 |
|
|
|
 |
 |
|||
 |
|
|||
 |
|
September 2008 |
||
|
|
||||
| Dr John Rasmussen, chief technology officer AnyBody is used by two major European OEMs for research only. The main uses are for discomfort modeling in terms of muscular fatigue and joint loads. The models are for ingress/egress, steering wheel operation, gear shifts, hand brakes, opening and closing doors and decks and that kind of thing. The characteristic feature of these situations is that the mechanical system becomes a combination of the human body and some artifact, for instance the steering wheel, in what we call closed kinematic chains. In this situation, what you do with one arm will affect the working situation of the other arm, so the mechanical situation is usually quite complex. You really cannot do this properly without software modeling. The issue as with all new technology is that it requires expertise to use it properly. If you compare with finite element analysis, you probably know that all mechanical and structural engineers today have comprehensive classes in understanding that technology in college. Biomechanics is equally complex but there is not the same educational backup, so getting the organization tuned to take advantage of the technology is more difficult.
On the overall level it is always a question of speeding development and saving cost. Musculoskeletal simulation can reduce the need for costly experiments with humans in mockups of car environments. A good example could be finding the combination of necessary seat adjustments and spring stiffness in the accelerator pedal; two parameters that may not be perceived as related by designers because the seat is quite different from the accelerator, but which are connected by the driver’s body. So a large male and a small female will have different needs in terms of force feedback from the accelerator because their legs have different weights and they will also need different seat adjustments. But the point is here that the two parameters influence each other. So a correctly selected accelerator spring may enable two people of very different sizes to be comfortable in the same car with a limited possible adjustment of the seat. Design for humans is full of that type of unexpected causalities and there really is not way to investigate them properly expect by proper modeling of the biomechanics of the situation. We are faced with the challenge of giving access to a difficult technology. So much of our efforts are devoted to improving the user interface and make it easier and quicker to get models up and running. We are also considering entirely different ways to wrap the technology into software. Our focus so far has been on generality; you can literally model anything with the AnyBody Modeling System, even a bird or a lizard if you want. We are thinking that sacrificing some of the generality could lead to simpler user interfaces. Experience and basic education in computational biomechanics are definitely lacking. It is not just because it would make it easier for users to harness our software, but I also think it would be much easier for a typical engineer to select a reasonable simplification of the problem. If a mechanical engineer is interested in the strength of a hinge in the door he does not go ahead and make a model of the entire car body. He immediately focuses his attention in the hinge and its environment. It’s the same with biomechanical models of, say, egress. The movement is very complex and involved the entire body, but if you think a little about it there are probably certain critical points in the movement that you could focus your attention on rather than trying to model everything in full detail. Biomechanics is an immature technology and we have not yet figured out how to do everything. Our full body model has a very high level of detail with almost 1000 individual muscles, so processor speed becomes an option if a user wants to try many different parameters and options. Robustness is also an issue. Biomechanical models are very nonlinear, so if they work for one situation, this is not a guarantee that they will work for the next posture or loading situation. We need to constantly work on robustness of the algorithms. We need to get biomechanics into the existing software platforms and preferable into the PLM software. I am sure that biomechanics will mature much like we saw finite element analysis mature through the nineties. The basic technology will take different forms addressing different groups of users. I believe that this technology will be as important in the future as FEM or CFD analysis and the major players in the CAE industry will be looking to integrate it into their product suites, likely by mergers when the technology and the market are mature.
|
LINKS Ansys: multi-physics analysis pays off. Read more... Integrated: Electromagnetic CAE tools that combine FEA and simulation. Read more... Lotus: vehicle dynamics made quick and easy. Read more...
|
|||
