神经网络方法在自相关过程控制中的应用

Using Neural Networks to Control Autocorrelated Process

将传统休哈特控制图应用于自相关过程控制时,会引发大量虚发报警。本文将使用时间序列模型模拟自相关过程并将神经网络方法引入自相关过程控制中。以神经网络特有的模式识别技术,对自相关过程中均值发生突变的情况进行监控,取得了良好效果。

Traditional Shewhart control chart is not applicable in many automatic process because data from these industries are autocorrelated. In this research, a BP neural network is utilized to identify shifts in process parameter values from time series models with varying values of the autocorrelation coefficient. As a result, BP neural network succeeds in a good identification rate.