Principle component kernel similarity immune mechanism based aircraft engine fault diagnosis method

Abstract

The present invention discloses a principle component kernel similarity immune mechanism based aircraft engine fault diagnosis method. A principle component kernel similarity immune mechanism based fault diagnosis method is proposed on the basis of a principle component kernel theory and an immune system mechanism. In the method, the similarity measurement in a principle component kernel form is adopted in an immune form space to regard each sample in a known fault mode as an antibody, regard a to-be-detected sample as an antigen, and convert a fault diagnosis problem into an antibody-to-antigen identification problem. The method is slightly influenced by a fault mode distribution structure, and when the dispersion degree of a fault sample is high and the clustering property is poor, a good diagnosis result still can be obtained.

Description

technical field [0001] The invention relates to a fault diagnosis method, in particular to an aeroengine fault diagnosis method based on an immune mechanism of principal component kernel similarity. Background technique [0002] At present, a variety of intelligent fault diagnosis methods for aero-engines have been developed at home and abroad, such as expert system methods, neural network methods, and support vector machine-based diagnosis methods. These methods have contributed to the establishment of intelligent diagnosis of aero-engines and the improvement of the decision-making efficiency of system operation and maintenance personnel. However, the above diagnostic methods need enough training samples of known fault modes to exert excellent performance, otherwise the results are often unsatisfactory, and the real-time performance of the algorithm is poor, which makes it difficult to meet the requirements of fast online diagnosis. [0003] The traditional fault diagnosis...

AI Technical Summary

Problems solved by technology

However, this method has a high diagnostic accuracy for fault samples with good clustering. When the fault samples are scattered and the clustering is poor, it will produce wrong diagnostic results, which limits it. Application in Aeroengine Fault Diagnosis

Images