Pump Failure Mode Forecasting Through the Use of an Integrated Diagnostic Methodology

This paper presents an Integrated Diagnostic Methodology (IDM) for analyzing hydraulic systems. The IDM fuses information from condition monitoring, system diagnosis, failure prognosis, and prescriptive action to provide a means to detect failure modes (in-situ), failure state, time to failure, and prescriptive maintenance (if needed). Information about the pump condition is determined through time series analysis of the pressure pulsation of the pump. Using fuzzy pattern recognition techniques, failure symptoms are detected and classified into known failure modes. Similarly, state of failure is categorized using a fuzzy reasoning method. Finally, a trend analysis is made to forecast time to failure, while a set of rules is used to decide the proper maintenance action. As an example of the IDM procedure, experimental results monitoring a gear pump are presented.