Innovative carbon emission estimation strategy for biomass materials using near-infrared (NIR) spectrum

Abstract

The world’s biomass resources amount to around 170.00 billion tons currently. In contrast, global biomass energy consumption only accounts for about 14 % of the world’s total energy consumption. Direct combustion of biomass raw materials and production of bio-natural gas are the two main methods for biomass utilization. In this study, 99 samples of 4 types of biomass materials from eight provinces of China were collected to determine the heat production efficiency of these two methods. Chemometric algorithms and deep learning (DL) were applied to connect the calorific value and methane yield (MY) of materials with their near-infrared (NIR) wavelength region. As a result, a model was proposed for predicting biomass heat release, which could be applied to the rapid evaluation of the calorific value utilization and MY of biomass resources. Meanwhile, the results showed that the GWP caused by direct combustion of straw and manure was estimated to be 1.8–3.1 times that of biogas in 20–500 years. The conversion of raw materials into biogas could significantly reduce greenhouse gas emissions relative to direct combustion. The NIR model by DL achieved excellent evaluation results to accurately predict calorific value and MY for the biomass feedstocks with determination coefficient and residual prediction deviation being 0.97 and 4.19, respectively.