Decision-making for Membrane Fouling Based on Knowledge Fuzzy Transfer in Municipal Wastewater Treatment

A decision-making method, based on a knowledge fuzzy transfer, was proposed to solve the problem of membrane fouling in wastewater treatment process. First, based on the data and experience collected from a real wastewater treatment process, the knowledge of membrane fouling decision-making was expressed in the form of fuzzy rules. Second, a knowledge reconstruction mechanism (KRM) was proposed to complete the knowledge reconstruction by balancing the matching accuracy and diversity between the source domain and the target domain with knowledge transfer method. Finally, a decision-making model based on data-knowledge-driven interval type-2 fuzzy neural network (DK-IT2FNN) was developed, the model parameters were designed by using fuzzy rules, and the transfer gradient descent algorithm was proposed to adjust the model weight. The decision-making accuracy has been improved. Results show that the proposed method can realize the decision of membrane fouling with high accuracy.