مدلسازی تصفیه فاضلاب لبنی بهروش انعقاد الکتریکی با استفاده از سامانه استنتاج فازي– عصبي تطبيقي

Abstract

Introduction: Forecasting of the wastewater quality parameters has great importance in modern wastewater treatment methods. One of the main problems in predicting the efficacy of a wastewater treatment is the complexity of physicochemical properties of raw sewage and data differences for different reasons. Modeling of wastewater treatment using Adaptive Neural Fuzzy Inference System (ANFIS)- can help to improve wastewater quality control process. The aim of the study was to modeling of treatment of dairy wastewaters by electrocoagulation process using ANFIS. Methods: In this study, ANFIS was used to estimate the chemical coagulation of dairy wastewater treatment. The input parameters to the ANFIS model were time, voltage, total suspended solids and boichemical oxygen demand and the output was chemical oxygen demand removal efficiency. Also the membership functions, the number of membership functions and number of learning cycles (Epochs) were used for optimization of different models by trial and error. Results: The best model was assessed by bell-shaped membership functions with number of membership functions as 3 3 3 3 3 and 300 epochs of training with lowest mean square error(MES) and the best coefficient of determination (R2). The coefficient of determination and MSE of the best ANFIS model were 0.9912 and 0.012, respectively. Conclusion: Analysis of the model revealed that the ANFIS is a powerful tool to predict the dairy wastewater treatment using electrical coagulation.