Research On Rolling Bearing Fault Diagnosis And Fault Trend Prediction Based On Chaos Theory

Rolling bearing is one of the most commonly used parts in rotating machinery,and almost all rotating machines inevitably use rolling bearings.With the development of high-speed trains and the rise of wind power,rolling bearings play an important role in more and more large rotating machinery and equipment.But at the same time,rolling bearings are also the most vulnerable parts.How to diagnose the fault location and the degree of fault when the rolling bearing fails,it is of great significance to ensure the safe and reliable operation of the mechanical equipment,avoid major accidents and reduce the cost of production.Aiming at the above problems,the rolling bearing fault diagnosis and fault trend prediction research is carried out in this paper.Firstly,the vibration signal of the rolling bearing is reconstructed in phase space,and the method of local projection noise reduction is studied in phase space.Then according to the chaos control principle,the Duffing chaotic oscillator is used to detect the characteristic frequency of rolling bearing,and a method based on the invariant moment of polar radius is proposed to identify the type of rolling bearing fault and the diagnosis of fault degree.The traditional particle swarm optimization is optimized by the ergodicity of chaos,and the adaptive mapping chaotic particle swarm optimization algorithm(AMCPSO)is proposed.It is applied to support vector regression model parameter optimization,and support vector regression model is used to predict rolling bearing fault trend.The main contents of this paper are as follows:(1)The failure mechanism of rolling bearing is expounded,and the feasibility of diagnosing fault form and fault degree of rolling bearing based on vibration signal is analyzed.The domestic and foreign research status of rolling bearing signal noise processing,feature extraction,fault diagnosis and fault trend prediction are discussed,and the existing problems and development direction in the field of rolling bearing fault diagnosis and trend prediction are summarized and discussed.The rolling bearing fault diagnosis and the full life acceleration test device are designed,and the single fault of rolling bearing,the compound fault diagnosis test with different degree of fault and the full life acceleration test are carried out on this device.The experimental data provide data support for this article.(2)A local projection denoising method for rolling bearing vibration signals based on phase space reconstruction is studied.The vibration signal of the rolling bearing is reconstructed in phase space.The standard deviation of the time series is chosen as the neighborhood radius,and the local projection denoising is carried out in the neighborhood.In view of the selection of noise space,the covariance matrix is solved by calculating the distance between the phase points and the centroid in the neighborhood,and the signal and noise are separated from the maximum change rate of the covariance matrix.Through the calculation of the signal to noise ratio of the simulated data and the fractal dimension and approximate entropy of the actual bearing monitoring data,it shows that the proposed method has better noise reduction ability than the traditional method.(3)A rolling bearing fault diagnosis method based on Duffing chaotic oscillator is studied.A chaotic oscillator state recognition method based on the pole radius invariant moments of chaotic oscillator phase diagram is proposed.According to the change of chaotic state of chaotic oscillator to large scale periodic state,the fault characteristic frequency component in the signal is detected,and the fault form is diagnosed.The driving force change value of chaotic oscillator system is used as the index of fault degree of rolling bearing to diagnose the degree of failure.For the problem of mixture signal in rolling bearing compound fault,the empirical wavelet decomposition(EWT)method is used to separate the signal,and the Duffing chaotic oscillator and the pole radius invariant moment method are combined to diagnose the rolling bearing compound fault and the fault degree.The results show that this method can better solve the fault form and degree diagnosis of rolling bearing single and compound fault.(4)The prediction method of rolling bearing fault trend is studied.In order to solve the problem that the traditional particle swarm optimization(PSO)algorithm is easily fall into the local minimum and slow convergence speed,an adaptive mapping chaotic particle swarm optimization(AMCPSO)algorithm is proposed,and some standard test functions are used to verify the optimization performance of the algorithm.The driving force dynamic change value of the rolling bearing chaotic oscillator is taken as the indicator of the fault trend of the rolling bearing,and the AMCPSO-SVM method is used to predict the failure trend of the rolling bearing.The results show that the support vector regression model optimized by AMCPSO method is better than the traditional method,which can effectively solve the problem of fault trend prediction of rolling bearing parts.In the end,the paper summarizes the work of this paper,and the rolling bearing fault diagnosis and fault trend prediction technology are prospected.