基于双通道特征融合与高斯过程回归的直流断路器寿命预测方法

doi:10.19595/j.cnki.1000-6753.tces.250868

Abstract
Figure/Table
References
Related Citation (15)

Download: PDF (5979 KB) HTML (1 KB)
Export: BibTeX | EndNote (RIS)

Abstract The operational lifetime of direct current (DC) circuit breakers is governed by cumulative degradation of the contact system caused by arc erosion, thermal stress, and morphological evolution during repeated interruption cycles. Accurate modeling and prediction of this degradation process are crucial for ensuring the long-term reliability and safe operation of DC switching equipment. This study establishes a physics-informed and data-driven hybrid framework for lifetime prediction of DC circuit breakers. The method integrates numerical simulation, experimental degradation measurement, and probabilistic regression modeling.
A two-dimensional axisymmetric finite element model is developed to simulate the erosion process of the contact system under multiple breaking conditions. The model accounts for transient heat conduction, surface-tension-driven fluid flow (Marangoni effect), and phase transition of the contact material. Spatial- temporal distributions of temperature, velocity, and pressure are solved to characterize molten-pool evolution and quantify the mass loss of the contact material. The simulated mass loss, a function of arc current and arcing duration, serves as a structurally interpretable degradation indicator that compensates for the lack of direct measurement during service.
Parallel experimental lifetime tests are performed on eight low-voltage DC circuit breakers rated at 500 V and 63 A, with test currents ranging from 63 A to 20 kA and time constants between 2 ms and 10 ms. During each breaking cycle, arc voltage and current are recorded to calculate the arc energy, while post-test measurements include contact resistance, surface roughness, and mass variation. The experiments reveal that contact degradation is accelerated by increasing current amplitude and prolonged arcing time, leading to dominant failure modes such as surface ignition, contact welding, and overtravel depletion.
To effectively represent the complex degradation behavior, a two-stage nonlinear feature extraction strategy combining kernel principal component analysis (KPCA) and autoencoder (AE) is employed. KPCA first maps heterogeneous degradation data into a high-dimensional feature space to capture nonlinear correlations, followed by AE-based compression that encodes temporal degradation trajectories into compact latent representations. The resulting features preserve both statistical and physical information, reducing redundancy among electrical, thermal, and morphological parameters.
A dual-channel Gaussian process regression (GPR) model is then constructed to perform lifetime prediction using dense-sampling features and sparse-sampling features. Independent regression channels are trained for each data type, and their predictions are fused through inverse-variance weighting to enhance stability under uncertain or incomplete information.
Model validation shows that the proposed framework achieves a prediction error below 5% when full degradation information is available. In cases with partial or missing features, the hybrid model maintains acceptable accuracy and robustness. Although the simulation-derived mass-loss feature increases the prediction error from 5.4% to 18.1%, it is verified to identify early-stage degradation before pronounced morphological changes occur.
The study establishes the nonlinear relationships between arc current, arcing time, and material erosion. The coupling between electrical and mechanical degradation mechanisms is elucidated. The hybrid modeling approach bridges the gap between physical interpretability and statistical generalization, providing a scalable route for predictive maintenance of electromechanical devices. Future work will focus on incorporating dynamic arc movement, material transfer effects, and multi-fault degradation pathways to extend the method toward real-time life assessment.

Key words： DC circuit breaker contact degradation lifetime prediction finite element simulation multi- source feature extraction

Received: 23 May 2025

PACS:

TM571

	Service

	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	Lin Jingyi
	Wu Jianwen
	Xia Shangwen
	Chen Ruang

Cite this article:

Lin Jingyi,Wu Jianwen,Xia Shangwen等. Lifetime Prediction Method of DC Circuit Breakers Based on Dual-Channel Feature Fusion and Gaussian Process Regression[J]. Transactions of China Electrotechnical Society, 2026, 41(10): 3550-3562.

URL:

https://dgjsxb.ces-transaction.com/EN/10.19595/j.cnki.1000-6753.tces.250868 OR https://dgjsxb.ces-transaction.com/EN/Y2026/V41/I10/3550

[1] 胡博凯, 李奎, 牛峰, 等. 低压断路器机械特性状态监测方法研究[J]. 电工技术学报, 2022, 37(13): 3317-3330.
Hu Bokai, Li Kui, Niu Feng, et al.Research on condition monitoring method of mechanical characte- ristics of low-voltage circuit breaker[J]. Transactions of China Electrotechnical Society, 2022, 37(13): 3317-3330.
[2] 邢云琪, 王子涵, 杨广正, 等. 低压直流断路器双金属片性能退化特性及机理分析[J]. 电工技术学报, 2024, 39(22): 7266-7277.
Xing Yunqi, Wang Zihan, Yang Guangzheng, et al.Degradation characteristics and mechanism analysis of bimetal sheet of low voltage DC circuit breaker[J]. Transactions of China Electrotechnical Society, 2024, 39(22): 7266-7277.
[3] 赵书涛, 黄伟杰, 刘会兰, 等. 高压断路器触头烧蚀及电寿命评估研究综述[J]. 高电压技术, 2024, 50(2): 489-502.
Zhao Shutao, Huang Weijie, Liu Huilan, et al.Review of research on contact ablation and electrical life assessment of high-voltage circuit breakers[J]. High Voltage Engineering, 2024, 50(2): 489-502.
[4] Cao Pei, Li Jiahao, Han Yili, et al.Mechanical remaining life prediction of circuit breaker based on long-term and short-term memory neural network[C]// 2023 IEEE 6th International Electrical and Energy Conference (CIEEC), Hefei, China, 2023: 2734-2738.
[5] 王毅钊, 陈兴贵, 韩寅峰, 等. 双断口真空断路器非同步关合直流预击穿特性[J]. 高压电器, 2025, 61(8): 61-66.
Wang Yizhao, Chen Xinggui, Han Yinfeng, et al.DC pre-breakdown characteristics of double-break vacuum circuit breaker in synchronous making operation[J]. High Voltage Apparatus, 2025, 61(8): 61-66.
[6] 侯彦杰, 贺小瑞, 刘伟, 等. 一种基于磁耦合转移原理的10kV直流断路器[J]. 电气技术, 2024, 25(5): 57-64.
Hou Yanjie, He Xiaorui, Liu Wei, et al.A 10kV DC circuit breaker based on magnetic coupling transfer principle[J]. Electrical Engineering, 2024, 25(5): 57-64.
[7] 赵成晨, 李奎, 胡博凯, 等. 变应力条件下低压断路器剩余电寿命预测[J]. 中国电机工程学报, 2022, 42(21): 8004-8016.
Zhao Chengchen, Li Kui, Hu Bokai, et al.Residual electrical life prediction of low-voltage circuit breakers under varied stress[J]. Proceedings of the CSEE, 2022, 42(21): 8004-8016.
[8] 陈殿生, 王田苗, 魏洪兴. 数控车床故障分布的两重威布尔分段模型[J]. 北京航空航天大学学报, 2005, 31(7): 766-769.
Chen Diansheng, Wang Tianmiao, Wei Hongxing.Sectional model involving two Weibull distributions for CNC lathe failure probability[J]. Journal of Beijing University of Aeronautics and Astronautics, 2005, 31(7): 766-769.
[9] Li Gaoyang, Wang Xiaohua, Rong Mingzhe, et al.Health status centered mechanical feature extraction for high voltage circuit breakers[C]//2017 4th Inter- national Conference on Electric Power Equipment- Switching Technology (ICEPE-ST), Xi'an, China, 2017: 911-915.
[10] Pecht M, Jaai R.A prognostics and health management roadmap for information and electronics-rich systems[J]. Microelectronics Reliability, 2010, 50(3): 317-323.
[11] 刘帼巾, 李想, 王泽, 等. 基于Wiener过程电子式漏电断路器的剩余寿命预测[J]. 电工技术学报, 2022, 37(2): 528-536.
Liu Guojin, Li Xiang, Wang Ze, et al.Remaining life prediction of electronic residual current circuit breaker based on Wiener process[J]. Transactions of China Electrotechnical Society, 2022, 37(2): 528-536.
[12] Guo Cheng, Li Gaoyang, Zhang Haojun, et al.Defect distribution prognosis of high voltage circuit breakers with enhanced latent Dirichlet allocation[C]//2017 Prognostics and System Health Management Con- ference, Harbin, China, 2017: 1-7.
[13] Lin Jingyi, Wu Jianwen, Xia Shangwen, et al.Multiobjective optimization of key parameters for the chamber of low-voltage DC circuit breakers based on MHD[J]. IEEE Transactions on Plasma Science, 2025, 53(1): 88-98.
[14] 李兴文, 吕启深, 田甜, 等. 直流空气电弧作用下触头烧蚀特性[J]. 高电压技术, 2020, 46(6): 1970-1977.
Li Xingwen, Lü Qishen, Tian Tian, et al.Contact erosion characteristics under direct current air arc[J]. High Voltage Engineering, 2020, 46(6): 1970-1977.
[15] 吴细秀, 李震彪. 电器电极材料喷溅侵蚀的理论计算[J]. 中国电机工程学报, 2003, 23(6): 96-101.
Wu Xixiu, Li Zhenbiao.Theoretic analysis on sputter erosion of electrode[J]. Proceedings of the CSEE, 2003, 23(6): 96-101.
[16] Hou Guobin, Zhang Dandan, Li Zhenbiao, et al.Study on magnetically blown arc characteristics and the resulting contact erosion mechanisms of high-power DC relays[J]. Journal of Physics D: Applied Physics, 2023, 56(21): 215201.
[17] 王其平. 电器电弧理论[M]. 北京: 机械工业出版社, 1991.
[18] 荣命哲, 王其平. 小电流下电触头材料转移的研究[J]. 中国电机工程学报, 1990, 10(3): 41-46.
Rong Mingzhe, Wang Qiping.The study of material transfer of electric contacts in low current[J]. Proceedings of the CSEE, 1990, 10(3): 41-46.
[19] Lee J M, Yoo C, Choi S W, et al.Nonlinear process monitoring using kernel principal component analysis[J]. Chemical Engineering Science, 2004, 59(1): 223-234.
[20] 杨旭, 周文, 程林, 等. 基于KPCA-SVM的变压器多源信息融合故障诊断研究[J]. 高压电器, 2025, 61(2): 54-62.
Yang Xu, Zhou Wen, Cheng Lin, et al.Research on multi-source information fusion fault diagnosis of transformer based on KPCA-SVM[J]. High Voltage Apparatus, 2025, 61(2): 54-62.
[21] 苏磊, 陈璐, 徐鹏, 等. 基于GRNN和KPCA组合模型的变压器油中气体体积分数短期预测[J]. 高压电器, 2021, 57(1): 82-88.
Su Lei, Chen Lu, Xu Peng, et al.Short-term prediction of gases dissolved in transformer oil based on GRNN and KPCA model[J]. High Voltage Apparatus, 2021, 57(1): 82-88.
[22] 梅飞, 梅军, 郑建勇, 等. 基于KPCA-SVM的断路器故障稳健诊断方法[J]. 电工技术学报, 2014, 29(增刊1): 50-58.
Mei Fei, Mei Jun, Zheng Jianyong, et al.The robust fault diagnostic method for circuit breaker based on KPCA and SVM[J]. Transactions of China Electro- technical Society, 2014, 29(S1): 50-58.