ABSTRACT
The heating/calorific value of municipal solid waste (MSW) is essential in selecting or designing the appropriate waste to energy (WTE) systems. Experimental evaluation of the heating value of solid fuels is labor intensive, costly, and subject to experimental errors. Different models have been established to predict the high heating values of MSW and other solid fuels, from the ultimate analysis. However, the reliability of OLS estimator used in the linear regression model depends on the non-violation of assumptions that include independency of the predictor variables and normality of the error term. In this study, a new technique of robust estimators is employed to solve the problem of non-normality and dependency of the predictor variables in the linear regression model. The Robust ridge, robust Liu and robust K-L estimators were applied to mitigate the problems of multicollinearity and non-normality in the linear regression model. Eight (8) models were developed, and the adequacies were evaluated using the coefficient of determination (R2), adjusted R2, Akaike criterion (AIC), the mean squared error and the Schwarz criterion (SBIC). The eighth model is considered as the best because it has the highest adjusted R2 (0.9710), the least mean squared error (1.9564), minimum AIC (133.2755) and SBIC (145.9437). The selected model with the robust K-L estimator is finally used to predict the high heating/calorific value of the ultimate analysis.
Acknowledgments
The authors appreciate the Technologists in the Laboratories of the International Institute of Tropical Agriculture (IITA) Ibadan, for their support during the laboratory analyses of the wastes samples. We equally thank the staff of Kwara Environmental Protection Agency for their support during Characterization of the wastes on the dumpsite.
Nomenclature
Adjusted R2: Adjusted coefficient of determination
HHV: High heating value
HV: Heating value
ICP-spec: Inductively Coupled Plasma – Optical Emission Spectrometer
JB: Jarque-Bera Test
KLestimator: Kibria-Lukman estimator
LHV: Low heating value
MSE: Mean Squared error
MSW: Municipal solid waste
OLS: Ordinary Least Squared
R2: Coefficient of determination
SBIC: Schwarz information criterion
AIC: Akaike criterion
SE: Standard error
SMSE: Scalar Mean Squared error
VIF: variance inflation factor
WTE: Waste to energy