Modellgestützte Fehlerdiagnose technischer Prozesse - Prinzipielles Vorgehen und Beispiele

TL;DRAbstract

For the early detection of faults in technical processes internal quantities are calculated on the basis of measured input and output signals and their causal dependencies. Parameter estimation or state estimation methods and parity models are used to determine these internal quantities which are considered as features of the process. Changes of these features are then analytical symptoms of faults. In addition heuristic symptoms are obtained by human operators. An inference mechanism determines possible faults on the basis of fault-symptom causalities and approximate reasoning methods by probabilistic or fuzzy-logic methods. Experimental investigations with electromechanical drives of a robot and a wheel suspension system show some applications.

Chat with Paper

AI Agents for this Paper