APU Degradation Prediction Based on EEMD and Gaussian Process Regression

Abstract

Prediction of aircraft Auxiliary Power Unit (APU) degradation plays an important role in aircraft health monitoring and condition-based maintenance. Due to complexity of the system and the variability and stochastic nature of the working environment, the degradation of the APU engine is hardly to be described by traditional deterministic time series analysis. Alternatively, advanced probabilistic methods are required to capture the randomness of the degradation signal. In this study, a probabilistic prediction method based on Gaussian Process Regress (GPR) combined with Ensemble Empirical Mode Decomposition (EEMD) is developed and four evaluation indices are selected from different aspects to evaluate the effectiveness of the EEMD-GPR. This method has the potential to increase the reliability and precision of the performance prediction for aircraft complex systems.