APPLICATION OF AN ADAPTIVE NEURO-FUZZY INFERENCE SYSTEM TO CONTROL THE WASTEWATER TREATMENT PROCESS

This study explores the application of an Adaptive Neuro-Fuzzy Inference System (ANFIS) for controlling wastewater treatment processes using ion-exchange resins. It addresses the critical challenges of water scarcity and pollution by enhancing the regulation of water hardness (H) and Total Dissolved Solids (TDS). Using a pilot laboratory device and experimental data from the mixed wastewater of the Kungrad Soda Plant in Uzbekistan, an ANFIS model was developed in MATLAB to automate process control. The model leverages water hardness and TDS as input parameters to regulate the water flow rate by servo valve opening degree, ensuring precise and efficient treatment. Compared to fuzzy logic models, ANFIS achieves a 3–4 second reduction in settling time and significantly lowers error rates. This enhanced control not only optimizes water hardness and TDS regulation but also ensures greater consistency and efficiency. Experimental results indicate that adopting ANFIS for real-time parameter adjustments can reduce energy consumption by up to 1.3%. These findings highlight the potential of ANFIS as a superior solution for controlling dynamic and nonlinear wastewater treatment processes.