Calorific value prediction models of processed refuse derived fuel 3 using ultimate analysis

Abstract

Various models have been developed to predict the calorific value of Biomass but only a few models exist to predict this measure for the urban waste like Refuse Derived Fuel (RDF). In this paper, new models are introduced to predict the calorific value of RDF, as more advanced studies are required to be conducted with a focus on a distinct group of RDFs for validating the robustness of the models in the existing literature. The calorific value based on ultimate (elemental) analysis considers the contents of C, H, N, S, and O elements in RDF. Using empirical and machine learning methods, the newly established models accurately predicted the calorific value of the samples provided by a local municipality situated in Edmonton, Alberta, Canada. Furthermore, these new models demonstrated a lower bias and average absolute error than the other twelve previously published models pertinent to RDF material. Based on the established workflow the ultimate analysis-based models gave a higher coefficient of determination (R²) value in the range 0.78 − 0.80, indicating that the developed model improves the prediction of calorific value for RDF. The newly developed machine-learning models showed better results than the empirical models developed in this study implying that complex correlations can be dealt effectively while predicting calorific values for RDF.

Keywords:

• Higher heating value
• ultimate analysis
• refused derive fuel
• American Society for Testing and Materials

Acknowledgments

The authors gratefully acknowledge the support of the Edmonton Waste Management Center for providing resources for this project.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The authors declare that the data supporting the findings of this study are available within the article and or its supplementary material.