A Study On Fault Diagnosis For Electrical Equipment Based On Infrared Thermography

With the development of automatic control and image processing, the fast and precise fault diagnosis for electrical equipment has become one of the necessary parts for everyone. If the diagnostic system can find the trouble before electrical equipment broken, it will be able to avoid the damage of the power system. And the fault diagnosis scheme for electrical equipment using infrared thermography in this thesis can achieve the objectives above with better operating efficiency and detection accuracy.A study on fault diagnosis for electrical equipment based on infrared thermography is analyzed and designed in this thesis. Firstly, the processing is discussed. It is divided into three steps: finding the region of interest, extracting the descriptive information in the regions, and classifying the power equipment state to give the result. For the three steps of the algorithm, the results evaluation criteria and test accuracy is analyzed. Then, the three steps of diagnostic procedures is detailed analyzed on there algorithms and conditions. In the step of finding the region of interest, the Otsu threshold segmentation algorithm is used to separate the foreground and background image. In the step of extracting the descriptive information in the regions, the expansion with restrictions algorithm is used to expand the hotspots. In the step of classifying the power equipment state to give the result, the qualitative and quantitative analysis method are combined to analyze the three results of the experiment, and the result which has the highest priority will be returned. Finally, 200 experiment samples which have the known results are used to test the fault diagnosis scheme in this thesis. The causes of wrong diagnostic happening are analyzed, and the way of improving the accuracy is discussed.Compared with the traditional maintenance, the fault diagnosis scheme for electrical equipment in this thesis needs less processing time and labor costs. And the power system can also run security efficiently in a long time.