Modelling COD concentration by using different artificial intelligence methods

Abstract

The determination of water quality is traditionally based on classification by considering physical, chemical, biological factors and heavy metals parameters within the context of the national or international standards according to the water usage purposes. More clearly, analysis of water quality involves some parameters such as dissolved oxygen (DO), pH, electrical conductivity (EC), temperature (T), biological oxygen demand (BOD), chemical oxygen demand (COD), chloride, total phosphate, nitrite, nitrate, ammonia, ions, heavy metals, total salt concentration, fecal coliform and so on. In this study, two different ANN methods, that is, the multi-layer perceptron (MLP) and radial basis neural network (RBNN) and the integrated fuzzy clustering and adaptive neuro fuzzy inference system (ANFIS-FCM) methods were developed to estimate COD concentration by using various combinations of daily input important variables water suspended solids (SS), Q, T and pH. The root mean square error (RMSE), mean absolute error (MAE) and determination coefficient (R²) statistics were used for the comparison criteria. Comparison of the results indicated that the MLP and RBNN performed slightly better than the ANFIS-FCM method in modelling COD. The MLP model, with one input water SS, was found to be the best model in estimation of COD according to the RMSE, MAE and R² criteria. However, it can be said that these techniques provide similar accuracy in estimating COD concentration.