Design method for planar transformers with balanced current density

Abstract

This invention relates to the field of power electronics technology and provides a planar transformer design method for current density balancing, comprising the following steps: S1. When the total number of turns of the planar transformer is odd, the odd number of turns is distributed to two layers of PCB windings, such that one layer has one more turn than the other; S2. The total winding width required for each turn is calculated based on the target current density, the operating current of the turn, and the PCB copper foil thickness; S3. The winding width of the extra turn is evenly distributed to the two PCB layers and arranged in parallel between the two layers. This planar transformer design method for current density balancing, by achieving balanced current density in each turn, avoids localized overheating, thereby slowing down the aging process of the insulation material, reducing the risk of hot spot temperatures exceeding safe thresholds, and significantly improving the long-term operational stability and service life of the planar transformer.

Description

Technical Field

[0001] This invention belongs to the field of power electronics technology, and in particular relates to a planar transformer design method for current density balancing. Background Technology

[0002] Planar transformers, with their advantages of small size, low height, high power density, and low leakage inductance, are widely used in switching power supplies, communication power supplies, on-board chargers for new energy vehicles, and aerospace power supplies. The winding structure of these transformers is typically implemented using multi-layer printed circuit board traces, and the uniformity of current density distribution in each turn directly affects the equipment's temperature rise control, energy conversion efficiency, and long-term operational reliability. However, in practical design, when the total number of turns of the planar transformer is odd, such as common configurations like 5, 7, or 9 turns, existing technologies reveal significant shortcomings.

[0003] Traditional design methods simply distribute the odd number of turns across two layers of printed circuit board windings, inevitably resulting in one layer having one more turn than the other. This extra turn exists only within a single-layer printed circuit board, and its effective current-carrying cross-sectional area is only half that of the other turns. This causes an abnormally high current density in this extra turn, forming a localized hot spot. This localized overheating not only accelerates the aging process of the insulation material but may also cause the hot spot temperature to exceed the safe threshold, thus threatening the overall lifespan and operational stability of the transformer.

[0004] While some improvement solutions attempt to alleviate the problem by increasing the thickness of the entire copper layer or widening the overall trace width, these measures lack specificity. They cannot accurately compensate for the insufficient current-carrying capacity of a single-turn coil and also increase material costs and manufacturing complexity. The existing technology system lacks a systematic mechanism to evenly distribute the winding width of an extra turn across two printed circuit board layers and achieve parallel connection through vias to balance current density. This has resulted in the long-standing problem of current density imbalance under odd-numbered turn configurations remaining unresolved.

[0005] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0006] The purpose of this invention is to provide a planar transformer design method with balanced current density, aiming to solve the above-mentioned problems.

[0007] This invention is implemented as follows: a planar transformer design method for current density equalization includes the following steps:

[0008] S1. When the total number of turns of the planar transformer is odd, the odd number of turns is distributed to the two layers of PCB windings, such that one layer has one more turn than the other layer;

[0009] S2. Calculate the total winding width required for each turn of the coil based on the target current density, the operating current of the coil, and the thickness of the PCB copper foil;

[0010] S3. The winding width of the extra turn of the coil is evenly distributed across the two PCB layers and arranged in parallel between the two layers;

[0011] S4. The remaining coils are wound in each layer according to the calculated total width;

[0012] S5. Connect the corresponding coils in the two layers in parallel through the vias, so that the total winding width of the coil is equal to the winding width of the other coils.

[0013] In a further technical solution, step S3, the additional turn of the coil is evenly distributed across the two PCB layers. Specifically, a coil trace with a width half of the original calculated width is drawn in each layer and connected in parallel through vias so that the total width of the coil is equal to the original calculated width.

[0014] A further technical solution is that this method is applicable to planar transformers with a total number of 5, 7, or 9 turns.

[0015] Further technical solutions also include step S6: equalization and control of current density, specifically including:

[0016] S61. A multi-parameter monitoring unit is arranged in or near each turn of the planar transformer to collect the following four parameters of each turn in real time: current density, local temperature, magnetic field strength, and voltage drop across the coil.

[0017] S62. Substitute the collected four parameter data into the normalization formula for dimensionless normalization processing, and convert them into standardized values ​​under a unified scale, which are respectively denoted as: normalized value of current density, normalized value of local temperature, normalized value of magnetic field strength, and normalized value of voltage drop across the coil.

[0018] The normalization formula is: the normalized value equals the original parameter data minus the sample mean within the current sliding time window, and then divided by the sample standard deviation;

[0019] S63. Perform nonlinear weighted fusion on the normalized values ​​of the four parameters to generate the current density equalization fitting index for each turn of the coil. Its expression is: weighted average divided by (1 plus the geometric mean of the product of the absolute values ​​of the four normalized parameters).

[0020] The weighted average is the sum of four factors: the normalized value of current density multiplied by the current density weight coefficient, the normalized value of local temperature multiplied by the local temperature weight coefficient, the normalized value of magnetic field strength multiplied by the magnetic field strength weight coefficient, and the normalized value of voltage drop across the coil multiplied by the voltage drop weight coefficient. The sum of each weight coefficient is 1.

[0021] The geometric mean of the product of the absolute values ​​of the four normalized parameters is obtained by multiplying the normalized absolute values ​​of current density, local temperature, magnetic field strength, and voltage drop by the fourth power.

[0022] S64. Compare the calculated current density equalization fitting index with the preset equalization threshold. If the fitting index is greater than or equal to the equalization threshold, it is determined that there is a risk of current density imbalance in the coil. Generate the corresponding control command according to the preset control strategy.

[0023] In a further technical solution, the current density weighting coefficient, local temperature weighting coefficient, magnetic field strength weighting coefficient, and voltage drop weighting coefficient are dynamically adjusted based on the operating state, load conditions, or historical fault data of the planar transformer; the nonlinear coefficient is preset to a fixed value based on the rated power or heat dissipation conditions of the planar transformer, or is adaptively adjusted during operation.

[0024] In a further technical solution, the control command includes at least one of the following:

[0025] Adjust the parallel via connection status to change the effective number of parallel layers of the coil.

[0026] Adjust the operating frequency;

[0027] Switching or adjusting the conduction path of the coil in this layer can achieve current shunting;

[0028] Output the warning signal to the system's host computer.

[0029] Further technical solutions, instructions The adjustment amount for the effective number of parallel layers is calculated using the following formula:

[0030] The target adjustment amount is equal to the maximum adjustable parallel layer variation range multiplied by (the difference between the fitting index and the equilibrium threshold divided by the difference between the maximum allowable fitting index and the equilibrium threshold) and then rounded down.

[0031] After adjustment, the equivalent winding width of the coil is equal to the single-layer winding width reference value multiplied by the effective number of parallel layers after adjustment.

[0032] Further technical solutions, instructions In this case, the operating frequency is calculated using the following formula:

[0033] The frequency adjustment amount is equal to the frequency adjustment coefficient multiplied by (the difference between the fitting index and the equilibrium threshold, divided by the difference between the maximum allowable fitting index and the equilibrium threshold, and the resulting quotient) and then multiplied by the current operating frequency; the adjusted operating frequency is equal to the current operating frequency minus the frequency adjustment amount, and is limited to not being lower than the minimum operating frequency allowed by the system.

[0034] In a further technical solution, the control command is executed in stages according to the magnitude of the current density equalization fitting index:

[0035] When the fit index is greater than or equal to the equilibrium threshold and less than the first-level control threshold, the instruction is executed. Output early warning signal;

[0036] When the fit index is greater than or equal to the first-level control threshold and less than the second-level control threshold, the instruction is executed. Adjust the parallel via connection status;

[0037] When the fit index is greater than or equal to the second-level control threshold, the instruction is executed simultaneously. ,instruction and instructions Joint regulation and control should be carried out.

[0038] The equilibrium threshold is less than the first-level control threshold, the first-level control threshold is less than the second-level control threshold, and the second-level control threshold is not greater than the maximum allowable fit index.

[0039] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0040] This invention provides a planar transformer design method for balancing current density. By evenly distributing the winding width of an extra turn of coil across two PCB layers and arranging them in parallel, and then connecting them in parallel through vias, the total winding width of this extra turn of coil is equal to the winding width of other turns of coil. This solves the problem of excessively high local current density caused by the distribution of odd number of turns (such as 5 turns, 7 turns, 9 turns) from the design source, and avoids local hot spots caused by single-layer winding.

[0041] The present invention provides a planar transformer design method with balanced current density. Since the current density of each coil is balanced, local overheating is avoided, thereby slowing down the aging process of the insulation material, reducing the risk of hot spot temperature exceeding the safety threshold, and significantly improving the long-term operational stability and service life of the planar transformer.

[0042] This invention provides a planar transformer design method for current density balancing. By arranging multi-parameter monitoring units to collect current density, local temperature, magnetic field strength, and voltage drop across the coil in real time, and using nonlinear weighted fusion to generate a current density balancing fitting index, it can dynamically identify unbalance risks and generate control commands (such as adjusting the parallel via connection status, adjusting the operating frequency, switching the conduction path, etc.) according to preset strategies, thereby realizing hierarchical control based on the multi-parameter fusion index.

[0043] This invention provides a planar transformer design method for current density balancing. Based on the magnitude of the current density balancing fitting index, it executes early warning, local control, or joint control commands in stages. This method can provide timely early warning when there is slight imbalance and can quickly intervene with multiple means when there is severe imbalance, avoiding unnecessary over-adjustment and improving the precision and intelligence of the control. Attached Figure Description

[0044] Figure 1 This is a flowchart illustrating the design method of the present invention;

[0045] Figure 2 This is a flowchart illustrating the current density equalization control step S6 in this invention.

[0046] Figure 3 This is a schematic diagram illustrating the control principle of hierarchical execution based on the magnitude of the current density equalization fitting exponent. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this invention clearer, 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 merely illustrative and not intended to limit the invention.

[0048] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0049] like Figure 1 As shown, a planar transformer design method for current density equalization provided in one embodiment of the present invention includes the following steps:

[0050] When the total number of turns in a planar transformer is odd, the first step is to distribute these odd-numbered turns across the two PCB winding layers. A common distribution method is to have one more turn in one layer than the other. The reason for this is that in a two-layer structure, the odd-numbered turns cannot be evenly distributed, inevitably resulting in one layer having an extra turn. If left unaddressed, this extra turn only has the current-carrying cross-sectional area of ​​a single layer, and its current density is twice that of the other turns, leading to localized overheating. For example, if the total number of turns is 5, then 3 turns can be allocated to the first PCB layer and 2 turns to the second PCB layer. This distribution is a fundamental step in planar transformer design, providing the structural basis for subsequent current density balancing.

[0051] Next, based on the target current density required by the design, the total winding width required for each turn of the coil is calculated. This calculation aims to ensure that the current density of each turn of the coil can be maintained within a safe and efficient range when carrying the expected current. For example, a maximum allowable current density value can be set, and combined with the maximum operating current of the coil, the required minimum conductor cross-sectional area can be calculated, and then the corresponding winding width can be determined based on the copper thickness of the PCB. This calculation provides a quantitative basis for subsequent coil routing design. Specifically, the total winding width required for each turn of the coil is obtained by dividing the operating current of that turn of the coil by the product of the target current density and the PCB copper foil thickness.

[0052] Subsequently, for the extra turn of coil after allocation, its winding width is evenly distributed across the two PCB layers and arranged in parallel between the two layers. The number of parallel vias is greater than or equal to the length of the turn of coil, and it is recommended that the number of parallel vias be no less than two and the spacing between adjacent vias be ≤5 mm. For example, if the calculated total winding width of the extra turn of coil is W, a trace with a width of W / 2 can be drawn on the first PCB layer, and a trace with a width of W / 2 can be drawn on the second PCB layer. These two traces together constitute the extra turn of coil. This evenly distributed arrangement provides the extra turn of coil with the same total current-carrying cross-sectional area potential as the other turns of coil from a physical structure perspective, avoiding the problem of insufficient cross-sectional area caused by a single-layer arrangement.

[0053] For the remaining coil turns, they are wound in each layer according to the calculated total width. Since these coil turns are usually designed with one turn per layer, winding them directly according to the total width calculated in step S2 ensures that their current carrying capacity meets the design requirements and that their current density meets the design expectations, thus not disrupting the overall current density balance.

[0054] Finally, corresponding coils in the two layers are connected in parallel using vias, making the total winding width of this coil equal to the winding width of the other coils. For example, multiple vias are placed between the two layers of traces of an extra coil, electrically connecting these two traces with a width of W / 2 to form a parallel loop. Through this parallel connection, the effective current-carrying cross-sectional area of ​​this coil is doubled, making its total winding width and total current-carrying cross-sectional area consistent with the remaining coils. Thus, the current density of all coils is balanced, solving the problem of uneven current density after an odd number of turns is distributed from a design perspective.

[0055] In this embodiment, the planar transformer is a type of transformer that uses a printed circuit board (PCB) as the winding carrier. Its characteristic lies in the fact that the windings are integrated on the PCB layer in a planar spiral or concentric ring structure. Compared to traditional wound transformers, it has advantages such as small size, low height, high power density, and low leakage inductance, and is widely used in switching power supplies, communication power supplies, new energy vehicle on-board chargers, and aerospace power supplies.

[0056] PCB windings refer to the primary and secondary coils of a transformer formed by etching copper foil traces onto a multilayer printed circuit board. This structure allows for precise control over the winding geometry, which is beneficial for achieving high-frequency characteristics and compact design.

[0057] The number of turns refers to the number of turns of a coil, a fundamental parameter in transformer design that directly affects the transformer's turns ratio and inductance. In planar transformers, one turn of the coil is typically composed of one or more traces on the PCB.

[0058] Current density refers to the amount of current flowing per unit cross-sectional area and is a key indicator of a conductor's current-carrying capacity. In a planar transformer, the uniformity of current density across all turns directly affects the transformer's temperature rise, efficiency, and long-term operational reliability. Excessive current density can lead to localized overheating, accelerate the aging of insulation materials, and consequently shorten the transformer's lifespan.

[0059] A via is a metallized hole in a printed circuit board used to connect conductive traces on different layers. In planar transformer design, vias are often used to achieve electrical connections between windings on different layers, such as connecting multiple layers of traces in parallel to increase the effective current-carrying cross-sectional area.

[0060] This application provides a refined current density balancing compensation mechanism. Specifically, the winding width of the extra turn is evenly distributed across two PCB layers and arranged in parallel. Corresponding coils in these two layers are connected in parallel via vias. This design ensures that the total effective winding width and current-carrying cross-sectional area of ​​the extra turn, which was originally arranged in a single layer, remain consistent with the remaining turns. This method overcomes the lack of a systematic compensation mechanism in existing technologies, avoiding the potential cost and manufacturing difficulty increases associated with traditional methods that crudely increase copper thickness or widen traces, and which fail to accurately address local imbalances. This application solves the current density imbalance problem caused by odd-numbered turn distribution from the design stage, achieving more precise and efficient current density balancing, and significantly improving the temperature rise characteristics, efficiency, and long-term operational reliability of the planar transformer.

[0061] In a preferred embodiment of the present invention, in step S3, the winding width of the extra turn of the coil is evenly distributed to the two PCB layers. Specifically, a coil trace with a width of half the original calculated width is drawn in each layer and connected in parallel through vias so that the total width of the coil is equal to the original calculated width.

[0062] In this embodiment, the current-carrying capacity of the multi-turn coil, originally concentrated on a single-layer PCB, is distributed across two layers to avoid excessively high local current density. This equal distribution ensures that the current share carried by each layer is relatively balanced, thus laying the physical foundation for subsequent parallel connections and current density balancing.

[0063] "Drawing" refers to the placement and routing in PCB design software, defining the geometry of the coils on the copper foil layer of the PCB. This can be done with the assistance of automated routing tools or by engineers drawing the coils manually with precision. The coil traces are the conductive paths that carry current, and their width directly affects the current-carrying capacity and current density.

[0064] Setting the width to "half the original calculated width" is the core method for achieving even distribution. The "original calculated width" is the total winding width calculated based on the target current density of a single-turn coil. This design ensures that half-width traces are formed on both PCB layers, providing a physical carrier for subsequent parallel connections.

[0065] A "via" is a metallized hole that connects conductive patterns on different layers of a PCB, and it is a key structure for achieving interlayer electrical connections. Vias can take the form of through holes, blind vias, or buried vias, and the specific choice depends on the number of PCB layers, manufacturing process, and cost requirements. For example, standard through-hole technology can be used to achieve electrical connections between upper and lower layer traces through drilling and electroplating; or micro blind via technology using laser drilling can be used to achieve higher density connections.

[0066] "Parallel connection" refers to electrically connecting half-width coil traces on two PCB layers to form a complete current path. This parallel connection allows the current to be freely distributed between the two layers of traces, thereby effectively utilizing the current-carrying capacity of the two copper foil layers.

[0067] Ultimately, through the above-mentioned equalization and parallel operation, the equivalent current-carrying cross-sectional area (or equivalent width) of the extra-turn coil is ensured to be consistent with the total width calculated for other turns during the design, thereby achieving a balance of current density.

[0068] As a preferred embodiment of the present invention, the method is applicable to planar transformers with a total number of 5, 7, or 9 turns.

[0069] In this embodiment, the solution of this application clarifies that the current density balancing design method is applicable to planar transformers with a total number of 5, 7, or 9 turns, thus providing designers with clear application boundaries and guidance. In the design process of a planar transformer, the total number of turns of the transformer must first be determined according to circuit requirements. When the total number of turns is exactly an odd number of 5, 7, or 9 turns, the designer can clearly know that the aforementioned current density balancing design method (including distributing the odd number of turns to two PCB winding layers, so that one layer has one more turn than the other; calculating the total winding width required for each coil based on the target current density; distributing the winding width of the extra coil equally to the two PCB layers, arranging them in parallel between the two layers, and ensuring that the number and position of vias are designed so that the current distribution deviation between the two layers does not exceed 10%; winding the remaining coils in each layer according to the calculated total width; connecting the corresponding coils in the two layers in parallel through vias so that the total winding width of that coil is equal to the winding width of the other coils, etc.) is fully applicable and recommended.

[0070] This clearly defined scope of application allows the aforementioned complex current density balancing control mechanism to be precisely applied to odd-turn configurations, which are the most common and prone to current density imbalance problems. It avoids blind experimentation or incorrect application by designers when unsure of the method's applicability, thereby improving design efficiency and success rate. It is precisely because of this clear applicability limitation that the aforementioned refined width-sharing and parallel compensation design for odd-turn allocation can effectively address current density imbalance problems in these specific scenarios, thereby significantly improving the temperature rise characteristics, efficiency, and long-term reliability of these commonly used planar transformer specifications.

[0071] like Figure 2 As shown, in a preferred embodiment of the present invention, step S6 is further included: equalizing and regulating the current density, specifically including:

[0072] S61. Multi-parameter monitoring units are arranged in or near each turn of the planar transformer to collect the following four parameters of each turn in real time: current density. Local temperature magnetic field strength Voltage drop across the coil ;

[0073] S62. Substitute the collected four parameter data into the normalization formula for dimensionless normalization, converting them into standardized values ​​under a unified scale, denoted as: normalized current density value. Local temperature normalized value Normalized value of magnetic field strength Normalized value of voltage drop across the coil The normalization formula is:

[0074]

[0075] in, This refers to raw parameter data such as current density, local temperature, magnetic field strength, or voltage drop across the coil. This represents the sample mean of the original parameter data within the current sliding time window. The sample standard deviation of the original parameter data within the sliding time window;

[0076] The sliding time window has a length of 0.1 seconds to 10 seconds and a sampling frequency of not less than 1 kHz, meaning each window contains no fewer than 100 sampling points; the sliding time window is not full or the sample standard deviation... ( When taking 0.5% of the corresponding sensor range, the factory-preset initial standard deviation value is used instead. The factory-preset initial standard deviation value is obtained through offline calibration or simulation.

[0077] S63. Perform nonlinear weighted fusion on the normalized values ​​of the four parameters to generate a current density equalization fitting index for each turn of the coil. Its expression is:

[0078]

[0079] in: The current density weighting coefficient is used. For local temperature weighting coefficients, These are the magnetic field strength weighting coefficients. Let be the voltage weighting factor, and satisfy . ;

[0080] These are nonlinear coefficients. , used to adjust the suppression strength under extreme coupling of multiple parameters;

[0081] The geometric mean of the absolute values ​​of the four normalized parameters is used to characterize the coupling degree of simultaneous deviation of multiple parameters. The geometric mean of the absolute values ​​of the four normalized parameters is defined as the fourth root of the product of the absolute values ​​of each normalized parameter. When the absolute value of any parameter is zero, the geometric mean is zero. At this time, the nonlinear term degenerates to 1, and the current density equalization fitting index is simplified to a weighted linear average value, which still maintains an effective characterization of the imbalance risk and does not lead to the failure of the control function.

[0082] S64. Calculated current density equalization fitting index With the preset equalization threshold If a comparison is made, If the current density of the coil is unbalanced, a corresponding control command is generated based on the preset control strategy.

[0083] In this embodiment, multi-parameter monitoring units are arranged in or near each turn of the planar transformer coil. These monitoring units can be integrated onto a PCB board. For example, the current density sensor can be a miniature Hall effect sensor or a current sensor based on the resistive voltage divider principle; the local temperature sensor can be a thermistor or thermocouple; the magnetic field strength sensor can be a magnetoresistive sensor or a miniature coil; and the voltage drop across the coil can be directly measured using a differential amplifier. Alternatively, the monitoring unit can be a standalone sensor module, integrated or tightly coupled to the planar transformer coil via flexible circuitry or miniature connectors. For example, a fiber optic temperature sensor can be used for local temperature monitoring, and a miniature fluxgate sensor can be used for magnetic field strength monitoring. These monitoring units acquire the current density of each turn of the coil in real time. Local temperature magnetic field strength Voltage drop across the coil Four parameters. Real-time acquisition ensures that the data can promptly reflect the current operating status of the transformer and potential imbalance risks. For example, the monitoring unit converts analog signals into digital signals through a high-speed data acquisition module and transmits the data to the processing unit at a preset sampling frequency (e.g., not less than 1kHz). These four parameters comprehensively reflect the operating status of the coil from three dimensions: electrical, thermal, and magnetic. The combination of multiple parameters can avoid misjudgments caused by monitoring a single parameter, ensure the comprehensiveness of the judgment on the imbalance state, and provide a reliable data foundation for subsequent control.

[0084] Subsequently, the collected four parameter data were substituted into the normalization formula for dimensionless normalization, converting them into standardized values ​​on a uniform scale. The normalization formula adopted... In the form of, This represents the sample mean of the original parameter data within the current sliding time window. This represents the sample standard deviation of the original parameter data within the sliding time window. This Z-score normalization method transforms the data into a distribution with a mean of 0 and a standard deviation of 1, better reflecting the degree of deviation of the data from its historical distribution. The length of the sliding time window is 0.1 seconds to 10 seconds, and the sampling frequency is no less than 1 kHz, meaning each window contains no fewer than 100 sampling points. This ensures that under dynamic operating conditions, the normalized parameters can reflect the current state in a timely and accurate manner, while avoiding interference from transient noise.

[0085] Based on this, the normalized values ​​of the four parameters are nonlinearly weighted and fused to generate the current density equalization fitting index for each turn of the coil. Nonlinear weighted fusion aims to comprehensively consider the influence of multiple parameters on current density equalization and introduces a nonlinear term to suppress evaluation bias under extreme coupling conditions. Specifically, , , , Let be the weight coefficients, and satisfy . These weighting coefficients reflect the relative importance of different parameters to the current density balance. The sum of 1 ensures the scale consistency of the fusion results. The weighting coefficients can be set and optimized based on expert experience, historical data analysis, or machine learning algorithms.

[0086] The nonlinear coefficient is used to adjust the suppression strength when multiple parameters are extremely coupled. When multiple parameters deviate significantly at the same time, the fitting index is appropriately suppressed by the nonlinear term to prevent the evaluation results from being overly sensitive or distorted. The nonlinear coefficient can be preset according to the rated power of the transformer, heat dissipation conditions, or through simulation experiments.

[0087] The geometric mean of the absolute values ​​of the four normalized parameters is used to characterize the degree of coupling when multiple parameters deviate simultaneously. The geometric mean can effectively reflect the degree of deviation of multiple parameters at the same time. In particular, when all parameters deviate, its value will increase significantly, thereby triggering nonlinear suppression.

[0088] Finally, the calculated current density equalization fitting index is used. With the preset equalization threshold If a comparison is made, If the current density is unbalanced, the coil is judged to have a risk of current density imbalance, and a corresponding control command is generated according to the preset control strategy. This comparison provides a clear judgment standard for identifying the risk of current density imbalance and initiating corrective measures to restore current density balance. The control strategy can be based on a rule-based expert system, which searches for the corresponding control command based on the magnitude of the fitting exponent and a preset threshold, or it can be based on a closed-loop controller using PID control or other advanced control algorithms, which generates continuous or discrete control signals based on the deviation of the fitting exponent.

[0089] For example, in a planar transformer with a total of 5 turns, the static design is first performed according to the method described above. The winding width of the extra turn is evenly distributed across two PCB layers, arranged in parallel between the two layers and connected in parallel vias, ensuring that the total winding width of this extra turn is equal to the winding width of the other turns. Building upon this, to achieve dynamic equalization during operation, a micro-sensor array can be integrated near the PCB traces of each turn as a multi-parameter monitoring unit. For example, a MEMS Hall sensor can be used to measure current density and magnetic field strength, a thin-film thermistor can be used to measure local temperature, and a high-precision differential amplifier and analog-to-digital converter can be used to measure the voltage drop across the coil. These sensors acquire data in real time and transmit the data to a high-performance embedded microcontroller.

[0090] The microcontroller receives data at a sampling frequency of 2 kHz and maintains a data buffer with a length of 2000 sampling points (corresponding to a 1-second sliding time window). When new sampled data arrives, the microcontroller calculates the sample mean of each parameter within the sliding time window in real time. and sample standard deviation And according to the normalization formula Calculate the normalized value of the current density Local temperature normalization value Normalized value of magnetic field strength and the normalized value of the voltage drop across the coil Subsequently, the microcontroller uses its built-in floating-point unit (FPU) to execute a nonlinear weighted fusion algorithm to calculate the current density equalization fitting index for each turn of the coil. Among them, the weighting coefficients , , , It can be preset according to the design characteristics and application scenarios of the transformer, for example =0.4, =0.3, =0.2, =0.1, and the nonlinear coefficient λ can be preset to 1. Finally, the microcontroller will calculate the fitting index. With the preset equalization threshold Compare. If Exceed The microcontroller will generate corresponding control instructions according to the preset control strategy. For example, it can output signals through the digital I / O port to trigger an external programmable switch array to adjust the parallel via connection state of the coil, thereby changing its effective parallel layer number, or send a warning signal to the system host computer.

[0091] As a preferred embodiment of the present invention, the current density weighting coefficient Local temperature weighting coefficient Magnetic field strength weighting coefficient Voltage reduction weighting factor Dynamic adjustments are made based on the operating status, load conditions, or historical fault data of the planar transformer.

[0092] The nonlinear coefficient The value is preset to a fixed value based on the rated power or heat dissipation conditions of the planar transformer, or it is adaptively adjusted during operation.

[0093] The sample mean and sample standard deviation of each original parameter data are updated synchronously in real time to ensure the responsiveness of the fitting index to the current working conditions.

[0094] In this embodiment, the current density weighting coefficient Local temperature weighting coefficient Magnetic field strength weighting coefficient Voltage reduction weighting factor Dynamic adjustment refers to modifying these coefficients during transformer operation based on real-time or historical data to more accurately reflect the contribution of each parameter to current density imbalance under current operating conditions. One implementation involves the system pre-setting multiple operating modes, such as light load, heavy load, high temperature, and low temperature, each corresponding to a pre-calibrated set of weighting coefficients. When a change in transformer operating status or load conditions is detected, the system automatically switches to the corresponding operating mode and loads the corresponding weighting coefficients. Another implementation involves the system utilizing machine learning algorithms to analyze historical operating and fault data, learn the nonlinear relationship between each parameter and current density imbalance, and predict and adjust the weighting coefficients in real time. For example, when historical data shows that local temperature has a stronger indicative effect on imbalance under high temperature and heavy load conditions, the system will correspondingly increase the local temperature weighting coefficient. .

[0095] Nonlinear coefficients Used to adjust the suppression strength under extreme multi-parameter coupling, its value directly affects the current density equalization fitting index. Sensitivity to extreme conditions. A preset fixed value means that during the transformer design phase, an optimal fixed value can be determined through simulation or experimentation based on the transformer's inherent characteristics such as rated power, heat dissipation capacity, and the heat resistance rating of the insulation material. The value remains constant throughout the entire operating cycle. For example, for transformers with strong heat dissipation capabilities, a relatively small value can be set. A value that allows for greater parameter fluctuations is set; for transformers with weak heat dissipation capabilities, a relatively large value is set. This value helps to suppress extreme coupling situations earlier.

[0096] Adaptive adjustment refers to the system's ability to dynamically adjust its operation during transformer operation based on real-time monitoring of the overall transformer operating status, such as ambient temperature, cooling system efficiency, and load rate, or on changes in transformer characteristics predicted by an internal model. Value. For example, when a transformer operates at high temperatures for an extended period, its heat dissipation capacity may decrease. In this case, the system can adaptively increase the value. The values ​​are used to improve sensitivity to extreme coupling of multiple parameters, thereby enabling earlier detection of potential risks. The sample mean of each original parameter data is... and sample standard deviation Synchronous real-time updates refer to the system's immediate recalculation of the sample mean of all data points within the current window as new data points enter the sliding time window and the oldest data points are removed from the window during each sampling period. and sample standard deviation This can be achieved by maintaining a first-in, first-out queue that stores data points within a sliding time window. Each time an update occurs, an old data point is removed from the head of the queue, a new data point is added to the tail of the queue, and then the statistics are recalculated based on the data in the queue.

[0097] In a preferred embodiment of the present invention, the control command includes at least one of the following:

[0098] Adjust the parallel via connection status to change the effective number of parallel layers of the coil. ;

[0099] Adjust the operating frequency ;

[0100] Switching or adjusting the conduction path of the coil in this layer can achieve current shunting;

[0101] Output the warning signal to the system's host computer.

[0102] In this embodiment, adjusting the parallel via connection state changes the effective number of parallel layers of the coil. This aims to directly adjust the equivalent current-carrying cross-sectional area by altering the coil's physical connection structure, thereby affecting the current density. When the current density of a particular coil is too high, increasing its effective number of parallel layers increases its equivalent conductive area, reducing the current density; conversely, if the current density is too low, the effective number of parallel layers can be reduced. This can be achieved by reserving multiple parallel via connection points in the PCB design and using microelectromechanical system switches or small relay arrays to selectively close or open some via connections according to control commands.

[0103] Adjusting the operating frequency is crucial because the operating frequency of a planar transformer significantly affects the current distribution within the coils, primarily through the skin effect and proximity effect. By adjusting the operating frequency, the intensity of these effects can be altered, thereby optimizing the current distribution across the conductor cross-section and achieving a more balanced current density. For example, lowering the frequency can reduce the skin effect, resulting in a more uniform current distribution within the conductor. This can be achieved by controlling the pulse width modulation signal generator or resonant controller in the transformer drive circuit to change its output frequency in real time.

[0104] Switching or adjusting the conduction path of the coil in this layer achieves current shunting. The aim is to redistribute current by changing the path through the coil, thereby reducing the current density of an overloaded coil. When the current density of a particular coil turn abnormally increases, a portion of its current can be diverted to a preset backup path or an adjacent, less loaded coil path to alleviate its burden. This can be achieved by reserving multiple parallel coil traces in the PCB design and configuring high-speed electronic switches to selectively connect or disconnect these paths according to control commands, thus realizing dynamic current shunting.

[0105] Sending early warning signals to the system's host computer is a non-intrusive control method, primarily used to alert operators or higher-level control systems that the planar transformer currently presents a potential risk of current density imbalance, but has not yet reached the point requiring immediate active physical or parameter adjustments. This provides a window of opportunity for manual intervention, inspection, or preventative maintenance. The early warning information can be sent to the connected host computer, monitoring system, or human-machine interface via a standard communication interface, displaying prompts on the interface using text, color, or sound.

[0106] As a preferred embodiment of the present invention, the instructions Adjustment amount of effective parallel layer Calculate using the following formula:

[0107]

[0108] in: To maximize the adjustable range of parallel layer number, The value can be 1 or 2;

[0109] The preset maximum allowable fitting index corresponds to the most severe unbalanced state allowed by the transformer and is determined through finite element simulation or factory calibration. This is the activation threshold; This is the floor function; The target adjustment amount is the number of effective parallel layers; ;

[0110] The equivalent winding width of the coil after adjustment It becomes:

[0111]

[0112] in This is the baseline value for the single-layer winding width. The adjusted effective number of parallel layers is equal to the current effective number of parallel layers plus the target adjustment amount of the effective number of parallel layers.

[0113] In this embodiment, the adjustment amount of the effective parallel layer number is as follows. The calculation formula is used to quantify the adjustment required to the effective parallel layer number of the coil when the current density is uneven. This formula comprehensively considers the current density balance fitting index. Start-up threshold Maximum allowable fit index and the range of maximum adjustable parallel layer number This is used to determine an integer adjustment step size. This calculation can be implemented in a microcontroller or digital signal processor inside the planar transformer, by embedding the formula logic into the firmware through programming, receiving parameters in real time and outputting the adjustment amount; or it can be implemented in an external monitoring system or host computer software, obtaining the transformer operating parameters through a data interface, performing calculations, and then sending the adjustment command to the transformer control unit.

[0114] in, The maximum step size that can be changed in the number of effective parallel layers in a single control operation is defined and limited to 1 or 2, in order to control the fineness and magnitude of the control and avoid over- or under-control. It can be used as a system design parameter and fixed in the controller's configuration register when the transformer leaves the factory, preset according to the transformer's design characteristics and application scenarios; or it can be dynamically configured or adjusted through the host computer software during system operation based on the transformer's historical operating data, aging degree, or specific operating condition requirements.

[0115] and It is a key reference value used to define the degree of current density imbalance. It is the minimum level of imbalance that triggers regulatory actions, and This represents the limit of unbalanced state that the system can withstand. These thresholds can be determined through detailed finite element simulation analysis of the planar transformer, simulating heat distribution, stress, etc., under different degrees of unbalance to find the critical point; alternatively, they can be determined through actual physical prototype testing and factory calibration, gradually increasing the degree of unbalance in a controlled environment and recording the fit index value before the transformer performance degrades or fails. The floor function is used to convert the calculated floating-point result into an integer, ensuring that the adjustment amount is an actual operable integer layer change. In digital circuit or microcontroller programming, the floor operation can be implemented using standard mathematical library functions; alternatively, it can be implemented through bitwise operations or custom integer truncation logic. The equivalent winding width of the coil after adjustment. The calculation formula is used to calculate the equivalent winding width that the coil should have after adjustment. It uses a single-layer winding width reference value... Multiply by the adjusted effective number of parallel layers This is used to determine the new total current-carrying cross-sectional area, thereby achieving the goal of current density balance. During the design phase of the planar transformer, the combination of physical trace widths corresponding to different effective parallel layers can be determined in advance, and the corresponding physical connection method can be selected during adjustment; alternatively, the actual number of PCB layers in operation can be changed by dynamically switching the switching elements in the parallel path, thereby physically achieving the adjustment of the equivalent winding width. It refers to the standard width of each PCB trace that makes up the coil. It is a basic parameter in the design and is determined during the manufacturing of the planar transformer. It is a fixed parameter in PCB design. The actual number of parallel layers that the coil should have after adjustment calculations directly affects the total current-carrying capacity of the coil. This is determined by adding the calculated adjustment amount to the initial design number of layers. The specific implementation can be achieved by controlling the opening or closing of parallel vias.

[0116] As a preferred embodiment of the present invention, the instructions In the middle, operating frequency Calculate using the following formula:

[0117]

[0118] in This is the current operating frequency; This is the frequency adjustment coefficient. ;

[0119] The adjusted operating frequency is: ;

[0120] and limit ,in The minimum operating frequency allowed by the system. The value is set to 40%-60% of the rated frequency. When the fitting index exceeds the equalization threshold, it is only applied at operating frequencies higher than [the specified frequency]. If the frequency is too low, the frequency adjustment operation is executed; otherwise, the frequency adjustment command is skipped to avoid further aggravation of the current density imbalance caused by core saturation or harmonic distribution distortion due to excessively low frequency.

[0121] In this embodiment, the operating frequency adjustment amount This refers to the adjustment to the current operating frequency required when a risk of current density imbalance is detected in a planar transformer. Its purpose is to influence the internal current distribution and losses by changing the transformer's operating frequency, thereby improving current density balance. This adjustment can be calculated based on a preset algorithm model or dynamically optimized using real-time monitoring data.

[0122] Frequency adjustment coefficient It is a function used to control the operating frequency adjustment amount. The scaling factor. Its range of values ​​is: This ensures that the magnitude of a single frequency adjustment is within a controllable and reasonable range, avoiding impacts on system stability due to excessive adjustments. This coefficient can be preset based on the specific application scenario, rated power, heat dissipation conditions, or the system's sensitivity to frequency changes of the planar transformer. For example, for systems with high frequency stability requirements, A smaller value can be chosen; for systems that allow for larger frequency fluctuations, A larger value can be selected.

[0123] Current operating frequency This refers to the real-time operating frequency of a planar transformer during current density equalization control. By introducing the current operating frequency as the benchmark for calculating the adjustment, the adjustment results can better reflect the actual operating conditions of the planar transformer, rather than relying on a fixed rated frequency, thereby improving the adaptability and accuracy of the control.

[0124] Adjusted operating frequency This refers to calculating the operating frequency adjustment amount. Then, by using the current operating frequency Subtract The resulting new operating frequency. This subtraction operation indicates that when a risk of current density imbalance is detected, the system tends to reduce the operating frequency to decrease the current, thereby alleviating the problem of excessive current density. This adjustment direction aligns with practical current density regulation logic.

[0125] Minimum operating frequency This is the minimum frequency threshold that the system allows the planar transformer to operate at. The purpose of setting this threshold is to prevent the operating frequency from being adjusted too low during frequency regulation, which could affect the normal operation of the planar transformer and its associated power system, or even lead to system failure. This minimum operating frequency... The value ranges from 40% to 60% of the rated frequency. For example, the specific value can be determined based on the transformer's design parameters, the characteristics of the core material, and the system's requirements for output power and efficiency.

[0126] like Figure 3 As shown, in a preferred embodiment of the present invention, the control command is based on a current density equalization fitting index. Size-based hierarchical execution:

[0127] when At that time, execute the instruction Output early warning signal;

[0128] when At that time, execute the instruction Adjust the parallel via connection status;

[0129] when At the same time, execute the instructions. ,instruction and instructions Joint regulation and control should be carried out.

[0130] in This is the first-level control threshold. This is the secondary control threshold. .

[0131] In this embodiment, "tiered execution" refers to execution based on the current density equalization fitting index. The numerical range of these ranges categorizes imbalance risks into different levels, with each level having its own preset control strategies and instructions. This mechanism ensures precise and targeted control, avoiding a "one-size-fits-all" approach. For example, measures of varying intensities, such as early warning, local adjustments, and global adjustments, can be implemented sequentially based on the risk level, from mild to severe. Alternatively, the system can dynamically adjust the thresholds for each level based on historical data or expert experience to adapt to the characteristics of the transformer in different operating stages or environments.

[0132] The instruction to "output a warning signal" refers to the current density equalization fitting index. When the risk level is low, the system does not immediately take physical control measures. Instead, it sends an alarm message to the operator or the host computer system. This instruction can manifest as displaying a warning message on the monitoring interface, issuing a prompt sound or flashing light via an audible and visual alarm, or sending a notification email or text message over the network. Its purpose is to alert the user to potential imbalance trends, but to allow the transformer to continue operating, avoiding unnecessary intervention.

[0133] The instruction to "adjust the connection status of parallel vias" refers to the current density equalization fitting index. When a medium-risk level is reached, the system triggers adjustments to the parallel via connection status of a specific coil. This can be achieved by controlling miniature relays or semiconductor switches integrated on the PCB, which selectively disconnect or connect certain parallel vias, thereby changing the effective number of parallel layers of that coil and thus adjusting its equivalent winding width and current carrying capacity. Alternatively, programmable logic devices can be used to control an array of digital switches, enabling dynamic configuration of the parallel paths.

[0134] "Simultaneous execution of joint control" refers to the simultaneous implementation of current density equilibrium fitting index. When a severe risk level is reached, the system will collaboratively activate multiple control measures to quickly and effectively resolve the imbalance problem. These include, but are not limited to: adjusting the parallel via connection status to change the coil width by altering the effective number of parallel layers; adjusting the operating frequency by controlling the PWM (Pulse Width Modulation) controller of the transformer drive circuit to change its output frequency, thereby affecting the transformer's flux density distribution and current distribution among the coil turns; and switching the conduction path. For planar transformers with redundant or switchable winding designs, this can be achieved by controlling switching elements to divert current from the overload path to a backup path, or by changing the specific flow direction of current in the multi-layer PCB windings to redistribute the current. The combined execution of these instructions aims to comprehensively intervene in current density from multiple dimensions to address more severe imbalances.

[0135] also, This is the first-level control threshold. These are the secondary control thresholds. These thresholds are preset key parameters used to classify the risk levels of current density imbalance. This is the minimum threshold for initiating regulation. When the fit index exceeds this value, the system begins to pay attention and may take action. This is the trigger point for primary regulation. When the fitting index exceeds this value, the system will upgrade from early warning to more aggressive local regulation. This is the trigger point for secondary control. When the fitting index exceeds this value, the system will initiate the most powerful joint control measures. The setting of these thresholds needs to be determined based on a large amount of experimental data, simulation analysis, and the safe operating boundary of the transformer to ensure the timeliness and effectiveness of control, while avoiding false triggering or over-control.

[0136] 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 planar transformer design method for current density equalization, characterized in that, Includes the following steps: S1. When the total number of turns of the planar transformer is odd, the odd number of turns is distributed to the two layers of PCB windings, such that one layer has one more turn than the other layer; S2. Calculate the total winding width required for each turn of the coil based on the target current density, the operating current of the coil, and the thickness of the PCB copper foil; S3. The winding width of the extra turn of the coil is evenly distributed across the two PCB layers and arranged in parallel between the two layers; S4. The remaining coils are wound in each layer according to the calculated total width; S5. Connect the corresponding coils in the two layers in parallel through the vias, so that the total winding width of the coil is equal to the winding width of the other coils.

2. The method according to claim 1, characterized in that, In step S3, the additional coil winding width is evenly distributed across the two PCB layers. Specifically, coil traces with a width half of the original calculated width are drawn in each layer and connected in parallel through vias so that the total width of the coil is equal to the original calculated width.

3. The method according to claim 1, characterized in that, This method is applicable to planar transformers with a total number of 5, 7, or 9 turns.

4. The method according to claim 3, characterized in that, It also includes step S6: equalization and regulation of current density, specifically including: S61. A multi-parameter monitoring unit is arranged in or near each turn of the planar transformer to collect the following four parameters of each turn in real time: current density, local temperature, magnetic field strength, and voltage drop across the coil. S62. Substitute the collected four parameter data into the normalization formula for dimensionless normalization processing, and convert them into standardized values ​​under a unified scale, which are respectively denoted as: normalized value of current density, normalized value of local temperature, normalized value of magnetic field strength, and normalized value of voltage drop across the coil. The normalization formula is: the normalized value equals the original parameter data minus the sample mean within the current sliding time window, and then divided by the sample standard deviation; S63. Perform nonlinear weighted fusion on the normalized values ​​of the four parameters to generate the current density equalization fitting index for each turn of the coil. Its expression is: weighted average divided by (1 plus the geometric mean of the product of the absolute values ​​of the four normalized parameters). The weighted average is the sum of four factors: the normalized value of current density multiplied by the current density weight coefficient, the normalized value of local temperature multiplied by the local temperature weight coefficient, the normalized value of magnetic field strength multiplied by the magnetic field strength weight coefficient, and the normalized value of voltage drop across the coil multiplied by the voltage drop weight coefficient. The sum of each weight coefficient is 1. The geometric mean of the product of the absolute values ​​of the four normalized parameters is obtained by multiplying the normalized absolute values ​​of current density, local temperature, magnetic field strength, and voltage drop by the fourth power. S64. Compare the calculated current density equalization fitting index with the preset equalization threshold. If the fitting index is greater than or equal to the equalization threshold, it is determined that there is a risk of current density imbalance in the coil. Generate the corresponding control command according to the preset control strategy.

5. The method according to claim 4, characterized in that, The current density weighting coefficient, local temperature weighting coefficient, magnetic field strength weighting coefficient, and voltage drop weighting coefficient are dynamically adjusted based on the operating state, load conditions, or historical fault data of the planar transformer; the nonlinear coefficient is preset to a fixed value based on the rated power or heat dissipation conditions of the planar transformer, or is adaptively adjusted during operation.

6. The method according to claim 5, characterized in that, The control command includes at least one of the following: Adjust the parallel via connection status to change the effective number of parallel layers of the coil. Adjust the operating frequency; Switching or adjusting the conduction path of the coil in this layer can achieve current shunting; Output early warning signals to the system's host computer.

7. The method according to claim 6, characterized in that, instruction The adjustment amount for the effective number of parallel layers is calculated using the following formula: The target adjustment amount is equal to the maximum adjustable parallel layer variation range multiplied by (the difference between the fitting index and the equilibrium threshold divided by the difference between the maximum allowable fitting index and the equilibrium threshold) and then rounded down. After adjustment, the equivalent winding width of the coil is equal to the single-layer winding width reference value multiplied by the effective number of parallel layers after adjustment.

8. The method according to claim 6, characterized in that, instruction In this case, the operating frequency is calculated using the following formula: The frequency adjustment amount is equal to the frequency adjustment coefficient multiplied by (the difference between the fitting index and the equilibrium threshold, divided by the difference between the maximum allowable fitting index and the equilibrium threshold, and the resulting quotient) and then multiplied by the current operating frequency; the adjusted operating frequency is equal to the current operating frequency minus the frequency adjustment amount, and is limited to not being lower than the minimum operating frequency allowed by the system.

9. The method according to claim 4, characterized in that, The control commands are executed in stages according to the magnitude of the current density equalization fitting index: When the fit index is greater than or equal to the equilibrium threshold and less than the first-level control threshold, the instruction is executed. Output early warning signal; When the fit index is greater than or equal to the first-level control threshold and less than the second-level control threshold, the instruction is executed. Adjust the parallel via connection status; When the fit index is greater than or equal to the second-level control threshold, the instruction is executed simultaneously. ,instruction and instructions Joint regulation and control should be carried out. The equilibrium threshold is less than the first-level control threshold, the first-level control threshold is less than the second-level control threshold, and the second-level control threshold is not greater than the maximum allowable fit index.

Images