基于多变量统计方法的产品质量控制

The Product Quality Control Based on Multivariable Statistical Methods

提出了一种多变量统计质量控制方法来减小由于过程扰动引起的产品质量变化.该方法首先通过正常工况的历史数据,建立一个部分最小二乘的回归模型,利用高采样频率的过程测量值来预测质量变量的值.预测误差作为部分最小二乘逆模型的输入得到过程操纵变量的调节量,通过调节过程操纵变量来抑制过程扰动,减小质量变量的变化.所提出的多变量统计质量控制方法在TE过程中得到了验证.仿真结果表明,与传统的PID质量控制方法相比,所提出的方法能减小由过程扰动引起的质量变化.

A multivariable statistical quality control method was presented to decrease the variance in product quality by the influence of process disturbance. A partial least squares regression model is used to predict the product quality based on high frequency process measurements. Prediction error as the input for the inversion of PLS regression models can obtain the adjustment of the process manipulated variables. By regulating the process manipulated variables, the process disturbance can be restrained and the variance of product quality can be reduced. The proposed multivariable statistical quality control method was demonstrated on the Tennessee Eastman benchmark process. The simulation result shows that the variance of product quality used by the proposed scheme is smaller than that utilized by conventional PID quality control.