Research On Abnormal Identification And Application Of Intelligent Workshop Production Based On Improved K-means

Intelligent manufacturing is the main trend of the development of the current manufacturing industry.As one of the main carriers of intelligent manufacturing,intelligent manufacturing workshop has emerged in a large number of industries.The production abnormality of intelligent manufacturing workshop has been widely concerned and studied by the academic and engineering circles.How to efficiently identify the production abnormality of intelligent manufacturing workshop has become one of the important problems to be solved in the field of intelligent manufacturing.Therefore,on the basis of indepth analysis and research of relevant theoretical knowledge of intelligent workshop and production anomaly identification technology,this paper carries out the following research work around the problem of production anomaly identification of intelligent workshop:(1)Research on improving K-means clustering algorithm.To solve the problem that Kmeans is overly dependent on the initial clustering center and it is difficult to determine the K value of the clustering number,based on the density combined with the Canopy clustering algorithm,the maximum weight product method of sample density was proposed.The initial clustering center was determined according to the weight product size of the sample density,and an adaptive mechanism of the input parameters of K-means was constructed.The rationality and feasibility of the improved K-means clustering algorithm are verified by comparative experiments.(2)Establish production anomaly identification model based on improved K-means intelligent workshop.On the basis of studying and analyzing the clustering method of time series data,an improved production anomaly identification model of intelligent shop based on K-means was established based on the dynamic time warping algorithm(DTW)as the similarity measurement method.In this model,Principal component Analysis(PCA)was used to determine the main production characteristic parameters of the intelligent workshop,and time series data were divided by time sliding window equidistance to construct the experimental sample data.By drawing the distribution map of abnormal samples,the clustering cluster was given practical production significance.(3)Research on the application of production anomaly identification in intelligent workshop.A domestic intelligent aluminum electrolysis manufacturing workshop was selected as the application object of the production anomaly identification model based on improved K-means in this paper,and abnormalities in five production factors such as working voltage and series current were identified in the production of electrolytic aluminum.To sum up,based on the improved algorithm improve identification precision,establish intelligent workshop production anomaly recognition model and intelligent workshop production abnormal recognition applied research from three aspects,for intelligent workshop production abnormal recognition problem specific exception recognition method is proposed,using k-means method,such as data mining technology was applied to intelligent workshop anomaly identification problem.Test based on the model proposed in this paper,the experiment shows that the application of the proposed intelligent workshop production abnormal recognition model of aluminum electrolysis intelligent manufacturing plant in the production of abnormal recognition result is consistent with the actual production of abnormal situation,shows that this model can provide intelligent workshop production abnormal recognition support and reference for theory and practice.