Comprehensive Performance Evaluation And Abnormal Detection Of Automobile Engine Parts Cleaning Machine

Efficient and accurate abnormal detection of Automobile engine parts cleaning machine is crucial to the stable and safe operation of cleaning equipment and the economic interests of enterprises.At present,the practical cleaning equipment health management methods has a lot of room for improvement in efficiency and detection accuracy.Therefore,the research of cleaning equipment health management has a strong theoretical value and practical value.By simulating the logic of transactions in the human brain,artificial intelligence can extract high-level features from data that often contain useful information that cannot be directly obtained by humans.Combined with the characteristics of monitoring data of cleaning equipment,this paper introduces operations research and artificial intelligence into the field of cleaning equipment health management,and proposes the comprehensive performance evaluation of cleaning equipment based on analytic hierarchy process,point anomaly detection based on auto-encoder and fluctuation anomaly detection based on k-mean clustering method.According to the characteristics of the evaluation index of Automobile engine parts cleaning machine,the hierarchical structure is established by using the analytic hierarchy process(AHP).Then,the consistency test is conducted on the pairwise comparison judgment matrix obtained by expert evaluation,and the judgment matrix will be modified until the consistency test is passed.Finally,the relative importance of each evaluation index of cleaning equipment can be obtained through the analytic hierarchy process,and the four equipment can be scored according to the monitoring parameters.Finally,according to the relative importance of the evaluation index,the original spot inspection suggestion of cleaning equipment can be modified.According to the characteristics of the abnormal data points of the Automobile engine parts cleaning machine,a method of detecting the abnormal points of the cleaning equipment based on the sparse auto-encoder and the advantage of unsupervised learning to extract the features is proposed.Firstly,the feature extraction based on sparse auto-encoder was carried out,and the original three-dimensional monitoring data was reduced to one-dimensional data.Then,the anomaly detection based on boxplot was carried out,and finally the abnormal points during the operation of cleaning equipment were found.Inspired by the clustering methods commonly used in the field of unsupervised learning,k-means clustering method was introduced into the parameter fluctuation anomaly detection of cleaning equipment,and a k-means based method for the operation parameter fluctuation anomaly detection of Automobile engine parts cleaning machine was proposed.Firstly,the normalized 3D input data is processed by data standardization,and then the elbow method is used to determine the initial clustering number.Then,k-means clustering is carried out by means of the normalized data and the initial clustering number.Finally,the threshold is set to screen the exceptions.Considering the importance of parameter fluctuation in cleaning equipment health management and fault diagnosis,parameter fluctuation anomaly detection based on k-means can obtain information ignored by ordinary point anomaly detection,which is of great significance for Automobile engine parts cleaning machine health management.