• Roman M. Bachurin, University of Tyumen (Tyumen, Russia)
• Irina G. Zakharova, University of Tyumen (Tyumen, Russia)

The article presents the results of a study aimed at the development and software implementation of methods for mathematical modeling of the plant biomass pyrolysis process. The main objective of the study is to create an intelligent system for analyzing experimental data, which allows us to study the technological parameters of the pyrolysis process. The key components of the developed system include a set of machine learning models (model quality metrics values RMSE < 3.9, R² > 0.8) for predicting the yield of final pyrolysis products; tools for analyzing differential scanning calorimetry (DSC) curves and assessing the thermal effect of reactions; a subsystem for identifying and visualizing thermogravimetric (TGA) and DSC curves. The predictive models were trained on a sample (750 records) compiled from open datasets on full-scale experiments on the pyrolysis of plant raw materials. The target variables were the percentage content of solids, liquids, and gases in the final pyrolysis products. Independent variables included the physicochemical characteristics of feedstock and the parameters of the pyrolysis process. The practical significance of this study lies in its potential for a deeper understanding of biomass decomposition processes. The developed system provides researchers and technologists with comprehensive tools for analyzing and identifying DSC curves, facilitating the interpretation of experimental data. The scientific novelty of this work lies in the creation of a unified research support platform based on the integration of machine learning methods with traditional approaches to pyrolysis process analysis. This significantly improves forecasting accuracy and optimizes the pyrolysis process for obtaining end products. The results of this study can be applied in research in the field of thermochemical processes, applied bioenergy, the chemical industry, and other areas related to biomass processing by pyrolysis. Future development of the system involves expanding the DSC curve database, improving machine learning algorithms, and integrating additional experimental data analysis methods for a detailed interpretation of the pyrolysis process, taking into account its dynamics.

biomass pyrolysis, mathematical modeling, machine learning, neural networks, differential scanning calorimetry

2026-03-05