Sign up for our newsletter

Fields marked with a * are required

COMSOL: A Different Multi-Physics Approach

BellowsMulti-physics simulation is a real quandry, isn’t it?

I find that almost everyone agrees that multi-physics simulations are more accurate. They more accurately predict the performance of products in the real world because, as we all know, reality is fundamentally multi-physics.

Knowing that multi-physics yields more useful results and actually conducting multi-physics simulations are two different things, however. Back in March, over at the NAFEMS blog, I published a post titled Does Multi-Physics Make Fools of Us All? on exactly this issue. Many  analysts specialize in one simulation field. Few master them all.

That’s one of the reasons I approached the folks at COMSOL about a briefing. I knew they boasted some unusual capabilities with respect to multi-physics. What I found definitely piqued my interest. In this post, I detail some of the critical capabilities of COMSOL’s software products and add in my own commentary and analysis.

Capabilities Provided by COMSOL Products

Let’s jump in. First off, let’s discuss the products that COMSOL offers.

Just to be clear, there is one software product with a lot of different modules. There are six sets of modules: four for physical domains, one mixed bag for optimization, materials and whatnot and the final set focuses on interfacing with other software products. Furthermore, there is one pre-processor, one solver, which we’ll talk about in a moment, and one post-processor. This actually surprised me a bit. Today, most software companies that have been around for any amount of time have products scattered hither and yon. COMSOL offers a single, nice, tidy software set.

Furthermore, COMSOL’s offering isn’t just broad, but it is also extensible because of how the software product is architected. Here’s how that works.

Now, let’s be clear: only certain types of elements can support specific types of physics. COMSOL doesn’t offer some magic sauce to circumvent that issue. However, right when you start setting up your simulation model, you are prompted to select which physics domains you want to include. This includes custom equations as noted above.

After the simulation is setup, the solver applies the equations, regardless of physics, to the model in the same way. It is a generic discretization approach that works for any kind of physics or set of equations. In COMSOL’s eyes, the common denominator are the equations.

A fallout of this approach is the consolidation of pre-processing as well. With any kind of physics genericized into equations, the setup actually becomes more generalized as well.

Of course, pre-processing wouldn’t be complete without leveraging existing 3D models and other  tools.

Now, the modules in this set don’t just import stuff from other software tools, they can associatively link to it. This works for Matlab, Excel, Solidworks, Inventor, Creo, Solid Edge, Spaceclaim, AutoCAD and CATIA.

Commentary and Analysis

Now, let’s step back. What does this all mean?

Advantages for the Simulation Analyst

One advantage provided through COMSOL is the ease of setting up a multi-physics simulation. With one pre-processor, solver and post-processor, simulation analysts don’t have to use a wide range of tools to get one job done. They don’t have to worry about mapping results of one simulation onto another simulation. It truly is a consolidated environment.

Furthermore, live-linking to such a range of other software tools offers a significant advantage as well. Obviously, this allows the simulation analyst to keep up to date on design changes, a significant challenge given the fast pace of today’s design cycles.

Areas for Improvement

Back in that NAFEMS blog post on multi-physics, I called for one specific area of improvement.

In my mind, however, I believe some additional capabilities are required to enable multi-physics. And I’m not talking 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.

Of course, any simulation toolset that is focused on multi-physics needs to have this kind of capability. COMSOL is not alone in that regard. But it is a valid point to include nonetheless.

Recap and Questions

  • COMSOL offers a pre-processor, solver and post-processor in a single integrated simulation software product.
  • COMSOL offers modules that fall into six major sets. Four sets focus on physics domains including: electrical, mechanical, fluids and chemical. Another module includes optimization and materials. The sixth focuses on import and associative connections with other software tools.
  • All physics domains as well as custom equations are applied to the same simulation model, where they are discretized into base equations and solved in one calculation engine.
  • This makes it easier for simulation analysts to setup multi-physics simulations. COMSOL could offer more collaboration capabilities, as could almost every other multi-physics simulation software, as a means for experts in different physics domains to collaborate together on one simulation model.

That’s my take folks. Want to weigh in? Is it advantageous to use a single set of tools to setup multi-physics simulations? Is an associative connection with CAD or other software tools important? Sound off and let us know what you think.

Take care. Talk soon. And thanks for reading.

Chad Jackson is an Industry Analyst at Lifecycle Insights and publisher of the engineering-matters blog. With more than 15 years of industry experience, Chad covers career, managerial and technology topics in engineering. For more details, visit his profile.

Like this post?

Sign up now to get more like it

Fields marked with a * are required
  • Ivar K

    So you have discovered this marvellous tool: finally!
    Multi-physics is here today.

    What I did not read clearly, above, is the difference of the direct solved COMSOL approach, and the fully segregated approach of many other FEM vendor tools, for different physics coupling. This is an essential plus too, for COMSOL.
    I have been using it for several years now, and I cannot think of going back to the old FEM tools.

    For the perfectionist, it is slightly weak on some internal post-processing capabilities for engineers, specifically per physics, but I’m sure that will come soon too, at the speed they adapt the software.
    Today, one must use Matlab or Java API to get the ultimate post-processing analysis done, if you are an advanced “classical” FEM specialist. But, by far, no other FEM tool l know about can propose so many physics in such a simple and identical methodological approach, and all fully coupled, or segregated, at your will :)

  • Dennis Nagy

    @Chad: excellent blog on COMSOL! One area you didn’t venture into (intentionally?) is the answer to this partly-rhetorical question: If COMSOL is really that great (technology and software architecture), and I am not necessarily saying it isn’t, why hasn’t it already blown away ALL the competition? You should blog on CAE (engineering simulation) market dynamics and the nature of new-technology/product uptake by engineers and engineering management (i.e., their culture). If I were a (more active) blogger, I’d do that, but if you do it, I’ll definitely chime in with my views.