Research On Precise Fault Location Method Of Substation Equipment Based On Infrared Image

With the rapid development of social economy in China,the demand for electricity from industrial production and residents' lives is increasing.In order to meet the increasing demand,the scale of power grid is also expanding,and the probability of various equipment failures in substations is also increasing.Therefore,it is of great significance to find out the faults of various electrical equipment in time and locate the specific location of the faults accurately so that the staff can check the faults in time and prevent the enlargement of the faults.Traditional substation fault detection is a method of equipment outage and line patrol by staff,which usually takes a lot of time,resulting in energy waste.This paper designs a fault detection system for substation equipment based on infrared image,which can locate the fault position accurately without power failure,contact or damage,detect and check the fault in time,prevent casualties and ensure reliable power supply for power system.The main contents of this paper include the following aspects:(1)The research status of infrared diagnosis technology at home and abroad is analyzed,and several important nodes in its development history are discussed.By comparing the development status of infrared image segmentation algorithms at home and abroad,and the theoretical basis of fault detection system segmentation is given.(2)Through the analysis of the imaging principle of infrared image,it is concluded that the noise pollution of infrared image is inevitable.Through the statistical analysis of experiments,it is concluded that impulse noise is the most common noise among the large amount of noise affecting infrared image,and it is also the biggest factor causing the low quality of infrared image.Then,several methods of impulse denoising are introduced,including mean filtering,median filtering,adaptive filtering,discrete cosine filtering and wavelet transform filtering.Then the filtering effect of several denoising methods is validated by infrared image of fault arrester with impulse noise.Secondly,the typical infrared image segmentation methods are discussed in terms of threshold,morphology,region segmentation and neural network,and their segmentation principles are introduced.It is pointed out that each method has some problems in infrared image segmentation and can not achieve accurate location of fault areas.(3)In view of the shortcomings of several segmentation methods proposed in the previous chapter,this paper proposes a segmentation method based on OTSU and improved region growing algorithm.In order to avoid the interference of complex large-area background in fault location,this paper adopts OTSU algorithm based on global threshold to segment high-temperature regions of interest including fault regions,and then uses improved region growing algorithm to segment fault regions to achieve accurate fault location.Because the traditional region growing algorithm needs to select seed points artificially through the interactive interface,which has a large workload and can not achieve automatic location,this paper uses a rectangular window to traverse the entire image,and finds the region with the largest mean value in the rectangular window.The middle point of the region is used as the initial seed point to start growing,thus eliminating the manual interaction process and improving the speed of segmentation.In order to compare the advantages of the improved algorithm,this paper uses the infrared image of fault CT and fault knife gate to verify the segmentation effect.From the segmentation results,we can see that the improved region growing algorithm has fast segmentation speed and accurate location,which is better than the segmentation results of several segmentation algorithms introduced in the previous chapter.In order to objectively evaluate the segmentation effect of several segmentation algorithms,this paper analyzed the segmentation effect from overlap area and error segmentation area.From the evaluation results table,it can be seen that the improved region growing algorithm can segment most of the fault areas without producing the wrong segmentation area,and the segmentation accuracy is much better than other segmentation algorithms.(4)The Qt image interface tool compiled by C++ is used to design the fault detection system,initialize the environmental configuration coefficient of the detection system,and use the mechanism of signal and slot to realize the communication between buttons and algorithm calls,so that the interactive interface can read the original infrared image,preprocess the image,detect the fault and locate the fault accurately.