摘要
高压电气设备是变电站安全稳定运行的重要保障。由于变电站通常会暴露在各种自然气候条件下,例如高温、低温、湿度、灰尘、盐雾等,这些因素都可能影响电气设备绝缘材料的性能和检测设备的准确性。为此,通过对变电站高压电气设备的红外辐射图像开展增强处理并对其特征进行校准,在改进YOLOv5网络算法中引入损失函数和数据增强技术,实现对变电站高压电气设备绝缘状态的有效识别。实验结果表明:研究方法对高压电气设备的绝缘状态的增强检测效果更好,且研究方法的IoU指标明显高于文献方法。
High-voltage electrical primary equipment is an important guarantee for safe and stable operation of transforming substations.Performances of electrical equipment insulation materials and accuracies of detection equipment are susceptive to vagaries of whether conditions of substations such as high and low temperatures,humidity,dust,salt spray,etc.In view of this the present work made a preliminary attempt to utilize advanced AI in the identification of insulation state of substation high-voltage electrical equipment.The main efforts entailed the enhanced treatment and feature calibration of equipment infrared radiation images,and the introduction of loss functions and data augmentation into YOLOv5 network algorithm.The proposed idea was indicated by experiment superior in enhancement detection of equipment insulation state and IoU index compared with the selected reference method.
作者
汪鹏
WANG Peng(China Energy Construction Group Yunnan Electric Power Design Institute Co.,Ltd.,Kunming 650011,China)
出处
《电工技术》
2024年第18期186-188,共3页
Electric Engineering
关键词
变电站
高压电气
电气设备
绝缘状态
检测方法
transforming substation
high-voltage electrics
electrical equipment
insulation state
detection method
作者简介
汪鹏(1988-),高级工程师,研究方向为变电站及光伏电站电气设计。
