Image segmentation and CNN-based deep learning architectures for the modelling on particulate matter formation during solid fuels combustion

Abstract

Three typical solid fuels (coal, biomass, and refuse-derived fuel) were individually combusted in a lab-scale drop tube furnace. The computer Vision method was employed to extract the morphological characteristics of particulate matter (PM) and establish a dataset of 12,637 particle features. Six convolutional neural network models were developed, and three transfer learning strategies were studied. The ResNet50 model achieved a peak accuracy of 96.6 % when fine-tuned across all layers, demonstrating its exceptional capability to identify irregular, angular, agglomerated, and completely melted particles. Based on this model, the predominant form of PM₁₀ produced from all three fuels was irregular, ranging from 28.91 to 81.37 wt%, whereas PM_10–200 consisted primarily of 9.92 to 49.44 wt% angular, 5.10 to 39.59 wt% agglomerated, and 0.08 to 39.06 wt% completely melted forms. By combining combustion experiments and thermodynamic equilibrium calculations, it was proven that Si, Na, K, and Cl form irregular particles as the major types of PM₁₀. SiAl readily forms ‘angular’ PM_10–200, whereas collisions with fine particles in the gas phase leads to the formation of agglomerated particles of Na/K-Al-Si. Ca/Fe-Al-Si formed completely melted particles of 59.06 wt% in total.