The Use of Artificial Neural Network (ANN) for Modeling of Ammonia Nitrogen Removal from Landfill Leachate by the Ultrasonic Process

Background: The study examined the implementation of artificial neural network (ANN) for the prediction of Ammonia nitrogen removal from landfill leachate by ultrasonic process.Methods: A three-layer backpropagation neural network was optimized to predict Ammonia nitrogen removal from landfill leachate by ultrasonic process. Considering the smallest mean square error (MSE), The configuration of the backpropagation neural network was three-layer ANN with tangent sigmoid transfer function (Tansig) at hidden layer with ۱۴ neurons, linear transfer function (Purelin) at output layer and Levenberg–Marquardt backpropagation training algorithm (LMA).Results: ANN predicted results were very close to the experimental results with correlation coefficient (R۲) of ۰.۹۹۳ and MSE ۰.۰۰۰۳۳۴. The sensitivity analysis showed that all studied variables (Contact time, ultrasound frequency and power and pH) had strong effect on Ammonia nitrogen removal. In addition, pH was the most influential parameter with relative importance of ۴۴.۹%.Conclusions: The results showed that neural network modeling could effectively predict Ammonia nitrogen removal from landfill leachate by ultrasonic process.Background: The study examined the implementation of artificial neural network (ANN) for the prediction of Ammonia nitrogen removal from landfill leachate by ultrasonic process. Methods: A three-layer backpropagation neural network was optimized to predict Ammonia nitrogen removal from landfill leachate by ultrasonic process. Considering the smallest mean square error (MSE), The configuration of the backpropagation neural network was three-layer ANN with tangent sigmoid transfer function (Tansig) at hidden layer with ۱۴ neurons, linear transfer function (Purelin) at output layer and Levenberg–Marquardt backpropagation training algorithm (LMA). Results: ANN predicted results were very close to the experimental results with correlation coefficient (R۲) of ۰.۹۹۳ and MSE ۰.۰۰۰۳۳۴. The sensitivity analysis showed that all studied variables (Contact time, ultrasound frequency and power and pH) had strong effect on Ammonia nitrogen removal. In addition, pH was the most influential parameter with relative importance of ۴۴.۹%. Conclusions: The results showed that neural network modeling could effectively predict Ammonia nitrogen removal from landfill leachate by ultrasonic process.