Two weeks ago I was invited to speak at Infolytica’s Engage 2016 conference (infolyticaengage.com). The company makes tools for finite element analysis (FEA). During this conference I listened to many interesting presentations, most notably one that showed where design tools are headed by crossing platforms and application areas.
For a long time we have combined magnetic analysis with thermal analysis. When you create your motor model you have almost everything that is required to solve the thermal modeling: sources of heat, materials with heat transfer characteristics, etc., that allow you not only to project the torque of a motor, but to also determine how hot it will get during operation — very important design criteria for many automotive and military projects. One important characteristic here is that such an analysis can now be performed as a co-simulation where the magnetic and thermal analysis are performed both concurrently and interrelatedly, which allows us to glean very accurate insights into the thermals during transient and peak conditions. This is a considerable leap from the past, when we had to transfer steady-state values from the magnetic analysis into the thermal model, and only then could we solve for the steady-state case.
Similar tools exist to calculate materials stresses and potentially to project mechanical problems due to magnetic stresses, although to my knowledge the commercially available interfaces are not quite as advanced, as they do not always allow for co-simulation. However, many software packages now come with powerful interface options such as Active-X scripting that allows even those users with only a very basic understanding of programming and scripting to design powerful interfaces that can be used to enable full co-simulation.
Now back to the presentation that interested me the most — a third-party module that was able to predict audible noise by using the data from the magnetic analysis to calculate the mechanical deformations of the motor body and then project the resulting audible noise signature where the users can place sensors (virtual microphones) in locations of interest. This particular software was developed to project the noise signature of electric motors for the automotive industry, but it can also be used to project motor noise in other applications such as military. What is interesting is that in the past, work had been done to project audible motor noise but everything was basic research; and now we have simple-to-use software that knowledgeable designers can use to model another dimension of their design matrix.
Why does that excite me so much? This software and others hopefully to come are another puzzle piece in the Holy Grail of virtual prototyping where we can design a system and see, feel, listen and test it before we ever build it and, most importantly, we will know whether it works as intended.
Many engineers will tell me that it is too costly or too time consuming to go through all this modeling and that it is faster and less expensive to “just build it.” I do not buy this argument. True, if you are only doing one design for a very low-volume application, you might make this case. After all, the required training to use modern simulation tools is still a factor, even though the interfaces are advancing rapidly to simplify these design tools.
But if you are looking to build very high volumes then the “just build it” argument is definitely not valid — the savings and the benefits from all the added knowledge are just too valuable to ignore. The cost savings from an optimized analysis far outweigh the associated costs.
If you only have a few design projects then you may wish to work with a consultant or a company that offers custom design services to perform this analysis for you. Many of the software vendors also offer design services for a fee, but I think that a custom design company or consultant is preferable since they often design and test many actual motors and they can often offer additional, helpful insights and advice. Some of the software vendors are starting to offer services where you can design and run their software remotely, and rather than purchasing or leasing the software tools you only pay as you use them, which can be a more cost effective alternative if you only do a few designs.
Over the years I have seen many cases where designers and inventors spent a large amount of money to build a motor or a motor invention, only to find out that it did not perform as expected and much money could have been saved if the designers had used simulations and “virtual prototyping” before attempting to actually build the design. It is hard enough to have your dreams a of a great invention shattered, since it will not work as expected, but do you also have to lose your home?
Our industry is changing and while we have to overcome a large amount of inertia, modeling and virtual prototyping are becoming increasingly commonplace; they change how our industry works and we must change our ways also and be open to these new possibilities that now exist.
As I have stated many times, our industry is becoming a very high-tech environment and it is exciting to work in it and to take advantage of all the new tools that are becoming available. The aircraft and defense industries have done it for many years; you design a new aircraft or system in a completely modeled design environment before you ever build it, and when you do the aircraft will fly reliably and (mostly) as predicted.
Now let us put on our virtual reality glasses and examine our latest design: are there any vibrations? Do we get enough torque? How precise will the robot arm move? The possibilities are endless. More importantly, if something goes wrong no major damage is done and nobody is injured.
Now if we could get some software folks to design operating systems the same way — they might actually work. And they might work reliably and predictably.