Adaptive Double Homogeneously Weighted Moving Average Control Chart Based On Auxiliary Information

Statistical process control is a quality management technology that applies statistical technology to evaluate and monitor all stages of the production process to maintain the production process at an acceptable and stable level,so as to ensure that the product meets the specified requirements.Among them,control charts are one of the most important tools used in the production process.In the real enterprise production,producers expect to detect the faults in the production process more and more quickly,eliminate abnormal factors in time,and avoid the production of more unqualified products,so as to avoid unnecessary resource consumption and reduce production costs.Therefore,it is necessary to continuously improve and optimize the existing control charts to improve the sensitivity of the control chart to detect small shifts in the process mean,so as to improve the production efficiency and product quality level of the enterprise.Firstly,auxiliary variables are introduced,and the regression estimator is designed by combining the process parameters of research variables and auxiliary variables,and a double homogeneously weighted moving average control chart based on auxiliary information is constructed;Then,an adaptive double homogeneously weighted moving average control chart can be constructed according to the control chart parameters that can be dynamically adjusted when continuously collecting and observing data;Finally,the introduction of auxiliary variables and the design of variable parameters that can be dynamically adjusted are combined.An adaptive double homogeneously weighted moving average control chart based on auxiliary information was constructed.The Monte Carlo stochastic simulation method is used to calculate the average running chain length of the three control charts in the runaway state,and compare it with the average running chain length of the double homogeneously weighted moving average control chart in this paper.The results show that the average running chain length of the proposed control chart in the runaway state is smaller,indicating that the small and medium deviation of the mean of the proposed control chart detection process is more sensitive,and it is relatively robust in non-normal state.Finally,the influence of parameter estimation on the proposed control chart is studied,and it is found that in order to make the controlled performance under parameter estimation close to the known state of parameters,a large sample subgroups size needs to be taken in the first stage of statistical process control.