Aiming at the problem of low precision of actuator fault identification under closed loop control of aircraft, a fault diagnosis algorithm was proposed, which extracted fault feature by ensemble empirical mode decomposition (EEMD) and principal component analysis (PCA), classified by AdaBoost and adaptive support vector machine (AdaBoost-ASVM). Firstly, in terms of feature extraction, the fault signal was decomposed into intrinsic mode functions (IMFs) of different frequency range by EEMD to avoid the modal mixing phenomenon in empirical mode decomposition (EMD). Secondly, the features of IMF were calculated from multiple dimensions. Thirdly, those high-dimensional features were analyzed by PCA to extract the main features that contain fault information. AdaBoost-ASVM algorithm was proposed to solve the problem of low accuracy of traditional multi-class SVM. Multiple SVM based classifiers with weak learning ability were constructed into a strong classifier through ensemble learning. Adaptive parameter optimization was carried out for each SVM based classifier. Simulation results shows that, compared with traditional multi-class SVM algorithm and some other machine learning algorithms, proposed method improves the accuracy of fault diagnosis and provides an effective and feasible method for the identification of aircraft actuator fault state under closed-loop control system.
Fault Diagnosis of Aircraft Actuators Based on AdaBoost-ASVM
Lect. Notes Electrical Eng.
2021-10-30
13 pages
Article/Chapter (Book)
Electronic Resource
English
Fault Diagnosis of Aircraft Actuators Based on AdaBoost-ASVM
TIBKAT | 2022
|Aircraft actuator fault diagnosis method based on AdaBoost-ASVM algorithm
European Patent Office | 2020
|Fault Diagnosis and Condition Monitoring of Aircraft Electro-Mechanical Actuators
Springer Verlag | 2021
|Fault Detection and Diagnosis of Aircraft Actuators using Fuzzy-Tuning IMM Filter
Online Contents | 2008
|