Biomass, biochar and hard coal: Data mining application to elemental composition and high heating values prediction

• Created predictive model for the determination, with good precision, of the heat of combustion and elemental composition on of a sample based on easy-to-perform chemical analyses.
• We developed an algorithm for different materials, which can predict values of the heat of combustion and elemental composition regardless of the origin of the material.
• Model giving the best results in terms of the prediction of carbon, hydrogen and oxygen content.
• In the research have been used method of standard regression tree model.

This paper investigates several methods of regression analysis to find relationships and develop predictive models that allow accurate calculation of the heat of combustion and content of elements from easy to perform chemical analyses. A selection of coals, biomass and bio-char were used for analysis. Bio-char can be produced from biomass through various processes and from various types of biomass. Chars from biomass were selected due to the increasing interest in this type of material as a solid carbon carrier that can be replaced with clean biomass for co-firing with coal. Bio-char is a potential alternative and environmentally friendly source of energy. The predictive models developed in this study have been validated on an independent data set using cross-validation, providing information about the accuracy of the obtained relations.