Optimization of biomass devolatilization stage component prediction models and their impact on civil boiler simulations

Biomass combustion is a pivotal carbon–neutral technology whose efficient conversion relies on accurate combustion modeling. However, the existing model has limitations in that it assumes hydrogen content is zero and requires devolatilization temperatures > 600 K. This study selects biomass fuel with a 589.4 K devolatilization temperature and proposes a framework for volatile prediction based on dynamic carbon-hydrogen–oxygen conservation. Thermogravimetric (TG) experiments calibrated the devolatilization temperature and char oxidation temperature. The mass and energy conservation equations were modified, and the Monte Carlo algorithm was used to optimize the distribution of the components in order to eliminate negative solutions in conventional models. Thermogravimetric-mass spectrometry (TG-MS) experiments show prediction errors for H 2 O, CO, and CO 2 reduced by 15.88 %, 10.12 %, and 3.54 % compared to existing models. Computational Fluid Dynamics (CFD) simulations demonstrate the modified model’s accuracy in temperature field prediction, with a maximum reduction in root mean square error (RMSE) of 48.9 %. Further analysis reveals CO oxidation as the primary exothermic mechanism in high-temperature zones, while H 2 O’s dual heat absorption/exothermic effects substantially impact temperature distribution. The air supply parameter directly affects the combustion efficiency and stability by regulating the residence time of combustible gases and mixing efficiency. The framework combines high-precision simulation and operational adaptability, providing theoretical support for designing clean biomass boiler systems.