Surface discharge analysis method based on nonlinear characteristic of leakage current of icing insulator

Abstract

The invention provides a surface discharge analysis method based on non-linear characteristics of leakage current of an icing insulator. The method is suitable for monitoring and guaranteeing the safe operation of the icing insulator. Firstly, the chaos of a leakage current time sequence is detected by estimating the maximum Lyapunov exponent by using an embedded space theory and a phase space reconstruction method. The dynamic behavior of the leakage current is studied by utilizing a recurrence plot technology, clear visual analysis is performed on the dynamic discharge behavior of the icing insulator through graphic texture transformation, and an index for quantifying the regularity of the leakage current is provided on the basis. Quantitative analysis is carried out on the cooperative relationship between the arc length of the electric arc extracted through image processing and the leakage current, and the change rule of the leakage current is determined. And the method is of great significance for perfecting an icing insulator flashover risk prediction technology.

Description

Technical Field

[0001] The invention relates to a method for analyzing surface discharge of ice-covered insulators, and in particular to a method for analyzing surface discharge based on nonlinear characteristics of leakage current of ice-covered insulators. Background Art

[0002] As my country vigorously develops world-leading ultra-high voltage (UHV) transmission and transformation projects, the number of ultra-high voltage (UHV) transmission lines continues to increase, and the risk of ice flash accidents is also increasing. In severe cold weather, the growth of ice ridges severely limits the electrical performance of insulators, which may cause corona discharge, arc development and even flashover, which will seriously endanger the operation of the power grid. Although my country has increased the investment in human and scientific and technological resources to build a strong and reliable power grid, flashover accidents caused by line icing are still inevitable. Therefore, it is necessary to reasonably and effectively monitor the discharge of insulators on transmission lines and improve the ability to judge the safe operation of iced insulators.

[0003] A great deal of research has been done to investigate the characteristics and processes of flashover characteristics of iced insulators. Mature experimental and mathematical models have been developed to predict the flashover voltage of iced insulators. Current research is devoted to developing state detection methods aimed at preventing ice flashover. The main objectives of these works are as follows: 1) to measure and analyze the surface leakage current during icing and melting; 2) to establish diagnostic criteria for detecting the severity of icing; and 3) to predict the probability of flashover occurrence. Previous studies have shown that leakage current can reflect the combined effects of icing severity, operating voltage, and environmental parameters. In addition, it can systematically provide information on the icing level on the insulator surface, and its amplitude fluctuations and waveform changes correspond to the irregular discharge behavior on the insulator surface, which is directly related to the surface characteristics of iced insulators. The ice flash process can be divided into four stages. In addition, extracting leakage current features in the time-frequency domain can reveal the correlation between icing rate, melting rate, threshold change pattern of arc discharge, flashover, and environmental factors such as ice thickness and icing type. The time-frequency characteristics of leakage current are closely related to arc characteristics. The intensity of the discharge arc changes periodically as the leakage current flows through the insulator surface. By using signal and image processing technology, the leakage current waveform and arc characteristics during ice flashover were obtained.

[0004] The study of leakage current characteristics of iced insulators is of great value in predicting the risk of ice flashover. However, during ice flashover, the surface discharge behavior occurs randomly, while the leakage current as a nonlinear time series is complex and irregular, which makes it difficult to accurately match the surface discharge activity of the insulator. In addition, the relationship between leakage current and surface discharge behavior is still in the qualitative description stage. Therefore, analyzing the nonlinear characteristics and establishing a quantitative relationship between surface discharge behavior and leakage current are the key to strengthening the detection technology of iced insulators. Summary of the invention

[0005] The purpose of the present invention is to propose a surface discharge analysis method based on the nonlinear characteristics of iced insulator leakage current. Based on the indoor ice flashover test under DC voltage, the dynamic characteristics of leakage current and arc characteristics obtained by image processing at different stages are studied. The chaotic and nonlinear characteristics of liquid crystal are studied respectively by applying embedded space theory combined with phase space reconstruction method and recursive graph technology. Through the coordinated quantitative analysis of leakage current and arc characteristics, it is determined that the regular changes of leakage current are suitable for monitoring and ensuring the safe operation of iced insulators.

[0006] To achieve the above object, the present invention adopts the following technical solution, a surface discharge analysis method based on the nonlinear characteristics of iced insulator leakage current, the specific steps are:

[0007] 1) Build an experimental platform, artificially simulate the flashover experiment of rime and icing, study the arc state of DC iced insulators in a laboratory environment, and obtain the original image data by shooting with a monitoring camera;

[0008] 2) Process the collected arc grayscale image and extract arc features.

[0009] 3) The flashover process on the surface of the ice-covered insulator is analyzed, and the time waveforms of the arc and the detected voltage and current changes during the period from stable discharge to flashover on the surface of the ice-covered insulator are analyzed.

[0010] 4) Perform nonlinear analysis on the leakage current waveform obtained by monitoring in the previous step.

[0011] 5) The recursive graph technology is used to transform the one-dimensional nonlinear time series of leakage current into two-dimensional matrix data, effectively extracting its nonlinear characteristics for outdoor insulator performance testing.

[0012] 6) Based on the recursive graph technology used in 5), the chaos of the leakage current time series is reflected by calculating the Lyapunov exponent and providing a quantitative description of the state of the chaotic system.

[0013] 7) The discharge characteristics of iced insulators during surface flashover are analyzed by recursive graph technology to reveal its relationship with the nonlinear change of leakage current.

[0014] Furthermore, the entire experimental process was carried out in an artificial low-temperature climate chamber, and its experimental device and parameter settings were in accordance with the IEEE 1783 / 2009 standard.

[0015] Beneficial effects:

[0016] The present invention proposes a surface discharge analysis method based on the nonlinear characteristics of the leakage current of iced insulators. In order to study the surface flashover characteristics of outdoor high-voltage DC insulators under extreme icing conditions, a laboratory ice flashover test was carried out under DC voltage. The dynamic characteristics of leakage current and arc at different stages during DC ice flashover were studied. The existence of chaos in the leakage current was confirmed by using embedded space theory and phase space reconstruction method, and its nonlinear characteristics were analyzed by recursive graph technology. The collaborative analysis of leakage current and arc characteristic quantities verified the variation law of leakage current. It has important value and significance for ice flashover warning. Its specific beneficial effects include:

[0017] 1. In order to improve the image quality, a combination of traditional image processing and morphological processing is used. First, the median filtering method of the 3*3 template is selected to denoise the image. The processed image is as follows Figure 1 (b) As the denoising process has a certain blurring effect on the image, the power transformation is used to perform nonlinear grayscale transformation to enhance the contrast and clarity of the image and ensure the difference in grayscale values ​​between pixels. The enhanced image is shown in Figure 1 (C) As shown. In edge segmentation, various edge detection operators in traditional image processing will mis-extract noise areas to varying degrees. Therefore, a watershed algorithm based on morphological marker control is used. A closed and accurate arc contour is obtained, as shown in Figure 1 (d) As shown. Finally, the filling algorithm is applied to fill the arc edge image to obtain the arc area as shown in Figure 1 As shown in (e), a more accurate arc circumference is extracted based on this area.

[0018] 2. In order to effectively extract nonlinear features from the leakage current time series, the recursion graph technique is used to convert the chaotic time series into a two-dimensional matrix. Each time point on the axis represents the moment of a different state reconstructed from the phase space of the leakage current series. The matrix elements (recursion points) correspond to the combination of time points associated with similar states in the dynamic system represented by the original time series. Therefore, the recursion graph intuitively displays the phase space trajectory of the original system at different comparable time points in the phase space. When surface discharge occurs on ice-covered insulators, the recursion graph technique can analyze the discharge characteristics of ice-covered insulators during surface flashover and reveal its relationship with the nonlinear change of leakage current. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Through the detailed implementation scheme and algorithm flow described below, the various advantages of the present invention can be presented more clearly to researchers in the field. The accompanying drawings will serve as auxiliary explanations. The accompanying drawings are as follows:

[0020] Figure 1 Image processing process diagram

[0021] Figure 2 Schematic diagram of the experimental setup

[0022] Figure 3 Schematic diagram of leakage current, measured voltage and arc length recorded for different stages

[0023] Figure 4 Schematic diagram for calculating m and τ using the CC method

[0024] Figure 5 Schematic diagram of calculation of maximum Lyapunov exponent of leakage current at each stage of flashover process

[0025] Figure 6 Recursion diagram at different stages DETAILED DESCRIPTION

[0026] The invention disclosed herein will be described in more detail below with reference to the accompanying drawings:

[0027] The purpose of the present invention is to propose a surface discharge analysis method based on the nonlinear characteristics of iced insulator leakage current. Firstly, based on the indoor ice flashover test under DC voltage, the dynamic characteristics of leakage current and arc characteristics obtained by image processing at different stages are studied. The chaotic and nonlinear characteristics of leakage current are studied respectively by applying embedded space theory combined with phase space reconstruction method and rapid prototyping technology. The regular changes of leakage current are determined by the coordinated quantitative analysis of leakage current and arc characteristics.

[0028] 1. The experiment was conducted in a cold room with dimensions of 4.8 m × 2.8 m × 3.5 m. Figure 2The schematic diagram of the experimental setup is described. The artificial cold chamber includes a grid ceiling for heat exchange, a water spraying system with air atomizing nozzles, a diffuser honeycomb panel with a fan system, and a high-voltage DC power system with a rated voltage of 120 kV and a rated power of 240 kVA. Before the start of the experiment, the surface of the insulator samples was cleaned with deionized water and then dried at room temperature (20 °C) for more than 24 hours. These tests were conducted according to the IEEE 1783 / 2009 standard. In order to ensure that the samples were at the same temperature as their environment, after placing them in the cold chamber, the temperature was reduced to -12 °C and kept constant for more than 15 hours. The samples were then exposed to a DC voltage of 75 kV, sodium chloride was mixed with deionized water to produce a water conductivity of 30 μS / cm, and cooled water droplets were sprayed on the surface at a wind speed of 3.3 m / s to simulate the icing process in the natural environment. The duration of the icing experiment was 60 minutes. The process was recorded using a high-speed camera. The Oscilloscope software that comes with the oscilloscope is used to reproduce the stored data. At the same time, a database is established to import all the collected magnetic field signal data, and the magnetic field data collected under different cable conditions are classified and stored to realize the integration of state information. After that, the voltage is gradually increased at a rate of 3.0kV / s. After reaching the estimated flashover voltage, the current-voltage level is maintained for 15 minutes. The experiment is carried out at a sampling frequency of 10kHz. A high-speed camera that captures 256 gray levels records the surface discharge phenomenon of the insulator at a rate of 6000 frames per second (fps).

[0029] 2. Process the grayscale image captured by the camera. The processing process is as follows: Figure 1 As shown in Figure 1, the image is first denoised. Since the image taken by the high-speed camera is a grayscale image, it can be directly processed for noise. By comparing the commonly used methods, it is found that the median filter is better than the smooth spatial linear filter, so the median filter method of the 3*3 template is selected; the image after denoising is as follows Figure 1 (b) As shown: Since denoising has a certain blurring effect on the image, it is necessary to enhance the filtered image. This paper uses power transformation to perform nonlinear grayscale transformation to enhance the contrast and clarity of the image and ensure the difference in grayscale values ​​between pixels, which is convenient for the next step of image processing. The enhanced image is shown in Figure 1(c) As shown. Then the most critical image edge segmentation is carried out. This paper uses the morphologically based labeled watershed segmentation algorithm. This is because on the one hand, it is difficult to select a suitable threshold in traditional image processing; on the other hand, there is still a lot of noise in the image after denoising. Various edge detection operators in traditional image processing methods will mistakenly extract noise areas of varying degrees, and even ignore certain arc edge features due to noise interference; therefore, the labeled watershed segmentation algorithm is selected. Its essence is to imitate the immersion process of topographic maps and segment the image according to the regional characteristics of the image. It combines the advantages of regional growth and edge detection to obtain closed, connected, and accurately positioned contours, such as Figure 1 Finally, the filling algorithm is applied to fill the arc edge image to obtain the arc area as shown in (d). Figure 1 As shown in (e), the arc features are extracted based on the area.

[0030] 3. Analysis of the flashover process of iced insulators: Figure 3 The time waveform of the arc and the changes in the detected voltage and current from the stable discharge on the surface of the ice-covered insulator to the flashover are shown. It goes through three stages: arc formation, arc development, and near flashover. In the first stage, the main discharge mode is glow discharge, and the appearance of filamentary partial discharge causes the voltage and current waveforms to have sawtooth characteristics, such as Figure 3 (a) and 3(b). Even though the transition from glow discharge to arc discharge begins at this stage, the duration of arc discharge is still very short, resulting in the leakage current being stably maintained at an amplitude of about 14 mA under DC voltage. In addition, the weak arc obtained by image processing maintains a stable length at this stage, with the average values ​​of 41.89 mm and 30.24 mm for positive and negative DC voltages, respectively, as shown in Figure 3(a). Figure 3 (c) is shown. Figure 3 During stage II from 3.15s to 3.29s in (b), LC increases from 6.97mA to 21.15mA, while the measured voltage decreases from 100.78kV to 84.95kV. The Joule heat of the leakage current increases the conductivity of the water film by melting the ice ridges. When a certain threshold is reached, the measured voltage drops sharply, indicating the transition from glow discharge to arc discharge. After that, the leakage current and the measured voltage return to normal levels, as shown in Figure 2. Figure 3 (b) This is because the arc reaches a critical size when the energy provided by the power supply to the arc plasma is lower than the heat loss of the arc column. Once this happens, the arc cannot continue to develop and the glow discharge mechanism is restored. Figure 3Comparison of (a) and 3(b) reveals that negative DC voltage leads to a significant increase in discharge activity during stage II. In fact, arcs appear and develop more frequently under negative DC voltage than observed under positive DC voltage. During stage III, the electric field intensity at the arc head increases significantly. As a result, a large amount of charge is transferred to the ice surface, causing leakage spikes and leading to flashover. Figure 3 In (b), the leakage current increases from 8.29 mA to 720.3 mA between 5.31 s and 5.41 s. Figure 3 In (a), the leakage current increases from -23.89mA to -522.1mA during 6.20s and 6.25s. In ten flashover experiments, when the leakage distance increases to 41%-52% and 46%-59% of the total distance under positive and negative DC voltages, respectively, spark discharge occurs spontaneously. Concentrated discharge channels appear, resulting in transient large currents. The inherent instability of spark discharge leads to obvious pulse characteristics. According to Figure 3 (a) and 3(b), during the flashover, the repetitive discharge frequencies under positive and negative DC voltages are 333 Hz and 833 Hz, respectively, while the corresponding maximum peak measurements are 720.3 mA and 522.1 mA, respectively.

[0031] 4. Analyze the nonlinear characteristics of the leakage current and calculate the maximum Lyapunov exponent of the leakage current time series. Chaotic systems refer to deterministic nonlinear systems that exhibit various characteristics. The main feature is their sensitivity to initial conditions, which causes the trajectories produced by closely spaced initial values ​​in the phase space to diverge exponentially over time. The Lyapunov exponent can reflect this characteristic and provide a quantitative description of the state evolution of the chaotic system, making multidimensional evaluation of the time series possible. It is defined as follows: Take the long-term evolution of an infinitesimal n-dimensional sphere in the n-dimensional phase space as an example. Due to the local instability of the system, the n-dimensional sphere will be transformed into an ellipsoid. The i-th Lyapunov index is defined by the length of the major axis of the ellipsoid, denoted by p i (t)

[0032]

[0033] The size of the Lyapunov exponent indicates the exponential rate of convergence or divergence of nearby trajectories in phase space. i <0 direction, the phase volume shrinks and the movement is stable. i >0 direction, the long-term behavior is sensitive to the initial conditions, and the motion is chaotic. i = 0 corresponds to the stability boundary, which is a critical case. Therefore, if the maximum L i If the value is greater than 0, it can be determined as a chaotic sequence.

[0034] This paper uses a small data method to calculate the maximum Lyapunov exponent of a chaotic time series. It does not need to calculate the entire Lyapunov exponent spectrum of the time series. It only needs to calculate the L max Whether it is greater than 0 determines the chaotic nature. Therefore, the small data method has higher calculation efficiency and accuracy. The calculation steps are as follows:

[0035] (1) Calculate the chaotic time series of leakage current with a delay of τ and an embedding dimension of m. According to Takens’ phase space reconstruction theory, the reconstructed phase space of the leakage current time series consists of point X i express:

[0036] X i =(x i , x i+τ , ..., x i+(m-1)τ ), X i ∈R m , (i=1, 2, ..., M) (2)

[0037] Where N represents the size of the data set, where N = M + (m-1) τ. The CC method is used to determine the values ​​of τ and m. The steps are as follows: First, define the correlation integral of the embedded time series:

[0038]

[0039] Among them, d ij =||X i -X j ||, when x < 0, θ(x) = 0, when x ≥ 0, θ(x) = 1. The distance in phase space is less than r(X i The vector of distance between them is called correlation vector. The proportion of these correlation vectors in all M(M-1) possible pairs is defined as correlation integral C(m, N, r, t). The study of CC algorithm is related to function S

[0040] S(m,N,r,t)=C(m,N,r,t)-C m (1, N, r, t) (4)

[0041] To choose a suitable τ, the time series is divided into t subsequences:

[0042]

[0043] When N→∞,

[0044]

[0045] When processing actual sequences, the time corresponding to the zero crossing point of S(m, r, t) or the minimum difference of S(m, r, t) with different radii is selected as the local maximum time interval. The difference between the maximum and minimum values ​​of r is defined as:

[0046] ΔS(m, t) = max{S(m, r j , t, N)}-min{S(m, r j , t, N)} (7)

[0047] According to the statistical conclusion, let m = 2, 3, 4, 5, The formula is as follows:

[0048]

[0049]

[0050]

[0051] Figure 4 The three corresponding curves of equations (8)-(10) are shown. Taking the leakage current data of the third stage under positive voltage as an example, the first zero point or the first local minimum The corresponding time τ = 14s is the appropriate delay time. cor The minimum point τ of (t) w = 28 s, which is also the selected time window value, from which m = 3. Similarly, for the first and second stages at 75 kV DC, the corresponding values ​​of t are 15 s and 10 s, and the values ​​of m are 3 and 4.

[0052] (2) Reconstruct the phase space: Based on τ and m{X j , j = 1, 2…, M} to reconstruct the phase space.

[0053] (3) Find the nearest neighbor: Find each point X in the phase space j The nearest neighbor of , and impose constraints on transient separation, that is,

[0054]

[0055] (4) Calculate the distance between adjacent points: For each point X in the phase space j , the distance between adjacent point pairs after the i-th discrete time step is calculated as:

[0056]

[0057] For each i, calculate the lnd of all j j The mean value of (i) is y(i):

[0058]

[0059] In (13), p represents the number of non-zero dj(i). A regression line is fitted using the least squares method. The slope of the line represents the Lmax of the ice flashover process. Figure 5 says It is clear that at each stage of the ice flashover process, L max are all greater than 0, which means that the LC time series of the flashover process of iced insulators has chaotic behavior and can be analyzed by chaos analysis. Figure 5 shows L max A downward trend from stage 1 to stage 2, and then an upward trend from stage 2 to stage 3. This change shows that stage 2 exhibits the most obvious chaotic characteristics of surface discharge on iced insulators, which is mainly due to the prevalence of glow discharge. However, as the leakage current increases, the local arc discharge becomes transient and unstable, which intensifies the nonlinear characteristics of the leakage current time series.

[0060] 5. Recursion graph analysis of leakage current. In step 4, it has been proved that the leakage current in the flashover process has chaotic characteristics. In order to effectively extract nonlinear features from the leakage current, the chaotic time series is converted into a two-dimensional matrix using a recursive technique. Each time point on the axis represents the moment of a different state obtained from the phase space reconstruction of the leakage current series. The matrix elements (recursive points) correspond to the combination of time points associated with similar states in the dynamic system represented by the original time series. As a result, the recursive graph intuitively displays the phase space trajectory of the original system at various comparable time instances in the phase space. When surface discharge occurs on the iced insulator, the recursive technique can analyze the discharge characteristics of the iced insulator during the surface flashover process and reveal its relationship with the nonlinear change of the leakage current. In addition, the dynamic changes in the arc discharge channel ultimately lead to flashovers that penetrate the two hardware accessories. Therefore, this paper focuses on the recursive characteristics of a large number of arc discharge stages. In addition, during the formation and development of arc discharge, the leakage current of the negative and positive arcs has a similar relationship with the arc characteristic quantity. Therefore, only the positive polarity characteristics are selected to analyze the flashover process. Figure 6 shows The RP diagrams of different stages under positive DC are shown. Typical waveforms T1 and T4 are extracted from stage I and stage III respectively.

[0061] Since the second stage lasts longer, two time intervals, T2 and T3, that exhibit significant arc discharge were selected. Each time interval is associated with 100ms of leakage current data. Using the calculation procedure mentioned in previous studies, a recurrence graph for each time interval was generated based on the calculated values ​​of m and τ.

[0062] The analysis shows that the recursive topology exhibits significant changes at different time periods. The gradual changes in the topology and fine structure of the recursive graph from the initial corona discharge to the near-flashover stage provide a valuable research method for studying the obvious characteristics and changes of the surface discharge of the insulator. The distribution of dense points on the recursive graph is related to the continuous arc burning, indicating the continuous behavior of the surface discharge of the iced insulator. In contrast, the blank area corresponds to the cessation of the surface discharge, indicating the transition between different discharge modes.

[0063] Figure 6 shows The results show a unique pattern of alternating large white areas and block areas composed of recursive points, indicating a significant change in the dynamic system and reflecting the transformation of the discharge mode on the surface of iced insulators. Specifically, Figure 6 a) Depicts the surface condition of the insulator during the transition from corona discharge to arc discharge in the second stage. The abundance of scattered recursive points indicates, to some extent, the randomness of corona discharge on the surface of the iced insulator. In addition, the presence of many vertical and horizontal line segments indicates that the leakage current state changes slowly within a specific time range, which indicates that the surface discharge on the insulator is not serious.

[0064] Figure 6 b) reflects the situation when a thick and bright white arc appears on the insulator surface. The number of scattered points decreases, while the time points are densely distributed along the main diagonal. The frequency of alternation between blank areas and dense recursive points increases. This feature effectively captures the periodic transformation occurring on the surface of the iced insulator and demonstrates the alternating formation and extinction of the arc discharge channel.

[0065] Figure 6 c) reflects the further development of the arc, in which the radius of the arc root and arc head in the arc channel increases, resulting in an increase in discharge energy. The gathering area of ​​the recursive point along the main diagonal is further expanded. This shows that the arc discharge intensity is gradually increasing and has become the main discharge mode at this stage. The discharge can be sustained and the discharge volume is gradually increasing.

[0066] Figure 6 d) depicts the proximity of flashover on the insulator surface. During this period, the arc can cross 2-3 insulator units, accompanied by observable arc drift. The dense points along the diagonal and the blank area show minimal alternation. This indicates that the nature of the transition is ambiguous at this stage and the arc continues to evolve. The blank area in the recurrence diagram further shrinks, indicating that the discharge phenomenon intensifies. This means that flashover is about to occur. However, during the flashover process, due to the unstable discharge of the spark, the instantaneous large current is accompanied by obvious pulse characteristics.

[0067] Based on the surface discharge analysis method of the nonlinear characteristics of iced insulator leakage current proposed above, the dynamic characteristics of leakage current and arc at different stages of DC ice flashover were studied. The existence of chaos in leakage current was confirmed by using embedded space theory and phase space reconstruction method, and its nonlinear characteristics were analyzed by recursive technology. The change of leakage regularity was verified by the coordinated analysis of leakage current and arc characteristic quantities, which is of great significance to the development of ice flashover warning technology for power equipment.

Claims

1. A surface discharge analysis method based on the nonlinear characteristics of iced insulator leakage current, characterized in that: The specific steps are: 1) Build an experimental platform, artificially simulate the flashover experiment of rime and icing, study the arc state of DC iced insulators in a laboratory environment, and obtain the original image data by shooting with a monitoring camera; 2) Processing the collected arc grayscale image and extracting arc features; 3) Analyze the flashover process on the surface of the ice-covered insulator, and analyze the time waveforms of the arc and the detected voltage and current changes during the period from stable discharge to flashover on the surface of the ice-covered insulator; 4) Perform nonlinear analysis on the leakage current waveform obtained by monitoring in the previous step; 5) The recursive graph technology is used to transform the one-dimensional nonlinear time series of leakage current into two-dimensional matrix data, effectively extracting its nonlinear characteristics for use Outdoor insulator performance testing; 6) Based on the recursive graph technology used in 5), the chaos of the leakage current time series is reflected by calculating the Lyapunov index and providing a quantitative description of the state of the chaotic system; 7) The discharge characteristics of iced insulators during surface flashover are analyzed by recursive graph technology, revealing its relationship with the nonlinear change of leakage current.

2. The surface discharge analysis method based on the nonlinear characteristics of iced insulator leakage current according to claim 1, characterized in that: In step 2), the grayscale image is processed by combining traditional image processing and morphological processing.

3. The surface discharge analysis method based on the nonlinear characteristics of iced insulator leakage current according to claim 1, characterized in that: The nonlinear characterization performed in step 4) is primarily characterized by sensitivity to initial conditions, resulting in trajectories for closely spaced initial values ​​in phase space that diverge exponentially over time.

4. The surface discharge analysis method based on the nonlinear characteristics of iced insulator leakage current according to claim 1, characterized in that: In step 6), the Lyapunov exponent (L i ) represents the exponential rate of convergence or divergence of nearby trajectories in phase space. i <0 direction, the phase volume shrinks and the movement is stable. i >0 direction, the long-term behavior is sensitive to the initial conditions, and the motion is chaotic. i = 0 corresponds to the stability boundary, which is a critical case. Therefore, if the maximum L i If the value is greater than 0, it can be determined that it is a chaotic sequence.

5. The surface discharge analysis method based on the nonlinear characteristics of iced insulator leakage current according to claim 1, characterized in that: In step 7), the topology of the recursion graph shows significant changes at different times. The gradual changes in the topology and fine texture of the recursion graph from the initial corona discharge to the stage close to the flashover provide a valuable method to study the unique characteristics and changes of the insulator surface discharge. The distribution of dense points in the recursion graph is related to the continuous arc burning, which indicates the continuous behavior of the surface discharge of the iced insulator. In contrast, the blank area corresponds to the cessation of the surface discharge, indicating the transition between different discharge modes.

Images