Data-driving control process monitoring method based on dynamic component analysis

Abstract

The invention provides a data-driving control process monitoring method based on dynamic component analysis. The method comprises the steps of (1) establishing ICA and DICA process statistics models according to multiple detection variables obtained from continuous detection, and defining statistics I2 relevant to independent variables belonging to non-Gaussian distribution in measured variables according to the DICA process statistics model; (2) establishing PCA and DPCA process statistics models for a process information matrix E left over after non-Gaussian distribution measured variable extraction, and defining statistics T2 and SPE of the detected variables in pivot element space and residual space; (3) obtaining similarity indexes ISM to serve as monitoring indexes through training by means of the support vector data description algorithm which is input by the performance indexes Lambda = (I2, T2, SPE); (4) comparing actual monitoring indexes ISM with IMAX and getting the conclusion that faults appear in the current control process if ISM>IMAX. According to the data-driving control process monitoring method based on dynamic component analysis, the number of monitoring diagrams can be reduced, and monitoring efficiency can be improved. The data-driving control process monitoring method based on dynamic component analysis can be widely used for multivariable control system monitoring such as industrial process control.

Description

technical field [0001] The present invention relates to the technical field of control systems, in particular to the technical field of control system monitoring, in particular to a data-driven control process monitoring method based on dynamic component analysis. Background technique [0002] Nowadays, the scale of industrial production process is getting larger and larger, especially for most chemical and biological factories. In order to increase product output, ensure product quality, and ensure factory safety, it is particularly important to monitor the production process online and provide timely and effective control process information. important. At present, the control process monitoring methods are mainly based on mathematical models and data-driven methods. Due to the existence of various unknown disturbances in the actual production process, it is difficult to obtain an accurate mathematical model, and the monitoring method based on the mathematical model is li...

AI Technical Summary

Problems solved by technology

However, in MSPC, it is often assumed that the measured variables of the control process completely obey the Gaussian distribution or completely obey the non-Gaussian distribution. However, in fact, most industrial process variable components are generally more complex, and besides mixed with various noises, they often contain Gaussian Distributed variables and non-Gaussian distributed variables, which makes the traditional MSPC limited to a certain extent in practical applications

[0003] After searching the public literature of the prior art, it was found that Z.Q.Ge, Z.H.Song. Process monitoring based on independent component analysis-principle component analysis (ICA-PCA) and Similarity Factors[J].Industrial and Engineering Chemistry Research, 2007,46( 7):2054-2063. (A process monitoring method based on ICA-PCA and similar factors, International Journal: Journal of Industrial and Engineering Chemistry Research, 2007,46(7):2054-2063), the author proposes a method based on ICA -PCA monitoring method, which combines ICA and PCA, extracts non-Gaussian distribution and Gaussian distribution information of process control measurement variables, and selects statistics for monitoring respectively, achieving simultaneous monitoring of process non-Gaussian distribution variables and Gaussian distribution variables purpose, but the author did not consider the dynamic characteristics of the control process. If the dynamic changes of the control process are considered in the model and a unified statistical variable is used for monitoring, a better monitoring effect can be achieved

Images