A gas-insulated transformer bushing and its optimization method

Abstract

This invention belongs to the field of power equipment insulation structure optimization technology, and discloses a gas-insulated transformer bushing and its optimization method. The gas-insulated transformer bushing includes a bushing cone and an extension for mounting a current transformer. The bushing cone and the extension are connected by a metal flange. A shielding ring is connected to the portion of the metal flange inside the bushing. The shielding ring includes a horizontal section, a vertical section, and an arc section. The horizontal section is perpendicular to the axis of the bushing cone and extends horizontally from the metal flange into the bushing cone. The junction of the metal flange and the bushing cone is flush with the bottom edge of the horizontal section. The vertical section extends vertically downward from the end of the horizontal section to the beginning of the arc section. The arc section starts from the end of the vertical section, bends outward from the bushing cone, and is entirely below the metal flange. This invention provides electric field shielding by setting a shielding structure at the three connection points, which can reduce the risk of partial discharge.

Description

Technical Field

[0001] This invention belongs to the field of power equipment insulation structure optimization technology, specifically relating to a gas-insulated transformer bushing and its optimization method. Background Technology

[0002] Bushings in transformers provide mechanical support and electrical insulation for energized conductors passing through the grounding box or insulating medium, while ensuring equipment sealing and withstanding electrical, thermal, and mechanical stresses during operation. Compared to oil-filled bushings, gas-insulated bushings (i.e., gas-insulated transformer bushings) offer the core advantage of using a high-insulation-strength, non-flammable, and stable compressed gas (such as SF6) as the primary insulating medium, achieving maintenance-free operation, high fire safety, and excellent shock and pollution resistance. However, due to the use of gas insulation, SF6 has a relatively low permittivity, resulting in a significantly concentrated electric field on the outer surface compared to traditional oil-paper bushings. In particular, the bushing cone 101 of this gas-insulated transformer is made of epoxy-glass fiber composite material. At the bushing cone 101, there is a three-way connection point: metal flange 201 (used to connect the extension of the current transformer to the bushing cone 101), transformer oil, and the bushing cone 101. At this point, the electric field strength undergoes abrupt changes, the geometry is sharp, and the electric field is highly concentrated, posing a risk of partial discharge to the external insulation. Summary of the Invention

[0003] To address the problems existing in the prior art, the present invention aims to provide a gas-insulated transformer bushing and its optimization method. The present invention provides a shielding structure at the three junction points of the gas-insulated transformer bushing's metal grounding flange, transformer oil, and bushing cone to shield the electric field at the three junction points, thereby reducing the risk of partial discharge.

[0004] To achieve the above objectives, the present invention adopts the following technical solution: A gas-insulated transformer bushing includes a bushing cone and an extension for mounting a current transformer. The bushing cone and the extension are connected by a metal flange. A shielding ring is connected to the portion of the metal flange inside the bushing. The shielding ring includes a horizontal section, a vertical section, and an arc section. The horizontal section is perpendicular to the axis of the bushing cone and extends horizontally into the bushing cone from the position of the metal flange. The junction of the metal flange and the bushing cone is flush with the bottom edge of the horizontal section. The vertical section extends vertically downward from the end of the horizontal section to the beginning of the arc section. The arc section starts from the end of the vertical section, bends outward toward the outside of the bushing cone, and the entire arc section is located below the metal flange.

[0005] Preferably, the length of the horizontal segment is 16mm to 20mm, the length of the vertical segment is 50mm to 80mm, and the radius of the arc segment is 8mm to 12mm.

[0006] Preferably, the arc segment adopts a single arc structure with the same curvature, or the arc segment is formed by a smooth transition between at least two arc structures with different curvatures and / or different central angles.

[0007] Preferably, the central angle of each arc segment in the arc segment ranges from 0° to 90°.

[0008] Preferably, the material of the sleeve cone is a composite material made of epoxy and glass fiber.

[0009] The present invention also provides an optimization method for a gas-insulated transformer bushing as described above, comprising the following steps: Establish a finite element analysis model of the electric field of the gas-insulated transformer bushing; The optimization objective and the range of the shielding ring's geometric parameters are set. The optimization objective is that when the core potential of the gas-insulated transformer bushing is a preset voltage and the potential of the shielding ring is 0V, the maximum value of the surface electric field of the shielding ring obtained by simulation does not exceed a preset value. Within the range of geometric parameters of the shielding ring, the electric field finite element analysis model is coarsely meshed and the maximum electric field intensity on the surface of the shielding ring is calculated to obtain a low-fidelity data set; the electric field finite element analysis model is finely meshed and the maximum electric field intensity on the surface of the shielding ring is calculated to obtain a high-fidelity data set. By using the low-fidelity data set as the source domain data and the high-fidelity data set as the target domain data for the transfer learning model, a surrogate model is constructed to predict the maximum electric field strength on the surface of the shielding ring under different combinations of geometric parameters. The surrogate model is optimized using a genetic algorithm to predict the combination of geometric parameters that minimizes the electric field on the surface of the shielding ring, thereby achieving optimization of the bushing of the gas-insulated transformer.

[0010] Preferably, the geometric parameters of the shielding ring include the horizontal extension length of the horizontal segment, the vertical insertion depth of the vertical segment, the radius of the arc segment, and the arc angle.

[0011] Preferably, the electric field finite element analysis model is established in Comsol software.

[0012] Preferably, within the range of geometric parameters of the shielding ring, the maximum electric field intensity on the surface of the shielding ring is calculated by parametric scanning in the Comsol software using the normal option in the mesh coarse setting, resulting in a low-fidelity data set; and the maximum electric field intensity on the surface of the shielding ring is calculated by parametric scanning in the Comsol software using the ultra-fine option in the mesh coarse setting, resulting in a high-fidelity data set.

[0013] Preferably, the low-fidelity data set is used as the source domain data of the transfer learning model, and the high-fidelity data set is used as the target domain data of the transfer learning model to construct a surrogate model for predicting the maximum electric field strength on the surface of the shielding ring under different combinations of geometric parameters. Specifically, this includes: The surrogate model is constructed by importing the low-fidelity dataset into the source domain of the Matlab transfer learning program and the high-fidelity dataset into the target domain of the Matlab transfer learning program.

[0014] The present invention has the following beneficial effects: In this field, the present invention addresses for the first time the problem of abrupt changes in electric field intensity and sharp geometric shapes leading to high electric field concentration and susceptibility to partial discharge of the external insulation at the three-way junction of the bushing conical tube made of epoxy-glass fiber composite material, metal flange, and transformer oil. The invention proposes for the first time to install a shielding ring at this junction to resolve the aforementioned risks. To this end, the present invention connects a shielding ring with a specific structure to the portion of the metal flange located inside the bushing in the gas-insulated transformer bushing. This provides targeted electric field shielding and optimizes the electric field distribution at the three-way junction. Specifically, the shielding ring includes a horizontal section, a vertical section, and an arc section. The horizontal section is perpendicular to the axis of the bushing conical tube, extends horizontally from the metal flange into the bushing conical tube, and its bottom edge is flush with the junction line between the metal flange and the bushing conical tube. Therefore, it can cover the electric field concentration area at the three-way junction, effectively buffering abrupt changes in electric field intensity, weakening the electric field concentration phenomenon, and fundamentally reducing the risk of partial discharge of the external insulation. The vertical section extends vertically downwards from the end of the horizontal section to the beginning of the arc section, further enclosing the area around the three junction points, widening the shielding range and reducing electric field leakage. The arc section bends outwards from the end of the vertical section toward the outside of the bushing cone and is entirely below the metal flange. This avoids the shielding ring itself forming sharp corners that would cause electric field concentration, and also guides the internal electric field to be evenly distributed. In addition, this shielding ring structure is highly compatible with the existing bushing structure, without the need for significant modifications to the original structure. While improving the bushing's pressure resistance and operational stability, it also ensures structural simplicity and ease of installation.

[0015] This invention discloses an optimization method for gas-insulated transformer bushings. It establishes a finite element analysis model of the electric field of the gas-insulated transformer bushing, clarifies the optimization objective and geometric parameter range of the shielding ring, and then divides the model into coarse and fine meshes to obtain low-fidelity and high-fidelity data. The low-fidelity data is used as the source domain data for a transfer learning model, and the high-fidelity data is used as the target domain data to construct a surrogate model. Finally, a genetic algorithm is used to optimize the surrogate model, quickly and accurately determining the geometric parameter combination that minimizes the electric field on the surface of the shielding ring. This effectively achieves the goal of optimizing the surface field strength of the shielding ring, ensuring that the optimized shielding ring meets the preset field strength requirements and reducing the risk of discharge faults under extreme operating conditions. The technical solution of this invention uses low-fidelity data to provide global characteristics and high-fidelity data to correct accuracy. The transfer learning algorithm significantly reduces the construction time cost of the high-precision surrogate model, and the genetic algorithm enables rapid optimization of the target. This effectively solves the core technical problems in insulation structure optimization, where traditional surrogate model construction and algorithm optimization methods rely on high-fidelity model statistical data, resulting in high computational costs and lengthy processing times. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the structure of the gas-insulated transformer bushing in an embodiment of the present invention.

[0017] Figure 2 This is a schematic diagram of the installation positions of the sleeve tapered tube connection, metal flange, and shielding ring in an embodiment of the present invention.

[0018] Figure 3 This is a schematic diagram illustrating the iteration of a genetic algorithm based on the surrogate model obtained through transfer learning in an embodiment of the present invention.

[0019] Figure 4 This is a comparison diagram of the voltage equalization characteristics before and after the optimization of the shielding ring structure in the embodiments of the present invention.

[0020] Figure 5 This is a flowchart of the optimization method for gas-insulated transformer bushings in an embodiment of the present invention.

[0021] Among them, 101 is the bushing cone, 102 is the extension, 103 is the top cover of the riser, 104 is the metal grounding flange, 105 is the double-layer shield in the middle section of the bushing, 106 is the external insulating skirt, and 107 is the center conductor. 201 is a metal flange, 202 is a shielding ring, 2021 is a horizontal section, 2022 is a vertical section, and 2023 is a circular arc section. Detailed Implementation

[0022] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0023] See Figure 1 and Figure 2 The gas-insulated transformer bushing of this invention differs from the traditional oil-paper bushing structure in that the gas bushing requires a shielding ring at the conical connection to shield the electric field at the three-way junction. The structural design of the shielding ring directly determines the withstand voltage performance of the gas bushing. While shielding the external electric field of the bushing, the grounded and inward-extending shielding ring leads to an uneven internal electric field. If the shielding ring structure is unreasonable, the interface between the shielding ring and the insulating gas still has a risk of discharge when subjected to high field strength. Therefore, a more reasonable shielding ring structure design can ensure the stable operation of the gas bushing. To this end, this invention also provides an optimization method for the gas-insulated transformer bushing. The main optimization direction of this method is the optimization of the geometric parameters of the shielding ring.

[0024] Specifically, the gas-insulated transformer bushing of the present invention includes a bushing cone 101 and an extension 102 for mounting a current transformer. The bushing cone 101 is made of epoxy-glass fiber composite material. The bushing cone 101 and the extension 102 are connected by a metal flange 201. A shielding ring 202 is connected to the portion of the metal flange 201 located inside the bushing. The shielding ring 202 includes a horizontal section 2021, a vertical section 2022, and an arc section 2023. The horizontal section 2021 is connected to the bushing cone 101. The axis is vertical. The horizontal segment 2021 starts from the position of the metal flange 201 and extends horizontally into the inside of the sleeve cone 101. The intersection of the metal flange 201 and the sleeve cone 101 is flush with the bottom edge of the horizontal segment 2021. The vertical segment 2022 starts from the end of the horizontal segment 2021 and extends vertically downward to the beginning of the arc segment 2023. The arc segment 2023 starts from the end of the vertical segment 2022 and bends towards the outside of the sleeve cone 101. The entire arc segment 2023 is located below the metal flange 201.

[0025] In the above-described scheme of the present invention, the arc segment 2023 can be a single arc structure with the same curvature; or the arc segment 2023 can be formed by smoothly connecting at least two arc structures with different curvatures and / or different central angles. In this case, the optimized geometric parameters are more extensive in the subsequent optimization configuration, which makes the optimization result of the arc segment 2023 more accurate.

[0026] Since optimizing the insulation structure is a crucial means of improving equipment insulation performance, the field has long relied on a surrogate model-algorithm optimization approach to find the optimal values ​​of structural parameters within the allowable range. However, constructing a high-precision surrogate model requires establishing an input-output mapping using statistical data from a high-fidelity model, resulting in high computational time costs. Therefore, a method is needed to reduce computational time costs by providing global characteristics with low-fidelity data and correcting accuracy with high-fidelity data. To this end, this invention also provides the optimization method for gas-insulated transformer bushings as described above. This method uses a genetic algorithm to optimize the surrogate model constructed through transfer learning, which can quickly predict the optimal structure with the lowest electric field strength on the surface of the shielding ring under lightning impulse voltage, thereby determining the optimal parameter selection. The optimization method of this invention solves the problems of significant external insulation hazards at the junction of the gas bushing cone-flange-transformer oil in existing structures and the high computational time cost of high-fidelity data during insulation structure optimization.

[0027] For details, see Figure 5 The optimization method for gas-insulated transformer bushings of the present invention includes the following process: In Comsol software, an electric field finite element analysis model of the gas-insulated transformer bushing as described above is established.

[0028] The optimization objective and the range of geometric parameters of the shielding ring 202 are set. The optimization objective is that when the core potential of the gas-insulated transformer bushing is a preset voltage and the potential of the shielding ring 202 is 0V, the maximum value of the surface electric field of the shielding ring 202 obtained by simulation does not exceed the preset value. The geometric parameters of the shielding ring 202 include the horizontal extension length of the horizontal segment 2021, the vertical insertion depth of the vertical segment 2022, the radius of the arc segment 2023, and the arc angle. When the arc segment 2023 is formed by the smooth transition connection of at least two arc structures, the radius of the arc and the arc angle in the geometric parameters of the shielding ring 202 include the radius of the arc and the arc angle of each arc structure.

[0029] In Comsol software, within the range of various geometric parameters of the shielding ring 202, the electric field finite element analysis model is coarsely meshed (specifically, the normal option in the mesh coarseness settings can be selected) and the maximum electric field intensity on the surface of the shielding ring 202 is calculated to obtain a low-fidelity data set; the electric field finite element analysis model is finely meshed (specifically, the ultra-fine option in the mesh coarseness settings can be selected) and the maximum electric field intensity on the surface of the shielding ring 202 is calculated to obtain a high-fidelity data set; By using the low-fidelity dataset as the source domain data of the transfer learning model (i.e., importing the low-fidelity dataset into the source domain of the Matlab transfer learning program) and the high-fidelity dataset as the target domain data of the transfer learning model (i.e., importing the high-fidelity dataset into the target domain of the Matlab transfer learning program), a surrogate model is constructed to predict the maximum electric field strength on the surface of the shielding ring 202 under different combinations of geometric parameters. The surrogate model is optimized using a genetic algorithm to predict the combination of geometric parameters that minimizes the surface electric field of the shielding ring 202, thereby achieving optimization of the gas-insulated transformer bushing.

[0030] Example The gas-insulated transformer bushing and its optimization method in this embodiment include the following steps: 1) Establish a finite element analysis model of the electric field of the gas-insulated transformer bushing in Comsol software. The shielding ring 202 is installed on the bushing for mounting the extension 102 of the current transformer and the epoxy resin. The location of the metal flange 201 between the glass fiber composite sleeve cone 101. The shielding ring 202 includes a horizontal section 2021, a vertical section 2022, and an arc section 2023. The horizontal section 2021 extends horizontally from the metal flange 201 into the sleeve, and its bottom edge is flush with the junction of the metal flange 201 and the sleeve cone 101. The vertical section 2022 extends vertically downward from the end of the horizontal section 2021 to the beginning of the arc section 2023. The arc section 2023 starts from the end of the vertical section 2022. In this embodiment, the arc section 2023 is formed by two smoothly connected arc segments and bends towards the outside of the sleeve. The entire arc section 2023 is located below the metal flange 201. Figure 2 As shown, the geometric parameters to be optimized in this embodiment include the horizontal extension length L1 of the horizontal segment 2021, the vertical penetration depth L2 of the vertical segment 2022, the first arc radius R1 and the first arc angle θ1 of the arc segment 2023, the second arc radius R2 and the second arc angle θ2 of the arc segment 2023.

[0031] 2) The optimization objectives for the shielding ring at the conical sleeve are as follows: when the core potential is 1800 kV and the shielding ring potential is 0 V, the maximum surface electric field strength of the shielding ring obtained from simulation should not exceed 24 kV / mm. The parameter variation range is as follows: the horizontal extension length L1 is 16 mm to 20 mm, the vertical encasing depth L2 is 50 mm to 80 mm, the radius R1 of the first arc segment at the bottom of the shielding ring is 8 mm to 12 mm, and the angle θ1 of the first arc segment is 0° to 90°; the radius R2 of the second arc segment at the bottom of the shielding ring is 8 mm to 12 mm, and the angle θ2 of the second arc segment is 0° to 90°.

[0032] 3) Within the above parameter variation range, in the Comsol software, select the "Normal" option in the mesh coarseness settings, and calculate the maximum electric field intensity on the shielding ring surface using parametric scanning. This yields 7000 sets of low-fidelity data, denoted as the low-fidelity data set. Select the "Extremely Fine" option in the mesh coarseness settings, and calculate 1000 sets of high-fidelity data using parametric scanning, denoted as the high-fidelity data set. Import the low-fidelity data set into the source domain of the Matlab transfer learning program, and import the high-fidelity data set into the target domain of the Matlab transfer learning program to construct a surrogate model that can predict the maximum electric field intensity on the shielding ring surface under different combinations of structural parameters.

[0033] Transfer learning is a machine learning method that optimizes model performance by reusing existing knowledge. It involves transferring data features learned from a source domain and task, which have relatively low computational difficulty and time cost, to a target domain and task with limited data and higher computational difficulty. Source domain knowledge corresponds to the basic data and general features already computed, while the target domain corresponds to the complex real-world scenario and optimization objective; together, they constitute the core system of transfer learning. With continuous optimization of transfer strategies, the learning effect and training efficiency of the target task can be significantly improved. By selecting the optimal transfer method and model structure based on set performance metrics, it provides an efficient and feasible technical solution for model optimization in small-sample scenarios.

[0034] Genetic algorithms are intelligent optimization algorithms that simulate biological evolution and natural selection. By encoding and evaluating the fitness of individuals in a population, and through iterative evolution via operations such as selection, crossover, and mutation, individuals with higher fitness are selected generation by generation, ultimately approaching the optimal solution.

[0035] 4) Import the surrogate model built based on transfer learning into the genetic algorithm program, such as... Figure 3 As shown, the optimal parameter combination that minimizes the electric field on the surface of the shielding ring is obtained after 100 iterations through optimization calculations using a genetic algorithm. Specifically, the parameters are: L1=15.7mm, L2=52.3mm, R1=11.4mm, R2=8.6mm, θ1=45°, and θ2=112°.

[0036] In this embodiment, the horizontal extension length, vertical insertion depth, radius of curvature, and angle of the shielding ring at the sleeve cone can be adjusted adaptively according to actual engineering needs, and this invention does not impose specific limitations.

[0037] This embodiment employs transfer learning to construct a proxy model, enabling the computation of the training dataset at a low time cost, and combines it with a genetic algorithm to achieve global optimization. The shielding ring designed in this invention can effectively reduce the maximum electric field value on the surface of the shielding ring and the electric field at the three junction points of the shielding cone and flange connection, reducing the probability of discharge faults at that point and contributing to the long-term safe operation of gas-insulated transformer bushings.

[0038] The transfer learning method used in this embodiment to build agent models differs from the general data-driven agent model building method, and can save computation time by predicting high-fidelity data.

[0039] After testing and verification, such as Figure 4 As shown, under the condition of a core potential of 1800 kV and a grounded shielding ring, the maximum electric field strength on the surface of the shielding ring decreased from 26.6469 kV / mm before optimization to 22.4248 kV / mm, a reduction of 16%; the electric field strength at the triple junction point was 6.8 kV / mm, which is within the safety margin range. Meanwhile, the overall optimization calculation time was shortened from approximately 12 hours to approximately 5 hours, a reduction of approximately 58.3%. This demonstrates a significant reduction in computational time costs while exhibiting a clear electric field optimization effect.

[0040] Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort should fall within the scope of protection of the present invention.

[0041] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A gas-insulated transformer bushing, characterized in that, The device includes a bushing cone (101) and an extension (102) for mounting a current transformer. The bushing cone (101) and the extension (102) are connected by a metal flange (201). A shielding ring (202) is connected to the portion of the metal flange (201) located inside the bushing. The shielding ring (202) includes a horizontal section (2021), a vertical section (2022), and an arc section (2023). The horizontal section (2021) is perpendicular to the axis of the bushing cone (101). The horizontal section (2021) extends from the metal flange (1022)... The 201) position starts to extend horizontally into the inside of the sleeve cone (101), the intersection of the metal flange (201) and the sleeve cone (101) is flush with the bottom edge of the horizontal section (2021), the vertical section (2022) extends vertically downward from the end of the horizontal section (2021) and extends to the beginning of the arc section (2023), the arc section (2023) starts from the end of the vertical section (2022) and bends towards the outside of the sleeve cone (101), and the entire arc section (2023) is located below the metal flange (201).

2. The gas-insulated transformer bushing according to claim 1, characterized in that, The horizontal segment (2021) has a length of 16mm to 20mm, the vertical segment (2022) has a length of 50mm to 80mm, and the arc segment (2023) has a radius of 8mm to 12mm.

3. A gas-insulated transformer bushing according to claim 1 or 2, characterized in that, The arc segment (2023) adopts a single arc structure with the same curvature, or the arc segment (2023) is formed by a smooth transition between at least two arc structures with different curvatures and / or different central angles.

4. A gas-insulated transformer bushing according to claim 3, characterized in that, The central angle of each arc segment in the arc segment (2023) ranges from 0° to 90°.

5. A gas-insulated transformer bushing according to claim 1, characterized in that, The sleeve cone (101) is made of a composite material of epoxy and glass fiber.

6. The optimization method for a gas-insulated transformer bushing according to any one of claims 1-5, characterized in that, The process includes the following: Establish a finite element analysis model of the electric field of a gas-insulated transformer bushing as described in any one of claims 1-5; The optimization objective and the range of geometric parameters of the shielding ring (202) are set. The optimization objective is that when the core potential of the gas-insulated transformer bushing is a preset voltage and the potential of the shielding ring (202) is 0V, the maximum value of the surface electric field of the shielding ring (202) obtained by simulation does not exceed the preset value. Within the range of geometric parameters of the shielding ring (202), the electric field finite element analysis model is coarsely meshed and the maximum field strength on the surface of the shielding ring (202) is calculated to obtain a low-fidelity data set. The electric field finite element analysis model was meshed with a fine mesh and the maximum field strength on the surface of the shielding ring (202) was calculated to obtain a high-fidelity data set; Using the low-fidelity data set as the source domain data of the transfer learning model and the high-fidelity data set as the target domain data of the transfer learning model, a surrogate model is constructed to predict the maximum field strength on the surface of the shielding ring (202) under different combinations of geometric parameters. The surrogate model is optimized using a genetic algorithm to predict the combination of geometric parameters that minimizes the electric field on the surface of the shielding ring (202), thereby achieving optimization of the gas-insulated transformer bushing.

7. The optimization method for a gas-insulated transformer bushing as described in claim 6, characterized in that, The geometric parameters of the shielding ring (202) include the horizontal extension length of the horizontal segment (2021), the vertical insertion depth of the vertical segment (2022), the radius of the arc segment (2023), and the arc angle.

8. The optimization method for a gas-insulated transformer bushing as described in claim 6, characterized in that, The electric field finite element analysis model was established in Comsol software.

9. The optimization method for a gas-insulated transformer bushing as described in claim 8, characterized in that, Within the range of geometric parameters of the shielding ring (202), in the Comsol software, select the regular option in the mesh coarse setting, parametrically scan to calculate the maximum field strength on the surface of the shielding ring (202) to obtain a low-fidelity data set; in the Comsol software, select the ultra-fine option in the mesh coarse setting, parametrically scan to calculate the maximum field strength on the surface of the shielding ring (202) to obtain a high-fidelity data set.

10. The optimization method for a gas-insulated transformer bushing as described in claim 6, characterized in that, Using the low-fidelity data set as the source domain data and the high-fidelity data set as the target domain data for the transfer learning model, a surrogate model is constructed to predict the maximum electric field intensity on the surface of the shielding ring (202) under different combinations of geometric parameters. Specifically, the model includes: The surrogate model is constructed by importing the low-fidelity dataset into the source domain of the Matlab transfer learning program and the high-fidelity dataset into the target domain of the Matlab transfer learning program.

Images