Process Quality Control Chart Bayesian Correction Model Based On Poor Information

Quality largely determines the use of the product performance and economical benefit of enterprises. Enterprises have attached importance to the product quality testing and monitoring. Process capability is one of the important indices for quality testing, it can response to the quality of products. The quality control chart can provide theoretical for quality monitoring. As customer demand for product differentiation and personalization, the process needs to update and improve, the production process shows the characteristics of small batch of poor information, based on the mass production control chart theory cannot fully apply. In this paper, a new process ability of quick judging for the target, the optimal sample size to choose as constraint, build the new process capacity of poor information fast judgment model. The paper researched poor information construction process quality control chart bayesian correction model. The main innovative points are as follows.(1) Building the new process capacity of poor information fast judgment model. Restricted to high production cost and complicated production process, new process capability trial production sample size is less. For such poor information decision problem under the background of process ability, this paper constructs the bayesian fast judgment model. Firstly, digging small sample information fully, obtain the mean and variance, poor and migration coefficient of quality characteristics, and take them as the constraint, Minimum(or maximum) to benchmark coefficient as the objective function optimization model was constructed. Finally judgment process capability index is converted into a constrained optimization problem.(2) Solving the new process capability bayesian benchmark quick decision model for decision table. To efficiently solve the bayesian rapid judgment model, Firstly proposed the benchmark qualified determination coefficient of the principle of caution and unqualified decision value principles of loose. And make the quickly determine feasible domain discretization of the model, using the theory of nonlinear benchmark decision table.(3) Building the new process dynamic bayesian quality control model. Considering the use of small sample to determine the new process ability, cannot erase the effect of the volatility, and in order to continue monitoring the new production process, this paper constructs the gray dynamic bayesian model for quality control. The model could divide into production steady quality control chart area, to determine area and uncontrolled area. With the increase of sample size, under the condition of bayesian correction function, to determine area is gradually degradation, and avoiding the abnormal fluctuations affect the results.