Electromagnetic interference prediction method, device and equipment under lightning stroke line working condition

Abstract

The invention provides an electromagnetic interference prediction method, device and equipment under a lightning stroke line working condition, and belongs to the field of electromagnetic compatibility, and the method comprises the steps: constructing a lightning current model, and building a wire coupling model; generating electromagnetic pulse energy of interference electromagnetic waves according to the total induced overvoltage and the coupling voltage of the cable, and performing simulation to obtain an electromagnetic interference data training sample and a verification sample; constructing a neural network, mapping the processed training sample to the neural network for training, and verifying the trained neural network by using the processed verification sample to obtain a lightning stroke line terminal electromagnetic interference pulse prediction model; and inputting a lightning stroke line terminal electromagnetic interference pulse prediction model based on each sampling point, wherein an output result of the lightning stroke line electromagnetic interference pulse prediction model is a prediction result. According to the method, the to-be-solved result is directly predicted by utilizing the existing simulation calculation parameters and results, so that the operation complexity of simulation calculation is greatly reduced, and the calculation time is saved.

Description

Technical Field

[0001] The present application relates to the field of electromagnetic compatibility, and in particular to a method for predicting electromagnetic interference under lightning-struck line conditions. The present application also relates to a device and equipment for predicting electromagnetic interference under lightning-struck line conditions. Background Art

[0002] In the actual operation of the power system, lightning accidents are an important factor causing the safe operation of the distribution network. At present, there are no mature lightning protection measures for low-voltage distribution line terminal equipment, such as smart meters and other transmission line secondary equipment. When the transmission line is struck by lightning, due to the wide spectrum, high energy and wide coverage of lightning electromagnetic waves, the rapidly changing transient electromagnetic field is more likely to cause electromagnetic interference to the distribution line through the coupling effect of overhead wires, and will also form an induced overvoltage at its terminal, which greatly affects the normal operation of the distribution line terminal equipment. When the distribution line is struck by lightning, the lightning surge invading along the line can easily cause damage to the electrical and electronic equipment on the low-voltage side of the transformer, and power supply failures caused by surge pulses that damage secondary equipment often occur. Therefore, whether the distribution line is operating reliably and safely is not only related to the benefits of the enterprise and the personal safety of electricity users, but also to whether the entire power system can operate normally.

[0003] Domestic and foreign scholars have done a lot of research on the problem of coupled lightning electromagnetic pulses generated after overhead lines are struck by lightning. The internationally recognized models of coupled electromagnetic pulses for overhead lines are mainly the Taylor model, Agrawal model and Rachidi model based on the Maxwell equation. In China, Wang Hao, Chen Yazhou and others have simulated the law of coupled lightning electromagnetic pulses of twisted pairs based on the finite difference time domain method, and analyzed the influence of twisted pair length, different terminal loads and electric field polarization direction on its coupled voltage characteristics. At present, there are many methods for simulation and calculation of electromagnetic transient processes based on instantaneous values, such as PSCAD / EMTDC, EMTP, PSAPAC, etc.

[0004] In the prior art, the electromagnetic transient process simulation calculation method based on instantaneous values ​​has a very complex solution process, a long calculation time and a large prediction error, and cannot meet the requirements of fast and accurate prediction. Summary of the invention

[0005] The purpose of this application is to overcome the technical problems in the above-mentioned prior art that the electromagnetic transient process simulation solution process is very complicated, the calculation time is long and the prediction error is large, and to provide an electromagnetic interference prediction method under lightning line conditions. The application also relates to an electromagnetic interference prediction device and equipment under lightning line conditions.

[0006] The present application provides a method for predicting electromagnetic interference under lightning-struck line conditions, comprising:

[0007] Construct lightning current model and establish conductor coupling model;

[0008] Simulating the electromagnetic interference data training samples and verification samples based on the total induced overvoltage of the cable of the conductor coupling model and the electromagnetic pulse energy of the interference electromagnetic wave generated by the coupling voltage, including selecting sample points on the input variable set of the electromagnetic interference pulse prediction model, and constructing samples using the induced overvoltage amplitude of the line corresponding to each sample point, the corresponding induced current and the electromagnetic interference pulse energy;

[0009] Construct a neural network, map the processed training samples to the neural network for training, and use the processed verification samples to verify the trained neural network to obtain a lightning line terminal electromagnetic interference pulse prediction model;

[0010] Based on each sample point input into the lightning stroke line terminal electromagnetic interference pulse prediction model, the output result of the lightning stroke line electromagnetic interference pulse prediction model is the prediction result.

[0011] Optional, including: using a double exponential wave to construct a lightning current model, the expression is:

[0012] i(t)=I0(e -αt -e -βt )

[0013] Where: α, β are time constants; I0 is the amplitude of lightning current.

[0014] Optionally, the scattered wave induced overvoltage and corresponding induced current on the transmission line include:

[0015] V S (a,ω)=Z0{[K1+P(a,ω)]e -γa -[K2+Q(x,ω)]e γa}

[0016] I(a,ω)=[K1+P(a,ω)]e -γa +[K2+Q(x,ω)]e γa

[0017]

[0018]

[0019] γ=jωC(R+jωL)

[0020] Among them, V S is the scattered wave induced overvoltage on the transmission line; I is the corresponding induced current; ω is the scattered wave angle; a is the distance from the lightning strike point; h is the overhead height of the transmission line; γ is the propagation constant; is the horizontal electric field generated by the lightning current; C and L are the capacitance and inductance per unit length of the cable respectively; Z0=10.4Ω is the characteristic impedance of the transmission line; K1=2, K2=5.

[0021] Optionally, the electromagnetic pulse energy of the interfering electromagnetic wave generated by the coupling voltage is calculated using the following formula:

[0022] e=∑V 2 Δt

[0023] Where V is the total induced overvoltage amplitude and Δt is the propagation time, which is 1s.

[0024] Optionally, the electromagnetic interference data training samples and verification samples are obtained by simulating the electromagnetic pulse energy of the interference electromagnetic wave generated by the total induced overvoltage of the wire coupling model cable and the coupling voltage, including:

[0025] Select the lightning current amplitude and distance from the lightning strike point;

[0026] The height of the transmission line from the ground is used as the input variable of the terminal electromagnetic interference pulse prediction model of the lightning-struck line;

[0027] S+N sample points are selected from the input variable set, where S and N are both natural numbers. The lightning current amplitude takes the single pulse impulse current amplitude, and sampling is performed within the range of the distance from the lightning strike point and the height of the transmission line from the ground.

[0028] Optionally, also include:

[0029] The induced overvoltage amplitude and corresponding induced current of the line corresponding to each sample point are obtained, and the electromagnetic interference pulse energy is used to construct the sample;

[0030] A total number of samples S+N is obtained, of which there are S training samples and N verification samples, and the electric field waveform of the incident plane wave is determined.

[0031] Optionally, build a neural network, including:

[0032] The input layer contains three elements: lightning current amplitude, distance from the lightning strike point, and transmission line height z from the ground. The output layer contains the induced overvoltage amplitude of the line and the corresponding induced current and electromagnetic interference pulse energy. The optional neural network is constructed, which also includes:

[0033] Normalize the S training samples and N validation samples so that their values ​​are in the range [-1,1].

[0034] The present application also provides an electromagnetic interference prediction device under lightning-struck line conditions, comprising:

[0035] Model module, used to construct lightning current model and establish conductor coupling model;

[0036] A sample module is used to simulate and obtain electromagnetic interference data training samples and verification samples according to the total induced overvoltage of the cable of the conductor coupling model and the electromagnetic pulse energy of the interference electromagnetic wave generated by the coupling voltage, including selecting sample points on the input variable set of the electromagnetic interference pulse prediction model, and constructing samples using the induced overvoltage amplitude, corresponding induced current and electromagnetic interference pulse energy of the line corresponding to each sample point;

[0037] A construction module is used to construct a neural network, map the processed training samples to train the neural network, and verify the trained neural network with the processed verification samples to obtain a lightning line terminal electromagnetic interference pulse prediction model;

[0038] The prediction module is used to input the electromagnetic interference pulse prediction model of the lightning stroke line terminal based on each sample point, and the output result of the electromagnetic interference pulse prediction model of the lightning stroke line is the prediction result.

[0039] The present application also provides an electromagnetic interference prediction device under lightning-struck line conditions, comprising:

[0040] A memory for storing an execution program for performing electromagnetic interference prediction under lightning-struck line conditions;

[0041] The processor is used to call the execution program from the memory and execute: constructing a lightning current model and establishing a conductor coupling model; simulating the electromagnetic pulse energy of the interference electromagnetic wave generated by the total induced overvoltage of the cable of the conductor coupling model and the coupling voltage to obtain electromagnetic interference data training samples and verification samples, including selecting sample points on the input variable set of the electromagnetic interference pulse prediction model, and constructing samples using the induced overvoltage amplitude, corresponding induced current and electromagnetic interference pulse energy of the line corresponding to each sample point; constructing a neural network, mapping the processed training samples to train the neural network, verifying the trained neural network with the processed verification samples, and obtaining a lightning line terminal electromagnetic interference pulse prediction model; inputting the lightning line terminal electromagnetic interference pulse prediction model based on each sample point, and the output result of the lightning line electromagnetic interference pulse prediction model is the prediction result.

[0042] The beneficial effects of this application are:

[0043] The present application provides a method for predicting electromagnetic interference under lightning-strike line conditions, including: constructing a lightning current model and establishing a conductor coupling model; simulating the electromagnetic pulse energy of the electromagnetic interference data training samples and verification samples based on the total induced overvoltage of the cable of the conductor coupling model and the electromagnetic pulse energy of the interference electromagnetic wave generated by the coupling voltage, including selecting sample points on the input variable set of the electromagnetic interference pulse prediction model, and constructing samples using the induced overvoltage amplitude, corresponding induced current and electromagnetic interference pulse energy of the line corresponding to each sample point; constructing a neural network, mapping the processed training samples to the training of the neural network, and verifying the trained neural network with the processed verification samples to obtain the electromagnetic interference pulse prediction model of the lightning-strike line terminal; inputting the electromagnetic interference pulse prediction model of the lightning-strike line terminal based on each sample point, the output result of the electromagnetic interference pulse prediction model of the lightning-strike line is the prediction result. The present application uses the existing simulation calculation parameters and results to directly predict the results to be sought, greatly reducing the operational complexity of the simulation calculation and saving calculation time. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a schematic diagram of the electromagnetic interference prediction process under the lightning strike line working condition in this application.

[0045] Figure 2 It is a schematic diagram of the electromagnetic interference prediction execution process under the lightning line working condition in this application.

[0046] Figure 3 It is a schematic diagram of the conductor coupling model under the lightning strike line condition in this application.

[0047] Figure 4 It is a schematic diagram of the electromagnetic pulse energy of the interfering electromagnetic wave under different lightning current amplitudes in this application.

[0048] Figure 5 It is a schematic diagram of the neural network structure in this application.

[0049] Figure 6 It is a schematic diagram of the comparison between the electromagnetic interference pulse prediction results and the simulation results in this application.

[0050] Figure 7 It is a schematic diagram of an electromagnetic interference prediction device under lightning-struck line conditions in this application. DETAILED DESCRIPTION

[0051] The present application is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present application and implement it.

[0052] The following contents are all examples of specific implementation processes provided for describing in detail the technical solutions to be protected by this application. However, this application may also be implemented in other ways different from the descriptions herein. Persons skilled in the art may, under the guidance of the concepts of this application, adopt different technical means to implement this application. Therefore, this application is not limited to the specific embodiments below.

[0053] The present application provides a method for predicting electromagnetic interference under lightning-strike line conditions, including: constructing a lightning current model and establishing a conductor coupling model; simulating the electromagnetic pulse energy of the electromagnetic interference data training samples and verification samples based on the total induced overvoltage of the cable of the conductor coupling model and the electromagnetic pulse energy of the interference electromagnetic wave generated by the coupling voltage, including selecting sample points on the input variable set of the electromagnetic interference pulse prediction model, and constructing samples using the induced overvoltage amplitude, corresponding induced current and electromagnetic interference pulse energy of the line corresponding to each sample point; constructing a neural network, mapping the processed training samples to the training of the neural network, and verifying the trained neural network with the processed verification samples to obtain the electromagnetic interference pulse prediction model of the lightning-strike line terminal; inputting the electromagnetic interference pulse prediction model of the lightning-strike line terminal based on each sample point, the output result of the electromagnetic interference pulse prediction model of the lightning-strike line is the prediction result. The present application uses the existing simulation calculation parameters and results to directly predict the results to be sought, greatly reducing the operational complexity of the simulation calculation and saving calculation time.

[0054] Figure 1 It is a schematic diagram of the electromagnetic interference prediction process under the working condition of the lightning-struck line in this application, Figure 2 It is a schematic diagram of the electromagnetic interference prediction execution process under the lightning line working condition in this application.

[0055] like Figure 2 As shown, the process described in this application can be summarized into three steps:

[0056] S201: Obtaining electromagnetic interference training samples through simulation calculation;

[0057] S202: Constructing a prediction model for electromagnetic interference pulses of lightning-struck line terminal equipment;

[0058] S203: Implement prediction based on the constructed electromagnetic interference pulse prediction model for lightning-strike line terminal equipment.

[0059] like Figure 1 As shown, the electromagnetic interference prediction process under the lightning-struck line working condition is as follows:

[0060] S101 constructs lightning current model and establishes conductor coupling model.

[0061] In this application, a double exponential wave is used to construct a lightning current model. The double exponential wave can accurately describe the actual lightning waveform among a variety of lightning current equivalent waveforms. Its expression is:

[0062] i(t)=I0(e -αt -e -βt )

[0063] Where: α, β are time constants; I0 is the amplitude of lightning current.

[0064] like Figure 3 As shown, Figure 3 It is a schematic diagram of the conductor coupling model under the lightning strike line condition in this application.

[0065] A conductor coupling model is established. Specifically, in this application, a 10kv overhead line is used for analysis, the simulated load is 470Ω, the conductor model is JKLYJ-10, the conductor outer diameter is 5.1mm, the line inductance per kilometer is 0.4Ω / km, and the susceptance is 2.71×10-4S / km.

[0066] The earth is taken as the horizontal axis x-axis; the vertical axis is the z-axis; the height of the transmission line from the ground is h = 10m; the horizontal distance from the lightning strike point is a = 1.5m; It is the horizontal electric field formed when the lightning current is at a horizontal distance from the lightning strike point and at a height of z from the ground.

[0067] S102 performs simulation to obtain electromagnetic interference data training samples and verification samples based on the total induced overvoltage of the cable of the conductor coupling model and the electromagnetic pulse energy of the interference electromagnetic wave generated by the coupling voltage, including selecting sample points on the input variable set of the electromagnetic interference pulse prediction model, and constructing samples using the induced overvoltage amplitude, corresponding induced current and electromagnetic interference pulse energy of the line corresponding to each sample point.

[0068] Calculation of total induced overvoltage on cables:

[0069] The calculation formula of the scattered wave induced overvoltage and the corresponding induced current on the transmission line is as follows: V S (a,ω)=Z0{[K1+P(a,ω)]e -γa -[K2+Q(x,ω)]e γa} I(a,ω)=[K1+P(a,ω)]e -γa +[K2+Q(x,ω)]e γa in,

[0070] γ=jωC(R+jωL)

[0071] Where V Sis the scattered wave induced overvoltage on the transmission line; I is the corresponding induced current; the scattered wave angle ω is 13.5°; a is the distance from the lightning strike point; h is the overhead height of the transmission line; γ is the propagation constant; is the horizontal electric field generated by the lightning current, which is calculated using the approximate mirror method formula; C and L are the capacitance and inductance per unit length of the cable respectively; Z0=10.4Ω is the characteristic impedance of the transmission line; K1=2, K2=5.

[0072] Specifically, let a1 = 6 km, a2 = 3 km (a1 > a2) at both ends of the transmission line, and assume that their impedances are Z1 = 0.8 + j1.4 Ω, Z2 = 0.4 + j1 Ω, and the calculation formulas for K1 and K2 are as follows:

[0073]

[0074]

[0075] Where ρ1, ρ2 are the reflection coefficients of the cable terminals, and the calculation formula is as follows:

[0076]

[0077] The total induced overvoltage on the cable is the algebraic sum of the scattered wave induced overvoltage and the excitation voltage, that is:

[0078]

[0079] In the formula is the vertical electric field generated by the lightning current and is calculated using the approximate mirror image method formula.

[0080] Calculate the electromagnetic pulse energy of the interfering electromagnetic wave generated by the coupling voltage:

[0081] The calculation formula is

[0082] e=∑V 2 Δt

[0083] Where V is the total induced overvoltage amplitude, and Δt is the propagation time, which is 1s.

[0084] Based on the finite difference time domain method, simulation is performed to obtain electromagnetic interference data training samples, and a total of S = 48,000 sets of training data and N = 4,690 sets of verification data are obtained. Each set of data contains the lightning current amplitude I0, the distance a from the lightning strike point, the height z of the transmission line from the ground, and the electromagnetic interference pulse energy e.

[0085] Please refer to Figure 4 As shown, Figure 4 It is a schematic diagram of the electromagnetic pulse energy of the interfering electromagnetic wave under different lightning current amplitudes in this application.

[0086] Specifically, the lightning current amplitude I0 and the distance a from the lightning strike point are selected; the height z of the transmission line above the ground is the input variable of the terminal electromagnetic interference pulse prediction model of the lightning strike line; S+N sample points are selected on the input variable set, where S and N are both natural numbers, the lightning current amplitude I0 takes the single pulse impulse current amplitude, the sampling range is from 10kA to 60kA, the step size is 5kA, and the distance a from the lightning strike point and the height h of the transmission line above the ground are sampled in the range of [0-10]km and [0-10]m.

[0087] The induced overvoltage amplitude V and the corresponding induced current I of the line corresponding to each sample point (I0,a,z) obtained by simulation, and the electromagnetic interference pulse energy e are used to construct samples.

[0088] A total number of samples S+N is obtained, of which there are S training samples and N verification samples, and the electric field waveform of the incident plane wave is determined.

[0089] In this application, the lightning current amplitude I0, the distance from the lightning strike point a, and the sampling points of the training samples of the transmission line height z from the ground are respectively:

[0090] I0=n×5kA,n=0,1,…,12

[0091] a=s×50m,s=0,1,…,200

[0092] z=w×0.5m,w=0,1,…,20

[0093] A total of 48,000 sample training data.

[0094] The sampling points for verification samples are:

[0095] I0=1kA+n×5kA,n=0,1,…,10

[0096] a=100m+s×50m,s=0,1,…,67

[0097] z=1m+w×0.5m,w=0,1,…,7

[0098] A total of 4690 samples of validation data.

[0099] According to the actual overhead line model and parameters, the overhead line length L = 15km, the overhead line radius r = 2.55mm, the height from the ground h = 10m, the simulated load is 470Ω, and the earth conductivity σ and relative dielectric constant ε are set to 0.001S / H and 5 respectively.

[0100] The electromagnetic interference pulse electric field waveform in this application is the electric field waveform of the IEC standard, specifically E0(t)=kE0[exp(-βt)-exp(-αt)], where k=1.2, E0=60kv / m, α=7.0×10 8 s -1 , β=5.0×10 -7 s -1 .

[0101] S103 constructs a neural network, maps the processed training samples to train the neural network, verifies the trained neural network with the processed verification samples, and obtains a lightning line terminal electromagnetic interference pulse prediction model.

[0102] like Figure 5 As shown, Figure 5 It is a schematic diagram of the neural network structure in this application.

[0103] Construct a neural network, where the input layer contains three elements: lightning current amplitude I0, distance a from the lightning strike point; transmission line height z from the ground; the output layer contains the induced overvoltage amplitude of the line and the corresponding induced current and electromagnetic interference pulse energy, and the tan-sigmoid function is used as the activation function of the neural network. Set the number of hidden layers, the number of nodes in each hidden layer, and the termination condition. In this application, the number of hidden layers is 10, and each hidden layer node has 15 nodes. The termination condition is set to 3000 training rounds and a termination gradient of 1×10 -8 , the loss function value is 1×10 -8 .

[0104] In order to eliminate the influence of data dimension, the S training samples and N validation samples are normalized so that the values ​​are in the interval [-1,1].

[0105] In this application, 48,000 training samples were used, and the training was terminated when the number of training rounds was 894. 4,690 sets of data were used to verify the neural network. The processed training samples were then mapped to train the neural network, and the processed verification samples were used to verify the trained neural network.

[0106] The output results of the neural network are inversely mapped into the induced overvoltage amplitude of the line and the corresponding induced current and terminal electromagnetic interference pulse energy, which are compared with the induced overvoltage amplitude of the line and the corresponding induced current and electromagnetic interference pulse energy at each sample point calculated by simulation to achieve verification and obtain the electromagnetic interference pulse prediction model for the lightning-struck line terminal.

[0107] S104 inputs a lightning stroke line terminal electromagnetic interference pulse prediction model based on each sample point, and the output result of the lightning stroke line electromagnetic interference pulse prediction model is the prediction result.

[0108] In step 3, the prediction is implemented based on the constructed electromagnetic interference pulse prediction model for the terminal of the lightning stroke line. The prediction is implemented by inputting (I0, a, z) of each sample point in the test sample into the electromagnetic interference pulse prediction model for the terminal of the lightning stroke line constructed in step 2, and predicting the electromagnetic interference pulse generated by the terminal when the lightning stroke line is struck. The output result of the electromagnetic interference pulse prediction model for the lightning stroke line is the prediction result.

[0109] Please refer to Figure 6 As shown, Figure 6 It is a schematic diagram of the comparison between the electromagnetic interference pulse prediction results and the simulation results in this application.

[0110] The results show that compared with simulation calculations, the present invention greatly simplifies the calculation process, reduces the calculation time and improves the prediction accuracy, and can meet the requirements of fast and accurate prediction.

[0111] like Figure 7 As shown, the present application also provides an electromagnetic interference prediction device under lightning line working conditions, which is used to execute the above method, including:

[0112] Model module, used to construct lightning current model and establish conductor coupling model;

[0113] A sample module is used to simulate and obtain electromagnetic interference data training samples and verification samples according to the total induced overvoltage of the cable of the conductor coupling model and the electromagnetic pulse energy of the interference electromagnetic wave generated by the coupling voltage, including selecting sample points on the input variable set of the electromagnetic interference pulse prediction model, and constructing samples using the induced overvoltage amplitude, corresponding induced current and electromagnetic interference pulse energy of the line corresponding to each sample point;

[0114] A construction module is used to construct a neural network, map the processed training samples to train the neural network, and verify the trained neural network with the processed verification samples to obtain a lightning line terminal electromagnetic interference pulse prediction model;

[0115] The prediction module is used to input the electromagnetic interference pulse prediction model of the lightning stroke line terminal based on each sample point, and the output result of the electromagnetic interference pulse prediction model of the lightning stroke line is the prediction result.

[0116] The present application also provides an electromagnetic interference prediction device under lightning-struck line conditions, comprising:

[0117] A memory for storing an execution program for performing electromagnetic interference prediction under lightning-struck line conditions;

[0118] The processor is used to call the execution program from the memory and execute: constructing a lightning current model and establishing a conductor coupling model; simulating the electromagnetic pulse energy of the interference electromagnetic wave generated by the total induced overvoltage of the cable of the conductor coupling model and the coupling voltage to obtain electromagnetic interference data training samples and verification samples, including selecting sample points on the input variable set of the electromagnetic interference pulse prediction model, and constructing samples using the induced overvoltage amplitude, corresponding induced current and electromagnetic interference pulse energy of the line corresponding to each sample point; constructing a neural network, mapping the processed training samples to train the neural network, verifying the trained neural network with the processed verification samples, and obtaining a lightning line terminal electromagnetic interference pulse prediction model; inputting the lightning line terminal electromagnetic interference pulse prediction model based on each sample point, and the output result of the lightning line electromagnetic interference pulse prediction model is the prediction result.

Claims

1. A method for predicting electromagnetic interference under lightning-struck line conditions, characterized in that: include: Construct lightning current model and establish conductor coupling model; Simulating the electromagnetic interference data training samples and verification samples based on the total induced overvoltage of the cable of the conductor coupling model and the electromagnetic pulse energy of the interference electromagnetic wave generated by the coupling voltage, including selecting sample points on the input variable set of the electromagnetic interference pulse prediction model, and constructing samples using the induced overvoltage amplitude of the line corresponding to each sample point, the corresponding induced current and the electromagnetic interference pulse energy; Construct a neural network, map the processed training samples to the neural network for training, and use the processed verification samples to verify the trained neural network to obtain a lightning line terminal electromagnetic interference pulse prediction model; Based on each sample point input into the lightning stroke line terminal electromagnetic interference pulse prediction model, the output result of the lightning stroke line electromagnetic interference pulse prediction model is the prediction result.

2. The electromagnetic interference prediction method under the working condition of a lightning-struck line according to claim 1 is characterized in that: include: The double exponential wave is used to construct the lightning current model, and its expression is: i(t)=I0(e -αt -Alright -βt ) Where: α, β are time constants; I0 is the amplitude of lightning current.

3. The electromagnetic interference prediction method under the working condition of a lightning-struck line according to claim 1 is characterized in that: The scattered wave induced overvoltage and corresponding induced current on the transmission line include: V S (a,ω)=Z0{[K1+P(a,ω)]e -γa -[K2+Q(x,ω)]e γa } I(a,ω)=[K1+P(a,ω)]e -γa +[K2+Q(x,ω)]e γa γ=jωC(R+jωL) Among them, V S is the scattered wave induced overvoltage on the transmission line; I is the corresponding induced current; ω is the scattered wave angle; a is the distance from the lightning strike point; h is the overhead height of the transmission line; γ is the propagation constant; is the horizontal electric field generated by the lightning current; C and L are the capacitance and inductance per unit length of the cable respectively; Z0=10.4Ω is the characteristic impedance of the transmission line; K1=2, K2=5.

4. The electromagnetic interference prediction method under lightning-struck line conditions according to claim 1, characterized in that: Calculate the electromagnetic pulse energy of the interfering electromagnetic wave generated by the coupling voltage. The calculation formula is: e=∑V 2 Δt Where V is the total induced overvoltage amplitude and Δt is the propagation time, which is 1s.

5. The electromagnetic interference prediction method under the lightning-struck line working condition according to claim 1 is characterized in that: According to the total induced overvoltage of the cable of the conductor coupling model and the electromagnetic pulse energy of the interference electromagnetic wave generated by the coupling voltage, the electromagnetic interference data training samples and verification samples are obtained by simulation, including: Select the lightning current amplitude and distance from the lightning strike point; The height of the transmission line from the ground is used as the input variable of the terminal electromagnetic interference pulse prediction model of the lightning-struck line; S+N sample points are selected from the input variable set, where S and N are both natural numbers. The lightning current amplitude takes the single pulse impulse current amplitude, and sampling is performed within the range of the distance from the lightning strike point and the height of the transmission line from the ground.

6. The electromagnetic interference prediction method under lightning-struck line conditions according to claim 5, characterized in that: Also includes: The induced overvoltage amplitude and corresponding induced current of the line corresponding to each sample point are obtained, and the electromagnetic interference pulse energy is used to construct the sample; A total number of samples S+N is obtained, of which there are S training samples and N verification samples, and the electric field waveform of the incident plane wave is determined.

7. The electromagnetic interference prediction method under the lightning-struck line working condition according to claim 1 is characterized in that: Build a neural network, including: The input layer contains three elements: lightning current amplitude, distance from the lightning strike point, and height z of the transmission line from the ground. The output layer contains the induced overvoltage amplitude of the line and the corresponding induced current and electromagnetic interference pulse energy. The tan-sigmoid function is used as the activation function of the neural network.

8. The electromagnetic interference prediction method under lightning-struck line conditions according to claim 7, characterized in that: Building a neural network also includes: Normalize the S training samples and N validation samples so that their values ​​are in the range [-1,1].

9. An electromagnetic interference prediction device under lightning-struck line conditions, characterized in that: include: Model module, used to construct lightning current model and establish conductor coupling model; A sample module is used to simulate and obtain electromagnetic interference data training samples and verification samples according to the total induced overvoltage of the cable of the conductor coupling model and the electromagnetic pulse energy of the interference electromagnetic wave generated by the coupling voltage, including selecting sample points on the input variable set of the electromagnetic interference pulse prediction model, and constructing samples using the induced overvoltage amplitude, corresponding induced current and electromagnetic interference pulse energy of the line corresponding to each sample point; A construction module is used to construct a neural network, map the processed training samples to train the neural network, and verify the trained neural network with the processed verification samples to obtain a lightning line terminal electromagnetic interference pulse prediction model; The prediction module is used to input the electromagnetic interference pulse prediction model of the lightning stroke line terminal based on each sample point, and the output result of the electromagnetic interference pulse prediction model of the lightning stroke line is the prediction result.

10. An electromagnetic interference prediction device under lightning-struck line conditions, characterized in that: include: A memory, used to store an execution program for executing any one of the methods described in claims 1 to 8; The processor is used to call the execution program from the memory and execute: constructing a lightning current model and establishing a conductor coupling model; simulating the electromagnetic pulse energy of the interference electromagnetic wave generated by the total induced overvoltage of the cable of the conductor coupling model and the coupling voltage to obtain electromagnetic interference data training samples and verification samples, including selecting sample points on the input variable set of the electromagnetic interference pulse prediction model, and constructing samples using the induced overvoltage amplitude, corresponding induced current and electromagnetic interference pulse energy of the line corresponding to each sample point; constructing a neural network, mapping the processed training samples to train the neural network, verifying the trained neural network with the processed verification samples, and obtaining a lightning line terminal electromagnetic interference pulse prediction model; inputting the lightning line terminal electromagnetic interference pulse prediction model based on each sample point, and the output result of the lightning line electromagnetic interference pulse prediction model is the prediction result.

Images