Research On Improved Singular Value Decomposition Method For Rolling Bearing Fault Signal Process

Rolling bearings are widely used in rotating machinery and equipment in various industries,such as aviation,power,coal,metallurgy,and automobiles,due to their low friction resistance and high mechanical efficiency.During the development of rotating machinery and equipment in the direction of high speed and heavy load,the rotor support component-rolling bearing often works under severe conditions such as variable load and impact load,resulting in an increase in the failure rate of the bearing.If the potential failure was not discovered and eliminated in time,then it will cause more serious secondary equipment damage or even catastrophic accidents.Therefore,the condition monitoring and fault identification of the rolling bearings of key equipment is of great significance for ensuring the safety of equipment and personnel and reducing unplanned downtime.This paper takes rolling bearing fault identification as the research goal and deeply studies the application of singular value decomposition in vibration signal analysis,including the extraction of weak fault features of bearings,the selection of fault sensitive signal components,and the research of compound faults of rolling bearings;The effect of matrix embedding dimension m and signal reconstruction order k of signal singular value decomposition are also analyzed.The main work of the thesis is as follows:(1)Based on the vibration mechanism of the rolling bearing,theoretically,this study summarized the most common faults of pitting corrosion on the inner and outer rings and the characteristics of the vibration signal when the failure occurred;the process of fault identification is summarized and the focus laies on the principles of envelope analysis.The singular value decomposition and its application in signal processing are described in detail.In addition,the principle and algorithm process of the variational model decomposition method are explained for the theoretical basis for the fault signal denoising,vibration signal optimization SVD and the study of bearing composite failure;(2)Aiming at the problem that it is difficult to identify the faults caused by the weak and early fault features of the rolling bearing under the large background noise,this study proposes a accumulative kurtosis-singular value decomposition method for extracting the weak fault signal features of the rolling bearing.This method is to use the accumulative kurtosis to measure the effect of a single singular component on the accumulative singular component from the impact characteristics of the signal singular value decomposition process,so as to effectively determine the reconstruction order k of the weak fault characteristic signal and then realize the vibration noise removal and the extraction of weak fault signal features.In addition,the amplitude kurtosis of the envelope spectrum and the envelope kurtosis of the Teager energy operator are combined with the signal singular value decomposition to form the singular value decomposition-square envelope spectrum and the singular value decomposition-kurtosis-Teager energy operator for the early fault identification of rolling bearings.Both simulation and experimental data verify the ability of accumulative kurtosis to extract weak fault signals.Teager energy operator is more suitable for extracting the characteristics of weak fault signals of rolling bearings;(3)In order to characterize how much the vibration signal component carries bearing fault information,the time domain and frequency domain of the fault signal are comprehensively considered,and the product of the root mean square value of the bearing vibration signal and its envelope spectrum amplitude kurtosis is defined as a comprehensive fault parameter.The parameter is used to solve the problem of screening the most sensitive component of the fault signal after the singular value decomposition of the vibration signal.The larger the comprehensive fault parameter,the more obvious the fault characteristic frequencies in the envelope spectrum of the corresponding signal component,from which it is easier to identify the characteristic frequencies of the rolling bearing fault.The comprehensive fault parameters are introduced into the singular value decomposition process of the vibration signal to realize the optimal determination of the embedding dimension m and the reconstruction order k.On this basis,a comprehensive fault parameter-singular value decomposition method is proposed.The simulated fault signals and experimental data of rolling bearings show that the comprehensive fault parameter-singular value decomposition method can extract the fault features in the vibration signal to the greatest extent;(4)Aiming at the problem of identifying compound faults of rolling bearings,the comprehensive fault parameter-singular value decomposition method is expanded to form a multi-signal component comprehensive fault parameter-singular value decomposition method.This method is based on the multi-signal component processing characteristics of singular value decomposition and determines the optimal embedding dimension and reconstruction order of the signal sub-components by synthesizing the distribution characteristics of the fault parameters,thereby realizing the decomposition of the composite fault signal into multiple signal subcomponents which aims at the separation of the composite fault signal.Finally,the multi-signal component comprehensive fault parameter-singular value decomposition method is applied to the vibration signal analysis of the compound fault of the rolling bearing,and the effectiveness of the method is verified.The over-decomposition and modal aliasing in the singular value decomposition of the vibration signal were also pointed out in the experimental signal analysis.