Sensor Fault Diagnosis and Classification in Aero-engine

In this paper, a model-based analytical redundancy method is used for sensor fault detection. The diagnosis system uses Kalman filters as state estimators, which can detect 6 kinds of typical sensor fault modes. Then Design a Multi-kernel SVM fault classification system, which makes use of PCA and WPEE method to extract fault characteristic. Compared to the traditional diagnostic and classification methods, Multi-kernel SVM is more effective