectrum of CAE usage from “early, rough trend” to “dead-on, .1% accuarcy against test data.” I tend to personally lean towards the “early, rough trend” end of that scale in terms of overall value.

    In those scenarios, it’s less likely that subtle differences between a simulation-ready and mfg-ready CAD model will matter.

    But when you get more to the other end of the scale where you are more using CAE as a validation tool, you are absolutely correct.

    Coincidentally, the modern solution for handling CAD in CAE will appeal to engineers at both ends of that spectrum.