Sewage plant sludge bulking detection method based on robust adaptive canonical correlation analysis

Abstract

The invention discloses a sewage plant sludge bulking detection method based on robust adaptive canonical correlation analysis. A robust adaptive canonical correlation analysis model fusing robust decomposition of a data matrix, canonical correlation analysis, reconstruction of a robust T2 detection graph and a robust SPE detection graph needs to be constructed. Online detection is conducted on the sludge bulking of a sewage plant through the constructed robust self-adaptive canonical correlation analysis model, specifically, data collected by the sewage plant is standardized, robust decomposition is conducted on the standardized data matrix, and robust decomposition mainly utilizes canonical correlation analysis and a constructed robust criterion function, a corresponding robust T2 detection graph and a corresponding robust SPE detection graph are constructed respectively by utilizing an obtained robust load matrix and a characteristic value matrix, and finally a robust self-adaptivecontrol line is constructed, thereby realizing effective monitoring of sludge bulking of the sewage plant. According to the invention, effective early warning can be performed on tiny faults in the initial stage of sludge bulking.

Description

technical field [0001] The invention relates to the technical field of sewage plant sludge expansion detection, in particular to a sewage plant sludge expansion detection method based on robust adaptive canonical correlation analysis. Background technique [0002] Filamentous bacteria sludge bulking is a type of drift failure that often occurs in sewage plants. Unlike other types of failures, the bulking of filamentous bacteria sludge in real sewage plants often restores equilibrium after microbial self-regulation. What is more special is that the fault signal at the initial stage of sludge bulking is usually relatively weak, which is a typical type of small fault. The general monitoring method is difficult to give timely alarm at the initial stage of sludge bulking, and the continuous sludge bulking will eventually bring huge losses to the sewage plant. [0003] With the introduction of Industry 4.0, intelligent fault diagnosis technology has received more and more attent...

AI Technical Summary

Problems solved by technology

However, when faced with sludge bulking, this method often has a large number of false positives

Images