Boiler Furnace Temperature Field Measurement and Reconstruction Error Elimination Based on Temperature Field Residual Correction Network
Yixin Duan,Liwei Chen,Xinzhi Zhou,Youan Shi,Nan Wu
DOI: https://doi.org/10.1109/tim.2024.3353873
IF: 5.6
2024-01-01
IEEE Transactions on Instrumentation and Measurement
Abstract:The temperature of the boiler furnace is an indicator of the combustion conditions within the boiler. Understanding the temperature distribution inside the boiler is crucial for enhancing combustion efficiency, detecting faults, and reducing pollutant emissions. However, traditional algorithms used for measuring and reconstructing the temperature field often introduce systematic errors that are challenging to eliminate. These errors accumulate step by step, hampering further improvements in accuracy. To address this issue, this article proposes a novel approach. First, an acoustic temperature field reconstruction simulation dataset (ATFRSD) is constructed. This dataset facilitates the accurate representation of temperature distribution in the target field. Additionally, a temperature field residual correction network (TRCN) is introduced. The TRCN has been extensively tested through simulation experiments and analysis of real engineering data. The core advantage of the TRCN is its ability to effectively eliminate errors while preserving the smoothness of the temperature field. Moreover, it demonstrates robustness and can be integrated with various traditional algorithms. By employing the TRCN, the reconstruction results more accurately reflect the temperature distribution in the measured field. This contributes to improved combustion efficiency, fault detection, and reduced emission of pollutants. The dataset and code are publicly available at: https://github.com/potatocell/TRCN.git.
engineering, electrical & electronic,instruments & instrumentation