Research On Helicopter Rotor Imbalance Fault Experiment And Diagnosis Method

Helicopter Health and Usage Monitoring System (HUMS) plays an important role in keeping a helicopter in the best condition of its Reliability, Maintainability and Supportability (RMS). However the current HUMS in service can not monitor and diagnose unbalanced faults in the main rotor system of helicopter. At past, researchers developed their methods based on rotor hub loads and blade tip displacement signal in order to detect the unbalanced faults of the rotor system. Measurement of rotor hub loads is so difficult that the methods are unfeasible in engineering practice. Based on the analysis of existing methods for fault diagnosis of helicopter rotor, the unbalanced faults in the main rotor system were studied using only information extracted from rotor-induced fuselage vibration. Through the fault simulation experimental study, diagnosis model for unbalanced faults in the main rotor system using only information extracted from rotor-induced fuselage vibration was established.The detailed work presented in the dissertation includes:(1) It is proved in theory that there exists an injection of trim tab imblance fault space into rotor induced fuselage vibration space. The work in this part provides a theoretical basis for the diagnosis of the trim tab unbalanced fault using only information from fuselage vibration.(2) A helicopter rotor unbalance fault simulation test was carried out on a rotor test rig, simulation faults include trim tab imbalance, pitch imbalance and mass imbalance in three kinds of single fault and composite faults, composite faults were designed and analysed using orthogonal experimental design method.(3) Power spectrum using the maximum entropy method (MEM) to estimate was applied to the data processing, and the imbalance fault feature was extracted by using principal component analysis (PCA). The result shows that the method can improve the imbalance fault diagnosis accuracy.(4) The power spectrum of fuselage vibration reflecting the rotor fault features was extracted, and the helicopter rotor unbalance fault diagnosis models based on general regression neural network (GRNN) and support vector machine (SVM) were established. Compared with the fault diagnosis model based on conventional neural network, the accuracy of fault classification of GRNN based model is 94.1% and the relative error of fault degree identification is approximately 5%, while the fault classification accuracy of SVM based model is 95.6%.