Aero-engine Fault Diagnosis Using a Feature Weighting Fuzzy Clustering Algorithm

Abstract

Aero-engine fault diagnosis based on multi-sensor data is still a challenge for lacking of fault samples and high complexity of data. This paper proposed a new aero-engine fault diagnosis method by using a feature weighting fuzzy compactness and separation clustering algorithm (WFCS). WFCS can handle high-dimensional data well byconsidering the feature weighting effect on clustering.Furthermore, WFCS can cluster the aero-engine multi-sensor data reasonably by getting the maximum separation of between-cluster and compactness of with-in clusters.The proposed method was validated both on simulated dataset and real dataset and had the best average results, which demonstrates the efficacy of the proposed approach.