Study On The Quality Control Of Working Procedure In Multi-Specification&Small-Batch Manufacturing

Quality is a prominent issue of common concern in modem industrial society and national economic construction, which plays an extremely important role in the national strategic level, the competitiveness of enterprises and customer demand levels. Research shows that the fundamental product quality control is the control of production processes, not only the finished products inspection. In addition, with the acceleration of the process of economic globalization and new technological development, the living environment of enterprises has changed enormously, and the individuation of part, the diversification of market and customers’requirement increased the speed of updating, which made multi-specification&small batch become the primary manufacturing style. However, Small Batch production’s few small-volume products and small sample make the control system complicated and instable, so it cannot get a stable value of quality characteristics. Moreover, because of low quality characteristics of data, the classic statistical process control(SPC) technical requirements of the statistics cannot be met. This dissertation studies the quality control of working procedure, based on the character of multi-specification&small-batch manufacturing mode and research and application situation of statistical process control in product quality control, with the quality control of working procedure as the goal. The main contents of this dissertation are as follows:1. The actual production process of multi-specification&small-batch manufacturing is analyzed, and a similarity recognition model of grouping procedures is constructed by using the principle of similarity cell. Then, manufacturing procedure similarity is evaluated, procedures accord with principles of similar are found out, and the same group of procedures is divided, which can take full advantage of the similarity between procedure information to expand the sample space and increase the sample size.2. Data transformation algorithm of the quality characteristics and the design method of control chart are proposed based on the group similar procedures, which lays the foundation of control chart pattern recognition. Then, a basic abnormal pattern recognition method of control chart is proposed. According to the problems of special abnormal pattern recognition method of present control chart such as poor adaptability, training slow and other issues, the special anomaly pattern recognition method of control chart is proposed based on Elman neural network, and then the Elman neural network is designed. 3. In order to provide working procedure quality diagnosis with more detailed information, three similar BP neural networks are constructed to estimate the characteristic parameters of special anomaly pattern of control chart. On the basis of the current analysis and comparison of the quality diagnostic programs of working procedure, neural network technology is introduced into the quality diagnosis of working procedure. Then, the method of neural network is used to make quality diagnosis of working procedure, and the neural network model of the quality diagnosis of working procedure is constructed. In addition, the process quality control strategy is put forward, according to the causes of exceptions of quality of working procedure.