Search citation statements
Paper Sections
Citation Types
Year Published
Publication Types
Relationship
Authors
Journals
The superheater in the boiler is the key of equipment connecting high-temperature steam to the turbine for power generation. At present, the problems of large variable fluctuations, strong timing coupling, and multi-power plant data utilization prevent the temperature, flow, and pressure prediction of the boiler superheater. In this paper, a method for predicting the parameters of boiler superheater based on a transfer learning model is proposed, which realizes the joint utilization of data from multiple power plants. The method first collects data from a waste incineration boiler power plant for pre-training the long short-term memory (LSTM)-transformer model, and then completes the transfer learning training on the new power plant. The proposed method has the advantages of high prediction accuracy, good robustness, and more reliable location prediction with drastic changes. The predictions on the test set are within ± 5% of the experimental value. Compared with the model not trained by the transfer learning, the proposed method achieves the lowest relative errors for all prediction intervals in the 3-15 min range. Compared to the linear regression (LR), support vector regression (SVR), and random forest (RF), the proposed method improves the average absolute percentage error (MAPE) by 30%, 13%, and 20%, respectively. Flatter loss sharpness value and better robust performance obtained from the transfer learning method is verified by an experimental verification. Finally, a digital system design for power plants with real-time data visualization monitoring, parameter prediction, and fault warning functions are implemented.
The superheater in the boiler is the key of equipment connecting high-temperature steam to the turbine for power generation. At present, the problems of large variable fluctuations, strong timing coupling, and multi-power plant data utilization prevent the temperature, flow, and pressure prediction of the boiler superheater. In this paper, a method for predicting the parameters of boiler superheater based on a transfer learning model is proposed, which realizes the joint utilization of data from multiple power plants. The method first collects data from a waste incineration boiler power plant for pre-training the long short-term memory (LSTM)-transformer model, and then completes the transfer learning training on the new power plant. The proposed method has the advantages of high prediction accuracy, good robustness, and more reliable location prediction with drastic changes. The predictions on the test set are within ± 5% of the experimental value. Compared with the model not trained by the transfer learning, the proposed method achieves the lowest relative errors for all prediction intervals in the 3-15 min range. Compared to the linear regression (LR), support vector regression (SVR), and random forest (RF), the proposed method improves the average absolute percentage error (MAPE) by 30%, 13%, and 20%, respectively. Flatter loss sharpness value and better robust performance obtained from the transfer learning method is verified by an experimental verification. Finally, a digital system design for power plants with real-time data visualization monitoring, parameter prediction, and fault warning functions are implemented.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.