Dry-type transformer temperature rise field simulation system and method

Abstract

The invention relates to the technical field of transformer simulation, in particular to a simulation system and method for a temperature rise field of a dry-type transformer. The method comprises the following steps: acquiring geometric parameters of an iron core lamination of the dry-type transformer to be simulated, spatial arrangement information of a winding and a positioning relation of an insulating component to form structural design data; the method comprises the following steps: constructing three-dimensional geometric modeling data based on structural design data, importing the three-dimensional geometric modeling data into a finite element electromagnetic simulation platform, setting a power-on condition of an excitation winding, and applying rated current or voltage, so as to obtain local magnetic flux density distribution data at each iron core lamination position in a whole operation period. According to the method, on the basis of three-dimensional structure data, three physical processes of an electromagnetic field, a thermal field and a fluid field are fused, a high-precision local heat source model and a dynamic temperature rise prediction model are constructed, and geometry-material-boundary all-parameter modeling is achieved from the whole structure of an iron core lamination, a winding and an insulation assembly.

Description

Technical Field

[0001] The present invention relates to the technical field of transformer simulation, and in particular to a dry-type transformer temperature rise field simulation system and method. Background Art

[0002] Dry-type transformers lack oil cooling. Heat is primarily generated by core and winding losses (copper and iron losses), which are dissipated to the surrounding environment through convection, radiation, and conduction. The temperature rise process can be divided into two stages: transient temperature rise, which occurs during startup or sudden load changes, and steady-state temperature rise, which stabilizes after reaching thermal equilibrium. However, over long-term operation, dry-type transformers are susceptible to factors such as localized overheating, difficult-to-detect hot spots in the windings, and poor ventilation due to compact structures. These factors can lead to safety risks such as excessive temperature rise, insulation degradation, and even breakdown. Therefore, accurately predicting and controlling the temperature rise distribution within dry-type transformers is crucial for ensuring their operational reliability and lifespan.

[0003] Traditional temperature-rise assessment methods, often based on experimental temperature measurements or empirically corrected formulas, fail to reflect the complex local heat distribution patterns within windings, cores, and insulation structures. Their accuracy and adaptability are particularly limited under conditions of non-uniform heat sources, localized air duct blockage, or structural modifications. Furthermore, while some numerical simulation methods incorporate electromagnetic-thermal coupling calculations, they often overlook the localized non-uniformity of magnetic flux density within the core or fail to consider the actual changes in thermal convection and heat transfer boundaries, making it difficult to truly reflect the dynamic temperature-rise field evolution of dry-type transformers under actual operating conditions. Summary of the Invention

[0004] Based on this, it is necessary for the present invention to provide a dry-type transformer temperature rise field simulation system and method to solve at least one of the above technical problems.

[0005] To achieve the above object, a dry-type transformer temperature rise field simulation method includes the following steps: Step S1: Collecting geometric parameters of the core laminations, spatial arrangement information of the windings, and positioning relationships of the insulation components of the dry-type transformer to be simulated to form structural design data; Step S2: constructing three-dimensional geometric modeling data based on the structural design data, importing the three-dimensional geometric modeling data into a finite element electromagnetic simulation platform, setting the energizing conditions of the excitation winding, and applying the rated current or voltage to obtain the local magnetic flux density distribution data at each core lamination position during the entire operating cycle; Step S3: Mapping the local magnetic flux density distribution data to spatial coordinates, calculating the eddy current loss and hysteresis loss respectively in combination with the directional parameters of the core material, and outputting the core loss distribution data for each spatial unit; Step S4: Using the core loss distribution data as heat source input, an electromagnetic-thermal-fluid coupling model is established in the multi-physics field simulation platform, a functional relationship between the thermal conductivity parameter and temperature is set, and the air flow rate, temperature, and convection heat transfer coefficient under natural air cooling or forced air cooling are set as boundary conditions; Step S5: Iteratively solve the electromagnetic-thermal-fluid coupling model using a transient finite element solver to obtain dynamic temperature rise field data including the global temperature distribution at each time point.

[0006] The present invention also provides a dry-type transformer temperature rise field simulation system for executing the above-mentioned dry-type transformer temperature rise field simulation method, the dry-type transformer temperature rise field simulation system comprising: The structural modeling preparation module is used to collect the geometric parameters of the core laminations of the dry-type transformer to be simulated, the spatial arrangement information of the windings, and the positioning relationship of the insulation components to form structural design data; The electromagnetic modeling and simulation module is used to construct three-dimensional geometric modeling data based on structural design data, import the three-dimensional geometric modeling data into the finite element electromagnetic simulation platform, set the energization conditions of the excitation winding, and apply the rated current or voltage to obtain the local magnetic flux density distribution data at each core lamination position during the entire operating cycle; The loss calculation mapping module is used to map the local magnetic flux density distribution data into spatial coordinates, calculate the eddy current loss and hysteresis loss respectively based on the directional parameters of the core material, and output the core loss distribution data for each spatial unit; The multi-field coupling modeling module is used to input core loss distribution data as a heat source, establish an electromagnetic-thermal-fluid coupling model in the multi-physics field simulation platform, set the functional relationship between thermal conductivity parameters and temperature, and set the air flow rate, temperature, and convective heat transfer coefficient under natural or forced air cooling as boundary conditions; The temperature rise field solving module is used to iteratively solve the electromagnetic-thermal-fluid coupling model using a transient finite element solver, and obtain dynamic temperature rise field data including the global temperature distribution at each time point through the solution.

[0007] The present invention can fully and accurately reproduce the temperature rise evolution law of the dry-type transformer during actual operation through the above-mentioned transient simulation steps based on electromagnetic-thermal-fluid coupling modeling, thereby providing a solid data foundation and highly consistent physical reference for thermal field analysis and structural optimization. First, in the meshing stage, by setting encrypted grids on the core boundary, winding surface and fluid-solid interface area, and constraining the skewness and aspect ratio of the overall grid, the geometric quality and stability of the solution unit are ensured. High-quality grids directly affect the capture accuracy of the thermal flow boundary layer and the ability to restore local turbulence characteristics in the fluid field, and are the key basis for whether the entire coupled simulation can reflect the real physical process.

[0008] Secondly, the reasonable configuration of transient solution parameters helps to strike a balance between computational efficiency and solution accuracy. The setting of simulation time and time step is determined according to the rate of change of the physical field and the grid size. Especially when the airflow velocity is high or there is a subtle heat source gradient in the structure, the use of a small step size solution can effectively capture the sudden temperature rise process. Setting the residual limit of the energy and momentum equations ensures the convergence of the solution of the heat-flow field at each time step, avoiding the misleading of the overall temperature rise trend by numerical drift. In addition, by reasonably initializing the temperature and velocity boundaries and setting the initial conditions that are physically consistent with the actual operating conditions, the simulation starting point has engineering authenticity, avoiding the deviation of simulation results from reality due to distorted assumptions.

[0009] Furthermore, when the heat source input is loaded into the model, the core iron losses (including hysteresis losses and eddy current losses) and the winding copper losses are converted into spatially distributed heat sources based on different physical mechanisms and accurately applied to their respective unit cells through mapping. This heat source loading method, driven by electromagnetic field simulation data, avoids the use of an equivalent average distribution of thermal power in previous simplified models. It can clearly reflect the spatial non-uniformity of losses, especially when there is an abnormal concentration of loss density at locations such as winding edges, ends, and ventilation blind spots, and can more keenly reveal potential overheating risk points. In addition, by setting continuity boundary conditions for temperature and heat flow at the fluid-solid coupling interface, a real heat exchange path is established between internal heat conduction in the solid and external convective heat transfer in the fluid, achieving dynamic feedback and coordinated response between the thermal field and the flow field. For natural cooling or forced air cooling scenarios, buoyancy models or inlet and outlet velocities, pressure boundaries, and Reynolds number-related turbulence parameters are set separately to make the boundary control of the fluid field adaptable to multiple working conditions. The introduction of a turbulence model is particularly crucial for analyzing heat exchange-enhanced regions, such as high-speed turbulence areas within the duct, recirculation zones at the winding ends, and convection-enhanced areas within the winding airway. This model fully reflects phenomena such as uneven air flow and variations in heat transfer efficiency found in actual equipment. Through the high-fidelity settings of physical parameters, boundary controls, material properties, and initial states, the model accurately represents the temperature response of various dry-type transformer components during long-term energized operation during an iterative solution process. The output dynamic temperature rise field data not only includes the full spatial temperature distribution but also preserves the transient evolution characteristics of the temperature rise process. Compared to traditional steady-state thermal simulation methods, this method enables analysis of temperature evolution at various stages before, during, and after a fault, particularly with greater sensitivity and resolution in identifying temperature mutation points, heat accumulation areas, and abnormal airway heat transfer efficiency. This full-field, high-resolution, and dynamically evolving simulation data provides a visual, quantifiable, and traceable foundation for subsequent identification of temperature anomaly types, attribution of heat sources, and optimization of structural heat dissipation channels. In engineering practice, it can help designers predict thermal control risks during operation at the initial stage of structural scheme formation, carry out targeted structural optimization in advance, significantly reduce failure risks such as winding overheating, local core ablation, and inter-turn insulation breakdown, and provide basic model data support for equipment intelligent monitoring and fault warning strategies, thereby realizing the transition of dry-type transformers from passive heat dissipation design to active thermal control optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments thereof made with reference to the following drawings: Figure 1 This is a schematic flow chart of the steps of a dry-type transformer temperature rise field simulation method according to the present invention; Figure 2 for Figure 1Detailed step flow diagram of step S1; Figure 3 A time series line graph of the magnetic flux density of a core lamination unit provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0011] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. It is obvious that the embodiments described are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts are within the scope of protection of the present invention.

[0012] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor and / or microcontroller approaches.

[0013] It should be understood that although the terms "first," "second," and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used solely to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0014] To achieve this, please refer to Figures 1 to 3 The present invention provides a dry-type transformer temperature rise field simulation method, the method comprising the following steps: Step S1: Collecting geometric parameters of the core laminations, spatial arrangement information of the windings, and positioning relationships of the insulation components of the dry-type transformer to be simulated to form structural design data; Step S2: constructing three-dimensional geometric modeling data based on the structural design data, importing the three-dimensional geometric modeling data into a finite element electromagnetic simulation platform, setting the energizing conditions of the excitation winding, and applying the rated current or voltage to obtain the local magnetic flux density distribution data at each core lamination position during the entire operating cycle; Step S3: Mapping the local magnetic flux density distribution data to spatial coordinates, calculating the eddy current loss and hysteresis loss respectively in combination with the directional parameters of the core material, and outputting the core loss distribution data for each spatial unit; Step S4: Using the core loss distribution data as heat source input, an electromagnetic-thermal-fluid coupling model is established in the multi-physics field simulation platform, a functional relationship between the thermal conductivity parameter and temperature is set, and the air flow rate, temperature, and convection heat transfer coefficient under natural air cooling or forced air cooling are set as boundary conditions; Step S5: Iteratively solve the electromagnetic-thermal-fluid coupling model using a transient finite element solver to obtain dynamic temperature rise field data including the global temperature distribution at each time point.

[0015] Furthermore, after step S5, the following steps are also included: Quantify the degree of local temperature mutation based on dynamic temperature rise field data and locate high temperature gradient areas; In some embodiments, the three-dimensional temperature distribution data at multiple time points are extracted from the transient simulation results, and the stage where the temperature rise tends to be stable (such as after running for 30 minutes or 1 hour) is selected as the analysis section to extract the temperature field tensor corresponding to this moment. Then, the three-dimensional gradient operator ( ) numerically differentiate the entire temperature field to obtain the temperature change rate of each unit in the X (radial), Y (tangential), and Z (axial) directions, generating a three-dimensional temperature gradient vector field. Next, calculate the temperature gradient modulus of each unit. By setting a gradient threshold (e.g., more than twice the average gradient), we can identify areas with significant temperature changes and mark their spatial extent as high-temperature gradient regions. The high-temperature gradient regions are then output as a closed grid set based on spatial coordinates, along with information such as the temperature, gradient vector, and component location (e.g., winding, core, insulation) for each grid cell.

[0016] It should be noted that the judgment of gradient mutation should be based on the relative gradient (temperature change per unit length) rather than the absolute temperature difference, otherwise misjudgment may occur in high-temperature but uniformly distributed areas.

[0017] Calculate the temperature gradient in the high temperature gradient area in the winding height direction as the axial gradient, and calculate the temperature gradient in the high temperature gradient area in the winding radius direction as the radial gradient; In some embodiments, the high temperature gradient area is limited to the "winding" component, eliminating interference between the core and the insulation structure; and a local cylindrical coordinate system is established with the winding center axis (Z axis) as the reference. Then, in this coordinate system, the partial derivative of the temperature field in the z direction ( ) as the axial gradient, the partial derivative in the r direction ( ) as the radial gradient, and discrete calculations are performed within the high-temperature gradient region using central difference or high-order finite difference methods. For each unit or slice, the axial and radial temperature gradient values ​​are calculated separately and statistically averaged or weighted averaged to form representative axial and radial temperature gradient indices.

[0018] It should be noted that the tangential direction ( direction) usually has little effect on temperature rise and can be ignored in this method; if the winding adopts a segmented structure, the gradient of each winding segment should be calculated separately and then averaged.

[0019] The ratio of the axial gradient to the radial gradient was calculated and recorded as the axial-radial gradient ratio; In some embodiments, the average axial temperature gradient obtained in the above steps is and the mean radial temperature gradient Substitute into the formula: axial-radial gradient ratio ; This ratio reflects whether the heat conduction directionality inside the winding is balanced. The larger the value, the heat accumulation is mainly in the axial direction and is difficult to dissipate. The smaller the value, the heat accumulation is mainly in the radial direction and is difficult to diffuse.

[0020] Optionally, this ratio is also calculated for each grid cell to form an axial-radial gradient ratio field (tensor field) for spatial visualization of thermal diagnosis.

[0021] When the axial-radial gradient ratio is greater than or equal to 3, it is determined that the axial heat dissipation is blocked due to heat dissipation obstruction of the winding end or blockage of the axial ventilation duct; When the axial-radial gradient ratio is greater than 0.5 and less than 3, it is determined to be local overheating caused by uneven distribution of heat sources inside the winding; When the axial-radial gradient ratio is less than or equal to 0.5, it is determined that the inter-turn air duct heat dissipation efficiency is low due to insufficient radial convection.

[0022] In some embodiments, the heat dissipation status of the winding region can be quantitatively classified and diagnosed based on the numerical distribution of the axial-radial temperature gradient ratio. When the ratio is greater than or equal to 3, it indicates that heat conduction in the axial direction is significantly hindered, resulting in heat being unable to diffuse smoothly along the height of the winding to the upper and lower ends. This is often manifested as the temperature in the center of the winding being significantly higher than at the ends. Such characteristics are generally attributed to heat dissipation obstruction at the winding ends or blockage of the axial ventilation duct. Possible causes include poor sealing of the end structure, obstruction of the ventilation duct by foreign matter, or overly tight insulation wrapping at the end. When the axial-radial gradient ratio is between 0.5 and 3, heat has a certain degree of diffusion ability in both the axial and radial directions, and there is no severe thermal retardation in a single direction. However, due to internal electromagnetic losses or structural unevenness, localized heat accumulation may occur, forming isolated hot spots. This situation is generally determined to be local overheating caused by uneven distribution of heat sources within the winding. Potential causes include local short-turn faults, concentrated current density, uneven winding stacking, or increased stray losses. If the axial-radial gradient ratio is less than or equal to 0.5, it indicates that heat is not effectively conducted and diffused in the radial direction, potentially leading to high-temperature accumulation areas at the outer diameter of the winding or in the inter-turn channels. This reflects insufficient radial convection and low heat dissipation efficiency in the inter-turn airways. This phenomenon is often caused by improper ventilation structure design, insufficient cooling air volume, or blockage of ventilation channels after winding varnishing.

[0023] It should be noted that the above judgment criteria are based on the reliability of the simulated temperature rise field. In practical applications, they should be cross-validated with thermometer measured data or infrared thermal images to improve diagnostic accuracy.

[0024] Furthermore, step S1 includes the following steps: Step S11: measuring the length, width, and thickness of the core laminations in the dry-type transformer to be simulated, and recording the lamination material, the thickness of the insulation layer between the laminations, the core window size, and the core leg spacing parameters to form the core geometry parameters; In some embodiments, the geometrical dimension parameters of the core laminations in the dry-type transformer to be simulated are obtained by precise measurement means, including the length, width and thickness of the laminations. The measured dimensions should cover the geometric differences that may exist in different parts of the core (such as the core column and the iron yoke). In addition, the model and type of the core lamination material used must be identified and recorded, such as whether it is oriented silicon steel, amorphous alloy, etc., for the subsequent extraction of the directional magnetic properties of the material. For the insulating coating or dielectric layer between the core laminations, a microscope or thin layer measurement tool should be used to confirm its average thickness, and use it as the insulation gap input parameter in the loss calculation. At the same time, the overall window size of the core is measured, including the window height and width, to determine the space area available for winding arrangement; It should be noted that for three-phase transformers, the center-to-center distance between each core leg should also be recorded to accurately reflect the relative arrangement of the multi-phase windings. All of the above parameters will be organized into a core geometry parameter dataset, which serves as the basic input for subsequent 3D modeling and electromagnetic simulation.

[0025] Step S12: collecting the inner diameter, outer diameter, and height dimensions of the high-voltage winding and the low-voltage winding of the dry-type transformer to be simulated, measuring the winding axial position coordinates and radial spacing, recording the number of winding turns, the cross-sectional area of ​​the conductor, and the distance between the windings, and forming the winding space layout parameters; In some embodiments, comprehensive spatial layout information for the high-voltage and low-voltage windings in dry-type transformers is required. First, the inner diameter, outer diameter, and height parameters of each winding are obtained through external measurements, and if necessary, verified with drawings. A three-dimensional coordinate measuring machine is then used to determine the axial center and radial position of the winding within the core window, clarifying the spatial cladding relationship and spacing arrangement of the high- and low-voltage windings. Simultaneously, the number of turns, current level, and conductor cross-sectional area parameters of the winding are read, and the conductor cross-sectional shape (e.g., round or flat wire) is recorded. This data will be used to calculate the current density distribution in electromagnetic simulations. During the winding layout, special attention should be paid to measuring the radial distance between windings (i.e., the insulation gap between windings) and the spacing between the winding and the core to reflect the actual electrical insulation structure and heat exchange path. All parameters are then organized into a unified winding spatial layout parameter dataset, providing a structural foundation for the construction of electromagnetic, thermal, and fluid coupling models.

[0026] Step S13: Identify the material type and thickness parameters of the inter-winding insulation, winding-to-ground insulation, and phase-to-phase insulation based on the core geometry parameters and the winding spatial layout parameters, measure the spatial position coordinates of each insulation component, record the relative position relationship between the insulation component and the core and winding, and establish the insulation component spatial positioning data; In some embodiments, the main insulation components in a dry-type transformer include insulation between high- and low-voltage windings (e.g., laminates, air gaps), ground insulation (e.g., bottom pads, supports), and interphase insulation between multi-phase windings (e.g., slot plates or insulation screens). Subsequently, standard thickness measurement tools (e.g., vernier calipers or precision rulers) are used to obtain the actual thickness parameters of each insulation component, and their material type and heat resistance rating are annotated based on the design data. Through physical positioning, structural drawing comparison, or 3D measurement, the 3D spatial coordinates of each insulator relative to the core and windings are further determined to identify its nested, attached, or suspended state within the overall structure. During the recording process, the relative position accuracy of all insulation components must be ensured to support heat flow path analysis and thermal resistance modeling. Ultimately, the aforementioned materials, thicknesses, and spatial relationships are organized into spatial positioning data for insulation components, providing input for setting heat conduction paths and boundary conditions in the electromagnetic-thermal-fluid coupling model.

[0027] It should be noted that although insulators have extremely low electrical conductivity in electromagnetic simulations and hardly participate in current channels, they often form dominant thermal resistance channels in thermal simulations. Therefore, this step is crucial in constructing the temperature rise field.

[0028] Step S14: combining the core geometry parameters, winding spatial arrangement parameters and insulation component spatial positioning data into structural design data based on the physical contact relationship among the dry-type transformer core, winding and insulation components to be simulated.

[0029] In some embodiments, the core geometry parameters, winding spatial arrangement parameters, and insulation component spatial positioning data obtained above are unified and integrated into a complete structural design data set. First, the spatial positions of all components should be aligned and spliced ​​based on the three-dimensional coordinate system to ensure that the relative relationship between the core, winding, and insulation parts is consistent with the actual structure. Next, the physical contact relationship between the structures is analyzed, such as whether there is a close fit between the winding and the core, whether there is an insulating support body in between, etc. This information will be used to assign contact thermal resistance and interface heat dissipation coefficient in subsequent modeling. The data integration process can be performed through geometric merging on a CAD software platform, or completed in a simulation pre-processing tool through a script. The final output structural design data should include a complete three-dimensional geometric model file (such as STEP format), as well as a physical property table associated with each structural component, which is used as a unified input interface for the joint solution of electromagnetic fields, thermal fields, and flow fields.

[0030] It should be noted that the structural design data is not only the basis for geometric modeling, but also the core basis for parameter binding, material property mapping and boundary setting in the coupled simulation system. Its accuracy will directly affect the physical credibility and engineering reference value of the simulation results.

[0031] Furthermore, constructing three-dimensional geometric modeling data based on the structural design data in step S2 includes: A right-handed coordinate system is established with the center of the core as the origin, the axial direction as the Z axis, and the radial direction as the X / Y axis. Based on the core lamination dimensions in the structural design data, silicon steel sheets are stacked layer by layer with stepped joints in a right-hand coordinate system to obtain a core lamination model; According to the inner / outer diameter dimensions of the winding in the structural design data, the winding is stretched into a cylinder and the axial / radial airway positions are marked to obtain the winding model; Insert insulation components between the core lamination model and the winding model according to the insulation positioning coordinates in the structural design data, and verify the gap with the winding to form an initial three-dimensional model; Assign magnetic material properties, conductor material properties, and insulation material properties to the initial three-dimensional model, set the XY plane conductivity to zero, and input the Z-direction conductivity according to the measured value to obtain three-dimensional geometric modeling data.

[0032] In some embodiments, a unified spatial coordinate system is established, with the core's geometric center as the origin. The Z axis is set along the axial direction of the core legs (i.e., the winding centerline), and the X and Y axes are defined as radial directions. These three axes form a right-handed rectangular coordinate system. Within this coordinate system, the core is three-dimensionally modeled based on the core lamination dimensions and stacking method provided in the structural design data. Based on the length, width, thickness, and interlayer insulation thickness of the individual silicon steel sheets, a "step-by-step stacking" method is used to simulate the actual core lamination structure. The core legs and yokes are then assembled to form a monolithic core frame model based on the spatial relationship between the core legs and yoke. Subsequently, based on the inner and outer diameters and axial height parameters of the high-voltage and low-voltage windings, a hollow cylindrical shell is extruded in the core window area to represent the winding geometry. Simultaneously, the axial / radial positions of air ducts or heat transfer plates are marked within the winding's axial height range based on the air duct data. Based on the number and thickness of the air ducts, the air duct areas are hollowed out or embedded in the geometric model to achieve detailed modeling of the winding's ventilation structure. Based on the positioning coordinates of the insulation components, they are inserted into the predetermined locations between the windings and the core, including inter-winding insulation, phase-to-phase insulation, and ground insulation, and their contact interfaces with the surrounding structure are precisely aligned. During assembly, the minimum clearance between the insulation components and the winding model should be verified in the 3D modeling platform to ensure that geometric conflicts do not affect subsequent meshing and solution stability. After the geometric arrangement of all components is initially completed, the entire structure is defined as an initial 3D model. Based on this initial 3D model, material properties are further assigned to each structural body. The core region is assigned ferromagnetic material properties with directional magnetic permeability and hysteresis loss coefficient; the winding region is assigned a high-conductivity conductor material (such as copper or aluminum); and the insulation is assigned electrical insulating material properties with low thermal conductivity and zero conductivity. Furthermore, to accurately reflect the interlaminar conductivity characteristics of the core material, the conductivity of the core material in the XY plane should be set to zero in the electromagnetic simulation, indicating that there is no current path between the silicon steel sheets. The Z-direction conductivity should be input according to measured data or manufacturer-provided material parameters to accurately assess the power-frequency eddy current distribution.

[0033] It should be noted that although the iron core behaves as an electromagnetic functional body as a whole, its lamellar structure and directional magnetic characteristics have a decisive influence on the magnetic flux distribution and loss. Therefore, the three-dimensional geometric modeling should not only accurately reflect the appearance, but also reflect the material directional properties and layered structure in detail, so as to ensure the physical accuracy of the magnetic field solution in subsequent simulations.

[0034] Furthermore, in step S2, the three-dimensional geometric modeling data is imported into the finite element electromagnetic simulation platform, the power-on conditions of the excitation winding are set, and the rated current or voltage is applied, so as to obtain the local magnetic flux density distribution data at each core lamination position during the entire operation cycle, including: Import the 3D geometric modeling data into the finite element electromagnetic simulation platform in a standard format, and verify the integrity of the geometry and the absence of overlap between components to obtain a complete 3D model tree; In the complete 3D model tree, assign the BH magnetization curve of the silicon steel sheet to the core lamination, input the nonlinear parameters of the change of magnetic permeability with magnetic field strength, set the resistivity and density properties of the core material, input the resistivity, density and cross-sectional area parameters of the high-voltage winding and low-voltage winding in the complete 3D model tree, set the geometric path of the winding end connection wire, define the energization conditions of the excitation winding, apply the rated current or voltage, and collect the local magnetic flux density tensor field; The local magnetic flux density tensor field is divided into multiple instantaneous magnetic flux density vector layers according to the simulation time step, and the spatial coordinate index is marked at each time point to obtain the time-space corresponding vector layer; The time-space corresponding vector diagram layer is sliced ​​in three axes, and the magnetic flux distribution characteristics on different cross sections inside each core lamination are extracted. The average magnetic flux density and gradient change rate of each lamination layer are calculated as the local magnetic flux density distribution data.

[0035] In some embodiments, the completed 3D geometry modeling data is imported into a finite element electromagnetic simulation platform in a standard format (e.g., .step, .iges, or .sat, a geometry exchange file compatible with CAD / CAE systems). The platform's built-in geometry diagnostic tools are then used to verify the integrity of the model structure, including boundary closure, facet normal consistency, and verification of overlap, intersection, or undefined connectors between components. After verification, a complete 3D model tree is automatically generated, with the core, windings, and insulation components assigned to distinct physical region nodes to facilitate subsequent property assignment and boundary condition setting. Next, material properties are assigned to each core lamination in the model tree. To accurately reflect the magnetic saturation and hysteresis behavior of the silicon steel lamination, the material's BH magnetization curve (i.e., the nonlinear relationship between magnetic flux density B and magnetic field intensity H) should be input. This can be done using a vector table or function fit. Furthermore, to account for power-frequency eddy current losses, the core material's resistivity and density should be set for subsequent coupled thermal-electrical simulations. In the high-voltage and low-voltage winding regions, the conductor material resistivity and density must be assigned, and the conductor cross-sectional area must be specified based on the winding's cross-sectional geometry to enable the simulation system to automatically calculate the current density distribution. Furthermore, the electrical connection paths at both ends of the winding must be clearly defined, and the geometric orientation of the incoming and outgoing wires must be defined to ensure that current flows accurately throughout the entire winding body under energized conditions. Regarding boundary condition settings, the simulation system specifies an "excitation winding" energization condition. This involves applying a sinusoidal AC current or voltage source of rated amplitude to the high-voltage or low-voltage winding. The simulation frequency is typically set to the power frequency (50 Hz or 60 Hz) to simulate the device's rated operating conditions. During the simulation, the system automatically calculates the magnetic flux density tensor field for the entire structure and records the time-varying three-dimensional vector distribution of the magnetic induction intensity B within all structural elements. This magnetic flux density tensor field is segmented according to the specified simulation time step (e.g., a 1 ms cycle is divided into 100 steps) to generate multiple instantaneous vector layers. Each layer is annotated with a corresponding timestamp and three-dimensional spatial index, that is, the position coordinates of each unit in the model volume and the magnetic flux density vector information at the corresponding moment, thus forming a complete time-space corresponding magnetic field map. The magnetic flux density layer data at each moment is then spatially sliced ​​in the X, Y, and Z directions to extract the magnetic flux distribution images on different axial, radial, and annular sections. Within each slice area, the modulus and vector direction of the magnetic flux density B are further extracted, and the average magnetic flux density is calculated according to the core lamination layer. The gradient change rate is estimated using the magnetic flux difference between adjacent lamination layers, thereby constructing the local magnetic flux density distribution data for each lamination during the entire operating cycle, providing an accurate basis for subsequent loss calculations and heat source assignments.

[0036] It should be noted that in this process, in order to ensure that the spatial resolution of the magnetic flux density data matches the laminate structure, a finite element meshing strategy equivalent to the thickness of the core laminate must be used, and appropriate local mesh encryption areas must be enabled. Otherwise, the magnetic flux information in high-gradient areas may be distorted, thereby affecting the accuracy of the local eddy current loss and hysteresis loss assessment.

[0037] Furthermore, the definition of the power-on conditions of the excitation winding and the application of the rated current or voltage include: for the voltage excitation mode, applying the rated voltage amplitude and frequency parameters at the end of the high-voltage winding, and setting the end of the low-voltage winding as the grounding boundary; for the current excitation mode, applying the rated current density distribution in the high-voltage winding, and automatically calculating the induced current distribution in the low-voltage winding through the ampere-turn relationship.

[0038] In some embodiments, depending on the simulation analysis objectives, voltage or current excitation can be selected to simulate the electromagnetic behavior of dry-type transformers operating at power frequency. Voltage excitation is typically suitable for studying the core's magnetic flux distribution and no-load losses under normal operating voltage.

[0039] In this approach, a voltage excitation boundary is set at one end of the high-voltage winding, with the rated voltage amplitude (e.g., 10 kV) and frequency (e.g., 50 Hz) input. The other end is set to free floating or zero ground potential. Both ends of the low-voltage winding are defined as grounded boundaries or reference potentials to stabilize the electric field distribution. The simulation platform automatically solves the time-series response of the induced voltage in the low-voltage winding based on the winding electrical parameters and mutual inductance coupling characteristics, thereby reproducing the coupled magnetic field between the winding and the core during no-load operation of an actual transformer.

[0040] The current excitation method is more commonly used to evaluate iron loss, temperature rise, or short-circuit characteristics during load operation. In this method, an equivalent current density distribution needs to be applied within the physical volume of the high-voltage winding. The current density per unit cross-section can be calculated by giving the total current amplitude and the number of winding turns (for example, applying a rated total current of 100A to a winding with a cross-sectional area of ​​50mm² and 500 turns corresponds to a current density of 2A / mm²). The simulation platform evenly distributes this current density within the conductor volume. At the same time, based on the conservation of excitation ampere-turns (i.e., the principle of magnetomotive force balance), the system automatically solves the equivalent current density distribution induced by the change in the main magnetic flux in the low-voltage winding. Even if the low-voltage winding is not externally energized, the system can accurately simulate the dynamic response of its induced current and its reaction to the magnetic field distribution.

[0041] For example, if simulating a 10kV / 400V dry-type transformer, in voltage excitation mode, a 10kV, 50Hz sinusoidal voltage source can be applied to the high-voltage winding, while the low-voltage winding is grounded. The system will then solve for the evolution of the main magnetic flux in the core and the relationship between the amplitude and phase of the voltage induced on the low-voltage side. If current excitation is used instead, a 100A AC current density can be input to the high-voltage winding. Based on the 500 turns of the high-voltage winding defined in the system, the simulation platform will automatically calculate the induced current on the low-voltage side according to the principle of flux consistency, closing the magnetomotive force loop in the system and completing the steady-state excitation solution.

[0042] It should be noted that in current excitation mode, since the simulation system must automatically match the inductive current, high requirements are placed on solver stability and mesh accuracy. Therefore, when simulating multi-turn windings, it is recommended to use a multi-region coupled vector boundary setting to ensure accurate ampere-turn control while avoiding high-frequency eddy current distribution distortion. In addition, voltage excitation is suitable for iron loss and magnetic saturation studies, while current excitation is more suitable for thermal simulation and load response analysis. The actual selection should be based on the simulation objectives.

[0043] Furthermore, step S3 includes the following steps: Step S31: extracting the magnetic flux density time series of each core lamination unit from the local magnetic flux density distribution data, calculating the peak value, effective value and change frequency of the magnetic flux density, and identifying the magnetic flux density components along the rolling direction and perpendicular to the rolling direction according to the rolling direction of the core material; In some embodiments, the time-varying magnetic flux density vector value of each core lamination unit is extracted from the local magnetic flux density distribution data to form a magnetic flux density time series of the unit within an industrial frequency operation cycle. This sequence usually records a value every millisecond or every angle step (such as every 5°) to form a complete time waveform curve. The peak value (i.e., maximum amplitude), effective value (root mean square RMS value) and main frequency component of the magnetic flux density can be calculated for this curve to determine whether there is a high-frequency disturbance in the magnetic flux change. In addition, combined with the rolling direction information of the silicon steel sheet recorded in the structural design, the three-dimensional magnetic flux density vector can be decomposed into two components, "along the rolling direction (RD) and perpendicular to the rolling direction (TD)" through vector projection, and subsequent loss analysis needs to be processed separately.

[0044] Step S32: Calculating the hysteresis loss density of the core lamination unit within one operating cycle using the BH magnetization curve of the silicon steel sheet based on the magnetic flux density component, wherein the hysteresis loss density is equal to the product of the area of ​​the hysteresis loop and the operating frequency; In some embodiments, based on the magnetic flux density time series in the two directions described above, the corresponding silicon steel sheet BH magnetization curve in the core material database is called to calculate the hysteresis loop within one cycle for each laminated unit. The hysteresis loss density is the area enclosed by the loop multiplied by the power supply frequency (e.g., 50 Hz). For example, if the hysteresis loop area formed by a unit in the rolling direction is 0.6 J / m³, the hysteresis loss density in that direction is 30 W / m³; if the loop area in the direction perpendicular to the rolling direction is 0.3 J / m³, the corresponding hysteresis loss density is 15 W / m³. The sum of the two is the total hysteresis loss density.

[0045] Step S33: Calculating the eddy current loss density along the rolling direction and the direction perpendicular to the rolling direction based on the magnetic flux density change rate in the magnetic flux density component and the thickness parameter and resistivity parameter of the core lamination, and performing vector synthesis of the eddy current loss density components in the two directions as the eddy current loss density; In some embodiments, to calculate the eddy current loss density, it is necessary to first calculate the time derivative (i.e., rate of change) of the magnetic flux density component and then substitute it into the classic eddy current loss formula. For a single core lamination unit, the eddy current loss density along the rolling direction (RD) and the perpendicular direction (TD) can be calculated using the following formulas:

[0046] in is the thickness of the laminate, is the frequency, is the peak magnetic flux density, The above calculation is applied to the RD and TD directions to obtain the eddy current loss density in the two directions, and then the total eddy current loss density of the unit is obtained by vector synthesis or energy superposition method.

[0047] Step S34: superimposing the hysteresis loss density and eddy current loss density of each core lamination unit to obtain the total core loss density of the unit, and multiplying it by the volume of the corresponding unit to calculate the core loss power value of each spatial unit; In some embodiments, the hysteresis loss density and eddy current loss density of each core lamination unit are added together to obtain the unit's total loss density (in W / m³). This total loss density is then multiplied by the unit's volume (typically the three-dimensional product of the unit's mesh size) to obtain the specific core loss power value (in W). For example, for a finite element with a volume of 1e-6 m³ and a loss density of 5000 W / m³, the core loss power of the unit is 5 mW.

[0048] Step S35: Arrange the core loss power values ​​of all core lamination units according to the spatial coordinate index to form a loss power distribution matrix in three-dimensional space, and output core loss distribution data containing the coordinates of each spatial unit and the corresponding loss power value.

[0049] In some embodiments, the power loss values ​​of all core lamination units are indexed and arranged according to their corresponding spatial coordinates in the model, forming a three-dimensional power distribution matrix. Each data record includes the spatial coordinates of the unit in the model (such as the x, y, and z position of the center point) and the corresponding power loss value at that point. This data set constitutes the core loss distribution data and can be exported to formats such as CSV and VTK for subsequent thermal simulation input, visualization, or analysis of high-loss areas.

[0050] It is important to note that during the hysteresis loop calculation process, the resolution of the BH data points must match the sampling rate of the time series. Otherwise, the loop will not be accurately closed, resulting in hysteresis loss errors. In addition, due to the orientation of actual core laminations, the BH curves along the rolling direction and the perpendicular direction are usually inconsistent. These curves must be modeled separately and cannot be simply replaced by an equivalent mean.

[0051] refer to Figure 3The time series line graph of the core lamination unit's magnetic flux density is shown as a typical time-domain magnetic flux density response diagram. This graph illustrates the changing trends in the rolling direction (B_RD), the perpendicular rolling direction (B_TD), and the resulting total magnetic flux density amplitude (|B|) for a specific core lamination unit over a complete operating cycle. These key analytical data are extracted from the local magnetic flux density tensor field. In step S31, the unit's magnetic flux density tensor is time-series expanded to obtain component curves along the rolling direction and perpendicular rolling direction (the blue and pink lines in the figure). This allows for the calculation of the peak value (e.g., a maximum of approximately 1.1 T for B_RD and 0.9 T for B_TD), the effective value (RMS calculated by integration), and the main frequency (corresponding to a period of 20 ms, or 50 Hz) of each component in each direction, laying the foundation for subsequent magnetic loss calculations. The magnetic flux density components in the B_RD and B_TD directions shown in the figure are fitted and integrated with the BH magnetization curves for the corresponding directions of the core material to obtain the hysteresis loop area per unit volume. This area is multiplied by the periodic frequency to obtain the hysteresis loss density. Due to the significant difference in the magnetic flux density in the two directions (B_RD is always higher than B_TD in the figure), hysteresis losses will exhibit significant anisotropy in actual calculations and should be calculated separately and then summed. By taking the slope of the time derivative of the magnetic flux density in the figure, dB / dt can be calculated in different directions. Based on the thickness and resistivity of the core laminations, the eddy current loss density per unit volume is calculated separately for the B_RD and B_TD directions according to the classical eddy current loss model. The eddy current losses in the two directions are then combined by vectorial superposition to obtain the total eddy current loss density. Subsequently, the hysteresis loss density and eddy current loss density are spatially superimposed and multiplied by the volume of the lamination unit to obtain the total core power loss corresponding to the point shown in the figure. This process not only reflects the frequency response characteristics of material loss, but also captures the loss gradient caused by the uneven local magnetic flux distribution in the core laminations.

[0052] Eventually it will be similar Figure 3 The time series data in the simulation is extended to all elements within the entire core lamination space. Through coordinate mapping, a three-dimensional power loss matrix is ​​formed, providing accurate input data for the subsequent construction of the heat source distribution field. This matrix not only records the coordinate position and corresponding power loss value of each element, but also outputs dynamic power distribution information based on the time step evolution trend.

[0053] It should be noted that Figure 3 The total magnetic flux density amplitude |B| curve (green) is not a simple algebraic sum, but the modulus of the magnetic flux density vector in three-dimensional space. It reflects the actual physical size of the magnetic flux density after superposition in two orthogonal directions. It should be used with caution as a reference value for the final composite loss. It should not be used directly in hysteresis or eddy current loss calculations. Instead, it should be processed using data in the B_RD and B_TD directions respectively.

[0054] Furthermore, step S4 includes the following steps: Step S41: Importing 3D geometric modeling data into the multi-physics simulation platform and establishing a heat transfer calculation domain, wherein the heat transfer calculation domain is specifically established by setting the iron core, windings, and insulation components as solid heat transfer domains, setting the air around the transformer as a fluid heat transfer domain, and setting fluid-solid coupling boundary conditions at the interface between the solid domain and the fluid domain; In some embodiments, the three-dimensional geometric modeling data constructed in step S2 is imported into a multi-physics simulation platform (such as COMSOL, ANSYS Fluent / CFX, etc.), and a heat transfer calculation domain is constructed on this basis. Specifically, the core laminations, windings (high voltage / low voltage), and insulation components in the model are respectively set as solid heat transfer domains (solid domains). The thermal conductivity calculations in these areas need to take into account the anisotropy of the material (such as the directional thermal conductivity of silicon steel sheets). At the same time, based on the air space or air duct area inside the dry-type transformer casing, a fluid heat transfer domain (fluid domain) is formed to simulate the process of cooling gas (usually air) flowing on the surface of the winding and the core to remove heat. The contact surface between the solid domain and the fluid domain (such as between the outer wall of the winding and the air) is defined as a fluid-solid coupling boundary (conjugate heat transfer interface), where the system automatically handles the continuity conditions of temperature and heat flux to ensure that heat transfer actually occurs.

[0055] Step S42: Load the core loss distribution data as a body heat source into the core lamination unit in the heat transfer calculation domain, convert the winding copper loss into a body heat source according to the current density distribution and load it into the corresponding winding unit to form a three-dimensional heat source distribution field; In some embodiments, the core loss distribution data calculated in the previous stage is applied as a volumetric heat source to all core lamination units in the heat transfer calculation domain. This means that each core unit applies local heat in the form of unit volumetric heat power. At the same time, the resistivity of each winding conductor in the structural design and the operating current density flowing through it must be combined according to the following formula:

[0056] The copper loss (ohmic loss) of each winding unit cell is converted into a thermal power value and applied to the corresponding winding unit cell as a volume heat source. Ultimately, a complete three-dimensional heat source distribution field that varies with spatial position is established in the model, including iron loss in the core area and copper loss in the winding area, which serves as the driving source for solving heat conduction and heat transfer.

[0057] Step S43: Setting the fluid domain boundary conditions in the electromagnetic-thermal-fluid coupling model. For the natural air cooling condition, set the buoyancy parameters and gravity direction. For the forced air cooling condition, set the air velocity and temperature at the air inlet and the pressure boundary at the air outlet. Calculate and set the turbulence model parameters based on the Reynolds number. In some embodiments, the boundary conditions of the fluid domain are set in the coupled model, and are divided into two typical working conditions: natural air cooling and forced air cooling. In the natural air cooling mode, the buoyancy force model of the simulation platform needs to be turned on, and the gravity direction and scalar gravity acceleration are set to simulate the natural convection process of hot air from bottom to top due to density differences; the initial air temperature can be set to the ambient temperature (such as 25°C). In the forced air cooling mode, the air inlet velocity vector (such as 2 m / s), inlet temperature (such as 20°C) and static pressure boundary of the air outlet (usually set to 0 Pa or atmospheric pressure) need to be specified on the air inlet surface of the model to ensure that the air flow can continue to pass through the inside of the transformer. For forced cooling scenarios with higher flow rates, the Reynolds number (Re) of the system should also be calculated to determine whether the flow is in a turbulent state, and the appropriate turbulence model (such as k- or SST model) to improve heat transfer accuracy.

[0058] Step S44: setting a convection heat transfer boundary condition on the coupling interface between the solid domain and the fluid domain, and establishing an electromagnetic-thermal-fluid coupling model including heat conduction, convection heat transfer, and radiation heat transfer through the three-dimensional heat source distribution field and turbulence model parameters.

[0059] In some embodiments, a convection heat transfer boundary condition is set at the interface between the fluid domain and the solid domain. The simulation platform will automatically couple the fluid heat transfer equation and the solid heat conduction equation to calculate the boundary heat flux to meet the heat transfer balance. To further improve the accuracy of the thermal coupling model, you can also choose to enable the radiation heat transfer module (Radiation to Ambient / Surface-to-Surface) at the interface between the high-temperature component (such as the outer layer of the winding) and the fluid, and input the material surface emissivity parameters (such as the copper surface =0.1, epoxy resin surface = 0.9) to account for the potentially significant radiation heat loss in natural cooling. By comprehensively considering the heat source, conduction, convection, and radiation effects, a complete electromagnetic-thermal-fluid three-field coupled model was ultimately developed for the dynamic solution of the subsequent temperature rise process.

[0060] It should be noted that when setting convection boundary conditions, if the duct geometry has complex turning points or dead corners, local recirculation or low-speed areas may form. It is recommended to locally refine the mesh of the fluid domain and enable steady-state initial field superposition transient iteration (such as SIMPLEC+Transient) in the solver. Otherwise, it will easily lead to initial convergence difficulties or distortion of temperature rise results.

[0061] Furthermore, step S5 includes: The electromagnetic-thermal-fluid coupling model was meshed, with a refined mesh at the core lamination boundary, winding surface, and fluid-solid coupling interface to obtain a finite element mesh model with a mesh skewness less than 0.8 and an aspect ratio less than 5. In some embodiments, after geometric modeling and physical field configuration are completed, the entire three-dimensional coupled model is meshed using finite element technology. To ensure solution accuracy and the ability to capture the thermal boundary layer, locally refined meshes (with a local minimum element size of 1–2 mm) are applied to the core lamination boundaries, the outer surface of the windings, the winding axial ventilation ducts, and the fluid-solid interface where the solid contacts the fluid. A relatively coarse volume mesh is used in the remaining areas to balance computational resources. After meshing, the global mesh quality is automatically evaluated to ensure that the skewness of the generated mesh is less than 0.8, meaning that the element shapes are not severely tilted. The mesh aspect ratio is also controlled to be less than 5 to prevent overstretching of elements that could degrade solver stability. The result is a finite element mesh model suitable for multi-physics transient solutions.

[0062] The transient solution parameters are set based on the finite element mesh model, including the total simulation time, time step, and convergence criterion. The time step is determined based on the minimum mesh size and the maximum flow velocity, and the convergence criterion is set to the energy residual being less than , the momentum residual is less than ; In some embodiments, the core parameters of the transient solver need to be configured on the finite element mesh model. First, set the total simulation time (for example, 600 seconds) and the time step (for example, 0.5 seconds). The time step should meet the CFL (Courant–Friedrichs–Lewy) stability criterion, as shown in the following formula: ;in is the minimum mesh element side length in the model, is the maximum possible velocity in the fluid domain (can be preliminarily estimated or determined by the forced wind speed input value). In addition, to improve simulation accuracy and convergence efficiency, a convergence criterion is set at each time step, that is, the residual limit of each physical field equation. Recommended setting: Energy residual (thermal field) is less than , the momentum residual (flow field) is less than , to ensure that the temperature field and flow velocity field are balanced with high accuracy at each time step.

[0063] Initial conditions are set on the nodes of the finite element mesh model. The initial temperature of the solid domain mesh nodes is set to the ambient temperature, the initial flow velocity of all fluid domain mesh nodes is set to zero, the initial heat source distribution of each unit is set according to the mapping relationship between the mesh unit and the heat source area, and temperature and heat flow continuity constraints are established on the mesh nodes of the fluid-solid coupling interface, thereby obtaining an iteratively solvable multi-physics field coupling initial state model. In some embodiments, the initial conditions of the transient simulation are set at the grid node level. The temperatures of all nodes in the solid domain (core, winding, insulator, etc.) are uniformly initialized to the ambient temperature (for example, 25°C) to simulate the initial cooling state. The velocities of all nodes in the fluid domain are set to zero, indicating that there is no airflow movement at the beginning. In addition, combined with the three-dimensional heat source distribution field defined in step S4, the local loss power corresponding to each grid unit is used as the body heat source input through a mapping relationship to form an initial heat source condition with consistent spatial distribution and zero excitation at time zero. In order to ensure that heat is correctly transferred between the solid-fluid interface, it is also necessary to set continuity constraints (conjugate boundary conditions) on the grid nodes on these coupled boundaries, that is, automatically balance the conduction and convection heat transfer terms on both sides of the boundary during the solution process to achieve seamless coupling. At this point, an electromagnetic-thermal-fluid multi-physics field initial state model is formed that can be iteratively solved by the transient simulation solver.

[0064] Based on the iteratively solvable multi-physics field coupling initial state model, the transient finite element solver is called to perform electromagnetic-thermal-fluid field coupling simulation, and the temperature distribution of each grid node in the model is iteratively calculated time step by time. The global temperature field results calculated in multiple time steps are integrated into dynamic temperature rise field data in a time series.

[0065] In some embodiments, a transient finite element solver in the platform (such as COMSOL's Time-DependentSolver and ANSYS Fluent's TransientSolver) is called to perform a step-by-step iterative solution to the multi-field model. In each time step, the temperature distribution of the solid domain under the action of the heat source is first calculated, and then the heat transfer effect of the air flow field on the surface is coupled and solved, and the temperature and velocity information of all nodes are updated. The iteration continues until the total simulation time is reached, and the complete temperature rise dynamic response process is obtained. Finally, the global three-dimensional temperature field results output at all time steps are organized into a time-varying data set to form the so-called dynamic temperature rise field data (Dynamic Temperature Field), which is used for subsequent high-temperature area identification, temperature gradient analysis, fault location, and other links.

[0066] It should be noted that if the simulation time span is long (e.g., ≥600s), it is recommended to adopt an adaptive time stepping strategy, which automatically reduces the step size when the temperature rise changes dramatically and automatically relaxes it during the stable phase to improve efficiency without sacrificing accuracy.

[0067] The present invention also provides a dry-type transformer temperature rise field simulation system for executing the above-mentioned dry-type transformer temperature rise field simulation method, the dry-type transformer temperature rise field simulation system comprising: The structural modeling preparation module is used to collect the geometric parameters of the core laminations of the dry-type transformer to be simulated, the spatial arrangement information of the windings, and the positioning relationship of the insulation components to form structural design data; The electromagnetic modeling and simulation module is used to construct three-dimensional geometric modeling data based on structural design data, import the three-dimensional geometric modeling data into the finite element electromagnetic simulation platform, set the energization conditions of the excitation winding, and apply the rated current or voltage to obtain the local magnetic flux density distribution data at each core lamination position during the entire operating cycle; The loss calculation mapping module is used to map the local magnetic flux density distribution data into spatial coordinates, calculate the eddy current loss and hysteresis loss respectively based on the directional parameters of the core material, and output the core loss distribution data for each spatial unit; The multi-field coupling modeling module is used to input core loss distribution data as a heat source, establish an electromagnetic-thermal-fluid coupling model in the multi-physics field simulation platform, set the functional relationship between thermal conductivity parameters and temperature, and set the air flow rate, temperature, and convective heat transfer coefficient under natural or forced air cooling as boundary conditions; The temperature rise field solving module is used to iteratively solve the electromagnetic-thermal-fluid coupling model using a transient finite element solver, and obtain dynamic temperature rise field data including the global temperature distribution at each time point through the solution.

[0068] The present invention is therefore intended to be illustrative and non-restrictive in all respects, with the scope of the invention being defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the application documents are intended to be embraced therein.

[0069] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A dry-type transformer temperature rise field simulation method, characterized in that: The following steps are involved: Step S1: Collecting geometric parameters of the core laminations, spatial arrangement information of the windings, and positioning relationships of the insulation components of the dry-type transformer to be simulated to form structural design data; Step S2: constructing three-dimensional geometric modeling data based on the structural design data, importing the three-dimensional geometric modeling data into a finite element electromagnetic simulation platform, setting the energizing conditions of the excitation winding, and applying the rated current or voltage to obtain the local magnetic flux density distribution data at each core lamination position during the entire operating cycle; Step S3: Mapping the local magnetic flux density distribution data to spatial coordinates, calculating the eddy current loss and hysteresis loss respectively in combination with the directional parameters of the core material, and outputting the core loss distribution data for each spatial unit; Step S4: Using the core loss distribution data as heat source input, an electromagnetic-thermal-fluid coupling model is established in the multi-physics field simulation platform, a functional relationship between the thermal conductivity parameter and temperature is set, and the air flow rate, temperature, and convection heat transfer coefficient under natural air cooling or forced air cooling are set as boundary conditions; Step S5: Iteratively solve the electromagnetic-thermal-fluid coupling model using a transient finite element solver to obtain dynamic temperature rise field data including the global temperature distribution at each time point.

2. The dry-type transformer temperature rise field simulation method according to claim 1, characterized in that: After step S5, the method further includes the following steps: Quantify the degree of local temperature mutation based on dynamic temperature rise field data and locate high temperature gradient areas; Calculate the temperature gradient in the high temperature gradient area in the winding height direction as the axial gradient, and calculate the temperature gradient in the high temperature gradient area in the winding radius direction as the radial gradient; The ratio of the axial gradient to the radial gradient was calculated and recorded as the axial-radial gradient ratio; When the axial-radial gradient ratio is greater than or equal to 3, it is determined that the axial heat dissipation is blocked due to heat dissipation obstruction of the winding end or blockage of the axial ventilation duct; When the axial-radial gradient ratio is greater than 0.5 and less than 3, it is determined to be local overheating caused by uneven distribution of heat sources inside the winding; When the axial-radial gradient ratio is less than or equal to 0.5, it is determined that the inter-turn air duct heat dissipation efficiency is low due to insufficient radial convection.

3. The dry-type transformer temperature rise field simulation method according to claim 2, characterized in that: Step S1 includes the following steps: Step S11: measuring the length, width, and thickness of the core laminations in the dry-type transformer to be simulated, and recording the lamination material, the thickness of the insulation layer between the laminations, the core window size, and the core leg spacing parameters to form the core geometry parameters; Step S12: collecting the inner diameter, outer diameter, and height dimensions of the high-voltage winding and the low-voltage winding of the dry-type transformer to be simulated, measuring the winding axial position coordinates and radial spacing, recording the number of winding turns, the cross-sectional area of ​​the conductor, and the distance between the windings, and forming the winding space layout parameters; Step S13: Identify the material type and thickness parameters of the inter-winding insulation, winding-to-ground insulation, and phase-to-phase insulation based on the core geometry parameters and the winding spatial layout parameters, measure the spatial position coordinates of each insulation component, record the relative position relationship between the insulation component and the core and winding, and establish the insulation component spatial positioning data; Step S14: combining the core geometry parameters, winding spatial arrangement parameters and insulation component spatial positioning data into structural design data based on the physical contact relationship among the dry-type transformer core, winding and insulation components to be simulated.

4. The dry-type transformer temperature rise field simulation method according to claim 3, characterized in that: Constructing 3D geometric modeling data based on the structural design data in step S2 includes: A right-handed coordinate system is established with the center of the core as the origin, the axial direction as the Z axis, and the radial direction as the X / Y axis. Based on the core lamination dimensions in the structural design data, silicon steel sheets are stacked layer by layer with stepped joints in a right-hand coordinate system to obtain a core lamination model; According to the inner / outer diameter dimensions of the winding in the structural design data, the winding is stretched into a cylinder and the axial / radial airway positions are marked to obtain the winding model; Insert insulation components between the core lamination model and the winding model according to the insulation positioning coordinates in the structural design data, and verify the gap with the winding to form an initial three-dimensional model; Assign magnetic material properties, conductor material properties, and insulation material properties to the initial three-dimensional model, set the XY plane conductivity to zero, and input the Z-direction conductivity according to the measured value to obtain three-dimensional geometric modeling data.

5. The dry-type transformer temperature rise field simulation method according to claim 4, characterized in that: In step S2, the three-dimensional geometric modeling data is imported into the finite element electromagnetic simulation platform, the power-on conditions of the excitation winding are set, and the rated current or voltage is applied to obtain the local magnetic flux density distribution data at each core lamination position during the entire operation cycle, including: Import the 3D geometric modeling data into the finite element electromagnetic simulation platform in a standard format, and verify the integrity of the geometry and the absence of overlap between components to obtain a complete 3D model tree; In the complete 3D model tree, assign the BH magnetization curve of the silicon steel sheet to the core lamination, input the nonlinear parameters of the change of magnetic permeability with magnetic field strength, set the resistivity and density properties of the core material, input the resistivity, density and cross-sectional area parameters of the high-voltage winding and low-voltage winding in the complete 3D model tree, set the geometric path of the winding end connection wire, define the energization conditions of the excitation winding, apply the rated current or voltage, and collect the local magnetic flux density tensor field; The local magnetic flux density tensor field is divided into multiple instantaneous magnetic flux density vector layers according to the simulation time step, and the spatial coordinate index is marked at each time point to obtain the time-space corresponding vector layer; The time-space corresponding vector diagram layer is sliced ​​in three axes, and the magnetic flux distribution characteristics on different cross sections inside each core lamination are extracted. The average magnetic flux density and gradient change rate of each lamination layer are calculated as the local magnetic flux density distribution data.

6. The dry-type transformer temperature rise field simulation method according to claim 5, characterized in that: The definition of the energizing conditions of the excitation winding and the application of the rated current or voltage includes: For the voltage excitation method, the rated voltage amplitude and frequency parameters are applied to the end of the high-voltage winding, and the end of the low-voltage winding is set to the ground boundary. For the current excitation method, the rated current density distribution is applied to the high-voltage winding, and the induced current distribution in the low-voltage winding is automatically calculated based on the ampere-turn relationship.

7. The dry-type transformer temperature rise field simulation method according to claim 6, characterized in that: Step S3 includes the following steps: Step S31: extracting the magnetic flux density time series of each core lamination unit from the local magnetic flux density distribution data, calculating the peak value, effective value and change frequency of the magnetic flux density, and identifying the magnetic flux density components along the rolling direction and perpendicular to the rolling direction according to the rolling direction of the core material; Step S32: Calculating the hysteresis loss density of the core lamination unit within one operating cycle using the BH magnetization curve of the silicon steel sheet based on the magnetic flux density component, wherein the hysteresis loss density is equal to the product of the area of ​​the hysteresis loop and the operating frequency; Step S33: Calculating the eddy current loss density along the rolling direction and the direction perpendicular to the rolling direction based on the magnetic flux density change rate in the magnetic flux density component and the thickness parameter and resistivity parameter of the core lamination, and performing vector synthesis of the eddy current loss density components in the two directions as the eddy current loss density; Step S34: superimposing the hysteresis loss density and eddy current loss density of each core lamination unit to obtain the total core loss density of the unit, and multiplying it by the volume of the corresponding unit to calculate the core loss power value of each spatial unit; Step S35: Arrange the core loss power values ​​of all core lamination units according to the spatial coordinate index to form a loss power distribution matrix in three-dimensional space, and output core loss distribution data containing the coordinates of each spatial unit and the corresponding loss power value.

8. The dry-type transformer temperature rise field simulation method according to claim 7, characterized in that: Step S4 includes the following steps: Step S41: Importing 3D geometric modeling data into the multi-physics simulation platform and establishing a heat transfer calculation domain, wherein the heat transfer calculation domain is specifically established by setting the iron core, windings, and insulation components as solid heat transfer domains, setting the air around the transformer as a fluid heat transfer domain, and setting fluid-solid coupling boundary conditions at the interface between the solid domain and the fluid domain; Step S42: Load the core loss distribution data as a body heat source into the core lamination unit in the heat transfer calculation domain, convert the winding copper loss into a body heat source according to the current density distribution and load it into the corresponding winding unit to form a three-dimensional heat source distribution field; Step S43: Setting the fluid domain boundary conditions in the electromagnetic-thermal-fluid coupling model. For the natural air cooling condition, set the buoyancy parameters and gravity direction. For the forced air cooling condition, set the air velocity and temperature at the air inlet and the pressure boundary at the air outlet. Calculate and set the turbulence model parameters based on the Reynolds number. Step S44: setting a convection heat transfer boundary condition on the coupling interface between the solid domain and the fluid domain, and establishing an electromagnetic-thermal-fluid coupling model including heat conduction, convection heat transfer, and radiation heat transfer through the three-dimensional heat source distribution field and turbulence model parameters.

9. The dry-type transformer temperature rise field simulation method according to claim 8, characterized in that: Step S5 includes: The electromagnetic-thermal-fluid coupling model was meshed, with a refined mesh at the core lamination boundary, winding surface, and fluid-solid coupling interface to obtain a finite element mesh model with a mesh skewness less than 0.8 and an aspect ratio less than 5. The transient solution parameters are set based on the finite element mesh model, including the total simulation time, time step, and convergence criterion. The time step is determined based on the minimum mesh size and the maximum flow velocity, and the convergence criterion is set to the energy residual being less than , the momentum residual is less than ; Initial conditions are set on the nodes of the finite element mesh model. The initial temperature of the solid domain mesh nodes is set to the ambient temperature, the initial flow velocity of all fluid domain mesh nodes is set to zero, the initial heat source distribution of each unit is set according to the mapping relationship between the mesh unit and the heat source area, and temperature and heat flow continuity constraints are established on the mesh nodes of the fluid-solid coupling interface, thereby obtaining an iteratively solvable multi-physics field coupling initial state model. Based on the iteratively solvable multi-physics field coupling initial state model, the transient finite element solver is called to perform electromagnetic-thermal-fluid field coupling simulation, and the temperature distribution of each grid node in the model is iteratively calculated time step by time. The global temperature field results calculated in multiple time steps are integrated into dynamic temperature rise field data in a time series.

10. A dry-type transformer temperature rise field simulation system, characterized in that: Used to execute the dry-type transformer temperature rise field simulation method according to claim 1, the dry-type transformer temperature rise field simulation system comprises: The structural modeling preparation module is used to collect the geometric parameters of the core laminations of the dry-type transformer to be simulated, the spatial arrangement information of the windings, and the positioning relationship of the insulation components to form structural design data; The electromagnetic modeling and simulation module is used to construct three-dimensional geometric modeling data based on structural design data, import the three-dimensional geometric modeling data into the finite element electromagnetic simulation platform, set the energization conditions of the excitation winding, and apply the rated current or voltage to obtain the local magnetic flux density distribution data at each core lamination position during the entire operating cycle; The loss calculation mapping module is used to map the local magnetic flux density distribution data into spatial coordinates, calculate the eddy current loss and hysteresis loss respectively based on the directional parameters of the core material, and output the core loss distribution data for each spatial unit; The multi-field coupling modeling module is used to input core loss distribution data as a heat source, establish an electromagnetic-thermal-fluid coupling model in the multi-physics field simulation platform, set the functional relationship between thermal conductivity parameters and temperature, and set the air flow rate, temperature, and convective heat transfer coefficient under natural or forced air cooling as boundary conditions; The temperature rise field solving module is used to iteratively solve the electromagnetic-thermal-fluid coupling model using a transient finite element solver, and obtain dynamic temperature rise field data including the global temperature distribution at each time point through the solution.

Images