Regularization GCCA model-based distributed chemical process monitoring method

Abstract

The invention discloses a regularization GCCA model-based distributed chemical process monitoring method which is used for solving the problem of distributed chemical process monitoring. According tothe method, the generality among variable blocks and the uniqueness in each block are simultaneously considered when a plurality of blocks are modeled by utilizing a related regularization generalizedcanonical correlation analysis (GCCA) algorithm, and the method can be simultaneously applied to overlapped and non-overlapped variable blocking modes. Compared with a traditional method, the regularization GCCA algorithm involved in the method extracts the common features among the variable quantum blocks through the square of the typical correlation coefficient, and further uses a least squareregression algorithm to describe the relationship among the correlation features among the blocks. In addition, according to the method, inter-block related features and intra-block features are considered separately, and two comprehensive monitoring indexes are used for monitoring separately. Therefore, the method provided by the invention is a more superior distributed chemical process operationstate monitoring method.

Description

technical field [0001] The invention relates to a data-driven process monitoring method, in particular to a distributed chemical process monitoring method based on a regularized GCCA model. Background technique [0002] The increasingly complex and large-scale development model of the modern chemical industrial production system pays more and more attention to the application of computer technology, advanced instrumentation technology, and artificial intelligence technology in management production, monitoring, scheduling and other issues. Nowadays, modern chemical process objects can store offline and measure massive sampling data online in real time. These data contain potentially useful information that can reflect the operating status of the production process, laying a solid data foundation for monitoring the operating status of the chemical process. In the past ten years, both academia and industry have invested a lot of manpower and material resources in researching d...

AI Technical Summary

Problems solved by technology

[0005] The main technical problem to be solved by the present invention is: for the distributed chemical process monitoring problem, how to simultaneously consider the commonality between variable blocks and the uniqueness of each block when implementing multi-block modeling, and how to apply overlapping and non-overlapping simultaneously variable chunking mode for

Images