Knowledge-Based Fuzzy Broad Learning Algorithm for Warning Membrane Fouling

Membrane fouling is a widespread problem that restricts the stable operation of membrane bioreactor (MBR) in wastewater treatment process (WWTP). However, it is difficult to avoid the occurrence of membrane fouling due to the lack of effective early warning methods. To deal with this problem, an intelligent early warning method, using a knowledge-based fuzzy broad learning (K-FBL) algorithm, is proposed for membrane fouling in this paper. First, the existing knowledge is extracted from the humanistic category of membrane fouling in the form of fuzzy rules. Then, the existing knowledge of membrane fouling can be used to compensate for the shortage of data sets. Second, a K-FBL algorithm is designed to train the fuzzy subsystems with the existing knowledge. Then, the uncertainties of membrane fouling process can be degraded to improve the learning performance. Third, a K-FBL-based early warning method is designed to realize the precise classification and provide the operational suggestions for membrane fouling. Finally, the experiment results of a real plant are given to demonstrate the effectiveness of this proposed K-FBL-based early warning method.

Automated summary

This paper proposes an intelligent early warning method for membrane fouling using a K-FBL algorithm, which extracts existing knowledge to compensate for the lack of data sets, trains fuzzy subsystems, reduces the uncertainties of membrane fouling process, and achieves precise classification and operational suggestions.