|
| Real-time Prediction of Rare Earth Molten Salt Electrolyzer Furnace Surface Temperature based on YOLOV8 Algorithm |
| Received:June 17, 2024 Revised:July 02, 2024 |
| View Full Text View/Add Comment Download reader |
| DOI：doi：10.20237/j.issn.1007-7545.2025.01.012 |
| KeyWord:Rare earth molten salt electrolyzer；Temperature detection; YOLOv8; Target classification |
| Author | Institution |
| HOU Wei |
江西理工大学 信息工程学院 |
| HUANG Jindi |
江西理工大学 冶金工程学院 |
| LI Mingzhou |
江西理工大学 冶金工程学院 |
| LI Jing |
江西理工大学 冶金工程学院 |
|
| Hits: 633 |
| Download times: 1314 |
| Abstract: |
| The electrolytic temperature is closely related to the current efficiency of the rare earth molten salt electrolyzer and the life of the furnace body. However, thermocouple temperature measurement, infrared thermal imaging, and other measurement methods are affected by the high temperature and strong corrosive environment in the electrolysis workshop, making it difficult to detect in real time. This paper predicts the furnace surface temperature of molten salt electrolyzer based on the YOLOv8 algorithm. First, the temperature interval classification model is obtained through the self-made temperature data set of the high-temperature experimental furnace and trained based on the YOLOv8 algorithm. Secondly, the temperature cloud map of the furnace surface image is reconstructed by using the relationship between image gray scale and temperature, Finally, based on YOLOv8-SSW algorithm, the image recognition model of furnace surface temperature was constructed, and its prediction accuracy was 93.4%, which could be used to monitor the furnace surface temperature of electrolytic cell. |
| Close |
|
|
|