Prognosis and Health Monitoring Systems for Aircraft Engines

摘要

Operational reliability of aircraft engines can be increased if one can detect signs of damages before failure. Prognosis and Health Monitoring (PHM) consists in detecting these signs, and giving the probability of equipment anomaly at a defined horizon. This article presents a global overview of PHM systems. Using a system engineering approach, the needs and the equipments to monitor are clearly specified. The dedicated algorithms and embedded / ground systems split are then defined. Embedded system is mainly limited to data acquisition and data reduction. Ground system contains most of the algorithmic part. First, embedded indicators are normalized in order to be compared flight by flight. Then several kinds of approach can be applied to these indicators: among them, trend analysis toward a predefined threshold or anomaly detection. When an anomaly is detected, a classification algorithm identifies failure signature and associated equipment. Lastly, Remaining Useful Life (RUL) is estimated. This algorithmic process requires a crucial learning phase, a topic dealt with this article. Another key point is to define failure signatures as aircraft engines are extremely reliable. If requested, probability of failure is assessed by an expert. In case of confirmation, customer is notified with the need for maintenance before the estimated RUL.