Method and device for predicting ground potential rise caused by switching of transformer substation
By obtaining and analyzing the incoming current and ground potential lift data in the substation back-off operation, combining the ground network model and the neural network model, efficient and accurate prediction of ground potential lift caused by the substation back-off operation is achieved, and the problem of difficulty in testing multiple measurement points at the same time in the prior art is solved.
Patent Information
- Application Number
- CN202411652215.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-19
- Publication Date
- 2025-03-28
AI Technical Summary
The prior art is difficult to simultaneously test the potential rise at multiple measurement point positions at the moment of the substation turning back operation, which makes it difficult to effectively predict the ground potential rise caused by the substation turning back.
By determining the substation's back-up operation position, the preset measurement point position, the interference source position and the zero potential reference point position, the incoming ground current measurement data and the ground potential lift measurement data are obtained, the fast Fourier transform is performed to obtain the key frequency of the waveform, the ground network model is constructed and the key frequency current is injected, the ground potential lift data of the preset measurement point position is calculated, and the neural network model is used for prediction.
It realizes efficient and accurate prediction of the transient ground potential rise of the ground grid caused by the substation turnover operation, and provides improvement in the anti-interference ability of the ground grid design and secondary equipment.
Smart Images
Figure CN119720740A_ABST
Abstract
Description
Technical Field
[0001] The present application belongs to the field of ground potential rise prediction, and in particular, relates to a method and device for predicting ground potential rise caused by substation switching. Background Art
[0002] In the field of electrical technology, the substation ground grid, as an important component of power plants and substations, plays a key role in current discharge and voltage balancing. However, since the substation ground grid is usually large in scale, when a line or primary equipment in the station fails, or a switching operation is performed, transient currents with large amplitude, high frequency and complex spectral components will be generated. These transient currents will flow along the ground grid and the soil around it, resulting in ground potential rises with different amplitudes and waveforms at different locations of the ground grid. This ground potential rise will generate electromagnetic interference throughout the ground grid, affecting the normal operation of secondary equipment.
[0003] In order to deal with this problem, it is necessary to test the transient ground potential rise generated by the substation switching operation at various locations of the ground network to determine its main frequency components and provide a reference for the anti-electromagnetic interference of secondary equipment and the optimization design of the ground network. However, due to the large scale of the substation ground network and the difficulty of transient testing, it is difficult to simultaneously test the potential rise at multiple measuring points in the substation at the moment of the switching operation. Therefore, it is necessary to propose an effective prediction method that can calculate the ground potential rise in the entire ground network based on the transient potential rise data at limited measuring points and predict the results under different interferences.
[0004] At present, there are some technical means for testing grounding devices and ground potential rise, such as grounding impedance testing of grounding devices and soil resistivity testing, but these methods mainly focus on the measurement of electrical parameters of the ground grid and do not directly solve the problem of predicting transient ground potential rise of the ground grid caused by switching operations. Summary of the invention
[0005] The purpose of the present application is to overcome the defects in the above-mentioned prior art and to provide a method and device for predicting the ground potential rise caused by substation switching.
[0006] The present application provides a method for predicting ground potential rise caused by substation switching, comprising:
[0007] Determine the substation switching operation position, preset measuring point position, interference source position and zero potential reference point position;
[0008] Acquire the ground potential rise measurement data of the preset measuring point position compared with the zero potential reference point position during the substation switching operation, and the ground current measurement data of the interference source position;
[0009] Performing a fast Fourier transform on the waveform of the ground current measurement data to obtain a key frequency of the waveform;
[0010] Constructing a substation ground grid model according to the substation ground grid structure, and injecting a current with the waveform key frequency as a unit amplitude at the substation switching operation position of the substation ground grid model;
[0011] According to the substation grounding network model, calculating the ground potential rise calculation data at the preset measuring point position from the current of the unit amplitude;
[0012] A neural network model is constructed by taking the interference source position, the ground potential rise calculation data and the preset measuring point position as the input layer and taking the ground potential rise measurement data of the preset measuring point position as the output layer;
[0013] The ground potential rise at the preset measuring point position is predicted according to the neural network model.
[0014] Optionally, determining the position of the zero potential reference point includes:
[0015] Select a point outside the substation grounding grid as the zero potential reference point position;
[0016] The electric potential at the zero potential reference point is led to each of the preset measuring points via a double-shielded cable.
[0017] Optionally, obtaining the ground potential rise measurement data of the preset measurement point position compared with the zero potential reference point position during the substation switching operation, and the ground current measurement data of the interference source position, includes:
[0018] A Rogowski coil and an oscilloscope are arranged at the interference source position to measure the ground current measurement data;
[0019] A high voltage probe and an oscilloscope are arranged at the preset measuring point to measure the ground potential rise measurement data.
[0020] Optionally, performing a fast Fourier transform on the waveform of the ground current measurement data to obtain a key frequency of the waveform includes:
[0021] The five frequencies with the largest amplitude data in the ground current measurement data waveform and their corresponding amplitude data are extracted as the waveform key frequencies.
[0022] Optionally, constructing a substation grounding network model according to the substation grounding network structure includes:
[0023] The substation ground grid model is built using CDEGS software according to the substation ground grid design drawing.
[0024] Optionally, predicting the ground potential rise at a preset measuring point location according to the neural network model includes:
[0025] Using the substation ground grid model to obtain ground potential rise calculation data at a preset measuring point location;
[0026] Based on the neural network model and the ground potential rise calculation data of the interference source position and the preset measuring point position, the ground potential rise of the preset measuring point position is predicted.
[0027] Optionally, also include error analysis:
[0028] The ground potential rise data at the preset measuring point location is compared with the actual ground potential rise measurement data at the preset measuring point location, and an error analysis is performed to evaluate the prediction accuracy of the neural network model.
[0029] Optionally, the interference source position, the ground potential rise calculation data and the preset measuring point position are used as the input layer, and the ground potential rise measurement data of the preset measuring point position is used as the output layer to construct a neural network model, including:
[0030] The number of intermediate layers and the number of nodes in each layer of the neural network model are adjusted until the expected prediction effect is achieved.
[0031] The present application also provides a device for predicting ground potential rise caused by substation switching, comprising:
[0032] Determination module, determining the substation switching operation position, preset measuring point position, interference source position and zero potential reference point position;
[0033] An acquisition module is used to acquire the ground potential rise measurement data of the preset measurement point position compared with the zero potential reference point position during the substation switching operation, and the ground current measurement data of the interference source position;
[0034] A transformation module performs a fast Fourier transform on the waveform of the ground current measurement data to obtain a key frequency of the waveform;
[0035] A simulation module, which constructs a substation ground grid model according to the substation ground grid structure, and injects a current with a unit amplitude of the waveform key frequency into the substation switching operation position of the substation ground grid model;
[0036] A calculation module, which calculates the ground potential rise calculation data at the preset measuring point position from the current of the unit amplitude according to the substation grounding network model;
[0037] A network module, which uses the interference source position, ground potential rise calculation data and preset measuring point position as input layer, and uses the ground potential rise measurement data of the preset measuring point position as output layer to construct a neural network model;
[0038] The prediction module predicts the ground potential rise at the preset measuring point position according to the neural network model.
[0039] Optionally, the network module further includes:
[0040] The number of intermediate layers and the number of nodes in each layer of the neural network model are adjusted until the best prediction effect is achieved.
[0041] The beneficial effects of this application are:
[0042] The present application provides a method for predicting ground potential rise caused by substation switching, including: determining the substation switching operation position, preset measuring point position, interference source position and zero potential reference point position; obtaining ground potential rise measurement data of the preset measuring point position compared with the zero potential reference point position during the substation switching operation, and ground current measurement data of the interference source position; performing fast Fourier transform on the waveform of the ground current measurement data to obtain the waveform key frequency; constructing a substation ground grid model according to the substation ground grid structure, and injecting a current with the waveform key frequency as a unit amplitude at the substation switching operation position of the substation ground grid model; calculating the ground potential rise calculation data of the preset measuring point position from the unit amplitude current according to the substation ground grid model; constructing a neural network model with the interference source position, the ground potential rise calculation data and the preset measuring point position as the input layer, and the ground potential rise measurement data of the preset measuring point position as the output layer; predicting the ground potential rise at the preset measuring point position according to the neural network model. This application achieves efficient and accurate prediction of the transient ground potential rise of the ground grid caused by the switching operation of the substation by combining measured data and neural network models, providing strong support for the design of the ground grid and the improvement of the anti-interference ability of secondary equipment. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] Figure 1 It is a schematic diagram of the process flow for predicting ground potential rise caused by substation switching in this application;
[0044] Figure 2 This is a logical diagram of the prediction of ground potential rise caused by substation switching in this application. DETAILED DESCRIPTION
[0045] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that various forms of implementing the present disclosure should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.
[0046] Please refer to Figure 1 and Figure 2 As shown, the present application provides a method for predicting ground potential rise caused by substation switching, the method comprising the steps of:
[0047] S101. Determine the substation switching operation position, the preset measuring point position, the interference source position and the zero potential reference point position.
[0048] The substation switching operation position refers to the specific location where the switching operation is performed, represented by the coordinates (X1, Y1). The switching operation is a common operation in the substation, involving the closing and opening of the switch, which will generate transient currents with large amplitude, high frequency and complex spectral components.
[0049] The preset measuring point (or measuring point) is a position preset on the substation ground network for measuring the ground potential rise, with coordinates (X i , Y i ), where i = 2, 3, ..., N, and N is the total number of measurement points.
[0050] The interference source location refers to the equipment or operation point that generates transient current. In this application, it refers to the location where the switching operation is performed. Since the switching operation generates transient current, it is considered to be a major source of electromagnetic interference.
[0051] The zero potential reference point refers to a selected reference point used to determine the potential of each point on the ground grid. In the substation ground grid test, a point outside the substation ground grid is selected as the zero potential reference point. The zero potential reference point is connected to each preset measuring point through a double-shielded cable so that the ground potential rise of each preset measuring point can be accurately measured.
[0052] When calculating and predicting the transient ground potential rise of the ground grid caused by the switching operation of the substation, it is first necessary to test the strength of the interference source and the ground potential rise at the measurement point.
[0053] Specifically, determine the switching operation position (X1, Y1), the number of preset measuring points N and their specific positions (X i , Y i ), where i = 2, 3,…, N.
[0054] At the same time, a point outside the substation grounding grid is selected as the zero potential reference point.
[0055] A Rogowski coil and an oscilloscope are set up at the interference source to test transient ground current. The Rogowski coil can accurately measure high-frequency transient current, while the oscilloscope can record the waveform of the current changing over time.
[0056] High-voltage probes and oscilloscopes are set up at the preset measuring points to test the ground potential rise. The high-voltage probe can safely measure high voltages, and the oscilloscope records the waveform of the ground potential rise over time.
[0057] Set the oscilloscope to trigger mode and wait for the switching operation to occur. Once the switching operation occurs, the oscilloscope will automatically trigger and record the waveforms of the current and ground potential rise.
[0058] S102, obtaining ground potential rise measurement data of the preset measurement point position compared with the zero potential reference point position during the substation switching operation, and ground current measurement data of the interference source position.
[0059] After the switching operation is completed, the test data is obtained and recorded, including: the ground potential rise measurement data of the preset measurement point compared with the zero potential reference point (waveform is VTi, i=2…N), and the ground current measurement data of the interference source location (waveform is I).
[0060] S103: Perform fast Fourier transform on the waveform of the ground current measurement data to obtain the waveform key frequency.
[0061] During the substation switching operation, transient currents with large amplitude, high frequency and complex spectral components will be generated. These currents will be injected into the ground grid and dispersed along the ground grid and the surrounding soil. In order to analyze the impact of these transient currents on the ground potential rise at various locations of the ground grid, it is necessary to accurately measure these currents and extract their key frequency components.
[0062] The ground current waveform data is subjected to fast Fourier transform (FFT).
[0063] FFT is a mathematical method that converts a time domain signal into a frequency domain signal. It decomposes a complex waveform into a series of sinusoidal wave components with different frequencies.
[0064] Through FFT transformation, the spectrum of the ground current waveform is obtained, that is, the amplitude data and phase information of different frequency components. The key frequency components are extracted from the spectrum. These key frequency components are frequency components with larger amplitude data, and they have the most significant impact on the ground potential rise.
[0065] According to actual needs, several frequency components with the largest amplitude data are selected as key frequencies, for example, five frequency components with the largest amplitude data are selected.
[0066] S104, constructing a substation ground grid model according to the substation ground grid structure, and injecting a current with the waveform key frequency as a unit amplitude at the substation switching operation position of the substation ground grid model.
[0067] By simulating the current injection during actual switching operations, the potential rise of each measuring point in the substation ground grid at the critical frequency is calculated, thus providing basic data for subsequent ground potential rise prediction.
[0068] Use CDEGS software to build a substation ground grid model based on the actual design drawing of the substation ground grid. CDEGS is a professional grounding and electromagnetic field analysis software that can accurately simulate the electrical behavior of the ground grid. By building a model, subsequent calculations and analysis can be easily performed.
[0069] In the substation grounding grid model that has been built, simulate the injection of a current of unit amplitude at the switching operation position (X1, Y1). Use the relevant modules of CDEGS software (such as MALZ) to inject current at the specified switching operation position.
[0070] S105 , calculating ground potential rise calculation data at the preset measuring point location from the current of unit amplitude according to the substation grounding network model.
[0071] Calculate the potential rise of other measurement points (2 to N-1) at each key frequency.
[0072] The above has injected a current of unit amplitude at the switching operation position ((X1, Y1).
[0073] Use the MALZ module in the CDEGS software to calculate the potential rise at 2 to N-1 preset measurement points at each key frequency.
[0074] The calculated potential rise data of each preset measuring point is subjected to inverse fast Fourier transform (FFT inverse transform). Using FFT inverse transform technology, the potential rise data in the frequency domain is converted back to the time domain to obtain the potential rise waveform V of each preset measuring point. Ci The calculation results are converted back into actual measurable time domain waveforms to provide basic data for subsequent ground potential rise prediction and model correction.
[0075] S106, constructing a neural network model with the interference source position, ground potential rise calculation data and preset measuring point position as input layer, and with the ground potential rise measurement data of the preset measuring point position as output layer.
[0076] Perform FFT on the calculation results of the preset measuring point positions, extract the five frequencies with the largest amplitude data and amplitude data, and record these data.
[0077] After extracting key data, a data set is created.
[0078] The data set includes a data set of the interference source location and a data set of each preset measurement point location.
[0079] Interference source location data set {F1, V1}: This data set contains the location information of the interference source (X1, Y1) and the key frequency and amplitude data of the ground current waveform ((f I1 , I1), (f I2 , I2),…(f I5 , I5)). These data reflect the intensity and frequency characteristics of the interference source.
[0080] Each preset measuring point location data set {F1, V1} (i = 2, 3 ... N). Each preset measuring point location data set contains the location information of the preset measuring point (X i , Y i ), as well as the key frequency and amplitude data of the ground potential rise test waveform and calculation waveform at the preset measuring point location.
[0081] Key frequency and amplitude data of the test waveform (f_i VT1 , V_i T1 ), (f_i VT2 , V_i T2 ), …(f_i VT5 , V_i T5 )) reflects the actual measured ground potential rise, while the key frequency and amplitude data of the calculated waveform (f_i VC1 , V_i C1 ), (f_i VC2 , V_i C2 ), …(f_i VC5 , V_i C5 )) is obtained through model calculation.
[0082] Together, these datasets form the base data used to refine the model.
[0083] The above data set contains not only location information, but also key parameters such as frequency and amplitude data, which are essential for the subsequent construction and training of neural network models. By comparing the differences between the test waveform and the calculated waveform, the model parameters can be continuously adjusted and optimized to improve the accuracy and reliability of the model.
[0084] Furthermore, the spatial position of the Nth preset measuring point and the ground potential rise calculation waveform data set {F in_N , V in_N}={(X N , Y N ), (f_N VC1 , V_N C1 ), (f_N VC2 , V_N C2 ), …(f_N VC5 , V_N C5)} and the spatial position of the first interference source and the ground current waveform data set {F in_1 , I in_1}={F1, V1}={(X1, Y1), (f I1 , I1), (f I2 , I2),…(f I5 , I5)}, and obtain the output of the model {F, V} and the test waveform data set {(f_i VT1 , V_i T1 ), (f_i VT2 , V_i T2 ), …(f_i VT5 , V_i T5 )} for comparison and error analysis.
[0085] S107: predicting the ground potential rise at the preset measuring point location according to the neural network model.
[0086] Determine the location information of the points to be detected, which will be used as the input of the neural network model.
[0087] According to the ground grid model and known key frequency components, the theoretical potential rise waveforms of the locations of these points to be detected under the action of the interference source are calculated.
[0088] The location information of the point to be detected, the potential rise calculation waveform of the point to be detected (key frequency components and their amplitudes extracted by FFT transformation, etc.), and the current waveform data set at the interference source location are used as inputs of the neural network. Based on these inputs, the neural network will calculate the ground potential rise prediction waveform at the location of the point to be detected using its internal weights, biases and other parameters.
[0089] Output the prediction result, that is, the transient ground potential rise waveform of the ground grid at the location of the detection point.
[0090] The present application also provides a device for predicting ground potential rise caused by substation switching, comprising:
[0091] Determination module, determining the substation switching operation position, preset measuring point position, interference source position and zero potential reference point position;
[0092] An acquisition module is used to acquire the ground potential rise measurement data of the preset measurement point position compared with the zero potential reference point position during the substation switching operation, and the ground current measurement data of the interference source position;
[0093] A transformation module performs a fast Fourier transform on the waveform of the ground current measurement data to obtain a key frequency of the waveform;
[0094] A simulation module, which constructs a substation ground grid model according to the substation ground grid structure, and injects a current with a unit amplitude of the waveform key frequency into the substation switching operation position of the substation ground grid model;
[0095] A calculation module, which calculates the ground potential rise calculation data at the preset measuring point position from the current of the unit amplitude according to the substation grounding network model;
[0096] A network module, which uses the interference source position, ground potential rise calculation data and preset measuring point position as input layer, and uses the ground potential rise measurement data of the preset measuring point position as output layer to construct a neural network model;
[0097] The prediction module predicts the ground potential rise at the preset measuring point position according to the neural network model.
[0098] Optionally, the network module constructs a neural network model, further comprising:
[0099] The number of intermediate layers and the number of nodes in each layer of the neural network model are adjusted until the best prediction effect is achieved.
[0100] The above are only specific embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A method for predicting ground potential rise caused by substation switching, characterized in that: include: Determine the substation switching operation position, preset measuring point position, interference source position and zero potential reference point position; Acquire the ground potential rise measurement data of the preset measuring point position compared with the zero potential reference point position during the substation switching operation, and the ground current measurement data of the interference source position; Performing a fast Fourier transform on the waveform of the ground current measurement data to obtain a key frequency of the waveform; Constructing a substation ground grid model according to the substation ground grid structure, and injecting a current with the waveform key frequency as a unit amplitude at the substation switching operation position of the substation ground grid model; According to the substation grounding network model, calculating the ground potential rise calculation data at the preset measuring point position from the current of the unit amplitude; A neural network model is constructed by taking the interference source position, the ground potential rise calculation data and the preset measuring point position as the input layer and taking the ground potential rise measurement data of the preset measuring point position as the output layer; The ground potential rise at the preset measuring point position is predicted according to the neural network model.
2. According to claim 1, a method for predicting ground potential rise caused by substation switching, characterized in that: Determine the zero potential reference point location, including: Select a point outside the substation grounding grid as the zero potential reference point position; The electric potential at the zero potential reference point is led to each of the preset measuring points via a double-shielded cable.
3. According to claim 1, a method for predicting ground potential rise caused by substation switching, characterized in that: Acquiring the ground potential rise measurement data of the preset measuring point position compared with the zero potential reference point position during the substation switching operation, and the ground current measurement data of the interference source position, including: A Rogowski coil and an oscilloscope are arranged at the interference source position to measure the ground current measurement data; A high voltage probe and an oscilloscope are arranged at the preset measuring point to measure the ground potential rise measurement data.
4. According to claim 1, a method for predicting ground potential rise caused by substation switching, characterized in that: Performing a fast Fourier transform on the waveform of the ground current measurement data to obtain the key frequency of the waveform includes: The five frequencies with the largest amplitude data in the ground current measurement data waveform and their corresponding amplitude data are extracted as the waveform key frequencies.
5. According to claim 1, a method for predicting ground potential rise caused by substation switching, characterized in that: The step of constructing a substation ground grid model according to the substation ground grid structure includes: The substation ground grid model is built using CDEGS software according to the substation ground grid design drawing.
6. A method for predicting ground potential rise caused by substation switching according to claim 1, characterized in that: Predicting the ground potential rise at a preset measuring point location according to the neural network model includes: Using the substation ground grid model to obtain ground potential rise calculation data at a preset measuring point location; Based on the neural network model and the ground potential rise calculation data of the interference source position and the preset measuring point position, the ground potential rise of the preset measuring point position is predicted.
7. A method for predicting ground potential rise caused by substation switching according to claim 1, characterized in that: Also includes error analysis: The ground potential rise data at the preset measuring point location is compared with the actual ground potential rise measurement data at the preset measuring point location, and an error analysis is performed to evaluate the prediction accuracy of the neural network model.
8. A method for predicting ground potential rise caused by substation switching according to claim 1, characterized in that: The interference source position, the ground potential rise calculation data and the preset measuring point position are used as the input layer, and the ground potential rise measurement data of the preset measuring point position is used as the output layer to construct a neural network model, including: The number of intermediate layers and the number of nodes in each layer of the neural network model are adjusted until the expected prediction effect is achieved.
9. A device for predicting ground potential rise caused by substation switching, characterized in that: include: Determination module, determining the substation switching operation position, preset measuring point position, interference source position and zero potential reference point position; An acquisition module is used to acquire the ground potential rise measurement data of the preset measurement point position compared with the zero potential reference point position during the substation switching operation, and the ground current measurement data of the interference source position; A transformation module performs a fast Fourier transform on the waveform of the ground current measurement data to obtain a key frequency of the waveform; A simulation module, which constructs a substation ground grid model according to the substation ground grid structure, and injects a current with a unit amplitude of the waveform key frequency into the substation switching operation position of the substation ground grid model; A calculation module, which calculates the ground potential rise calculation data at the preset measuring point position from the current of the unit amplitude according to the substation grounding network model; A network module, which uses the interference source position, ground potential rise calculation data and preset measuring point position as input layer, and uses the ground potential rise measurement data of the preset measuring point position as output layer to construct a neural network model; The prediction module predicts the ground potential rise at the preset measuring point position according to the neural network model.
10. A device for predicting ground potential rise caused by substation switching according to claim 9, characterized in that: The network module also includes: The number of intermediate layers and the number of nodes in each layer of the neural network model are adjusted until the best prediction effect is achieved.