A 10kV distribution line wildfire fault early warning method based on zero-sequence voltage signal

Abstract

The present invention discloses a 10kV distribution line wildfire fault early warning method based on a zero-sequence voltage signal, comprising: obtaining zero-sequence voltage data from a 10kV power grid zero-sequence voltage transformer and performing empirical mode decomposition to obtain each intrinsic mode function; obtaining Hilbert transform thereof to obtain time-frequency distribution data of each mode, and calculating instantaneous frequency, instantaneous amplitude, and energy spectrum according to the time-frequency distribution data; performing time-frequency analysis based on the instantaneous frequency to obtain time-varying characteristics of fluctuations during a zero-sequence voltage fault, and extracting main frequency components based on the zero-sequence voltage time-varying characteristics; constructing fault characteristic parameters based on the main frequency, instantaneous amplitude, and energy spectrum in the zero-sequence voltage time-varying characteristics, setting a threshold value based on the characteristic parameters, comparing a characteristic signal during abnormal system operation with the threshold value, and issuing an alarm and disconnecting the faulty line if the system fault signal exceeds the discrimination threshold value. The present invention is beneficial to reducing the risk of 10kV line wildfire faults on the normal operation of the system.

Description

Technical Field

[0001] The present invention relates to the technical field of 10kV line wildfire early warning tripping, and in particular to a 10kV distribution line wildfire fault early warning method based on a zero-sequence voltage signal. Background Art

[0002] Due to the imbalance in the distribution of power generation and power load centers in my country, large-capacity, long-distance power transmission is required. Transmission and distribution lines often cross dense forests or agricultural areas, where wildfires are common, severely impacting the safe and stable operation of power networks. When a wildfire occurs beneath a 10kV power line, the high temperature, high conductivity, and ash content of the vegetation flames can significantly reduce the insulation strength of the air gap beneath the transmission line. If the flames bridge the conductors, there is a high risk of conductor-to-ground breakdown. Summary of the Invention

[0003] In order to solve the problems existing in the prior art, the purpose of the present invention is to provide a 10kV distribution line wildfire fault warning method based on zero-sequence voltage signal. The present invention can be used to identify faults when a wildfire occurs under the 10kV line conductor. The grid-side zero-sequence voltage is processed by empirical mode decomposition and Hilbert transform to achieve extraction through characteristic components in the time domain, and ultimately achieve early warning and tripping before the conductor breaks down to the ground.

[0004] To achieve the above object, the present invention adopts a technical solution: a 10kV distribution line wildfire fault early warning method based on zero-sequence voltage signal, comprising the following steps:

[0005] Step 1: Obtain zero-sequence voltage data from a running 10 kV power grid zero-sequence voltage transformer;

[0006] Step 2: Perform empirical mode decomposition on the zero-sequence voltage to obtain each intrinsic mode function;

[0007] Step 3: Calculate the Hilbert transform of each intrinsic mode function to obtain the time-frequency distribution data of each mode, and calculate the instantaneous frequency, instantaneous amplitude, and energy spectrum according to the time-frequency distribution data;

[0008] Step 4: Perform time-frequency analysis on the instantaneous frequency to obtain the time-varying characteristics of the zero-sequence voltage fluctuation during the fault, and extract the main frequency components based on the time-varying characteristics of the zero-sequence voltage.

[0009] Step 5: Construct fault characteristic parameters through the main frequency, instantaneous amplitude, and energy spectrum of the zero-sequence voltage time-varying characteristics, and set thresholds based on the characteristic parameters. By comparing the characteristic signal when the system is operating abnormally with the threshold, if the system fault signal exceeds the judgment threshold, an alarm is issued and the fault line is disconnected.

[0010] As a further improvement of the present invention, the step 2 specifically includes the following steps:

[0011] Step 2.1, extract local features of the original signal and find its local maximum and local minimum;

[0012] Step 2.2: Use spline interpolation to connect all the maximum and minimum points to form the upper envelope e max (t), lower envelope e min (t);

[0013] Step 2.3, calculate the average value of the upper and lower envelopes to obtain the mean curve:

[0014] Step 2.4: Subtract the mean curve from the original signal to obtain the preliminary intrinsic mode function (IMF): h(t) = x(t) - m(t);

[0015] Step 2.5: Check whether h(t) satisfies the intrinsic modal component IMF restriction condition. If not, repeat steps 2.1 to 2.4 until the IMF condition is satisfied. ci(t) ;

[0016] Step 2.6: Subtract the IMF from the original signal ci(t) We get: r(t) = x(t) - c1(t), and repeat steps 2.1 to 2.5 to get the IMF. c2 , until the residual signal r(t) is monotonic or no IFM can be extracted.

[0017] As a further improvement of the present invention, in step 2.5, the intrinsic modal component IMF restriction condition is specifically as follows:

[0018] (1) In the entire zero-sequence voltage data set, the number of extreme points and the number of zero points must be equal or differ by at most one; that is:

[0019] (Nz-1)≤Ne≤(Nz+1)

[0020] (2) At any point, the average value of the envelope formed by local maxima and minima must be zero; that is:

[0021]

[0022] As a further improvement of the present invention, the step 3 specifically includes the following steps:

[0023] Step 3.1: For each IMF ci(t) Perform Hilbert transform to obtain the real and imaginary parts of the analytical signal. The calculation formula is as follows:

[0024]

[0025] Where, The analytical signal consists of IMF and its Hilbert transform, H[x(t)] is the Hilbert transform of IMF, where P and V are principal value integrals used to deal with singular points in the integral;

[0026] Step 3.2: Calculate the instantaneous amplitude A i (t) and instantaneous phase φ i (t), the calculation formula is as follows:

[0027]

[0028] Step 3.3, calculate the instantaneous frequency f i (t), the calculation formula is as follows:

[0029]

[0030] Where, is the derivative of the instantaneous phase with respect to time;

[0031] Step 3.4: Calculate the energy spectrum E i (t), the calculation formula is as follows:

[0032] E i (t) = A i (t) 2 .

[0033] As a further improvement of the present invention, the step 4 is specifically as follows:

[0034] A time-frequency analysis is performed on each effective IMF instantaneous frequency obtained after Hilbert transform. The time-frequency two-dimensional image and Hilbert energy spectrum are drawn by Matlab to obtain the time-varying characteristics of the zero-sequence voltage fluctuation when a fault occurs, and the main low-frequency components are extracted based on the time-varying characteristics of the zero-sequence voltage.

[0035] As a further improvement of the present invention, the step 5 is specifically as follows:

[0036] According to the main frequency components, Hilbert energy spectrum E i (t), through the Hilbert spectrum amplitude distribution characteristics and Hilbert energy spectrum IMF index, the main frequency component and Hilbert energy spectrum are compared with the normal operation of the system without wildfire disaster; due to the imbalance of the system itself, the peak value of the zero-sequence voltage change is set as the discrimination threshold, and the main frequency component and Hilbert energy spectrum E i (t), Hilbert spectrum is compared with the threshold, if the main frequency component IMF, Hilbert energy map E i(t) and the Hilbert spectrum are out of limit, it is determined whether a wildfire disaster has occurred and whether the faulty line needs to be removed in time.

[0037] When a flame bridges the transmission line, the leakage current will increase, but the leakage current parameters cannot be obtained in the actual operating system. Only the three-phase voltage, three-phase current, and zero-sequence component can be obtained from the grid side. Since the zero-sequence current value changes very weakly when the flame causes discharge, the present invention analyzes the change in zero-sequence voltage to determine whether a wildfire has occurred under the conductor or whether a wildfire has occurred near the conductor and spread to the bottom of the conductor, causing a trip. By performing empirical mode decomposition and Hilbert transform on the zero-sequence voltage component, empirical mode decomposition can overcome the problem of non-adaptability of the basis function, without the need for pre-analysis and processing of the signal, and is convenient for application in real-time monitoring of power systems. By performing Hilbert transform on each inherent mode function after empirical mode decomposition, meaningful instantaneous frequencies can be extracted, avoiding the disadvantage of meaningless instantaneous frequencies in traditional methods. By analyzing the fluctuation of the instantaneous frequency of the zero-sequence voltage, the function of distinguishing and warning can be achieved.

[0038] The beneficial effects of the present invention are:

[0039] The present invention takes into account that the transmission line passes through forest areas, cultivated land and other areas where wildfires are frequent. When a wildfire occurs below the line and the flame bridges the conductor, the insulation of the air gap below the conductor will drop significantly, which can easily cause the conductor to directly break down to the ground. If an early warning can be issued based on the electrical characteristic quantity when a wildfire occurs, and the fire department and forestry department can be coordinated in time to isolate the high-risk areas near the line. Since the success rate of line tripping and reclosing caused by wildfire disasters is low, it is necessary to consider whether to shut down the line to avoid risks based on the electrical characteristic quantity of the line under the wildfire disaster. Based on the electrical characteristic components under wildfire faults, the leakage current is the most intuitive electrical quantity reflecting the air insulation characteristics, but it is impossible to detect its leakage current when a wildfire occurs, so the zero-sequence component is then subjected to fault feature extraction. The zero-sequence voltage component includes power frequency, low-frequency, harmonic, and high-frequency components. Empirical mode decomposition (EMD) is the intrinsic mode function (IMF) derived by subtracting the upper and lower envelopes of the signal in the time domain. This results in a step-by-step decomposition of the real fluctuations or trends of varying scales in the signal, from high to low frequency. Empirical mode decomposition no longer decomposes the signal into a series of sinusoidal waves; its frequency definition is based on the instantaneous frequency. The Hilbert-Huang transform decomposes based on changes in the signal envelope. Its time resolution remains constant and its accuracy is high, while its frequency resolution can be adaptively adjusted, making it suitable for processing fault signals that fluctuate over time. Fault signatures are extracted from the instantaneous frequency and amplitude of the IMF components, which are then coupled with the Hilbert energy spectrum and Hilbert spectrogram. A threshold is set for comparison and judgment. Ultimately, the operations manager determines whether to perform a shutdown or risk avoidance measure based on the comparison results. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 is a flow chart of an embodiment of the present invention;

[0041] Figure 2 This is a diagram showing the original zero-sequence voltage waveform and the IMFs superimposed waveform when no fault occurs in an embodiment of the present invention;

[0042] Figure 3 : This is a waveform diagram of each inherent modal component of the zero-sequence voltage during normal operation according to an embodiment of the present invention;

[0043] Figure 4 The Hilbert energy spectrum and Hilbert spectrum diagram for normal operation in an embodiment of the present invention;

[0044] Figure 5 This is a diagram showing the original zero-sequence voltage waveform and the IMFs superimposed waveform under a wildfire fault in an embodiment of the present invention;

[0045] Figure 6 : This is a waveform diagram of each inherent modal component of the zero-sequence voltage under a wildfire fault in an embodiment of the present invention;

[0046] Figure 7 Graphs of the Hilbert energy spectrum and Hilbert spectrum under wildfire fault conditions in an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0048] Example

[0049] Figure 1 As shown, a 10kV distribution line wildfire fault early warning method based on zero-sequence voltage signal includes the following steps:

[0050] Step 1: Obtain zero-sequence voltage data from the operating 10kV power grid zero-sequence voltage transformer. When a wildfire fault occurs, the zero-sequence voltage will fluctuate, such as Figure 5 As shown;

[0051] Step 2: Perform empirical mode decomposition (EMD) on the zero-sequence voltage to obtain each intrinsic mode function IMF, such as Figure 6 As shown, the steps to obtain each intrinsic mode function IMF are as follows:

[0052] a. Extract local features from the original signal and find its local maximum and local minimum;

[0053] b. Use spline interpolation to connect all the maximum and minimum points to form the upper e max (t), lower envelope e min (t);

[0054] c. Calculate the average value of the upper and lower envelopes to obtain the mean curve,

[0055] d. Subtract the mean curve from the original signal to obtain the preliminary intrinsic mode function IMF candidate: h(t) = x(t) - m(t);

[0056] e. Check whether h(t) satisfies the IMF condition. If not, repeat ad until the IMF condition is satisfied. ci(t) ;

[0057] f. Subtract IMF from the original signal ci(t) We get: r(t) = x(t) - c1(t), and repeat ae to get IMF c2 , until the residual signal r(t) is monotonic or no IFM can be extracted.

[0058] The intrinsic mode component IMF satisfies two constraints:

[0059] (1) In the entire zero-sequence voltage data set, the number of extreme points and the number of zero points must be equal or differ by at most one; that is:

[0060] (Nz-1)≤Ne≤(Nz+1)

[0061] (2) At any point, the average value of the envelope formed by local maxima and minima must be zero; that is:

[0062]

[0063] Step 3: For each IMF ci(t) The Hilbert transform (HHT) is calculated to obtain the time-frequency distribution data of each mode. The instantaneous frequency, instantaneous amplitude, and energy spectrum are calculated based on the time-frequency distribution data, as follows:

[0064] a. For each IMF ci(t) Perform Hilbert transform to obtain the real and imaginary parts of the analytical signal. The calculation formula is as follows:

[0065]

[0066] Where, The analytical signal consists of an IMF and its Hilbert transform. H[x(t)] is the Hilbert transform of the IMF, where P and V are principal value integrals used to handle singular points in the integral.

[0067] b. Calculate the instantaneous amplitude A i (t) and instantaneous phase φ i (t), the calculation formula is as follows:

[0068]

[0069] c. Calculate the instantaneous frequency f i (t), the calculation formula is as follows:

[0070]

[0071] Where, is the time derivative of the instantaneous phase.

[0072] d. Calculate the energy spectrum E i (t), the calculation formula is as follows:

[0073] E i (t) = A i (t) 2

[0074] Step 4: Perform time-frequency analysis on each effective IMF instantaneous frequency obtained after Hilbert transform in step 3, and draw a two-dimensional time-frequency image and Hilbert energy spectrum through Matlab, as shown in the following example: Figure 7 As shown in the figure, the time-varying characteristics of the zero-sequence voltage fluctuation when a fault occurs are obtained, and the main low-frequency components are extracted according to the time-varying characteristics of the zero-sequence voltage;

[0075] Step 5: Based on the main frequency components and Hilbert energy spectrum E in step 4 i (t), through the Hilbert spectrum amplitude distribution characteristics and Hilbert energy spectrum IMF index, and when the system is operating normally and no wildfire disaster occurs (such as Figure 2 、 Figure 3 The main frequency component (shown as Figure 4 As the system itself is unbalanced, the peak value of the zero-sequence voltage change is set as the discrimination threshold, and the main frequency component, Hilbert energy diagram E i (t), Hilbert spectrum is compared with the threshold, if the main frequency component IMF, Hilbert energy map E i (t) and the Hilbert spectrum are out of limit, it is judged that a wildfire disaster has occurred and the fault line needs to be removed in time.

[0076] In this embodiment, the zero-sequence voltage signal that is easy to collect on the grid side is first analyzed, and the average value of the upper and lower envelopes is subtracted from the zero-sequence voltage signal to obtain the intrinsic mode function. It is determined whether the obtained intrinsic mode function meets the IMF condition. If it does, it is marked as an IMF component. The extracted IMF component is subtracted from the voltage signal to obtain a residual signal. It is checked whether the residual signal is monotonic or constant. If it does, each intrinsic mode function is arranged in descending order of frequency. Then, each intrinsic mode function is Hilbert transformed to solve the instantaneous frequency, instantaneous phase, instantaneous amplitude, and Hilbert spectrum. Each effective IMF instantaneous frequency obtained after the Hilbert transform is subjected to time-frequency analysis. A two-dimensional time-frequency image and a Hilbert energy spectrum are plotted using MATLAB to obtain the time-varying characteristics of the zero-sequence voltage fluctuation when a fault occurs. The main low-frequency component is extracted based on the time-varying characteristics of the zero-sequence voltage. The main frequency component, the amplitude distribution characteristics of the Hilbert spectrum, and the IMF index of the Hilbert energy spectrum are compared with the main frequency component and Hilbert energy spectrum when the system is operating normally without a wildfire disaster. The peak value of zero-sequence voltage change caused by the imbalance of system factors is set as the discrimination threshold, and the main frequency component, Hilbert energy diagram E i (t), the Hilbert spectrum is compared with the threshold to determine whether a wildfire disaster has occurred and whether the faulty line needs to be removed in time, so as to reduce the risk of line damage caused by wildfire disasters during power grid operation.

[0077] This embodiment has the following advantages: In real-world situations, when wildfires occur beneath power lines, sparsely populated areas lack sufficient early warning of wildfires. However, as one of the fixed electrical signals collected on the grid side, the zero-sequence voltage component offers the advantage of being easily acquired and analyzed. According to the sinusoidal frequency definition, when a power system fault occurs, its frequency characteristics are almost impossible to identify due to waveform distortion. Therefore, the commonly used Fourier decomposition and wavelet transform methods for extracting high-frequency components cannot reflect the fault characteristics, and the wavelet transform is overly dependent on the selection of the wavelet basis. HHT is based on the instantaneous frequency definition and decomposes the high-frequency components according to the signal envelope. Fault-induced changes in the zero-sequence voltage amplitude cause changes in the signal envelope, allowing the signal envelope to contain fault information. This eliminates the need for selecting basis functions, provides constant and high-precision time resolution, and allows adaptive frequency resolution. Because EMD decomposition no longer decomposes the signal into a series of sinusoidal waves, its frequency definition is based on the instantaneous frequency. This provides significant advantages in situations where the fault can cause changes in the signal envelope but subtle changes in the sinusoidal frequency. By comparing the decomposed main frequency components and the Hilbert energy spectrum with the established threshold, a basis for judgment is provided when power companies determine whether wildfire failures require shutdown to avoid risks.

[0078] The above-described embodiments merely represent specific implementations of the present invention. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make numerous variations and improvements without departing from the spirit of the present invention, and all such variations and improvements fall within the scope of protection of the present invention.

Claims

1. A 10kV distribution line wildfire fault early warning method based on zero-sequence voltage signal, characterized in that: The following steps are involved: Step 1: Obtain zero-sequence voltage data from a running 10 kV power grid zero-sequence voltage transformer; Step 2: Perform empirical mode decomposition on the zero-sequence voltage to obtain each intrinsic mode function; Step 3: Calculate the Hilbert transform of each intrinsic mode function to obtain the time-frequency distribution data of each mode, and calculate the instantaneous frequency, instantaneous amplitude, and energy spectrum according to the time-frequency distribution data; Step 4: Perform time-frequency analysis on the instantaneous frequency to obtain the time-varying characteristics of the zero-sequence voltage fluctuation during the fault, and extract the main frequency components based on the time-varying characteristics of the zero-sequence voltage. Step 5: Construct fault characteristic parameters based on the main frequency, instantaneous amplitude, and energy spectrum of the zero-sequence voltage time-varying characteristics, and set thresholds based on the characteristic parameters. By comparing the characteristic signal when the system is operating abnormally with the threshold, if the system fault signal exceeds the judgment threshold, an alarm is issued and the faulty line is disconnected; The step 2 specifically includes the following steps: Step 2.1, extract local features of the original signal and find its local maximum and local minimum; Step 2.2: Use spline interpolation to connect all the maximum and minimum points to form the upper envelope e max (t), lower envelope e min (t); Step 2.3, calculate the average value of the upper and lower envelopes to obtain the mean curve: Step 2.4: Subtract the mean curve from the original signal to obtain the preliminary intrinsic mode function (IMF): h(t) = x(t) - m(t); Step 2.5: Check whether h(t) satisfies the intrinsic modal component IMF restriction condition. If not, repeat steps 2.1 to 2.4 until the IMF condition is satisfied. ci(t) ; Step 2.6: Subtract the IMF from the original signal ci(t) We get: r(t) = x(t) - c1(t), and repeat steps 2.1 to 2.5 to get the IMF. c2 , until the residual signal r(t) is monotonic or no IFM can be extracted; In step 2.5, the intrinsic modal component IMF restriction conditions are as follows: (1) In the entire zero-sequence voltage data set, the number of extreme points and the number of zero points must be equal or differ by at most one; that is: (Nz-1)≤Ne≤(Nz+1) (2) At any point, the average value of the envelope formed by local maxima and minima must be zero; that is: The step 3 specifically includes the following steps: Step 3.1: For each IMF ci(t) Perform Hilbert transform to obtain the real and imaginary parts of the analytical signal. The calculation formula is as follows: Where, The analytical signal consists of the IMF and its Hilbert transform, is the Hilbert transform of IMF, where P and V are principal value integrals used to handle singular points in the integral; Step 3.2: Calculate the instantaneous amplitude A i (t) and instantaneous phase φ i (t), the calculation formula is as follows: Step 3.3, calculate the instantaneous frequency f i (t), the calculation formula is as follows: Where, is the derivative of the instantaneous phase with respect to time; Step 3.4: Calculate the energy spectrum E i (t), the calculation formula is as follows: E i (t)=A i (t) 2 The step 4 is specifically as follows: Perform time-frequency analysis on each effective IMF instantaneous frequency obtained after Hilbert transform, draw a two-dimensional time-frequency image and Hilbert energy spectrum using Matlab, and obtain the time-varying characteristics of zero-sequence voltage fluctuations when a fault occurs. Then, extract the main low-frequency components based on the time-varying characteristics of zero-sequence voltage. The step 5 is specifically as follows: According to the main frequency components, Hilbert energy spectrum E i (t), through the Hilbert spectrum amplitude distribution characteristics and Hilbert energy spectrum IMF index, the main frequency component and Hilbert energy spectrum are compared with the normal operation of the system without wildfire disaster; due to the imbalance of the system itself, the peak value of the zero-sequence voltage change is set as the discrimination threshold, and the main frequency component and Hilbert energy spectrum E i (t), Hilbert spectrum is compared with the threshold, if the main frequency component IMF, Hilbert energy map E i (t) and the Hilbert spectrum are out of limit, it is determined whether a wildfire disaster has occurred and whether the faulty line needs to be removed in time.

Images