Entropy-based sensor selection for condition monitoring and prognostics of aircraft engine

The condition of the aircraft engine directly affects the reliability and safety of the aircraft. To monitor its condition, many sensors are deployed on or in the aircraft engine. However, only parts of the sensors can provide useful information for failure prognostics. An entropy-based sensor selection method which can provide quantitative description of the information contained in sensor data is proposed in this paper. Selecting the appropriate sensors cannot only help reduce the amount of data processing, but also increase the performance of condition monitoring. The experiments implemented with sensor data sets from NASA Ames Research Center prove the effectiveness of the proposed method.