GIS partial discharge ultrahigh frequency electromagnetic wave signal propagation characteristic simulation method, system, medium and equipment

Abstract

The invention discloses a GIS partial discharge ultrahigh frequency electromagnetic wave signal propagation characteristic simulation method and system, a medium and equipment in the technical field of GIS partial discharge detection, and aims to solve the optimization problem of the existing detection technology. The method comprises the following steps: constructing a GIS model according to the size of a GIS and electrical parameters of a material of the GIS; carrying out grid division on the GIS model by adopting a self-adaptive grid division strategy; according to the grid density, partial discharge signal source parameters and data acquisition points in the GIS model are set, data acquisition is carried out, and a partial discharge signal data set is obtained; performing adaptive signal processing on the partial discharge signal data set, and inputting the processed signal data set into the trained CNN-LSTM combination model: performing convolution operation on the processed signal data set through a CNN layer to obtain partial discharge signal features; and according to the partial discharge signal characteristics, performing characteristic analysis by using an LSTM layer to obtain GIS partial discharge ultrahigh frequency propagation characteristics.

Description

Technical Field

[0001] The present invention relates to the technical field of GIS partial discharge detection, and particularly to a simulation method, system, medium and device for the propagation characteristics of UHF electromagnetic wave signals generated by GIS partial discharge. Background Art

[0002] GIS is an important part of the high-voltage power grid, and its insulation performance plays a crucial role in the stability of the system. GIS adopts a closed metal shell structure, and the internal electric field distribution is highly concentrated. During the manufacturing and operation processes, local discharge (PD) may occur due to process defects or external contamination. The interface between the insulating material and the high-voltage conductor inside GIS is complex, and the discharge phenomenon is not only affected by the insulating medium but also restricted by the metal shielding effect. Conducting in-depth research on the propagation characteristics of the UHF electromagnetic wave signals generated by GIS partial discharge can not only optimize the on-site partial discharge detection scheme but also provide theoretical support for the structural design and on-line monitoring of GIS equipment.

[0003] The partial discharge signals inside GIS are generated in the form of nanosecond-level pulse currents and are accompanied by high-frequency electromagnetic radiation. The signal frequency range usually reaches several GHz. Since GIS adopts a fully enclosed structure, the propagation of electromagnetic waves is limited by the conductive characteristics of the equipment shell. At the same time, the complex dielectric environment inside it will cause obvious interference and attenuation effects of electromagnetic waves between multiple layers of structures. Compared with transformers, the high-voltage conductor and the grounded metal shell inside GIS form a unique shielding structure, making the propagation path of electromagnetic signals highly dependent on the relative positions and arrangement methods of internal components. Therefore, studying the propagation characteristics of GIS partial discharge UHF signals, such as signal attenuation, spectral characteristics, and multipath interference, is crucial for optimizing the sensor layout and improving the detection accuracy.

[0004] The ultra-high frequency (UHF) partial discharge detection method is the core technology for current GIS partial discharge detection. This method relies on UHF sensors to receive high-frequency electromagnetic wave signals and analyze their characteristic parameters to determine the location, intensity, and type of partial discharge sources. The instantaneous current changes caused by internal partial discharges in GIS generate high-frequency electromagnetic waves. However, due to the closed metal shell structure of GIS, the propagation characteristics of signals are affected by the metal shielding effect, and some signals may be absorbed by the shell or reflected back into the GIS, forming complex multipath propagation. The typical GIS sensor arrangement is to install sensors at grounding flanges, pressure relief valves, or other parts that are easy to couple electromagnetic waves to optimize the signal reception effect. After filtering, amplifying, and feature extraction of the signals, location methods based on time delay differences or signal intensity attenuation models can be used to estimate the location of partial discharge sources. Due to the strong shielding effect in the GIS partial discharge detection environment, this method usually combines a multi-sensor collaborative detection strategy to improve the signal capture efficiency and reduce the misjudgment probability caused by the weakening of signals at a single measurement point.

[0005] Due to the closed metal shell structure of GIS, the propagation path of partial discharge signals inside is restricted, and some signals may not be effectively transmitted to the detection point due to the metal shielding effect. This leads to a reduction in detection accuracy and may cause misjudgment. In addition, components such as high-voltage conductors and basin insulators in the GIS structure form a complex dielectric environment, causing the electromagnetic wave signals to be refracted and reflected multiple times during propagation, resulting in signal arrival time aliasing and affecting the accurate positioning of the discharge source. To solve these problems, it is necessary to optimize the sensor arrangement and adopt multi-sensor fusion technology to improve the signal capture rate and positioning accuracy. Summary of the Invention

[0006] The purpose of the present invention is to overcome the deficiencies in the prior art and provide a simulation method, system, medium, and device for the propagation characteristics of UHF electromagnetic wave signals in GIS partial discharge, to study the propagation behavior of UHF electromagnetic wave signals generated by partial discharges in high-voltage GIS equipment, and to analyze the attenuation law, propagation path, and frequency spectrum distribution of UHF electromagnetic wave signals inside GIS through simulation methods to guide the optimization of GIS partial discharge detection technology.

[0007] To solve the above technical problems, the present invention is implemented by adopting the following technical solutions:

[0008] In the first aspect, the present invention provides a simulation method for the propagation characteristics of UHF electromagnetic wave signals in GIS partial discharge, including:

[0009] Construct a GIS model according to the size of the GIS and the electrical parameters of its materials;

[0010] An adaptive mesh generation strategy is adopted to mesh the GIS model, and different mesh densities are set in different regions inside the GIS model;

[0011] According to the mesh density, set the partial discharge signal source parameters and data acquisition points in the GIS model, perform data acquisition, and obtain a partial discharge signal data set;

[0012] Perform adaptive signal processing on the partial discharge signal data set to obtain a processed signal data set;

[0013] Input the processed signal data set into the trained CNN-LSTM combined model: perform convolution operations on the processed signal data set through the CNN layer to obtain partial discharge signal features; according to the partial discharge signal features, use the LSTM layer for feature analysis to obtain the UHF propagation characteristics of GIS partial discharge.

[0014] Optionally, the GIS model includes a high-voltage conductor model, a basin insulator model, and a GIS housing model; the high-voltage conductor model is used to transmit electric energy to ensure the efficient flow of current inside the GIS model; the basin insulator model is used to support and isolate the high-voltage conductor to prevent arcs and short circuits; the GIS housing model is used to seal the internal equipment of the GIS model.

[0015] Optionally, the mesh generation is realized by reinforcement learning, and dense meshes are used within a preset distance from the partial discharge source and in areas where the electric field strength change exceeds the threshold, while sparse meshes are used in other areas.

[0016] Optionally, the partial discharge signal source parameters include pulse type, time width, voltage amplitude, and signal source position; the data acquisition points are arranged in the tangential and axial directions of the GIS housing, and the GIS model generates partial discharge signals using a generative adversarial network.

[0017] Optionally, the adaptive signal processing includes wavelet transform and empirical mode decomposition to enhance the detection accuracy of high-frequency components in the partial discharge signal data set and remove low-frequency noise.

[0018] Optionally, the CNN-LSTM combined model uses the mean squared error as the loss function and uses an adaptive gradient optimization method to update the weights of the neural network in the CNN-LSTM combined model.

[0019] In a second aspect, the present invention provides a simulation system for the UHF electromagnetic wave signal propagation characteristics of GIS partial discharge, including:

[0020] A model construction module, configured to: construct a GIS model according to the size of the GIS and the electrical parameters of its material; adopt an adaptive mesh generation strategy to perform mesh generation on the GIS model, and set different mesh densities in different regions inside the GIS model;

[0021] A simulation setting module, configured to: according to the mesh density, set the partial discharge signal source parameters and data acquisition points in the GIS model, perform data acquisition to obtain a partial discharge signal data set; perform adaptive signal processing on the partial discharge signal data set to obtain a processed signal data set;

[0022] A result analysis module, configured to: input the processed signal data set into a trained CNN-LSTM combined model; perform convolution operations on the processed signal data set through the CNN layer to obtain partial discharge signal features; according to the partial discharge signal features, use the LSTM layer to perform feature analysis to obtain the UHF propagation characteristics of GIS partial discharge.

[0023] In a third aspect, the present invention provides a computer-readable storage medium, on which computer instructions are stored. When the computer instructions are executed by a processor, the steps of the GIS partial discharge UHF electromagnetic wave signal propagation characteristic simulation method described in any one of the first aspects are implemented.

[0024] In a fourth aspect, the present invention provides a computer device, including:

[0025] A memory, configured to store computer instructions;

[0026] A processor, configured to execute the computer instructions to implement the steps of the GIS partial discharge UHF electromagnetic wave signal propagation characteristic simulation method described in any one of the first aspects.

[0027] In a fifth aspect, the present invention provides a computer program product, including computer instructions, characterized in that when the computer instructions are executed by a processor, the steps of the GIS partial discharge UHF electromagnetic wave signal propagation characteristic simulation method described in any one of the first aspects are implemented.

[0028] Compared with the prior art, the beneficial effects achieved by the present invention:

[0029] 1. The GIS partial discharge UHF electromagnetic wave signal propagation characteristic simulation method provided by the present invention, through neural network structure optimization, reinforcement learning to optimize simulation parameters, generative adversarial network to generate data augmentation, and adaptive signal processing, provides a precise analysis means for the UHF signal propagation of GIS partial discharge, and at the same time provides a scientific basis for the partial discharge detection scheme of GIS equipment, improving the reliability and sensitivity of detection;

[0030] 2. The GIS partial discharge UHF electromagnetic wave signal propagation characteristic simulation system provided by the present invention can perform simulation calculations based on the GIS structure characteristics by setting up a model construction module, a simulation setting module, and a result analysis module. Combining with the deep neural network (CNN-LSTM) modeling, it optimizes the arrangement of detection sensors and provides an accurate local discharge source positioning scheme, thereby improving the fault detection efficiency and reliability of GIS equipment;

[0031] 3. The computer-readable storage medium, computer device, and computer program product provided by the present invention can execute the steps of the GIS partial discharge UHF electromagnetic wave signal propagation characteristic simulation method provided by the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 is a flowchart of the GIS partial discharge UHF electromagnetic wave signal propagation characteristic simulation method provided according to an embodiment of the present invention;

[0033] Figure 2 is a schematic diagram of the GIS model and its external space adaptive mesh division provided according to an embodiment of the present invention;

[0034] Figure 3 is a schematic diagram of the local discharge source parameter setting provided according to an embodiment of the present invention;

[0035] Figure 4 is a schematic diagram of the data acquisition point setting provided according to an embodiment of the present invention;

[0036] Figure 5 is a schematic diagram of the neural network confusion matrix provided according to an embodiment of the present invention;

[0037] Figure 6 is a waveform diagram of the measured electric field strength modulus value changing with time provided according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0038] The technical solution of the present invention will be described in detail below through the drawings and specific embodiments. It should be understood that the specific features in the embodiments of the present application and the embodiments are detailed descriptions of the technical solution of the present application, rather than limitations on the technical solution of the present application. Without conflict, the technical features in the embodiments of the present application and the embodiments can be combined with each other.

[0039] It should be noted that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this article generally represents an "or" relationship between the associated objects before and after.

[0040] Embodiment 1:

[0041] An embodiment of the present invention discloses a simulation method for the propagation characteristics of UHF electromagnetic wave signals of partial discharge in GIS. Refer to Figure 1 as shown, the specific steps are as follows:

[0042] The present invention provides a simulation method for the propagation characteristics of UHF electromagnetic wave signals of partial discharge in GIS, including:

[0043] S1. Construct a GIS model according to the size of the GIS and the electrical parameters of its materials;

[0044] S2. Adopt an adaptive mesh generation strategy to mesh the GIS model, and set different mesh densities in different regions inside the GIS model;

[0045] S3. According to the mesh density, set the partial discharge signal source parameters and data acquisition points in the GIS model, perform data acquisition, and obtain a partial discharge signal data set;

[0046] S4. Perform adaptive signal processing on the partial discharge signal data set to obtain a processed signal data set;

[0047] S5. Input the processed signal data set into a trained CNN-LSTM combined model: perform convolution operations on the processed signal data set through the CNN layer to obtain partial discharge signal features; according to the partial discharge signal features, use the LSTM layer for feature analysis to obtain the UHF propagation characteristics of partial discharge in GIS.

[0048] Specifically,

[0049] In step S1, build a high-voltage conductor model, a pot-type insulator model, and a GIS housing model according to the sizes of the components of the gas-insulated switchgear (GIS), and set the electrical parameters of the materials of each component to accurately reflect the internal electromagnetic environment of the GIS; the GIS model includes a high-voltage conductor model, a pot-type insulator model, and a GIS housing model; the high-voltage conductor model is used to transmit electric energy to ensure the efficient flow of current in the GIS model; the pot-type insulator model is used to support and isolate the high-voltage conductor to prevent arc and short circuit; the GIS housing model is used to seal the internal equipment of the GIS model.

[0050] In step S2, refer to Figure 2As shown, the mesh generation is achieved by reinforcement learning. Denser meshes are used within a preset distance from the local power source and in areas where the change in electric field strength exceeds the threshold, while larger-sized meshes are used in more distant areas to optimize the utilization of computing resources while ensuring calculation accuracy. Reinforcement learning is a machine learning method for making optimal decisions based on a trial and feedback mechanism. By examining the simulation performance under different mesh generation strategies, the mesh density distribution is continuously adjusted to achieve the best balance between calculation accuracy and calculation time, improving the simulation accuracy of the complex electromagnetic environment inside the GIS.

[0051] In step S3, refer to Figure 3 As shown, a local discharge signal source is set in the simulation environment, and parameter configuration is carried out according to typical discharge positions and possible discharge types inside the GIS. The parameters of the local discharge signal source include pulse type, time width, voltage amplitude, and signal source position. Refer to Figure 4 As shown, multiple data acquisition points are set at key areas of the GIS structure to ensure multi-dimensional acquisition of signal characteristic data. The acquisition points are distributed in different directions of the GIS shell and arranged with reasonable spacing to comprehensively obtain the propagation characteristics of the signal in the GIS and improve the analysis accuracy of the UHF signal propagation path. The GIS model uses a generative adversarial network to generate local discharge signals. The generative adversarial network is a neural network system composed of a generator and a discriminator. The generator is used to learn the characteristics of real GIS partial discharge signals and generate synthetic data, while the discriminator is responsible for evaluating whether the generated data is real. Through continuous adversarial training, the generated synthetic data becomes increasingly close to the actual GIS partial discharge signals, thereby enhancing the dataset and improving the similarity of the model to real GIS equipment.

[0052] In step S4, the adaptive signal processing includes wavelet transform and empirical mode decomposition, enhancing the detection accuracy of high-frequency components in the local discharge signal dataset and removing low-frequency noise to ensure the reliability of the input data. Wavelet transform is a mathematical transformation method that can decompose a signal into sub-signals of different scales to analyze the characteristics of different frequency components, making high-frequency features more prominent. Empirical mode decomposition is a signal decomposition method that can split a complex signal into multiple independent intrinsic mode functions, helping to identify key information in the partial discharge signal and effectively removing low-frequency interference to improve the accuracy of the data.

[0053] In step S5, the CNN-LSTM combined model uses a convolutional neural network (CNN) and a long short-term memory network (LSTM) to extract the spatio-temporal characteristics of GIS signals in order to analyze the propagation characteristics of GIS ultra-high frequency signals; CNN is a neural network specifically designed for processing grid data. It automatically extracts the spatial and frequency features of signals by performing convolutional operations on the input data, improving the feature recognition ability; while LSTM is a special recurrent neural network that can store and utilize long-term and short-term time information, so it has a stronger capture ability when analyzing time-varying signals; the CNN layer is used to automatically learn the time-frequency distribution features of signals, and the LSTM layer is used to analyze the dynamic changes of signals over time to improve the model's understanding ability of complex propagation patterns.

[0054] The CNN-LSTM combined model uses the mean squared error as the loss function and uses the adaptive gradient optimization method to update the weights of the neural network in the CNN-LSTM combined model to accelerate convergence and improve training efficiency; the adaptive gradient optimization method is an optimization algorithm used to adjust the learning rate of a neural network. It can automatically select an appropriate learning step size for different parameters, thus avoiding the problem of too fast or too slow convergence during the training process.

[0055] After the simulation runs, first collect the signals at each data collection point and use the CNN-LSTM combined model for data analysis; first, convert the collected signal data (such as the waveform of the electric field strength changing with time, spectral distribution, etc.) into a training data set, and establish an input feature space according to the characteristics of the signal source (position, amplitude, pulse parameters); then, through the neural network model, learn the propagation mode of the internal signals of the GIS and identify the complex relationships between different signal sources and collection points.

[0056] During training, focus on learning the metal shielding effect of the GIS housing and the multi-path characteristics of signal propagation, and adjust the model weights to minimize the error function; at the same time, use some samples for verification to avoid overfitting and optimize the model performance; by inputting the simulation conditions (such as signal source parameters), the model can quickly predict the signal distribution at different collection points, including the attenuation law of the electric field strength, the spectral change trend and the propagation delay, and optimize it in combination with the multiple reflection characteristics of the internal insulation structure and the metal housing of the GIS; by optimizing the simulation parameters, the calculation accuracy can be further improved to make the simulation results more consistent with the actual operating environment of the GIS.

[0057] After the training is completed, the neural network can quickly predict the signal intensity and spectral characteristics at each data collection point under different signal source conditions; combined with the simulation data analysis, the propagation path, energy attenuation characteristics and delay law of the ultra-high frequency signals of GIS partial discharge can be deduced, so as to optimize the layout of the data collection points and improve the sensitivity of partial discharge detection.

[0058] This embodiment further analyzes the prediction results, identifies possible signal capture blind spots or over-dense areas in the current acquisition point layout, and simulates different acquisition point layout schemes to optimize their quantity and position. In addition, by changing the signal source characteristics or GIS structure conditions for simulation, the influence of design improvements on signal propagation characteristics can be evaluated, providing a theoretical basis and guidance for GIS structure optimization and the rational arrangement of UHF detection schemes. The model performance is evaluated by the mean square error and positioning error, and its robustness is tested by introducing noise or structural changes to ensure the prediction reliability of complex signal propagation laws.

[0059] This embodiment uses Pro / E software to construct a geometric model and assigns material electrical parameters in the XFDTD simulation software. XFDTD is based on the finite-difference time-domain (FDTD) method and can achieve high-precision full-wave electromagnetic field simulation. The GIS simulation modeling process includes: Step S100a: Construct geometric models of each component according to the GIS structure size and ensure that the models accurately match the actual equipment structure. The main components include high-voltage conductors, pot-type insulators, and GIS shells. Step S100b: Assign corresponding material electrical parameters to each component to accurately simulate the electromagnetic characteristics inside the GIS.

[0060] During the simulation process, the partial discharge signal source needs to truly reflect the internal discharge phenomenon of the GIS. Therefore, the signal source settings are optimized, including selecting appropriate pulse types, time widths, and voltage amplitudes to ensure that the characteristics of the signal match the actual discharge situation.

[0061] The layout of data acquisition points is crucial for the accuracy of simulation results. Therefore, multiple data acquisition points are arranged in both the tangential and axial directions of the GIS shell, and a reasonable interval distance is maintained to ensure that the acquired data can fully reflect the spatial propagation characteristics of the signal.

[0062] After the simulation is completed, simulation software such as XFDTD can be used to analyze the time-domain waveform of the electric field intensity at the data acquisition points and compare the data at each acquisition point. Multiple observation points (250 cm, 500 cm, 750 cm, 1000 cm, 1250 cm, 1500 cm) are arranged at different axial positions of the GIS shell. By comparing the changes in the field strength modulus values, the signal attenuation characteristics inside the GIS are analyzed. By setting partial discharge signal sources with different parameters (such as Gaussian pulses with time widths of 1 ns, 2 ns, 3 ns, and 4 ns), the propagation laws of signals under different conditions can be studied.

[0063] To further optimize the GIS partial discharge detection scheme, the spectrum analysis method can also be combined. Using data processing tools such as Python, the spectrum data at the acquisition points are analyzed, the spectrum characteristics of the GIS UHF signals are extracted, and the cut-off frequency is calculated to optimize the selection and layout strategy of sensors.

[0064] To verify the effectiveness of the simulation method for the propagation characteristics of UHF electromagnetic wave signals of partial discharge in GIS provided in this embodiment, a simulation experiment was conducted on a GIS model of a straight cylinder with pot-type insulators, the propagation characteristics of partial discharge signals inside the GIS structure were analyzed, and the signal distribution at different measurement points was evaluated; Figure 6 The waveform of the magnitude of the electric field strength at a specific measurement point changing with time is shown, revealing the attenuation mode of partial discharge signals and the influence of the GIS shell on signal propagation.

[0065] Based on the data generated by the simulation, neural networks were used for training and prediction; due to the obvious shielding effect of the metal shell of GIS on UHF signals, multi-path reflection modes are formed when electromagnetic waves propagate inside, resulting in uneven attenuation of signal intensity; in this embodiment, a CNN-LSTM combined model was constructed, and by analyzing the spatial distribution of signals at data acquisition points, a mapping relationship between signal source parameters (position, amplitude) and acquisition point characteristics (magnitude of electric field strength, spectral characteristics) was established.

[0066] During the training process, the model focused on learning the characteristics of GIS partial discharge signals, including the multiple reflection attenuation of signals inside the shell, the influence of the dielectric layer on signal propagation, and the spectral distribution of UHF signals; referring to Figure 5 As shown, the confusion matrix generated according to the training results shows that the global prediction accuracy of the model is between 75% - 85%, and the diagonal elements (true positive ratio) are stable at 65% - 90%, indicating that the model can effectively capture the signal propagation law inside GIS. Especially when the amplitude of the signal source is high or the measurement point is close to the discharge source, the prediction accuracy can reach over 90%.

[0067] The analysis results of the model show that the GIS shell and insulation structure have an important influence on the propagation path of partial discharge signals, resulting in signal shielding or attenuation in some areas; based on this, using neural networks for reverse analysis can deduce the possible discharge positions of signal sources and optimize the layout of data acquisition points; the prediction results of the model help to adjust the GIS partial discharge detection scheme, improve the detection sensitivity, and provide a theoretical basis for the optimal placement position of UHF sensors.

[0068] In practical engineering applications, the value of this method lies in intelligently predicting the relationship between the signal source parameters and the signal characteristics at the acquisition points, thereby optimizing the layout of the data acquisition points and improving the capture efficiency of GIS partial discharge signals. In addition, this model can be applied to simulations under different GIS structure design conditions to evaluate the propagation characteristics of the signal source in a specific structure, thereby optimizing the detection strategy and sensor layout scheme of GIS equipment. Utilizing the fast inference ability of the neural network, this method can effectively improve the efficiency of analyzing signal propagation laws and provide theoretical support and engineering practice guidance for the partial discharge detection of GIS equipment.

[0069] In summary, the GIS partial discharge UHF electromagnetic wave signal propagation characteristic simulation method provided in this embodiment, through CNN-LSTM combined model neural network structure optimization, reinforcement learning to optimize simulation parameters, generative adversarial network to generate data augmentation, and adaptive signal processing, provides a precise analysis means for the UHF signal propagation of GIS partial discharge. At the same time, it provides a scientific basis for the partial discharge detection scheme of GIS equipment, improving the reliability and sensitivity of detection.

[0070] Embodiment 2:

[0071] Based on the same inventive concept as Embodiment 1, the present invention discloses a GIS partial discharge UHF electromagnetic wave signal propagation characteristic simulation system, including:

[0072] A model construction module, used for: constructing a GIS model according to the size of the GIS and the electrical parameters of its materials; adopting an adaptive grid division strategy to divide the grid of the GIS model, and setting different grid densities in different regions inside the GIS model;

[0073] A simulation setting module, used for: setting the partial discharge signal source parameters and data acquisition points in the GIS model according to the grid density, performing data acquisition to obtain a partial discharge signal data set; performing adaptive signal processing on the partial discharge signal data set to obtain a processed signal data set;

[0074] A result analysis module, used for: inputting the processed signal data set into the trained CNN-LSTM combined model: performing convolution operations on the processed signal data set through the CNN layer to obtain partial discharge signal characteristics; performing feature analysis using the LSTM layer according to the partial discharge signal characteristics to obtain the UHF propagation characteristics of GIS partial discharge.

[0075] Specifically,

[0076] The model construction module is used to construct a GIS model based on the dimensions of the GIS and the electrical parameters of the materials to ensure that the simulation model can accurately reflect the electromagnetic environment inside the GIS; it includes building a high-voltage conductor model, a pot-type insulator model, and a GIS housing model according to the dimensions of each component of the GIS respectively, and setting their material electrical parameters.

[0077] The simulation setting module is used to set the partial discharge signal source in the simulation environment and optimize the layout of the data acquisition points to comprehensively capture the signal propagation characteristics; it includes setting the type, time width, voltage amplitude of the pulse and the signal source position, and optimizing the parameters in combination with the shielding effect of the GIS metal housing and the internal reflection situation; this module uses the reinforcement learning method to enable the grid division to be dynamically adjusted according to the simulation requirements. In terms of the layout of the data acquisition points, multiple data acquisition points are arranged in the tangential and axial directions of the GIS housing respectively to ensure that the propagation path and energy attenuation characteristics of the UHF electromagnetic wave signal can be completely captured.

[0078] The result analysis module is used to analyze the simulation data and train the neural network model to optimize the analysis ability of the signal propagation characteristics; it includes an electric field intensity modulus analysis unit and a spectrum analysis unit, which are respectively used to analyze the change trend of the electric field intensity and its spectrum characteristics at the same data acquisition point under different signal source conditions; this module optimizes the layout schemes of different acquisition points in combination with the shielding effect of the GIS metal structure and the multipath propagation characteristics of the electromagnetic wave; at the same time, this module uses adaptive signal processing technologies such as wavelet transform and empirical mode decomposition to enhance the signal characteristics, remove low-frequency noise, and improve the data reliability.

[0079] In other embodiments, an application analysis module can also be set up to optimize the installation position of the UHF sensor according to the UHF propagation characteristics of the GIS partial discharge signal, ensuring that the partial discharge signal can be effectively captured in the on-site detection of the GIS. This module combines reinforcement learning to optimize the signal source parameters and the layout of the data acquisition points, enabling the model to quickly predict the signal distribution at each acquisition point inside the GIS and optimize the detection scheme, thereby improving the accuracy and anti-interference ability of signal detection.

[0080] The GIS partial discharge UHF electromagnetic wave signal propagation characteristic simulation system provided by this embodiment constructs a high-precision simulation model through the structural parameters and material characteristics of the GIS, and adopts a hierarchical adaptive grid division strategy, making the simulation results have both calculation efficiency and accuracy; aiming at the propagation characteristics of the partial discharge signal inside the GIS, it optimizes the setting method of the partial discharge signal source, and combines the layout of the data acquisition points in different directions to make the collected data more representative and analyzable; based on these data, a CNN-LSTM combined model is further constructed to automatically extract the signal propagation law and optimize the detection scheme.

[0081] For the specific function implementation of each of the above modules, refer to the relevant content in the method of Embodiment 1, which will not be elaborated here.

[0082] Embodiment 3:

[0083] This embodiment provides a computer-readable storage medium, on which computer instructions are stored. When the computer instructions are executed by a processor, the steps of the GIS partial discharge UHF electromagnetic wave signal propagation characteristic simulation method described in any one of Embodiment 1 are implemented.

[0084] Embodiment 4:

[0085] This embodiment provides a computer device, including:

[0086] A memory for storing computer instructions;

[0087] A processor for executing the computer instructions to implement the steps of the GIS partial discharge UHF electromagnetic wave signal propagation characteristic simulation method described in any one of the first aspect.

[0088] Embodiment 5:

[0089] This embodiment provides a computer program product, including computer instructions, characterized in that when the computer instructions are executed by a processor, the steps of the GIS partial discharge UHF electromagnetic wave signal propagation characteristic simulation method described in any one of Embodiment 1 are implemented.

[0090] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0091] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the specified functions in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0092] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to operate in a particular manner, such that the instructions stored in the computer-readable memory produce a manufacture including an instruction means that implements the functions specified in one or more of the processes and / or blocks Figure 1 of the process or processes and / or boxes Figure 1 of the box or boxes.

[0093] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, such that a series of operational steps are performed on the computer or other programmable apparatus to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in one or more of the processes and / or blocks Figure 1 of the process or processes and / or boxes Figure 1 of the box or boxes.

[0094] The embodiments of the present invention have been described above in conjunction with the accompanying drawings. However, the present invention is not limited to the above specific embodiments. The above specific embodiments are merely illustrative and not restrictive. Under the inspiration of the present invention, those of ordinary skill in the art can also make many forms without departing from the spirit of the present invention and the scope protected by the claims. All of these are within the protection scope of the present invention.

Claims

1. A method for simulating propagation characteristics of ultra-high frequency electromagnetic wave signals of partial discharge in GIS, characterized in that: include: Construct a GIS model based on the dimensions of the GIS and the electrical parameters of its materials; Adopting an adaptive grid division strategy to grid the GIS model, and setting different grid densities in different areas within the GIS model; According to the grid density, setting the local discharge signal source parameters and data collection points in the GIS model, performing data collection, and obtaining a local discharge signal data set; Performing adaptive signal processing on the partial discharge signal data set to obtain a processed signal data set; The processed signal data set is input into the trained CNN-LSTM combined model: the processed signal data set is convolved by the CNN layer to obtain the local discharge signal characteristics; according to the local discharge signal characteristics, the LSTM layer is used to perform feature analysis to obtain the GIS local discharge ultra-high frequency propagation characteristics.

2. The method for simulating propagation characteristics of UHF electromagnetic wave signals of GIS partial discharge according to claim 1 is characterized in that: The GIS model includes a high-voltage conductor model, a pot-type insulator model and a GIS shell model; the high-voltage conductor model is used to transmit electric energy and ensure the efficient flow of current in the GIS model; the pot-type insulator model is used to support and isolate the high-voltage conductor to prevent arcing and short circuit; the GIS shell model is used to seal the internal equipment of the GIS model.

3. The method for simulating propagation characteristics of UHF electromagnetic wave signals of GIS partial discharge according to claim 1 is characterized in that: The grid division is achieved by using reinforcement learning, and a dense grid is used in areas within a preset distance from a local discharge source and in areas where the electric field intensity changes beyond a threshold, and a sparse grid is used in other areas.

4. The method for simulating propagation characteristics of UHF electromagnetic wave signals of GIS partial discharge according to claim 1 is characterized in that: The local discharge signal source parameters include pulse type, time width, voltage amplitude and signal source position; the data collection points are set in the tangential and axial directions of the GIS shell; the GIS model uses a generative adversarial network to generate local discharge signals.

5. The method for simulating propagation characteristics of UHF electromagnetic wave signals of GIS partial discharge according to claim 1 is characterized in that: The adaptive signal processing includes wavelet transform and empirical mode decomposition, which enhances the detection accuracy of high-frequency components in the partial discharge signal data set and removes low-frequency noise.

6. The method for simulating propagation characteristics of UHF electromagnetic wave signals of GIS partial discharge according to claim 1 is characterized in that: The CNN-LSTM combination model adopts mean square error as the loss function and uses an adaptive gradient optimization method to update the weights of the neural network in the CNN-LSTM combination model.

7. A GIS partial discharge ultra-high frequency electromagnetic wave signal propagation characteristics simulation system, characterized in that: include: Model building module, used to: build GIS model according to the size of GIS and electrical parameters of its materials; Adopting an adaptive grid division strategy to grid the GIS model, and setting different grid densities in different areas within the GIS model; The simulation setting module is used to: set the local discharge signal source parameters and data collection points in the GIS model according to the grid density, perform data collection, and obtain a local discharge signal data set; perform adaptive signal processing on the local discharge signal data set to obtain a processed signal data set; The result analysis module is used to: input the processed signal data set into the trained CNN-LSTM combination model; perform convolution operation on the processed signal data set through the CNN layer to obtain local discharge signal characteristics; and perform feature analysis using the LSTM layer based on the local discharge signal characteristics to obtain the GIS local discharge ultra-high frequency propagation characteristics.

8. A computer-readable storage medium having computer instructions stored thereon, characterized in that: When the computer instruction is executed by the processor, the steps of the method for simulating the propagation characteristics of GIS partial discharge ultra-high frequency electromagnetic wave signals described in any one of claims 1-6 are implemented.

9. A computer device, characterized in that: include: Memory, for storing computer instructions; A processor is used to execute the computer instructions to implement the steps of the GIS partial discharge ultra-high frequency electromagnetic wave signal propagation characteristics simulation method described in any one of claims 1-6.

10. A computer program product comprising computer instructions, characterized in that: When the computer instructions are executed by the processor, the steps of the method for simulating the propagation characteristics of ultra-high frequency electromagnetic wave signals of partial discharge in GIS described in any one of claims 1-6 are implemented.

Images