基于红外图像的电力变压器故障的在线检测

发布时间：2018-07-20 21:51

【摘要】：电力变压器作为电力系统运行的重要设备之一,为保证供电的可靠性和安全性,对运行的变压器进行故障在线检测非常必要。红外诊断技术作为一种有效的故障检测手段被广泛采用。它可以检测和诊断电力变压器大量的内部、外部缺陷,快速对设备热状态进行红外成像,通过对电力变压器故障的红外图像的分析,从而对运行的变压器存在的事故隐患和缺陷进行定位、定性的故障诊断。在对变压器的红外检测方法进行研究,总结国内外研究成果的基础上,结合课题的实际要求,提出了红外诊断技术对电力变压器的故障进行在线检测的方案。通过设计的变压器故障在线检测与诊断系统,完成变压器故障的在线检测的效果首先对采集的变压器红外图像进行预处理。通过线性变换和直方图均衡化算法,实现红外图像的图像增强；然后根据红外图像的噪声特点,讨论了几种经典的去噪算法,采用小波包阈值算法对红外图像进行消噪。仿真结果显示,该算法有效地抑制了图像中的噪声信号,很好的改善了图像的质量。其次,研究了红外图像分割技术中两种经典的分割算法,即边缘检测和Ostu分割法。对比实验结果图,采取了Canny算子边缘检测的图像分割法。然后根据电力变压器各部位图像的特征,采用改进的Hu不变矩提取特征值,并利用最近邻分类器进行图像识别。最后,选用Visual Basic6.0设计了电力变压器故障在线检测与诊断系统。该系统针对某区域进行重点监测,通过故障温度阈值与温度变化率的双重判断预测设备的运行情况；或根据“图像特征判断法”与“模糊温差法”,并结合红外图像数据库中的信息判断变压器故障的类型,基本实现了变压器红外图像故障的在线检测。
[Abstract]:Power transformer is one of the most important equipments in power system operation. In order to ensure the reliability and safety of power supply, it is necessary to detect the fault of the running transformer on line. Infrared diagnosis technology is widely used as an effective method of fault detection. It can detect and diagnose a large number of internal and external defects of power transformers, and fast infrared imaging of the thermal state of power transformers, through the analysis of infrared images of power transformer faults, In order to locate the hidden trouble and defect of the running transformer and diagnose the fault qualitatively. On the basis of studying the infrared detection method of transformer and summarizing the research results at home and abroad, combined with the practical requirements of the subject, the paper puts forward a scheme of on-line detection of power transformer fault by infrared diagnosis technology. Through the design of transformer fault on-line detection and diagnosis system, the effect of on-line detection of transformer fault is completed. First, the infrared image of transformer is preprocessed. Infrared image enhancement is realized by linear transformation and histogram equalization algorithm. Then according to the noise characteristics of infrared image, several classical denoising algorithms are discussed, and wavelet packet threshold algorithm is used to de-noise infrared image. Simulation results show that the algorithm can effectively suppress the noise signal in the image and improve the image quality. Secondly, two classical segmentation algorithms in infrared image segmentation, namely edge detection and Ostu segmentation, are studied. Compared with the experimental results, the image segmentation method based on Canny operator edge detection is adopted. Then, according to the features of power transformer image, the improved Hu moment invariant moment is used to extract the feature value, and the nearest neighbor classifier is used to recognize the image. Finally, the on-line fault detection and diagnosis system of power transformer is designed with Visual basic 6.0. The system focuses on monitoring a certain area and predicts the operation of the equipment by double judgment of fault temperature threshold and temperature change rate, or according to "image feature judgment method" and "fuzzy temperature difference method". Combined with the information in the infrared image database to judge the type of transformer fault, the on-line detection of transformer infrared image fault is basically realized.
【学位授予单位】：安徽理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM41

【参考文献】