Co-combustion performances of biomass pyrolysis semi-coke and rapeseed cake: PCA, 2D-COS and full range prediction of M-DAEM via machine learning

Biomass pyrolysis semi-coke (PC) is a refractory byproduct of biomass refinery system, and its efficient downstream disposal is of significance to improve the bioenergy utilization efficiency. Herein, we proposed an environmentally friendly co-combustion strategy to explore the synergistic conversion of PC with agricultural waste rapeseed cake (RC). The co-combustion interaction, kinetics, prediction and emission responses were determined through principal component analysis (PCA), multiple distributed activation energy model (M-DAEM), artificial neural network (ANN) and two-dimensional correlation spectroscopy (2D-COS) analysis. The results indicated that the P/R (mixed) ratio in 1:1 strengthened the reaction rate for main incineration stage, whereas lower P/R ratio advanced the peak temperature and cooperated with temperature range in 730–790 K to facilitate co-combustion synergies. The prediction of the kinetic distribution under all mixed ratios was successfully conducted (R2 = 0.999863) through M-DAEM coupling with ANN, from which the activation energy distribution centers E0 and standard deviation σ were in the ranges of 115.23–270.84 kJ/mol and 2.89–42.11 kJ/mol, respectively. Co-combustion resulted in centralized activation energy distribution, and successfully lowered the reaction energy barriers while augmenting the instantaneous energy release. Meanwhile, the temperature dependency responses of flue gas were varied significantly as a function of mixed ratios and temperatures.