关键词: 控制图, 智能诊断, 人工神经网络, 模式识别, 参数估计

Abstract: To control the machining process quality more effectively under the Contemporary Integrated Manufacturing System (CIMS) environment, an on-line intelligent diagnosis and analysis system for quality control charts was proposed, which consisted of four modules, namely, pattern recognition module, parameter estimation module for abnormal patterns, diagnosis and analysis expert system and process adjustment system. To recognize the control chart patterns and estimate the parameters of abnormal patterns effectively in this intelligent diagnosis and analysis system, a new neural network model named regional supervised feature mapping (RSFM) network was proposed. The topology structure and training algorithm of this network were represented, and its basic performance was also studied. Numerical simulation and an engineering application example are provided to demonstrate that RSFM network is fit for pattern recognition and parameter estimation of control charts in a real time SPC system.

Key words: control charts, intelligent diagnosis, artificial neural network, pattern recognition, parameter estimation