A method, device and medium for electric field regulation of a GIS pot-type insulator

Abstract

This invention relates to the field of high-voltage equipment insulation technology, specifically providing a method, device, and medium for electric field control of GIS basin insulators. The method includes: constructing a geometric model of the target GIS basin insulator, defining the target region and target electric field distribution; constructing a potential prediction model based on a physical information neural network architecture, and setting the conductivity parameters of the coating sub-region as trainable parameters; inputting the coordinates of sampling points in the computational domain into the potential prediction model, and outputting the predicted potential values ​​of the sampling points; updating the potential prediction model and conductivity parameters through backpropagation by calculating the comprehensive loss function value, thereby achieving electric field control of the target region. This invention uses a physical information neural network as the solution method for coating conductivity, directly embedding physical laws into the neural network training process, achieving integration of forward solution and parameter inversion; and actively designing the electric field strength on the insulator surface by adjusting the coating conductivity.

Description

Technical Field

[0001] This invention relates to the field of high-voltage equipment insulation technology, and in particular to a method, device and medium for electric field control of GIS basin insulators. Background Technology

[0002] Epoxy resin insulators in gas-insulated switchgear (GIS) are key components ensuring the reliability of equipment insulation. As GIS equipment becomes more compact and operates at higher voltage levels, the electric field strength borne by the insulators is continuously increasing, leading to more severe charge accumulation at the gas-solid interface and a more prominent problem of uneven electric field distribution. Surface electric field distortion is a major cause of partial discharge and surface flashover, directly affecting the operational reliability of GIS and seriously threatening power grid safety. Therefore, regulating and optimizing the surface electric field of insulators is of great significance.

[0003] Insulator partitioning coating technology is an effective means of controlling the surface electric field. This technology reduces the surface electric field by coating different regions of the insulator surface with materials having different electrical conductivities. Electric field control techniques based on partitioned coatings are mainly divided into two categories: 1) Manual screening method. Several coating materials with significantly higher or lower conductivity than the epoxy resin matrix are coated on the insulator surface. By calculating or measuring their impact on the electric field, the conductivity parameters that can reduce the surface electric field are manually selected. 2) Electric field minimization method nested with finite element method and optimization method. First, using the coating conductivity as input, the electric field distribution is calculated using the finite element method or response surface method as a forward solver. Then, the coating conductivity is used as an optimization variable, and an objective function is constructed with the desired electric field objective (e.g., "minimize the maximum field strength" or "the field strength at a certain point reaches a certain value"). Optimization algorithms such as gradient descent, genetic algorithm, and particle swarm optimization are combined to find the optimal solution, and finally, the parameter that makes the electric field meet the target requirement is taken as the optimal solution, thus determining the coating conductivity value.

[0004] However, manual selection methods are inefficient, have limited optimization effects, and struggle to obtain optimal parameters. For electric field minimization methods that nest the finite element method and optimization methods, the solution process is computationally expensive, and large-scale finite element simulations result in extremely long computation times and low efficiency. Furthermore, this method can only search under a given simple objective function (such as "minimizing the maximum value"), and cannot determine the conductivity parameter that makes the electric field exhibit a desired distribution in a certain region, making it difficult to achieve proactive electric field design. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method, device and medium for electric field control of GIS basin insulators, so as to realize the active design of local electric field intensity on the surface of the insulator, thereby improving insulation performance and enhancing the reliability of GIS equipment.

[0006] To achieve the above objectives, the present invention is implemented using the following technical solution:

[0007] In a first aspect, embodiments of the present invention provide a method for controlling the electric field of a GIS basin-type insulator, comprising:

[0008] A geometric model of the target GIS basin insulator is constructed to determine the computational domain of the target GIS basin insulator and the coating sub-region within the computational domain; on the convex surface of the basin of the geometric model, the target region of the electric field to be controlled and the corresponding target electric field distribution are set.

[0009] Based on the physical information neural network architecture, a potential prediction model is constructed, and the conductivity parameters of the coating sub-region are set as trainable parameters.

[0010] The computational domain is sampled at multiple points, the coordinate data of the sampled points are input into the potential prediction model, and the potential prediction value of the sampled points is output.

[0011] Based on the predicted potential value, the conductivity parameter, the target region, and the target electric field distribution, a comprehensive loss function value is calculated; wherein, the comprehensive loss function value includes a physical residual loss term, a data matching loss term, a boundary loss term, and a regularization loss term;

[0012] The potential prediction model and the conductivity parameter are updated by backpropagation using the comprehensive loss function value, so as to achieve electric field control of the target region.

[0013] In some embodiments of the present invention, a geometric model of the target GIS basin insulator is constructed to determine the computational domain of the target GIS basin insulator and the coating sub-regions within the computational domain, including:

[0014] Based on the structure and dimensions of the target GIS basin insulator, a two-dimensional axisymmetric geometric model is constructed.

[0015] Based on the geometric model, the ranges of radial and axial coordinates are defined to determine the computational domain of the target GIS pot insulator and multiple sub-regions within the computational domain.

[0016] The sub-regions include: a high-voltage electrode sub-region, a grounding electrode sub-region, an insulator substrate sub-region, an insulating gas sub-region, and at least one coating sub-region; the coating sub-regions are defined by radial coordinate intervals set on the convex surface of the insulator substrate sub-regions.

[0017] In some embodiments of the present invention, multi-point sampling of the computational domain includes:

[0018] Uniform random sampling is performed within the computational domain to obtain multiple first sampling points;

[0019] Uniform sampling is performed within the target area to obtain multiple second sampling points;

[0020] Multiple third sampling points are obtained by sampling along the Chebyshev node distribution on the first boundary;

[0021] Multiple fourth sampling points were obtained by sampling along the second boundary according to the Chebyshev node distribution;

[0022] Wherein, the first boundary is the interface between the high-voltage electrode sub-region and the insulator substrate sub-region, and the interface between the high-voltage electrode sub-region and the insulating gas sub-region; the second boundary is the interface between the grounding electrode sub-region and the insulator substrate sub-region, and the interface between the grounding electrode sub-region and the insulating gas sub-region.

[0023] In some embodiments of the present invention, before calculating the comprehensive loss function value based on the predicted potential value, the conductivity parameter, the target region, and the target electric field distribution, the method further includes: constructing the conductivity distribution function of the calculation domain through a partitioning indicator function;

[0024] The expression for the conductivity distribution function in the computational domain is as follows:

[0025] ;

[0026] in, Indicates the number of coating sub-regions; Indicates the first Conductivity parameters of each coating sub-region The logarithm of ; Indicates the first Partitioning indicator function for each coating sub-region; This represents the conductivity distribution in a subregion of the insulator matrix; A partitioning indicator function representing a sub-region of the insulator matrix; This represents the conductivity distribution of the high-voltage electrode sub-region and the grounding electrode sub-region; A partitioning indicator function for the high-voltage electrode sub-region and the ground electrode sub-region.

[0027] In some embodiments of the present invention, a comprehensive loss function value is calculated based on the predicted potential value, the conductivity parameter, the target region, and the target electric field distribution, including:

[0028] Based on the predicted potential value of the first sampling point and the conductivity distribution function of the computational domain, physical constraints are constructed through the current continuity equation to obtain the physical residual loss term;

[0029] Based on the potential prediction value of the second sampling point, the electric field amplitude is calculated; data matching constraints are constructed according to the deviation between the electric field amplitude and the target electric field distribution to obtain the data matching loss term;

[0030] Based on the potential prediction values ​​of the third and fourth sampling points and the preset boundary conditions, boundary constraints are constructed to obtain the boundary loss term;

[0031] Based on the conductivity parameter, a regularization loss term is obtained through regularization constraints.

[0032] In some embodiments of the present invention, the calculation formula for the physical residual loss term is as follows:

[0033] ;

[0034] in, Indicates the number of the first sampling points; Indicates the first The radial coordinates of the first sampling point in the computational domain; Indicates the first The axial coordinates of the first sampling point in the computational domain; Represents coordinate points The conductivity value at that point; This represents the predicted potential distribution output by the potential prediction model. express First-order partial derivatives with respect to radial coordinates; express The first-order partial derivative with respect to the axial coordinates.

[0035] In some embodiments of the present invention, the calculation formula for the data matching loss term is as follows:

[0036] ;

[0037] in, Indicates the number of second sampling points; Indicates the first The arc length parameter of the second sampling point in the target area; This indicates the target electric field distribution.

[0038] In some embodiments of the present invention, the calculation formula for the boundary loss term is as follows:

[0039] ;

[0040] in, Indicates the number of third sampling points; Indicates the number of fourth sampling points; The potential prediction model indicates the first The third sampling point The predicted potential value; The potential prediction model indicates the first The fourth sampling point The predicted potential value;

[0041] The formula for calculating the regularization loss term is as follows:

[0042] ;

[0043] in, This represents the linear rectifier function.

[0044] Secondly, the present invention also provides an electronic device, including: a processor, and a memory storing a program, the program including instructions that, when executed by the processor, cause the processor to perform the above-described electric field control method for GIS basin insulators.

[0045] Thirdly, the present invention also provides a non-transient machine-readable medium storing computer instructions for causing the computer to execute the above-described electric field control method for GIS basin insulators.

[0046] Compared with the prior art, the above-described technical solution of the present invention has the following advantages:

[0047] In this embodiment of the invention, the proposed electric field control method for GIS basin insulators sets the conductivity parameter of the coated sub-region as a trainable variable and introduces a data matching loss term, enabling the electric field amplitude in the target region to accurately follow the preset target electric field distribution. Compared to related technologies that can only passively "shaving off peaks" to reduce the maximum electric field value, this invention can actively shape the electric field shape according to design requirements, achieving multi-point, on-demand precise control, providing an effective means for complex insulation designs such as electric field homogenization and electric stress control. Furthermore, based on a physical information neural network architecture, this invention directly embeds physical constraints such as the current continuity equation into the physical residual loss term, allowing for simultaneous optimization of network parameters and conductivity parameters through a single training iteration, avoiding repeated calls to external solvers. Under the same computing resources, this invention can reduce the computation time for such inversion problems from several days in traditional methods to several hours, achieving an efficiency improvement of one to two orders of magnitude. Attached Figure Description

[0048] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other embodiments based on these drawings without creative effort.

[0049] Figure 1 This is a flowchart illustrating an electric field control method for a GIS basin insulator provided in an embodiment of the present invention;

[0050] Figure 2 This is a schematic diagram of the geometric model of the GIS basin insulator according to an embodiment of the present invention;

[0051] Figure 3 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention.

[0052] The above figures include the following reference numerals:

[0053] 1—High-voltage electrode sub-region; 2—Insulating gas sub-region; 3—Insulator substrate sub-region; 4—Grounding electrode sub-region; 5—First coating sub-region; 6—Second coating sub-region. Detailed Implementation

[0054] Embodiments of the present invention will now be described in more detail with reference to the accompanying drawings. While some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the present invention. It should be understood that the drawings and embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0055] like Figure 1 As shown, this embodiment of the invention provides a method for controlling the electric field of a GIS basin insulator. Figure 1 This is a flowchart illustrating the electric field control method for GIS basin insulators. This flowchart only shows the logical sequence of the method in this embodiment. Provided there are no conflicts, other possible embodiments of the invention may use different methods. Figure 1 Complete the steps shown or described in the order indicated.

[0056] See Figure 1 The method of this invention specifically includes the following steps:

[0057] Step S101: Construct a geometric model of the target GIS basin insulator to determine the computational domain of the target GIS basin insulator and the coating sub-region within the computational domain; on the convex surface of the basin of the geometric model, set the target region of the electric field to be controlled and the corresponding target electric field distribution.

[0058] The GIS basin-type insulator is a typical two-dimensional axisymmetric insulation structure, mainly composed of five parts: a high-voltage electrode, a grounding electrode, an insulator substrate, an insulating gas domain, and coating zones. The high-voltage electrode is typically located at the center of the insulator substrate and is directly connected to the high-voltage conductor inside the GIS, while the grounding electrode is located at the outer edge of the insulator substrate. The high-voltage electrode and the grounding electrode are mechanically supported and electrically isolated by the basin-shaped insulator substrate, which is made of high-performance epoxy resin composite material, possessing high dielectric and mechanical strength. On the convex surface of the insulator substrate, between the high-voltage electrode and the grounding electrode, multiple annular coating zones are radially divided according to the electric field control requirements. Each coating zone is coated with a material with different conductivity to optimize the surface electric field distribution. The space surrounding the insulator substrate and electrodes is the insulating gas domain, typically filled with... Alternatively, environmentally friendly insulating gases can be used to provide a good gas insulation environment.

[0059] See Figure 2 A geometric model of the target GIS pot insulator is constructed to determine the computational domain of the target GIS pot insulator and the coating sub-regions within the computational domain, including:

[0060] Based on the structure and dimensions of the target GIS pot insulator, a two-dimensional axisymmetric geometric model is constructed. Based on the geometric model, the ranges of radial and axial coordinates are defined to determine the computational domain of the target GIS pot insulator and multiple sub-regions within the computational domain.

[0061] The computational domain is defined in a two-dimensional plane, and the coordinate system adopted is a cylindrical coordinate system. ,in, Radial coordinates, For axial coordinates. The computational domain is... , The computational domain includes multiple sub-regions in the geometric model, namely, high-voltage electrode sub-region 1, grounding electrode sub-region 4, insulating gas sub-region 2, insulator substrate sub-region 3, and at least one coating sub-region.

[0062] The coating sub-region is defined by a radial coordinate interval set on the convex surface of the insulator substrate sub-region. Each coating sub-region is a convex surface of an insulator substrate sub-region that satisfies... The remaining portion of the convex side of the insulator substrate sub-region is an uncoated area.

[0063] In this embodiment of the invention, the target GIS basin insulator is provided with two coating zones, corresponding to the first coating sub-region 5 and the second coating sub-region 6 in the geometric model.

[0064] On the convex surface of the insulator substrate, between the high-voltage electrode and the grounding electrode, a target region is defined. The target region is a specific arc-length interval of the electric field to be regulated on the insulator surface. Its spatial position is independent of the division of the coating sub-regions. The target region can completely coincide with one or more coating sub-regions, partially overlap with one or more coating sub-regions, or not coincide with any coating sub-region at all. It is entirely up to the designer to freely select according to the electric field regulation requirements.

[0065] In this embodiment of the invention, the radial range of the target region is: And mapped to arc length parameter ,in, .

[0066] The desired electric field distribution to be achieved in the target region is defined as the target electric field distribution. The target electric field distribution, as the expectation of active control, is a known function, denoted as . The target electric field distribution can be uniform or have a specific gradient, and its specific form is determined by the insulation design requirements.

[0067] Considering the continuous distribution of electric field intensity along the arc length, in actual calculations, this continuous distribution is discretized into a finite number of sampling points.

[0068] Step S102: Based on the physical information neural network architecture, construct a potential prediction model and set the conductivity parameters of the coating sub-region as trainable parameters.

[0069] Physics-Informed Neural Network (PINN) is a deep learning architecture that directly embeds physical laws into the neural network training process as constraints in the loss function. Its core idea is to make the network not only fit the data, but also satisfy the physical control equations described by partial differential equations.

[0070] In this embodiment of the invention, the potential prediction model is constructed using a deep fully connected neural network. The network structure includes: an input layer with two neurons, corresponding to the radial and axial coordinates of the sampling point; eight hidden layers, each with 128 neurons, using a hyperbolic tangent function to ensure the smoothness and boundedness of the output, satisfying the second-order differentiability requirement and facilitating stable gradient propagation; and an output layer with one neuron, outputting the predicted potential value of the sampling point. Weights are initialized using a Xavier normal distribution, and biases are initialized to zero.

[0071] The conductivity parameters of each coating sub-region are set as trainable parameters. To ensure that the conductivity parameters are positive and to compress the range of conductivity parameters spanning multiple orders of magnitude to a reasonable interval, the conductivity parameters are converted into logarithmic form. The logarithmic conductivity parameter of each coated sub-region is expressed as follows: ,in, Indicates the first Conductivity parameters of each coating sub-region.

[0072] Furthermore, this invention maps the logarithmic conductivity parameters corresponding to each coating sub-region, as well as the fixed conductivity constants of the insulator substrate and electrode regions, to every spatial coordinate point in the entire computational domain through a partitioning indicator function, thereby forming a spatially continuous conductivity distribution function. This conductivity distribution function forms the basis for subsequent calculations of the physical residual loss term. It combines material parameters with the potential gradient output by the potential prediction model, enabling the current continuity equation to apply physical constraints point-by-point within the computational domain. This allows for optimization of trainable parameters. The goal is to invert the coating conductivity and actively regulate the electric field distribution.

[0073] The expression for the conductivity distribution function is as follows:

[0074] ;

[0075] in, Indicates the number of coating sub-regions; Indicates the first Conductivity parameters of each coating sub-region The logarithm of ; Indicates the first Partitioning indicator function for each coating sub-region; This represents the conductivity distribution in a subregion of the insulator matrix; A partitioning indicator function representing a sub-region of the insulator matrix; This represents the conductivity distribution of the high-voltage electrode sub-region and the grounding electrode sub-region; A partitioning indicator function for the high-voltage electrode sub-region and the ground electrode sub-region.

[0076] The partition indicator function is a logical function used to determine which subregion any coordinate point within the computational domain belongs to. The input of the partition indicator function is a spatial coordinate point. If the input spatial coordinates point is located within the specified sub-region, output 1; otherwise, output 0.

[0077] Step S103: Perform multi-point sampling on the computational domain, input the coordinate data of the sampling points into the potential prediction model, and output the potential prediction value of the sampling points.

[0078] This includes multi-point sampling of the computational domain, including:

[0079] Uniform random sampling is performed within the computational domain to obtain multiple first sampling points; uniform sampling is performed within the target region to obtain multiple second sampling points; sampling is performed along the first boundary according to the Chebyshev node distribution to obtain multiple third sampling points; and sampling is performed along the second boundary according to the Chebyshev node distribution to obtain multiple fourth sampling points. The first boundary is the interface between the high-voltage electrode sub-region and the insulator substrate sub-region, and the interface between the high-voltage electrode sub-region and the insulating gas sub-region; the second boundary is the interface between the grounding electrode sub-region and the insulator substrate sub-region, and the interface between the grounding electrode sub-region and the insulating gas sub-region.

[0080] Chebyshev node distribution sampling is a non-uniform sampling strategy that densifies sampling at the endpoints of an interval and makes it relatively sparse in the middle. The location of its sampling points is determined by the zeros or extrema of the Chebyshev polynomial, for example, within an interval... Upward The second sampling, the first A Chebyshev node is defined as... + or + .

[0081] All the first, second, third, and fourth sampling points are merged to form the training sample set for the potential prediction model. The radial and axial coordinates of the sampling points in the training sample set are organized into a two-dimensional coordinate matrix and input into the potential prediction model. After nonlinear transformations and activation function mappings through multiple hidden layers, the potential prediction value corresponding to each sampling point is generated in the output layer.

[0082] Step S104: Calculate the comprehensive loss function value based on the predicted potential value, conductivity parameter, target area, and target electric field distribution.

[0083] The comprehensive loss function value includes physical residual loss, data matching loss, boundary loss, and regularization loss.

[0084] The calculation of the comprehensive loss function value specifically includes:

[0085] Step S1041: Based on the predicted potential value of the first sampling point and the conductivity distribution function of the computational domain, physical constraints are constructed through the current continuity equation to obtain the physical residual loss term.

[0086] The formula for calculating the physical residual loss term is as follows:

[0087] ;

[0088] in, Indicates the number of the first sampling points; Indicates the first The radial coordinates of the first sampling point in the computational domain; Indicates the first The axial coordinates of the first sampling point in the computational domain; Represents coordinate points The conductivity value at that point; This represents the predicted potential distribution output by the potential prediction model. express First-order partial derivatives with respect to radial coordinates; express The first-order partial derivative with respect to the axial coordinate. In electromagnetic field theory, the electric field strength... u, therefore, and The negative values ​​correspond to the radial and axial components of the electric field strength.

[0089] The physical residual loss term is calculated using the first sampling point, which is uniformly and randomly sampled throughout the entire computational domain, to compute the current continuity equation. The residual mean square of u). Its function is to directly embed the fundamental physical laws of electromagnetic fields into the training process of the potential prediction model, forcing the predicted potential value and the trainable conductivity distribution function to jointly satisfy the steady-state current conservation condition. By minimizing the physical residual loss, the potential distribution learned by the potential prediction model no longer depends solely on data fitting, but naturally conforms to physical laws, thus ensuring the physical rationality of the inversion results.

[0090] Step S1042: Calculate the electric field amplitude based on the potential prediction value of the second sampling point; construct data matching constraints based on the deviation between the electric field amplitude and the target electric field distribution to obtain the data matching loss term.

[0091] The formula for calculating the data matching loss term is as follows:

[0092] ;

[0093] in, Indicates the number of second sampling points; Indicates the first The arc length parameter of the second sampling point in the target area; This represents the target electric field distribution. The target region is a segment of a curve representing the electric field to be controlled on the insulator surface, with an arc length of... From the starting point ( ) to the destination ( A unique identifier for the position on the curve. (Through...) It can be determined that the first Spatial coordinates of the second sampling point and the corresponding target electric field value .

[0094] The data matching loss term treats the target electric field distribution as a hard constraint, driving the trainable conductivity parameter to optimize in the direction that can generate the target electric field distribution. This is the key difference between this invention and the traditional "passive peak clipping" method, enabling the local electric field to be shaped as needed.

[0095] Step S1043: Based on the potential prediction values ​​of the third and fourth sampling points and the preset boundary conditions, construct boundary constraints to obtain the boundary loss term.

[0096] The formula for calculating the boundary loss term is as follows:

[0097] ;

[0098] in, Indicates the number of third sampling points; Indicates the number of fourth sampling points; The potential prediction model indicates the first The third sampling point The predicted potential value; The potential prediction model indicates the first The fourth sampling point The predicted potential value.

[0099] The purpose of the boundary loss term is to constrain the potential prediction value output by the potential prediction model to satisfy the first type of boundary condition (Dirichlet condition) as much as possible: the surface potential of the high-voltage electrode is the rated voltage, and the surface potential of the ground electrode is zero.

[0100] In this embodiment of the invention, the rated voltage is kV.

[0101] Step S1044: Based on the conductivity parameter, the regularization loss term is obtained through regularization constraints.

[0102] The formula for calculating the regularization loss term is as follows:

[0103] ;

[0104] in, This represents the linear rectifier function.

[0105] The regularization loss term applies to the logarithmic conductivity parameter of the coating sub-region. A penalty term is constructed and constrained within a preset reasonable physical range to prevent the conductivity parameter from diverging to a physically meaningless value during the optimization process, thereby improving the stability of training and the engineering usability of the inversion results.

[0106] The formula for calculating the overall loss function value is as follows:

[0107] ;

[0108] in, , , , This represents the corresponding loss weight.

[0109] In this embodiment of the invention, the initial loss weight is: , , ; .

[0110] During training, the weights with larger losses are reduced, and the weights with smaller losses are increased. The loss weights are adjusted proportionally, and the corresponding adjustment strategy is as follows:

[0111] ;

[0112] in, express , , or ; Indicates the corresponding , , or .

[0113] Step S105: By integrating the loss function value, the potential prediction model and conductivity parameters are updated through backpropagation to achieve electric field control of the target area.

[0114] Based on the comprehensive loss function value, the adaptive moment estimation (Adam) optimizer is used to backpropagate and update the network parameters and conductivity parameters of the potential prediction model according to the preset learning rate:

[0115] ; ;

[0116] Set learning rate The initial value is 0.001, and it decays by a factor of 0.9 every 5000 steps. This is the first-order moment decay rate parameter of the Adam optimizer. Second-order moment decay rate Numerical stability constant Learning rate Set it to 0.01.

[0117] When the training converges, the obtained conductivity parameters This corresponds to the optimal coating conductivity value that enables active electric field design. At this point, the conductivity parameter will be obtained by inversion. The material preparation for each coating zone on the surface of GIS basin insulators enables the equipment to automatically present a preset electric field distribution during operation, thereby completing the closed-loop control from "desired electric field" to "material parameters" and then to "actual electric field".

[0118] An embodiment of the present invention also provides a non-transitory machine-readable medium storing a computer program, wherein the computer program, when executed by a computer processor, is used to cause the computer to perform the electric field control method for a GIS basin insulator according to an embodiment of the present invention.

[0119] An embodiment of the present invention also provides a computer program product, including a computer program, wherein the computer program, when executed by a computer processor, is used to cause the computer to perform the electric field control method for GIS basin insulators according to an embodiment of the present invention.

[0120] An embodiment of the present invention also provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor. The memory stores a computer program executable by the at least one processor, which, when executed by the at least one processor, causes the electronic device to perform the electric field control method for a GIS basin insulator according to an embodiment of the present invention.

[0121] refer to Figure 3 The present invention will now describe a structural block diagram of an electronic device that can serve as an embodiment of the present invention, serving as an example of a hardware device applicable to various aspects of the present invention. The electronic device is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0122] like Figure 3 As shown, the electronic device includes a computing unit 101, which can perform various appropriate actions and processes based on a computer program stored in a read-only memory (ROM) 102 or a computer program loaded from a storage unit 108 into a random access memory (RAM) 103. The RAM 103 may also store various programs and data required for the operation of the electronic device. The computing unit 101, ROM 102, and RAM 103 are interconnected via a bus 104. An input / output (I / O) interface 105 is also connected to the bus 104.

[0123] Multiple components in the electronic device are connected to I / O interface 105, including: input unit 106, output unit 107, storage unit 108, and communication unit 109. Input unit 106 can be any type of device capable of inputting information into the electronic device. Input unit 106 can receive input digital or character information and generate key signal inputs related to user settings and / or function control of the electronic device. Output unit 107 can be any type of device capable of presenting information and may include, but is not limited to, a display, speaker, video / audio output terminal, vibrator, and / or printer. Storage unit 108 may include, but is not limited to, disks and optical discs. Communication unit 109 allows the electronic device to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks, and may include, but is not limited to, modems, network cards, infrared communication devices, and / or wireless communication transceivers, such as Bluetooth devices, WiFi devices, WiMax devices, cellular communication devices, and / or the like.

[0124] The computing unit 101 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 101 include, but are not limited to, CPUs, graphics processing units (GPUs), various special-purpose artificial intelligence (AI) computing units, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. The computing unit 101 performs the various methods and processes described above. For example, in some embodiments, the method embodiments of the present invention can be implemented as computer programs tangibly contained in a machine-readable medium, such as storage unit 108. In some embodiments, part or all of the computer program can be loaded and / or installed on an electronic device via ROM 102 and / or communication unit 109. In some embodiments, the computing unit 101 can be configured to perform the methods described above by any other suitable means (e.g., by means of firmware).

[0125] Computer programs for implementing the methods of embodiments of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0126] In the context of embodiments of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable signal medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, or infrared systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0127] It should be noted that the term "comprising" and its variations used in the embodiments of this invention are open-ended, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The modifications of "one" and "a plurality" mentioned in the embodiments of this invention are illustrative and not restrictive, and those skilled in the art should understand that unless explicitly indicated otherwise in the context, they should be understood as "one or more".

[0128] The user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the embodiments of this invention are subject to strict compliance with relevant laws, regulations, and regulatory requirements in their collection, storage, use, processing, transmission, provision, and disclosure, and adhere to the principles of legality, legitimacy, necessity, and good faith. The acquisition of relevant information and data is premised on the user's explicit consent or other legitimate reasons, and a clear and convenient authorization management approach is provided to the user, allowing the user to independently choose to consent, withdraw consent, or refuse to provide relevant information. For functions that rely on user information, if the user does not authorize or withdraws authorization, the corresponding technical function cannot be implemented, and the technical solution of this invention is not applicable in this scenario.

[0129] The steps described in the method embodiments provided by the present invention can be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of protection of the present invention is not limited in this respect.

[0130] The term "embodiment" in this specification refers to a specific feature, structure, or characteristic described in connection with an embodiment that may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily imply the same embodiment, nor does it imply independence or alternativeity from other embodiments. The various embodiments in this specification are described in a related manner, with reference to each other for similar or identical parts. In particular, for apparatus, device, and system embodiments, since they are substantially similar to method embodiments, the description is relatively simple, and relevant details are referred to in the description of the method embodiments.

[0131] The above embodiments merely illustrate several implementation methods of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of protection. It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept of the present invention, and these all fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the appended claims.

Claims

1. A method for electric field regulation of a GIS pot insulator, characterized in that, include: Construct a geometric model of the target GIS pot insulator to determine the computational domain of the target GIS pot insulator and the coating sub-regions within the computational domain; On the convex surface of the basin of the geometric model, the target region for electric field control and the corresponding target electric field distribution are defined. Based on the physical information neural network architecture, a potential prediction model is constructed, and the conductivity parameters of the coating sub-region are set as trainable parameters. The computational domain is sampled at multiple points, the coordinate data of the sampled points are input into the potential prediction model, and the potential prediction value of the sampled points is output. Based on the predicted potential value, the conductivity parameter, the target region, and the target electric field distribution, a comprehensive loss function value is calculated; wherein, the comprehensive loss function value includes a physical residual loss term, a data matching loss term, a boundary loss term, and a regularization loss term; The potential prediction model and the conductivity parameter are updated by backpropagation using the comprehensive loss function value, so as to achieve electric field control of the target region.

2. The electric field control method for GIS basin-type insulators according to claim 1, characterized in that, Construct a geometric model of the target GIS pot insulator to determine the computational domain of the target GIS pot insulator and the coating sub-regions within the computational domain, including: Based on the structure and dimensions of the target GIS basin insulator, a two-dimensional axisymmetric geometric model is constructed. Based on the geometric model, the ranges of radial and axial coordinates are defined to determine the computational domain of the target GIS pot insulator and multiple sub-regions within the computational domain. The sub-regions include: a high-voltage electrode sub-region, a grounding electrode sub-region, an insulator substrate sub-region, an insulating gas sub-region, and at least one coating sub-region; the coating sub-regions are defined by radial coordinate intervals set on the convex surface of the insulator substrate sub-regions.

3. The electric field control method for GIS basin-type insulators according to claim 2, characterized in that, Multi-point sampling of the computational domain includes: Uniform random sampling is performed within the computational domain to obtain multiple first sampling points; Uniform sampling is performed within the target area to obtain multiple second sampling points; Multiple third sampling points are obtained by sampling along the Chebyshev node distribution on the first boundary; Multiple fourth sampling points were obtained by sampling along the second boundary according to the Chebyshev node distribution; Wherein, the first boundary is the interface between the high-voltage electrode sub-region and the insulator substrate sub-region, and the interface between the high-voltage electrode sub-region and the insulating gas sub-region; the second boundary is the interface between the grounding electrode sub-region and the insulator substrate sub-region, and the interface between the grounding electrode sub-region and the insulating gas sub-region.

4. The electric field control method for GIS basin-type insulators according to claim 3, characterized in that, Before calculating the comprehensive loss function value based on the predicted potential value, the conductivity parameter, the target region, and the target electric field distribution, the method further includes: constructing the conductivity distribution function of the calculation domain through a partitioning indicator function; The expression for the conductivity distribution function in the computational domain is as follows: ； in, Indicates the number of coating sub-regions; Indicates the first Conductivity parameters of each coating sub-region The logarithm of ; Indicates the first Partitioning indicator function for each coating sub-region; This represents the conductivity distribution in a subregion of the insulator matrix; A partitioning indicator function representing a sub-region of the insulator matrix; This represents the conductivity distribution of the high-voltage electrode sub-region and the grounding electrode sub-region; A partitioning indicator function for the high-voltage electrode sub-region and the ground electrode sub-region.

5. The electric field control method for a GIS basin-type insulator according to claim 3, characterized in that, Based on the predicted potential value, the conductivity parameter, the target region, and the target electric field distribution, the comprehensive loss function value is calculated, including: Based on the predicted potential value of the first sampling point and the conductivity distribution function of the computational domain, physical constraints are constructed through the current continuity equation to obtain the physical residual loss term; Based on the potential prediction value of the second sampling point, the electric field amplitude is calculated; data matching constraints are constructed according to the deviation between the electric field amplitude and the target electric field distribution to obtain the data matching loss term; Based on the potential prediction values ​​of the third and fourth sampling points and the preset boundary conditions, boundary constraints are constructed to obtain the boundary loss term; Based on the conductivity parameter, a regularization loss term is obtained through regularization constraints.

6. The electric field control method for a GIS basin-type insulator according to claim 5, characterized in that, The formula for calculating the physical residual loss term is as follows: ； in, Indicates the number of the first sampling points; Indicates the first The radial coordinates of the first sampling point in the computational domain; Indicates the first The axial coordinates of the first sampling point in the computational domain; Represents coordinate points The conductivity value at that point; This represents the predicted potential distribution output by the potential prediction model. express First-order partial derivatives with respect to radial coordinates; express The first-order partial derivative with respect to the axial coordinates.

7. The electric field control method for a GIS basin-type insulator according to claim 6, characterized in that, The formula for calculating the data matching loss term is as follows: ； in, Indicates the number of second sampling points; Indicates the first The arc length parameter of the second sampling point in the target area; This indicates the target electric field distribution.

8. The electric field control method for a GIS basin-type insulator according to claim 6, characterized in that, The formula for calculating the boundary loss term is as follows: ； in, Indicates the number of third sampling points; Indicates the number of fourth sampling points; The potential prediction model indicates the first The third sampling point The predicted potential value; The potential prediction model indicates the first The fourth sampling point The predicted potential value; The formula for calculating the regularization loss term is as follows: ； in, This represents the linear rectifier function.

9. An electronic device, comprising: A processor and a memory storing a program, characterized in that the program includes instructions that, when executed by the processor, cause the processor to perform the electric field control method for a GIS basin insulator according to any one of claims 1 to 8.

10. A non-transitory machine-readable medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to execute the electric field control method for GIS basin insulators according to any one of claims 1 to 8.

Images