基于ARMA新息模型辨识的工序质量诊断方法

A Method for Process Quality Diagnosis Based on ARMA Innovation Model

本文基于辨识 ARMA新息模型生成估计残差序列 ,再对残差序列的平均值和无偏方差进行假设检验 ,可实现工序质量的异常诊断。

In this paper, a method of process quality diagnosis using hypothesis testing for residual sequence of ARMA innovation model estimation error by recursive maximum likelihood method was studied. It involves T testing the difference between present and past unbiased variance and F testing the difference between present and past mean for residual sequence of ARMA innovation model estimation error. A simulation example shows its usefulness.