Research On Fault Diagnosis Methods Of Rotor Blades Based On Non-Contact Measurement And Control System

The rotor is a prominent symbol of the helicopter,which combines the functions of the common fixed-wing aircraft wing,aileron,rudder,elevator and other components.It is the direct source of the helicopter's ascending,hovering,steering and other operational forces.In view of the important position of the blade in the helicopter flight process,to achieve early diagnosis and prediction before a catastrophic accident,real-time monitoring of its status is needed to help improving the safety of helicopter flight and provide basis and reference for manual protection and parts replacement after flight.Considering the working environment of the rotor blades,it is necessary to carry out research on the rotor blade fault diagnosis method based on vibration signal analysis with the non-contact measurement and control system.Firstly,the paper deeply studies the damage type of the rotor,the signal decomposition method and the feature extraction method,which are important part of a general flow of fault diagnosis based on vibration signal.Secondly,the modal and random vibration analysis of the rotor blades were carried out by Ansys finite element software,and the response of the blades under different fault conditions was compared.Subsequently,the vibration test of the blades in 9 different situations was completed by means of wireless data transmission.In order to remove the noise,the adaptive noise reduction of the acceleration signal is realized by the sparse representation.Then on the basis of the noise reduction data,the acceleration power spectrum of the blade under different fault conditions are calculated,and the judgment of the blade health state is proposed accordingly.Finally,the types of faulted blades are classified by variational mode decomposition and improved cross-sample entropy.The research results show that the noise reduction method based on sparse representation can effectively identify the main components in the signal and remove noise.The acceleration power spectrum can well reflect the health of the blade,which is consistent with the results of the finite element analysis.Variational mode decomposition can accurately separate the signal into multiple modal components,and the improved cross-sample entropy characteristics based on these modal components have a good discrimination for blade faults.