Research on Defect Identification Method of Main Transformer Core Clamping Based on the Collaborative Strategy of Large and Small Models

DOI: https://doi.org/10.62517/jsse.202508302

Author(s)

Peng Li, Yunlong Liu*, Baiyuan Liu, Xiangcheng Kong, Weifeng Wang, Chuanhui Zhang

Affiliation(s)

State Grid Shangqiu Power Supply Company, Shangqiu, Henan, China *Corresponding Author

Abstract

With the intelligent development of the power system, the main transformer, as a key piece of equipment in the power system, the health of its operating state directly affects the stability of the power grid. Due to its complex structure and harsh operating environment, the core clamping parts of the main transformer are prone to defects such as loosening, cracking and rusting. The traditional defect identification methods have obvious deficiencies in terms of accuracy and efficiency. To this end, this paper proposes an intelligent recognition method based on the collaborative strategy of large and small models. By combining the rapid response capability of small models on the on-site side with the high-precision analysis capability of large models in the cloud, an efficient and intelligent main transformer core clamping defect recognition system is constructed. Experiments show that this method significantly improves the response speed and system practicability while ensuring the recognition accuracy.

Keywords

Defect Identification; Collaborative Strategy of Large and Small Models; Main Transformer Core Clamping Piece; Artificial Intelligence

References

[1] Dang, L. M., Wang, H., Li, Y., Nguyen, T. N., & Moon, H. (2022). DefectTR: End‑to‑end defect detection for sewage networks using a transformer. Construction and Building Materials, 325, 126584. [2] Gao, L., Zhang, J., Yang, C., & Zhou, Y. (2022). Cas‑VSwin transformer: A variant Swin transformer for surface‑defect detection. Computers in Industry, 140, 103689. [3] Zhang, Y., Tang, Y., Liu, Y., & Liang, Z. (2022) Fault diagnosis of transformer using artificial intelligence: A review. Front. Energy Res. 10:1006474 [4] Wu, H. Y., Matthew, J. T., & John W. S. (2023) A transformer-based approach for novel fault detection and fault classification/diagnosis in manufacturing: A rotary system application. Journal of Manufacturing Systems, 67, 439-452 [5] Lei, Z., Zhang, Y., Wang, J., & Zhou, M. (2024). Cloud‑Edge Collaborative Defect Detection Based on Efficient Yolo Networks and Incremental Learning. Sensors, 24(18), 5921. [6] Li, Y., Xiang, Y., Guo, H., Liu, P., & Liu, C. (2022). Swin Transformer Combined with Convolution Neural Network for Surface Defect Detection. Machines, 10(11), 1083. [7] Gao, P., Wu, T., & Song, C. (2024). Cloud–Edge Collaborative Strategy for Insulator Recognition and Defect Detection Model Using Drone‑Captured Images. Drones, 8(12), 779. [8] Wang, J., Xu, G., Yan, F., Wang, J., & Wang, Z. (2022). Defect Transformer: An Efficient Hybrid Transformer Architecture for Surface Defect Detection. arXiv preprint arXiv:2207.08319. [9] Hu, C., Yao, J., Wu, W., Qiu, W., & Zhu, L. (2022). A Lightweight Reconstruction Network for Surface Defect Inspection. arXiv preprint arXiv:2212.12878. [10] Xu, Y., Khan, T.M., Song, Y. et al. (2025) Edge deep learning in computer vision and medical diagnostics: a comprehensive survey. Artif Intell Rev 58, 93.