As we go through our training as Mechanical and Structural Engineers we participate in a rigorous and incremental program which gradually exposes us to disciplines such as statics, strength of materials, dynamics, thermodynamics, fluid mechanics and progresses into more advanced applications of these disciplines. As we progress we each are drawn to certain of the disciplines and tend to focus our work in those areas. Frequently we learn CAE techniques and methods for dealing with practical problems in our specific areas of interest.

In my personal case the area of interest was dynamics, and my earliest application of numerical methods in dynamics was to helicopter rotors.

I quickly learned that in such an application the analyst had to include many different disciplines into the dynamics model to include the physics of aerodynamics, the flexibility of the helicopter blade and the required control systems.

The focus area of my work was in the area of multi-body dynamics – this is what paid my stipend – the other disciplines were required as inputs into my dynamics model. In retrospect, I realize that incorporating aerodynamics, flexibility and control systems was a secondary objective to my primary objective of building my multi-body dynamics model. As my career progressed, I found that this is very common. I found that a dynamics analysis of a naval structure required ocean wave motion as input. I found that loads generated from my analysis of a chain saw were required for fatigue analyses. Modern Automotive CAE frequently involves interaction between several disciplines such as vehicle dynamics, noise vibration and harshness, and durability.

It seems that CAE analyses frequently (perhaps always) require data and other inputs from disciplines that are not necessarily within the core subject matter expertise of a particular analyst.

To meet the needs of these specialized analysts, the CAE industry has developed a very broad and comprehensive set of software applications that focus on their particular areas of specialization. For each discipline the analyst generally has multiple software solutions available – tailored to her needs and developed with methodology that she learned in University. Each analyst has strong preferences within his or her area of specialization. Each of these tools has varying degrees interoperability with other areas specialization. For example, in the case of multi-body dynamics we have fantastic solvers available for the equations of motion, and we can input pressure loads for aerodynamics, we can input modal representations for flexible bodies, and we can export loads from these models for use in fatigue analyses.

In each of these steps the analyst makes simplifying assumptions which have historically been required for a practical solution. When I import pressure loads into my dynamics model I implicitly assume that the dynamics model does not influence the aerodynamics loads. When I use a modal representation of a flexible body I implicitly assume that the body remains linearly elastic. When I use a spline or table lookup method to approximate the loads generated by an elastomeric component (such as a bushing) in a dynamics model I make a set of simplifying assumptions about that component.

These simplifying assumptions introduce errors into the simulation. Relying on other CAE tasks to be performed before a particular analysis can be initiated creates bottlenecks and opportunities for confusion.

Today we collectively have the opportunity to step beyond traditional boundaries and CAE software applications, and look toward solutions that integrate the different solution disciplines. By this we mean bringing together solver technologies and solving the governing equations simultaneously.

Consider the example of modeling an elastomeric component in a typical multi-body application. Instead of accounting for the elastomeric component on the right hand side of the equations of motion as a force element (and making simplifying assumptions in the process) imagine representing this component as a finite element model and solving it simultaneously with the muti-body equations of motion. Clearly, this may consume more CPU time, but it may also increase the accuracy of the results by accounting for previously neglected dynamic effects. This is a vision that I had in graduate school while working on my helicopter rotor problem. A large number of years later, this exciting technology is finally becoming possible. The results will be more accurate simulations and better airplanes, cars, tractors, chainsaws and naval structures.

Patrick J. O’Heron

Manager Automotive Center of Competence