The Online Detection Of Power Transformer Fault Based On Infrared Images

To ensure the reliability and security of power supply, the power transformer as one of the important equipments of power system, the operation of transformer fault on-line detection is very necessary. As a kind of effective means of fault detection, infrared diagnosis technology is widely used. It can detect and diagnosis of power transformer both internal and external defects, than fast imaging of the thermal state by infrared. At last by the analysis of the infrared images of the power transformer faults, it helps to defect the existing positioning, qualitative fault diagnose of the operation transformer.By studying of transformer with the infrared detection method, on the basis of summarizing the research results at home and abroad, combining the actual requirement of project, we put forward the technique of infrared diagnosis of power transformer on-line detection scheme. It can complete the effect of transformer faults online detection through the design of transformer faults online detection and diagnosis system.Firstly, transformer preprocessing of infrared images should be done. Through linear transformation and histogram equalization algorithm, to realize the image enhancement of infrared image; Then according to the noise characteristics of infrared image, the several classic denoising algorithm has been discussed, then using the algorithm of wavelet packet threshold to denoise the infrared image. Simulation results show that the algorithm effectively suppress the noise in the image signal, it is a good way to improve the quality of the image. Secondly, studied the infrared image segmentation technology in two kinds of classic segmentation algorithm, namely Ostu segmentation and edge detection method. The contrast experiment result, a Canny operator edge detection of image segmentation method has been adopted. Then according to the characters of each part of the power transformer of the image, with the improved Hu invariant moments to extract the characteristic value and use the nearest neighbor classifier to recognize image. Finally, the Visual Basic6.0has been chosen to design power transformer online fault detection and diagnosis system. The system for a certain area of monitoring, highlighted by fault threshold temperature and temperature gradient of double judgment to predict equipment running status; Or according to the "image characteristics determine method" and "fuzzy temperature difference method", and combining with the type of infrared image in the database information to judge the type of transformer faults, thus, transformer fault on-line detection of infrared image has been basically achieved.