Research On Detection And Identification Algorithms Of Contactor Electrical Life Bad Data Based On Data Mining

With the development of the national economy and the wide application of AC contactors in electrical circuits,it is possible to predict the electrical life of AC contactors before failure,so as not to cause significant losses to the national economy.The bad data of the system operation affects the electrical life prediction result of the AC contactor.Therefore,the requirements for the electrical life data quality of the AC contactor are getting higher and higher,and the accuracy of the detection and identification of the bad data is also more important.This thesis studies the method of data detection and identification of AC contactor data based on data mining,which eliminates the bad data to a certain extent,which is of great significance to the reliability of AC contactor electrical life prediction.The research content is divided into the following points:Firstly,according to the basic principle and degradation mechanism of AC contactor,the corresponding characteristic parameters are calculated by defining: contact resistance,suction time,bounce time,arc phase angle,arc time,arc energy,average arc power and so on.Combining the new wavelet threshold theory with the principle of empirical mode decomposition and denoising method,an improved method for extracting degenerate characteristic parameters of EMD AC contactor is proposed,which extracts and denoises the electrical life test data,and prepares the data for bad data detection and identification.Secondly,association rules are very important data mining technology.The core idea,basic properties,algorithm steps and implementation of association rules are deeply studied.According to their defects and shortcomings,an improved association rules algorithm is proposed based on the basic association rules.The reliability and effectiveness of the improved association rules algorithm are verified by experiments.Finally,data mining is performed on the characteristic parameters of the AC contactor electrical life preprocessing by using the improved association rule algorithm.For the parameters such as contact resistance,suction time and bounce time,the association rules are mined;according to the arcing time,arcing energy,average arcing power and other parameters,the arc phase angle has periodic changes,and the association rules are mined.Get the corresponding rules.In order to verify the validity of the recognition algorithm,the bad data is manually set for detection and identification,which proves that the algorithm can detect and identify bad data of AC contactor electrical life monitoring data.Finally,the data of the new experimental array to be identified by the AC contactor is predicted to be bad data.The above series of steps are used to finally identify the bad data and eliminate the bad data.