Abstract：Although data-driven models for mechanical fault diagnosis of high-voltage circuit breakers（HVCBs）have demonstrated the ability to extract health-related features from monitoring signals, their performance in practical field scenarios is often constrained by the scarcity of fault samples, insufficient labeling, and the lack of mechanistic constraints. These limitations lead to insufficient feature representation, limited generalization, and reduced interpretability, hindering the deployment of these models in engineering applications. This paper proposes a self-supervised contrastive domain adaptation framework that leverages prior knowledge to achieve high-precision mechanical fault diagnosis of HVCBs under small-sample, unlabeled conditions. First, a Fourier feature learning network incorporating prior knowledge is developed to extract more informative, physically meaningful features. The network integrates insights into fault mechanisms derived from the excitation-response behavior of HVCB mechanical components, guiding the extraction and refinement of frequency-domain features from both the amplitude and phase spectra. It effectively enhances the model's robustness to strong noise and reduces its dependence on large-scale labeled datasets. By fusing prior knowledge with deep Fourier features, the network achieves physically interpretable feature representations that accurately reflect the underlying fault mechanisms. Second, a contrastive domain adaptation strategy is introduced to address distributional differences between the source and target domains, which may arise from variations in voltage levels, structural configurations, and operating mechanisms across different HVCBs. The proposed strategy employs contrastive learning to simultaneously align features in the source and target domains. By constructing positive and negative sample pairs based on class correspondence, the model enhances intra-class compactness and inter-class separability. This approach enables effective feature representation and cross-domain generalization even with unlabeled target data and small sample sizes. Multi-round training with data augmentation further strengthens the model's adaptability to real-world operating conditions. The proposed framework integrates Fourier feature learning, guided by prior knowledge, and contrastive domain adaptation into a unified model. The model is validated on two independently developed HVCB diagnostic platforms, covering multiple mechanical fault types and normal operating states. The results demonstrate that the proposed approach achieves diagnostic accuracy, recall, and F1 scores of 97.36%, 95.28%, and 94.89%, respectively. Notably, the model maintained high performance under strong noise and sample imbalance. A comparative analysis with baseline and conventional methods confirms that integrating prior knowledge and contrastive domain adaptation significantly improves diagnostic performance across multiple fault categories. In conclusion, by combining physically guided Fourier feature extraction with domain-adaptive contrastive learning, the proposed framework not only achieves high-precision fault recognition but also ensures interpretability and practical applicability. The experimental results highlight its potential for deployment in real-world engineering environments, providing a reliable and robust tool for intelligent monitoring and predictive maintenance of HVCBs.
王艳新, 闫静, 荆乾震, 耿英三, 王建华. 基于先验知识约束和对比域适应学习的高压断路器机械故障诊断方法研究[J]. 电工技术学报, 2026, 41(16): 5698-5711.
Wang Yanxin, Yan Jing, Jing Qianzhen, Geng Yingsan, Wang Jianhua. Mechanical Fault Diagnosis of High-Voltage Circuit Breakers Based on Prior Knowledge Constraints and Contrastive Domain Adaptation. Transactions of China Electrotechnical Society, 2026, 41(16): 5698-5711.
[1] 赵陈琛, 张国钢, 刘洁, 等. 基于物理信息和迁移网络的高压断路器弹簧操动机构机械参数辨识及故障诊断[J]. 电工技术学报, 2026, 41(6): 2073-2085. Zhao Chenchen, Zhang Guogang, Liu Jie, et al.Mechanical parameter identification and fault diagnosis of spring operating mechanism of high-voltage circuit breaker based on physical information and transfer network[J]. Transactions of China Electrotechnical Society, 2026, 41(6): 2073-2085. [2] 范兴明, 许洪华, 张思舜, 等. 基于改进蜣螂优化算法深度混合核极限学习机的高压断路器故障诊断[J]. 电工技术学报, 2025, 40(12): 3994-4003. Fan Xingming, Xu Honghua, Zhang Sishun, et al.Fault diagnosis of high voltage circuit breaker based on improved dung beetle optimizer algorithm deep hybrid kernel extreme learning machine[J]. Transactions of China Electrotechnical Society, 2025, 40(12): 3994-4003. [3] Yan Jing, Wang Yanxin, Yang Zhou, et al.Few-shot mechanical fault diagnosis for a high-voltage circuit breaker via a transformer-convolutional neural net-work and metric meta-learning[J]. IEEE Transactions on Instrumentation and Measurement, 2023, 72: 1-11. [4] 范兴明, 许洪华, 张思舜, 等. 基于改进蜣螂优化算法深度混合核极限学习机的高压断路器故障诊断[J]. 电工技术学报, 2025, 40(12): 3994-4003. Fan Xingming, Xu Honghua, Zhang Sishun, et al.Fault diagnosis of high voltage circuit breaker based on improved dung beetle optimizer algorithm deep hybrid kernel extreme learning machine[J]. Transactions of China Electrotechnical Society, 2025, 40(12): 3994-4003. [5] Yang Qiuyu, Wang Zixuan, Ruan Jiangjun, et al.Small-sample fault diagnosis method for high-voltage circuit breakers via data augmentation and deep learning[J]. IEEE Transactions on Instrumentation and Measurement, 2024, 73: 3537411. [6] 王艳新, 闫静, 王建华, 等. 基于多级二阶注意力孪生网络的小样本GIS局部放电诊断方法[J]. 电工技术学报, 2023, 38(8): 2255-2264. Wang Yanxin, Yan Jing, Wang Jianhua, et al.Few-shot partial discharge diagnosis for gas-insulated switchgear using a novel multi-level second-order attention Siamese network[J]. Transactions of China Electrotechnical Society, 2023, 38(8): 2255-2264. [7] Wang Yanxin, Yan Jing, Qi Meirong, et al.A novel meta-learning and network architecture search approach for few-shot high-voltage circuit breaker fault diagnosis[C]//2023 IEEE 6th International Electrical and Energy Conference（CIEEC), Hefei, China, 2023: 122-127. [8] Wang Yanxin, Yan Jing, Wang Jianhua, et al.A novel hybrid transfer learning approach for small-sample high-voltage circuit breaker fault diagnosis on-site[J]. IEEE Transactions on Industry Applications, 2023, 59(4): 4942-4950. [9] Wang Yanxin, Yan Jing, Ye Xinyu, et al.Few-shot transfer learning with attention mechanism for high-voltage circuit breaker fault diagnosis[J]. IEEE Transactions on Industry Applications, 2022, 58(3): 3353-3360. [10] 阎光伟, 张云馨, 符哲源, 等. 基于改进集合预测网络的输变电设备故障知识图谱构建方法[J]. 电工技术学报, 2025, 40(15): 4976-4987. Yan Guangwei, Zhang Yunxin, Fu Zheyuan, et al.The construction method of fault knowledge graph for transmission and transformation equipment based on improved set prediction network[J]. Transactions of China Electrotechnical Society, 2025, 40(15): 4976-4987. [11] 杨夏洁, 林子谦, 徐丹丹, 等. 基于联邦对比学习的航天器故障诊断[J]. 计算机应用研究, 2025, 42(8): 2467-2473. Yang Xiajie, Lin Ziqian, Xu Dandan, et al.Spacecraft fault diagnosis based on federated comparative learning[J]. Application Research of Computers, 2025, 42(8): 2467-2473. [12] An Yiyao, Zhang Ke, Chai Yi, et al.Domain adaptation network base on contrastive learning for bearings fault diagnosis under variable working con-ditions[J]. Expert Systems with Applications, 2023, 212: 118802. [13] Yang Chaoying, Liu Jie, Xu Qi, et al.A generalized graph contrastive learning framework for few-shot machine fault diagnosis[J]. IEEE Transactions on Industrial Informatics, 2024, 20(2): 2692-2701. [14] Wang Yanxin, Yan Jing, Jing Qianzhen, et al.Self-supervised representation learning for GIS partial discharge condition assessment[C]//Proceedings of the 1st Electrical Artificial Intelligence Conference, Nanjing, China, 2024: 332-342. [15] Gao Honghao, Jiang Wangyang, Ran Q, et al.Vision-language interaction via contrastive learning for surface anomaly detection in consumer electronics manufacturing[J]. IEEE Transactions on Consumer Electronics, 2024, 70(3): 6119-6130. [16] Zhou Yuhang, Zhang Qiang, Huang Ting, et al.Prior knowledge-augmented meta-learning for fine-grained fault diagnosis[J]. IEEE Transactions on Industrial Informatics, 2024, 20(6): 8115-8124. [17] Zhang Tianci, Chen Jinglong, He Shuilong, et al.Prior knowledge-augmented self-supervised feature learning for few-shot intelligent fault diagnosis of machines[J]. IEEE Transactions on Industrial Electronics, 2022, 69(10): 10573-10584. [18] Wang Huan, Luo Wenjun, Liu Zhiliang, et al.Fourier feature refiner network with soft thresholding for machinery fault diagnosis under highly noisy conditions[J]. IEEE Internet of Things Journal, 2024, 11(13): 22880-22891. [19] Yan Jing, Wang Yanxin.High-voltage circuit breaker intelligent diagnosis technology for mechanical faults under power Internet of Things context[C]//2020 IEEE International Conference on High Voltage Engineering and Application（ICHVE), Beijing, China, 2020: 1-4. [20] 王涛, 杨远, 申冰洁. 面向运行场景变化的判别式域适应暂态稳定评估[J]. 电工技术学报, 2025, 40(10): 3301-3314. Wang Tao, Yang Yuan, Shen Bingjie.A discrimi-native domain-adaptive transient stability assessment framework for operating scenario variations[J]. Transactions of China Electrotechnical Society, 2025, 40(10): 3301-3314. [21] 王艳新, 闫静, 王建华, 等. 基于域对抗迁移卷积神经网络的小样本GIS绝缘缺陷智能诊断方法[J]. 电工技术学报, 2022, 37(9): 2150-2160. Wang Yanxin, Yan Jing, Wang Jianhua, et al.Intelligent diagnosis for GIS with small samples using a novel adversarial transfer learning in convolutional neural network[J]. Transactions of China Electro-technical Society, 2022, 37(9): 2150-2160. [22] Ning Yujie, Peng Jiangtao, Liu Quanyong, et al.Contrastive learning based on category matching for domain adaptation in hyperspectral image classifica-tion[J]. IEEE Transactions on Geoscience and Remote Sensing, 2023, 61: 5301814. [23] 王艳新, 闫静, 王建华, 等. 基于特征融合度量学习的高压断路器机械故障诊断[J]. 仪器仪表学报, 2022, 43(9): 98-105. Wang Yanxin, Yan Jing, Wang Jianhua, et al.Mechanical fault diagnosis for high voltage circuit breaker via a novel feature fusion metric learning[J]. Chinese Journal of Scientific Instrument, 2022, 43(9): 98-105.
2026, Vol. 41 
