| As the core component of the electric rear-view mirror,the driver of electric rear-view mirror is more prone to failure than the traditional mechanical driver,leading to the electric rear-view mirror being unable to rotate to the target position,which seriously affects the driver’s field of view.At present,most manufacturers that produce electric rear-view mirrors lack automated failure detection methods and facilities to test electric rear-view mirror driver.Employees excute mechanical auditory diagnosis with human ears.Since judgment with human hearing is highly subjective,it is possible to cause malfunctioning electric rear-view mirror drivers to flow into the market,causing negative effects.This dissertation takes the driver of electric rearview mirror as research object.The vibration signal is analyzed to diagnosis the fault and develop the corresponding fault diagnosis system.The main research contents are as follows:(1)Aiming at the problem that the vibration signal of the electric rearview mirror driver of the automobile features nonlinear and non-stationary,and the fault characteristics are difficult to extract,an improved EEMD-AR power spectrum estimation model combined with support vector machine is established for fault diagnosis.In the improved empirical mode decomposition method,the kurtosis coefficient and the correlation coefficient are used to filter the decomposed intrinsic mode function to abandon false components.The components are subjected to adaptive power spectrum estimation,and the accumulated spectrum is obtained;the boundary points and amplitudes of the increasing frequency bands in the accumulated spectrum are extracted as feature vectors for support vector machine training and classification.The results show that the method can achieve an average overall recognition accuracy rate of 94.3%,and the average time consumption is 46.1s.Compared with other traditional empirical mode decomposition methods,the improved model scores a higher accuracy rate.(2)Use the signal decomposed by the EEMD method to reconstruct the vibration signal and perform the symmetrized dot patten transformation to obtain the mirror-symmetrical snowflake image.Extract the features of the gray-level co-occurrence matrix and K-means clustering algorithm is used to provide a visualization method to classify the vibration signals of the three types of electric rearview mirror driver samples.(3)Discuss the selection of the hardware parameters for the fault diagnosis system of the electric rearview mirror driver and the construction of the experimental bench.The parameters of the signal acquisition board,signal amplifier,and vibration signal sensor are determined;the control law of the electric rear-view mirror driver is clarified,and the measurement point position of the vibration signal is determined through experiments,which is the hardware basis for the software programming of the host computer.(4)Designed and compiled the host computer program of the fault diagnosis system of the electric rearview mirror driver,and completed the design of the three main interfaces of the parameter setting interface,the historical result judgment interface,and the waveform graphic analysis interface.Corresponding the procedures of electric rearview mirror driver motion control and vibration signal sampling with actual automation facility,debugging system parameters,designing and conducting experiments comparing automation facility with manual testing.The experimental results show that the automated test equipment designed in this paper has a higher recognition accuracy rate than manual labor,and it can be concluded that the automated equipment is suitable for actual enterprise production and quality inspection. |
