Optimization Of PSS Parameters Based On Low Frequency Oscillation Mode Identification In Power System

With the continuous development of China’s national power industry,the workload of power inspection is also gradually increasing.The traditional human inspection has problems such as low work efficiency and great security risks.In order to solve this problem,this paper is in the background of the booming power inspection of drones,the insulator fault is diagnosed by image processing technology.First of all,the existing problems of electric power manual inspection and the advantages of electric inspection of drones are analyzed.Then,the research status of composite insulator string-breaking identification in aerial images at home and abroad is analyzed and studied.With the phenomenon of composite insulator stringing as the research goal.From the image insulator image pre-processing,insulator image segmentation,insulator broken string diagnosis of the current research situation in-depth analysis of domestic and foreign research situation.On the basis of comparison with the relevant technology of image processing of insulator break-string recognition algorithm.A grabcut algorithm-based split insulator is proposed,the method of identifying whether an insulator is broken by comparing the size of the filler area.1.The current situation of electric inspection of drones at home and abroad is analyzed.Depending on the characteristics of the drone aerial image,the following actions were selected for the appropriate image pre-processing method:(1)Different ways of grayscales and RGB conversion to HSV color space for the original image.The underlying picture is split by the conversion of these color spaces as a threshold.(2)The grayscales transformation of the original image is carried out.These include linear transformation of grayscales,nonlinear variation of grayscales,and equalization of histograms.(3)The truncated linear variation and histogram equalization in the linear transformation are selected by contrast as the preprocessing method for image-enhanced contrast.Different grayscales and HSV space conversions are paved for the threshold segmentation that needs to be done later.2.plot the part in the composite insulator.In this paper,the Grabcut algorithm in the graph theory is used to split the composite insulator.The principle of the Grabcut algorithm and the implementation process in opencv are analyzed.The method of automation of Grabcut algorithm is derived from theoretical analysis.In this paper,the initial automation of the Grabcut algorithm is used to achieve the segmentation of the target object,the process includes threshold segmentation and maximum connectivity domain extraction.In the implementation process,the appropriate threshold segmentation algorithm is selected first.This paper compares the basic threshold segmentation method with the best threshold segmentation based on Otsu,getting the best threshold segmentation based on Otsu works better.Otsu is then split in turn based on the previous preprocessed images.Find out the complementary split pictures to achieve the effect of drying,and then through theoretical analysis to verify the correctness of the experimental effect.The black target area used in this article.So the use of morphological corrosion to further dry.Finally,the initialization of the Grabcut algorithm is completed by the maximum connectivity domain extraction.The segmentation of composite insulators is realized by this method.3.During the diagnosis of composite insulator break-strings,in this paper,a new method of area segmentation is used to diagnose whether the split insulator is broken.First of all,the process of drying this paper compares the different morphological methods of the insulator split image processing effect.Choose a method that combines good and simple-to-operate morphological corrosion with Otsu’s optimal threshold segmentation as a drying algorithm.Secondly,the insulators are framed according to the resulting insulator picture,Flood fill,Logical operations to extract the area filled into the space of the insulator.According to the formula of the area of the filler,the composite insulator is broken.Finally,it analyzes the shortcomings and shortcomings of the algorithm in this paper.It is obtained through theoretical analysis and legend analysis.When the target composite insulator has multiple broken strings and a single broken string fault exists at the same time,it may occur that the largest striking position can only be positioned.4.In the process of verifying the validity of the algorithm in this paper.This article selected90 pictures include: 30 single insulator strings,30 double insulator strings,and 30 random sheets.To verify the validity of the algorithm in this article.In the paper implementation process,Threshold segmentation and Human-computer interaction Grabcut algorithm are applied.In this paper,two methods are compared with the experimental effect of this method.The advantages and disadvantages of single pair insulator string in the application of this algorithm in different backgrounds are analyzed in detail.At the same time,the laboratory data of this paper are compared with the experimental data of other papers,the superiority of the algorithm to the segmentation of composite insulators is obtained.The Grabcut automation algorithm is a good way to separate the target insulators,the accuracy of segmentation can reach more than96.6%.The proposed area-judgment method can detect more than 98% of insulators with broken string faults,the error detection rate for trouble-free is less than 6.7%.