Switching frequency control method and LLC full-half-bridge resonant converter

Abstract

The invention discloses a switching frequency control method and an LLC full-bridge and half-bridge resonant converter. The method comprises the following steps: acquiring an input voltage and an output current of the converter in real time; based on a comparison result of the input voltage and a preset mode switching threshold value, the working mode of the converter is determined, and the working mode comprises a full-bridge mode and a half-bridge mode; calculating circuit parameters according to the input voltage, the output voltage set value, the output current and device parameters of the converter; determining an initial normalization frequency according to the value range of the target gain; based on the circuit parameters, the initial normalized frequency and a gain mathematical model of the converter, solving a target normalized frequency through an iterative algorithm to obtain a feed-forward switching frequency; and generating a driving signal of each switching tube according to the feed-forward switching frequency and the working mode. According to the invention, the optimal switching frequency can be calculated in real time according to the current working condition, the sudden change of the output voltage during full-bridge and half-bridge mode switching is effectively inhibited, and the dynamic response speed of the converter is improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of resonant converters, and in particular to a switching frequency control method and an LLC full-half-bridge resonant converter. Background Technology

[0002] LLC resonant converters are widely used in server power supplies, electric vehicle chargers, and communication power supplies due to their advantages such as soft-switching characteristics, high efficiency, and high power density. In practical applications, to meet the requirements of a wide input voltage range or a wide output voltage range, it is necessary to extend the voltage gain range of the LLC resonant converter.

[0003] A common technical solution is to extend the voltage gain range by switching between full-bridge and half-bridge topologies. Since there is approximately a 2:1 gain ratio between full-bridge and half-bridge topologies, switching the operating mode under different input voltage conditions can effectively expand the converter's gain adjustment range. When the input voltage is low, the converter operates in full-bridge mode to obtain higher voltage gain; when the input voltage is high, the converter switches to half-bridge mode to reduce voltage gain.

[0004] However, traditional LLC resonant converter control methods typically employ proportional-integral (PI) regulators based on output voltage feedback to adjust the switching frequency. This single voltage-loop feedback control method inherently exhibits response lag; at the instant of switching between full-bridge and half-bridge modes, the controller cannot anticipate and compensate for the gain jump caused by the topology change. Due to the significant gain difference between full-bridge and half-bridge modes, abrupt output voltage changes can occur during topology switching.

[0005] Sudden changes in output voltage can cause voltage surges to downstream loads or energy storage batteries. This can range from affecting the normal operation of the load to shortening the lifespan of the equipment or even damaging the components. Summary of the Invention

[0006] The main technical problem solved by the embodiments of the present invention is to provide a switching frequency control method and an LLC full-half-bridge resonant converter, which can solve at least some of the defects of the existing LLC full-half-bridge resonant converter.

[0007] In a first aspect, embodiments of the present invention provide a switching frequency control method applied to an LLC full-bridge / half-bridge resonant converter, comprising: real-time acquisition of the input voltage and output current of the converter; determining the operating mode of the converter based on a comparison result of the input voltage and a preset mode switching threshold; wherein the operating mode includes a full-bridge mode and a half-bridge mode; calculating circuit parameters based on the input voltage, the output voltage setpoint, the output current, and the device parameters of the converter; the circuit parameters including a target gain; determining an initial normalized frequency based on the range of the target gain; obtaining a feedforward switching frequency by iteratively calculating and solving the target normalized frequency based on the circuit parameters, the initial normalized frequency, and the gain mathematical model of the converter; and generating drive signals for each switch in the converter based on the feedforward switching frequency and the operating mode.

[0008] In a second aspect, embodiments of the present invention provide an LLC full-half-bridge resonant converter, comprising: a power conversion circuit including a switching bridge arm on the input side and a rectifier circuit on the output side, wherein the switching bridge arm includes a first switching transistor, a second switching transistor, a third switching transistor, and a fourth switching transistor; a resonant network including a resonant inductor, a resonant capacitor, and a magnetizing inductor; and a controller for executing the switching frequency control method as described in the first aspect.

[0009] The beneficial effects of the embodiments of the present invention are as follows: Unlike the prior art, the embodiments of the present invention calculate circuit parameters by real-time acquisition of input voltage, output current and device parameters of the converter, and obtain the feedforward switching frequency by numerical iterative algorithm based on the mathematical model of the converter gain, and superimpose the feedforward switching frequency with the output of the feedback controller to obtain the final switching frequency. It can calculate the optimal switching frequency in real time according to the current operating conditions, effectively suppress the sudden change of output voltage during the switching of full and half bridge modes, and improve the dynamic response speed of the converter. Attached Figure Description

[0010] One or more embodiments are illustrated by way of example with reference to the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements having the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0011] Figure 1 This is the circuit schematic of a traditional LLC full-half-bridge converter circuit;

[0012] Figure 2 This is a waveform diagram of the output voltage of a single voltage loop (PI loop) in a traditional LLC full-half-bridge converter circuit. Figure 3 This is a flowchart illustrating a switching frequency control method provided by an embodiment of the present invention; Figure 4 This is a control block diagram of a switching frequency control method provided by an embodiment of the present invention; Figure 5 This is a waveform diagram of the output voltage of the switching feedforward control loop of the LLC full-half-bridge converter circuit under the switching frequency control method provided by the embodiments of the present invention. Figure 6 This is a waveform diagram of the feedforward frequency and the actual switching frequency under the switching frequency control method provided by the embodiments of the present invention. Figure 7 This is a schematic diagram of the structure of an LLC full-half-bridge resonant converter provided by an embodiment of the present invention. Detailed Implementation

[0013] To facilitate understanding of this application, a more detailed description is provided below with reference to the accompanying drawings and specific embodiments. It should be noted that when an element is described as being "fixed to" another element, it can be directly on the other element, or one or more intermediate elements may exist between them. When an element is described as being "connected" to another element, it can be directly connected to the other element, or one or more intermediate elements may exist between them. The terms "upper," "lower," "inner," "outer," "bottom," etc., used in this specification indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application. Furthermore, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0014] Unless otherwise defined, all technical and scientific terms used in this specification have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. The term "and / or" as used in this specification includes any and all combinations of one or more of the associated listed items.

[0015] Furthermore, the technical features involved in the different embodiments of this application described below can be combined with each other as long as they do not conflict with each other.

[0016] The technical solutions in this application will be described below with reference to the accompanying drawings.

[0017] Please see Figure 1 , Figure 1This is a circuit schematic of a traditional LLC full-half-bridge converter circuit. In some embodiments of this application, the LLC full-half-bridge resonant converter includes a power conversion circuit and a resonant network. The power conversion circuit includes a switching bridge arm on the input side and a rectifier circuit on the output side. The switching bridge arm includes a first switch Q1, a second switch Q2, a third switch Q3, and a fourth switch Q4. The resonant network includes a resonant inductor Lr, a resonant capacitor Cr, and a magnetizing inductor Lm.

[0018] Specifically, the first switch Q1 and the second switch Q2 are connected in series to form the first bridge arm, and the third switch Q3 and the fourth switch Q4 are connected in series to form the second bridge arm. The first and second bridge arms are connected in parallel and then connected to the input voltage Vi. The resonant inductor Lr, the resonant capacitor Cr, and the magnetizing inductor Lm form a resonant network, which is connected between the midpoint of the first bridge arm and the midpoint of the second bridge arm. The primary winding of the transformer Tr is connected in parallel with the magnetizing inductor Lm, and the secondary winding of the transformer Tr is connected to the rectifier circuit. The output capacitor Co is connected in parallel to the output terminal of the rectifier circuit, and the load Ro is connected in parallel with the output capacitor Co.

[0019] In some embodiments of this application, the driving signal for the first switch Q1 is g1, the driving signal for the second switch Q2 is g2, the driving signal for the third switch Q3 is g3, and the driving signal for the fourth switch Q4 is g4.

[0020] Understandably, the LLC full-bridge / half-bridge resonant converter has two operating modes: full-bridge mode and half-bridge mode. When the converter operates in full-bridge mode, the first switch Q1 and the second switch Q2 are complementary conductions, as are the third switch Q3 and the fourth switch Q4. Furthermore, the drive signals for the first switch Q1 and the third switch Q3 are in phase, and the drive signals for the second switch Q2 and the fourth switch Q4 are in phase. When the converter operates in half-bridge mode, the first switch Q1 and the second switch Q2 are complementary conductions, the third switch Q3 remains off, and the fourth switch Q4 remains on.

[0021] In some embodiments of this application, there is an approximately two-fold gain ratio between full-bridge mode and half-bridge mode. When the input voltage Vi is low, the converter operates in full-bridge mode to obtain a higher voltage gain; when the input voltage Vi is high, the converter switches to half-bridge mode to reduce the voltage gain. By switching between full-bridge mode and half-bridge mode, the voltage gain range of the converter can be expanded.

[0022] Specifically, the converter regulates the output voltage by modulating the switching frequency fs. In both operating modes, the duty cycle of the drive signal for each switch is 50%.

[0023] It is understandable that the duty cycle mentioned above is a theoretical value ignoring dead time. In practical applications, to prevent short circuits caused by simultaneous conduction of the upper and lower switches on the same bridge arm, a dead time needs to be set between the two complementary switches. During the dead time, both the upper and lower switches on the same bridge arm are in the off state.

[0024] Please see Figure 2 , Figure 2 This is a waveform diagram of the output voltage of the single-voltage loop (PI loop) of a traditional LLC full-half-bridge converter circuit.

[0025] In some embodiments of this application, Figure 2 The upper part displays the waveform of the input voltage Vi. Figure 2 The lower half of the display shows the waveform of the corresponding output voltage Vo.

[0026] Understandably, traditional LLC resonant converter control methods use proportional-integral regulators based on output voltage feedback to adjust the switching frequency. During the switching between full-bridge and half-bridge modes, a single voltage loop feedback control cannot anticipate and compensate for the gain jump caused by the topology switch.

[0027] As an example rather than a limitation, such as Figure 2 As shown, when the input voltage Vi changes and triggers the full-half-bridge mode switching, the output voltage Vo exhibits a significant abrupt change. The magnitude of this abrupt change in output voltage Vo is substantial, deviating from the expected stable output voltage value.

[0028] In some embodiments of this application, sudden changes in output voltage can cause voltage surges to downstream loads or energy storage batteries, affecting the normal operation of the load, shortening the service life of the equipment, or even causing device damage.

[0029] To address the aforementioned issues, this application provides a switching frequency control method applied to an LLC full-half-bridge resonant converter. This method can pre-compensate for gain step changes during full-half-bridge mode switching, effectively suppressing output voltage surges. A flowchart illustrating this method is shown below. Figure 3 As shown, the specific steps include the following: Step S100: Real-time acquisition of the converter's input voltage and output current.

[0030] In some embodiments of this application, the controller acquires the input voltage Vi and output current Io of the LLC full-half-bridge resonant converter in real time.

[0031] Specifically, the input voltage Vi is the DC bus voltage on the input side of the converter, and the output current Io is the current flowing from the output side of the converter to the load. In some embodiments of this application, the input voltage Vi is used to determine the operating mode of the converter and calculate the target gain, and the output current Io is used to calculate the current load conditions and circuit parameters.

[0032] Step S200: Determine the operating mode of the converter based on the comparison result between the input voltage and the preset mode switching threshold.

[0033] In some embodiments of this application, the controller compares the acquired input voltage Vi with a preset mode switching threshold and determines the operating mode of the converter based on the comparison result.

[0034] Specifically, the operating modes include full-bridge mode and half-bridge mode.

[0035] Understandably, full-bridge mode is suitable for operating conditions with lower input voltages and can provide higher voltage gain; half-bridge mode is suitable for operating conditions with higher input voltages and can provide lower voltage gain. By switching modes based on the input voltage, the voltage gain range of the converter can be expanded.

[0036] Step S300: Calculate the circuit parameters based on the input voltage, output voltage setting, output current, and converter device parameters.

[0037] In some embodiments of this application, the controller calculates circuit parameters based on the input voltage Vi, output voltage setpoint Vo_set, output current Io, and the converter's own device parameters collected in step S100.

[0038] Specifically, the device parameters include the transformer turns ratio N, resonant inductance Lr, resonant capacitance Cr, and magnetizing inductance Lm. These parameters are inherent parameters of the converter, determined during the converter design phase and stored in the controller.

[0039] Specifically, circuit parameters are used to characterize the operating state of the converter under its current conditions and serve as input to the numerical iterative algorithm. In some embodiments of this application, the circuit parameters include inductance ratio, target gain, and quality factor. The target gain is determined based on the output voltage, input voltage setpoint, and transformer turns ratio.

[0040] Specifically, when the operating mode is full-bridge mode, the formula for calculating the target gain Mtarget is:

[0041] When the operating mode is half-bridge mode, the formula for calculating the target gain Mtarget is:

[0042] Understandably, the target gain characterizes the voltage conversion ratio that the converter needs to achieve under current operating conditions. The value of the target gain is determined by both the desired output voltage (i.e., the output voltage setpoint) and the actual input voltage. The target gain calculation formulas for full-bridge mode and half-bridge mode differ because the equivalent voltage applied to the resonant network differs between the two modes. In full-bridge mode, the two bridge arms work alternately, and the voltage amplitude applied to the resonant network is equal to the input voltage Vi; in half-bridge mode, only one bridge arm participates in the switching action, and the voltage amplitude applied to the resonant network is equal to half the input voltage Vi / 2. Therefore, to obtain the same output voltage, the gain required for half-bridge mode is twice that of full-bridge mode.

[0043] In some embodiments of this application, the quality factor is determined based on the resonant inductance, resonant capacitance, and equivalent AC resistance.

[0044] Specifically, the formula for calculating the quality factor Q is:

[0045] Where Lr is the resonant inductance, Cr is the resonant capacitance, and Rac is the equivalent AC resistance.

[0046] In some embodiments of this application, the equivalent AC resistance is determined based on the output voltage and output current.

[0047] Specifically, the formula for calculating the equivalent AC resistance Rac is:

[0048] Where Ro is the DC load resistance and N is the transformer turns ratio. The DC load resistance Ro can be calculated from the output voltage Vo and the output current Io:

[0049] Understandably, the quality factor reflects the quality characteristics of a resonant network and is related to the values ​​of the resonant inductor and capacitor, as well as the load conditions. When the load changes, the equivalent AC resistance changes accordingly, and the quality factor changes accordingly as well.

[0050] In some embodiments of this application, the inductance ratio is the ratio of the magnetizing inductance to the resonant inductance.

[0051] Specifically, the formula for calculating the inductance ratio k is:

[0052] Where Lm is the magnetizing inductance and Lr is the resonant inductance.

[0053] It is understandable that the inductance ratio is an inherent parameter of the converter, determined by the design of the resonant network, and remains unchanged during the operation of the converter.

[0054] Step S400: Determine the initial normalized frequency based on the range of the target gain.

[0055] In some embodiments of this application, the gain mathematical model of the converter is expressed as a functional relationship between normalized frequency, inductance ratio and quality factor.

[0056] Specifically, the normalized frequency fn is the ratio of the switching frequency to the resonant frequency, and its calculation formula is as follows:

[0057] Where fs is the switching frequency and fr is the resonant frequency.

[0058] In some embodiments of this application, the voltage gain M of the LLC resonant converter can be expressed as a function of the normalized frequency fn, the inductance ratio k, and the quality factor Q.

[0059] Specifically, the formula for calculating the voltage gain M is:

[0060] Understandably, the above gain mathematical model describes the relationship between the voltage gain and the normalized frequency of an LLC resonant converter. Given the target gain Mtarget, inductance ratio k, and quality factor Q, the corresponding target normalized frequency can be obtained by solving the above gain mathematical model.

[0061] Since the gain mathematical model is a complex nonlinear function of the normalized frequency fn, its analytical expression cannot be directly obtained using algebraic methods. Therefore, this invention employs a numerical iterative algorithm to solve for the target normalized frequency. The numerical iterative algorithm sets an initial guess value and then iteratively approximates the error function. The root equal to zero is used to obtain the corresponding target normalized frequency. Using the above numerical solution method, the correct solution can be quickly converged to be found under given accuracy requirements, overcoming the difficulty of solving the gain equation analytically.

[0062] In some embodiments of this application, the controller determines the initial normalized frequency based on the range of the target gain.

[0063] It is understandable that the initial normalized frequency is the initial guess value when the numerical iterative algorithm starts iterative calculation.

[0064] Specifically, different initial normalization frequencies are selected based on the different ranges of the target gain to ensure that the numerical iterative algorithm can quickly converge to the correct target normalization frequency.

[0065] In some embodiments of this application, when the target gain is greater than a first threshold, the converter operates in the deep capacitive region, and the initial normalized frequency is taken as a first preset value.

[0066] Understandably, when the target gain is high, the converter needs to operate in a lower normalized frequency range to achieve high-gain output. The first preset value corresponds to the typical normalized frequency under high-gain conditions.

[0067] In some embodiments of this application, when the target gain is greater than the second threshold and not greater than the first threshold, the converter operates in the capacitive region, and the initial normalized frequency is taken as the second preset value.

[0068] Specifically, the second preset value is greater than the first preset value, corresponding to the typical normalized frequency when the target gain is slightly lower than the first threshold.

[0069] In some embodiments of this application, when the target gain is greater than the third threshold and not greater than the second threshold, the converter operates in the resonant region and the initial normalized frequency is taken as the third preset value.

[0070] Specifically, the third preset value is greater than the second preset value, corresponding to the typical normalized frequency when the target gain is in the medium range.

[0071] In some embodiments of this application, when the target gain is greater than the fourth threshold and not greater than the third threshold, the converter operates in the inductive region, and the initial normalized frequency is taken as the fourth preset value.

[0072] Specifically, the fourth preset value is greater than the third preset value, corresponding to the typical normalized frequency when the target gain is low.

[0073] In some embodiments of this application, when the target gain is not greater than the fourth threshold, the converter operates in the deep sensing region, and the initial normalized frequency is taken as the fifth preset value.

[0074] Specifically, the fifth preset value is greater than the fourth preset value, corresponding to the typical normalized frequency under low-gain conditions.

[0075] It is understandable that the first threshold is greater than the second threshold, the second threshold is greater than the third threshold, and the third threshold is greater than the fourth threshold. The first preset value is less than the second preset value, the second preset value is less than the third preset value, the third preset value is less than the fourth preset value, and the fourth preset value is less than the fifth preset value.

[0076] In some embodiments of this application, there is a negative correlation between the target gain and the initial normalized frequency. The larger the target gain, the smaller the initial normalized frequency; the smaller the target gain, the larger the initial normalized frequency.

[0077] It is understandable that the above-mentioned selection rule for the initial normalized frequency is designed based on the gain curve characteristics of the LLC resonant converter.

[0078] Specifically, according to the gain mathematical model, when the target gain M is greater than 1, the normalized frequency fn needs to be less than 1 in order to achieve a gain greater than 1, and the converter operates in the capacitive region; when the target gain M is less than 1, the normalized frequency fn needs to be greater than 1 in order to achieve a gain less than 1, and the converter operates in the inductive region.

[0079] As an example rather than a limitation, the table below shows the correspondence between the target gain range, the operating region, and the initial normalized frequency preset value:

[0080] It is understood that the specific values ​​in the table above are only examples, and in actual applications they can be adjusted according to the converter's design parameters and gain curve characteristics.

[0081] In some embodiments of this application, by selecting an appropriate initial normalization frequency based on the range of the target gain, an initial guess value close to the true solution can be provided for the numerical iterative algorithm, thereby accelerating the iteration convergence speed and reducing the number of iterations.

[0082] Understandably, an inappropriate selection of the initial normalization frequency may lead to slow or even non-convergence of the numerical iterative algorithm. By dividing the target gain into multiple value ranges and setting a corresponding preset initial normalization frequency for each range, it is possible to ensure that the numerical iterative algorithm can quickly and accurately solve for the target normalization frequency under various operating conditions.

[0083] In some embodiments of this application, the initial normalized frequency can also be used as the default output frequency when the numerical iteration algorithm fails to run properly.

[0084] Understandably, if the normalized frequency fn is too low, the converter will operate in the deep capacitive region, leading to an exponential increase in resonant current and switching current stress, potentially damaging the device. Conversely, if the normalized frequency fn is too high, switching losses will increase linearly with frequency, resulting in a sharp drop in efficiency. Therefore, the initial normalized frequency selected based on the target gain not only provides an accurate initial guess for the iterative algorithm but also provides a physically reasonable switching frequency that ensures basic safety when the iterative algorithm cannot run.

[0085] Step S500: Based on the circuit parameters, the initial normalized frequency, and the mathematical model of the converter's gain, the target normalized frequency is solved through iterative calculation, thereby obtaining the feedforward switching frequency.

[0086] In some embodiments of this application, the controller uses a numerical iterative algorithm to solve for the target normalized frequency based on the calculated circuit parameters and the mathematical model of the converter's gain.

[0087] Understandably, the gain mathematical model describes the relationship between the voltage gain of an LLC resonant converter and the normalized frequency. The normalized frequency is the ratio of the switching frequency to the resonant frequency.

[0088] Specifically, the numerical iterative algorithm solves the target normalized frequency through iterative calculation, which can overcome the shortcomings of the traditional table lookup method, such as poor adaptability and large storage resource consumption.

[0089] In some embodiments of this application, the controller obtains the feedforward switching frequency based on the target normalized frequency.

[0090] It is understandable that the feedforward switching frequency is the optimal switching frequency calculated based on the current operating conditions, which enables the converter to output the desired voltage value under the current input voltage and load conditions.

[0091] Step S600: Generate drive signals for each switch in the converter based on the feedforward switching frequency and operating mode.

[0092] In some embodiments of this application, the controller generates drive signals for each switch in the converter based on the feedforward switching frequency obtained in step S500 and the operating mode determined in step S200.

[0093] Specifically, the feedforward switching frequency determines the period of the drive signal, and the operating mode determines the drive logic relationship between each switch.

[0094] Understandably, the driving methods for each switch differ between full-bridge and half-bridge modes. The controller generates corresponding drive signals based on the current operating mode to ensure the converter operates normally in that mode.

[0095] In some embodiments of this application, the drive signals include drive signal g1 for the first switch Q1, drive signal g2 for the second switch Q2, drive signal g3 for the third switch Q3, and drive signal g4 for the fourth switch Q4.

[0096] It is understood that, through steps S100 to S600 above, the switching frequency control method provided in this application can calculate the feedforward switching frequency based on the real-time acquired voltage and current signals using a numerical iterative algorithm, and generate the final switching frequency by combining it with the output of the feedback controller. During full-half-bridge mode switching, the feedforward switching frequency can pre-compensate for the gain step caused by topology changes, thereby effectively suppressing sudden changes in output voltage and improving the dynamic response speed of the converter.

[0097] In some embodiments of this application, step S200 specifically includes the following steps: The input voltage is compared with the full-bridge mode switching threshold and the half-bridge mode switching threshold respectively; the operating mode of the converter is determined based on the comparison result or the current operating mode is maintained.

[0098] In some embodiments of this application, the mode switching threshold includes a full-bridge mode switching threshold and a half-bridge mode switching threshold.

[0099] Specifically, the full-bridge mode switching threshold is the input voltage threshold that triggers the converter to enter full-bridge mode. The half-bridge mode switching threshold is the input voltage threshold that triggers the converter to enter half-bridge mode.

[0100] It is understandable that the half-bridge mode switching threshold is greater than the full-bridge mode switching threshold, and a hysteresis interval is formed between the full-bridge mode switching threshold and the half-bridge mode switching threshold.

[0101] In some embodiments of this application, the controller compares the acquired input voltage with the full-bridge mode switching threshold and the half-bridge mode switching threshold respectively, and executes the corresponding mode decision logic based on the comparison result.

[0102] Specifically, when the input voltage is below the full-bridge mode switching threshold, the operating mode is determined to be full-bridge mode. It is understandable that when the input voltage is low, the converter needs to operate in high-gain mode to maintain the desired output voltage. Full-bridge mode provides higher voltage gain and is suitable for low input voltage conditions.

[0103] Specifically, when the input voltage exceeds the half-bridge mode switching threshold, the operating mode is determined to be half-bridge mode. It is understandable that when the input voltage is high, the converter needs to operate in low-gain mode to avoid excessively high output voltage. Half-bridge mode provides lower voltage gain and is suitable for high input voltage conditions.

[0104] Specifically, when the input voltage is between the full-bridge mode switching threshold and the half-bridge mode switching threshold, the current operating mode is maintained. This means that when the input voltage is within the hysteresis range, the converter does not switch modes but maintains the current operating mode. If the current operating mode is full-bridge mode, it continues to maintain full-bridge mode; if the current operating mode is half-bridge mode, it continues to maintain half-bridge mode.

[0105] In some embodiments of this application, by setting the hysteresis range between the full-bridge mode switching threshold and the half-bridge mode switching threshold, the problem of frequent mode switching caused by fluctuations in the input voltage near the critical point can be effectively prevented.

[0106] Understandably, if only a single threshold is used for mode switching, the converter will frequently switch between full-bridge and half-bridge modes when the input voltage fluctuates slightly around the threshold, leading to system instability. By introducing a hysteresis range, mode switching is only triggered when the input voltage clearly exceeds the boundary of the hysteresis range, thus ensuring stable system operation.

[0107] In some embodiments of this application, the logic flow of pattern determination is as follows: First, the controller samples the input voltage.

[0108] Then, the controller determines whether the input voltage is lower than the full-bridge mode switching threshold. If the input voltage is lower than the full-bridge mode switching threshold, the operating mode is determined to be full-bridge mode.

[0109] If the input voltage is not lower than the full-bridge mode switching threshold, the controller further determines whether the input voltage is higher than the half-bridge mode switching threshold. If the input voltage is higher than the half-bridge mode switching threshold, the operating mode is determined to be half-bridge mode.

[0110] If the input voltage is neither lower than the full-bridge mode switching threshold nor higher than the half-bridge mode switching threshold, the current operating mode remains unchanged.

[0111] Understandably, the above mode decision logic is repeatedly executed in each control cycle to track changes in the input voltage in real time and adjust the operating mode accordingly.

[0112] As an example, and not a limitation, the full-bridge mode switching threshold V_FBth can be set to 300V, and the half-bridge mode switching threshold V_HBth can be set to 310V. Under these settings, when the input voltage Vi is below 300V, the converter operates in full-bridge mode; when the input voltage Vi is above 310V, the converter operates in half-bridge mode; and when the input voltage Vi is between 300V and 310V, the converter maintains its current operating mode.

[0113] Understandably, the difference between the full-bridge mode switching threshold and the half-bridge mode switching threshold is 10V, forming a 10V hysteresis range. The size of the hysteresis range can be adjusted according to the input voltage fluctuation range and system stability requirements.

[0114] Understandably, by employing a hysteresis comparison mode switching strategy, this application can smoothly switch between full-bridge and half-bridge modes when the input voltage changes, avoiding system instability caused by frequent mode switching and providing reliable mode decision results for subsequent feedforward frequency calculation and drive signal generation.

[0115] In some embodiments of this application, step S500 involves iteratively calculating the target normalized frequency based on circuit parameters, the initial normalized frequency, and the converter's gain mathematical model. This specifically includes the following steps: The initial normalized frequency is used as the current normalized frequency; the current gain is calculated based on the current normalized frequency and the gain mathematical model; the error function value between the current gain and the target gain is calculated based on the error function; it is determined whether the error function value meets the convergence condition; if it does, the target normalized frequency is output; if it does not, the current normalized frequency is updated according to the iterative formula and the iteration continues.

[0116] In some embodiments of this application, the controller first verifies the input parameters before starting iterative calculation.

[0117] Specifically, the controller verifies the rationality of input parameters such as inductance ratio k, target gain Mtarget, and quality factor Q.

[0118] In some embodiments of this application, when the input parameters exceed a reasonable range, the controller does not perform iterative calculations but directly uses the initial normalized frequency fn_guess selected based on the target gain as the target normalized frequency output. Once the input parameters have been verified, the controller performs initialization operations.

[0119] Understandably, parameter verification can avoid iterative calculation errors caused by abnormal parameters, ensuring that the system can still output a safe and reasonable switching frequency under abnormal operating conditions.

[0120] In some embodiments of this application, the controller uses the initial normalized frequency as the current normalized frequency.

[0121] Specifically, the initial normalized frequency fn_guess is determined based on the range of the target gain, and the controller assigns the initial normalized frequency to the current normalized frequency fn_current, that is:

[0122] In some embodiments of this application, the controller initializes the iteration count counter iterations_used to 0, initializes the best normalization frequency best_fn to the initial normalization frequency fn_guess, and initializes the minimum error function value best_error to infinity.

[0123] Specifically, the best normalization frequency best_fn is used to record the normalization frequency corresponding to the minimum error function value during the iteration process, so as to output the optimal result when the iteration fails to converge normally.

[0124] In some embodiments of this application, the controller calculates the current gain based on the current normalized frequency and the gain mathematical model.

[0125] Specifically, the current normalized frequency fn_current is substituted into the gain mathematical model to calculate the current gain M_calc. The calculation formula for the gain mathematical model is:

[0126] Where k is the inductance ratio and Q is the quality factor.

[0127] In some embodiments of this application, the controller calculates the error function value between the current gain and the target gain based on the error function.

[0128] Specifically, the error function F(fn) is defined as the difference between the current gain and the target gain, and its calculation formula is as follows:

[0129] in, Mtarget is the target gain calculated from the mathematical model of gain, and fn is the normalized frequency.

[0130] In some embodiments of this application, the controller calculates the error function value corresponding to the current normalized frequency fn_current, denoted as error:

[0131] Where M_calc is the current gain calculated by substituting the current normalized frequency fn_current into the gain mathematical model.

[0132] It can be understood that `error` is the error function value of the error function F(fn) at the current normalized frequency. The error function F(fn) describes the functional relationship between gain deviation and normalized frequency, while the error function value `error` is the specific value of this function at a particular frequency point. When the error function value is positive, it indicates that the current gain is greater than the target gain; when the error function value is negative, it indicates that the current gain is less than the target gain; when the error function value is zero, it indicates that the current gain is equal to the target gain, and the corresponding normalized frequency is the target normalized frequency.

[0133] In some embodiments of this application, the controller updates the best result after each iteration of calculating the error function value.

[0134] Specifically, if the absolute value of the current error function is less than the minimum error function value best_error, then the current normalized frequency fn_current is updated to the optimal normalized frequency best_fn, and the absolute value of the current error function is updated to the minimum error function value best_error. like ,but:

[0135]

[0136] It is understandable that by recording the normalization frequency corresponding to the minimum error function value in each iteration, it is possible to ensure that a relatively optimal result can still be output even when the iteration fails to converge normally or reaches the maximum number of iterations.

[0137] In some embodiments of this application, the controller determines whether the absolute value of the error function is less than a preset tolerance.

[0138] Specifically, when the absolute value of the error function is less than the preset tolerance tol, the iteration is considered to have converged, and the current normalized frequency fn_current is output as the target normalized frequency.

[0139] As an example rather than a limitation, the preset tolerance tol can be set to 0.001 or other suitable values, and the specific value can be adjusted according to the control accuracy requirements.

[0140] In some embodiments of this application, when the absolute value of the error function is less than a preset tolerance, the iteration is successful, the controller outputs the current normalized frequency fn_current as the target normalized frequency fn, and ends the iteration process. In some embodiments of this application, when the absolute value of the error function is not less than the preset tolerance, the controller calculates the derivative of the error function to update the current normalized frequency.

[0141] Specifically, the derivative F'(fn) of the error function represents the rate of change of the error function F(fn) with respect to the normalized frequency fn, reflecting the sensitivity of the gain deviation to frequency changes. According to the definition of the error function, its derivative is mathematically expressed as:

[0142] Due to the gain function The analytical derivative expression is quite complex, and directly finding its analytical derivative is tedious. Based on the gain mathematical model:

[0143] Finding the derivative of fn requires applying the chain rule, and its analytical derivative expression involves multiple nested operations, resulting in high computational complexity. Therefore, this invention employs the central difference method to calculate the derivative of the error function, simplifying the calculation process.

[0144] Specifically, the central difference method approximates the derivative by calculating the difference between the function values ​​of the error function before and after the current frequency point. The calculation formula is as follows:

[0145] Where h is the difference step size, Let fn be the function value shifted forward by h. Let fn be the function value shifted backward by h.

[0146] Substituting the definition of the error function into the above equation, we get:

[0147] Understandably, the central difference method uses the function values ​​on both sides of a point to estimate the derivative at that point, resulting in higher accuracy compared to the forward or backward difference methods. The selection of the difference step size h requires a trade-off between accuracy and numerical stability; a step size that is too large will increase the approximation error, while a step size that is too small may introduce numerical computational errors.

[0148] As an example rather than a limitation, the difference step size h can be set to 0.001 or other suitable values, and the specific value can be adjusted according to the requirements of calculation accuracy and numerical stability.

[0149] In some embodiments of this application, the controller calculates the derivative value deriv corresponding to the current normalized frequency fn_current:

[0150] It is understandable that the derivative value *deriv* is the derivative of the error function F(fn) at the current normalized frequency *fn_current*. This derivative value reflects the trend of the error function at the current frequency: a positive derivative value indicates that the error function increases with increasing frequency, while a negative derivative value indicates that the error function decreases with increasing frequency; the larger the absolute value of the derivative, the more sensitive the error function is to frequency changes, and the smaller the absolute value of the derivative, the less sensitive the error function is to frequency changes.

[0151] In some embodiments of this application, the controller performs exception handling before updating the normalized frequency.

[0152] Specifically, when the absolute value of the derivative of the error function is less than a preset threshold, the iteration is terminated, and the optimal normalization frequency is taken as the target normalization frequency.

[0153] Understandably, when the derivative value is close to zero, continuing the iteration may lead to errors such as division by zero or numerical instability. By detecting whether the derivative value is close to zero, numerical calculation anomalies can be avoided.

[0154] In some embodiments of this application, when the number of iterations reaches a preset maximum number of iterations, the normalization frequency corresponding to the minimum error function value during the iteration process is taken as the target normalization frequency.

[0155] As an example rather than a limitation, the preset maximum number of iterations max_iter can be set to 10, 20 or other suitable values. The specific value can be adjusted according to the convergence speed and computing resource requirements.

[0156] Understandably, by setting a maximum number of iterations, it is possible to prevent the iteration process from getting stuck in an infinite loop and ensure that the controller outputs results within a limited time.

[0157] In some embodiments of this application, when the absolute value of the error function is not less than the preset tolerance and no exception handling is triggered, the controller updates the current normalized frequency according to the iterative formula.

[0158] Specifically, the iterative formula is:

[0159] in, Let be the normalized frequency of the i-th iteration. Let be the normalized frequency of the (i+1)th iteration. Let be the error function value of the i-th iteration. Let be the derivative value of the error function in the i-th iteration; In some embodiments of this application, the controller updates the current normalized frequency based on the ratio of the error function value to the derivative of the error function, obtaining the updated normalized frequency fn_new:

[0160] It is understandable that the above iterative formula approximates the root of the equation based on the tangent line of the function at the current point, and through continuous iteration, the normalized frequency gradually approaches the target value.

[0161] In some embodiments of this application, the controller performs boundary checks on the updated normalized frequency.

[0162] Specifically, when the updated normalized frequency fn_new exceeds the preset frequency range, a binary search method is used to adjust the updated normalized frequency to the preset frequency range.

[0163] In some embodiments of this application, the preset frequency range is [fn_min, fn_max].

[0164] As an example, not a limitation, the preset frequency range can be set to [0.5, 1.7]. When fn_new is less than 0.5, a binary search method is used to adjust fn_new towards the center of the interval; when fn_new is greater than 1.7, the same binary search method is used to adjust fn_new towards the center of the interval.

[0165] In some embodiments of this application, the preset frequency range is [fn_min, fn_max].

[0166] As an example, not a limitation, the preset frequency range can be set to [0.5, 1.7]. When the updated normalized frequency fn_new exceeds this range, a binary search method is used to adjust it back into the range.

[0167] Specifically, when fn_new is less than 0.5, it indicates that the frequency of the iterative calculation is too low. In this case, the bisection method is used to adjust fn_new to the midpoint between the current frequency fn_current and the lower boundary 0.5, i.e., fn_new = (fn_current + 0.5) / 2. For example, if fn_current is 0.6 and the iterative calculation yields fn_new of 0.3, then after adjustment, fn_new = (0.6 + 0.5) / 2 = 0.55.

[0168] When fn_new is greater than 1.7, it indicates that the frequency of iterative calculation is too high. In this case, the bisection method is used to adjust fn_new to the midpoint between the current frequency fn_current and the upper boundary 1.7, i.e., fn_new = (fn_current + 1.7) / 2. For example, if fn_current is 1.5 and the iterative calculation yields fn_new of 2.0, then after adjustment, fn_new = (1.5 + 1.7) / 2 = 1.6.

[0169] Understandably, using a binary search approach to adjust the frequency, rather than directly setting fn_new as the boundary value, allows the adjusted frequency to be closer to the true solution, which is beneficial for faster convergence in subsequent iterations. At the same time, this gradual adjustment method avoids drastic frequency jumps, improving the stability of the iterative algorithm.

[0170] Understandably, if the normalized frequency is too low, the converter will operate in the deep capacitive region, leading to an exponential increase in resonant current and switching transistor current stress, potentially damaging the devices. Conversely, if the normalized frequency is too high, switching losses will increase linearly with frequency, resulting in a sharp drop in efficiency. Boundary checks ensure that the normalized frequency remains within a safe and reasonable range.

[0171] In some embodiments of this application, the controller updates the normalized frequency after boundary check to the current normalized frequency fn_current, increments the iteration count counter by 1, and returns to the step of calculating the current gain to continue the next iteration.

[0172] Specifically, the iterative process continues until it terminates when any of the following conditions are met: (1) The absolute value of the error function is less than the preset tolerance, and the iteration converges successfully; (2) The number of iterations reaches the preset maximum number of iterations; (3) The absolute value of the derivative of the error function is less than the preset threshold.

[0173] In some embodiments of this application, the complete flow of the numerical iteration algorithm is as follows: Step 1: Input parameters include inductance ratio k, target gain Mtarget, quality factor Q, resonant frequency fr, preset tolerance tol, and preset maximum number of iterations max_iter.

[0174] Step 2: Determine the initial normalization frequency fn_guess based on the range of the target gain Mtarget.

[0175] Step 3: Verify the rationality of the input parameters. If the parameters are outside the reasonable range, do not iterate and directly use the initial normalized frequency fn_guess as the output.

[0176] Step 4: Initialize the current normalized frequency fn_current to fn_guess, and initialize the iteration count counter iterations_used to 0.

[0177] Step 5: Start the iterative loop.

[0178] Step 6: Calculate the current gain M_calc at the current normalized frequency fn_current.

[0179] Step 7: Calculate the error .

[0180] Step 8: Determine if convergence has occurred. If the iteration is successful, the current normalized frequency fn_current will be output.

[0181] Step 9: Calculate the derivative value deriv of the error function using the central difference method.

[0182] Step 10: Check if the derivative value is close to zero. If so, terminate the iteration and return the optimal result.

[0183] Step 11: Update the normalized frequency according to the iterative formula .

[0184] Step 12: Perform boundary checks. If fn_new exceeds the preset frequency range, adjust it using a binary search method.

[0185] Step 13: Update fn_new to fn_current and increment the iteration count by 1.

[0186] Step Fourteen: Determine if the maximum number of iterations has been reached. If not, return to Step Six to continue iterating. If the maximum number of iterations has been reached, output the normalized frequency corresponding to the minimum error function value.

[0187] Understandably, the numerical iterative algorithm described above can automatically and accurately solve for the corresponding target normalized frequency based on the circuit parameters sampled and calculated in real time. This overcomes the shortcomings of traditional lookup table methods, such as poor adaptability, large storage resource consumption, and difficulty in covering all operating points, and achieves precise frequency control with parameter adaptation.

[0188] In some embodiments of this application, the specific implementation of obtaining the feedforward switching frequency in step S500 is as follows: Specifically, the controller uses the product of the target normalized frequency and the resonant frequency as the feedforward switching frequency.

[0189] In some embodiments of this application, the formula for calculating the feedforward switching frequency fs_ff is as follows:

[0190] Where fs_ff is the feedforward switching frequency, fn is the target normalized frequency, and fr is the resonant frequency.

[0191] It is understandable that the target normalized frequency fn is obtained through the numerical iterative algorithm described in the above implementation method, and the resonant frequency fr is an inherent parameter of the converter resonant network. Multiplying the target normalized frequency by the resonant frequency yields the optimal feedforward switching frequency under the current operating condition.

[0192] In some embodiments of this application, the resonant frequency fr is determined by the resonant inductance Lr and the resonant capacitance Cr, and the calculation formula is as follows:

[0193] It is understandable that the feedforward switching frequency fs_ff is the optimal switching frequency calculated based on the current operating conditions, which enables the converter to output the desired voltage value under the current input voltage and load conditions.

[0194] In some embodiments of this application, in order to further improve the steady-state accuracy of the output voltage, feedback control can be introduced on the basis of feedforward control to form a composite control structure of feedforward plus feedback.

[0195] Please see Figure 4 , Figure 4 This is a control block diagram of a switching frequency control method provided by an embodiment of the present invention. Specifically, the feedforward control path includes a feedforward frequency calculation module. The feedforward frequency calculation module dynamically calculates the circuit parameters under the current operating condition by sampling the input voltage Vi and output current Io in real time, combined with the preset output voltage setpoint Vo_set and the device parameters of the converter, including the inductance ratio k, target gain Mtarget, quality factor Q, and resonant frequency fr, and then calculates the feedforward switching frequency fs_ff in real time through a numerical iterative algorithm.

[0196] Specifically, the feedback control path includes a feedback controller. The feedback controller acquires the converter's output voltage Vo in real time and adjusts it based on the deviation between the output voltage setpoint Vo_set and the actual output voltage Vo, outputting a frequency compensation amount fs_pi. In some embodiments of this application, the feedback controller is a PI controller, which adjusts based on the deviation between the output voltage setpoint Vo_set and the output voltage Vo, outputting a frequency compensation amount fs_pi.

[0197] As is understandable, a PI controller is a proportional-integral controller, consisting of a proportional element and an integral element. The proportional element generates a proportional control action based on the magnitude of the deviation, while the integral element generates an integral control action based on the accumulation of the deviation. Together, they output a frequency compensation value.

[0198] In some embodiments of this application, the feedforward switching frequency is superimposed with the output of the feedback controller, including: adding the feedforward switching frequency fs_ff to the frequency compensation amount fs_pi to obtain the final switching frequency fs.

[0199] Specifically, the formula for calculating the final switching frequency fs is:

[0200] Where fs is the final switching frequency, fs_ff is the feedforward switching frequency, and fs_pi is the frequency compensation amount output by the PI controller.

[0201] It is understandable that the above superposition formula reflects the master-slave relationship between feedforward control and feedback control. The feedforward switching frequency fs_ff, as the basic control variable, undertakes the main task of frequency regulation and determines the main part of the switching frequency; the frequency compensation variable fs_pi, as the correction variable, is used to eliminate the influence of factors such as feedforward model error, parameter deviation, and parasitic parameters, and makes fine adjustments based on the feedforward frequency.

[0202] It should be noted that there is a clear timing relationship between feedforward control and feedback control in the composite control structure, with feedback control lagging behind feedforward control by one control cycle.

[0203] Specifically, in the k-th control cycle, the feedforward control module calculates the feedforward switching frequency fs_ff(k) based on the input voltage Vi(k) and output current Io(k) acquired at the current moment, combined with the output voltage setpoint Vo_set, using a gain mathematical model. Feedforward control is calculated based on the input conditions at the k-th moment, enabling it to respond immediately to changes in the input at the current moment.

[0204] Simultaneously, the feedback controller acquires the actual output voltage Vo(k) at the current moment. It should be noted that the actual output voltage Vo(k) at the current moment is the output result after the final switching frequency fs(k-1) is applied to the converter at the previous moment; that is, Vo(k) reflects the effect of the control action at moment k-1. The feedback controller calculates the frequency compensation amount fs_pi(k) based on the deviation between the output voltage setpoint Vo_set and the actual output voltage Vo(k). Therefore, feedback control adjusts based on the result of the control action at moment k-1, lagging behind feedforward control by one control cycle.

[0205] In some embodiments of this application, the final switching frequency fs(k) at time k is obtained by superimposing the feedforward switching frequency fs_ff(k) and the frequency compensation amount fs_pi(k):

[0206] Where fs_ff(k) is the feedforward calculation result based on the input conditions at time k, and fs_pi(k) is the feedback correction amount based on the control effect at time k-1.

[0207] When the input voltage undergoes a step change or the full-bridge / half-bridge mode switches, feedforward control can immediately calculate the corresponding switching frequency based on the new input conditions in the current control cycle, without waiting for the output voltage to deviate. Feedback control, however, lags behind by one control cycle, and its adjustment is a correction of the control effect from the previous moment. Therefore, in dynamic processes, feedforward control undertakes the primary task of rapid response, while feedback control gradually eliminates residual deviations in subsequent cycles.

[0208] In some embodiments of this application, under steady-state operating conditions, the input voltage and load conditions remain stable, and the feedforward switching frequency fs_ff is maintained at a fixed value. After several control cycles, if the feedforward model is accurate enough, the deviation between the actual output voltage Vo and the set value Vo_set converges to a very small range, and the frequency compensation amount fs_pi approaches zero or is maintained at a very small steady-state value, which is used to compensate for the inherent error of the feedforward model.

[0209] In some embodiments of this application, during the converter startup phase or dynamic processes such as sudden input voltage changes, feedforward control can quickly calculate the required switching frequency based on the gain model and plays a dominant role. Since the output voltage may not yet be stable or is changing rapidly at this time, the output of the PI controller may be limited, resulting in a small output, or it may be in the process of adjustment. As the output voltage gradually stabilizes, if there is a voltage deviation caused by feedforward calculation errors, feedback control gradually intervenes and outputs a compensation amount to correct it, so that the output voltage is accurately stabilized at the set value.

[0210] In some embodiments of this application, during the steady-state operation of the converter, the input voltage and load conditions remain stable, and the feedforward switching frequency fs_ff is maintained at a fixed value. At this time, if the feedforward model is sufficiently accurate, the deviation between the actual output voltage and the set output voltage is very small, and the frequency compensation amount fs_pi output by the PI controller approaches zero. The system's operating point is mainly determined by feedforward control. If model errors or external disturbances cause a deviation between the output voltage and the set value, the PI controller will output a corresponding frequency compensation amount to correct this deviation, causing the output voltage to return to the set value.

[0211] In some embodiments of this application, the frequency compensation amount fs_pi is a fine-tuning compensation amount that can be positive or negative, and its sign and magnitude are determined by the actual deviation of the output voltage.

[0212] Specifically, when the feedforward switching frequency fs_ff is lower than the actual required frequency, the output voltage Vo will be lower than the output voltage setpoint Vo_set. At this time, the PI controller detects the negative voltage deviation and outputs a positive frequency compensation amount fs_pi to positively compensate the feedforward switching frequency, thereby increasing the final switching frequency fs and thus improving the output voltage.

[0213] When the feedforward switching frequency fs_ff is higher than the actual required frequency, the output voltage Vo will be higher than the output voltage set value Vo_set. At this time, the PI controller detects a positive voltage deviation and outputs a negative frequency compensation amount fs_pi to perform negative compensation on the feedforward switching frequency, thereby reducing the final switching frequency fs and thus reducing the output voltage.

[0214] Understandably, the positive and negative nature of the frequency compensation amount fs_pi allows feedback control to bidirectionally correct the error function value calculated by the feedforward. Regardless of whether the feedforward frequency is too high or too low, the PI controller can eliminate the deviation in the output voltage by outputting a compensation amount of the corresponding sign, ensuring that the output voltage remains stable near the set value.

[0215] In some embodiments of this application, the amplitude of the frequency compensation amount fs_pi is typically much smaller than the amplitude of the feedforward switching frequency fs_ff. This is because the feedforward calculation is based on the mathematical model of the converter's gain, which can accurately estimate the switching frequency required for the current operating condition. The frequency compensation amount only needs to correct minor deviations caused by factors such as model errors, parameter deviations, and parasitic parameters. Therefore, the frequency compensation amount fs_pi plays a fine-tuning compensation role in the control system, rather than a dominant control role.

[0216] As an example, and not a limitation, assuming the calculated feedforward switching frequency fs_ff is 100kHz, while the actual required frequency is 101kHz, the frequency compensation amount fs_pi output by the PI controller will be approximately +1kHz, and the final switching frequency fs = 100kHz + 1kHz = 101kHz. Conversely, if the actual required frequency is 99kHz, the frequency compensation amount fs_pi will be approximately -1kHz, and the final switching frequency fs = 100kHz - 1kHz = 99kHz.

[0217] In some embodiments of this application, by introducing a feedforward frequency calculation module, the system can calculate the optimal switching frequency that matches the current operating condition in advance when the input voltage changes. At the instant when the full-half-bridge topology switching occurs, the feedforward switching frequency can compensate for the gain step caused by the topology change in advance, thereby fundamentally avoiding the voltage surge problem that traditional pure feedback control cannot suppress due to response lag.

[0218] In some embodiments of this application, the PI controller operates in parallel, and its output fs_pi is used as a fine-tuning compensation amount to correct small deviations in the feedforward model, thereby further improving the steady-state accuracy and disturbance rejection capability of the output voltage.

[0219] Understandably, the composite control framework that combines feedforward control and feedback control can not only achieve rapid adjustment of the output voltage, but also effectively suppress the sudden changes in output voltage during the switching process of the full-half-bridge, thus balancing dynamic response speed and steady-state control accuracy.

[0220] In some embodiments of this application, step S600 specifically includes the following steps: In some embodiments of this application, the controller determines the period of the drive signal based on the final switching frequency.

[0221] Specifically, the relationship between the period Ts of the driving signal and the final switching frequency fs is as follows:

[0222] Where Ts is the period of the drive signal and fs is the final switching frequency.

[0223] It is understandable that the higher the final switching frequency fs, the shorter the period Ts of the drive signal; the lower the final switching frequency fs, the longer the period Ts of the drive signal.

[0224] In some embodiments of this application, when the operating mode is full-bridge mode, the controller generates drive signals for each switch according to the drive logic of full-bridge mode.

[0225] Specifically, when the operating mode is full-bridge mode, the first switch Q1 and the second switch Q2 are complementaryly turned on, the third switch Q3 and the fourth switch Q4 are complementaryly turned on, and the drive signals of the first switch Q1 and the third switch Q3 are in phase, and the drive signals of the second switch Q2 and the fourth switch Q4 are in phase.

[0226] It is understandable that complementary conduction means that two switching transistors alternately conduct within a switching cycle. When one switching transistor is on, the other switching transistor is off, and vice versa.

[0227] In some embodiments of this application, the drive signal g1 of the first switch Q1 and the drive signal g2 of the second switch Q2 are complementary, and the drive signal g3 of the third switch Q3 and the drive signal g4 of the fourth switch Q4 are complementary.

[0228] Specifically, drive signal g1 and drive signal g3 are in phase, that is, g1 and g3 are both high or both low; drive signal g2 and drive signal g4 are in phase, that is, g2 and g4 are both high or both low.

[0229] In some embodiments of this application, the duty cycle of each drive signal is 50% in full-bridge mode.

[0230] It is understandable that the duty cycle mentioned above is a theoretical value ignoring dead time.

[0231] In some embodiments of this application, in order to prevent short circuits caused by simultaneous conduction of the upper and lower switches of the same bridge arm, dead time is set between the first switch Q1 and the second switch Q2, and between the third switch Q3 and the fourth switch Q4.

[0232] Specifically, during the dead time, both the upper and lower switches of the same bridge arm are in the off state. The length of the dead time can be set according to the turn-on and turn-off delay times of the switches.

[0233] It is understandable that, ignoring the dead time, in one switching cycle, the first switch Q1 and the third switch Q3 are turned on for half a cycle, and the second switch Q2 and the fourth switch Q4 are turned on for the other half cycle.

[0234] Understandably, in full-bridge mode, the first and second bridge arms work simultaneously, and the input voltage Vi is applied to the resonant network through the alternating switching of the two bridge arms, which can achieve a higher voltage gain.

[0235] In some embodiments of this application, when the operating mode is half-bridge mode, the controller generates drive signals for each switch according to the drive logic of half-bridge mode.

[0236] Specifically, when the operating mode is half-bridge mode, the first switch Q1 and the second switch Q2 are turned on complementaryly, the third switch Q3 is kept off, and the fourth switch Q4 is kept on.

[0237] In some embodiments of this application, the drive signal g1 of the first switch Q1 and the drive signal g2 of the second switch Q2 are complementary, and the duty cycle of both drive signals g1 and g2 is 50%.

[0238] Understandably, in half-bridge mode, a dead time also needs to be set between the first switch Q1 and the second switch Q2 to prevent shoot-through short circuit.

[0239] Specifically, the drive signal g3 of the third switch Q3 is kept at a low level, so that the third switch Q3 is always in the off state; the drive signal g4 of the fourth switch Q4 is kept at a high level, so that the fourth switch Q4 is always in the on state.

[0240] Understandably, in half-bridge mode, only the first bridge arm participates in the switching action, while the fourth switch Q4 in the second bridge arm is normally open, forming a direct path. The input voltage Vi is applied to the resonant network through the switching of the first bridge arm, resulting in a voltage gain that is approximately half that of the full-bridge mode.

[0241] In some embodiments of this application, the controller outputs drive signals g1, g2, g3, and g4 through the PWM module to drive the first switch Q1, the second switch Q2, the third switch Q3, and the fourth switch Q4, respectively.

[0242] Specifically, the PWM module generates corresponding PWM waveforms as drive signals for each switching transistor based on the final switching frequency fs and the current operating mode.

[0243] Understandably, by generating different drive logic according to the operating mode, the converter can flexibly switch between full-bridge mode and half-bridge mode to achieve a wide range of voltage gain adjustment.

[0244] Please see Figure 5 and Figure 6 , Figure 5 This is a waveform diagram of the output voltage of the switching feedforward control loop of the LLC full-half-bridge converter circuit under the switching frequency control method provided by the embodiments of the present invention. Figure 6 This is a waveform diagram of the feedforward frequency and the actual switching frequency under the switching frequency control method provided by the embodiments of the present invention.

[0245] In some embodiments of this application, the switching frequency control method provided in this application is simulated and verified.

[0246] As an example rather than a limitation, the full-bridge mode switching threshold V_FBth is set to 300V and the half-bridge mode switching threshold V_HBth is set to 310V in the simulation test.

[0247] Specifically, the input voltage Vi changes in the following order: 150V → 250V → 350V → 300V → 150V.

[0248] Understandably, in the above test scenario, the input voltage Vi underwent a process of changing from low to high and then back to low, covering various operating conditions such as full-bridge mode, half-bridge mode, and mode switching.

[0249] In some embodiments of this application, the conventional LLC single-voltage loop PI loop control is compared with the LLC full-half-bridge switching feedforward control provided in this application.

[0250] Please see Figure 2 , Figure 2 This is a waveform diagram of the output voltage of the single-voltage loop (PI loop) of a traditional LLC full-half-bridge converter circuit.

[0251] Specifically, under traditional LLC single-voltage-loop PI control, the output voltage Vo changes significantly when the input voltage Vi changes. This is especially true during the full-half-bridge switching phase, where the output voltage Vo exhibits a noticeable abrupt change.

[0252] Understandably, traditional single-voltage-loop PI control has an inherent response lag characteristic, which cannot anticipate and compensate for the gain step caused by topology switching, resulting in a sudden change in output voltage at the moment of mode switching.

[0253] Please see Figure 5 , Figure 5 This is a waveform diagram of the output voltage of the switching feedforward control loop of the LLC full-half-bridge converter circuit under the switching frequency control method provided in this application.

[0254] Specifically, Figure 5 The upper part displays the waveform of the input voltage Vi. Figure 5 The lower half of the display shows the waveform of the corresponding output voltage Vo.

[0255] In some embodiments of this application, after employing the switching frequency control method provided in this application, the output voltage Vo remains basically stable under the same input voltage variation conditions. No significant abrupt change in output voltage Vo occurs during the full-half-bridge mode switching.

[0256] It is understood that the feedforward control method provided in this application can calculate the optimal feedforward switching frequency in advance through an iterative algorithm when the input voltage changes, and pre-compensate for the gain step caused by the topology change, thereby effectively suppressing the sudden change in output voltage.

[0257] In some embodiments of this application, the switching frequency control method provided in this application can also significantly improve the dynamic response speed of the converter when the input voltage changes.

[0258] Please see Figure 6 , Figure 6 This is a waveform diagram of the feedforward frequency and the actual switching frequency under the switching frequency control method provided in this application.

[0259] Specifically, Figure 6 The upper part shows the curve of feedforward frequency fs_ff and the waveform of actual system switching frequency fs. Figure 6 The lower half displays the changes in the working mode.

[0260] In some embodiments of this application, Mode=1 represents full-bridge mode and Mode=0 represents half-bridge mode.

[0261] Specifically, under the set input voltage variation and full-half-bridge mode switching test conditions, the feedforward frequency fs_ff curve calculated in real time by the iterative algorithm proposed in this application basically coincides with the actual switching frequency fs waveform of the converter.

[0262] Understandably, the high consistency between the feedforward frequency fs_ff and the actual switching frequency fs indicates that the numerical iterative algorithm proposed in this application can accurately calculate the optimal switching frequency under the current operating conditions. The frequency compensation amount fs_pi output by the PI controller only plays a fine-tuning role, further verifying the dominant position and accuracy of the feedforward control.

[0263] Simulation results in some embodiments of this application show that: (1) The switching frequency control method provided in this application can effectively suppress sudden changes in output voltage during full-half-bridge mode switching, and the output voltage remains basically stable.

[0264] (2) The switching frequency control method provided in this application can significantly improve the dynamic response speed of the converter when the input voltage changes.

[0265] (3) The feedforward frequency calculated in real time by the iterative algorithm proposed in this application is highly consistent with the actual switching frequency of the converter, which verifies the accuracy and effectiveness of the algorithm.

[0266] It is understandable that by introducing a frequency feedforward calculation module based on an iterative algorithm, this application can calculate the optimal switching frequency that matches the current operating condition in advance and accurately when the input voltage changes, thus compensating for the gain step caused by topology changes in advance and fundamentally avoiding the voltage surge problem that traditional pure feedback control cannot suppress due to response lag.

[0267] Please see Figure 7 , Figure 7 This is a schematic diagram of an LLC full-half-bridge resonant converter provided by an embodiment of the present invention. In some embodiments of this application, an LLC full-half-bridge resonant converter is provided, including a power conversion circuit, a resonant network, and a controller 300.

[0268] In some embodiments of this application, the power conversion circuit includes a switching bridge arm 110 on the input side and a rectifier circuit 120 on the output side.

[0269] Specifically, the switch bridge arm 110 includes a first switch Q1, a second switch Q2, a third switch Q3, and a fourth switch Q4.

[0270] In some embodiments of this application, a first switch Q1 and a second switch Q2 are connected in series to form a first bridge arm, and a third switch Q3 and a fourth switch Q4 are connected in series to form a second bridge arm. The first bridge arm and the second bridge arm are connected in parallel and then connected to the input voltage Vi.

[0271] Specifically, the drive signal for the first switch Q1 is g1, the drive signal for the second switch Q2 is g2, the drive signal for the third switch Q3 is g3, and the drive signal for the fourth switch Q4 is g4.

[0272] In some embodiments of this application, the rectifier circuit 120 is connected to the secondary winding of the transformer Tr to rectify AC voltage into DC voltage. The output capacitor Co is connected in parallel to the output terminal of the rectifier circuit 120, and the load Ro is connected in parallel with the output capacitor Co.

[0273] In some embodiments of this application, the resonant network 200 includes a resonant inductor Lr, a resonant capacitor Cr, and a magnetizing inductor Lm.

[0274] Specifically, the resonant inductor Lr, the resonant capacitor Cr, and the magnetizing inductor Lm constitute a resonant network 200, which is connected between the midpoint of the first bridge arm and the midpoint of the second bridge arm.

[0275] In some embodiments of this application, the primary winding of transformer Tr is connected in parallel with the magnetizing inductance Lm.

[0276] Understandably, the resonant frequency fr of the resonant network 200 is determined by the resonant inductance Lr and the resonant capacitance Cr. The inductance ratio k is the ratio of the magnetizing inductance Lm to the resonant inductance Lr.

[0277] In some embodiments of this application, the controller 300 is used to perform the switching frequency control method as described in any of the above embodiments.

[0278] Specifically, the controller 300 acquires the input voltage Vi and output current Io of the converter in real time; based on the comparison result of the input voltage Vi and the preset mode switching threshold, it determines the working mode of the converter; it calculates the circuit parameters according to the input voltage Vi, the output voltage setpoint Vo_set, and the output current Io, and based on the circuit parameters and the gain mathematical model of the converter, it uses a numerical iterative algorithm to solve for the target normalized frequency, thereby obtaining the feedforward switching frequency; it superimposes the feedforward switching frequency with the output of the feedback controller to obtain the final switching frequency; and based on the final switching frequency and the working mode, it generates the drive signals g1, g2, g3, and g4 for each switch in the converter.

[0279] In some embodiments of this application, the controller 300 may be implemented using a digital signal processor (DSP).

[0280] It is understood that the LLC full-half-bridge resonant converter provided in this application, by executing the aforementioned switching frequency control method through the controller 300, can pre-compensate for gain step during full-half-bridge mode switching, effectively suppress output voltage surges, and improve the dynamic response speed of the converter.

[0281] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and not to limit them; under the concept of this application, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of this application as described above. For the sake of brevity, they are not provided in detail; although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A switching frequency control method applied to a converter, the converter being an LLC full- half bridge resonant converter, characterized in that, The method comprises: real-time acquisition of input voltage and output current of the converter; determination of the working mode of the converter based on the comparison result of the input voltage and the preset mode switching threshold; wherein the working mode comprises full-bridge mode and half-bridge mode; calculation of circuit parameters according to the input voltage, output voltage set value, output current and device parameters of the converter; the circuit parameters comprise target gain; determination of initial normalized frequency according to the value range of the target gain; solving of target normalized frequency through iterative calculation based on the circuit parameters, the initial normalized frequency and the gain mathematical model of the converter, and then obtaining of feedforward switching frequency; generation of driving signals of each switch tube in the converter according to the feedforward switching frequency and the working mode.

2. The method of claim 1, wherein, The determination of the working mode of the converter based on the comparison result of the input voltage and the preset mode switching threshold comprises: comparison of the input voltage with full-bridge mode switching threshold and half-bridge mode switching threshold respectively; wherein the half-bridge mode switching threshold is greater than the full-bridge mode switching threshold; determination of the working mode as full-bridge mode when the input voltage is lower than the full-bridge mode switching threshold; determination of the working mode as half-bridge mode when the input voltage is higher than the half-bridge mode switching threshold; maintaining of the current working mode when the input voltage is between the full-bridge mode switching threshold and the half-bridge mode switching threshold.

3. The method of claim 1, wherein, The circuit parameters further comprise inductance ratio and quality factor; the calculation formula of the target gain Mtarget when the working mode is full-bridge mode is: the calculation formula of the target gain Mtarget when the working mode is half-bridge mode is: wherein Vo_set is the output voltage set value, Vi is the input voltage, and N is transformer turns ratio; the calculation formula of the quality factor Q is: wherein Lr is resonant inductance, Cr is resonant capacitance, and Rac is equivalent AC resistance; the calculation formula of the equivalent AC resistance Rac is: wherein Ro is DC load resistance, and the DC load resistance is determined according to the ratio of output voltage Vo and output current Io; the calculation formula of the inductance ratio k is: wherein Lm is excitation inductance.

4. The method of claim 1, wherein, The determination of initial normalized frequency according to the value range of the target gain comprises: the initial normalized frequency takes a first preset value when the target gain is greater than a first threshold value; the initial normalized frequency takes a second preset value when the target gain is greater than a second threshold value and not greater than the first threshold value; the initial normalized frequency takes a third preset value when the target gain is greater than a third threshold value and not greater than the second threshold value; the initial normalized frequency takes a fourth preset value when the target gain is greater than a fourth threshold value and not greater than the third threshold value; the initial normalized frequency takes a fifth preset value when the target gain is not greater than the fourth threshold value. The first threshold value is greater than the second threshold value, the second threshold value is greater than the third threshold value, and the third threshold value is greater than the fourth threshold value; the first preset value is less than the second preset value, the second preset value is less than the third preset value, the third preset value is less than the fourth preset value, and the fourth preset value is less than the fifth preset value.

5. The method of claim 4, wherein, When the target gain is greater than the first threshold value, the converter operates in a deep capacitive region; When the target gain is greater than the second threshold value and not greater than the first threshold value, the converter operates in a capacitive region; When the target gain is greater than the third threshold value and not greater than the second threshold value, the converter operates in a resonant region; When the target gain is greater than the fourth threshold value and not greater than the third threshold value, the converter operates in an inductive region; When the target gain is not greater than the fourth threshold value, the converter operates in a deep inductive region.

6. The method of claim 3, wherein, The gain mathematical model is expressed as a function relationship of a normalized frequency, an inductance ratio and a quality factor; The normalized frequency fn is a ratio of a switching frequency fs and a resonant frequency fr: The calculation formula of the voltage gain M is: 。 7. The method of claim 6, wherein, The target normalized frequency is solved by iterative calculation based on the circuit parameters, the initial normalized frequency and the gain mathematical model of the converter, comprising: Taking the initial normalized frequency as a current normalized frequency; Calculating a current gain according to the current normalized frequency and the gain mathematical model; computing an error function value between the current gain and the target gain, the error function is computed as wherein, Mcurrent is the current gain calculated for the gain mathematical model, and Mtarget is the target gain. When the absolute value of the error function value is less than a preset tolerance, taking the current normalized frequency as the target normalized frequency; When the absolute value of the error function value is not less than the preset tolerance, updating the current normalized frequency according to the error function value and a derivative value of the error function, and the iterative formula is: wherein, is the normalized frequency for the i-th iteration, is the normalized frequency for the i+1-th iteration, is the error function value for the i-th iteration, is the derivative value of the error function for the i-th iteration; Returning to the step of calculating the current gain according to the current normalized frequency and the gain mathematical model.

8. The method of claim 7, wherein, The derivative value of the error function is calculated by using a central difference method, and the calculation formula is: where h is the difference step size, is the function value offset by h forward at fn, is the function value offset by h backward at fn.

9. The method of claim 7, wherein, The updating of the current normalized frequency according to the error function value and the derivative value of the error function comprises: Updating the current normalized frequency according to a ratio of the error function value and the derivative value of the error function to obtain an updated normalized frequency; When the updated normalized frequency is out of a preset frequency range, adjusting the updated normalized frequency to be within the preset frequency range by using a dichotomy method.

10. The method of claim 7, wherein, When the number of iterations reaches a preset maximum number of iterations, taking a normalized frequency corresponding to the minimum error function value in the iteration process as the target normalized frequency; When the absolute value of the derivative value of the error function is less than a preset threshold value, terminating the iteration, and taking an optimal normalized frequency as the target normalized frequency.

11. The method of claim 1, wherein, Further comprising: Collecting an output voltage of the converter in real time; Superimposing the feedforward switching frequency and an output of the feedback controller to obtain a final switching frequency; Generating a driving signal of each switch tube in the converter according to the final switching frequency and the working mode.

12. The method of claim 11, wherein, The feedback controller is a PI controller; The PI controller adjusts according to the deviation between the output voltage set value and the output voltage, and outputs a frequency compensation amount; The superposition of the feedforward switching frequency and the output of the feedback controller comprises: adding the feedforward switching frequency and the frequency compensation amount to obtain the final switching frequency.

13. The method of claim 12, wherein, The frequency compensation amount is a fine-tuning compensation amount that can be positive or negative; When the feedforward switching frequency is lower than the actual required frequency, the frequency compensation amount is positive, used for positively compensating the feedforward switching frequency; When the feedforward switching frequency is higher than the actual required frequency, the frequency compensation amount is negative, used for negatively compensating the feedforward switching frequency.

14. An LLC full- and half-bridge resonant converter, characterized by, The method comprises: a power conversion circuit comprising a switching bridge arm on the input side and a rectifier circuit on the output side, the switching bridge arm comprising a first switch tube, a second switch tube, a third switch tube and a fourth switch tube; a resonant network comprising a resonant inductor, a resonant capacitor and an excitation inductor; a controller for executing the switching frequency control method according to any one of claims 1 to 13.

Images