Chemical production process monitoring method based on multi-block projection non-negative matrix factorization

Abstract

The invention discloses a chemical production process monitoring method based on multi-block projection non-negative matrix factorization (MPNMF). The chemical production process monitoring method comprises the steps that n historical normal samples of physical quantities monitored by m physical quantity monitoring points in the chemical production process are acquired, and data are divided into Bsub-blocks after being preprocessed, and statistics Nb2 and SPEb of each sub-block and control limits SPEb and lim of the statistics Nb2 and SPEb are calculated; and physical quantity data of m physical quantity monitoring points in the chemical production process are collected on line, and the data are preprocessed in the same preprocessing mode and then correspond to B sub-blocks, and statistics SPEb and new of each sub-block are calculated, and Bayesian reasoning is adopted to construct statistics and BICSPE. If the statistics BICSPE exceeds the control limit of the statistics BICSPE, a fault occurs. After a fault is monitored, a weighted reconstruction contribution value of the physical quantity data in each sub-block is calculated and a maximum value is taken, wherein the physical quantity data corresponding to the maximum value is the physical quantity data when the fault occurs. According to the chemical production process monitoring method, the number of statistics needing tobe monitored is reduced, and the monitoring result is more visual, and the monitoring cost is saved.

Description

technical field [0001] The invention belongs to the technical field of industrial production process fault monitoring, and in particular relates to a chemical production process monitoring method, which is used to improve the accuracy rate of fault detection and identification in the chemical production process. Background technique [0002] With the rapid development of data acquisition and computer technology, the chemical production process has become more automated and intelligent. In recent years, frequent accidents occurred in the chemical production process, production safety has become very important, and process monitoring has received more and more attention. Multivariate statistical process monitoring (MSPM) methods, such as principal component analysis (PCA), partial least squares (PLS), etc., are widely used in industrial processes because they can extract effective feature information from process data for process monitoring. [0003] MSPM is mainly used to pr...

AI Technical Summary

Problems solved by technology

[0005] The technical problem to be solved by the present invention is that the number of statistics required to be monitored is large when applying the existing block-type chemical production process fault detection method, which is not conducive to reducing the technical problem of calculation speed and use cost. For this reason, the present invention provides A method for monitoring chemical production process based on multi-block projected non-negative matrix factorization (MPNMF)

Images