D8.1 - Detecting and Compensating Sensor Faults in a Hydraulic Condition Monitoring System
- AMA Conferences 2015
2015-05-19 - 2015-05-21
- Proceedings SENSOR 2015
- D8 - Signals & Systems
- N. Helwig, E. Pignanelli - ZeMA – Centre for Mechatronics and Automation gGmbH, Saarbrücken (Germany), A. Schütze - Saarland University, Saarbrücken (Germany)
- 641 - 646
In this work, we present a statistical condition monitoring system allowing the qualitative and quantitative determination of typical faults related to the hydraulic system as well as sensor faults of the installed sensor network. Single sensor faults such as constant offset, drift, noise, and signal peaks can be identified before leading to false alarms. Furthermore, the system is able to compensate the failure of up to 5 sensors with sufficient classification performance exploiting inherent correlations of sensor signals. Feature extraction, selection, and classification using LDA are fully automated. Thus, the system can react to a detected sensor fault and adapt the sensor data used for analysis significantly improving the robustness of condition monitoring system.