The Research On Dynamic Quality Control Method During The Manufacturing Procedure

Quality is the basis of the enterprises, and also one of the core competitivenesses of enterprises. In order to gain a competitive advantage under the market economy, enterprises must carry out an effective quality control strictly during the course of production. But as the improvement of the market's demand for quality, the traditional quality control can not meet the requirements, it is necessary to implement the dynamic quality control to complete the quality control in manufacting process, and it can play a role of the quality in prevention.Besides, the diversification of customers' needs prompt the change from sellers' market to buyer's market and the growth of personalization of products. Many varieties, small batch production has gradually becomes the main mode of operation. On this basis, the concept of SPC based on small-batch has been proposed, and widely used in the enterprise. At the same time, the thesis also researches the problem of pattern recognition of control chart in small-batch, and gives a new method of pattern recognition of control chart based on support vector machine.The main work in this paper:First of all, the thesis summarized the meaning of quality, the significance and the development process of quality control, on this basis, the purpose and the significance of the study was put forward. And the thesis analyzed and summarized the status quo and trends in the field of quality control, and gave the research work and framework of the dynamic quality control in the manufacting process.Secondly, under the production mode of many varieties and small-batch, the thesis put forward the concept of dynamic quality control, focusing on the characteristics of the quality information in this mode. Based on analyzing and summarizing the method and the sphere of application in the current dynamic tolerance arrangement, the thesis established a new model. On this basis, it will adjust directly the related one of procedures or procedures and tolerance, according to the process out-of-tolerance, and make the manufacturing process to gain the optimum solution and perfect.Thirdly, the thesis analyzed the problems of applying SPC directly in small-batch, summed up the scheduling research at home and abroad, and summarized and compared the existing methods. On this basis, introducting the Bayesian prediction theory, setting up the dynamic model of process quality, and researching the parameter estimation, model intervention and implement steps of this method. Next optimizing the control parameters, and finally the thesis carried out the Case Study about this method.Finally, the thesis introduced the basic principles of control chart, detection rules, and introduced easily the traditional pattern recognition and its shortcomings, and pointed out the advantages of control chart pattern recognition method based on SVM. Through selecting the proper kernel function, establishing the model based on SVM, and comparing with the method based neural network, at last proving the feasibility and superiority of the model.