Modeling Leachate Generation Using Artificial Neural Networks

In this study, a neural network model is proposed for modeling leachate flow-rate in a municipal solid waste landfill site. After training, the neural network model predicts leachate generation based on meteorological data and leachate characteristics. Parameters such as pH, temperature, conductivity and meteorological data were used as input data. To validate the proposed method, a case study was carried out based on the data obtained from city of Beirut landfill site. While waste disposal at the site started in October 1997, measuring leachate generation rates was not initiated until April 1998. The Levenberg-Marquardt algorithm was selected as the best of thirteen backpropagation algorithms. The optimal neuron number for Levenberg-Marquardt algorithm is 10. The performance of modeling was determined. According to the statistical performance indices (R=0.976, ARE=0.089), the results of the forecast model were in good agreement with measured data.