Enhance and improve modelling prediction by using an adaptive neuro-fuzzy inference system-based model to predict pollution removal efficacy in wastewater treatment plants

Alnajjar H. Y. H., Üçüncü O.

Desalination and Water Treatment, vol.286, pp.52-63, 2023 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 286
  • Publication Date: 2023
  • Doi Number: 10.5004/dwt.2023.29320
  • Journal Name: Desalination and Water Treatment
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Aqualine, Aquatic Science & Fisheries Abstracts (ASFA), Biotechnology Research Abstracts, CAB Abstracts, Environment Index, Geobase, Pollution Abstracts, Veterinary Science Database, Civil Engineering Abstracts
  • Page Numbers: pp.52-63
  • Keywords: Adaptive network, Biological oxygen demand, Fuzzy inference, Neural networks, Total nitrogen, Wastewater treatment
  • Karadeniz Technical University Affiliated: Yes


An adaptive network-based fuzzy inference system (ANFIS) was used to create models for predicting the removal of biological oxygen demand (BOD), total nitrogen (TN), total phosphorus (TP), and total suspended solids (TSS) in a wastewater treatment plant treating process wastewaters. Temperature (T), hydraulic retention time, and dissolved oxygen were used as input variables for the BOD, TN, TP, and TSS models, using linear correlation matrices between input and output vari-ables. The results show that the created system has provided reasonable forecasting and control performance. The minimum root mean square errors of 1.4816, 1.9558, 0.2299 and 0.4733 for effluent BOD, TN, TP and TSS could be achieved using ANFIS. The maximum R-square values for BOD, TN, TP and TSS were 0.9137, 0.9204, 0.9865 and 0.9231, respectively. ANFIS’s architecture consists of both artificial neural networks and fuzzy logic including linguistic expression of membership functions and if–then rules, consequently it can overcome the limitations of traditional neural networks and increase the prediction performance.