Enhance modelling predicting for pollution removal in wastewater treatment plants by using an adaptive neuro-fuzzy inference system

Authors : Hussein ALNAJJAR, Osman ÜÇÜNCÜ

Pages : 213-226

Doi:10.35208/ert.1106463

View : 16 | Download : 8

Publication Date : 2022-09-30

Article Type : Research Paper

Abstract :Biological and physical treatment in wastewater treatment plants appears to be one of the most important variables in water quality management and planning. This crucial characteristic, on the other hand, is difficult to quantify and takes a long time to obtain precise results. Scientists have sought to devise several solutions to address these issues. Artificial intelligence models are one technique to monitor the pollutant parameters more consistently and economically at treatment plants and regulate these pollution elements during processing. This study proposes using an adaptive network-based fuzzy inference system insert ignore into journalissuearticles values(ANFIS); model to regulate primary and biological wastewater treatment and used it to model the nonlinear interactions between influent pollutant factors and effluent variables in a wastewater treatment facility. Models for the prediction of removal efficiency of biological oxygen demand insert ignore into journalissuearticles values(BOD);, total nitrogen insert ignore into journalissuearticles values(TN);, total phosphorus insert ignore into journalissuearticles values(TP);, and total suspended solids insert ignore into journalissuearticles values(TSS); in a wastewater treatment plant were developed using ANFIS. Hydraulic retention time insert ignore into journalissuearticles values(HRT);, temperature insert ignore into journalissuearticles values(T);, and dissolved oxygen insert ignore into journalissuearticles values(DO); were input variables for BOD, TN, TP, and TSS models, as determined by linear correlation matrices between input and output variables. The findings reveal that the developed system is capable of accurately predicting and controlling outcomes. For BOD, TN, TP, and TSS, ANFIS was able to achieve minimum mean square errors of 0.1673, 0.0266, 0.0318, and 0.0523, respectively. The correlation coefficients for BOD, TN, TP, and TSS are all quite strong. In the wastewater treatment plant, ANFIS` prediction performance was satisfactory and the ANFIS model can be used to predict the efficiency of removing pollutants from wastewater.
Keywords : ANFIS, artificial neural networks, biological oxygen demand, biological treatment, hydrolic retention time, total suspended solids