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.

膜污染是一个普遍存在的问题，它限制了膜生物反应器（MBR）在废水处理过程（WWTP）中的稳定运行。但是，由于缺乏有效的预警方法，很难避免发生膜污染。针对这一问题，本文提出了一种基于知识的模糊广义学习（K-FBL）算法的智能预警方法。首先，以模糊规则的形式从人为的膜污染类别中提取现有知识。然后，可以使用膜污染的现有知识来弥补数据集的不足。其次，设计了一种K-FBL算法，以利用现有知识训练模糊子系统。然后，膜污染过程的不确定性可以降低，以提高学习性能。第三，设计了一种基于K-FBL的预警方法，以实现精确的分类并为膜污染提供操作建议。最后，给出了真实植物的实验结果，以证明该基于K-FBL的预警方法的有效性。