Application of Gradient Boosting and Digital Simulation for Burner Optimization

Machine learning algorithms are currently actively employed in the combustion field, streamlining the analysis of extensive combustion data derived from experiments or simulations across various space-time parameters. This facilitates the identification of latent patterns inherent in the data. Achieving optimal combustion processes involves stabilizing the flame and minimizing emissions of various harmful gases, including greenhouse gases. This study explores the application of a machine learning approach to classify the operational modes of a gas-oil burner by utilising experimental data obtained from a digital model within a Computational Fluid Dynamics (CFD) system. The machine learning techniques proposed in this investigation aim to advance technologies that ensure optimal operational conditions while mitigating the release of harmful substances. Specifically, a boosting model was selected and subsequently trained in this study.