Good day! I am Zachary Origenes, a mechanical engineering student in the Philippines and I am currently doing my thesis which involves validating my simulation model of an in-pipe inspection robot made in MSC Adams with the physical model of a previous study.
I have searched on what range the error can be acceptable when validating simulation results with the physical results of multibody systems, but I've only found that it's 5-20% only when it comes to FEA (Finite Element Analysis).
Hello
After deriving the equations with Lagrange or Newton-Euler, set the control inputs equal to zero and simulate the system. After the simulation, check for the total energy and other constants of motion (if applicable). For example, if there is no friction in the system the energy should be conservative. Of course due to numerical inaccuracies and truncations the error is not exactly zero. My many years experience in multibody dynamic modelling says that even with an adaptive ode solver and a fine discretization the error won't be better than say 1e-14 percent, approximately.
For complex systems the error even increases to 1e-10 or more. Anyways, I don't think 5-20 percent error is acceptable in your case.
Regards,
- Badges
- Science topic