Image segmentation and CNN-based deep learning architectures for the modelling on particulate matter formation during solid fuels combustion
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 co...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-03-01
|
| Series: | Fuel Processing Technology |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S0378382024001462 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841528022098771968 |
|---|---|
| author | Yanchi Jiang Lanting Zhuo Xiaojiang Wu Zhongxiao Zhang Xinwei Guo Wei Wang Cunjiang Fan |
| author_facet | Yanchi Jiang Lanting Zhuo Xiaojiang Wu Zhongxiao Zhang Xinwei Guo Wei Wang Cunjiang Fan |
| author_sort | Yanchi Jiang |
| collection | DOAJ |
| description | 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 PM10–200 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’ PM10–200, 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. |
| format | Article |
| id | doaj-art-b9456c86e838494bbb03306c7bf5aa53 |
| institution | Kabale University |
| issn | 0378-3820 |
| language | English |
| publishDate | 2025-03-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Fuel Processing Technology |
| spelling | doaj-art-b9456c86e838494bbb03306c7bf5aa532025-01-15T04:11:33ZengElsevierFuel Processing Technology0378-38202025-03-01267108176Image segmentation and CNN-based deep learning architectures for the modelling on particulate matter formation during solid fuels combustionYanchi Jiang0Lanting Zhuo1Xiaojiang Wu2Zhongxiao Zhang3Xinwei Guo4Wei Wang5Cunjiang Fan6College of Smart Energy, Shanghai Jiao Tong University, Shanghai 200240, China; Corresponding author at: College of Smart Energy, Shanghai Jiao Tong University, Shanghai 200240, China.College of Smart Energy, Shanghai Jiao Tong University, Shanghai 200240, ChinaSchool of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai 200240, ChinaSchool of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai 200240, ChinaSchool of Energy and Power Engineering, North China University of Water Resources and Electric Power, Zhengzhou 450045, ChinaSchool of Energy and Power Engineering, North China University of Water Resources and Electric Power, Zhengzhou 450045, ChinaSchool of Energy and Power Engineering, North China University of Water Resources and Electric Power, Zhengzhou 450045, ChinaThree 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 PM10–200 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’ PM10–200, 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.http://www.sciencedirect.com/science/article/pii/S0378382024001462Deep learningPM formationMorphology transformationObject detectionTransfer learning |
| spellingShingle | Yanchi Jiang Lanting Zhuo Xiaojiang Wu Zhongxiao Zhang Xinwei Guo Wei Wang Cunjiang Fan Image segmentation and CNN-based deep learning architectures for the modelling on particulate matter formation during solid fuels combustion Fuel Processing Technology Deep learning PM formation Morphology transformation Object detection Transfer learning |
| title | Image segmentation and CNN-based deep learning architectures for the modelling on particulate matter formation during solid fuels combustion |
| title_full | Image segmentation and CNN-based deep learning architectures for the modelling on particulate matter formation during solid fuels combustion |
| title_fullStr | Image segmentation and CNN-based deep learning architectures for the modelling on particulate matter formation during solid fuels combustion |
| title_full_unstemmed | Image segmentation and CNN-based deep learning architectures for the modelling on particulate matter formation during solid fuels combustion |
| title_short | Image segmentation and CNN-based deep learning architectures for the modelling on particulate matter formation during solid fuels combustion |
| title_sort | image segmentation and cnn based deep learning architectures for the modelling on particulate matter formation during solid fuels combustion |
| topic | Deep learning PM formation Morphology transformation Object detection Transfer learning |
| url | http://www.sciencedirect.com/science/article/pii/S0378382024001462 |
| work_keys_str_mv | AT yanchijiang imagesegmentationandcnnbaseddeeplearningarchitecturesforthemodellingonparticulatematterformationduringsolidfuelscombustion AT lantingzhuo imagesegmentationandcnnbaseddeeplearningarchitecturesforthemodellingonparticulatematterformationduringsolidfuelscombustion AT xiaojiangwu imagesegmentationandcnnbaseddeeplearningarchitecturesforthemodellingonparticulatematterformationduringsolidfuelscombustion AT zhongxiaozhang imagesegmentationandcnnbaseddeeplearningarchitecturesforthemodellingonparticulatematterformationduringsolidfuelscombustion AT xinweiguo imagesegmentationandcnnbaseddeeplearningarchitecturesforthemodellingonparticulatematterformationduringsolidfuelscombustion AT weiwang imagesegmentationandcnnbaseddeeplearningarchitecturesforthemodellingonparticulatematterformationduringsolidfuelscombustion AT cunjiangfan imagesegmentationandcnnbaseddeeplearningarchitecturesforthemodellingonparticulatematterformationduringsolidfuelscombustion |