Fault positioning method based on neural network and related equipment

Abstract

The invention discloses a fault positioning method based on a neural network and related equipment, relates to the technical field of power transmission line fault positioning, and solves the problem of relatively large fault positioning error. According to the method, the correction coefficient of the traveling wave speed is calculated based on the single-ended fault traveling wave signal, so that errors generated by a traditional fixed wave speed method under the conditions of complex terrains or weather and the like can be eliminated; and then based on an LSTM network, capturing a non-linear relationship among a correction coefficient, a fault point distance and a double-end traveling wave signal time difference, and obtaining a more accurate fault positioning result.

Description

Technical Field

[0001] The present invention relates to the technical field of transmission line fault location, and particularly to a fault location method and related equipment based on a neural network. Background Art

[0002] Accurate fault ranging of transmission lines in a power system can quickly narrow down the fault location range, reduce the patrol burden, and shorten the fault elimination time, which is of great significance for improving the power supply reliability of the power system and reducing power outage losses.

[0003] As one of the most effective methods for fault location, the traveling wave location method is widely used. According to different traveling wave location principles, the traveling wave location method is divided into the single - end traveling wave method and the double - end traveling wave method. Among them, the single - end traveling wave method uses the time when the initial fault traveling wave arrives at one end of the line and the time when the reflected traveling wave from the fault point arrives at this end, and combines the transmission speed of the fault traveling wave to calculate the fault point location; the double - end traveling wave method uses the time when the initial fault traveling wave arrives at both ends of the line and combines the transmission speed of the fault traveling wave to calculate the fault point location.

[0004] The traveling wave velocity has a certain influence on the fault location accuracy. The actual wave velocity is uncertain due to many factors such as line parameters, terrain, geographical location, and climate. The traveling wave velocity obtained based on the assumption of a uniform line and symmetric line structure in traditional methods has a certain error compared with the actual wave velocity. Since the wave velocity is close to the speed of light, it will lead to a large location error.

[0005] In view of this, a fault location method and related equipment based on a neural network are needed. Summary of the Invention

[0006] Aiming at the problem of large location error in the existing fault location methods, the present invention provides a fault location method and related equipment based on a neural network, which can reduce the fault location error. The specific technical solutions are as follows:

[0007] In a first aspect, an embodiment of the present application provides a fault location method based on a neural network, including:

[0008] Obtain the first fault traveling wave signal received at the first end of the transmission line and the second fault traveling wave signal received at the second end of the transmission line. The first fault traveling wave signal and the second fault traveling wave signal are traveling wave signals corresponding to the same fault. Based on the first fault traveling wave signal and the second fault traveling wave signal, obtain the first time when the wavefront of the first fault traveling wave signal arrives at the first end and the second time when the wavefront of the second fault traveling wave signal arrives at the second end. Based on the time data of three consecutive reflected waves in the first fault traveling wave signal or the second fault traveling wave signal, calculate the correction coefficient of the traveling wave velocity. Input the time difference between the first time and the second time and the correction coefficient into a preset long-short term memory (LSTM) network to obtain the fault point distance output by the LSTM network. Map the fault point distance to geographical coordinates based on the preset transmission line topology to generate a fault location result.

[0009] Preferably, calculating the correction coefficient of the traveling wave velocity based on the time data of three consecutive reflected waves in the first fault traveling wave signal or the second fault traveling wave signal includes: obtaining the first reflection wavefront time and the third reflection wavefront time of the three consecutive reflected waves; calculating the estimated distance of the fault point based on the time difference between the first time and the second time; calculating the correction coefficient based on the estimated distance of the fault point, the first reflection wavefront time, and the third reflection wavefront time.

[0010] Preferably, the expression for calculating the correction coefficient includes:

[0011] α = 4D init / [v o *(t 3 -t 1 )];

[0012] Where α represents the correction coefficient, t 1 represents the first reflection wavefront time, t 3 represents the third reflection wavefront time, D init is the estimated distance of the fault point, and v o is the theoretical wave velocity.

[0013] Preferably, calculating the correction coefficient based on the estimated distance of the fault point, the first reflection wavefront time, and the third reflection wavefront time includes: establishing an objective function, and the expression of the objective function includes:

[0014] minf(α) = |4D' / (t 3 -t 1 ) - v 0 α'|;

[0015] Among them, α' is the intermediate correction coefficient, D' is the distance to the intermediate fault point, and L is the length of the transmission line;

[0016] Taking this estimated distance to the fault point as the distance to the intermediate fault point, based on the head time of the first reflected wave and the head time of the third reflected wave, solve this objective function to obtain a feasible solution; when the intermediate correction coefficient of this feasible solution is within the corresponding value range, obtain this intermediate correction coefficient as this correction parameter; when the intermediate correction coefficient of this feasible solution is not within the corresponding value range, vary the distance to the intermediate fault point and solve this objective function again.

[0017] Preferably, the value range of α' is [0.98, 1.02], and the value range of D' is [0, L].

[0018] Preferably, after obtaining the distance to the fault point output by this LSTM network, this method further includes: obtaining information about impedance mutation points in this transmission line topology; based on this information about impedance mutation points and this distance to the fault point, determining whether the distance between this impedance mutation point and the fault point is less than a preset threshold; when the distance between this impedance mutation point and the fault point is less than this preset threshold, adjusting this correction coefficient based on this information about impedance mutation points; the expression for adjusting this correction coefficient is:

[0019] α' = α·(1 + βΓ);

[0020] Among them, α' is the adjusted correction coefficient, β is the compensation weight coefficient, and Γ is the reflection coefficient of the impedance mutation point; based on this adjusted correction coefficient, the time difference between this first time and this second time, and this LSTM network, calculate to obtain the adjusted distance to the fault point; mapping this distance to the fault point to geographical coordinates based on a preset transmission line topology includes: mapping this adjusted distance to the fault point to geographical coordinates based on this transmission line topology.

[0021] Preferably, mapping this distance to the fault point to geographical coordinates based on a preset transmission line topology includes: obtaining the longitude and latitude of this first end and the longitude and latitude of this second end from the line tower coordinate database; based on the longitude and latitude of this first end and the longitude and latitude of this second end, perform linear interpolation along the transmission line topology with this distance to the fault point as the path length to obtain the longitude and latitude of the fault point as geographical coordinates.

[0022] Preferably, before calculating the correction coefficient of the traveling wave velocity, this method further includes: performing Hilbert transform on this first fault traveling wave signal or this second fault traveling wave signal to generate an analytic signal; calculating the instantaneous frequency of this analytic signal to determine the frequency mutation point; based on this frequency mutation point and a preset energy threshold, obtaining the time data of these three consecutive reflected waves.

[0023] Preferably, obtaining the first time when the wavefront of the first fault traveling wave signal arrives at the first end and the second time when the wavefront of the second fault traveling wave signal arrives at the second end based on the first fault traveling wave signal and the second fault traveling wave signal includes: performing continuous wavelet transform on the first fault traveling wave signal and the second fault traveling wave signal through a Morlet wavelet basis function to obtain a first wavelet coefficient and a second wavelet coefficient; determining the first time based on the modulus maximum value of the first wavelet coefficient, and determining the second time based on the modulus maximum value of the second wavelet coefficient.

[0024] In a second aspect, an embodiment of the present application provides a fault location system based on a neural network, which is applied to the method described in the first aspect. The system includes:

[0025] An acquisition module, configured to acquire a first fault traveling wave signal received at a first end of a transmission line and a second fault traveling wave signal received at a second end of the transmission line, where the first fault traveling wave signal and the second fault traveling wave signal are traveling wave signals corresponding to the same fault;

[0026] A wavefront detection module, configured to obtain the first time when the wavefront of the first fault traveling wave signal arrives at the first end and the second time when the wavefront of the second fault traveling wave signal arrives at the second end based on the first fault traveling wave signal and the second fault traveling wave signal;

[0027] A calculation module, configured to calculate a correction coefficient of the traveling wave velocity based on time data of three consecutive reflected waves in the first fault traveling wave signal or the second fault traveling wave signal;

[0028] A neural network module, configured to input the time difference between the first time and the second time and the correction coefficient into a preset LSTM network to obtain a fault point distance output by the LSTM network;

[0029] A location module, configured to map the fault point distance to geographical coordinates based on a preset transmission line topology to generate a fault location result.

[0030] In a third aspect, an embodiment of the present application provides a computing device, including: a memory, configured to store a program; a processor, configured to load the program to execute the method described in the first aspect.

[0031] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, where the computer-readable storage medium includes a stored program, and when the program runs, it controls the device where the computer-readable storage medium is located to execute the method described in the first aspect.

[0032] Compared with the prior art, the beneficial effects of the present invention are as follows: By calculating the correction coefficient of the traveling wave velocity based on the single-ended fault traveling wave signal, the error generated by the traditional fixed wave velocity method under conditions such as complex terrain or weather can be eliminated; then, based on the LSTM network to capture the non-linear relationship between the correction coefficient, the fault point distance, and the time difference of the double-ended traveling wave signal, a more accurate fault location result can be obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the specific embodiments or the prior art. In all the drawings, similar elements or parts are generally denoted by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0034] Figure 1 It is a schematic flowchart of a fault location method based on a neural network provided by an embodiment of the present application;

[0035] Figure 2 It is a schematic structural diagram of a fault location system based on a neural network provided by an embodiment of the present application;

[0036] Figure 3 It is a schematic structural diagram of a computing device provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0037] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0038] It should be understood that when used in this specification and the appended claims, the terms "comprises" and "comprising" indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their combinations.

[0039] It should also be understood that the terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an", and "the" are intended to include the plural forms.

[0040] It should be further understood that the term "and / or" used in the specification and appended claims of the present invention refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0041] To solve the problem of large errors in fault location caused by calculating based on a fixed wave velocity in traditional fault location methods, the present invention provides a fault location method based on a neural network and related devices, which can reduce the fault location error.

[0042] Please refer to Figure 1 , Figure 1 This application provides an embodiment of a fault location method based on a neural network, and this method is applied to a computing device; as Figure 1 shown, this method includes:

[0043] Step 101, the computing device acquires a first fault traveling wave signal received at the first end of the transmission line, and a second fault traveling wave signal received at the second end of the transmission line.

[0044] Among them, the computing device can be a computing module, a control module or a monitoring and acquisition module arranged on the transmission line, which can directly acquire the traveling wave signal on the transmission line; it can also be a server, or an intelligent terminal such as a personal computer or a tablet directly operated by a power system manager or a maintenance person, which is communicatively connected to the first end and the second end by wire or wirelessly, and acquires the traveling wave signals received at the first end and the second end from them.

[0045] Among them, the first fault traveling wave signal and the second fault traveling wave signal are traveling wave signals corresponding to the same fault. That is to say, after a fault occurs on the transmission line, the traveling wave signals generated by the fault point where the fault is located are respectively transmitted to the first end and the second end of the transmission line, so that the first end receives the first fault traveling wave signal and the second end receives the second fault traveling wave signal.

[0046] Step 102, the computing device obtains a first time when the wavefront of the first fault traveling wave signal reaches the first end, and a second time when the wavefront of the second fault traveling wave signal reaches the second end based on the first fault traveling wave signal and the second fault traveling wave signal.

[0047] Among them, the computing device can extract the arrival time of the wavefront of the fault traveling wave signal from the fault traveling wave signal by means of a threshold detection method, a slope detection method, a wavelet transform method, etc.

[0048] Preferably, the computing device can perform a continuous wavelet transform on the first fault traveling wave signal and the second fault traveling wave signal to obtain a first wavelet coefficient and a second wavelet coefficient; determine the first time based on the modulus maximum value of the first wavelet coefficient, and determine the second time based on the modulus maximum value of the second wavelet coefficient.

[0049] By suppressing noise interference through joint time-frequency analysis, the accuracy of the wavefront detection time can be improved.

[0050] Among them, the computing device can adopt wavelet basis functions such as Morlet wavelet, Daubechies wavelet, Haar wavelet, Symlets wavelet, etc.

[0051] Step 103: The computing device calculates a correction coefficient of the traveling wave velocity based on the time data of three consecutive reflected waves in the first fault traveling wave signal or the second fault traveling wave signal.

[0052] Among them, the computing device can calculate the actual traveling wave velocity of the fault traveling wave from the fault point to the corresponding single end through the single-ended fault traveling wave signal data, and then calculate the correction coefficient of the traveling wave velocity based on the actual traveling wave velocity and the theoretical traveling wave velocity.

[0053] It can be understood that the traveling wave velocity calculated through the single-ended three-time reflected wave essentially only reflects the wave velocity of the section from the fault point to the measurement end. However, the actual line may include sections such as cable-overhead line hybrid sections and transposition towers, and the wave velocity shows sectional characteristics. The single-ended velocity cannot directly represent the global wave velocity; through the correction coefficient, the local wave velocity can be mapped to the global equivalent wave velocity, which can be compatible with the fluctuations of line parameters and improve the robustness of the algorithm. In addition, as a dimensionless parameter, the correction coefficient is more suitable for training in the long-short term memory (LSTM) network in the subsequent steps.

[0054] The detection of the single-ended three-time reflected wave may be affected by noise interference (such as measurement errors caused by non-uniformity caused by line parameters). Through the iterative correction of the correction coefficient and the LSTM data fusion, the influence of local outliers can be smoothed, and the noise influence of the single-ended signal calculation can be suppressed.

[0055] It can be understood that the computing device can calculate the correction coefficient based on the first fault traveling wave signal or the second fault traveling wave signal. Correspondingly, the calculated distance of the fault point obtained subsequently is the distance from the first end or the second end.

[0056] Preferably, the computing device can perform a Hilbert transform on the first fault traveling wave signal or the second fault traveling wave signal to generate an analytic signal; calculate the instantaneous frequency of the analytic signal to determine the frequency mutation point; and obtain the time data of the three consecutive reflected waves based on the frequency mutation point and a preset energy threshold.

[0057] Among them, the computing device can perform Hilbert transform on the single-ended traveling wave signal S(t) to generate an analytic signal Z(t) = S(t) + jH[S(t)]; where t is time, j is the imaginary unit, and H[S(t)] represents the Hilbert transform of the single-ended traveling wave signal S(t).

[0058] Then, the computing device can calculate the instantaneous frequency of the analytic signal and detect the frequency mutation points therein. The calculation formula for the instantaneous frequency is:

[0059]

[0060] where f(t) represents the instantaneous frequency of the signal, d / dt is the operator for taking the derivative with respect to time t, and arg[Z(t)] is the phase function of the analytic signal Z(t).

[0061] Then, the computing device can calculate the reflected wave energy of each frequency mutation point. When the reflected wave energy is greater than the energy threshold, the computing device determines that the frequency mutation point is a reflected wave. Then, the computing device can obtain the time data of the consecutive 3 reflected waves with the earliest time. The wave head times of the 3 reflected waves are sequentially denoted as t 1 、t 2 、t 3 .

[0062] Among them, the computing device can calculate the correction coefficient based on 2 or 3 of the three wave head times.

[0063] It can be understood that it is better to calculate the correction coefficient using the time difference between the first reflected wave and the third reflected wave, because a longer interval improves the measurement accuracy and reduces the noise interference.

[0064] Preferably, the computing device can obtain the wave head time t 1 of the first reflected wave and the wave head time t 3 of the third reflected wave of the consecutive three reflected waves; then calculate the estimated distance to the fault point based on the time difference between the first time T1 and the second time T2; and then calculate the correction coefficient based on the estimated distance to the fault point, and the wave head time of the first reflected wave and the wave head time of the third reflected wave.

[0065] It can be understood that when the computing device calculates the correction coefficient based on the first fault traveling wave signal, the wave head time t 1 of the first reflected wave is equivalent to the first time T1.

[0066] Among them, the computing device can calculate a correction coefficient with a lower precision, input it into the LSTM network in step 104, and correct the non-linear relationship between the correction coefficient and the time difference of the double-ended fault traveling wave model through the LSTM network; or it can calculate a correction coefficient with a higher precision in advance to reduce the computational workload of the LSTM network.

[0067] Preferably, the expression for calculating the correction coefficient includes:

[0068] α = 4D init / [v o *(t 3 -t 1 )];

[0069] Among them, α represents the correction coefficient, t 1 represents the head time of the first reflected wave, t 3 represents the head time of the third reflected wave, D init is the estimated distance of the fault point, and v o is the theoretical wave velocity. The correction coefficient with a lower precision calculated in this way hands over the main computational work to the LSTM network. In the case of a high training completion degree of the LSTM network, computational resources can be saved.

[0070] Preferably, the computing device can establish an objective function, and the expression of the objective function includes:

[0071] minf(α) = |4D' / (t 3 -t 1 ) - v 0 α'|;

[0072] Among them, α' is the intermediate correction coefficient, D' is the intermediate fault point distance, and L is the length of the transmission line; it can be understood that "intermediate" here means the intermediate process, that is, the variable in the iterative process. Preferably, the preset value range of α' is [0.98, 1.02], and the value range of D' is [0, L].

[0073] Then, the computing device can use the estimated distance of the fault point as the intermediate fault point distance, and based on the head time of the first reflected wave and the head time of the third reflected wave, solve the objective function to obtain a feasible solution; in the case where the intermediate correction coefficient of the feasible solution is within the preset value range, obtain the intermediate correction coefficient as the correction parameter; in the case where the intermediate correction coefficient of the feasible solution is not within the value range, change the intermediate fault point distance and solve the objective function again.

[0074] It can be understood that the computing device changes the intermediate correction coefficient within the corresponding value range of the intermediate correction coefficient.

[0075] Step 104: The computing device inputs the time difference between the first time and the second time, and the correction coefficient into a preset LSTM network to obtain the fault point distance output by the LSTM network.

[0076] Among them, the LSTM network is trained with historical data to learn the non-linear compensation relationship between the synchronization error of the double-ended traveling wave head time and the wave velocity fluctuation. Specifically, the historical data includes the historical fault data of the transmission line, as well as the corresponding correction coefficient and the double-ended time synchronization error.

[0077] Among them, the LSTM network includes an input layer, a hidden layer and an output layer; the input layer contains 2 neurons, which are respectively used to process the input time difference and correction coefficient; the hidden layer includes 128 LSTM cells; the output layer contains 1 neuron, which is used to output the fault point distance.

[0078] Specifically, the LSTM network can normalize the input time difference and correction coefficient, then perform the forward propagation calculation of the LSTM, and finally denormalize the calculation result to obtain the fault point distance.

[0079] Step 105: The computing device maps the fault point distance to geographical coordinates based on the preset transmission line topology to generate a fault location result.

[0080] Among them, after obtaining the fault point distance, the computing device needs to further map the fault point distance to actual geographical coordinates so that maintenance personnel can go to the fault point for maintenance and repair.

[0081] Before mapping the actual geographical coordinates, the computing device can further correct the fault point location to reduce the time for maintenance personnel to perform precise positioning on site. Preferably, after obtaining the fault point distance output by the LSTM network, the computing device can obtain the information of the impedance mutation point in the transmission line topology; based on the information of the impedance mutation point and the fault point distance, determine whether the distance between the impedance mutation point and the fault point is less than a preset threshold; when the distance between the impedance mutation point and the fault point is less than the preset threshold, adjust the correction coefficient based on the information of the impedance mutation point; the expression for adjusting the correction coefficient is:

[0082] α' = α·(1 + βΓ)

[0083] Among them, α' is the adjusted correction coefficient, β is the compensation weight coefficient, Γ is the reflection coefficient of the impedance mutation point; based on the adjusted correction coefficient, the time difference between the first time and the second time, and the LSTM network, calculate the adjusted fault point distance; then, the computing device can map the adjusted fault point distance to geographical coordinates based on the transmission line topology.

[0084] Among them, the impedance mutation points on the transmission line include the branch points of the transmission line, the connection points between the transmission line and the cable, the connection points between the lightning protection wire and the cross arm of the transmission tower, etc.; the impedance mutation points will cause the mutation of the traveling wave reflection coefficient, resulting in a reflection situation similar to that of a fault point. Therefore, when the fault point may be relatively close to the impedance mutation point, it is necessary to adjust the correction coefficient and the distance of the fault point to obtain a more accurate fault point positioning.

[0085] Among them, the computing device can obtain impedance mutation points such as the branch points of the transmission line based on the transmission line topology, and obtain the corresponding reflection coefficients from the relevant database; the computing device can also obtain the image of the area corresponding to the distance of the fault point, identify whether there are new impedance mutation points and their types (such as bird nests) in the area, and then obtain the reflection coefficients of this type of impedance mutation points.

[0086] After obtaining the reflection coefficient of the impedance mutation point, the computing device can adjust the correction coefficient, and then input the adjusted correction coefficient and the time difference into the LSTM network again to obtain the adjusted distance of the fault point; finally, based on the adjusted distance of the fault point, the actual geographical coordinates are mapped.

[0087] Preferably, the computing device can obtain the longitude and latitude of the first end and the longitude and latitude of the second end from the line tower coordinate database; then, based on the longitude and latitude of the first end and the longitude and latitude of the second end, linear interpolation is performed along the transmission line topology with the distance of the fault point as the path length, combined with the geographic information system (GIS), to obtain the longitude and latitude of the fault point as the geographical coordinates; then the computing device can generate a fault location result based on the geographical coordinates.

[0088] Among them, the line tower coordinate database can be an online database or an internal database of the computing device or the system where the computing device is located; it stores the longitude and latitude coordinates of the first end and the second end of the transmission line.

[0089] In the embodiment of the present application, by calculating the correction coefficient of the traveling wave velocity based on the single-ended fault traveling wave signal, the error generated by the traditional fixed wave velocity method under conditions such as complex terrain or weather can be eliminated; then, based on the LSTM network to capture the non-linear relationship between the correction coefficient, the distance of the fault point and the time difference of the double-ended traveling wave signal, a more accurate fault location result can be obtained.

[0090] A specific embodiment can be:

[0091] The scenario parameters are set as follows: the total length L of the transmission line is 200 KM, the theoretical wave velocity v 0 is 298 m / μs, the double-ended synchronization error Δt sync is 0.8 μs, and the actual wave velocity v = v0 ·α, the correction coefficient α ∈ [0.98, 1.02]. The distance D between the true fault point and the first end real = 85.6 KM.

[0092] The computing device can obtain the first fault traveling wave signal S A (t) of the first end, and the second fault traveling wave signal S B (t) of the second end; the signal-to-noise ratio of S A (t) is 25 dB, and the signal-to-noise ratio of S B (t) is 28 dB.

[0093] The computing device can perform continuous wavelet transform on S A (t) and S B (t), and the scale range corresponds to the 0.1 - 10 MHz frequency band; then the extreme points of the time-frequency ridge line of the wavelet transform are extracted to obtain the wave head time t A = 286.3 μs, t B = 291.7 μs; then the time synchronization error Δt is calculated to be 5.4 μs.

[0094] Then, the computing device can obtain the continuous 3 reflection wave times t 1 = 286.3 μs, t 2 , t 3 = 287.1 μs; then the time difference of the three reflection waves is calculated to be 0.8 μs. Since the distance D between the fault point and the first end is unknown, the estimated distance between the initial fault point and the first end is estimated using the theoretical speed and the double-end time synchronization error

[0095] D init = (L - v 0 Δt) / 2.

[0096] In practical applications, since the total length L of the transmission line is much larger than the fault point distance D, the influence of the total length L of the line is usually ignored, and the simplified calculation formula is: D init ≈ v 0 Δt / 2 = 804.6 KM. It can be seen that this distance detail exceeds the reasonable range. Therefore, the distance of the intermediate fault point is changed, and the iterative calculation in step 103 of the embodiment shown in Figure 1 is performed to obtain the correction coefficient α = 0.99.

[0097] Input Δt = 5.4 μs and α = 0.99 into the trained LSTM network. The LSTM network first normalizes the input data to obtain Δt norm = 0.8 and α norm = -0.5; then update the hidden state based on the normalized data:

[0098] ht = LSTM(h t-1 , [0.8, -0.5]);

[0099] Where h t represents the hidden state of the LSTM network at time t.

[0100] Then, the LSTM network performs the output layer calculation: D pred = W o ·h t + b o = 5.3; where is the fault point distance predicted by LSTM, W o and b o represent the weight matrix and bias vector of the output layer respectively.

[0101] Then, the LSTM network performs denormalization based on the normalized output result, and the calculation formula is:

[0102] D = D pred * σ D + μ D ;

[0103] Where D is the fault point distance finally output by the LSTM network, σ D is the standard deviation of the fault point distance in the training samples, μ D is the mean of the fault point distance in the training samples. Assuming that the standard deviation is 1.0 and the mean is 80, the fault point distance finally output by LSTM is 85.3 KM.

[0104] After obtaining the fault point distance, the calculation device can determine the first end coordinate as (116.4°E, 39.9°N) and the second end coordinate as (117.2°E, 40.5°N) according to the line tower coordinate database; then the fault point coordinate can be calculated as (116.7412°E, 40.1559°N).

[0105] The above elaborates on the method part provided by the embodiments of the present application. Next, the system part provided by the embodiments of the present application will be described.

[0106] Please refer to Figure 2 , Figure 2 which is a schematic structural diagram of a fault location system based on a neural network provided by the embodiments of the present application. As Figure 2 shown, the system 20 includes:

[0107] An acquisition module 201, configured to acquire a first fault traveling wave signal received at the first end of the transmission line and a second fault traveling wave signal received at the second end of the transmission line. The first fault traveling wave signal and the second fault traveling wave signal are traveling wave signals corresponding to the same fault;

[0108] A wavefront detection module 202, configured to obtain a first time when the wavefront of the first fault traveling wave signal reaches the first end and a second time when the wavefront of the second fault traveling wave signal reaches the second end based on the first fault traveling wave signal and the second fault traveling wave signal;

[0109] A calculation module 203, configured to calculate a correction coefficient of the traveling wave velocity based on time data of three consecutive reflected waves in the first fault traveling wave signal or the second fault traveling wave signal;

[0110] A neural network module 204, configured to input a time difference between the first time and the second time and the correction coefficient into a preset LSTM network to obtain a fault point distance output by the LSTM network;

[0111] A positioning module 205, configured to map the fault point distance to geographical coordinates based on a preset transmission line topology to generate a fault location result.

[0112] Preferably, the calculation module 203 is specifically configured to obtain a first reflected wavefront time and a third reflected wavefront time of the three consecutive reflected waves; calculate a predicted fault point distance based on a time difference between the first time and the second time; and calculate the correction coefficient based on the predicted fault point distance, the first reflected wavefront time, and the third reflected wavefront time.

[0113] Preferably, an expression for calculating the correction coefficient includes:

[0114] α = 4D init / [v o *(t 3 -t 1 )];

[0115] wherein, α represents the correction coefficient, t 1 represents the first reflected wavefront time, t 3 represents the third reflected wavefront time, D init is the predicted fault point distance, and v o is the theoretical wave velocity.

[0116] Preferably, the calculation module 203 is specifically configured to establish an objective function, and an expression of the objective function includes:

[0117] minf(α) = |4D' / (t 3 -t 1 ) - v 0 α'|;

[0118] Among them, α' is the intermediate correction coefficient, D' is the distance to the intermediate fault point, the value range of α' is [0.98, 1.02], the value range of D' is [0, L], and L is the length of the transmission line;

[0119] The calculation module 203 uses this estimated fault point distance as the distance to the intermediate fault point, and based on this first reflected wavefront time and this third reflected wavefront time, solves this objective function to obtain a feasible solution; when the intermediate correction coefficient of this feasible solution is within the corresponding value range, obtain this intermediate correction coefficient as this correction parameter; when the intermediate correction coefficient of this feasible solution is not within the corresponding value range, change this distance to the intermediate fault point and solve this objective function again.

[0120] Preferably, the acquisition module 201 is further configured to acquire information about impedance mutation points in the transmission line topology; the system 20 further includes a determination module 206, configured to determine whether the distance between the impedance mutation point and the fault point is less than a preset threshold based on the information about the impedance mutation point and the fault point distance; an adjustment module 207, configured to, when the distance between the impedance mutation point and the fault point is less than the preset threshold, adjust this correction coefficient based on the information about the impedance mutation point; the expression for adjusting this correction coefficient is:

[0121] α' = α·(1 + βΓ)

[0122] Among them, α' is the adjusted correction coefficient, β is the compensation weight coefficient, and Γ is the reflection coefficient of the impedance mutation point; the calculation module 203 is specifically configured to calculate the adjusted distance to the fault point based on this adjusted correction coefficient, the time difference between this first time and this second time, and this LSTM network; the positioning module 205 is specifically configured to map this adjusted distance to the fault point to geographical coordinates based on the transmission line topology.

[0123] Preferably, the positioning module 205 is specifically configured to obtain the longitude and latitude of this first end and this second end from the line tower coordinate database; perform linear interpolation along the transmission line topology with this distance to the fault point as the path length, and output the longitude and latitude of the fault point.

[0124] Preferably, the calculation module 203 is further configured to perform Hilbert transform on this first fault traveling wave signal or this second fault traveling wave signal to generate an analytic signal; calculate the instantaneous frequency of this analytic signal to determine the frequency mutation point; based on this frequency mutation point and a preset energy threshold, obtain the time data of these three consecutive reflected waves.

[0125] Preferably, the wavefront detection module 202 is specifically configured to perform continuous wavelet transform on the first fault traveling wave signal and the second fault traveling wave signal through a Morlet wavelet basis function to obtain a first wavelet coefficient and a second wavelet coefficient; determine the first time based on the modulus maximum value of the first wavelet coefficient, and determine the second time based on the modulus maximum value of the second wavelet coefficient.

[0126] The fault location system based on a neural network provided by the embodiments of the present application can be understood by referring to the corresponding content in the foregoing method embodiment section, and will not be repeated here.

[0127] As Figure 3 shown, Figure 3 FIG. 10 is a schematic diagram of a possible logical structure of a computing device provided by an embodiment of the present application. The computing device 300 includes: a processor 301, a communication interface 302, a memory 303, and a bus 304. The processor 301, the communication interface 302, and the memory 303 are interconnected through the bus 304. In the embodiment of the present application, the processor 301 is used to control and manage the actions of the computing device 300. For example, the processor 301 is used to execute Figure 1 the steps in the embodiment and / or other processes for the technologies described herein. The communication interface 302 is used to support the computing device 300 to communicate. The memory 303 is used to store the program code and data of the computing device 300.

[0128] Among them, the processor 301 may be a central processing unit, a general-purpose processor, a digital signal processor, an application-specific integrated circuit, a field-programmable gate array, or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute various exemplary logical blocks, modules, and circuits described in combination with the disclosure of the present application. The processor may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a digital signal processor and a microprocessor, and so on. The bus 304 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For the sake of simplicity of representation, Figure 3 only a thick line is shown in FIG. 10, but it does not mean that there is only one bus or one type of bus.

[0129] In another embodiment of the present application, a computer-readable storage medium is further provided. The computer-readable storage medium includes instructions. When the instructions are run on a computer, the computer is caused to execute the above Figure 1 method described in the embodiment.

[0130] Those of ordinary skill in the art can realize that the units of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the components of the examples have been generally described according to their functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0131] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0132] In several embodiments provided by the embodiments of the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or units can be in electrical, mechanical, or other forms.

[0133] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0134] In addition, the functional units in various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0135] When the above-mentioned functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, read-only memories (ROM), random access memories (RAM), mobile hard disks, magnetic disks, or optical discs.

[0136] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the various embodiments of the present invention, and they should all be covered within the scope of the claims and the description of the present invention.

Claims

1. A fault location method based on neural network, characterized in that: include: Acquire a first fault traveling wave signal received at a first end of a transmission line, and a second fault traveling wave signal received at a second end of the transmission line, wherein the first fault traveling wave signal and the second fault traveling wave signal are traveling wave signals corresponding to the same fault; Based on the first fault traveling wave signal and the second fault traveling wave signal, acquiring a first time when the wave head of the first fault traveling wave signal reaches the first end, and a second time when the wave head of the second fault traveling wave signal reaches the second end; Calculating a correction coefficient of traveling wave velocity based on time data of three consecutive reflected waves in the first fault traveling wave signal or the second fault traveling wave signal; Inputting the time difference between the first time and the second time, and the correction coefficient into a preset long short-term memory (LSTM) network to obtain the fault point distance output by the LSTM network; The fault point distance is mapped into geographic coordinates based on a preset transmission line topology to generate a fault location result.

2. The method according to claim 1, characterized in that The calculating the correction coefficient of the traveling wave velocity based on the time data of three consecutive reflection waves in the first fault traveling wave signal or the second fault traveling wave signal comprises: Obtaining the first reflection wave head time and the third reflection wave head time of the three consecutive reflection waves; Calculate the estimated distance of the fault point based on the time difference between the first time and the second time; The correction coefficient is calculated based on the estimated distance of the fault point, the first reflection wave head time and the third reflection wave head time.

3. The method according to claim 2, characterized in that The expression for calculating the correction coefficient includes: α=4D init / [v o *(t3-t1)]; Among them, α is the correction coefficient, t1 is the time of the first reflected wave head, t3 is the time of the third reflected wave head, D init is the estimated distance to the fault point, v o is the theoretical wave speed.

4. The method according to claim 2, characterized in that: The calculation of the correction coefficient based on the estimated distance of the fault point, the first reflection wave head time and the third reflection wave head time includes: Establish an objective function, the expression of which includes: minf(α)=|4D' / (t3-t1)-v0α'|; Among them, α' is the intermediate correction coefficient, D' is the distance of the intermediate fault point, and L is the length of the transmission line; Taking the estimated distance of the fault point as the intermediate fault point distance, solving the objective function based on the first reflection wave head time and the third reflection wave head time to obtain a feasible solution; When the intermediate correction coefficient of the feasible solution is within a preset value range, obtaining the intermediate correction coefficient as the correction parameter; When the intermediate correction coefficient of the feasible solution is not within the preset value range, the intermediate fault point distance is changed and the objective function is solved again.

5. The method according to any one of claims 1 to 4, characterized in that After obtaining the fault point distance output by the LSTM network, the method further includes: Obtaining information of impedance mutation points in the transmission line topology; Based on the information of the impedance mutation point and the distance to the fault point, determining whether the distance between the impedance mutation point and the fault point is less than a preset threshold; When the distance between the impedance mutation point and the fault point is less than the preset threshold, the correction coefficient is adjusted based on the information of the impedance mutation point; the expression for adjusting the correction coefficient is: α'=α·(1+βΓ); Among them, α' is the adjusted correction coefficient, β is the compensation weight coefficient, and Γ is the reflection coefficient of the impedance mutation point; An adjusted fault point distance is calculated based on the adjusted correction coefficient, the time difference between the first time and the second time, and the LSTM network; The mapping of the fault point distance to geographic coordinates based on a preset transmission line topology includes: The adjusted fault point distance is mapped into geographic coordinates based on the transmission line topology.

6. The method according to any one of claims 1 to 4, characterized in that The mapping of the fault point distance to geographic coordinates based on a preset transmission line topology includes: Acquire the longitude and latitude of the first end and the longitude and latitude of the second end from a line tower coordinate database; Based on the longitude and latitude of the first end and the longitude and latitude of the second end, linear interpolation is performed along the transmission line topology with the distance to the fault point as the path length to obtain the longitude and latitude of the fault point as the geographic coordinates.

7. The method according to any one of claims 1 to 4, characterized in that Before calculating the correction coefficient of the traveling wave velocity, the method further includes: Performing Hilbert transform on the first fault traveling wave signal or the second fault traveling wave signal to generate an analytical signal; Calculating the instantaneous frequency of the analytical signal and determining the frequency mutation point; Based on the frequency mutation point and a preset energy threshold, time data of the three consecutive reflected waves are acquired.

8. A neural network-based fault location system, characterized in that: The method applied to any one of claims 1 to 7, wherein the system comprises: an acquisition module, configured to acquire a first fault traveling wave signal received at a first end of a transmission line, and a second fault traveling wave signal received at a second end of the transmission line, wherein the first fault traveling wave signal and the second fault traveling wave signal are traveling wave signals corresponding to the same fault; a wave crest detection module, configured to acquire, based on the first fault traveling wave signal and the second fault traveling wave signal, a first time when the wave crest of the first fault traveling wave signal reaches the first end, and a second time when the wave crest of the second fault traveling wave signal reaches the second end; A calculation module, used for calculating a correction coefficient of traveling wave velocity based on time data of three consecutive reflection waves in the first fault traveling wave signal or the second fault traveling wave signal; A neural network module, used for inputting the time difference between the first time and the second time, and the correction coefficient into a preset long short-term memory (LSTM) network to obtain a fault point distance output by the LSTM network; The positioning module is used to map the fault point distance into geographic coordinates based on a preset transmission line topology to generate a fault positioning result.

9. A computing device, characterized in that include: Memory, used to store programs; A processor, configured to load the program to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 7.

Images