Prediction of the heating value of municipal solid waste: a case study of the city of Johannesburg

In this study, a municipality-based model was developed for predicting the Lower heating value (LHV) of waste, which is capable of overcoming the demerit of generalised model in capturing the peculiarity and characteristics of waste generated locally. The city of Johannesburg was used as a case study. Artificial Neural Network (ANN) and Adaptive Neuro Fuzzy-Inference System (ANFIS) models were developed using the percentage composition of waste streams such as paper, plastics, organic, textile and glass as input variables and LHV as the output variable. The ANFIS model used three clustering techniques, namely Grid Partitioning (ANFIS-GP), Fuzzy C-means (ANFIS-FCM) and Subtractive Clustering (ANFIS-SC). ANN architectures with a range of 1–30 neurons in a single hidden layer were tested with three training algorithms and activation functions. The GP-clustered ANFIS model (ANFIS-GP) outperformed all other models with root mean square error (RMSE), mean absolute deviation (MAD), and mean absolute percentage error (MAPE) values of 0.1944, 0.1389 and 4.2982, respectively. Based on the result of this study, a GP-clustered ANFIS model is viable and recommended for predicting LHV of waste in a municipality.