Paper
5 June 2024 Research on the estimation of SF6 gas recovery rate based on kNN-IChOA-KELM
Lijun Zhang, Kecheng Liu, Hesong Han, Rongxue Shi, Yingnan Wang
Author Affiliations +
Proceedings Volume 13163, Fourth International Conference on Mechanical, Electronics, and Electrical and Automation Control (METMS 2024); 1316321 (2024) https://doi.org/10.1117/12.3030495
Event: International Conference on Mechanical, Electronics, and Electrical and Automation Control (METMS 2024), 2024, Xi'an, China
Abstract
SF6 gas, commonly used as an insulation and arc-quenching medium in high-voltage electrical equipment, has garnered considerable attention regarding its recovery and treatment after use. Addressing the challenge of calculating and measuring gas recovery rates during on-site operations, this paper proposes an estimation model based on KNN-IChOA-KELM. The approach begins by establishing an SF6 recovery rate estimation model using Kernel Extreme Learning Machine (KELM). To enhance the model's generalization capability, an improved chimpanzee algorithm is employed to optimize the KELM parameters. Subsequently, in cases where certain samples exhibit significant error deviations, the KNearest Neighbors (KNN) algorithm is applied for classification. After classification, the IChOA-KELM model is utilized for SF6 recovery rate estimation. Finally, the proposed method's accuracy is analyzed through comparison with four other estimation methods: IChOA-BP, IChOA-SVM, IChOA-RF, and IChOA-LSTM. Case study results demonstrate that the SF6 recovery rate estimation method presented in this paper achieves at least a 0.07% improvement in Mean Absolute Percentage Error (MAPE) and a 0.11% improvement in Root Mean Square Error (RMSE) compared to the other methods.
(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.
Lijun Zhang, Kecheng Liu, Hesong Han, Rongxue Shi, and Yingnan Wang "Research on the estimation of SF6 gas recovery rate based on kNN-IChOA-KELM", Proc. SPIE 13163, Fourth International Conference on Mechanical, Electronics, and Electrical and Automation Control (METMS 2024), 1316321 (5 June 2024); https://doi.org/10.1117/12.3030495
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Data modeling

Mathematical optimization

Error analysis

Machine learning

Measurement devices

Back to Top