A transformer heat dissipation structure and a multi-physical field simulation optimization method for transformer heat dissipation

Abstract

This invention discloses a transformer heat dissipation structure and a multiphysics simulation optimization method for transformer heat dissipation. The transformer heat dissipation structure includes: an iron core, a low-voltage coil, a high-voltage coil, and a main channel disposed between the low-voltage coil and the high-voltage coil. Within the main channel, several levels of Z-shaped insulating flow guide baffles are fixed alternately from bottom to top along the axial direction. The Z-shaped insulating flow guide baffles cut off the vertical convection path of the cooling oil, forcing the cooling oil to repeatedly deflect and laterally flush the oil passages of the high-voltage coil. An insulating corner ring is provided at the end of the high-voltage coil. The mounting reference plane of the insulating corner ring is lowered, so that the physical lower edge of the insulating corner ring is flush with the physical upper edge of the uppermost Z-shaped insulating flow guide baffle in axial height, forming a thermal plume interception and guiding unit that closes the axial short-circuit gap of the fluid, thereby forcibly intercepting the rising cooling oil and deflecting it laterally along the lower surface of the insulating corner ring. This invention can effectively improve the operating efficiency of compact transformers.

Description

Technical Field

[0001] This invention relates to the field of thermal management and computer-aided engineering technology for power transformers, and in particular to a transformer heat dissipation structure and a multiphysics simulation optimization method for transformer heat dissipation. Background Technology

[0002] In compact transformer design, the phase spacing is compressed to the extreme in order to reduce the size of the equipment. This compact arrangement reduces the cross-sectional area of ​​the main oil passage, resulting in insufficient lateral scouring of the oil flow at the winding ends. The thermal boundary layer is prone to stagnation, thus forming a cooling dead zone that is difficult to dissipate heat deep inside the coil, which accelerates insulation aging.

[0003] To enhance heat dissipation within a limited space, existing technologies commonly employ Z-shaped staggered baffles within the main oil passage to extend the oil flow path and increase turbulence. However, in such conventional Z-shaped baffle layouts, an axial installation gap (typically around 12mm) is generally present between the top insulating corner ring and the uppermost stepped baffle. This gap constitutes a low-pressure short-circuit channel for the fluid, allowing a large amount of cooling oil to flow directly longitudinally through this gap without lateral backflow. As a result, the area at the top of the winding that most requires cooling lacks effective lateral scouring, the top thermal boundary layer cannot be broken down, and the temperature rise at the hottest spot is not effectively suppressed.

[0004] To accurately evaluate the heat dissipation effect of the complex oil channels, electromagnetic-thermal-fluid multiphysics coupling simulation is required. However, existing simulation processes face two typical computational difficulties when dealing with transformer models containing Z-shaped staggered baffles: transient thermal shock divergence, where the initialization method of directly applying a full-power step load at the beginning of the transient calculation has a second-order discontinuity in time, which will cause drastic nonlinear abrupt changes in the flow field and temperature field, leading to a sharp reduction in the solver time step or even divergence and deadlock; and steady-state numerical oscillation, where in steady-state thermal-fluid coupling solutions, the fluid velocity changes abruptly due to the abrupt change in geometric cross-section at the sharp corner of the Z-shaped baffle, which will induce strong nonlinear numerical oscillations, generating false high-temperature spots in the temperature field that do not exist physically, causing the residual curve oscillations to fail to converge to the steady-state equilibrium solution, and the simulation results lack credibility.

[0005] To address the aforementioned simulation challenges, some researchers have proposed alternative evaluation approaches. These include indirectly verifying the design by comparing simulated temperature images with historical fault images, or using historical simulation data to train neural network models for rapid temperature field prediction. However, these approaches attempt to circumvent rather than overcome the problem of simulation convergence itself. They do not offer any substantial improvements to the core simulation deficiencies, such as the loading method of full-power step loads or the suppression mechanism for numerical oscillations at baffle corners. Therefore, they cannot provide high-fidelity, convergent multiphysics simulation information during the design phase and cannot serve as a reliable basis for optimizing heat dissipation structures. Therefore, there is an urgent need for a solution that can be applied to transformer oil duct cooling structure with a compact phase spacing arrangement. Summary of the Invention

[0006] This invention aims to provide a transformer heat dissipation structure and a multiphysics simulation optimization method for transformer heat dissipation. The transformer heat dissipation structure of this invention improves upon existing oil channel structures where axial installation gaps between the top insulating corner ring and the uppermost baffle lead to insufficient lateral oil flow scouring at the winding ends, stagnation in the top thermal boundary layer, and excessively high hotspot temperature rise. The method of this invention, in multiphysics transient simulations of transformers with staggered baffle oil channels, avoids the problem of computational divergence and deadlock caused by nonlinear thermal shocks to the flow and temperature fields due to the direct application of a full-power step heat source. It also reduces the problem in steady-state thermal-fluid coupling solutions where geometrical abrupt changes at the Z-shaped baffle corners induce rapid velocity changes and nonlinear numerical oscillations, resulting in spurious temperature spikes and preventing residuals from converging to a steady-state equilibrium solution.

[0007] According to one aspect of the present invention, a transformer heat dissipation structure is provided, comprising: an iron core, a low-voltage coil sleeved outside the iron core, a high-voltage coil sleeved outside the low-voltage coil, and a main channel disposed between the low-voltage coil and the high-voltage coil. Within the main channel, several Z-shaped insulating flow guide baffles are fixed alternately from bottom to top along the axial direction. The Z-shaped insulating flow guide baffles cut off the vertical convection path of the cooling oil, forcing the cooling oil to repeatedly turn back and laterally flush the oil passage of the high-voltage coil. The high-voltage coil end is provided with an insulating corner ring. The mounting reference surface of the insulating corner ring is moved down so that the physical lower edge of the insulating corner ring is flush with the physical upper edge of the uppermost Z-shaped insulating guide baffle in axial height. The insulating corner ring and the uppermost Z-shaped insulating guide baffle together form a thermal plume interception and guide unit that closes the axial short-circuit gap of the fluid, so as to forcibly intercept the rising cooling oil and fold it back laterally along the lower surface of the insulating corner ring.

[0008] According to some embodiments, the high-voltage coil adopts a segmented cylindrical structure, with horizontal transverse oil channels formed between the coils; The Z-shaped insulating baffle forces the cooling oil to deflect along an S-shaped path and laterally flush the transverse oil passage.

[0009] According to some embodiments, the transformer adopts a phase spacing of 600mm~700mm, and the radial spacing of the Z-shaped insulating current guide baffles is adapted to the total width of the main channel under the phase spacing.

[0010] According to some embodiments, each Z-shaped insulating baffle includes a Z-shaped bend formed by connecting a vertical section and a horizontal section, wherein the angle of the Z-shaped bend ranges from 85° to 95°. The horizontal sections of the Z-shaped insulating baffles in adjacent stages extend in opposite directions. The radial distance between the Z-shaped insulating guide baffle and the main insulation of the coil is 20% to 35% of the total width of the main channel.

[0011] According to some embodiments, the downward shift of the mounting reference plane of the insulating corner ring further includes: The lower physical edge of the insulating corner ring is lower than the upper physical edge of the uppermost Z-shaped insulating baffle, and the two partially overlap in axial height by 0 to 5 mm; and / or, The insulating corner ring has an arc-shaped cross-section to accommodate the electric field distribution at the winding ends.

[0012] According to another aspect of the present invention, a multiphysics simulation optimization method for transformer heat dissipation is provided, comprising: Structural parameters of the transformer under tight phase spacing are extracted, and a three-dimensional geometric model is established, including high and low voltage coils, Z-shaped insulating baffles, and insulating corner rings flush with the uppermost Z-shaped insulating baffles. The inner cavity of the transformer tank is set as a complete closed fluid domain to construct a natural convection physical environment driven by real buoyancy force. Non-uniform loss mapping loading is performed, and three-dimensional loss density distribution data including winding skin effect and end eddy current effect are calculated in the electromagnetic field frequency domain solver. The loss values ​​on the electromagnetic field grid nodes are mapped to the corresponding spatial location nodes of the thermal-fluid coupling model through element shape function interpolation, which serves as a non-uniform body heat source. Mesh the fluid and solid domains. In the transient solution phase, an exponential soft-start function is applied to the non-uniform heat source to smoothly pass the load ramp-up period and eliminate the transient thermal shock caused by the full power step. In the steady-state solution phase, for the abrupt change in flow velocity caused by the sudden change of the cooling fluid at the corner of the Z-shaped insulating guide baffle, a joint solution is performed using local mesh refinement and pseudo-transient time step strategy. By using the converged 3D temperature cloud map and flow field velocity vector map, the effect of the thermal plume interception and guiding unit on suppressing the temperature of the local hot spot and improving the lateral scouring velocity is verified, and a heat dissipation structure scheme is output.

[0013] According to some embodiments, mapping the loss values ​​on the electromagnetic field grid nodes to the corresponding spatial location nodes of the thermal-fluid coupling model through element shape function interpolation includes: By using element shape function interpolation mapping, the loss values ​​on the electromagnetic field grid nodes calculated by the electromagnetic field frequency domain solver are mapped to the corresponding spatial location nodes of the thermal-fluid coupling model.

[0014] According to some embodiments, an exponential soft-start function is applied to the non-uniform heat source, wherein the exponential soft-start function is:

[0015] Where P(t) is the actual applied heating power at time t. The rated heating power is τ, which is a time constant used to eliminate solver divergence deadlock caused by step loading of heating power. The value range is 10s~60s.

[0016] According to some embodiments, the joint solution using local mesh refinement combined with a pseudo-transient time-stepping strategy includes: Static meshing was performed at each level on all corners of the Z-shaped insulating flow guide baffle and the adjacent oil passage walls. In the steady-state solution of the discrete control equations, an inertial damping term with an adaptive virtual time step is introduced. By adaptively adjusting the virtual time step according to the local residual gradient, the nonlinear numerical oscillations caused by abrupt changes in the geometric cross-section are smoothed out, thereby eliminating spurious temperature spikes in the flow field. The discrete form expression is as follows:

[0017] Where Δτ is the virtual time step in seconds (s), which is adaptively adjusted by the solver based on the gradient magnitude of the local residuals at each grid node. High-oscillation regions are automatically reduced to increase damping, while smooth regions are gradually expanded to accelerate convergence; ρ is the density of the insulating oil in kg / m³. 3 Δu is the velocity vector difference between the current iteration step and the previous iteration step, in m / s; is the gradient operator; u is the three-dimensional velocity vector of the insulating oil, in m / s; p is the hydrostatic pressure of the fluid element, in Pa; μ is the dynamic viscosity of the insulating oil, in Pa. g is the gravitational acceleration constant; β is the volumetric expansion coefficient of transformer oil, in units of... T represents the real-time temperature field distribution within the fluid domain, expressed in Kelvin (K). ref For reference temperature, the initial ambient temperature of the transformer oil or the average oil temperature under rated operating conditions is taken, and the unit is K.

[0018] According to another aspect of the present invention, a computing device is provided, including a processor; and a memory storing a computer program that, when executed by the processor, implements the method as described in any of the preceding claims.

[0019] The technical solutions provided by the embodiments of the present invention have the following beneficial effects: The transformer heat dissipation structure of this invention, through the flush connection of the insulating corner ring and the uppermost Z-shaped insulating guide baffle, together constitutes a "thermal plume interception and guiding unit" that seals the axial short-circuit gap of the fluid. This physically blocks the 12mm axial installation gap present in conventional designs, forcing the originally linearly rising cooling oil to deflect at the top region, folding back laterally along the lower surface of the insulating corner ring, thereby directionally flushing the hottest spots at the winding ends and completely breaking the stagnant top thermal boundary layer. The Z-shaped insulating guide baffles, arranged axially in a staggered pattern within the main channel, force the cooling oil to fold back multiple times along the path, ensuring sufficient contact between the cooling oil and the winding surface and improving the convective heat transfer coefficient. This structure achieves efficient heat dissipation without significantly increasing the transformer volume. By precisely controlling the radial distance between the baffle and the main coil insulation (designed according to 20%~35% of the total width of the main channel), it maintains good heat dissipation adaptability within a compact phase spacing range of 600mm~700mm, solving the problem of limited heat dissipation space in miniaturized transformers.

[0020] The simulation method of this invention mathematically eliminates the second-order discontinuity caused by full-power step loading by introducing an exponential soft-start heat source strategy in the transient solution stage. This allows the heating power to gradually climb to the rated value along a smooth curve, avoiding time step collapse caused by drastic changes in the flow and temperature fields. Testing has shown that this method enables transient calculations to smoothly pass through the load ramp-up period within 30 seconds, eliminating the divergence problem that is easily caused by conventional step loading.

[0021] To address the nonlinear numerical oscillations induced by geometric abrupt changes at the corners of Z-shaped insulated flow guide baffles, a joint convergence mechanism combining local mesh refinement with pseudo-transient time stepping is employed. By introducing an inertial damping term with an adaptive virtual time step into the discrete control equations, numerical fluctuations caused by abrupt changes in flow velocity are effectively mitigated. This eliminates physically non-existent spurious high-temperature spots, ensuring the physical reliability of the temperature field distribution results.

[0022] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0023] Figure 1 This is a three-dimensional structural diagram of the transformer heat dissipation structure in an example embodiment of the present invention.

[0024] Figure 2 This is a cross-sectional schematic diagram of the transformer heat dissipation structure according to an example embodiment.

[0025] Figure 3 is a partial cross-sectional three-dimensional schematic diagram of the arrangement position of the Z-type insulating current-conducting baffle of the transformer in an embodiment of the present invention.

[0026] Figure 4 is a half-sectional three-dimensional schematic diagram of the oil passage and oil flow direction in an embodiment of the present invention.

[0027] Figure 5 This is a flowchart illustrating a multiphysics simulation optimization method according to an example embodiment of the present invention.

[0028] Figure 6 This is a multi-faceted velocity distribution cloud map of the internal flow field of the transformer in an embodiment of the present invention.

[0029] Figure 7 This is a cross-sectional view of the internal temperature field distribution of a transformer (comparative example) using a conventional installation gap structure.

[0030] Figure 8 This is a cross-sectional view of the internal temperature field distribution of a transformer using the flush connection structure of the corner ring-baffle according to an embodiment of the present invention.

[0031] Figure 9 A block diagram of a computing device according to an exemplary embodiment of the present invention is shown. Detailed Implementation

[0032] To further understand the content of this invention, a detailed description of the invention is provided in conjunction with the accompanying drawings and embodiments. The invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only for explaining the relevant invention and are not intended to limit the invention. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0033] like Figure 1 As shown, this embodiment proposes a transformer heat dissipation structure.

[0034] According to an example embodiment, the transformer heat dissipation structure includes an iron core 01, a low-voltage coil 02 sleeved outside the iron core 01, a high-voltage coil 03 sleeved outside the low-voltage coil 02, and a main channel disposed between the low-voltage coil 02 and the high-voltage coil 03.

[0035] Within the main channel, several levels of Z-shaped insulating baffles 04 are fixed axially from bottom to top in an alternating pattern. These Z-shaped insulating baffles 04 interrupt the vertical convection path of the cooling oil, forcing the cooling oil to repeatedly deflect and laterally flush the oil channels of the high-voltage coil. See the cross-sectional view of the transformer's heat dissipation structure. Figure 2 .

[0036] According to some embodiments, the high-voltage coil 03 is provided with an insulating corner ring 05 at its end. The insulating corner ring 05 is an annular insulating corner ring (with an L-shaped or oblique cross-section). The mounting reference plane of the insulating corner ring 05 is lowered so that the lower edge of the insulating corner ring 05 is flush with the upper edge of the uppermost Z-shaped insulating guide baffle 04 in axial height. The insulating corner ring 05 and the uppermost Z-shaped insulating guide baffle 04 together constitute a thermal plume interception and guide unit 06 that closes the axial short-circuit gap of the fluid, so as to forcibly intercept the rising cooling oil and fold it back laterally along the lower surface of the insulating corner ring. The spiral cooling oil channel and oil flow direction of the transformer coil are as follows. Figure 3 As shown. Specifically, the Z-shaped insulating baffle 04 is arranged in a ring within the main channel, see [reference]. Figure 4 The Z-shaped insulating flow guide baffles 04 are arranged in layers and staggered along the coil axis, with the radial extension directions of adjacent baffles alternating in opposite directions, forming a three-dimensional folded cooling channel in the main channel, so that the cooling oil can be transversely introduced into the transverse oil channel of the high-voltage coil, thereby realizing layered heat dissipation of the coil.

[0037] Furthermore, the thermal plume interception and diversion unit 06 is composed of an insulating corner ring 05 installed downward and an uppermost Z-shaped insulating diversion baffle 04 axially flush. The lower edge of the insulating corner ring 05 and the upper edge of the uppermost Z-shaped insulating diversion baffle 04 form a continuous axial blocking interface on the radial section of the main channel, and there is no axially penetrating fluid short-circuit gap at the joint. When the transformer is running, the heat generated by the coil layers of the high-voltage coil 03 reduces the heat density of the surrounding cooling oil, forming a hot plume that rises axially along the main channel. When this hot plume reaches the top of the main channel, it is intercepted by the hot plume interception and guiding unit 06 and cannot escape directly through the axial gap. Instead, it is forced to fold back laterally along the lower surface of the insulating corner ring 05 towards the high-voltage coil 03 side, and then enter the uppermost transverse oil channel to participate in heat exchange. This eliminates the axial short circuit phenomenon of the cooling oil at the top of the main channel, ensures the cooling oil flow and heat exchange efficiency in the uppermost coil area, and improves the uniformity of the coil's axial temperature rise.

[0038] According to some embodiments, the high-voltage coil 03 adopts a segmented cylindrical structure, with horizontal transverse oil channels formed between the coil coils. Specifically, the high-voltage coil 03 is composed of multiple coil coils (i.e., disc-shaped coil segments wound with flat or round copper wire) stacked sequentially along the axial direction. To ensure insulation and heat dissipation, adjacent coil coils are separated by insulating blocks, thereby forming horizontal transverse oil channels between adjacent coil coils. When the cooling oil flows through the main channel, it can penetrate deep into the coil through the transverse oil channels, carrying away the heat generated by the coil coil conductors.

[0039] The Z-shaped insulating baffle 04 forces the cooling oil to deflect along the S-shaped path 041 and laterally flush the transverse oil passage. This "S-shaped deflection path" significantly extends the residence time of the cooling oil in the area of ​​concentrated heat load and enhances fluid turbulence, preventing localized overheating, ensuring sufficient contact between the cooling oil and the winding surface, and improving the convective heat transfer coefficient. By setting the Z-shaped insulating baffle 04, this invention effectively disrupts the fluid thermal boundary layer within the oil passage, significantly improving the convective heat transfer coefficient of the coil surface.

[0040] In this embodiment, the transformer adopts a phase spacing of 600mm~700mm. The radial spacing of the Z-shaped insulating guide baffles is adapted to the total width of the main channel under the phase spacing. Each stage of the Z-shaped insulating guide baffle includes a Z-shaped bend formed by connecting a vertical section and a horizontal section. The angle range of the Z-shaped bend is 85° to 95°, and more preferably a 90° right angle structure. The 90° right angle structure of the Z-shaped bend not only facilitates the processing and manufacturing of the insulating baffle, but also forms a stable fluid separation point at the bend, guiding the cooling oil to change direction more smoothly. The horizontal extension directions of the Z-shaped insulating guide baffles of adjacent stages are alternately opposite. The radial spacing between the Z-shaped insulating guide baffle and the main insulation of the coil is 20% to 35% of the total width of the main channel. If this proportion is less than 20%, it may lead to the oil flow channel being too narrow, increasing the flow resistance; if it is greater than 35%, it may lead to the cooling oil short circuit, failing to fully flush the coil surface.

[0041] In an alternative embodiment, the lower physical edge of the insulating corner ring is lower than the upper physical edge of the uppermost Z-shaped insulating baffle, and the two partially overlap in axial height by 0 to 5 mm, achieving the same interception and conduction effect. The cross-sectional shape of the insulating corner ring can be arc-shaped to adapt to the electric field distribution at the winding end.

[0042] In this embodiment, the transformer heat dissipation structure precisely lowers the mounting reference plane of the annular insulating corner ring at the top of the winding, ensuring that its lower edge is flush with the upper edge of the top Z-shaped insulating baffle in axial height. The axial assembly tolerance of both is controlled within ±2mm. Thus, the insulating corner ring and the Z-shaped insulating baffle together form a physical "thermal plume interception and guiding unit," completely sealing the fluid short-circuit gap. The cold oil rising from the bottom is forcibly intercepted in the top region and laterally folded back along the lower surface of the corner ring, directionally and laterally flushing the hottest spot at the coil end, thereby breaking the stagnant top thermal boundary layer and eliminating the heat dissipation dead zone.

[0043] By physically aligning the lower edge of the insulating corner ring 05 with the uppermost Z-shaped insulating guide baffle 04 (tolerance ±2mm, or 0~5mm overlap), the axial gap is geometrically sealed, preventing the rising hot plume from having a direct path at the top. It is forced to turn 90° along the lower surface of the insulating corner ring 05, creating a lateral pressure gradient. This pressure gradient forces the plume into the horizontal lateral oil channel at the end of the high-voltage coil 03, directionally scouring the hottest spot on the coil. Simultaneously, the lower Z-shaped insulating guide baffle 04 causes the oil flow to fold back in an S-shape and pre-accelerate it. The lateral foldback flow at the top overlaps with the mainstream S-shape at the end region, completely breaking through the stagnant thermal boundary layer. The hot plume interception and guide unit in this embodiment is essentially a passive guide structure that seals gaps and forcibly changes flow direction. It requires no external pump and relies solely on buoyancy and natural convection to function. Unlike conventional structures, where the high-voltage coil end area forms an upward-accumulating heat plume (low-density high-temperature oil column) due to concentrated losses, this heat plume is usually lost through a direct longitudinal short circuit via the axial gap between the corner ring and the uppermost baffle.

[0044] like Figure 5 As shown, this invention proposes a multiphysics simulation optimization method for transformer heat dissipation.

[0045] In S101, the structural parameters of the transformer under a compact phase spacing are extracted, and a three-dimensional geometric model is established, including high and low voltage coils, Z-shaped insulating guide baffles, and an insulating corner ring flush with the uppermost Z-shaped insulating guide baffle. The inner cavity of the transformer tank is set as a complete closed fluid domain to construct a natural convection physical environment driven by real buoyancy force.

[0046] According to some embodiments, structural parameters of the transformer are extracted under a compact phase spacing, preferably 600mm~700mm. In this embodiment, a phase spacing of 625mm is used as the specific implementation dimension. Using the geometric module built into 3D CAD software or simulation software, a 3D geometric model including the core, low-voltage coil, and high-voltage coil is established.

[0047] In the model construction, the main channel and end structure are handled. The high-voltage coil adopts a segmented cylindrical structure, clearly defining the horizontal transverse oil channels between the coil coils. Within the main channel, 3 to 6 levels of staggered and fixed Z-shaped insulating baffles are constructed axially from bottom to top, with the Z-shaped angle set to 85° to 95° (preferably 90°). Alternating reverse extensions and radial overlap areas are set for the horizontal sections of adjacent baffles. The horizontal sections of adjacent baffles extend in opposite directions, causing the oil flow to form multiple S-shaped return paths axially. The radial distance between the baffles and the main insulation of the coil is designed to be 20% to 35% of the total width of the oil channel to ensure that the oil flow fully contacts the winding surface.

[0048] An annular insulating corner ring is constructed at the end of the high-voltage coil. Through Boolean operations or assembly constraints, its physical lower edge is made flush with the physical upper edge of the uppermost Z-shaped insulating guide baffle in axial height (or forms an axial overlap of no more than 5mm), thereby sealing the axial short-circuit gap of the fluid at the geometric level and constructing a three-dimensional solid model of the thermal plume interception and guide unit.

[0049] This simulation method employs a natural convection physical environment driven by real buoyancy, without relying on artificially set forced inlet velocity or outlet pressure boundary conditions, treating the transformer tank cavity as a complete closed fluid domain. Fluid movement is driven solely by the density change of the transformer oil itself due to temperature differences. The coupled calculation of non-isothermal flow and heat transfer follows the Bousinesko approximation, which ignores the velocity term in the solid domain, degenerating into pure heat conduction, and considers the buoyancy term driving the flow field in the fluid domain, thus realistically reproducing the thermally induced flow nature inside the transformer.

[0050] According to the energy conservation equations for solids and fluids, the temperature field distribution T inside the transformer is determined by both heat conduction and heat convection, satisfying the following heat transfer equation:

[0051] In the formula, C is the material density. p Let u be the constant pressure heat capacity, u be the flow velocity vector, and k be the thermal conductivity. Let be the gradient operator and Q be the heat source. In solid domains (such as iron cores or baffles), since the flow velocity u = 0, the convection term on the left side of the equation is zero, and it automatically degenerates into a pure heat conduction equation.

[0052] According to the fluid mass conservation equation, the flow of insulating oil satisfies the continuity requirement of incompressible (or weakly compressible) fluids throughout the entire flow and heat dissipation process:

[0053] According to the buoyancy-driven and Businsk approximation equations, after the insulating oil is heated by the aforementioned heat source Q, the local temperature increase leads to a decrease in density, thereby generating a buoyancy-driven volume force F that drives the oil flow upward. Its density change is calculated using the Businsk approximation formula:

[0054] The corresponding buoyancy driving equation is expressed as:

[0055] in, The density of the insulating oil is given by the reference temperature T0, and g is the acceleration due to gravity. Let T be the volumetric expansion coefficient of the transformer oil, and T be the local real-time temperature.

[0056] This invention realistically reproduces the natural convection nature of "thermally induced buoyancy drive" inside the transformer at the physical field level, avoiding the calculation errors caused by the artificial assumption of inlet flow velocity in traditional simulations, and laying a physical foundation for subsequent high-precision multi-physics coupling solutions.

[0057] In S103, non-uniform loss mapping loading is performed. The three-dimensional loss density distribution data, including the winding skin effect and end eddy current effect, is calculated in the electromagnetic field frequency domain solver. The loss values ​​on the electromagnetic field grid nodes are mapped to the corresponding spatial location nodes of the thermal-fluid coupling model through element shape function interpolation, serving as a non-uniform body heat source.

[0058] According to some embodiments, the heat generation of transformer windings is not uniformly distributed due to significant skin effect, proximity effect, and eddy current effect caused by end leakage flux under alternating magnetic fields. In the electromagnetic field frequency domain solver, a rated current excitation is applied to calculate the three-dimensional loss density distribution data, which includes the aforementioned effects. Through element shape function interpolation mapping, the loss values ​​on the electromagnetic field mesh nodes calculated by the electromagnetic field frequency domain solver are mapped to the corresponding spatial location nodes of the thermal-fluid coupling model.

[0059] Specifically, in the electromagnetic field frequency domain solver, three-dimensional loss density distribution data, including winding skin effect and end eddy current effect, are calculated. The loss values ​​at the electromagnetic field mesh nodes are mapped to the corresponding spatial location nodes of the heat-fluid model using the element shape function interpolation formula. The specific formula is as follows:

[0060] in, This represents the mapping loss value at the i-th grid node in the heat-flow model. Let j be the local coordinates of the j-th shape function of the electromagnetic field element at the mapping point. The value below, n is the total number of nodes in the electromagnetic field element.

[0061] In this embodiment, the total internal heat source Q is composed of the Joule loss (copper loss) of the coil and the alternating magnetic loss (iron loss) generated by the laminated iron core, satisfying the electromagnetic loss internal heat source coupling equation:

[0062] Among them, Q cu Q is the coil heat generation rate based on current density. fe The losses are due to eddy currents and hysteresis generated in the iron core. The heat source Q is directly used as an input term in the energy conservation equation.

[0063] This embodiment provides a method to traverse each target node in the heat-fluid coupling model, locate its corresponding electromagnetic field grid cell, transform the global coordinates of the target node to the local coordinate system of that electromagnetic grid cell, calculate the isoparametric shape function weights of the node within the cell, and then perform a weighted summation of the loss values ​​of all nodes in that electromagnetic grid cell based on these weights to obtain the non-uniform volumetric heat source value of the target node. This mapping method avoids the heat smoothing phenomenon caused by the traditional volume averaging method and achieves high-precision, faithful mapping of the spatial distribution of losses.

[0064] In S105, the fluid domain and solid domain are meshed. In the transient solution stage, an exponential soft-start function is applied to the non-uniform heat source to smoothly pass the load ramp-up period and eliminate the transient thermal shock caused by the full power step. In the steady-state solution stage, for the rapid change in flow velocity caused by the sudden change at the corner of the Z-shaped insulating guide baffle, a joint solution is performed using local mesh refinement and pseudo-transient time step strategy.

[0065] According to some embodiments, in the multiphysics transient coupling solution setup, a smoothing time function is applied to the global volume heat source to simulate the soft-start process of the transformer, and an exponential soft-start function is applied to the non-uniform volume heat source. The exponential soft-start function is as follows:

[0066] Where P(t) is the actual applied heating power at time t. τ is the rated heating power, and τ is a time constant used to eliminate solver divergence deadlock caused by step loading of heating power. The value range is 10s~60s, preferably 30s.

[0067] The function's first derivative is continuous and its second derivative is bounded at t=0, thus mathematically eliminating the second-order discontinuity of the step load. The heating power gradually increases from zero to its rated value along a smooth exponential curve, causing the flow field and temperature field to be approximately... Within a time window of ×3, the system smoothly transitions to rated operating conditions, fundamentally avoiding time step collapse and computational divergence deadlock caused by sudden load changes in the early stages of transient solutions.

[0068] According to some embodiments, based on the joint convergence mechanism of local mesh refinement and pseudo-transient damping, to address the problem in steady-state solutions where the flow velocity changes abruptly due to abrupt changes in geometric cross-section at the corners of the Z-shaped insulated flow guide baffle, resulting in strong nonlinear numerical oscillations and false temperature spikes, a joint solution is performed using local mesh refinement combined with a pseudo-transient time stepping strategy. The static mesh is gradually refined at all corners of the Z-shaped insulated flow guide baffle and the adjacent oil passage walls. When solving the discrete control equations in steady-state, an inertial damping term with an adaptive virtual time step is introduced. By adaptively adjusting the virtual time step according to the local residual gradient, the nonlinear numerical oscillations caused by abrupt changes in geometric cross-section are smoothed out, thereby eliminating false temperature spikes in the flow field.

[0069] Specifically, local mesh refinement is first performed. For high-gradient areas such as all corners of the baffle and adjacent oil passage walls, at least two levels of static mesh refinement are pre-processed. At the fluid-solid interface, a first boundary layer mesh with a thickness limit controlled at approximately 0.5 mm is generated. The mesh size of each refinement layer is gradually reduced by about 1 / 3 of the previous layer to accurately analyze local drastic velocity and temperature gradient changes.

[0070] Secondly, enabling pseudo-transient solving technology in the steady-state solver introduces adaptive virtual time steps into the discretized control equations. The inertial damping term, in discrete form, is expressed as:

[0071] Where Δτ is the virtual time step in seconds (s), which is adaptively adjusted by the solver based on the gradient magnitude of the local residuals at each grid node. High-oscillation regions are automatically reduced to increase damping, while smooth regions are gradually expanded to accelerate convergence; ρ is the density of the insulating oil in kg / m³. 3 Δu is the velocity vector difference between the current iteration step and the previous iteration step, in m / s; Here, represents the gradient operator; u is the three-dimensional velocity vector of the insulating oil, in m / s; (ρuu) represents the convection tensor, where ρ is the fluid density. In fluid mechanics, the standard vector form of the convection term in the momentum equation is... For ease of discrete calculation, it is written in divergence form. .

[0072] p is the static pressure of the fluid element, in Pa; μ is the dynamic viscosity of the insulating oil, in Pa. g is the gravitational acceleration constant; β is the volumetric expansion coefficient of transformer oil, in units of... T represents the real-time temperature field distribution within the fluid domain, in K; Tref represents the reference temperature, taken as the initial ambient temperature of the transformer oil or the average oil temperature under rated operating conditions, in K.

[0073] In this embodiment, the damping term gradually smooths out the numerical oscillations caused by abrupt changes in the geometric cross-section, forcing the temperature residual norm to decrease monotonically and eventually converge to a certain value. The following method completely eliminates spurious temperature spikes that do not exist physically, mathematically.

[0074] In S107, the three-dimensional temperature cloud map and flow field velocity vector map obtained by solving are used to verify the effect of the thermal plume interception and guiding unit on suppressing the temperature of the local hot spot and the effect of improving the lateral scouring velocity, and output the heat dissipation structure scheme.

[0075] According to some embodiments, the physical field distribution cloud map after the flow field and temperature field reach a steady state, as well as the residual convergence history curve of key variables, are extracted to verify the heat dissipation advantages and computational reliability of the present invention.

[0076] Specifically, after the solution converges, the residual convergence curve is extracted. This is done by observing whether the residual curve smoothly decreases to a set threshold (e.g., ...). (The following) and without high-frequency spikes, the effect of the pseudo-transient strategy on the removal of false temperature spikes is verified, and a three-dimensional temperature field cloud map and flow field velocity vector map are generated.

[0077] In the flow field velocity vector diagram, verify whether the cooling oil is forced to turn back along an S-shaped path by the Z-shaped insulating guide baffle, and observe whether the top hot plume is successfully intercepted by the insulating corner ring and turns back laterally along its lower surface, thus verifying the effect of increasing the lateral scouring velocity. In the temperature field contour diagram, extract the temperature of the hottest spot in the winding (usually located in the top end region of the winding) to verify the effect of the hot plume interception and guide unit on suppressing the local hottest spot temperature.

[0078] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and prior art. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0079] Example 1: Comprehensive Verification of Structure and Simulation Methods Based on Limiting Phase Spacing This embodiment provides a 35kV compact transformer with a phase spacing strictly limited to 625mm, which is designed and verified using the heat dissipation structure and multiphysics simulation optimization method described in this invention.

[0080] A 3D simulation model of a 35kV transformer with a phase spacing strictly limited to 625mm was established. Four levels of Z-shaped stepped staggered baffles were installed axially within the main oil channel between the high-voltage and low-voltage coils. Each baffle had a 90° bend angle, and the radial distance between the baffles and the main coil insulation was designed to be 25% of the total width of the oil channel. The top insulating corner ring adopted an L-shaped cross-section, with its lower edge flush with the upper edge of the top baffle in axial height. The axial assembly tolerance was controlled within ±1mm, forming a thermal plume interception and guiding unit. During the mesh generation stage, approximately 2.24 million unstructured meshes were generated across the entire fluid-solid coupling computational domain. Boundary layer meshes were generated at the fluid-solid interface, and two levels of static local mesh refinement were performed at all corners of the Z-shaped insulating guiding baffles.

[0081] Simulation process: First, the electromagnetic field is solved in the frequency domain to obtain a three-dimensional non-uniform loss distribution including winding skin effect and end effect. This distribution is then mapped to the mesh nodes of the thermal-fluid coupling model as a volume heat source through element shape function interpolation. In the multiphysics transient coupling solution stage, an exponential soft-start function is applied to the global non-uniform volume heat source. That is, the time constant τ = 30s.

[0082] The outer wall of the tank is insulated, while the interior is completely sealed for natural convection. Density changes are calculated using the Boussinesq approximation. After reaching rated operating conditions, a pseudo-transient solver is activated during the steady-state solution phase, with an adaptive virtual time step set. Following solution initiation, the system smoothly starts oscillating within 30 seconds under the soft-start function, without any computational divergence. Upon reaching rated operating conditions, under the damping effect of the joint convergence strategy, no spurious temperature spikes appear at the corner of the Z-shaped insulated baffle, and the temperature residual norm monotonically decreases and stably converges to [value missing]. .

[0083] See the multi-faceted velocity distribution cloud map of the internal flow field of the transformer using the heat dissipation structure and simulation method of this invention. Figure 6 Simulation results show that the temperature rise of the hottest spot of the winding under this structure is 62.5K, and the peak value of the transverse oil flow velocity at the top reaches 0.052 m / s.

[0084] Example 2: Verification of structural suitability under different phase spacing ranges To verify whether the improved heat dissipation structure described in this invention has universality for different phase spacing designs, this embodiment establishes three-dimensional models of two transformers of the same voltage level (35kV) with phase spacing of 650mm and 700mm, respectively.

[0085] The internal Z-shaped insulating baffle structure, the flush connection of the corner rings, and the grid division strategy are all consistent with those in Example 1. The radial distance between the baffle and the main insulation of the coil scales proportionally with the width of the oil channel.

[0086] The simulation strategy is exactly the same as in Example 1, employing a combined convergence mechanism of electromagnetic-thermal non-uniform loss mapping, an exponential soft-start function (τ=30s), local mesh refinement, and pseudo-transient damping. The tank walls are all adiabatic boundaries, with the interior utilizing buoyancy-driven natural convection.

[0087] 650mm phase spacing model: After simulation convergence, the temperature rise of the winding hot spot is 46.7K. Compared with the empirically estimated temperature rise (approximately 52K) using conventional alignment for the same size, the temperature rise is reduced by approximately 5.3K.

[0088] 700mm phase spacing model: After simulation convergence, the temperature rise of the winding hot spot is 38.2K. Compared with the empirically estimated temperature rise (approximately 42K) using conventional alignment for the same size, the temperature rise is reduced by approximately 3.8K.

[0089] The results show that although the heat dissipation conditions of the windings are improved as the phase spacing increases, the "angle ring-baffle flush connection structure" described in this invention still brings additional and stable temperature rise benefits, proving that this structural improvement has universal engineering application value for compact transformers of different sizes and phase spacings.

[0090] Comparative Example 1: Comparative Verification of a Structure Without Top Corner Ring Alignment (Conventional Structure with Gaps) To highlight the necessity of the "thermal plume interception and guidance unit" and the "pseudo-transient time stepping strategy" in this invention, a set of control experiments was set up in this comparative example.

[0091] like Figure 7 The figure shows a cross-sectional view of the internal temperature field distribution of a transformer (Comparative Example 1) using a conventional installation gap structure after reaching steady-state operation. It can be clearly observed from the figure that, due to the 12mm axial short-circuit gap between the top insulating corner ring and the uppermost Z-shaped insulating baffle, the high-temperature cooling oil that should have entered the transverse oil passage of the coil for flushing is directly lost longitudinally through this low-resistance flow channel.

[0092] In terms of thermal characteristics, the following are observed: First, a severe high-temperature dead zone (near region B) appears. The inner end region at the very top of the left and right phases and the middle phase high-voltage coil exhibits a deep red, highly heated state, with the temperature of the local hottest spots soaring. This indicates severe stagnation in the fluid thermal boundary layer at the top. Second, a significant heat plume accumulation phenomenon occurs at the top (near region A in the figure). Due to the lack of active interception, a large amount of heat is directly ejected upwards with the short-circuit fluid, forming a large-area, continuous, laterally penetrating, red-orange high-temperature field region in the upper yoke and the fluid domain above it. This extremely uneven temperature field distribution and severe local overheating directly demonstrate the heat dissipation disadvantage of conventional structures with tight phase spacing.

[0093] The same 3D model of a 35kV transformer with a 625mm phase spacing as Example 1, along with the same number of baffle levels and mesh generation strategy (approximately 2.24 million meshes), was used. The only difference was the installation position of the top insulating corner ring. In Comparative Example 1, the corner ring maintained a conventional installation height, and there was a 12mm axial gap between its lower edge and the upper edge of the top Z-shaped insulating current-guiding baffle, forming a fluid short-circuit channel.

[0094] The simulation strategy is exactly the same as in Example 1: a combined convergence mechanism of non-uniform loss mapping loading, exponential soft-start function, local mesh refinement, and pseudo-transient damping is adopted. The tank wall uses an adiabatic boundary, and the interior is completely natural convection. That is, the two models differ only in structure, and the simulation conditions are completely identical.

[0095] A cross-sectional view of the internal temperature field distribution of a transformer with a conventional installation gap structure (Comparative Example 1) shows that due to the 12mm axial gap, a large amount of rising cooling oil does not laterally flush the oil passages at the top of the coil, but instead flows out directly through the gap short circuit. Simulation results show that a significant fluid stagnation zone appears at the winding ends, and the temperature rise of the hottest spot in the winding reaches 70.8K (8.3K higher than 62.5K in Example 1).

[0096] Figure 8 This is a cross-sectional view of the internal temperature field distribution of a transformer employing the flush-joint corner ring-baffle structure described in this invention under the same rated operating conditions. Figure 7 In comparison, this invention constructs a "thermal plume interception and guiding unit" that seals the fluid short-circuit gap by lowering the mounting reference surface of the insulating corner ring. Its temperature field cloud map shows an absolute advantage in heat dissipation optimization both visually and in terms of data. The thermal simulation diagram clearly shows that the high-temperature heat source region (region D) inside the coil is significantly suppressed and fragmented. There are no large, connected, dark red overheated spots in the diagram; the overall temperature field of the coil transforms into a uniform, gentle gradient of blue-green and pale yellow (relative to...). Figure 7 In region B of the winding, the temperature rise at the hottest spot was precisely and directionally suppressed, and the absolute value of the temperature rise was significantly reduced, fully verifying the physical fact that the thermal boundary layer was completely broken down. The phenomenon of concentrated heat plume at the top (near region C) was fundamentally reversed, and the red-orange high-temperature cloud in the upper yoke and the fluid domain above (compared to) Figure 7 In area A of the medium, the color is changed to a lower-grade, low-temperature blending color. This strongly demonstrates that the rising cooling oil is successfully forced to undergo lateral backflush at the top, achieving a virtuous cycle of comprehensive thermal management.

[0097] Simulation results from Example 1 and Comparative Example 1 verify that, in a compact transformer with a phase spacing of 625mm, the flush-joint corner ring-baffle structure of this invention can effectively eliminate steady-state false temperature spikes. Compared with the conventional Z-type insulating baffle scheme with a 12mm axial gap under the same operating conditions, the conventional scheme results in stagnation of the thermal boundary layer at the top of the winding, forming a significant high-temperature dead zone, with the highest oil temperature rise stabilizing at 70.8K. The present invention constructs a thermal plume interception-lateral deflection mechanism, forcibly guiding the cold oil to the hottest area at the winding end for targeted scouring. The highest oil temperature at the same location drops to 62.5K, and the absolute value of the hot spot temperature rise decreases by 8.3K, a reduction of 11.7%. Simultaneously, the peak lateral velocity at the top of the winding increases from the order of 0.02m / s in the conventional scheme to over 0.05m / s, significantly enhancing the lateral scouring capability. This is the direct hydrodynamic basis for the temperature rise reduction.

[0098] The method of this invention significantly improves the stability of transient solutions. By adopting an exponential soft-start heat source strategy, the transient calculation smoothly passes through the load ramp-up period within 30 seconds, without any step size collapse or divergence deadlock. Compared with the conventional full-power step loading method (which usually reports divergence errors within 5 seconds), this method mathematically eliminates the transient thermal shock divergence problem.

[0099] In steady-state thermo-fluid coupling solutions with Z-shaped insulating baffle corners, without applying the joint convergence mechanism of this invention, the residual curves at... The magnitude of the temperature residual norm remained oscillating without decreasing, and isolated high-temperature spots (false temperature spikes) that did not physically exist repeatedly appeared behind the corner of the baffle in the temperature contour map. However, after enabling the pseudo-transient solver and combining it with local mesh refinement, the temperature residual norm changed from an oscillating state to a monotonically decreasing state, and finally converged stably to a constant within about 350 iterations. In the following, the false temperature spikes completely disappeared, and the resulting temperature field distribution was smooth and physically explainable.

[0100] Compared to the trial-and-error approach of infinitely shrinking the physical time step to barely suppress oscillations, this invention uses a joint convergence strategy that eliminates the need for manual intervention in the step size. The total number of iterations required to achieve the same convergence criteria can be reduced by approximately 40% to 60%, significantly improving the computational efficiency and engineering usability of multiphysics simulation.

[0101] Figure 9 A block diagram of a computing device according to an exemplary embodiment of the present invention is shown.

[0102] To implement the above simulation optimization method, this invention also provides a specific hardware execution environment. For example... Figure 9As shown, the computing device 30 includes a processor 12 and a memory 14. The computing device 30 may also include a bus 22, a network interface card 16, and an I / O interface 18. The processor 12, memory 14, network interface card 16, and I / O interface 18 can communicate with each other via the bus 22. The computing device of this invention can be applied to monitoring equipment such as drones.

[0103] Processor 12 may include one or more general-purpose CPUs (Central Processing Units), microprocessors, or application-specific integrated circuits, for executing relevant program instructions. According to some embodiments, computing device 30 may also include a high-performance display adapter (GPU) 20 for accelerating processor 12.

[0104] Memory 14 may include a machine system readable medium in the form of volatile memory, such as random access memory (RAM), read-only memory (ROM), and / or cache memory. Memory 14 is used to store one or more programs containing instructions, as well as data. Processor 12 may read the instructions stored in memory 14 to perform the methods described above according to embodiments of the present invention.

[0105] The computing device 30 can also communicate with one or more networks via the network interface card 16. The network interface card 16 can be a DPU smart network card.

[0106] Bus 22 can include address bus, data bus, control bus, etc. Bus 22 provides a path for exchanging information between components.

[0107] If the technical solution of this invention involves personal information processing, the relevant products have established a sound user authorization mechanism: before collecting, using, or sharing personal information, the obligation to inform is fulfilled in accordance with the law, and the individual's voluntary and explicit consent is obtained; if sensitive personal information is involved, the user's separate and explicit consent is further obtained. Specific measures include, but are not limited to: setting up prominent prompts in the information collection area, or clearly displaying the processing rules through electronic interfaces such as pop-ups, checkboxes, and active submissions. The processing rules include, but are not limited to, the processor, purpose, method, and information type, ensuring that the user voluntarily authorizes the process based on informed consent. All personal information processing activities strictly comply with national laws and regulations, especially the relevant provisions of the "Personal Information Protection Law of the People's Republic of China," to effectively protect the legitimate rights and interests of personal information subjects.

[0108] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0109] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A transformer heat dissipation structure, characterized in that, include: An iron core, a low-voltage coil sleeved around the iron core, a high-voltage coil sleeved around the low-voltage coil, and a main channel disposed between the low-voltage coil and the high-voltage coil: Within the main channel, several Z-shaped insulating flow guide baffles are fixed alternately from bottom to top along the axial direction. The Z-shaped insulating flow guide baffles cut off the vertical convection path of the cooling oil, forcing the cooling oil to repeatedly turn back and laterally flush the oil passage of the high-voltage coil. The high-voltage coil end is provided with an insulating corner ring. The mounting reference surface of the insulating corner ring is moved down so that the physical lower edge of the insulating corner ring and the physical upper edge of the uppermost Z-shaped insulating guide baffle are flush with each other in axial height. The two form a partial overlap of 0 to 5 mm in axial height, and the axial assembly tolerance is ±2 mm. The insulating corner ring and the uppermost Z-shaped insulating baffle together constitute a thermal plume interception and guiding unit that closes the axial short-circuit gap of the fluid. The cross-sectional shape of the insulating corner ring is any one of L-shaped, oblique, or arc-shaped, and it forms a continuous axial blocking interface with the uppermost Z-shaped insulating baffle on the radial cross section to forcibly intercept the rising cooling oil and fold it back laterally along the lower surface of the insulating corner ring.

2. The transformer heat dissipation structure according to claim 1, characterized in that, The high-voltage coil adopts a segmented cylindrical structure, with horizontal transverse oil channels formed between the coil coils; The Z-shaped insulating baffle forces the cooling oil to deflect along an S-shaped path and laterally flush the transverse oil passage.

3. The transformer heat dissipation structure according to claim 1, characterized in that, The transformer adopts a phase spacing of 600mm~700mm, and the radial spacing of the Z-shaped insulating guide baffles is adapted to the total width of the main channel under the phase spacing.

4. The transformer heat dissipation structure according to claim 1, characterized in that, Each Z-shaped insulating baffle includes a Z-shaped bend formed by connecting a vertical section and a horizontal section, and the angle of the Z-shaped bend ranges from 85° to 95°. The horizontal sections of the Z-shaped insulating baffles in adjacent stages extend in opposite directions. The radial distance between the Z-shaped insulating guide baffle and the main insulation of the coil is 20% to 35% of the total width of the main channel.

5. A multiphysics simulation optimization method for transformer heat dissipation, applied to a transformer including the heat dissipation structure as described in any one of claims 1-4, characterized in that, include: Structural parameters of the transformer under tight phase spacing are extracted, and a three-dimensional geometric model is established, including high and low voltage coils, Z-shaped insulating baffles, and insulating corner rings flush with the uppermost Z-shaped insulating baffles. The inner cavity of the transformer tank is set as a complete closed fluid domain to construct a natural convection physical environment driven by real buoyancy force. Non-uniform loss mapping loading is performed, and three-dimensional loss density distribution data including winding skin effect and end eddy current effect are calculated in the electromagnetic field frequency domain solver. The loss values ​​on the electromagnetic field grid nodes are mapped to the corresponding spatial location nodes of the thermal-fluid coupling model through element shape function interpolation, which serves as a non-uniform body heat source. Mesh the fluid and solid domains. In the transient solution phase, an exponential soft-start function is applied to the non-uniform heat source to smoothly pass the load ramp-up period and eliminate the transient thermal shock caused by the full power step. In the steady-state solution phase, for the abrupt change in flow velocity caused by the sudden change of the cooling fluid at the corner of the Z-shaped insulating guide baffle, a joint solution is performed using local mesh refinement and pseudo-transient time step strategy. By using the converged 3D temperature cloud map and flow field velocity vector map, the effect of the thermal plume interception and guiding unit on suppressing the temperature of the local hot spot and improving the lateral scouring velocity is verified, and a heat dissipation structure scheme is output.

6. The multiphysics simulation optimization method according to claim 5, characterized in that, The step of mapping the loss values ​​on the electromagnetic field grid nodes to the corresponding spatial location nodes of the thermal-fluid coupling model through element shape function interpolation includes: By using element shape function interpolation mapping, the loss values ​​on the electromagnetic field grid nodes calculated by the electromagnetic field frequency domain solver are mapped to the corresponding spatial location nodes of the thermal-fluid coupling model.

7. The multiphysics simulation optimization method for transformer heat dissipation according to claim 5, characterized in that, An exponential soft-start function is applied to the non-uniform heat source, and the exponential soft-start function is: Where P(t) is the actual applied heating power at time t. The rated heating power is τ, which is a time constant used to eliminate solver divergence deadlock caused by step loading of heating power. The value range is 10s~60s.

8. The multiphysics simulation optimization method according to claim 5, characterized in that, The method employs a joint solution using local mesh refinement and a pseudo-transient time-stepping strategy, including: Static meshing was performed at each level on all corners of the Z-shaped insulating flow guide baffle and the adjacent oil passage walls. In the steady-state solution of the discrete control equations, an inertial damping term with an adaptive virtual time step is introduced. By adaptively adjusting the virtual time step according to the local residual gradient, the nonlinear numerical oscillations caused by abrupt changes in the geometric cross-section are smoothed out, thereby eliminating spurious temperature spikes in the flow field. The discrete form expression is as follows: in, The virtual time step, in seconds (s), is adaptively adjusted by the solver based on the gradient magnitude of the local residuals at each grid node. High-oscillation regions are automatically reduced to increase damping, while smooth regions are gradually expanded to accelerate convergence. ρ is the density of the insulating oil, in kg / m³. 3 Δu is the velocity vector difference between the current iteration step and the previous iteration step, in m / s; Here, is the gradient operator; u is the three-dimensional velocity vector of the insulating oil, in m / s; (ρuu) represents the convection tensor; ρ is the fluid density; p is the hydrostatic pressure of the fluid element, in Pa; and μ is the dynamic viscosity of the insulating oil, in Pa. g is the gravitational acceleration constant; β is the volumetric expansion coefficient of transformer oil, in units of... T represents the real-time temperature field distribution within the fluid domain, expressed in Kelvin (K). ref For reference temperature, the initial ambient temperature of the transformer oil or the average oil temperature under rated operating conditions is taken, and the unit is K.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the multiphysics simulation optimization method for transformer heat dissipation as described in any one of claims 5 to 8.

Images