Insulator fault diagnosis method for renewable energy generation systems based on LAF-YOLOv11

Abstract

Insulators in the transmission lines of renewable energy generation systems are susceptible to faults caused by contamination, mechanical damage, and surface cracks. Therefore, accurate insulator fault diagnosis is essential for ensuring the safe and stable operation of such systems. Optical image-based detection has attracted increasing attention due to its non-contact nature, visual interpretability, and suitability for unmanned aerial vehicle (UAV) inspection. However, existing methods still suffer from limited detection accuracy under complex illumination, background interference, weak-texture defects, and small-scale fault conditions. To address these issues, this paper proposes an optical image-based insulator fault diagnosis method for renewable energy generation systems using an improved YOLOv11 algorithm, termed light‑adaptive enhancement and multi‑scale attention fusion YOLOv11 (LAF‑YOLOv11). Based on YOLOv11, a light-adaptive enhancement module is introduced to improve image feature quality under strong light, shadow, and low-illumination conditions. Meanwhile, a multi-scale attention mechanism is integrated to enhance the optical texture representation of insulator cracks, flashover traces, and damaged regions. Experimental results show that the proposed method achieves a mean average precision (mAP) of 96.5% on an insulator optical inspection image dataset, which is 4.1% higher than that of the original YOLOv11. The detection speed also meets the requirements of real-time inspection, demonstrating the effectiveness of the proposed method for safe operation and maintenance of renewable energy generation systems.