Co-Combustion Behavior of Coal Gangue and Corn Straw: Kinetics, Gaseous Product Evolution, and SHAP-Interpreted ANN Modeling

Abstract Coal gangue (CG) has limited fuel value because of its low carbon and high ash contents, while cocombustion with biomass may improve its thermal utilization. This study investigated the cocombustion of CG and corn straw (CS) using thermogravimetry/derivative thermogravimetry (TG/DTG), TG coupled with Fourier transform infrared spectroscopy (TG–FTIR), isoconversional kinetic analysis, and an artificial neural network (ANN) interpreted using Shapley additive explanations (SHAP). CS addition improved ignition and overall combustion performance. CG1CS3 exhibited a pronounced synergistic effect above approximately 480 °C and had the lowest mean apparent activation energy (131.7 kJ mol–1), which was 26.8% lower than that of CG (179.8 kJ mol–1). The main FTIR signals occurred between 200 and 600 °C. CS addition strengthened volatile-related bands at 300–450 °C, shifted the NO-related response toward lower temperatures, and weakened high-temperature signals tentatively assigned to CO, NO, and SO2. The ANN was developed using the 10 and 20 °C min–1 data sets and evaluated against the held-out DTG curves at 15 °C min–1, yielding R2 = 0.8566. SHAP identified temperature as the dominant predictor and indicated stage-dependent feature effects. These results support CS blending as a means of improving CG combustion and show that the ANN can interpolate DTG behavior within the investigated compositions and heating-rate range.