UTILIZING NEURO FUZZY LOGIC FOR THE AUTOMATION OF COAGULANT DOSING IN WATER TREATMENT SYSTEMS

Abstract

This thesis explores the application of Adaptive Neuro-Fuzzy Inference System (ANFIS) for optimizing coagulant dosing in water treatment systems. Utilizing a dataset with variables including temperature, pH, Total Dissolved Solids (TDS), and turbidity, the study compares different ANFIS models to identify the most effective configuration. The four-variable model, incorporating all the mentioned parameters, demonstrated the best performance with a Root Mean Square Error (RMSE) of 0.320207 for training data and 0.272577 for testing data. Results showed that models with multiple input variables enhance prediction accuracy, with R² values improving from 0.047 for temperature alone to 0.43 for the comprehensive model. The findings highlight the model's capability to predict coagulant dosage, thereby enhancing the automation and efficiency of water treatment processes. Continuous monitoring and regular updates of the model are recommended to adapt to changing water quality. Further research is suggested to explore additional water quality parameters and integrate the ANFIS model with advanced control systems for improved real-time monitoring and management. This study concludes that ANFIS is a reliable tool for optimizing coagulant dosing, contributing to more efficient and consistent water treatment operations.