Research On Fault Feature Extraction Method Of Planetary Gear-box Based On EEMD

With the rapid development of industry, multifunctional mechanical equipm-ent is playing an important role in modern manufacturing. At the same time, mechanical equipment tends to be more complicated structure. In order to achie-ve the safe operation of equipment, and to avoid the loss of production, so in the actual project often need to the real-time monitoring of equipment or regul-ar fault diagnosis.However, due to the complexity of the working environment of the gear box, method to evaluat the fault detection and damage degree is extremely limited. And for different types of gear box, the gear box fault featu-re extraction method is equally effective, it is often the focus of research. The-refore, the research on the gear box feature extraction method has a good eng-ineering application prospects and important academic value.In this paper, Firstly proposed a method which based on the difference bet-ween the signal itself standard set decomposition parameters, to solve the EE MD decomposition parameters set of experience dependent problems, such as the signal EEMD choice decomposition number of NE and white noise amplitu-de coefficient k, Secondly, through the introduction of correlation coefficient, k-urtosis, variance contribution rate as IMF refactoring guidelines to original sig-nal pulse on extraction and de-noising purpose; Then, according to detection p-rinciple, it proposed EEMD and FWT combing method to achieve second den-oising and feature extraction of gear box. Compared with the single use of E-EMD or FWT method, the efficiency of the integration method is verified by comparing with the single which use of the integration method.The main idea of this paper is to study the method of EEMD and the rel-atively mature wavelet analysis which has the ability of adaptive decompositio-n, and finally achieve the goal of noise reduction and realize the fault feature extraction of the gear box. First of all, through the gear simulation experimen-ts verify the effectiveness of the proposed method; secondly, through building QPZZ-Ⅱ type test bench test platform of spur gear, the artificial tooth crack fault, the vibration data and using acceleration sensor fault gear box was collec-ted to verify the validity of the method the gear box gear local fault signal denoising and feature extraction; Finally, through the test platform of 75KW pl-anetary gear box, and artificial sun gear root crack fault vibration signal using acceleration sensor acquisition fault of planetary gear box, and the processing method of the signal data, verify the validity the method of fault feature extrac-tion of planetary gear box. Analysis with MATLAB software as a platform, th-rough simulation and real data validate the proposed method, it shows that me-thod presented which can effectively realize partial cylindrical gear box and a planetary gear box fault feature extraction.