Research On Process Capability Evaluation Model For Complex Products Of Mass Varieties And Small Batches

With the development of industrial technology, market environment in which complex products manufacturing enterprises existed in has undergone a tremendous change. Market for the complex products of precision, coordination, delivery time and cost has many strict requirements. Consequently, the patterns of producing in small batches has played an important role in the manufacturing industry. But it also often faces with the problem of small sample and lacking of information, which undoubtedly poses a challenge to the traditional production process quality control method based on large quantities. Based on the quality control of mass variety and small batches production, and also with the help of knowledge in Operation research, Mathematical Statistics, Maximum Entropy Principle and Virtual Sample, the thesis will study the process capability evaluation model for complex products of mass varieties and small batches in two aspects.① The method of process capability evaluation for mass varieties and small batches production which is based on maximum entropy virtual generation model is proposed. The traditional process capability evaluation commonly compares parameter estimation point with the recommended values of process capability indices(PCI).It easily lead the accuracy of estimation to depend on the size of sample. But in the mass varieties and small batches production, it often cannot obtain enough effective samples, especially in the initial stages of processing, which undoubtedly greatly increase the uncertainty of results. And it also will lead the estimation of PCI to have a large variations. To solve these problems, firstly, the thesis proposes a virtual products generation model to generate many virtual overall. Then to do sample test for the virtual overall which will produce some sampling sample. Finally, the thesis looks for a virtual sample that has the highest similarity with the actual small sample which can reflect the PCI of actual production. Through the above steps, the thesis will effectively solve the problem of PCI evaluation for the mass varieties and small batches production.② The thesis studies a Bayes estimation method of the PCI for mass varieties and small batches production, proposing a PCI Bayes inference model. Firstly, to analyze the limitations of the Bayes estimation under circumstances that has no information priori or conjugated priori. On this basis, the thesis establish the Bayes inference model of PCI to obtain the posterior expectations estimation of PCI. Finally, an empirical analysis is given to demonstrate that the method is more simple and the results more accurate and reliable.