Research On Process Monitoring And Abnormal Pattern Recognition Of Multi-variety And Small Batch Production Based On TK-EWMA Control Chart

With the ever-increasing level of social and economic development,consumer demand is becoming more diversified,which makes companies gradually adopt multi-variety and small batch production mode to replace mass production mode,and the quality control problem in the process of multi-variety and small batch production has increasingly become the urgent problem of manufacturing enterprises.Based on this background,with the purpose of improving the process quality control ability and the quality of the process output,this paper combines the characteristics of the multi-variety and small batch production mode,and studies the problem of process quality monitoring and abnormal pattern recognition under multi-variety and small batch production mode.First,for the problem that the small sample size of the multi-variety and small-batch production process leads to a large deviation in the estimated value of the standard deviation of the quality characteristic parameter,which affects the monitoring performance of the control chart,and the defect of T-K multi-variety and small batch control chart with low sensitivity to small and medium offset,this article combines the exponential weighted moving average(EWMA)statistical process control theory with the T-K control chart to generate the TKEWMA control chart to monitor the offset of the mean and standard deviation of the key quality characteristic parameter during the multi-variety and small batch production process.The TKEWMA control chart is compared with the T-K control chart by Monte Carlo method.The comparison result shows that the abnormal detection ability of the TK-EWMA control chart,especially for small and medium offset,is significantly better than the T-K control chart.At the same time,examples are used to illustrate the effectiveness of the TK-EWMA control chart to monitor the multi-variety and small batch production process.Secondly,this paper establishes the Markov chain model of the TK-EWMA control chart to analyze the performance of the control chart,analyze the main parameters' influence on the performance of the TK-EWMA control chart.Then according to the analysis result,this paper provides effective suggestions for the parameter selection of the TK-EWMA control chart under different offset of the mean and standard deviation.After that,a general method and steps for TK-EWMA control chart parameter optimization are given under the condition of a predetermined offset degree,which provides guidance for more precise selection of TK-EWMA control chart parameters to further increase the actual application value of the TK-EWMA control chart during the multi-variety and small batch production process.After the process abnormality is detected,it is necessary to further identify the abnormal pattern characteristics to take targeted corrective measures.Therefore,this paper finally proposes a TK-EWMA control chart abnormal pattern recognition method based on the PCA-IMPSO-PNN model.The simulation results show that the recognition model proposed in this paper has better convergence effect and the recognition accuracy rate is as high as 97.67%,which has certain advantages in recognition ability.