Novel multi-modal chemical process abnormal state detection method

Abstract

The invention discloses a novel multi-modal chemical process abnormal state detection method, and aims to invent a process modeling technology capable of mining related characteristics among multiplemodals and apply the process modeling technology to multi-modal chemical process abnormal state detection. Specifically, the method implements multi-modal chemical process abnormal state detection based on the following two characteristics of multi-modal sampling data: the multi-modal sampling data are correlated; the multi-modal sampling data have respective uniqueness. Compared with the prior art, the multi-modal PCA method independently builds a model for each piece of modal sampling data, and only the uniqueness of each modal is considered; secondly, related characteristic components and independent characteristic components are respectively used for carrying out online abnormal state detection, and mined potential characteristics are comprehensively used; and finally, the superiorityof the method are compared and verified in the aspect of detecting the abnormal running state of a multi-mode continuous stirring reaction kettle through a specific embodiment.

Description

technical field [0001] The invention relates to a method for monitoring the operating state of a chemical process, in particular to a novel multi-mode abnormal state detection method for the chemical process. Background technique [0002] During the operation of the modern chemical process, the temperature, pressure, flow and other data information in the production process are usually collected continuously and in real time through measuring instruments and meters. These massively stored sampling data and real-time updated online sampling data have laid a solid data foundation for the current chemical "big data" construction and intelligent chemical applications. In the past ten years, the use of sampling data to detect abnormal states in the operation of chemical processes has received more and more attention in the field of safe chemical production. Safe production and maintaining the stability of product quality. Both academia and industry have invested a lot of manpow...

AI Technical Summary

Problems solved by technology

From this perspective, the classic multimodal PCA method does not consider the relationship between the various modalities, so the multimodal PCA method cannot mine the correlation features between multimodal sampling data from the perspective of implementation route.

Images