ResearchResearch Group: Robotics & Autonomous SystemsResearch Projects
Online-Identifikation regelungstechnisch relevanter Systemparameter elektrisch angetriebener mechanischer Systeme

On-Line identification of control relevant system parameters electrically driven mechanical systems

Team:  M. Sc. Dipl.-Ing. (FH) Daniel Beckmann
Year:  2011
Funding:  Forschungsvereinigung Antriebstechnik (FVA)
Is Finished:  yes

Systemparameters which significantly influence electrical drive systems are generally only imprecisely known. During the initial startup or after swapping of components of the system, the controller parameters are often complicated by heuristic Setting criteria optimized by hand to achieve the desired system behavior.

Furthermore, system states and parameters of the control system can change during operation by leaps and bounds (eg change of loading condition), or change slowly over time (eg due to thermal influences, wear). To capture this change, a real-time model of the mechatronic system is necessary.

With the help of the model the excitation of the parameters to be estimated were computed at each step on the fly. This information is used to calculate the relevant parameters only when there is a sufficiently good suggestion. This ensures that a robust identification is made of the model parameters. The resulting updated models are the basis for an optimized tracking of the controller setting.

The observability of the parameters and states should be based on a test bench. It is a belt drive, which can be described according to the degree of model as a linear (or nonlinear) two or three mass oscillator. The belt drive can be further extended by static nonlinearities.