Research On Thermal Fault Diagnosis Method Of Substation Equipment Based On Infrared Image

With the proposal and development of smart grids,higher requirements are put forward for the safety,economy,and flexibility of the operation of each component of the grid.Substation equipment is an important equipment required for power transmission and conversion,and its operating status directly affects the overall working conditions of the power grid.However,most of the current maintenance work for substation equipment uses traditional methods,which are inefficient and time-consuming and labor-intensive.Based on the advantages of infrared imaging technology and the characteristics of thermal faults in substation equipment,this paper proposes a method for diagnosing thermal faults in substation equipment,which realizes the location and diagnosis of thermal faults in substation equipment,and gives treatment suggestions based on the results to improve This improves the stability and reliability of the work of substation equipment.The main contents of this paper are as follows:(1)The reasons and types of noise generated during the shooting and transmission of infrared images of substation equipment are analyzed.Various denoising algorithms are used for comparison experiments according to the characteristics of the noise.Finally,the median filter algorithm based on wavelet transform is used to complete the denoising,and the Retinex algorithm is completed.Image enhancement makes the edge details of the image more obvious.The experimental results show that the denoising and enhancement algorithm used in this paper can better complete the image preprocessing.(2)Aiming at the characteristics of the infrared image of substation equipment,an infrared image segmentation method based on the improved Lazy Snapping algorithm is proposed to complete the feature extraction of substation equipment,and compared with other segmentation methods.The experimental results show that the algorithm proposed in this paper can realize the transformation of standard Lazy Snapping algorithm from interactive to non-interactive operation,and can accurately extract the device from the background and achieve good segmentation results.(3)For the pre-processed and segmented image,according to the geometric invariance of the substation equipment,the method of combining the geometric invariant moment and the support vector machine is used to complete the recognition of the substation equipment.Two invariant moments of Hu matrix and Zernike matrix are used for sample feature extraction and support vector machine database formation.The support vector machine training model is used to identify substation equipment.The experimental results show that Zernike invariant moments have higher recognition The accuracy has reached 95%.(4)The temperature recognition of the infrared image of the substation equipment has been completed.According to the colorimetric bar and temperature information provided by the infrared imager,the fitting of the temperature field is realized by the square matching of the pixel value Euclidean distance.The realization results show that the temperature field fitted in this paper can intuitively reflect the temperature distribution during the operation of the equipment and the temperature change characteristics during faults,and provide a basis for the subsequent fault diagnosis plan.(5)Combining the thermal fault characteristics of substation equipment and related document requirements,a fault diagnosis plan is given and an online diagnosis platform is designed to complete the fault diagnosis of substation equipment.