Modeling The Toxicity of Textile Industry Wastewater Using Artificial Neural Networks

Abstract

Toxicity tests are required to detect the possible effects of pollutants on organisms. This study investigates the effect of Chemical Oxygen Demand (COD), suspended solid (SS) and pH parameters on toxicity of textile industry wastewaters except for the color parameter, effect of which is well known. Fish bioassay taking place in legal regulation of Turkey was used as toxicity test. At the end of the toxicity test, various values of the parameters were predicted through Artificial Neural Networks (ANN). In addition, Artificial Neural Networks were used to calculate the effect of each parameter on toxicity (%). Accordingly, COD is the parameter which mostly affects toxicity following color parameter and SS is the parameter which has the minimum effect. It is found that results deviate at the rate of 15.41% when values of COD parameter are excluded from the model input data and the error rate becomes 5.07% when SS parameter is excluded. In this study, the effect of each input of each parameter, which is an open ecosystem, based on selected parameters is successfully predicted through Artificial Neural Networks which is a heuristic method.