Fault Diagnosis Method Of Diffused Silicon Pressure Sensor Based On Multi Classification Of Acquisition Signals

With the continuous development of information technology and electronic technology,sensors have been increasingly applied to automatic control systems and industrial systems,its running state is directly related to the normal operation of the whole system,therefore,the reliability of the sensor is crucial.To a certain extent,the sensor fault diagnosis technology can increase the reliability of the sensor,to ensure the normal operation of the system.The research on fault diagnosis of sensors has important academic significance and engineering application value.Although the existing sensor fault diagnosis technology has achieved good results,but there are still many problems and shortcomings.Especially the advantages of fault diagnosis technology based on data driven,make it a research hotspot and focus in the field of sensor fault diagnosis,there are still many problems that need to be further solved.In this paper,a series of researches on the feature extraction and fault classification of diffused silicon pressure sensor fault diagnosis technology are carried out based on the acquisition signal of pressure sensor.The research mainly includes the following aspects:(1)Elaborate the research background and significance of sensor fault diagnosis technology,taking the diffused silicon pressure sensor as the specific object,the fault types and causes of the sensors are analyzed,from the perspective of data driving,we compare and analysis the method of feature extraction and classification for pressure sensor signals.(2)An improved pressure sensor fault feature extraction method based on improved HHT marginal spectrum energy is proposed,firstly,the modal aliasing and end-effect problems of the HHT algorithm are analyzed;through the improvement of empirical mode decomposition adding white Gaussian noise,the modal aliasing phenomenon in signal decomposition is solved;by extending the endpoint of the signal to be decomposed,the endpoint effect in the algorithm is eliminated;the improved HHT method is applied to the pressure sensor data signal to obtain the marginal spectrum energy of the sensor state and use it as the characteristic information of the pressure sensor signal.(3)A method for solving multi-classification problems based on the least squares support vector machines is introduced,this method cascades multiple SVMs to achieve multi-classification of faults,and applies this method to pressure sensor fault classification;the least squares support vector machine used in this paper,it can convert quadratic programming problems into solving linear equations,improving calculation speed and convergence accuracy of support vector machines,at the same time,the relevant parameters are parallel optimized by the particle swarm optimization algorithm,so that the classification performance of the support vector machine is optimized,and the effectiveness of the algorithm is verified through open source data.(4)Taking the pressure sensor as the specific experimental object,the pressure sensor fault diagnosis experimental platform was established.By using the method proposed in this paper,the feature extraction and fault classification of the pressure sensor signal were achieved,and the status recognition and the fault diagnosis of the pressure sensor were realized.