Research On Circuit Breaker Fault Detection Technology Based On Infrared Imaging

The circuit breaker is used to interrupt the current in the normal working state of the power system and the short-circuit current when a fault occurs.It is an important part of the power grid system.Therefore,the thermal failure of the circuit breaker has become one of the main factors affecting the normal operation of the power grid system.Infrared detection technology has the advantages of strong adaptability,non-contact,etc.It is widely used in the maintenance of circuit breakers and other power equipment,but most of the domestic use of manual detection methods,its efficiency is low and time-consuming.This article takes the ceramic pillar type SF₆ high-voltage circuit breaker as the research object,and proposes an automatic analysis and processing method of the infrared image of the circuit breaker,which can realize the automatic positioning and discrimination of the thermal fault of the high-voltage circuit breaker.The specific contents are as follows:?1?Study the method of infrared image preprocessing of circuit breakers.By analyzing the noise characteristics of infrared images of high voltage circuit breakers and common types of noise,compare several commonly used denoising algorithms,and use non-local mean filtering based on integral image technology The algorithm plus median filtering algorithm filters the noise present in the infrared image of the circuit breaker,and then enhances the edge of the infrared image of the high voltage circuit breaker through the L0norm smoothing algorithm.?2?In view of the defect that the random walk image segmentation algorithm needs to manually calibrate the seed points,a fast fuzzy C-means clustering algorithm based on the maximum and minimum distance method is proposed for the pre-segmentation of infrared images of high-voltage circuit breakers and processed by morphology It provides seed points for the random walk algorithm and realizes the automatic image segmentation of the random walk algorithm.After experimental verification,the image segmentation effect of the algorithm adopted in this paper is closer to the complete image segmentation,which can achieve accurate image segmentation of the high-voltage circuit breaker.?3?Hu invariant moments are used to extract the eigenvalues??of the segmented images,and then the support vector machine is used for recognition training to obtain the high-voltage circuit breaker recognition model,which realizes the infrared image recognition of each component of the high-voltage circuit breaker.The experimental results show that the recognition The recognition accuracy of the model reaches 98.75%,which meets the expected requirements.?4?Study the heating characteristics of high-voltage circuit breakers,and combine with the"Code for Infrared Diagnosis of Live Equipment"to construct a circuit breaker thermal fault diagnosis process?5?Using Matlab APP designer to develop a software platform for automatic diagnosis of circuit breaker thermal faults,and its accuracy was verified through experiments.