A high-frequency variable transformer line loss calculation method and system

Abstract

The application discloses a high-frequency transformer Litz wire loss calculation method and system, relates to the electrical engineering technical field, and comprises the following steps: calculating the reference loss value and the actual coordinate of the Litz wire, calculating the magnetic field gradient characteristic quantity based on the actual coordinate of each filament, substituting the magnetic field gradient characteristic quantity into the impedance mapping function, calculating the loss correction factor and correcting the reference loss value, obtaining the final total loss value and performing iterative correction, and the like. The application accurately analyzes the magnetic field gradient characteristic quantity around the single wire by introducing the actual coordinate of the filament and the local micro-element model. The impedance mapping function substituted by the magnetic field gradient characteristic quantity is innovatively constructed. In combination with the closed-loop iterative correction mechanism of the measured data, the final total loss value with high precision is output, the two-dimensional visualization of the filament-level loss distribution and the magnetic field gradient is realized, and accurate data support is provided for the local hotspot positioning and targeted structure optimization of the Litz wire winding of the high-frequency transformer.

Description

Technical Field

[0001] This invention relates to the field of electrical engineering technology, and in particular to a method and system for calculating the loss of a high-frequency transformer line (Lieds line). Background Technology

[0002] In recent years, with the development of power electronics technology towards higher frequencies and higher power densities, the operating frequency of high-frequency transformers has been continuously increasing, leading to a sharp increase in the skin effect and proximity effect inside the windings. Litz wire, through the specific twisting and transposition of multiple insulating filaments, effectively breaks the eddy current loop and suppresses the skin effect, becoming an ideal choice for reducing high-frequency AC losses. The interior of Litz wire is a highly complex multi-strand electromagnetic coupling system. The actual twisting structure causes the filaments to deviate from the ideal position, forming an extremely non-uniform spatial magnetic field distribution inside the wire bundle. This intricate local magnetic field environment at the microscale, intertwined with the proximity effect, constitutes the core difficulty in the loss characteristic analysis of high-frequency transformers.

[0003] In existing technologies, computational models are usually based on the idealized assumption of an absolutely uniform distribution, ignoring the positional deviation of the filaments caused by actual twisting. They cannot characterize the local magnetic field non-uniformity and gradient features at the microscale, resulting in extremely large errors at high frequencies. While conventional finite element simulation can solve for macroscopic reference losses, it lacks a refined analysis of the rate of change of the magnetic field around the single filament and fails to establish a deep mapping mechanism between magnetic field gradient and impedance loss. Existing methods generally lack the ability to perform closed-loop iterative correction based on measured data and can only output a single total loss value, failing to visualize the loss distribution at the filament level and making it difficult to effectively guide the targeted optimization of winding structures. Summary of the Invention

[0004] The technical problem solved by this invention is that: the calculation model is usually based on the idealized assumption of absolute uniform distribution, ignoring the positional deviation of the filaments caused by actual twisting, and cannot characterize the local magnetic field non-uniformity and gradient characteristics at the microscale. The error is extremely large at high frequencies. Although conventional finite element simulation can solve the macroscopic reference loss, it lacks a refined analysis of the rate of change of the magnetic field around the single filament and fails to establish a deep mapping mechanism between the magnetic field gradient and impedance loss. Existing methods generally lack the ability to perform closed-loop iterative correction of measured data and can only output a single total loss value. They cannot achieve visualization of the loss distribution at the filament level and are difficult to effectively guide the targeted optimization of the winding structure.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for calculating the loss of a high-frequency transformer line (Litz line), comprising the following steps:

[0006] Step S1: Obtain the multi-strand strand structure parameters of the Litz wire, and calculate the reference loss value of the Litz wire based on the multi-strand strand structure parameters;

[0007] Step S2: Calculate the actual coordinates of each filament based on the parameters of the multi-strand twisted structure, and calculate the superimposed magnetic field strength based on the actual coordinates of each filament and the current value. Summate the superimposed magnetic field strength with the intrinsic magnetic field strength vector to obtain the local magnetic field strength. Calculate the ratio of the magnetic field difference between adjacent sampling points to the distance to obtain the magnetic field strength change rate. Use the average value of the magnetic field strength change rate at each angular position as the magnetic field gradient feature quantity.

[0008] Step S3: Construct an impedance mapping function, substitute the magnetic field gradient characteristic into the impedance mapping function, and calculate the loss correction factor;

[0009] Step S4: Correct the baseline loss value using the loss correction factor, calculate the final total loss value, perform iterative correction based on the final total loss value, and store and visualize the result.

[0010] As a preferred embodiment of the high-frequency transformer line Litz line loss calculation method of the present invention, step S1 specifically includes:

[0011] Obtain the parameters of the multi-strand twisted structure of Leeds wire;

[0012] The parameters of the multi-strand stranded structure include the number of stranded layers, the number of filaments in each layer, the stranding direction, and the stranding pitch.

[0013] Based on the number of strands and the number of filaments in each layer, a Cartesian coordinate system is established with the geometric center of the Litz wire cross-section as the origin, and a two-dimensional geometric model of the Litz wire cross-section is generated in the Cartesian coordinate system.

[0014] The two-dimensional geometric model is divided into finite element mesh units according to a preset mesh size;

[0015] Obtain the total current value, the total number of filaments, and the cross-sectional area of ​​a single filament of the Litz wire. Calculate the quotient of the total current value and the total number of filaments, and then divide it by the cross-sectional area of ​​a single filament to obtain the initial current density of a single filament.

[0016] The initial current density is used as a boundary condition, and the boundary condition is assigned to the nodes of each finite element mesh element.

[0017] The preset two-dimensional electromagnetic field solver is invoked, and the initial current density on the nodes of each finite element mesh element is used as input to solve the Maxwell equations of the two-dimensional geometric model to obtain the magnetic flux density of each finite element mesh element.

[0018] The square of each magnetic flux density is multiplied by the area of ​​each finite element mesh cell, and the results of these multiplications are summed. The sum of these sums is then multiplied by the axial unit length of the Litz line, and the result is used as the reference loss value.

[0019] As a preferred embodiment of the high-frequency transformer line Litz line loss calculation method of the present invention, step S2 specifically includes:

[0020] Based on the twist pitch and number of twist layers of the Litz wire, a polar coordinate system is established with the geometric center of the Litz wire cross section as the origin, and the theoretical distribution radius of each layer of filaments is calculated.

[0021] Based on the theoretical distribution radius of each layer of filaments and the number of filaments in each layer, the theoretical angular coordinates of each filament are calculated using the principle of equal angular division.

[0022] Substitute the preset angle deviation coefficient and the preset radial deviation coefficient into the angle deviation correction function and the radial deviation correction function respectively to correct the theoretical angle coordinates and the theoretical distribution radius, and obtain the actual coordinates of the filament;

[0023] Traverse each filament inside the Lids line and use the currently traversed filament as the target filament;

[0024] Based on the actual coordinates of the target filament, an annular region is divided in the polar coordinate system with the geometric center of the target filament as the center and a preset distance as the radius. The annular region is used as a local micro-element model around the target filament.

[0025] Obtain the actual coordinates of each filament other than the target filament and the current value of each filament. Substitute the actual coordinates of each filament, the current value of each filament and the actual coordinates of the target filament into the preset magnetic field superposition analytical formula for calculation, and output the superimposed magnetic field strength generated by the filaments other than the target filament at the position of the target filament.

[0026] Obtain the current value and diameter of the target filament, calculate the ratio of the current value to the product of the filament diameter and pi, and use the ratio as the intrinsic magnetic field strength of the target filament.

[0027] The vector component of the superimposed magnetic field strength is added to the vector component of the intrinsic magnetic field strength, and the modulus operation is performed on the added vector component to obtain the local magnetic field strength around the target filament.

[0028] A predetermined number of angle positions are uniformly selected along the circumference of the target filament. A predetermined number of sampling points are selected along the radial direction at each angle position. The difference in magnetic field strength between adjacent radial sampling points at each angle position is calculated.

[0029] Calculate the ratio of the difference in magnetic field strength to the distance between adjacent sampling points, and use the ratio as the rate of change of magnetic field strength at that angular position;

[0030] Calculate the average value of the rate of change of magnetic field intensity at each angular position, and use the average value as the magnetic field gradient characteristic quantity of the target filament.

[0031] In a preferred embodiment of the high-frequency transformer line Litz line loss calculation method of the present invention, step S3 specifically includes:

[0032] An impedance mapping function is constructed, the magnetic field gradient characteristic quantity is substituted into the impedance mapping function, and the integrand in the impedance mapping function is discretized and summed using a numerical integration algorithm to obtain a numerical solution.

[0033] Extract the real part of the numerical solution and obtain the resistance coefficient corresponding to the reference loss value. Use the resistance coefficient as a proportionality coefficient, calculate the product of the real part and the proportionality coefficient, and use the product result as a loss correction factor.

[0034] As a preferred embodiment of the high-frequency transformer line Litz line loss calculation method described in this invention, the construction of the impedance mapping function specifically includes:

[0035] Obtain the type identifier of the fractional differential operator, and invoke the corresponding fractional differential operator calculation rule based on the type identifier;

[0036] Obtain the upper limit of the operating frequency range of the Litz wire and the coupling strength value between the filaments, and calculate the product of the upper limit of the operating frequency range and the coupling strength value, and use the product result as the initial value of the fractional order.

[0037] Using the magnetic field gradient feature as the base variable, the power of the magnetic field gradient feature value is calculated to obtain the kernel function;

[0038] Obtain the time step of the historical change of the magnetic field, calculate the negative power of the exponent with the initial value as the base of the time step of the historical change of the magnetic field, and use the calculation result as the weighting coefficient of the memory effect.

[0039] Multiply the memory effect weighting coefficient by the kernel function, and use the result of the multiplication as the integrand;

[0040] The integrand is calculated according to the fractional differential operator calculation rules to obtain the impedance mapping function.

[0041] In a preferred embodiment of the high-frequency transformer line Litz line loss calculation method described in this invention, step S4 specifically includes:

[0042] Obtain the total number of filaments in the Litz wire and create a reference loss component variable equal to the total number of filaments;

[0043] Traverse each finite element mesh element in the two-dimensional geometric model to obtain the center point coordinates, magnetic flux density, and area of ​​the currently traversed finite element mesh element;

[0044] The distance between the coordinates of the center point and the geometric center coordinates of each filament in the two-dimensional geometric model is calculated, and the filament with the smallest distance is selected as the current filament.

[0045] Calculate the product of the square of the magnetic flux density of the currently traversed finite element mesh cell and the area, and add the product result to the reference loss component variable corresponding to the current filament, and update the value of the reference loss component variable;

[0046] After traversing each finite element mesh, the final value of the reference loss component variable corresponding to each filament is used as the reference loss component value of each filament.

[0047] Calculate the product of the reference loss component value of each filament and the loss correction factor corresponding to each filament, and use the product result as the single filament loss value;

[0048] The loss values ​​of each single filament are summed up, and the summation result is used as the total loss correction value.

[0049] Calculate the product of the total loss correction value and the preset experimental calibration coefficient, and use the product result as the final total loss value;

[0050] Iterative corrections are performed based on the final total loss value, and the results are stored and visualized.

[0051] As a preferred embodiment of the high-frequency transformer line Litz line loss calculation method described in this invention, the iterative correction specifically includes:

[0052] The actual loss values ​​of the Litz wire under specific operating conditions were collected using a power analyzer.

[0053] Calculate the difference between the actual loss value and the final total loss value, and divide the absolute value of the difference by the actual loss value to obtain the relative error value;

[0054] The relative error value is compared with a preset error threshold.

[0055] When the relative error value is greater than the preset error threshold, the parameter iteration step is executed to obtain the final output result;

[0056] When the relative error value is less than or equal to the preset error threshold, the current fractional order and the current proportional coefficient are used as the final output result.

[0057] As a preferred embodiment of the high-frequency transformer line Litz line loss calculation method described in this invention, the parameter iteration step specifically includes:

[0058] Get the current value of the fractional order as the first value, and get the current value of the proportional coefficient as the second value.

[0059] Calculate the first partial derivative of the relative error value with respect to the fractional order, and calculate the second partial derivative of the relative error value with respect to the proportionality coefficient;

[0060] Calculate the product of the first partial derivative and the preset first step length to obtain the first result, and calculate the first value minus the first result to obtain the updated fractional order;

[0061] Calculate the product of the second partial derivative and the preset second step size to obtain the second result, and calculate the second value minus the second result to obtain the updated scaling factor;

[0062] The final total loss value is recalculated based on the updated fractional order and the updated scaling factor to obtain the updated final total loss value.

[0063] The absolute value of the difference between the updated final total loss value and the actual loss value is divided by the actual loss value to obtain the updated relative error value;

[0064] The updated relative error value is compared with a preset error threshold.

[0065] If the updated relative error value is greater than the preset error threshold, the parameter iteration step is re-executed.

[0066] When the updated relative error value is less than or equal to the preset error threshold, the iteration stops and the currently updated fractional order and the updated proportional coefficient are used as the final output results.

[0067] As a preferred embodiment of the high-frequency transformer line Litz line loss calculation method described in this invention, the step of storing and visualizing the output specifically includes:

[0068] The actual coordinates of each filament, local magnetic field strength, magnetic field gradient characteristic quantity, loss correction factor, single filament loss value and final total loss value are written into the database according to the preset structure.

[0069] Read the single filament loss value corresponding to the actual coordinates of each filament in the database, match the corresponding color value, and generate a two-dimensional loss image on the display terminal;

[0070] Calculate the direction angle of the magnetic field gradient feature quantity in the database, and draw vector arrows by superimposing the direction angle at the actual coordinates of each filament.

[0071] A high-frequency transformer line Litz line loss calculation system is applied to a high-frequency transformer line Litz line loss calculation method, including a reference module, an analysis module, a correction module and an output module;

[0072] The reference module is used to obtain the multi-strand strand structure parameters of the Leeds wire and calculate the reference loss value of the Leeds wire based on the multi-strand strand structure parameters.

[0073] The analysis module is used to calculate the actual coordinates of each filament based on the parameters of the multi-strand twisted structure, and to calculate the superimposed magnetic field strength based on the actual coordinates of each filament and the current value. The superimposed magnetic field strength is summed with the intrinsic magnetic field strength vector to obtain the local magnetic field strength. The ratio of the magnetic field difference between adjacent sampling points to the distance is calculated to obtain the magnetic field strength change rate. The average value of the magnetic field strength change rate at each angular position is used as the magnetic field gradient feature quantity.

[0074] The correction module is used to construct an impedance mapping function, and to calculate the loss correction factor by substituting the magnetic field gradient characteristic into the impedance mapping function.

[0075] The output module is used to correct the baseline loss value using the loss correction factor, calculate the final total loss value, perform iterative correction based on the final total loss value, and then store and visualize the result.

[0076] The beneficial effects of this invention are as follows: By introducing the actual coordinates of the filament and a local micro-element model, this invention accurately analyzes the magnetic field gradient characteristics around the filament, breaking through the limitations of the idealized uniform distribution assumption. It truly restores the local magnetic field non-uniformity at the microscale under high frequency. By innovatively constructing an impedance mapping function that substitutes the magnetic field gradient characteristics, a deep mapping mechanism between the magnetic field gradient and impedance loss is established, realizing refined single-filament-level correction of the reference loss. This significantly reduces high-frequency calculation errors. Combined with a closed-loop iterative correction mechanism based on measured data, it can dynamically optimize and approximate the real physical loss, eliminating calculation deviations. Ultimately, it not only outputs a high-precision total loss value but also realizes two-dimensional visualization of the filament-level loss distribution and magnetic field gradient, providing accurate data support for the local hotspot location and targeted structural optimization of the Litz wire winding of high-frequency transformers. Attached Figure Description

[0077] Figure 1 This is a schematic diagram of the basic process for calculating the loss of a high-frequency transformer line, provided in one embodiment of the present invention.

[0078] Figure 2 This is a basic flowchart of a high-frequency transformer line Litz line loss calculation system provided in one embodiment of the present invention. Detailed Implementation

[0079] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0080] Reference Figure 1 As an embodiment of the present invention, a method for calculating the loss of a high-frequency transformer line Litz wire is provided, comprising the following steps:

[0081] Step S1: Obtain the multi-strand strand structure parameters of the Leeds wire, and calculate the reference loss value of the Leeds wire based on the multi-strand strand structure parameters;

[0082] Step S2: Calculate the actual coordinates of each filament based on the parameters of the multi-strand twisted structure, and calculate the superimposed magnetic field strength based on the actual coordinates of each filament and the current value. Summate the superimposed magnetic field strength with the intrinsic magnetic field strength vector to obtain the local magnetic field strength. Calculate the ratio of the magnetic field difference between adjacent sampling points to the distance to obtain the magnetic field strength change rate. Use the average value of the magnetic field strength change rate at each angular position as the magnetic field gradient feature quantity.

[0083] Step S3: Construct the impedance mapping function, substitute the magnetic field gradient characteristic into the impedance mapping function, and calculate the loss correction factor.

[0084] Step S4: Correct the baseline loss value using the loss correction factor, calculate the final total loss value, perform iterative correction based on the final total loss value, and then store and visualize the result.

[0085] This paper defines a comprehensive computational architecture that includes baseline loss calculation, magnetic field gradient characteristic quantity analysis, impedance mapping function correction, and iterative correction. This architecture abandons the conventional method of relying solely on the average magnetic flux density of the macroscopic cross section to evaluate losses. For the first time, it extracts the rate of change of magnetic field strength between adjacent sampling points in the microscopic space as a key characteristic indicator and directly participates in the construction of total loss. While comprehensively reflecting the high-frequency electromagnetic coupling characteristics of multi-strand stranded structures, it significantly improves the accuracy and reliability of loss evaluation results under complex operating conditions, laying a new underlying logical foundation for the thermal design and efficiency evaluation of high-frequency transformers.

[0086] Step S1 specifically includes:

[0087] Obtain the parameters of the multi-strand twisted structure of Leeds wire;

[0088] The parameters of a multi-strand stranded structure include the number of stranded layers, the number of filaments in each layer, the stranding direction, and the stranding pitch.

[0089] Based on the number of strands and the number of filaments in each layer, a Cartesian coordinate system is established with the geometric center of the Litz wire cross-section as the origin, and a two-dimensional geometric model of the Litz wire cross-section is generated in the Cartesian coordinate system.

[0090] The two-dimensional geometric model is divided into finite element mesh elements according to the preset mesh size;

[0091] The preset mesh size is set to one-fifth of the diameter of a single filament.

[0092] Obtain the total current value, the total number of filaments, and the cross-sectional area of ​​a single filament of the Litz wire. Calculate the quotient of the total current value and the total number of filaments, and then divide it by the cross-sectional area of ​​a single filament to obtain the initial current density of a single filament.

[0093] The initial current density is used as the boundary condition, and the boundary conditions are assigned to the nodes of each finite element mesh element.

[0094] Call the preset two-dimensional electromagnetic field solver, take the initial current density on the nodes of each finite element mesh element as input, solve Maxwell's equations of the two-dimensional geometric model, and obtain the magnetic flux density of each finite element mesh element.

[0095] The preset two-dimensional electromagnetic field solver is set as a harmonic magnetic field solver based on the magnetic vector potential A.

[0096] The square of each magnetic flux density is multiplied by the area of ​​each finite element mesh cell, and the results of these multiplications are summed. The sum of these sums is then multiplied by the axial unit length of the Litz line, and the result is used as the reference loss value.

[0097] The two-dimensional geometric model uses the absolute boundary equivalence method to treat the insulation layer, and the insulation thickness of a single filament is equivalent to the physical distance between adjacent conductors. The circular boundary of the single filament is discretized according to a node every 20 degrees to capture edge field distortion.

[0098] In the process of converting the total current to the initial current density of a single filament, a skin depth weighting coefficient based on the operating frequency is introduced, so that nodes closer to the outer surface are given a higher proportion of initial current density. When solving Maxwell's equations, the outer solution domain truncation boundary is set to a distance of five times the outer diameter of the Litz wire from the outermost filament surface, and a second-kind homogeneous boundary condition with zero magnetic vector potential is applied to completely eliminate the interference of boundary reflection on the calculation of the internal micro-eddy current distribution.

[0099] By establishing a two-dimensional geometric model and performing finite element analysis according to a preset mesh size, the total current is accurately converted into the initial current density of a single filament as the boundary condition for solving Maxwell's equations at each node. Compared with the simulation scheme that directly applies equivalent excitation to the entire macroscopic cross-section of the Litz wire, this microscopic processing method can more realistically reflect the independent eddy current distribution state of multiple independent insulating filaments under an alternating magnetic field, avoids local oversaturation distortion caused by overall averaging, and provides extremely rigorous and accurate initial reference data for subsequent refined loss correction. This effectively overcomes the inherent accuracy defects of existing macroscopic finite element calculations in the initial calculation stage.

[0100] Step S2 specifically includes:

[0101] Based on the twist pitch and number of twist layers of the Litz wire, a polar coordinate system is established with the geometric center of the Litz wire cross section as the origin, and the theoretical distribution radius of each layer of filaments is calculated.

[0102] Using the geometric center of the target multi-core cable cross-section as the reference origin, the theoretical distribution radius of each layer of filaments is calculated layer by layer from the inside out, based on the diameter of a single filament, the thickness of the insulation medium between adjacent layers, and the preset number of filaments in each layer.

[0103] The theoretical distribution radius is the radius of the trajectory of the concentric circles containing the center points of each layer of filaments.

[0104] Based on the theoretical distribution radius of each layer of filaments and the number of filaments in each layer, the theoretical angular coordinates of each filament are calculated using the principle of equal angular division.

[0105] Substitute the preset angle deviation coefficient and preset radial deviation coefficient into the angle deviation correction function and radial deviation correction function respectively to correct the theoretical angle coordinates and theoretical distribution radius, and obtain the actual coordinates of the filament;

[0106] The preset angle deviation coefficient is set to the statistical mean of angle offset derived from the normal distribution disturbance term.

[0107] The preset radial deviation coefficient is set to the statistical mean of radial offset derived from the normal distribution disturbance term.

[0108] Traverse each filament inside the Lids line and use the currently traversed filament as the target filament;

[0109] Based on the actual coordinates of the target filament, an annular region is divided in the polar coordinate system with the geometric center of the target filament as the center and a preset distance as the radius. The annular region is used as a local micro-element model around the target filament.

[0110] The preset distance is set to one and a half times the diameter of a single filament.

[0111] Obtain the actual coordinates of each filament other than the target filament and the current value of each filament. Substitute the actual coordinates of each filament, the current value of each filament and the actual coordinates of the target filament into the preset magnetic field superposition analytical formula for calculation, and output the superimposed magnetic field strength generated by the filaments other than the target filament at the position of the target filament.

[0112] The preset analytical formula for superposition of magnetic fields is set as a two-dimensional degenerate form of the Biot-Savart law.

[0113] Obtain the current value and diameter of the target filament, calculate the ratio of the current value to the product of the filament diameter and pi, and use the ratio as the intrinsic magnetic field strength of the target filament.

[0114] The vector component of the superimposed magnetic field strength is added to the vector component of the intrinsic magnetic field strength, and the modulus of the added vector component is calculated to obtain the local magnetic field strength around the target filament.

[0115] A predetermined number of angle positions are uniformly selected along the circumference of the target filament. A predetermined number of sampling points are selected along the radial direction at each angle position. The difference in magnetic field strength between adjacent radial sampling points at each angle position is calculated.

[0116] A predetermined number of angle positions are evenly selected at equal angular intervals along the circumference of the target filament. These angle positions are divided into 360 degrees with the center of the target filament as the pole, and the microscopic magnetic field characteristics are accurately extracted in the normal direction of each division point.

[0117] The default quantity is set to eight.

[0118] Calculate the ratio of the difference in magnetic field strength to the distance between adjacent sampling points, and use the ratio as the rate of change of magnetic field strength at that angular position;

[0119] Calculate the average value of the rate of change of magnetic field intensity at each angular position, and use the average value as the magnetic field gradient characteristic quantity of the target filament.

[0120] The angular deviation coefficient and radial deviation coefficient are derived from the pitch parameters and tension fluctuation threshold in the actual Litz wire stranding process. The mathematical expressions are constructed by substituting the theoretical helical equation into the normal distribution disturbance term.

[0121] The local micro-element model takes the center of the target filament as the pole, divides the polar diameter into ten equally spaced distance steps, and divides the polar angle into fan-shaped regions at 30-degree intervals. This scale preserves the characteristics of drastic changes in the micro magnetic field while keeping the amount of computation per operation at an extremely low level.

[0122] The analytical formula for superposition of magnetic fields adopts a two-dimensional degenerate form of Biot-Savart's law, which vector-sums the tangential magnetic field strength generated by all surrounding current-carrying filaments at the target micro-element node, thereby accurately extracting the magnetic field gradient characteristic quantity that reflects the degree of drastic spatial changes.

[0123] By introducing angular deviation coefficients and radial deviation coefficients to correct the theoretical coordinates, the actual coordinates of the filament are obtained. Combining the local micro-element model and the analytical formula for magnetic field superposition, the rate of change of magnetic field intensity at each angular position around the target filament is accurately calculated, and the average value is taken as the magnetic field gradient characteristic quantity. This scheme accurately replicates the unavoidable filament position offset phenomenon in actual manufacturing process, fills the gap of existing numerical models that ignore physical processing errors and cause microscopic magnetic field analysis distortion, and enables the loss calculation model to deeply fit the internal geometry of real physical samples, greatly enhancing engineering adaptability and bridging the physical gap between theoretical derivation and industrial manufacturing.

[0124] Step S3 specifically includes:

[0125] An impedance mapping function is constructed, and the magnetic field gradient characteristic quantity is substituted into the impedance mapping function. The integrand in the impedance mapping function is discretized and summed using a numerical integration algorithm to obtain the numerical solution of the kernel function under the action of the fractional differential operator.

[0126] The integration interval of the integrand in the impedance mapping function is divided into defined discrete sub-intervals. Within each sub-interval, the corresponding integration node is selected and the function value at that node is calculated. Each function value is multiplied by the integration step size of its respective sub-interval. Finally, the multiplications of all sub-intervals are summed to complete the discretized numerical integration calculation of the integrand.

[0127] Extract the real part of the numerical solution and obtain the resistance coefficient corresponding to the reference loss value. Use the resistance coefficient as the proportional coefficient, calculate the product of the real part and the proportional coefficient, and use the product result as the loss correction factor.

[0128] The numerical integration algorithm uses the Gauss-Legend integration method, which discretizes the integration path of the impedance mapping function in the complex plane into sixteen Gaussian integration nodes, effectively solving the problem of singularities in the integrand at high frequencies.

[0129] During the extraction of real part data, a phase-shift-free Butterworth low-pass filter was applied to the complex impedance sequence obtained by discretization to filter out high-frequency computational noise introduced by higher-order spatial harmonics, ensuring that the extracted real part purely reflects the active power loss characteristics.

[0130] The resistivity is matched and calibrated using the single-wire AC resistance reference value obtained by the four-wire method at a constant operating temperature, so that the final calculated loss correction factor strictly corresponds to the actual physical heating state.

[0131] By using a numerical integration algorithm to discretize and solve the impedance mapping function containing magnetic field gradient characteristics and extract the real part data, and combining it with the resistivity corresponding to the reference loss to calculate the loss correction factor, the abstract microscopic spatial magnetic field gradient characteristics are accurately transformed into concrete resistance increment parameters. This correction mechanism based on real part impedance extraction effectively eliminates the interference of reactive components in the complex domain calculation process, making the stripping of pure active power loss increment caused by proximity effect purer. The correction accuracy and physical meaning are significantly better than the calculation performance of the existing fixed empirical coefficient compensation method, and the adverse erosion of complex plane error on the final active power calculation result is eliminated.

[0132] Constructing the impedance mapping function specifically includes:

[0133] Obtain the type identifier of the fractional differential operator, and call the corresponding fractional differential operator calculation rule based on the type identifier;

[0134] Obtain the upper limit of the operating frequency range of the Litz wire and the coupling strength between the filaments, and calculate the product of the upper limit of the operating frequency range and the coupling strength. Use the product result as the initial value of the fractional order.

[0135] Using the magnetic field gradient feature as the base variable, the power of the magnetic field gradient feature is calculated to obtain the kernel function.

[0136] The frequency dependence index of the Litz wire conductor material is obtained as the power exponent value, and the magnetic field gradient characteristic quantity is used as the base variable of the power exponent. The power exponent value of the magnetic field gradient characteristic quantity is calculated to obtain the kernel function.

[0137] Obtain the time step of the historical changes in the magnetic field, calculate the negative power of the initial value with the time step of the historical changes in the magnetic field as the base, and use the calculation result as the weighting coefficient of the memory effect.

[0138] Multiply the memory effect weighting coefficients by the kernel function, and use the result of the multiplication as the integrand;

[0139] The integrand is calculated according to the rules for fractional differential operators to obtain the impedance mapping function.

[0140] The order of the fractional differential operator is strictly defined in the non-integer domain between 0 and 1. Its value undergoes dynamic step transitions based on the relative relationship between the skin depth driven by the operating frequency and the Litz line radius: when in the low frequency range and the skin depth is greater than the Litz line radius, the order approaches 0; when in the high frequency range and the skin depth is less than the Litz line radius, the order approaches 1. This precisely matches the nonlinear shrinkage process of the skin depth.

[0141] The memory effect weighting coefficient is calculated by introducing a recursive formula containing a decay exponent. The gradient information of the past ten consecutive magnetic field history time steps is weighted according to an exponential law, with the latest time step having the largest weight. Earlier time steps decay rapidly through a forgetting factor. The product of the upper limit of the operating frequency and the coupling strength is used as the initial value of the fractional order because this product directly characterizes the complexity of the electromagnetic transient process of the system. Using this as the starting point can enable the subsequent gradient descent optimization process to converge to the true physical order with the shortest path.

[0142] By introducing a fractional-order differential operator and a memory effect weighting coefficient calculated based on the time step of the magnetic field history when constructing the impedance mapping function, and using the product of the upper limit of the operating frequency and the coupling strength between the filaments as the initial value of the fractional-order, this design profoundly reveals the physical essence of the significant time memory effect of high-frequency electromagnetic fields in complex stranded structures. It overcomes the inherent limitations of existing integer-order calculus models in describing the dynamic changes of magnetic flux diffusion and skin depth, and achieves a higher-level and more realistic mathematical fitting of the evolution process of nonlinear strongly coupled magnetic fields. This represents a cutting-edge breakthrough in the field of high-frequency electromagnetic loss analysis, moving from the classical integer order to the modern fractional order.

[0143] Step S4 specifically includes:

[0144] Obtain the total number of filaments in the Litz wire and create a baseline loss component variable equal to the total number of filaments;

[0145] Traverse each finite element mesh element in the two-dimensional geometric model to obtain the center point coordinates, magnetic flux density, and area of ​​the currently traversed finite element mesh element;

[0146] The distance between the center point coordinates and the geometric center coordinates of each filament in the two-dimensional geometric model is calculated, and the filament with the smallest distance is selected as the current filament.

[0147] Calculate the product of the square of the magnetic flux density and the area of ​​the finite element mesh element being traversed, and add the product result to the reference loss component variable corresponding to the current filament, and update the value of the reference loss component variable.

[0148] After traversing each finite element mesh, the final value of the reference loss component variable corresponding to each filament is used as the reference loss component value of each filament.

[0149] Calculate the product of the reference loss component value of each filament and the loss correction factor corresponding to each filament, and use the product result as the single filament loss value;

[0150] The loss values ​​of each single filament are summed up, and the summation result is used as the total loss correction value.

[0151] Calculate the product of the total loss correction value and the preset experimental calibration coefficient, and use the product result as the final total loss value;

[0152] The preset experimental calibration coefficient is set to a constant of one.

[0153] Iterative corrections are performed based on the final total loss value, and the results are stored and visualized.

[0154] The spatial distance is accumulated by using a kd-tree-based spatial nearest neighbor search algorithm. The Euclidean distance from the geometric center of each finite element mesh cell to the center of each filament is calculated, and the mesh loss is multiplied by the reciprocal of the distance as a weight to accurately distribute the filaments to the nearest filament.

[0155] The storage structure of the baseline loss component variables adopts a sparse matrix compressed storage format, which only records the filament index and corresponding value that contribute to the mesh loss. This reduces the memory usage by an order of magnitude when mapping thousands of mesh cells to hundreds of filaments. An extreme value filtering mechanism is set before global summation. When the loss value of a single filament exceeds three times the standard deviation of the average loss of all filaments, a secondary microscopic analysis based on local mesh refinement is triggered to eliminate calculation anomalies caused by mesh distortion.

[0156] By traversing all finite element mesh elements in the two-dimensional geometric model, the product of the square of the magnetic flux density and the area of ​​each element is accumulated to the reference loss component variable of the corresponding filament according to the spatial distance. Combined with the loss correction factor specific to each filament, the loss value of each filament is calculated and then summed globally. This achieves a seamless and accurate mapping from macroscopic finite element mesh loss to microscopic individual filament loss, completely breaking the limitation of existing calculation schemes that can only give the average loss of the entire wire bundle. This provides a quantitative basis for accurately identifying hot filaments inside the wire bundle that are subjected to extreme electromagnetic stress.

[0157] Iterative correction specifically includes:

[0158] The actual loss values ​​of the Litz wire under specific operating conditions were collected using a power analyzer.

[0159] The specific operating conditions are the specific electromagnetic and physical operating conditions applied when calculating the impedance mapping function, including the set frequency of the alternating current, the current amplitude flowing through the cable, and the ambient temperature during detection.

[0160] Calculate the difference between the actual loss value and the final total loss value, and divide the absolute value of the difference by the actual loss value to obtain the relative error value;

[0161] Compare the relative error value with a preset error threshold;

[0162] When the relative error value is greater than the preset error threshold, the parameter iteration step is executed to obtain the final output result;

[0163] When the relative error value is less than or equal to the preset error threshold, the current fractional order and the current proportional coefficient are used as the final output result.

[0164] The actual data acquisition process was carried out in a constant temperature sealed chamber. The temperature of the Litz wire sample under test was locked at the constant value set by theoretical calculation through real-time feedback of thermocouples. At the same time, the excitation frequency and current amplitude were strictly tested and calculated synchronously using an arbitrary waveform generator.

[0165] The preset error threshold adopts an adaptive dynamic setting rule based on frequency bands. The threshold is set to 2% in the low-frequency band below 100 kHz and relaxed to 5% in the high-frequency band above 500 kHz to match the uncertainty level of physical measurements under different frequency bands. When calculating the relative error, the absolute value of the actual loss measured by the power analyzer is used as the normalized denominator, which completely eliminates the distortion problem of the error percentage being abnormally amplified under light load or low-frequency extremely low loss conditions.

[0166] By directly acquiring the actual loss value of Litz wire under specific operating conditions using a power analyzer, comparing it with the final total loss value to obtain the relative error, and setting a preset error threshold to trigger parameter iteration, a complete closed-loop feedback channel from pure theoretical derivation to physical entity testing is constructed. This self-calibration mechanism, driven by real measured data, can autonomously eliminate the accumulated systematic errors caused by material property fluctuations, environmental temperature changes, and simplified boundary conditions, ensuring that the loss calculation model maintains a very high approximation degree under different operating conditions and different batches of Litz wire products.

[0167] The parameter iteration steps specifically include:

[0168] Get the current value of the fractional order as the first value, and get the current value of the proportional coefficient as the second value.

[0169] Calculate the first partial derivative of the relative error value with respect to the fractional order, and calculate the second partial derivative of the relative error value with respect to the proportionality coefficient;

[0170] Calculate the product of the first partial derivative and the preset first step length to obtain the first result, and calculate the first value minus the first result to obtain the updated fractional order;

[0171] Calculate the product of the second partial derivative and the preset second step size to obtain the second result, and calculate the second value minus the second result to obtain the updated scaling factor;

[0172] The default step length is set to one percent of the current score order value.

[0173] The preset second step size is set to one percent of the current scaling factor value.

[0174] The final total loss value is recalculated based on the updated fractional order and the updated scaling factor to obtain the updated final total loss value.

[0175] The absolute value of the difference between the updated final total loss value and the actual loss value is divided by the actual loss value to obtain the updated relative error value.

[0176] Compare the updated relative error value with the preset error threshold;

[0177] If the updated relative error value is greater than the preset error threshold, the parameter iteration step is re-executed.

[0178] When the updated relative error value is less than or equal to the preset error threshold, the iteration stops and the updated fractional order and the updated proportional coefficient are used as the final output.

[0179] The partial derivatives are calculated using the adjoint derivative method for complex functions. Through a backpropagation network, the gradient of the final relative error is precisely decomposed into two independent parameters: the fractional order and the proportionality coefficient. This avoids the truncation error caused by directly performing numerical differencing on expressions containing fractional calculus.

[0180] By accurately calculating the partial derivatives of the relative error with respect to the fractional order and the proportional coefficient, and dynamically updating these two core parameters using gradient descent with a preset step size, the loss calculation system is endowed with a powerful multidimensional parameter self-optimization capability. Compared with the existing method that relies on engineers' personal experience to repeatedly manually adjust parameters, this mechanism can automatically lock the global optimal solution with an extremely fast convergence speed, and maintain the high efficiency and absolute stability of the algorithm even when facing complex high-frequency operating conditions with strong coupling of multiple physics fields and highly nonlinear interweaving.

[0181] The specific steps for storing and visualizing the output include:

[0182] The actual coordinates of each filament, local magnetic field strength, magnetic field gradient characteristic quantity, loss correction factor, single filament loss value and final total loss value are written into the database according to the preset structure.

[0183] The default structure is set as a standard twisted layered arrangement of multi-core cables.

[0184] Read the single filament loss value corresponding to the actual coordinates of each filament in the database, match the corresponding color value, and generate a two-dimensional loss image on the display terminal;

[0185] Calculate the direction angle of the magnetic field gradient feature quantity in the database, and draw vector arrows by superimposing the direction angle at the actual coordinates of each filament.

[0186] A local polar coordinate system is established with the center of the target filament as the pole and the preset reference baseline as the polar axis. The magnetic field intensity change vector of the filament at adjacent sampling points is extracted. The direction angle of the magnetic field gradient characteristic quantity is obtained by calculating the arctangent function value between the change vector and the polar axis.

[0187] Color value matching uses a classic heatmap color band that transitions from dark blue to bright red, and is dynamically calibrated with linear normalization based on the maximum and minimum values ​​of all single filament loss values ​​to ensure that color contrast can clearly reflect loss differences.

[0188] The vector arrow drawing introduces a logarithmic scaling factor, which maps the magnitude of the magnetic field gradient characteristic quantity to the geometric length of the arrow after taking the logarithm. At the same time, a semi-transparent diffused halo is added to the tail of the arrow to effectively avoid visual occlusion caused by dense overlap of arrows in areas of high magnetic field distortion.

[0189] The coordinate alignment of the two-dimensional image is achieved by extracting the extreme boundary of the actual coordinates of the filament as the canvas reference box, and rigidly scaling the pixel coordinates of each filament according to a fixed ratio of physical size and image resolution, thus realizing pixel-level seamless alignment between the visualized heat map and the real physical cross section.

[0190] By precisely matching the actual coordinates of the filament with the color values ​​of the single filament loss to generate a two-dimensional loss cloud map, and simultaneously extracting the direction and angle of the magnetic field gradient feature quantity and superimposing vector arrows at the corresponding coordinates, the originally completely invisible microscopic electromagnetic field spatial evolution process is transformed into an intuitive and information-dense graphical representation. This multi-dimensional data fusion presentation method completely surpasses the output form of existing single numerical reports, enabling R&D personnel to instantly perceive the geometric distribution of loss distortion areas and the direction of magnetic field interference, and significantly shortening the optimization cycle of high-frequency transformer winding layout schemes.

[0191] Reference Figure 2 This is another embodiment of the present invention. Unlike the first embodiment, this embodiment provides a high-frequency transformer line Litz line loss calculation system, including a reference module, an analysis module, a correction module and an output module.

[0192] The reference module is used to obtain the multi-strand strand structure parameters of the Leeds wire and calculate the reference loss value of the Leeds wire based on the multi-strand strand structure parameters;

[0193] The analysis module is used to calculate the actual coordinates of each filament based on the parameters of the multi-strand twisted structure, and to calculate the superimposed magnetic field strength based on the actual coordinates of each filament and the current value. The superimposed magnetic field strength and the intrinsic magnetic field strength vector are summed and the modulus is obtained to obtain the local magnetic field strength. The ratio of the magnetic field difference between adjacent sampling points to the distance is calculated to obtain the rate of change of magnetic field strength. The average value of the rate of change of magnetic field strength at each angular position is used as the magnetic field gradient feature quantity.

[0194] The correction module is used to construct the impedance mapping function, substitute the magnetic field gradient characteristics into the impedance mapping function, and calculate the loss correction factor.

[0195] The output module is used to correct the baseline loss value using the loss correction factor, calculate the final total loss value, perform iterative correction based on the final total loss value, and then store and visualize the output.

[0196] This invention, by introducing the actual coordinates of the filament and a local micro-element model, accurately analyzes the magnetic field gradient characteristics around the filament, breaking through the limitations of the idealized uniform distribution assumption. It realistically restores the local magnetic field non-uniformity at the microscale under high frequency. By innovatively constructing an impedance mapping function that substitutes the magnetic field gradient characteristics, a deep mapping mechanism between the magnetic field gradient and impedance loss is established, realizing refined single-filament-level correction of the reference loss. This significantly reduces high-frequency calculation errors. Combined with a closed-loop iterative correction mechanism based on measured data, it can dynamically optimize and approximate the real physical loss, eliminating calculation deviations. Ultimately, it not only outputs a high-precision total loss value but also achieves two-dimensional visualization of the filament-level loss distribution and magnetic field gradient, providing precise data support for the local hotspot location and targeted structural optimization of the Litz wire winding of high-frequency transformers.

[0197] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0198] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the protection scope of the present invention.

Claims

1. A method for calculating the loss of a high-frequency transformer line (Litz wire), characterized in that, Includes the following steps: Step S1: Obtain the multi-strand strand structure parameters of the Litz wire, and calculate the reference loss value of the Litz wire based on the multi-strand strand structure parameters; Step S2: Calculate the actual coordinates of each filament based on the parameters of the multi-strand twisted structure, and calculate the superimposed magnetic field strength based on the actual coordinates of each filament and the current value. Summate the superimposed magnetic field strength with the intrinsic magnetic field strength vector to obtain the local magnetic field strength. Calculate the ratio of the magnetic field difference between adjacent sampling points to the distance to obtain the magnetic field strength change rate. Use the average value of the magnetic field strength change rate at each angular position as the magnetic field gradient feature quantity. Step S3: Construct an impedance mapping function, substitute the magnetic field gradient characteristic into the impedance mapping function, and calculate the loss correction factor; Step S4: Correct the baseline loss value using the loss correction factor, calculate the final total loss value, perform iterative correction based on the final total loss value, and store and visualize the result.

2. The method for calculating the loss of a high-frequency transformer line as described in claim 1, characterized in that, Step S1 specifically includes: Obtain the parameters of the multi-strand twisted structure of Leeds wire; The parameters of the multi-strand stranded structure include the number of stranded layers, the number of filaments in each layer, the stranding direction, and the stranding pitch. Based on the number of strands and the number of filaments in each layer, a Cartesian coordinate system is established with the geometric center of the Litz wire cross-section as the origin, and a two-dimensional geometric model of the Litz wire cross-section is generated in the Cartesian coordinate system. The two-dimensional geometric model is divided into finite element mesh units according to a preset mesh size; Obtain the total current value, the total number of filaments, and the cross-sectional area of ​​a single filament of the Litz wire. Calculate the quotient of the total current value and the total number of filaments, and then divide it by the cross-sectional area of ​​a single filament to obtain the initial current density of a single filament. The initial current density is used as a boundary condition, and the boundary condition is assigned to the nodes of each finite element mesh element. The preset two-dimensional electromagnetic field solver is invoked, and the initial current density on the nodes of each finite element mesh element is used as input to solve the Maxwell equations of the two-dimensional geometric model to obtain the magnetic flux density of each finite element mesh element. The square of each magnetic flux density is multiplied by the area of ​​each finite element mesh cell, and the results of these multiplications are summed. The sum of these sums is then multiplied by the axial unit length of the Litz line, and the result is used as the reference loss value.

3. The method for calculating the loss of a high-frequency transformer line as described in claim 2, characterized in that, Step S2 specifically includes: Based on the twist pitch and number of twist layers of the Litz wire, a polar coordinate system is established with the geometric center of the Litz wire cross section as the origin, and the theoretical distribution radius of each layer of filaments is calculated. Based on the theoretical distribution radius of each layer of filaments and the number of filaments in each layer, the theoretical angular coordinates of each filament are calculated using the principle of equal angular division. Substitute the preset angle deviation coefficient and the preset radial deviation coefficient into the angle deviation correction function and the radial deviation correction function respectively to correct the theoretical angle coordinates and the theoretical distribution radius, and obtain the actual coordinates of the filament; Traverse each filament inside the Lids line and use the currently traversed filament as the target filament; Based on the actual coordinates of the target filament, an annular region is divided in the polar coordinate system with the geometric center of the target filament as the center and a preset distance as the radius. The annular region is used as a local micro-element model around the target filament. Obtain the actual coordinates of each filament other than the target filament and the current value of each filament. Substitute the actual coordinates of each filament, the current value of each filament and the actual coordinates of the target filament into the preset magnetic field superposition analytical formula for calculation, and output the superimposed magnetic field strength generated by the filaments other than the target filament at the position of the target filament. Obtain the current value and diameter of the target filament, calculate the ratio of the current value to the product of the filament diameter and pi, and use the ratio as the intrinsic magnetic field strength of the target filament. The vector component of the superimposed magnetic field strength is added to the vector component of the intrinsic magnetic field strength, and the modulus operation is performed on the added vector component to obtain the local magnetic field strength around the target filament. A predetermined number of angle positions are uniformly selected along the circumference of the target filament. A predetermined number of sampling points are selected along the radial direction at each angle position. The difference in magnetic field strength between adjacent radial sampling points at each angle position is calculated. Calculate the ratio of the difference in magnetic field strength to the distance between adjacent sampling points, and use the ratio as the rate of change of magnetic field strength at that angular position; Calculate the average value of the rate of change of magnetic field intensity at each angular position, and use the average value as the magnetic field gradient characteristic quantity of the target filament.

4. The method for calculating the loss of a high-frequency transformer line as described in claim 3, characterized in that, Step S3 specifically includes: An impedance mapping function is constructed, the magnetic field gradient characteristic quantity is substituted into the impedance mapping function, and the integrand in the impedance mapping function is discretized and summed using a numerical integration algorithm to obtain a numerical solution. Extract the real part of the numerical solution and obtain the resistance coefficient corresponding to the reference loss value. Use the resistance coefficient as a proportionality coefficient, calculate the product of the real part and the proportionality coefficient, and use the product result as a loss correction factor.

5. The method for calculating the loss of a high-frequency transformer line as described in claim 4, characterized in that, The construction of the impedance mapping function specifically includes: Obtain the type identifier of the fractional differential operator, and invoke the corresponding fractional differential operator calculation rule based on the type identifier; Obtain the upper limit of the operating frequency range of the Litz wire and the coupling strength value between the filaments, and calculate the product of the upper limit of the operating frequency range and the coupling strength value, and use the product result as the initial value of the fractional order. Using the magnetic field gradient feature as the base variable, the power of the magnetic field gradient feature value is calculated to obtain the kernel function; Obtain the time step of the historical change of the magnetic field, calculate the negative power of the exponent with the initial value as the base of the time step of the historical change of the magnetic field, and use the calculation result as the weighting coefficient of the memory effect. Multiply the memory effect weighting coefficient by the kernel function, and use the result of the multiplication as the integrand; The integrand is calculated according to the fractional differential operator calculation rules to obtain the impedance mapping function.

6. The method for calculating the loss of a high-frequency transformer line as described in claim 5, characterized in that, Step S4 specifically includes: Obtain the total number of filaments in the Litz wire and create a reference loss component variable equal to the total number of filaments; Traverse each finite element mesh element in the two-dimensional geometric model to obtain the center point coordinates, magnetic flux density, and area of ​​the currently traversed finite element mesh element; The distance between the coordinates of the center point and the geometric center coordinates of each filament in the two-dimensional geometric model is calculated, and the filament with the smallest distance is selected as the current filament. Calculate the product of the square of the magnetic flux density of the currently traversed finite element mesh cell and the area, and add the product result to the reference loss component variable corresponding to the current filament, and update the value of the reference loss component variable; After traversing each finite element mesh, the final value of the reference loss component variable corresponding to each filament is used as the reference loss component value of each filament. Calculate the product of the reference loss component value of each filament and the loss correction factor corresponding to each filament, and use the product result as the single filament loss value; The loss values ​​of each single filament are summed up, and the summation result is used as the total loss correction value. Calculate the product of the total loss correction value and the preset experimental calibration coefficient, and use the product result as the final total loss value; Iterative corrections are performed based on the final total loss value, and the results are stored and visualized.

7. The method for calculating the loss of a high-frequency transformer line as described in claim 6, characterized in that, The iterative correction specifically includes: The actual loss values ​​of the Litz wire under specific operating conditions were collected using a power analyzer. Calculate the difference between the actual loss value and the final total loss value, and divide the absolute value of the difference by the actual loss value to obtain the relative error value; The relative error value is compared with a preset error threshold. When the relative error value is greater than the preset error threshold, the parameter iteration step is executed to obtain the final output result; When the relative error value is less than or equal to the preset error threshold, the current fractional order and the current proportional coefficient are used as the final output result.

8. The method for calculating the loss of a high-frequency transformer line as described in claim 7, characterized in that, The parameter iteration steps specifically include: Get the current value of the fractional order as the first value, and get the current value of the proportional coefficient as the second value. Calculate the first partial derivative of the relative error value with respect to the fractional order, and calculate the second partial derivative of the relative error value with respect to the proportionality coefficient; Calculate the product of the first partial derivative and the preset first step length to obtain the first result, and calculate the first value minus the first result to obtain the updated fractional order; Calculate the product of the second partial derivative and the preset second step size to obtain the second result, and calculate the second value minus the second result to obtain the updated scaling factor; The final total loss value is recalculated based on the updated fractional order and the updated scaling factor to obtain the updated final total loss value. The absolute value of the difference between the updated final total loss value and the actual loss value is divided by the actual loss value to obtain the updated relative error value; The updated relative error value is compared with a preset error threshold. If the updated relative error value is greater than the preset error threshold, the parameter iteration step is re-executed. When the updated relative error value is less than or equal to the preset error threshold, the iteration stops and the currently updated fractional order and the updated proportional coefficient are used as the final output results.

9. The method for calculating the loss of a high-frequency transformer line as described in claim 8, characterized in that, The specific steps of storing and visualizing the output include: The actual coordinates of each filament, local magnetic field strength, magnetic field gradient characteristic quantity, loss correction factor, single filament loss value and final total loss value are written into the database according to the preset structure. Read the single filament loss value corresponding to the actual coordinates of each filament in the database, match the corresponding color value, and generate a two-dimensional loss image on the display terminal; Calculate the direction angle of the magnetic field gradient feature quantity in the database, and draw vector arrows by superimposing the direction angle at the actual coordinates of each filament.

10. A high-frequency transformer line Litz line loss calculation system, applied in the high-frequency transformer line Litz line loss calculation method as described in any one of claims 1-9, characterized in that, It includes a baseline module, a parsing module, a correction module, and an output module; The reference module is used to obtain the multi-strand strand structure parameters of the Leeds wire and calculate the reference loss value of the Leeds wire based on the multi-strand strand structure parameters. The analysis module is used to calculate the actual coordinates of each filament based on the parameters of the multi-strand twisted structure, and to calculate the superimposed magnetic field strength based on the actual coordinates of each filament and the current value. The superimposed magnetic field strength is summed with the intrinsic magnetic field strength vector to obtain the local magnetic field strength. The ratio of the magnetic field difference between adjacent sampling points to the distance is calculated to obtain the magnetic field strength change rate. The average value of the magnetic field strength change rate at each angular position is used as the magnetic field gradient feature quantity. The correction module is used to construct an impedance mapping function, and to calculate the loss correction factor by substituting the magnetic field gradient characteristic into the impedance mapping function. The output module is used to correct the baseline loss value using the loss correction factor, calculate the final total loss value, perform iterative correction based on the final total loss value, and then store and visualize the result.

Images