A fault detection approach for aero-engines based on PCA

Traditional fault detection approaches for aeroengines based on PCA cannot effectively detect faults when the data does not follow the normal distribution. Meanwhile, there are few effective methods for the elimination of outliers during the modeling thus the model precision cannot be guaranteed. Aiming at a solution of the problems above, a new fault detection approach for aero-engines based on PCA is proposed, including the PCA-KDE fault detection approach and the R-PCA outliers´ elimination approach. The instance analysis on a turbofan engine shows that this approach is able to detect potential faults accurately and can provide the maintenance staff with evidences for early elimination.