Research Andapplication Of BSS And HHT Algorithm For The On Line Diagnosis Of Main Reduce

Main reducer is the key component of a vehicle’s transmission. It is the mainproducer of malfunction, directly influencing evaluation of quality of a vehicle.Nowadays, there are two common ways to assess the quality of main reduceradopted by many famous large car manufactures at home and abroad. One way isexperienced engineers’ examing and or listening to running main reducer driven byelectrical motor on manufacturing line. Another is evaluating by experienced andprofessional engineers who drive the tested vehicle. Obviously, this job has to bedone by experienced and professional engineers whose evaluation is highly objective.This method both produces errors which cannot reflect features of malfunctions, andincreases the intensity of workers dramatically. Therefore, it is greatly important todevelop new high-tech system of automatic malfunction analysis and relatedequipment.Vibration is a common indicator of malfunction of a machine. It also oftenleads to other malfunctions. Vibration is also an important indicator of safety andstability of a machine. Therefore, it is plausible to obtain the vibration signal usingrelated apparatus, analyze and process information behind, and accomplishintelligent and automatic analysis of main reducer’s quality.My dissertation is mainly composed of the following parts.First part is about the research of extraction of main reducer’s vibration signal.This part begins with algorithm of obtaining typical features of malfunction’sfrequency by exploring malfunction mechanism of main reducer. Then, a3D modelis built based on lab platform using UG software, trying to find and optimize thestrategy of location of sensor by loading ANSYS. Based on the signal of mainreducer obtained by acceleration sensor put at best vibration location, the processand analysis is prepared to proceed.The second part is devoted to research of optimizing blind source separation(BSS) algorithm. This part begins with research of BSS algorithm’s theory andevaluation criteria. Then, based on the research, obtain mixed signal from sensor byreplacing objective function’s nonlinear functions of Fast ICA algorithm withHuber-M estimation function. And then extract signal of running main reducer using improved ICA (independent component analysis) algorithm. Lastly, distinguishsignal of running main reducer needed to research by comparing the energy ofdifferent components.The third part is devoted to research application of HHT algorithm. This partbegins with analysis of difficulty of unstable signal. Then based on research of HHTalgorithm’s mathematic principle and steps, analyze the instant frequency,time-varying frequency and its law. And evaluating criteria are set for influence ofend effect to compare different inhibit effects by adopting different intelligentmethods to prevent end effects after researching defects of end effect of HHTalgorithm. Lastly, in order to avoid the analysis difficulty rising from great change ofmalfunction frequency caused by different speed of vehicle, choose the energy ofIMF component as vector of malfunction features, and conduct analysis ofmalfunction of main reducer using nonlinear relationship of vibration signal obtainedby intelligent methods.The last part is about presentation and analysis of software and hardwaresystem of evaluation platform developed by a project based on this dissertation. Thissystem has two main function: distribution and analysis. Main reducers aretransferred to a mute room. After distribution, qualified main reducers are producedfor seamless connection, while unqualified ones are transferred to malfunctionrepairing platform after labeling. The vibration signal of unqualified main reducerswill be analyzed intelligently and repaired manually. Currently, the system beendeployed in the manufacturing line of Wuling Axle Plant of Liuzhou city.Supporting digital testing equipment is also used in related car parts makers. Iteffectively resolves the issue of conventional method of rear axle evaluation andover reliance on human experience, improving the efficiency of car makers notably.The feedbacks are quite positive.