The Study On Multivariate Statistical Process Quality Control For Small-batch That Based On Kalman Filter

With the continuous development of economic globalization,competition among enterprises become more and more intense.Competition comes down to the final product quality competition.In this paper,new product prototype and small volume production of less than the amount of data exist and the impact analysis of such issues as product quality,combined with the annual Science and Technology in Anhui Province key project "Intelligent control of the quality of diagnostic technology in the integrated manufacturing system in the applied basic research" (Item Number:07021075) to the Kalman prediction theory,try to use the Kalman reasoning functions to solve the small batch production process for modeling the problem of insufficient data.Based on the study of multivariate statistical process quality control chart and small-batch manufacturing procession the article bring forward a kind of synthetical modeling method to deal with multivariate statistical process for small batch manufacturing.By Kalman forecast to theory,try to solve the problems of small batch production with insufficient data.To provide a new way of small-batch manufacturing.When the system equation of state is known,with a view to significantly improve the accuracy of the model;Mainly involves the following aspects:1.Analysis of the low-volume,multiple quality control of the status quo and existing problems;2.In the analysis of multiple small batches of existing quality control techniques,based on the T2 control chart for the lack of existence,the application of Kalman filter theory to design and implementation of small quantities of quality control model of multi-KT~2;3.Combined with the practical application of quality control problem, expatiate KT~2 model on the implementation and application,Verify the feasibility and effectiveness of this model;Practical application/emulational instance shows that this modeling method can make the best of data we have gotten,build control model dynamicly, consequently it could resolve the problem of lacking modeling data in small-batch manufacturing procession.