To address the hidden heating hazards in high-density server racks at power distribution units, UPS terminals and busbar connections, this paper introduces an infrared thermal imaging-based equipment fault detection and multi-level early warning system. The system has integrated long-wave infrared cameras, visible-light assisted asset mapping, rack power, ambient temperature and humidity, local airflow, and maintenance feedback to form a closed-loop architecture of "thermal image acquisition-radiation calibration-target area location-temperature rise anomaly identification-graded warning linkage". Methodologically, this paper builds a radiance-corrected temperature inversion model, a load-corrected temperature-rise residual model and a risk integral model with duration filtering. Literature indicators are redrawn according to task type, and classification accuracy, detection mAP and segmentation Jaccard are no longer shown in the same chart. A simulated rack-aisle case further confirms that transient hot spots are filtered, and sustained PDU terminal loosening can be upgraded from attention to an alarm based on the temperature-rise and duration standards.

Xinyi Sun, Design of Infrared Thermal Imaging Fault Detection and Early Warning System for Data Center Equipment. International Journal of Neural Network (2026), Vol. 5, Issue 1: 119-129. https://doi.org/10.38007/NN.2026.050113.

[1] Zhao R., Du Y., Yang X., Zhou Z., Wang W., Yang X. A critical review on the thermal management of data center for local hotspot elimination. Energy and Buildings, 2023, 297: 113486. DOI: 10.1016/j.enbuild.2023.113486.

[2] Chen X., Tu R., Li M., Yang X., et al. Hot spot temperature prediction and operating parameter estimation of racks in data center using machine learning algorithms based on simulation data. Building Simulation, 2023, 16: 2159-2176. DOI: 10.1007/s12273-023-1022-4.

[3] Wiysobunri B. N., Erden H. S., Toreyin B. U. CNN-based server state monitoring and fault diagnosis using infrared thermal images. Soft Computing, 2024. DOI: 10.1007/s00500-024-09792-y.

[4] Tang Z., Jian X. Thermal fault diagnosis of complex electrical equipment based on infrared image recognition. Scientific Reports, 2024, 14: 5547. DOI: 10.1038/s41598-024-56142-x.

[5] Zhou S., Liu J., Fan X., Fu Q., Goh H. H. Thermal fault diagnosis of electrical equipment in substations using lightweight convolutional neural network. IEEE Transactions on Instrumentation and Measurement, 2023, 72: 5005709. DOI: 10.1109/TIM.2023.3240210.

[6] Li J., Xu Y., Nie K., Cao B., Zuo S., Zhu J. PEDNet: A lightweight detection network of power equipment in infrared image based on YOLOv4-Tiny. IEEE Transactions on Instrumentation and Measurement, 2023, 72: 5004312. DOI: 10.1109/TIM.2023.3235416.

[7] Xiang S., Chang Z., Liu X., Luo L., Mao Y., Du X., Li B., Zhao Z. Infrared image object detection algorithm for substation equipment based on improved YOLOv8. Energies, 2024, 17(17): 4359. DOI: 10.3390/en17174359.

[8] Siami M., Barszcz T., Wodecki J., et al. Semantic segmentation of thermal defects in belt conveyor idlers using thermal image augmentation and U-Net-based convolutional neural networks. Scientific Reports, 2024, 14: 5748. DOI: 10.1038/s41598-024-55864-2.

[9] Resendiz-Ochoa E., et al. Application of thermography and convolutional neural network to diagnose mechanical faults in induction motors and gearbox wear. Applied System Innovation, 2024, 7(6): 123. DOI: 10.3390/asi7060123.

[10] Yao S., Liu Z., Zhao Y., et al. Automatic diagnosis of heating in oil-filled terminals of cables. Scientific Reports, 2025, 15: 38733. DOI: 10.1038/s41598-025-22506-0.