Accurately predicting pollutant emission factors (EFs) from woody biomass fuels remains challenging because small-scale combustion tests are fuel-specific, time-consuming, and highly sensitive to operating conditions. This study combines controlled laboratory combustion experiments with a physics-informed artificial neural network (ANN–PINN) to estimate the emission factors of particulate matter (${\mathrm{EF}}_{PM}$), carbon monoxide (${\mathrm{EF}}_{CO}$), and nitrogen oxides (${\mathrm{EF}}_{NOx}$) using only laboratory-scale fuel characterization. Three pelletized woody biomass, Pinus radiata, Acacia dealbata, and Nothofagus obliqua, were analyzed through ultimate and proximate composition, lignin content, and TGA-derived parameters and tested in a residential pellet stove under identical control setpoints, resulting in a narrow and well-defined operating regime. A medium-depth ANN–PINN was constructed by integrating mechanistic constraints, monotonicity based on known emission trends and a weak carbon balance penalty, into a feed-forward neural network trained and evaluated using Leave-One-Out Cross-Validation. The model accurately reproduced the experimental behavior of ${\mathrm{EF}}_{CO}$ and captured structured variability in ${\mathrm{EF}}_{PM}$, while the limited nitrogen variability of the fuels restricted generalization for ${\mathrm{EF}}_{NOx}$. Sensitivities derived via automatic differentiation revealed physically coherent relationships, demonstrating that PM emissions depend jointly on fuel chemistry and aero-thermal conditions, CO emissions are dominated by mixing and temperature, and ${\mathrm{NO}}_x$ formation is primarily governed by fuel-bound nitrogen. When applied to external biomass fuels characterized independently in the literature, the ANN–PINN produced physically plausible predictions, highlighting its potential as a rapid, low-cost screening tool for assessing new biomass feedstocks and supporting cleaner residential heating technologies. The integrated experimental–PINN framework provides a physically consistent and data-efficient alternative to classical empirical correlations and purely data-driven ANN models.

The global energy crisis drives the search for sustainable biomass resources. Microalgae, particularly Chlorella vulgaris, represent a promising third-generation feedstock for bi-ochar and biofuels. However, detailed kinetic schemes for its slow devolatilization are still scarce. This work compares the thermogravimetric behavior of commercially Chlorella vulgaris with data reported in the literature under identical experimental conditions and develops a multi-stage kinetic scheme using model-free methods and simultaneous global optimization. A complete set of kinetic parameter is provided in conjunction with a mass wights in order to close the reaction scheme. Biological composition of microalgae was experimentally determined resulting in 21.20, 59.30 and 19.50% for carbohydrates, protein and lipids. Thermogravimetric (TG/DTG) analyses were conducted with 5, 10 and 20 °C/min heating rates. Activation energy distribution was obtained through isoconversional model-free methods (Fried-man, FWO, KAS and Starink). A parallel multi-stage kinetic model was subsequently optimized globally against the experimental data to determine the complete kinetic tri-plet (E, A, n). TG/DTG profiles exhibited in general a good agreement with the literature refer-ence in the number and the temperature of features, peaks and shoulders, however different in intensity probably due to the different amount of biological components, carbohydrates, proteins and lipids. The multi-stage model achieved excellent fitting quality accounting for 5 reactions. Activation energies for the principal devolatilization stages ranged from 140 to 220 kJ/mol, while ln(A) values lay between 20 and 35 s⁻¹. The findings an results provided by this study is considered useful for the community con-tributing with discussion and a robust kinetic scheme suitable for example for slow pyrolysis process simulation.