Artificial Neural Network Based Control of Electrocoagulation based Automobile Wastewater Treatment Plant

Abstract

Wastewater from vehicle garages and workshops is an important contributor to water pollution. Oil is the major content of wastewater in vehicle garages. The project focuses on the use of Electrocoagulation technique (EC) for the removal of oil content in wastewater from vehicle garages. We took samples from KSRTC Thrissur depot since they use more than 10000 litres of water per day in a continuous manner and this huge amount of water is wasted. The different parameters affecting the EC process will be quickly varying. Hence a non linear model of the process is required for further automation and control of input parameters for the EC process. Artificial Neural Network (ANN) technique is used for the non linear modeling purpose. An ANN model is developed relating important parameters affecting the Eletrocoagulation and the oil removal. The removal of oil is observed in terms of Chemical Oxygen Demand of experiment feed and water sample after Electrocoagulation.. The parameters are Current Density, time of EC, salt concentration and pH of the sample. The combination of inputs is designed by Design of Experiment tool in MINITAB software. The percentage COD removal is predicted using ANN. The Regression Coefficient is obtained in the range of 0.8-0.9 by ANN model and comparison of ANN predicted COD removal and experimental removal has also shown closed result. We concluded that EC can give about 90 % removal of oil in terms of COD and the ANN can predict percentage removal of oil. Hence in practice the adjustment of operating parameters will result in greater removal of oil content in wastewater and to allow automation of the Electrocoagulation process