A Decentralized Dynamic Process Monitoring Method Based on Dynamic Optimal Selection

Abstract

The invention discloses a distributed dynamic process monitoring method based on dynamic optimal selection. According to the method, a plurality of delay measuring values at sampling time are introduced for all measurement variables; on the basis of a genetic algorithm, dynamic characteristics corresponding to all measurement variables are selected optimally; and then on-line monitoring on a dynamic process is performed by using a prediction error of a regression model. Compared with the conventional methods, the method disclosed by the invention has the following advantages: the dynamic characteristic is selected optimally for each measurement variable and the process monitoring is implemented by using the prediction error of the regression model, wherein the error is a result after rejection of autocorrelation and cross-correlation in the regression model; on the basis of the novel idea, the autocorrelation problem in the dynamic process monitoring is solved; because of the single monitoring result, a phenomenon of several kinds of process monitoring results because of multiple statistic index combinations is avoided; and thus the method is an optimal dynamic process monitoring method.

Description

Technical field [0001] The invention relates to a data-driven process monitoring method, in particular to a distributed dynamic process monitoring method based on dynamic optimal selection. Background technique [0002] The purpose of process monitoring is to find faults in a timely and accurate manner, which is of great significance for ensuring safe production and maintaining stable product quality. At present, the mainstream implementation technology of process monitoring is a data-driven method, which is mainly due to the large-scale construction of modern industrial processes and the extensive application of advanced instrumentation and computer technology. Mass data can be collected during the production process. The development method that sampling data is easy to obtain and mechanism model is difficult to obtain makes traditional fault detection methods based on mechanism model gradually decline. In contrast, the data-driven fault detection method does not require a mech...

AI Technical Summary

Problems solved by technology

However, different variables have different dynamic characteristics, and the order of the delay measurement values ​​required is also inconsistent. Using a traditional regression model alone cannot achieve this goal.

Images