To address the three major challenges in fault diagnosis of main circulating pumps in converter station valve cooling systems—small-sample scenarios, cross-domain distribution shift, and lack of physical interpretability—a physics-data dual-channel collaborative cross-domain transfer learning framework is proposed. The physics channel designs a physics-constrained conditional generative adversarial network (PC-cGAN), which explicitly embeds the pump's Euler equation into the loss function to constrain the hydraulic characteristics of generated samples. The data channel constructs a multi-source signal attention fusion network (MSAF-Net), achieving cross-modal fusion of vibration, temperature, and pressure features through a physically guided mechanism. Theoretically, this paper establishes a quantitative relationship between the Wasserstein distance and the HΔH divergence, deriving an explicit upper bound for domain adaptation error. Validation on measured data from 69 main circulating pumps across five UHV converter stations shows that under the stringent condition where only 5% of target domain samples are labeled, the cross-domain diagnostic accuracy reaches 89.7%, an improvement of 13.4 percentage points over the optimal baseline. The Pearson correlation coefficient between the model's decision logic and the fault physical mechanism is 0.82 (p < 0.001), indicating that the diagnostic results possess clear physical interpretability.
| Published in | American Journal of Science, Engineering and Technology (Volume 11, Issue 3) |
| DOI | 10.11648/j.ajset.20261103.21 |
| Page(s) | 212-224 |
| Creative Commons |
This is an Open Access article, distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution and reproduction in any medium or format, provided the original work is properly cited. |
| Copyright |
Copyright © The Author(s), 2026. Published by Science Publishing Group |
Converter Station, Valve Cooling System, Main Circulating Pump, Fault Diagnosis, Transfer Learning, Small-sample Learning, Physical Constraints, Cross-domain Adaptation
| [1] | Zhang G H, Ma G P. Analysis of main pump switching logic in valve cooling system under voltage faults [J]. Power System Technology, 2021, 45(1): 189-197. |
| [2] | Yu J Y, Qiao Z Q, Wang Y J, et al. Research on fault diagnosis technology of converter valve cooling system based on CVA [J]. High Voltage Apparatus, 2023, 59(11): 32-39. |
| [3] | Wang L Y, Yang Y Q, Shi L, et al. Parameter prediction of converter station valve cooling system based on long short-term memory method [J]. Science, Technology and Engineering, 2022, 22(2): 547-555. |
| [4] | Wei Y, Li X W, Zhang Y L, et al. Research on fault location method for converter station valve area based on metric learning and knowledge reasoning [J]. Electrical Measurement and Instrumentation, 2025, 62(6): 134-142. |
| [5] | Hu J W. Research on intelligent diagnosis algorithms for rotating machinery equipment in small-sample cross-domain complex scenarios [D]. China Wuhan: Wuhan University of Science and Technology, 2024. |
| [6] | Wang S, Shuai H Q, Hu H H, et al. Few-shot fault diagnosis of axial piston pump based on prior knowledge-embedded meta learning vision transformer under variable operating conditions [J]. Expert Systems with Applications, 2025, 269: 126452. |
| [7] | Jiang C X, Chen H, Xu Q F, et al. Few-shot fault diagnosis of rotating machinery with two-branch prototypical networks [J]. Journal of Intelligent Manufacturing, 2023, 34(4): 1667-1681. |
| [8] | Liu S, Li J, Zhou N, et al. Intelligent fault diagnosis of rotating machine via expansive dual-attention fusion transformer enhanced by semi-supervised learning [J]. Expert Systems with Applications, 2025, 260: 125398. |
| [9] | Ye M, Yan X, Hua X, et al. MRCFN: A multi-sensor residual convolutional fusion network for intelligent fault diagnosis of bearings in noisy and small sample scenarios [J]. Expert Systems with Applications, 2025, 259: 125214. |
| [10] | Zhang S, Ma X, Nie Z, et al. Integration of multi-relational graph oriented fault diagnosis method for nuclear power circulating water pumps [J]. Measurement, 2025, 242: 115811. |
| [11] | Zhao C, Zio E, Shen W. Domain generalization for cross-domain fault diagnosis: An application-oriented perspective and a benchmark study [J]. Reliability Engineering and System Safety, 2024, 245: 109964. |
| [12] | Liang P, Wang X, Ai C, et al. SRSGCN: A novel multi-sensor fault diagnosis method for hydraulic axial piston pump with limited data [J]. Reliability Engineering and System Safety, 2025, 253: 110563. |
| [13] | Zhou K, Lu N, Jiang B, et al. FEV-Swin: Multi-source heterogeneous information fusion under a variant Swin Transformer framework for intelligent cross-domain fault diagnosis [J]. Knowledge-Based Systems, 2025, 310: 112982. |
| [14] | Ren L, Mo T, Cheng X. Meta-learning based domain generalization framework for fault diagnosis with gradient aligning and semantic matching [J]. IEEE Transactions on Industrial Informatics, 2024, 20: 754-764. |
| [15] | Jia L, Chow T W S, Yuan Y. Causal disentanglement domain generalization for time-series signal fault diagnosis [J]. Neural Networks, 2024, 172: 106099. |
| [16] | Su Z, Lian P, Shang P, et al. Semi-supervised source-free domain adaptation method via diffusive label propagation for rotating machinery fault diagnosis [J]. Reliability Engineering and System Safety, 2024, 252: 110408. |
| [17] | Lin T, Ren Z, Zhu L, et al. A systematic review of multi-sensor information fusion for equipment fault diagnosis [J]. IEEE Transactions on Instrumentation and Measurement, 2025, 74: 3507848. |
| [18] | Jiang X, Li X, Wang Q, et al. Multi-sensor data fusion-enabled semi-supervised optimal temperature-guided PCL framework for machinery fault diagnosis [J]. Information Fusion, 2024, 101: 102005. |
| [19] | Yan X, Yan W J, Xu Y, et al. Machinery multi-sensor fault diagnosis based on adaptive multivariate feature mode decomposition and multi-attention fusion residual convolutional neural network [J]. Mechanical Systems and Signal Processing, 2023, 202: 110664. |
| [20] | Huo Z, Martínez-García M, Zhang Y, et al. A multisensor information fusion method for high-reliability fault diagnosis of rotating machinery [J]. IEEE Transactions on Instrumentation and Measurement, 2022, 71: 1-12. |
| [21] | Zhong J, Zheng Y, Ruan C, et al. M-IPISincNet: An explainable multi-source physics-informed neural network based on improved SincNet for rolling bearings fault diagnosis [J]. Information Fusion, 2025, 115: 102761. |
| [22] | Sun D, Li Y, Jia S, et al. Physical knowledge-driven feature fusion and reconstruction network for fault diagnosis with incomplete multisource data [J]. Mechanical Systems and Signal Processing, 2025, 225: 112222. |
| [23] | Wang Z, Zhou Z, Xu W, et al. Physics informed neural networks for fault severity identification of axial piston pumps [J]. Journal of Manufacturing Systems, 2023, 71: 421-437. |
| [24] | Yang J F, Zhang N, He Y L, et al. Novel dual-network autoencoder based adversarial domain adaptation with wasserstein divergence for fault diagnosis of unlabeled data [J]. Expert Systems with Applications, 2024, 238: 122393. |
| [25] | M. Arjovsky, S. Chintala, and L. Bottou, "Wasserstein generative adversarial networks," in Proc. 34th Int. Conf. Mach. Learn. (ICML), Sydney, Australia, 2017, pp. 214-223. |
| [26] | Ben-David S, Blitzer J, Crammer K, et al. A theory of learning from different domains [J]. Machine Learning, 2010, 79: 151-175. |
| [27] | Lundberg S M, Lee S I. A unified approach to interpreting model predictions [C]//Advances in Neural Information Processing Systems, 2017: 4765-4774. |
| [28] | Zhou S, Qin L, Yang Y, Wei Z, Wang J, Wang J, Ruan J, Tang X, Wang X, Liu K. A novel ensemble fault diagnosis model for main circulation pumps of converter valves in VSC-HVDC transmission systems [J]. Sensors, 2023, 23(11): 5082. |
| [29] | Mei F, Zhang X G, Li J W, Lu J H, Feng T T. A lightweight fault diagnosis model for the main circulating pump of an ultra high voltage converter valve based on a multidimensional vibration feature graph [J]. Power System Protection and Control, 2024, 52(16): 83-96. |
| [30] | Zhao Z M, Gao H D, Teng M F, Zhang Q Y, Liu Y. Fault cause analysis and hidden danger management of main circulating pump for valve cooling system in a flexible DC converter station [J]. Hunan Electric Power, 2025, 45(3): 89-94. |
APA Style
Lei, G., Wenchao, W., Heng, W., Yu, Z., Biao, H. (2026). A Physics-Data Dual-Channel Collaborative Cross-Domain Transfer Learning Fault Diagnosis Method for Main Circulation Pump in Converter Valve Cooling Systems. American Journal of Science, Engineering and Technology, 11(3), 212-224. https://doi.org/10.11648/j.ajset.20261103.21
ACS Style
Lei, G.; Wenchao, W.; Heng, W.; Yu, Z.; Biao, H. A Physics-Data Dual-Channel Collaborative Cross-Domain Transfer Learning Fault Diagnosis Method for Main Circulation Pump in Converter Valve Cooling Systems. Am. J. Sci. Eng. Technol. 2026, 11(3), 212-224. doi: 10.11648/j.ajset.20261103.21
AMA Style
Lei G, Wenchao W, Heng W, Yu Z, Biao H. A Physics-Data Dual-Channel Collaborative Cross-Domain Transfer Learning Fault Diagnosis Method for Main Circulation Pump in Converter Valve Cooling Systems. Am J Sci Eng Technol. 2026;11(3):212-224. doi: 10.11648/j.ajset.20261103.21
Share This Article
Twitter
Linked In
Facebook
Copy
Download@article{10.11648/j.ajset.20261103.21,
author = {Gao Lei and Wang Wenchao and Wu Heng and Zhang Yu and He Biao},
title = {A Physics-Data Dual-Channel Collaborative Cross-Domain Transfer Learning Fault Diagnosis Method for Main Circulation Pump in Converter Valve Cooling Systems},
journal = {American Journal of Science, Engineering and Technology},
volume = {11},
number = {3},
pages = {212-224},
doi = {10.11648/j.ajset.20261103.21},
url = {https://doi.org/10.11648/j.ajset.20261103.21},
eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.ajset.20261103.21},
abstract = {To address the three major challenges in fault diagnosis of main circulating pumps in converter station valve cooling systems—small-sample scenarios, cross-domain distribution shift, and lack of physical interpretability—a physics-data dual-channel collaborative cross-domain transfer learning framework is proposed. The physics channel designs a physics-constrained conditional generative adversarial network (PC-cGAN), which explicitly embeds the pump's Euler equation into the loss function to constrain the hydraulic characteristics of generated samples. The data channel constructs a multi-source signal attention fusion network (MSAF-Net), achieving cross-modal fusion of vibration, temperature, and pressure features through a physically guided mechanism. Theoretically, this paper establishes a quantitative relationship between the Wasserstein distance and the HΔH divergence, deriving an explicit upper bound for domain adaptation error. Validation on measured data from 69 main circulating pumps across five UHV converter stations shows that under the stringent condition where only 5% of target domain samples are labeled, the cross-domain diagnostic accuracy reaches 89.7%, an improvement of 13.4 percentage points over the optimal baseline. The Pearson correlation coefficient between the model's decision logic and the fault physical mechanism is 0.82 (p 0.001), indicating that the diagnostic results possess clear physical interpretability.},
year = {2026}
}
TY - JOUR T1 - A Physics-Data Dual-Channel Collaborative Cross-Domain Transfer Learning Fault Diagnosis Method for Main Circulation Pump in Converter Valve Cooling Systems AU - Gao Lei AU - Wang Wenchao AU - Wu Heng AU - Zhang Yu AU - He Biao Y1 - 2026/09/22 PY - 2026 N1 - https://doi.org/10.11648/j.ajset.20261103.21 DO - 10.11648/j.ajset.20261103.21 T2 - American Journal of Science, Engineering and Technology JF - American Journal of Science, Engineering and Technology JO - American Journal of Science, Engineering and Technology SP - 212 EP - 224 PB - Science Publishing Group SN - 2578-8353 UR - https://doi.org/10.11648/j.ajset.20261103.21 AB - To address the three major challenges in fault diagnosis of main circulating pumps in converter station valve cooling systems—small-sample scenarios, cross-domain distribution shift, and lack of physical interpretability—a physics-data dual-channel collaborative cross-domain transfer learning framework is proposed. The physics channel designs a physics-constrained conditional generative adversarial network (PC-cGAN), which explicitly embeds the pump's Euler equation into the loss function to constrain the hydraulic characteristics of generated samples. The data channel constructs a multi-source signal attention fusion network (MSAF-Net), achieving cross-modal fusion of vibration, temperature, and pressure features through a physically guided mechanism. Theoretically, this paper establishes a quantitative relationship between the Wasserstein distance and the HΔH divergence, deriving an explicit upper bound for domain adaptation error. Validation on measured data from 69 main circulating pumps across five UHV converter stations shows that under the stringent condition where only 5% of target domain samples are labeled, the cross-domain diagnostic accuracy reaches 89.7%, an improvement of 13.4 percentage points over the optimal baseline. The Pearson correlation coefficient between the model's decision logic and the fault physical mechanism is 0.82 (p 0.001), indicating that the diagnostic results possess clear physical interpretability. VL - 11 IS - 3 ER -
State Grid Shaanxi Electric Power Co., Ltd., Xi'an, China
State Grid Shaanxi Electric Power Co., Ltd., Xi'an, China
State Grid Shaanxi Electric Power Co., Ltd., Xi'an, China
State Grid Shaanxi Electric Power Co., Ltd., Xi'an, China
State Grid Shaanxi Electric Power Co., Ltd., Xi'an, China
Information