基于电力大数据的变电主设备缺陷演化规律红外成像分析方法

针对目标区域与背景区域混杂，变电主设备缺陷缺陷演化规律不明的问题，提出了基于电力大数据的变电主设备缺陷演化规律红外成像分析方法。应用红外成像技术采集变电主设备红外图像后，使用Otsu算法分割变电主设备红外图像内目标区域和背景区域；以分割后的变电主设备红外图像和电力大数据作为输入，通过混合深度学习神经网络模型输出变电主设备缺陷检测结果；将变电主设备缺陷检测结果输入到RFPA2D软件内，分析变电主设备缺陷检测结果基元破裂情况，得到变电主设备缺陷演化规律分析结果。实验结果表明：该方法采集变电主设备图像与其实际图像吻合度较高；分割变电主设备红外图像目标区域与背景区域时，受对比度影响较小；可有效检测变电主设备缺陷类型和分析其缺陷演化规律。

Infrared Imaging Analysis Method of Defect Evolution Law of Substation Main Equipment Based on Power Big Data

Aiming at the problem that the target area and background area are mixed, and the evolution law of substation main equipment defects is unknown, an infrared imaging analysis method based on power big data is proposed. After collecting the infrared image of the main substation equipment with infrared imaging technology, Otsu algorithm is used to segment the target area and background area in the infrared image of the main substation equipment. Taking the segmented infrared image of the main substation equipment and power big data as inputs, the defect detection results of the main substation equipment are output through the hybrid deep learning neural network model. Input the defect detection results of the main substation equipment into RFPA2D software, analyze the primitive fracture of the defect detection results of the main substation equipment, and obtain the analysis results of the defect evolution law of the main substation equipment. The experimental results show that the image of the main substation equipment collected by this method is highly consistent with the actual image. When segmenting the target area and background area of the infrared image of the main substation equipment, it is less affected by the contrast. It can effectively detect the defect type of the main equipment of the substation and analyze its defect evolution law.