Does Multi-Physics Make Fools of Us All?
March 12th 2013
Quite a long time ago, I remember the first time I visited a hard-core simulation analyst's group. It was at NASA Marshall Space Flight Center and I, fresh out of college, was showing them some new simulation technology. There were lots of questions and conversations that day. But first and foremost in my memory, I remember the… well… reverence some of the other engineers had for these simulation analysts. They were the unequivocal experts. Their authority was unquestionable. They were on a pedestal.
Fast forward to today, and I believe that the standing of simulation analysts is changing in the engineering community. No, there's been no terrible fall from grace. The change has been subtle, but noticeable. Let me explain.
Multi-Physics: The Good and the Bad
Have you ever heard of the term success-disaster?One of my friends startled me with it one day. Basically, it's when something works so good, that the expectations of others are significantly increased. And that translates into a lot more work. It's good overall. But bad if you can't meet expectations.
Today's simulation software tools have become ridiculously powerful when it comes to multi-physics. You can perform simulations that couple disparate physical domains including structural, fluids, electro-magnetics and more. Furthermore, simulations are expanding into the domain of system engineering and modeling. The capability to accurately predict product behaviours is there, much more so than there has ever been in the past.
Executives and managers with some awareness of multi-physics have had their expectations raised. They see new means to eliminate more rounds of testing, cut out more physical prototypes and reduce the volume of change orders. And all that is great. However, expectations do need to be tempered. Because all of that can't happen with the wave of a magic wand.
Issues Holding Multi-Physics Back
So, why the heck can't multi-physics cure all the ills of an engineering organization? Well, it comes down to two issues: expertise and time.Most of the simulation analysts I know have invested a significant amount of time both in their post graduate education as well as numerous years at work in becoming an expert. However, most simulation analysts are experts in a specific, or a few related specific physical domains. You find great structural guys. You find awesome fluids guys. But you rarely find simulation analysts that are experts in all of those fields. And frankly, any expectations of that are completely unreasonable. It takes years for someone to be solid in a single field.
Why is that an issue for multi-physics? Well, with multi-physic simulations, you have to know how to avoid the pitfalls in all of the applicable physics domains, not just one. You have to instrument the model(s) accordingly.
And that brings us to the second issue: time. Even if you had simulation experts with expertise across all applicable physical domains, these multi-physics simulations require a significant amount of time to setup. Then they require verification and validation once they run. Multi-physics currently is not, and may never be, a high volume thing.
New Technology Enablers?
Now, let me pull back a bit. I'm not trying to rain on anyone's parade. This situation isn't in fact hopeless. In my mind, however, I believe some additional capabilities are required to enable multi-physics. And I'm not tailing about new simulation capabilities.One area that needs to be enabled is collaboration. Yes, I know. Forgive me for using such a generic term. But basically I mean various simulation analysts need to be able to share model preparation sessions and communicate, both in real time as well as asynchronously. Why? Because different simulation analysts with expertise in different physical domains need to work together on multi-physics models. That's one approach.
An alternative, or supplemental, approach is knowledge management. Simulation analysts lacking expertise in a certain domain could reference a knowledge base to enhance their ability to prepare multi-physics models. In fact, strange as this may sound, this can be incorporated into HR skills and training improvements.
Again, I don't believe multi-physics is doomed. It holds great promise for the organization. However, I believe to truly enable it, you need some capabilities that address the reality that every simulation analysts will have some holes in their expertise across physical domains. I suggested a couple ways. But I'm sure that smarter folk than I can think of additional and better ones.
So, those are my thoughts. What are yours? What has you experience been with respect to multi-physics? How did you and your organization address the need for expertise across physical domains? Did you employ any ad-hoc collaboration or knowledge systems?
Sound off. Would love to hear what you have to say.
Take care. Talk soon. Thanks for reading.