Konferensartikel

important role since many years. Nowadays; attention is also paid to maintenance times. Maintenance standstills are to be reduced as far as possible. On the other hand; technical systems are subject to signs of wear and tear which are in general growing slowly and imperceptibly. The gradual abrasion of applied tools may lead to poor production tolerances or to a component’s standstill. Hence; a condition-based maintenance strategy will be of increasing importance. Such a strategy requires a permanent condition monitoring during operation. To this end; reliable high-performance algorithms for signal processing; feature extraction; and classification are needed. Modeling the process of wear and tear may be useful to find the particular steps of the condition monitoring system’s signal processing. This strategy was investigated by means of one very important device from automation engineering; a hydraulic axial piston pump. The procedure of getting signals by an appropriate Modelica model of the main parts of the pump is shown within the paper. Additionally; the manipulation process for the signals and the steps of classification are shortly presented to give an overview to the possibilities of model-based signal generation based on a Modelica model. The advantages of the multi-physics modeling language are emphasized because the axial piston pump model combines the mechanical and the hydraulic domain in a very efficient way.

Condition monitoring; classification; signal processing