Research on the prediction method of SO3 in the emission flue gas of thermal power plant based on machine learning
Year: 2023
Type: preprint
Abstract: To reduce nitrogen oxide emissions, many thermal power plants have adopted Selective Catalytic Reduction (SCR) denitrification technology. However, during the SCR process, Due to the denitrification reaction process will produce a certain amount of ammonia escape, the actual production operation, denitrification flue gas in part of the SO2 in the role of the catalyst is converted to SO3, and difficult to monitor.This paper proposes a migratory learning model based on the VGG16 network to predict the concentration of sulfur trioxide (SO3) in the flue gas of power plants. The model was validated using flue gas data collected from experimental power plants and compared with classical machine learning models. The results show that the migratory learning model outperforms the comparison models, providing a reference for using machine learning techniques to predict air pollution from power plant emissions.
Source: Research Square
Authors: Weiting Jiang, Kai Wang, Tiancheng Ge, Lei Zhao, Weiguo Pan
Institution: Shanghai University of Electric Power
Language: English
Subfield: Mechanical Engineering
Field: Engineering
Domain: Physical Sciences
SDG: Climate action
Open Access status: green
OpenAlex ID: W4389664887