关键词: 统计工序质量控制, 小波变换, 概率神经网络, 模式识别, 专家系统

Abstract: Aiming at the problem of statistical process control, basic framework of intelligence control system was presented. Patterns and expression of the control chart were discussed. Three major fields of intelligent statistical process control including control chart patterns identification, abnormal patterns parameters estimation, and the diagnose expert system were introduced and analyzed. The solution scheme and system model were also presented. In the modeling of the structure, the wavelet probabilistic neural network was used to recognize the control chart patterns and estimate the abnormal patterns parameters. Simulation and application results showed that the performance of the proposed method had many advantages such as simple structure, fast convergence, high identification rate, and it could be used in on-line detection and analysis of control chart.

Key words: statistical process control, wavelet transform, probabilistic neural network, pattern recognition, expert system