PREDICTION OF HEAT ENERGY FOR ELECTRICITY ESTIMATION ON MUNICIPAL SOLID WASTE USING ARTIFICIAL NEURAL NETWORK

Abstract

This work is aimed at prediction of the heat energy from municipal solid waste properties and its potential for electricity generation using artificial neural network statistical tool. Random truck sampling was used according to American Society for Testing and Materials (ASTM) in the collection of waste to the disposal site and characterized into four (4) parameters; food waste, plastic waste, wood waste, and cotton waste, A sample of the waste was measured for experimental analysis to determine the range of input parameters. Statistical design of experiment (DOE) using central composite design (CCD) matrix version (13.0.5.0) was employed to predict the heating value of each input parameters that will minimize the rate of solid waste disposal and generate heat energy
responses using Artificial Neural Network model. Reliability was produced to test the networks adequacy. The central composite design (CCD) matrix predicted heat energy value of 26,102.6kJ/kg. A regression plot showing the correlation between the input and output was produced with R2 values of 97% for the training, 87% for the validation, 98% for testing and 97% for the overall. A reliability plot of 81.7% was obtained for artificial neural network (ANN). The study established the heating values for electricity potential of the municipal solid waste components in the area. The results showed that recovery heat energy for electricity estimation s a feasible option as part of an integrated municipal solid waste management plan in Nigeria.