A method for dynamic power distribution for multiple charging interfaces
By dynamically tuning the switching frequency and drive phase of the asymmetric half-bridge converter, the problem of matching frequency changes with heat dissipation paths in dynamic power distribution across multiple charging interfaces is solved, thereby reducing hard switching losses and balancing thermal stress, and improving the stability and reliability of charging equipment.
Patent Information
- Application Number
- CN202611407016.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-11
- Publication Date
- 2026-11-10
AI Technical Summary
In the dynamic power distribution process of multiple charging interfaces, the existing technology cannot effectively match the rate of change of the operating frequency with the transient heat conduction capability of the multi-layer heat dissipation path. This causes the asymmetric half-bridge conversion circuit to easily enter the capacitive turn-on state, resulting in hard switching losses and local heat accumulation, which affects the thermal stress balance of the device and the power supply stability.
By collecting protocol messages and real-time electrical parameters from each output charging interface, and combining them with temperature and thermal resistance parameters, the switching frequency of the asymmetric half-bridge converter is dynamically tuned. The driving phase adjustment of the secondary synchronous rectifier tube is used to smoothly transition to the target frequency, ensuring zero-voltage turn-on of the inductive circuit and balancing thermal stress.
This technology reduces hard switching losses, minimizes heat buildup, improves the thermal stability and operational reliability of charging equipment, and avoids the risk of device damage during the power redistribution process across multiple charging interfaces.
Smart Images

Figure CN123074032A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of power conversion and power control technology for charging equipment, and more specifically, to a dynamic power allocation method for multiple charging interfaces. Background Technology
[0002] Multi-interface charging devices typically include a power factor correction circuit, an asymmetric half-bridge converter, and a multi-layer heat dissipation structure formed by power devices, heat sinks, and a housing. The device allocates output power according to the protocol messages of each charging interface. When a charging interface is connected to the device or the charging protocol changes, the control unit needs to redetermine the target output power of each interface in a short period of time.
[0003] When responding to power redistribution requests, existing charging equipment typically changes its operating frequency directly based on a pre-set power allocation relationship. While the change in electrical parameters can be completed in microseconds, the heat generated by the power devices must travel through a heat dissipation path with thermal resistance and capacitance to transfer from the device to the heat sink and then to the casing. This means temperature changes lag significantly behind changes in electrical parameters. During high-power redistribution, if the operating frequency changes abruptly, the conversion circuit is prone to deviating from its inductive operating region and entering a capacitive turn-on state. This causes the primary-side power switch to lose its zero-voltage conduction condition, resulting in capacitive turn-on losses. Simultaneously, the high-frequency current generated by the frequency jump causes significant Joule heating in the die junction region within a short time. Before this heat can be transferred to the heat sink and casing, it tends to accumulate near the die, causing a sudden increase in localized thermal stress. In severe cases, this can lead to device thermal failure, as illustrated in publication number CN113394861A. A Chinese invention patent application discloses a dynamically adjustable LLC smart charger, which dynamically adjusts the PFC output voltage to match the single-channel battery voltage, thereby keeping the LLC circuit operating at a fixed resonant frequency. However, this type of solution is only suitable for single-load, slowly changing battery charging conditions. Its control mechanism relies entirely on the static tracking of the front-end PFC voltage, without considering the rapid frequency conversion requirements caused by the frequent power reconfiguration of each channel in the multi-interface concurrent fast charging scenario of the asymmetric half-bridge architecture. Furthermore, it ignores the transient heat storage and thermal hindrance effects of multiple physical heat conduction channels such as power devices, heat sinks, and casings, as well as the dynamic constraint of transformer excitation energy storage on zero-voltage turn-on margin during frequency conversion transition. When there are sudden power changes in multiple interfaces, if the frequency modulation slope cannot be dynamically constrained in real time based on the junction temperature margin of the devices and the transient heat conduction rate of multiple layers, it is very easy to cause local heat accumulation in the power switching tubes or loss of soft switching characteristics due to insufficient excitation energy.
[0004] Therefore, the technical problem to be solved by this invention is how to match the rate of change of the operating frequency with the transient thermal conductivity of the multi-layer heat dissipation path during the dynamic process of power redistribution through multiple charging interfaces, while ensuring that the asymmetric half-bridge converter circuit has the margin required to maintain zero voltage turn-on in the inductive region, thereby reducing hard switching losses and local heat accumulation in the die, and maintaining stable power supply through multiple interfaces and balanced thermal stress of the devices. Summary of the Invention
[0005] To address the problems in the background art, the present invention provides a dynamic power allocation method with multiple charging interfaces, comprising the following steps:
[0006] Step S1: Collect the protocol handshake messages and real-time electrical parameters of each output charging interface, calculate the target power allocation of each output charging interface and the target output voltage of the bus, and determine the target switching frequency of the asymmetric half-bridge converter and the dynamic tuning window of the switching frequency to maintain inductive zero voltage turn-on.
[0007] Step S2: Collect the package temperature of the gallium nitride power switch die, the temperature of the heat sink substrate, and the outer surface temperature of the outer shell. Combine the preset thermal resistance and thermal capacity parameters of each layer to calculate the transient thermal conduction rate of the first layer from the gallium nitride power switch to the heat sink and the transient thermal conduction rate of the second layer from the heat sink to the outer shell.
[0008] Step S3: Determine the turn-on margin correction factor based on the ratio of transformer excitation energy storage to parasitic capacitance pumping energy of the half-bridge node in the asymmetric half-bridge converter. Multiply the ratio of the first-layer transient thermal conductivity to the second-layer transient thermal conductivity, the remaining junction temperature margin of the gallium nitride power switch, and the turn-on margin correction factor to calculate the maximum allowable change slope of the switching frequency.
[0009] Step S4: Drive the operating frequency to smoothly transition to the target switching frequency within the dynamic tuning window of the switching frequency according to the maximum allowable change slope of the switching frequency, and apply dynamic feedforward compensation to the drive-off phase of the secondary synchronous rectifier of the asymmetric half-bridge converter according to the real-time switching cycle during the transition.
[0010] Preferably, step S1, determining the target switching frequency of the asymmetric half-bridge converter and the dynamic tuning window for maintaining inductive zero-voltage turn-on, includes the following sub-steps: Step S11, calculating the target DC voltage gain based on the target output voltage of the bus and the bus input voltage output by the power factor correction circuit, and calculating the normalized gain equation of the asymmetric half-bridge by combining the equivalent AC load impedance of the transformer primary side corresponding to the target power allocation and the asymmetric duty cycle, and obtaining the target switching frequency; Step S12, calculating the critical state of the transformer excitation current extracting the parasitic capacitance charge of the half-bridge node during the dead time at the moment the primary side switch is turned off, extracting the lower and upper limits of the inductive operating frequency, and forming the dynamic tuning window for the switching frequency.
[0011] Preferably, step S2 includes the following sub-steps: Step S21, calculate the first layer temperature difference between the gallium nitride power switch die package temperature and the heat sink substrate temperature, divide the first layer temperature difference by the first layer equivalent thermal resistance from the device to the heat sink to obtain the first static heat flow component, multiply the device's own equivalent heat capacity by the temperature rise rate of the gallium nitride power switch die package temperature to obtain the first transient heat storage component, and determine the first layer transient thermal conductivity rate by the sum of the first static heat flow component and the first transient heat storage component; Step S22, calculate the second layer temperature difference between the heat sink substrate temperature and the outer surface temperature of the outer casing, divide the second layer temperature difference by the second layer equivalent thermal resistance from the heat sink to the outer casing to obtain the second static heat flow component, multiply the heat sink's equivalent heat capacity by the temperature rise rate of the heat sink substrate temperature to obtain the second transient heat storage component, and determine the second layer transient thermal conductivity rate by the sum of the second static heat flow component and the second transient heat storage component.
[0012] Preferably, the maximum allowable change slope of the switching frequency calculated in step S3 satisfies: ,in, The slope is adjusted to the switching frequency reference. This refers to the maximum allowable operating junction temperature of gallium nitride power switches. The real-time acquisition of the gallium nitride power switch die package temperature. To preset the reference temperature difference margin, The transient thermal conductivity of the first layer, For the transient thermal conductivity of the second layer, It is a positive bias constant. This is the opening margin correction factor.
[0013] Preferably, determining the turn-on margin correction factor in step S3 includes the following sub-steps: Step S31, calculate the peak value of the transformer excitation current based on the real-time switching cycle and the current asymmetric conduction duty cycle, and calculate the inductive reactive power stored in the transformer excitation inductance at the moment of turn-off; Step S32, calculate the minimum pumping energy required for the parasitic capacitance of the half-bridge node to complete charging and discharging; Step S33, calculate the ratio of the inductive reactive power to the minimum pumping energy; Step S34, in response to the ratio being greater than a preset energy margin threshold, configure the turn-on margin correction factor to a preset upper limit saturation value; in response to the ratio not being greater than the preset energy margin threshold, linearly reduce the turn-on margin correction factor to a preset lower limit cutoff value as the ratio decreases.
[0014] Preferably, the dynamic feedforward compensation applied to the drive-off phase of the secondary synchronous rectifier based on the real-time switching cycle in step S4 includes the following sub-steps: Step S41, during the operating frequency transition, the turn-off advance angle is calculated based on the real-time switching frequency, drive propagation delay time, and body diode reverse recovery time margin; Step S42, the drive pulse conduction width of the secondary synchronous rectifier is clamped to the difference electrical angle obtained by subtracting the turn-off advance angle from the reference electrical angle corresponding to the secondary conduction, and the secondary synchronous rectifier is driven to turn off in advance before the secondary current crosses zero in the reverse direction.
[0015] Preferably, the method further includes: in response to detecting that the temperature difference between the heat sink base temperature and the outer surface temperature of the housing decreases to 0°C and the temperature rise rate of the heat sink base temperature is 0°C / s, determining that the transient thermal conductivity of the second layer is 0W; locking the maximum allowable change slope of the switching frequency to 0Hz / s, and stopping the jump adjustment of the operating frequency to the target switching frequency, maintaining the current power distribution state.
[0016] Preferably, the method further includes: in response to a sudden drop in total power demand caused by disconnection of any output charging interface, shifting the target switching frequency to a higher frequency end; reducing the turn-on margin correction factor based on the reduced excitation energy storage to pumping energy ratio to narrow the maximum allowable change slope of the switching frequency; and driving the operating frequency to linearly increase to the target high-frequency operating point by a preset step size according to the narrowed maximum allowable change slope of the switching frequency.
[0017] Preferably, step S4, which involves driving the operating frequency to transition to the target switching frequency within the switching frequency dynamic tuning window according to the maximum permissible change slope of the switching frequency, includes the following sub-steps: Step S43, calculating the absolute difference between the real-time operating frequency and the target switching frequency to be adjusted; Step S44, dividing the absolute difference between the frequency to be adjusted by the maximum permissible change slope of the switching frequency to calculate the total smooth frequency modulation transition time; Step S45, within the total smooth frequency modulation transition time, dividing multiple interpolation control intervals with a preset control cycle as the step size, updating the operating frequency according to the product of the maximum permissible change slope of the switching frequency and the control cycle within each control step size, until the operating frequency reaches the target switching frequency.
[0018] Preferably, the method further includes: during the transition of the operating frequency to the target switching frequency, simultaneously acquiring the real-time output sampling values of the step-down converter circuits after each output charging interface; calculating the target duty cycle of each step-down converter circuit based on the target allocated power of each output charging interface and the real-time output voltage of the bus; adjusting the duty cycle of each step-down converter circuit according to a timing gradient synchronized with the operating frequency change step size, smoothing transient fluctuations in the bus voltage, and completing the power redistribution of multiple interfaces.
[0019] The embodiments of the present invention have at least the following beneficial effects:
[0020] 1. In the dynamic power distribution of multiple charging interfaces, by making the operating frequency change smoothly within the dynamic tuning window of the switching frequency, and in conjunction with the turn-off phase adjustment of the secondary synchronous rectifier, the asymmetric half-bridge converter can remain in the inductive operating region during the power redistribution process of multiple charging interfaces, avoiding the entry into the capacitive turn-on state due to sudden changes in the operating frequency. At the same time, the secondary synchronous rectifier can be turned off in time before the secondary current crosses zero in reverse, thereby reducing the hard switching loss of the primary gallium nitride power switch and reducing the bus voltage fluctuation caused by the reverse current of the secondary side, making the power switching process of multiple interfaces more stable.
[0021] 2. By utilizing the temperature changes between the power switch, heat sink, and casing, as well as the thermal resistance and thermal capacitance of each layer to reflect the actual heat dissipation state, and thereby constraining the maximum allowable change slope of the switching frequency, the adjustment speed of the operating frequency can be adapted to the speed of heat transfer from the device to the outside. This reduces the accumulation of heat in the junction region of the power switch in a short time during high-power operation and power redistribution, reduces the risk of over-temperature false protection and device thermal breakdown caused by sudden increase in local thermal stress, and helps to improve the thermal stability of charging equipment under high-power conditions.
[0022] 3. By adjusting the switching frequency change rate according to the relationship between the transformer excitation energy storage and the energy required for charging and discharging the parasitic capacitance of the half-bridge node, the operating frequency can be prevented from changing to the high-frequency side too quickly during light load switching processes such as disconnection of the charging interface and sudden reduction of total load. This can reserve sufficient excitation energy for charging and discharging of the parasitic capacitance of the half-bridge node, reduce the situation of losing the zero-voltage turn-on condition due to insufficient excitation energy storage under light load conditions, reduce the risk of device damage caused by hard turn-on, and improve the operational reliability of multi-interface charging equipment when the load changes rapidly. Attached Figure Description
[0023] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description taken in conjunction with the accompanying drawings, in which several embodiments of the invention are illustrated by way of example and not limitation, wherein:
[0024] Figure 1 This is a flowchart of the smooth frequency conversion control of the present invention, which combines temperature rise thermal conductivity and energy storage margin.
[0025] Figure 2 This is a diagram showing the thermoelectric coordinated control architecture and signal interaction of the multi-channel charging device of the present invention. Detailed Implementation
[0026] The specific technical solutions of the present invention will now be described with reference to the accompanying drawings and several exemplary embodiments. It should be understood that these embodiments are provided only to enable those skilled in the art to better understand and implement the present invention, and are not intended to limit the scope of the present invention in any way. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.
[0027] A dynamic power allocation method with multiple charging ports includes the following steps:
[0028] Step S1: Collect the protocol handshake messages and real-time electrical parameters of each output charging interface, calculate the target power allocation of each output charging interface and the target output voltage of the bus, and determine the target switching frequency of the asymmetric half-bridge converter and the dynamic tuning window of the switching frequency to maintain inductive zero voltage turn-on.
[0029] Step S2: Collect the package temperature of the gallium nitride power switch die, the temperature of the heat sink substrate, and the outer surface temperature of the outer shell. Combine the preset thermal resistance and thermal capacity parameters of each layer to calculate the transient thermal conduction rate of the first layer from the gallium nitride power switch to the heat sink and the transient thermal conduction rate of the second layer from the heat sink to the outer shell.
[0030] Step S3: Determine the turn-on margin correction factor based on the ratio of transformer excitation energy storage to parasitic capacitance pumping energy of the half-bridge node in the asymmetric half-bridge converter. Multiply the ratio of the first-layer transient thermal conductivity to the second-layer transient thermal conductivity, the remaining junction temperature margin of the gallium nitride power switch, and the turn-on margin correction factor to calculate the maximum allowable change slope of the switching frequency.
[0031] Step S4: Drive the operating frequency to smoothly transition to the target switching frequency within the dynamic tuning window of the switching frequency according to the maximum allowable change slope of the switching frequency, and apply dynamic feedforward compensation to the drive-off phase of the secondary synchronous rectifier of the asymmetric half-bridge converter according to the real-time switching cycle during the transition.
[0032] Preferably, step S1, determining the target switching frequency of the asymmetric half-bridge converter and the dynamic tuning window for maintaining inductive zero-voltage turn-on, includes the following sub-steps: Step S11, calculating the target DC voltage gain based on the target output voltage of the bus and the bus input voltage output by the power factor correction circuit, and calculating the normalized gain equation of the asymmetric half-bridge by combining the equivalent AC load impedance of the transformer primary side corresponding to the target power allocation and the asymmetric duty cycle, and obtaining the target switching frequency; Step S12, calculating the critical state of the transformer excitation current extracting the parasitic capacitance charge of the half-bridge node during the dead time at the moment the primary side switch is turned off, extracting the lower and upper limits of the inductive operating frequency, and forming the dynamic tuning window for the switching frequency.
[0033] Preferably, step S2 includes the following sub-steps: Step S21, calculate the first layer temperature difference between the gallium nitride power switch die package temperature and the heat sink substrate temperature, divide the first layer temperature difference by the first layer equivalent thermal resistance from the device to the heat sink to obtain the first static heat flow component, multiply the device's own equivalent heat capacity by the temperature rise rate of the gallium nitride power switch die package temperature to obtain the first transient heat storage component, and determine the first layer transient thermal conductivity rate by the sum of the first static heat flow component and the first transient heat storage component; Step S22, calculate the second layer temperature difference between the heat sink substrate temperature and the outer surface temperature of the outer casing, divide the second layer temperature difference by the second layer equivalent thermal resistance from the heat sink to the outer casing to obtain the second static heat flow component, multiply the heat sink's equivalent heat capacity by the temperature rise rate of the heat sink substrate temperature to obtain the second transient heat storage component, and determine the second layer transient thermal conductivity rate by the sum of the second static heat flow component and the second transient heat storage component.
[0034] Preferably, the maximum allowable change slope of the switching frequency calculated in step S3 satisfies: ,in, The slope is adjusted to the switching frequency reference. This refers to the maximum allowable operating junction temperature of gallium nitride power switches. The real-time acquisition of the gallium nitride power switch die package temperature. To preset the reference temperature difference margin, The transient thermal conductivity of the first layer, For the transient thermal conductivity of the second layer, It is a positive bias constant. This is the opening margin correction factor.
[0035] Preferably, determining the turn-on margin correction factor in step S3 includes the following sub-steps: Step S31, calculate the peak value of the transformer excitation current based on the real-time switching cycle and the current asymmetric conduction duty cycle, and calculate the inductive reactive power stored in the transformer excitation inductance at the moment of turn-off; Step S32, calculate the minimum pumping energy required for the parasitic capacitance of the half-bridge node to complete charging and discharging; Step S33, calculate the ratio of the inductive reactive power to the minimum pumping energy; Step S34, in response to the ratio being greater than a preset energy margin threshold, configure the turn-on margin correction factor to a preset upper limit saturation value; in response to the ratio not being greater than the preset energy margin threshold, linearly reduce the turn-on margin correction factor to a preset lower limit cutoff value as the ratio decreases.
[0036] Preferably, the dynamic feedforward compensation applied to the drive-off phase of the secondary synchronous rectifier based on the real-time switching cycle in step S4 includes the following sub-steps: Step S41, during the operating frequency transition, the turn-off advance angle is calculated based on the real-time switching frequency, drive propagation delay time, and body diode reverse recovery time margin; Step S42, the drive pulse conduction width of the secondary synchronous rectifier is clamped to the difference electrical angle obtained by subtracting the turn-off advance angle from the reference electrical angle corresponding to the secondary conduction, and the secondary synchronous rectifier is driven to turn off in advance before the secondary current crosses zero in the reverse direction.
[0037] Preferably, the method further includes: in response to detecting that the temperature difference between the heat sink base temperature and the outer surface temperature of the housing decreases to 0°C and the temperature rise rate of the heat sink base temperature is 0°C / s, determining that the transient thermal conductivity of the second layer is 0W; locking the maximum allowable change slope of the switching frequency to 0Hz / s, and stopping the jump adjustment of the operating frequency to the target switching frequency, maintaining the current power distribution state.
[0038] Preferably, the method further includes: in response to a sudden drop in total power demand caused by disconnection of any output charging interface, shifting the target switching frequency to a higher frequency end; reducing the turn-on margin correction factor based on the reduced excitation energy storage to pumping energy ratio to narrow the maximum allowable change slope of the switching frequency; and driving the operating frequency to linearly increase to the target high-frequency operating point by a preset step size according to the narrowed maximum allowable change slope of the switching frequency.
[0039] Preferably, step S4, which involves driving the operating frequency to transition to the target switching frequency within the switching frequency dynamic tuning window according to the maximum permissible change slope of the switching frequency, includes the following sub-steps: Step S43, calculating the absolute difference between the real-time operating frequency and the target switching frequency to be adjusted; Step S44, dividing the absolute difference between the frequency to be adjusted by the maximum permissible change slope of the switching frequency to calculate the total smooth frequency modulation transition time; Step S45, within the total smooth frequency modulation transition time, dividing multiple interpolation control intervals with a preset control cycle as the step size, updating the operating frequency according to the product of the maximum permissible change slope of the switching frequency and the control cycle within each control step size, until the operating frequency reaches the target switching frequency.
[0040] Preferably, the method further includes: during the transition of the operating frequency to the target switching frequency, simultaneously acquiring the real-time output sampling values of the step-down converter circuits after each output charging interface; calculating the target duty cycle of each step-down converter circuit based on the target allocated power of each output charging interface and the real-time output voltage of the bus; adjusting the duty cycle of each step-down converter circuit according to a timing gradient synchronized with the operating frequency change step size, smoothing transient fluctuations in the bus voltage, and completing the power redistribution of multiple interfaces.
[0041] Example 1: In a multi-port fast charging device, the first output charging port provides 65W fast charging at 20V and 3.25A, while the second and third output charging ports are in an unloaded state. The second output charging port connects to the device and completes a protocol handshake, requesting 35W power supply at 20V and 1.75A. The system control unit determines that the first output charging port maintains a target power allocation of 65W based on the rated total output power of 100W, and the second output charging port initiates a target power allocation of 35W. The total output power is reconfigured to 100W, and the target output voltage on the bus is maintained at 20V. During this power redistribution transient process, the asymmetric half-bridge converter... If a discrete operating frequency step command is directly used, the microsecond-level abrupt change in the switching state can easily cause the asymmetric half-bridge conversion circuit to deviate from the inductive operating region and enter the capacitive turn-on region. As a result, the primary-side gallium nitride power switch loses the zero-voltage turn-on condition and is subjected to capacitive hard-turn-on discharge impact. At the same time, the Joule heat released instantaneously in the micro-junction region of the power device by the sudden increase in high-frequency current is affected by the time constant and heat capacity of the three-layer solid heat conduction channel, forming local heat accumulation at the die interface and triggering over-temperature protection malfunction. In this embodiment, the system achieves smooth power redistribution by multi-node temperature sensing, transient heat conduction rate calculation, dynamic tuning window extraction of switching frequency, and limiting the maximum allowable change slope of switching frequency.
[0042] The main circuit of the multi-interface charging device adopts a bridgeless interleaved boost power factor correction circuit to output a 400V DC bus voltage. The primary side of the subsequent asymmetric half-bridge converter is equipped with an enhancement-mode gallium nitride power switch. and The asymmetric half-bridge circuit configuration includes the loop inductance of leakage inductance and external inductance. 15μH, DC blocking and resonant capacitor For 22nF and transformer magnetizing inductance The inductance is 90μH, and the preset inductance value is... Transformer turns ratio Half-bridge dead zone time The equivalent output capacitance of a primary-side gallium nitride transistor is set to 100ns. The total parasitic capacitance at the half-bridge node is 120pF. The value is 240pF, and the corresponding intrinsic resonant frequency is calculated as follows: When total demand power is reconfigured And the target output voltage of the bus Equivalent DC load resistance when maintaining 20V The asymmetric equivalent AC load impedance converted to the primary side of the transformer. The approximation calculated based on the turns ratio and topological transmission impedance mapping is approximately... Combined with the characteristic impedance of the circuit Calculate the equivalent quality factor The target DC voltage gain is calculated by combining the bus input voltage and the equivalent input voltage of the transformer primary side. Substituting the values into the normalized gain equation of the asymmetric half-bridge, the normalized target switching frequency is calculated. Corresponding to the target switching frequency Simultaneously, the control unit calculates the amount of parasitic charge that can be extracted from the half-bridge node during the dead time based on the transformer excitation current and dead time at the moment the primary-side switch is turned off. This is compared with the amount of charge required for the parasitic capacitance of the half-bridge node to complete charging and discharging. Based on the critical state, the lower and upper limits of the inductive operating frequency are extracted, and the dynamic tuning window for maintaining zero-voltage switching of the inductive circuit is determined. This allows the current 65W operating frequency to be adjusted. To target switching frequency The adjustment trajectory is located within the dynamic tuning window of the switching frequency.
[0043] The three-layer solid thermal channel consists of the device package, heat sink, and housing in sequence. The die packages of the two gallium nitride power switching transistors are connected to an integrated stamped aluminum alloy heat sink via a copper-clad ceramic substrate and a phase-change thermal grease pad. The first layer of equivalent thermal resistance from the device to the heat sink... The effective heat capacity of the device is 1.8 K / W. The thermal resistance is 0.08 J / K; the aluminum alloy heat sink is bonded to the inner wall of the flame-retardant engineering plastic shell with insulating and thermally conductive sealant, and the second layer of equivalent thermal resistance between the heat sink and the shell is 0.08 J / K. It is 3.2K / W, and the equivalent heat capacity of the heatsink is... The temperature sensor array, with a sampling period of 100 μs, synchronously acquires the starting time of temperature data at temperature measurement nodes corresponding to the device package, heat sink substrate, and outer surface of the housing, with a sampling rate of 2.5 J / K. The node temperature and temperature rise rate were used to obtain the package temperature of the gallium nitride power switch die. The temperature was 72.4℃ and the rate of temperature change was... The heatsink base temperature is 0.25 K / s. The temperature was 56.2℃ and the rate of temperature change was... The speed is 0.12 K / s, and the outer surface temperature of the casing is... The temperature is 42.1℃, from which the temperature difference of the first layer can be calculated. Temperature difference with the second layer The transient thermal conductivity of the first layer was calculated based on the heat conduction equation. With the transient thermal conductivity of the second layer : , ,in, The real-time acquisition of the gallium nitride power switch die package temperature. The temperature of the heat sink substrate is collected in real time. The outer surface temperature of the casing is collected in real time; the parameters are substituted to calculate the current transient thermal conductivity of the first layer. The transient thermal conductivity of the second layer The ratio of the transient thermal conductivity of the first layer to the transient thermal conductivity of the second layer is: Where the positive bias constant The power consumption is 0.01W, while the maximum allowable operating junction temperature of the gallium nitride power switch is [not specified]. With real-time acquisition of gallium nitride power switch die package temperature The difference is used as the residual junction temperature margin for gallium nitride power switches; The temperature is set at 115℃, with a preset reference temperature difference margin. Normalized to 40K, the ratio of remaining junction temperature margin is obtained. In the equivalent model of a multi-layer thermal network, the transient thermal conduction rate of the first layer characterizes the total equivalent heat generation power input to the first layer thermal channel by the power switch during transient operation. The first static heat flow component reflects the conduction heat flow driven by the interface temperature difference to the heat sink, and the first transient heat storage component reflects the energy change rate absorbed and stored by the heat capacity of the die body. The sum of the two characterizes the total instantaneous heat load generated in the junction region of the device, which is used to compare with the thermal conduction rate dissipated from the second layer to the outer shell to quantitatively evaluate the degree of heat accumulation in the intermediate heat dissipation structure.
[0044] At the current operating frequency Corresponding switching cycle Under operating conditions, the microcontroller calculates the peak value of the transformer excitation current. ;in The real-time duty cycle of the primary power switch is set to 0.5 at the preset symmetrical calibration reference operating point corresponding to the current nominal test condition. In actual asymmetrical operation under varying conditions, the duty cycle D is adaptively updated by the bus input voltage and the target output voltage. Here, combining the real-time switching cycle and the asymmetrical conduction timing, the current peak excitation current is calculated to be approximately 2.096A. The transformer excitation energy stored in the transformer excitation inductance at the moment of turn-off is then calculated. Simultaneously calculate the energy required to complete the charging and discharging of the parasitic capacitance at the half-bridge node. The ratio of transformer excitation energy storage to the energy pumped from the parasitic capacitance of the half-bridge node is obtained. Since this ratio is greater than the preset energy margin threshold of 2, the margin correction factor is activated. The preset upper limit saturation value is set to 0.88; the slope is then adjusted based on the switching frequency reference. Calculate the maximum permissible slope for switching frequency changes. : In the calculation of the maximum allowable change slope of the switching frequency, the ratio of the remaining junction temperature margin to the ratio of the transient thermal conductivity of the first layer to the transient thermal conductivity of the second layer are each compared with 1. The minimum value among the three is taken as the thermal constraint term, which is then multiplied by the turn-on margin correction factor and the reference adjustment slope of the switching frequency, and substituted into the numerical solution to obtain the result. Therefore, based on the absolute difference in frequency to be adjusted Calculate the total transition time of smooth frequency modulation. The switching frequency reference adjustment slope of 50kHz / ms is obtained based on the calibration of the maximum distortion-free tracking rate of the digital phase-locked loop in a single control cycle; the preset reference temperature margin of 40K is taken from the design safety temperature difference range between the maximum allowable operating junction temperature of gallium nitride power switch tube (115℃) and the rated ambient temperature of the equipment (75℃); the positive bias constant of 0.01W is set as the base value for preventing zero bias in the calculation of the first layer transient thermal conductivity; the preset upper limit saturation value of the turn-on margin correction factor of 0.88 is the safety factor after reserving a 12% component parameter tolerance margin.
[0045] Within the total smooth frequency modulation transition time of 1.003ms, the microcontroller unit divides the time into 10 interpolation control intervals with a control cycle of 100μs. Within each control step, the driving operating frequency is shifted by a single step frequency shift. Linearly decreasing update; during the smooth transition of operating frequency, based on the real-time switching cycle. Dynamic feedforward compensation is applied to the drive turn-off phases of the secondary-side synchronous rectifier diodes SR1 and SR2, combined with the drive propagation delay time. The 40ns reverse recovery time margin is the same as that of the body diode. Calculate the turn-off advance angle for 30ns The drive pulse width of the secondary synchronous rectifier diode is clamped to the reference electrical angle corresponding to the secondary rectification conduction. With turn-off advance angle The difference in electrical angle, i.e. The reference electrical angle Determined by the mapping between the current duty cycle and the switching cycle of the asymmetric half-bridge (if it satisfies...). The control unit conducts the corresponding reference electrical angle via the secondary side. As a reference for conduction width, the actual drive pulse of the synchronous rectifier is generated by subtracting the turn-off advance angle from the reference conduction width. The synchronous rectifier is driven to turn off 70ns before the secondary-side rectified current crosses zero in reverse. At the same time, the real-time output sampling values of the buck converter circuits after the first and second output charging interfaces are collected. The target power distribution of each output charging interface and the real-time output voltage of the bus are used as the calculation input for the target conduction duty cycle. The target conduction duty cycle of each buck converter circuit is calculated by combining the real-time output sampling values of the corresponding interfaces. Then, the conduction duty cycle of each buck converter circuit is adjusted according to the timing gradient synchronized with the step size of the operating frequency change. When the converter is operating at a frequency lower than the intrinsic resonant frequency, the actual conduction angle of the secondary-side rectified current in the secondary-side conduction interval changes with the ratio of the operating frequency to the intrinsic resonant frequency. The control unit still uses the corresponding reference conduction angle of the secondary side conduction. As a reference conduction width, the actual drive pulse of the synchronous rectifier is generated by subtracting the turn-off advance angle from this reference conduction width, so that the drive signal is cut off in advance before the secondary current naturally oscillates and crosses zero. During the entire 1.003ms frequency modulation transition, the transient thermal conductivity of the first layer is... The transient thermal conductivity of the second layer smoothly transitioned from 9.02W to 10.85W. Synchronous changes in the package temperature of gallium nitride power switches The transient temperature rise during the entire power redistribution process does not exceed 0.5℃. The asymmetrical half-bridge circuit maintains zero inductive voltage switching throughout and suppresses the secondary side reverse current surge. The bus voltage remains stable during multi-interface coordinated power supply.
[0046] Example 2: The current dynamic power distribution test platform for multi-interface charging equipment includes an AC programmable frequency converter, a three-channel independent programmable DC electronic load, a multi-channel temperature recorder, and a digital oscilloscope with a bandwidth of not less than 500MHz and a sampling rate of not less than 2.5GS / s. The AC input terminal supplies rated 220V AC power, which is output as a 400V DC bus voltage through a bridgeless interleaved boost power factor correction circuit. The subsequent asymmetric half-bridge converter, after center-tapped secondary-side rectification and three-way synchronous buck converter circuits, leads out the first output charging interface Port1, the second output charging interface Port2, and the third output charging interface Port3, respectively. Each output charging interface is connected to the corresponding electronic load channel to simulate the charging protocol handshake and power surge of different electrical devices. The temperature sensing array uses surface-mount thermistor units with a temperature measurement accuracy better than 0.1℃ and a response time of less than 10ms, which are respectively set on the primary-side gallium nitride power switch transistors. and The temperature measurement positions of the package, the geometric center of the aluminum alloy heat sink, and the center of the outer surface of the flame-retardant engineering plastic shell are measured synchronously at a fixed control cycle of 100μs. The temperature of the gallium nitride power switch die package, the temperature of the heat sink base, and the temperature of the outer surface of the shell are collected synchronously. The control unit reads the temperature sampling data in each 100μs discrete control step and stores 100 consecutive sampling points in a sliding window buffer queue. The temperature change trend across a 10ms time scale is extracted by a first-order low-pass filter, and the temperature rise rate is obtained accordingly. This is used for real-time calculation of the maximum allowable change slope of the switching frequency. During the test, a broadband ripple disturbance with a signal-to-noise ratio of 25dB is superimposed at the input. Under the initial steady state, Port1 outputs 20V and 3.25A, corresponding to 65W fast charging. Port2 and Port3 are in an unloaded state. The target output voltage of the bus is... The initial steady-state operating frequency of the asymmetric half-bridge converter is 20V. It is 265kHz.
[0047] The experiment included the sample group of the present invention and multiple control sample groups. The sample group of the present invention tested the transient thermal conductivity of the first layer. With the transient thermal conductivity of the second layer Multiply the ratio, the remaining junction temperature margin of the gallium nitride power switch, and the turn-on margin correction factor to calculate the maximum allowable slope of the switching frequency change. The driving frequency smoothly transitions within a dynamic tuning window of 185kHz to 320kHz switching frequency, and a turn-off advance angle is applied to the secondary synchronous rectifier tube according to the real-time switching cycle. Dynamic feedforward compensation: Control group 1 uses discrete step frequency modulation of the target switching frequency directly after static table lookup; Control group 2 adjusts the slope based on the switching frequency reference. For 50kHz / ms, a fixed-slope linear frequency modulation is performed without using the ratio constraint between the transient thermal conductivity of the first layer and the transient thermal conductivity of the second layer; control group 3 uses a fixed half-cycle turn-on signal to drive the secondary-side synchronous rectifier tube without applying dynamic feedforward compensation for the turn-off advance angle; control group 4 uses a positive bias constant. Set it to 1W, and adjust the slope of the switching frequency reference. Set to 150 kHz / ms; control group 5 adjusts the slope of the switching frequency reference. Set to 5kHz / ms, in addition to the above control conditions, the sample group of this invention was also subjected to dynamic tests under the following conditions: heavy load at 25℃, high temperature at 45℃, thermal hysteresis at 60℃, sudden drop from 100W to 5W light load unloading, and sudden light load processing. The frequency shift, the ratio of the transient thermal conductivity of the first layer to the transient thermal conductivity of the second layer, the ratio of the transformer excitation energy storage to the energy of the parasitic capacitance of the half-bridge node, the frequency modulation slope, the turn-off advance angle, the transient temperature rise of the gallium nitride power switch die package, the primary drain-source voltage spike, the secondary reverse current, and the bus voltage fluctuation.
[0048] During standard loading at 25℃, the Port2 connection increased the total output power from 65W to 100W, and the operating frequency shifted from 265kHz to 242kHz. The ratio of the transient thermal conductivity of the first layer to the transient thermal conductivity of the second layer recorded in the experiment was 0.521, and the real-time solution value of the control unit was 0.5212. The ratio of transformer excitation energy storage to the energy of the parasitic capacitance of the half-bridge node was 5.15, and the turn-on margin correction factor was... The value is 0.88. Based on this, the maximum allowable change slope of the switching frequency is 22.93kHz / ms. The operating frequency is smoothly shifted according to this slope. After applying a 6.3° turn-off advance angle, the transient temperature rise of the gallium nitride power switch die package is 0.46℃, the primary-side drain-source voltage spike is 0V, the secondary-side reverse current is 0.04A, and the bus voltage fluctuation is 0.18V.
[0049] When maintaining the same frequency range of 265kHz to 242kHz for comparison, the ratio of the first-layer transient thermal conductivity to the second-layer transient thermal conductivity in control sample group 1 was 0.521, the ratio of transformer excitation energy storage to half-bridge node parasitic capacitance pumping energy was 5.15, and the frequency modulation adopted a step change with a turn-off advance angle of 0°. Under these conditions, the transient temperature rise of the gallium nitride power switch die package reached 8.74℃, the primary-side drain-source voltage spike was 88.5V, the secondary-side reverse current was 3.42A, and the bus voltage... The voltage fluctuation was 1.85V. For control group 2, the ratios of the two corresponding terms within the same frequency range remained 0.521 and 5.15. With frequency modulation using the fixed slope, the turn-off advance angle was 6.3°. The measured transient temperature rise of the gallium nitride power switch die was 3.65℃, the primary-side drain-source voltage spike was 18.2V, the secondary-side reverse current was 0.28A, and the bus voltage fluctuation was 0.76V. For control group 3, the ratios of the two terms were also 0.521 and 5.15, and the maximum allowable slope for the switching frequency change was 22.93kHz. The actual frequency modulation slope was increased to 65.8 kHz / ms while maintaining a 6.3° turn-off advance angle. The corresponding transient temperature rise of the gallium nitride power switch die package was 1.12°C, the primary-side drain-source voltage spike was 0V, the secondary-side reverse current increased to 2.85A, and the bus voltage fluctuation was 1.42V. In control group 4, under the conditions of ratios of 0.521 and 5.15, the transient temperature rise of the gallium nitride power switch die package was 4.92°C, and the primary-side drain-source voltage spike was... At 34.6V, the secondary-side reverse current was 0.45A, and the bus voltage fluctuation was 0.98V. The control sample group 5 was still operating within the frequency migration range of 265kHz to 242kHz and under the conditions of 0.521 and 5.15 ratios. The actual frequency modulation slope dropped to 2.15kHz / ms, the turn-off advance angle was 6.3°, the transient temperature rise of the gallium nitride power switch die package was 0.25℃, the primary-side drain-source voltage spike was 0V, and the secondary-side reverse current was 0.04A, but the bus voltage fluctuation reached 1.65V.
[0050] Under high-temperature loading conditions of 45℃, the operating frequency shifts from 265kHz to 242kHz. The ratio of the transient thermal conductivity of the first layer to that of the second layer decreases to 0.385. The ratio of the transformer excitation energy storage to the energy of the parasitic capacitance of the half-bridge node remains at 5.15. The maximum allowable change slope of the switching frequency is 16.94kHz / ms, and the turn-off advance angle is 6.3°. During this process, the transient temperature rise of the gallium nitride power switch die package is 0.58℃, the primary-side drain-source voltage spike is 0V, the secondary-side reverse current is 0.05A, and the bus voltage fluctuation is 0.21V.
[0051] Under thermal hindrance conditions at 60℃, when the temperature difference between the heat sink substrate and the outer surface of the casing drops to zero, the transient thermal conductivity of the second layer... When the power output drops to 0W, the ratio of the transient thermal conductivity of the first layer to that of the second layer is 0, the ratio of the transformer excitation energy storage to the energy of the parasitic capacitance of the half-bridge node is 5.15, the maximum allowable change slope of the switching frequency is locked at 0Hz / s, the corresponding frequency modulation slope recorded in the test is 0kHz / ms, the starting and ending operating frequencies are both 265kHz, the turn-off advance angle is 6.3°, the transient temperature rise of the gallium nitride power switch die package is 0.12℃, the primary drain-source voltage spike is 0V, the secondary reverse current is 0.04A, and the bus voltage fluctuation is 0.05V.
[0052] Under light load unloading conditions at 25℃ and with the total output power suddenly dropping from 100W to 5W, the operating frequency shifts from 242kHz to 300kHz. The ratio of the transient thermal conductivity of the first layer to that of the second layer is 0.518, the ratio of transformer excitation energy storage to the energy of the half-bridge node parasitic capacitance being removed decreases to 1.09, and the turn-on margin correction factor... Compressed to 0.1, the maximum allowable change slope of the switching frequency is narrowed to 5kHz / ms, and the turn-off advance angle is extended to 7.6°. During this migration process, the transient temperature rise of the gallium nitride power switch die package is 0.32℃, the primary-side drain-source voltage spike is 0V, the secondary-side reverse current is 0.03A, and the bus voltage fluctuation is 0.24V. Subsequently, a 25℃ light load surge test was performed, and the operating frequency dropped from 300kHz to 242kHz. The first layer transient conduction... The ratio of thermal rate to the transient thermal conductivity of the second layer is 0.492, the ratio of transformer excitation energy storage to half-bridge node parasitic capacitance pumping energy rises to 4.86, the maximum allowable change slope of the switching frequency is 21.65kHz / ms, and the turn-off advance angle is 6.3°; the corresponding transient temperature rise of the gallium nitride power switch die package is 0.49℃, the primary drain-source voltage spike is 0V, the secondary reverse current is 0.05A, and the bus voltage fluctuation is 0.22V.
[0053] In the above experiments, the standard load and each control sample group adopted the same frequency migration start and end conditions and corresponding thermal conductivity and energy storage states. Discrete step frequency modulation caused the operating point of the asymmetric half-bridge converter circuit to cross the inductive boundary, resulting in a simultaneous increase in the primary drain-source voltage spike, the transient temperature rise of the gallium nitride power switch die package, and the secondary reverse current. When fixed slope frequency modulation did not introduce the constraint of the ratio of the first-layer transient thermal conductivity rate to the second-layer transient thermal conductivity rate, the temperature rise and drain-source voltage spike still increased. After canceling the dynamic feedforward compensation of the turn-off advance angle, the secondary reverse current and bus voltage fluctuation increased. Under different environmental conditions, when the ratio of the first-layer transient thermal conductivity rate to the second-layer transient thermal conductivity rate decreased, the maximum allowable change slope of the switching frequency narrowed synchronously. During thermal hysteresis, the slope returned to zero and frequency migration stopped. During light load unloading, the ratio of transformer excitation energy storage to half-bridge node parasitic capacitance deloading energy decreased, the turn-on margin correction factor decreased accordingly, and the operating frequency migrated to the higher frequency end according to the narrowed slope.
[0054] Example 3: When the multi-port charging device is in a dual-port concurrent fast charging steady state, the first output charging interface Port1 provides 65W output corresponding to 20V and 3.25A, and the second output charging interface Port2 provides 35W output corresponding to 20V and 1.75A, with a total output power of 100W. The operating frequency of the asymmetric half-bridge converter is... At 242kHz, after the device connected to the first output charging port (Port1) is momentarily disconnected, the total power demand of the device drops sharply from 100W to 35W maintained by the second output charging port (Port2), and can further enter a 5W light load condition. When the total output power drops to 5W and the target output voltage of the bus is... Equivalent DC load resistance when maintaining 20V The equivalent AC load impedance, calculated to the primary side of the transformer, is 80Ω. Increased to 6484.6Ω, equivalent impedance quality factor The target switching frequency of the asymmetric half-bridge converter drops to 0.004. When migrating to a higher frequency range of 300kHz, during the high-frequency transition from heavy load to light load, directly issuing a high-frequency step command would shorten the switching cycle, reducing the peak value of the primary excitation current of the transformer. Consequently, the transformer excitation energy stored at the moment of turn-off of the primary switch would decrease, making it difficult to utilize the 100ns dead time. Complete the total parasitic capacitance of the half-bridge node. pF charge extraction, primary-side gallium nitride power switch and This could result in the switch being turned on before the midpoint voltage of the half-bridge drops to zero potential. Simultaneously, the zero-crossing point of the secondary current commutation shifts forward with the change in operating frequency, which can easily lead to backflow current and cause bus voltage overshoot. In this state, the control unit reduces the turn-on margin correction factor based on the ratio of transformer excitation energy storage to the energy pumped from the parasitic capacitance of the half-bridge node, narrows the maximum allowable change slope of the switching frequency, and simultaneously adjusts the turn-off advance angle of the secondary synchronous rectifier tube and the duty cycle of the subsequent step-down converter circuit.
[0055] The microcontroller detects load tripping via the physical layer interrupt signal of the protocol chip at the first output charging interface Port1, recalculates the target power allocation for each output charging interface, and sets the target switching frequency accordingly. The operating frequency is set at 300kHz; the shift range of the operating frequency is affected by the dynamic tuning window of the switching frequency. Constraints: When the operating frequency reaches 300kHz, the corresponding switching cycle... The value is 3.3333 μs, and the transformer magnetizing inductance is... μH, the peak excitation current calculated according to the switching cycle and the current asymmetric conduction sequence (taking the reference conduction duty cycle D as 0.5) is: The corresponding transformer excitation energy storage is: When the total output power further stabilizes at the steady-state limit point of 5W, the peak excitation current decays to 0.966A, corresponding to a reduction in transformer excitation energy storage at the moment the primary-side switch is turned off to 42μJ; the energy storage at this steady-state limit point is used to determine the turn-on margin and the total parasitic capacitance of the half-bridge node. Input voltage at 400V bus The energy required to complete the charging and discharging of the parasitic capacitance of the half-bridge node is: Thus, the ratio of transformer excitation energy storage to the energy pumped from the parasitic capacitance of the half-bridge node is obtained: Since this ratio is less than the preset energy margin threshold of 2, the microcontroller activates the soft-switching suppression reduction model and adjusts the turn-on margin correction factor. Linear compression from a preset upper saturation value of 0.88 to a preset lower cutoff value of 0.1: In this case, both the denominator and the subtrahend are set to 1, so that when the energy ratio decreases to the critical pump-off equilibrium point 1, the turn-on margin correction factor transitions to a preset lower limit cutoff value of 0.1, while retaining the basic adjustment step size. At this time, the temperature sensor array collects the package temperature of the gallium nitride power switch die. The temperature of the heat sink base is 68.2℃. The outer surface temperature of the casing is 54.1℃. The ratio of the transient thermal conductivity of the first layer to that of the second layer is 41.5℃. Approximately 0.53, obtained based on the remaining junction temperature margin of gallium nitride power switches. The switching frequency reference adjustment slope is 1.17. Given a frequency of 50 kHz / ms, the maximum permissible slope for the switching frequency change is calculated at the current time step: The absolute difference between the real-time operating frequency and the target switching frequency to be adjusted is: Therefore, the total transition time for smooth frequency modulation is obtained: During the 21.887ms smooth frequency modulation total transition time, the microcontroller unit continuously divides the control into 219 interpolation control intervals with a control period of 100μs. Within each control step, the operating frequency is updated according to the product of the maximum allowable change slope of the switching frequency and the control period. The corresponding single-step frequency shift is: The operating frequency increases linearly from 242kHz to 300kHz in this step size. During the continuous frequency increase, the turn-off advance angle of the secondary synchronous rectifier diodes SR1 and SR2 is updated according to the real-time switching frequency, drive propagation delay time, and body diode reverse recovery time margin. As the real-time switching frequency increases from 242kHz to 300kHz, the turn-off advance angle... Extending from 6.3° to 7.6°, the drive pulse conduction width clamp of the secondary-side synchronous rectifier is the difference between the reference electrical angle corresponding to the secondary-side conduction and the turn-off advance angle, i.e. ,in The synchronous rectifier is turned off 70ns before the secondary side rectified current crosses zero in reverse. At the same time, the microcontroller acquires the real-time output sampling value of the buck converter circuit after the second output charging interface Port2. Based on the target power distribution of the second output charging interface Port2 and the real-time output voltage of the bus, the target duty cycle of the buck converter circuit is calculated. Then, the duty cycle is adjusted according to the timing gradient synchronized with the operating frequency change step size to reduce the transient fluctuation of the bus voltage during the frequency shift and maintain the continuous regulated output of the charging interface that is not disconnected.
[0056] When the charging device operates under closed, high-temperature conditions for an extended period and the heat dissipation channel experiences thermal blockage, the temperature sensing array detects the temperature of the heat sink substrate. The outer surface temperature of the casing is 95℃. When the temperature reaches 95℃, the temperature difference of the second layer drops to 0℃, and the temperature rise rate of the heat sink substrate is 0℃ / s. Based on this, the microcontroller unit determines the transient thermal conductivity of the second layer. The ratio of the transient thermal conductivity of the first layer to the transient thermal conductivity of the second layer is 0W. Setting it to 0 will set the maximum permissible slope for changing the switching frequency. Locked at 0Hz / s, the frequency transition to the target switching frequency is halted, maintaining the current power distribution state. During the frequency modulation slope lock-in period, the asymmetric half-bridge converter maintains its current steady-state operating frequency and continues to output power. The control unit sends power limiting commands to each output charging interface via the protocol chip, constraining the total output power of each output port to within the threshold that the current bus voltage can bear. The subsequent step-down converter circuits of each interface continue to maintain local regulated output. After the temperature difference between the heat sink base temperature and the outer surface temperature of the casing recovers and the thermal hysteresis state is released, then... By restoring dynamic frequency conversion regulation, during the light load sudden change process caused by the disconnection of the output charging interface, the operating frequency is always constrained by the dynamic tuning window of the switching frequency and the maximum allowable change slope of the switching frequency. The secondary synchronous rectifier tube performs dynamic feedforward compensation according to the real-time turn-off advance angle. Under thermal hysteresis conditions, the maximum allowable change slope of the switching frequency returns to zero and frequency migration stops, thereby maintaining the zero-voltage turn-on state of the primary gallium nitride power switch tube, reducing the impact of capacitive hard turn-on losses, local heat accumulation in the die, and transient fluctuations in bus voltage on the operational stability and hardware lifespan of multi-interface charging equipment.
[0057] Example 4: This example combines Figures 1 to 2 This paper describes a dynamic power allocation method with multiple charging interfaces, such as... Figure 1 As shown, step S1 collects the protocol handshake messages and real-time electrical parameters of each output charging interface, calculates the target power allocation and target output voltage of each output charging interface based on the acquired protocol handshake messages and real-time electrical parameters, and determines the target switching frequency of the asymmetric half-bridge converter and the dynamic tuning window of the switching frequency to maintain inductive zero voltage turn-on based on the calculation results; step S2 collects the die package temperature of gallium nitride power switch, the substrate temperature of the heat sink and the outer surface temperature of the casing, and calculates the first transient thermal conductivity rate from the gallium nitride power switch to the heat sink and the second transient thermal conductivity rate from the heat sink to the casing based on the pre-set thermal resistance and thermal capacitance parameters of each layer. Step S3 determines the turn-on margin correction factor based on the ratio of transformer excitation energy storage to the pumping energy of parasitic capacitance at the half-bridge node of the asymmetric half-bridge converter. The ratio of the transient thermal conductivity of the first layer to the transient thermal conductivity of the second layer, the remaining junction temperature margin of the gallium nitride power switch, and the turn-on margin correction factor are multiplied to calculate the maximum allowable change slope of the switching frequency. Step S4 drives the operating frequency to smoothly transition to the target switching frequency within the dynamic tuning window of the switching frequency according to the maximum allowable change slope of the switching frequency. During the transition of the operating frequency, dynamic feedforward compensation is applied to the drive-off phase of the secondary synchronous rectifier of the asymmetric half-bridge converter according to the real-time switching cycle.
[0058] like Figure 2As shown, the power factor correction circuit provides the DC bus voltage to the asymmetric half-bridge converter. During the power conversion process, the asymmetric half-bridge converter inputs the first layer transient heat conduction rate to the three-layer solid heat conduction channel. The three-layer solid heat conduction channel transmits the temperature rise rate to the temperature sensing array. The temperature sensing array collects the node temperature and inputs it to the microcontroller unit. The output charging interface transmits the protocol handshake message to the microcontroller unit. The microcontroller unit calculates and generates control quantity based on the protocol handshake message and node temperature, outputs the target switching frequency to the asymmetric half-bridge converter to drive frequency tuning, and outputs the target duty cycle to the buck converter circuit. The asymmetric half-bridge converter outputs the real-time bus output voltage to the buck converter circuit. The buck converter circuit combines the real-time bus output voltage and adjusts the power conversion according to the target duty cycle, and provides the target allocated power to the output charging interface to realize dynamic power reconfiguration of multiple interfaces.
[0059] Example 5: After the multi-interface charging device is powered on and initialized, a step power excitation of a set power is applied to the asymmetric half-bridge converter. Temperature sensor arrays are respectively deployed at the temperature measurement positions of the gallium nitride power switch die package, the center of the heat sink substrate, and the geometric surface of the shell, and the temperature of the gallium nitride power switch die package is recorded synchronously with a period of 100μs. Heat sink base temperature With the outer surface temperature of the casing The time-step response curve is used to extract the thermal equilibrium temperature difference and heating time constant from the transient response model of the dual-node thermal network of the microcontroller unit. The equivalent thermal resistance of the first layer is calculated by using the temperature response between the gallium nitride power switch die package and the heat sink. and the device's own equivalent heat capacity The equivalent thermal resistance of the second layer is calculated based on the temperature response between the heat sink and the outer casing. Equivalent heat capacity of heat sink The above parameters are stored in the internal static register; simultaneously, the control unit samples the time difference between the gate drive flipping moment of the secondary-side synchronous rectifier and the falling edge of the primary-side switching transistor through the pulse capture port to determine the drive transmission delay time. Reverse recovery time margin of the body diode Therefore, the compensation benchmark for the turn-off advance angle is determined as follows: The above calibration results serve as the basis for real-time calculation of the thermal resistance, thermal capacity parameters, and synchronous rectification turn-off advance angle of the three-layer solid heat conduction channel.
[0060] During continuous power supply and dynamic redistribution of multiple output charging interfaces, the microcontroller unit continuously monitors the temperature rise rate and steady-state thermal equilibrium point of each temperature measurement node using a sliding time window. When device aging or thermal grease degradation causes the actual thermal resistance value obtained by dividing the static temperature difference of the first layer by the static heat generation power to exceed the internally stored nominal thermal resistance of 1.8K / W by more than 15%, the microcontroller unit reads a correction factor of 0.85 from the non-volatile memory, multiplies the real-time ratio of the transient thermal conductivity rate of the first layer to the transient thermal conductivity rate of the second layer by the correction factor of 0.85 to reduce the value, and writes the reduced ratio into the working register to directly narrow the maximum allowable change slope of the switching frequency. When any charging interface is unplugged or plugged in, causing a step change in the total output power of the device within the range of 5W to 100W, the system determines the real-time switching cycle based on the real-time operating frequency and calculates the peak value of the transformer excitation current in combination with the real-time bus output voltage. Then, the transformer excitation energy storage is calculated based on this peak value. When the ratio of transformer excitation energy storage to the energy pumped by the parasitic capacitance of the half-bridge node is... When the range is between 1 and 2, the opening margin correction factor is adjusted. As this ratio decreases linearly, the microcontroller calculates the maximum permissible slope for changing the switching frequency. It is written in real time to the frequency shift step register of the pulse width modulation controller, driving the operating frequency within the switching frequency dynamic tuning window. Internal smooth migration ensures that even after device aging or changes in the performance of thermally conductive materials, the control process still updates the corresponding control quantities based on real-time temperature, transformer excitation energy storage, and the energy of pumping out parasitic capacitors at half-bridge nodes. This keeps the primary-side gallium nitride power switch on at zero voltage and enables the secondary-side synchronous rectifier to perform dynamic feedforward compensation according to the real-time turn-off advance angle.
[0061] Example 6: In the engineering calibration process of constructing the dynamic tuning window of the switching frequency of an asymmetric half-bridge converter in a multi-charging-interface fast charging device, the microcontroller unit, based on the bus input voltage of 400V, the target bus output voltage of 20V, and the transformer ratio of 10:1:1, calculates the equivalent AC load impedance of the transformer primary side, which continuously varies in the range of 5W to 100W. By combining the loop inductance of 15μH, the DC blocking and resonant capacitor of 22nF, and the excitation inductance of 90μH, the normalized gain equation of the asymmetric half-bridge is solved to determine the target switching frequency corresponding to each power point. To calibrate the safe frequency boundary for maintaining the inductive zero-voltage turn-on of the primary-side gallium nitride power switch, the microcontroller unit uses a dead time... Under constraints, solve for the total parasitic capacitance of the half-bridge node completely extracted by the excitation current within the dead time. The critical state for storing charge, and its critical charge-discharge relationship, are as follows: Based on this critical state, the lower limit of the inductive operating region frequency is extracted respectively. and the upper limit of the frequency of the sensitive working area The extracted frequency boundaries are verified by combining the asymmetric half-bridge gain characteristics and the light-load voltage regulation requirements, thereby generating a dynamic tuning window for the switching frequency covering a load gradient from 5W to 100W. The frequency limit of the inductive operating region, 185kHz, corresponds to the critical input impedance state of the asymmetric half-bridge converter circuit under a rated 100W full-load condition. When the operating frequency is reduced to 185kHz, the equivalent input impedance angle of the asymmetric half-bridge decreases to the critical positive phase angle boundary of 0°. When the operating frequency is further reduced, the circuit enters the capacitive impedance region. At the same time, within the 100ns dead time, the excitation current calculated based on the excitation inductance and dead time can provide the amount of charge required to complete the charging and discharging of the total parasitic capacitance of the 240pF half-bridge node. This critical state, together with the input impedance boundary of the asymmetric half-bridge converter circuit, is used to determine the lower limit of the inductive operating region frequency of 185kHz.
[0062] When a power redistribution event is triggered at the output charging interface, the microcontroller unit reads the current operating frequency and the target switching frequency in real time and calculates the absolute difference in frequency to be adjusted. Based on the ratio of the transient thermal conductivity of the first layer to the transient thermal conductivity of the second layer and opening margin correction factor Calculate the maximum permissible change slope of the switching frequency corresponding to the current time step. Then, the total transition time for smooth frequency modulation is calculated from the absolute difference of the frequency to be adjusted. Subsequently, the total transition time of smooth frequency modulation is discretized with a control period of 100 μs as the step size. Interpolation control intervals: Within each interpolation control interval, the operating frequency is adjusted according to a single-step frequency shift increment. Smoothly increasing or decreasing: While the operating frequency is updated in steps, the microcontroller unit simultaneously acquires the real-time output voltage and output current of the independent synchronous buck converter circuits following the first output charging interface Port1, the second output charging interface Port2, and the third output charging interface Port3. For each output charging interface, the target output voltage corresponding to the current power allocation state is first determined based on the target allocated power and the real-time output current. Combined with the real-time output voltage of the bus Calculate the target duty cycle of each buck converter circuit: The microcontroller updates the duty cycle registers of each buck converter circuit driver chip according to a timing gradient synchronized with the operating frequency change step size. During the timing gradient update process, the dynamic change of the bus voltage in the next 100μs control cycle is estimated based on the current single-step frequency shift increment. The target duty cycle of each buck converter circuit is decomposed into discrete increments or discrete decrements corresponding to the frequency shift step size. The pulse width register of the buck converter circuit driver chip is synchronously triggered by a microsecond-level clock interrupt, so that the duty cycle adjustment of the subsequent buck converter circuit is synchronized with the operating frequency step of the preceding stage. During the voltage change of the preceding asymmetric half-bridge converter circuit, the equivalent input impedance of the subsequent stage is synchronously adjusted to offset the transient charge and discharge charge difference on the bus filter capacitor, reduce the port voltage overshoot and drop during the power redistribution transition, and enable each output charging interface to complete the power supply state reconstruction according to the updated target power allocation.
[0063] The above description is only a few preferred embodiments of the present invention and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present invention is not limited to the technical solutions formed by a specific combination of the above-mentioned technical features, but should also cover other technical solutions formed by any combination of the above-mentioned technical features or their equivalent features without departing from the above-mentioned inventive concept. For example, technical solutions formed by equivalent substitution of the above-mentioned features with technical features with similar functions disclosed in the embodiments of the present invention.
Claims
1. A dynamic power allocation method with multiple charging interfaces, characterized in that, Includes the following steps: Step S1: Collect the protocol handshake messages and real-time electrical parameters of each output charging interface, calculate the target power allocation of each output charging interface and the target output voltage of the bus, and determine the target switching frequency of the asymmetric half-bridge converter and the dynamic tuning window of the switching frequency to maintain inductive zero voltage turn-on. Step S2: Collect the package temperature of the gallium nitride power switch die, the temperature of the heat sink substrate, and the outer surface temperature of the outer shell. Combine the preset thermal resistance and thermal capacity parameters of each layer to calculate the transient thermal conduction rate of the first layer from the gallium nitride power switch to the heat sink and the transient thermal conduction rate of the second layer from the heat sink to the outer shell. Step S3: Determine the turn-on margin correction factor based on the ratio of transformer excitation energy storage to parasitic capacitance pumping energy of the half-bridge node in the asymmetric half-bridge converter. Multiply the ratio of the first-layer transient thermal conductivity to the second-layer transient thermal conductivity, the remaining junction temperature margin of the gallium nitride power switch, and the turn-on margin correction factor to calculate the maximum allowable change slope of the switching frequency. Step S4: Drive the operating frequency to smoothly transition to the target switching frequency within the dynamic tuning window of the switching frequency according to the maximum allowable change slope of the switching frequency, and apply dynamic feedforward compensation to the drive-off phase of the secondary synchronous rectifier of the asymmetric half-bridge converter according to the real-time switching cycle during the transition.
2. The dynamic power allocation method with multiple charging interfaces according to claim 1, characterized in that, Step S1, determining the target switching frequency of the asymmetric half-bridge converter and the dynamic tuning window for maintaining inductive zero-voltage turn-on, includes the following sub-steps: Step S11, calculate the target DC voltage gain based on the target output voltage of the bus and the bus input voltage output by the power factor correction circuit, and calculate the normalized gain equation of the asymmetric half-bridge by combining the equivalent AC load impedance of the transformer primary side corresponding to the target power allocation and the asymmetric duty cycle, and obtain the target switching frequency; Step S12, calculate the critical state of the transformer excitation current extracting the parasitic capacitance charge of the half-bridge node during the dead time at the moment the primary side switch is turned off, extract the lower and upper limits of the inductive operating frequency, and form the dynamic tuning window for the switching frequency.
3. The dynamic power allocation method with multiple charging interfaces according to claim 1, characterized in that, Step S2 includes the following sub-steps: Step S21, calculate the first layer temperature difference between the gallium nitride power switch die package temperature and the heat sink substrate temperature, divide the first layer temperature difference by the first layer equivalent thermal resistance from the device to the heat sink to obtain the first static heat flow component, multiply the device's own equivalent heat capacity by the temperature rise rate of the gallium nitride power switch die package temperature to obtain the first transient heat storage component, and determine the first layer transient thermal conductivity by the sum of the first static heat flow component and the first transient heat storage component. Step S22: Calculate the second layer temperature difference between the heat sink base temperature and the outer surface temperature of the outer casing. Divide the second layer temperature difference by the second layer equivalent thermal resistance from the heat sink to the outer casing to obtain the second static heat flow component. Multiply the heat sink equivalent heat capacity by the temperature rise rate of the heat sink base temperature to obtain the second transient heat storage component. The sum of the second static heat flow component and the second transient heat storage component is determined as the second layer transient thermal conductivity.
4. The dynamic power allocation method with multiple charging interfaces according to claim 1, characterized in that, In step S3, the maximum allowable slope for changing the switching frequency is calculated to satisfy: ,in, The slope is adjusted to the switching frequency reference. This refers to the maximum allowable operating junction temperature of gallium nitride power switches. The real-time acquisition of the gallium nitride power switch die package temperature. To preset the baseline temperature difference margin, The transient thermal conductivity of the first layer, For the transient thermal conductivity of the second layer, It is a positive bias constant. This is the opening margin correction factor.
5. The dynamic power allocation method with multiple charging interfaces according to claim 4, characterized in that, Step S3, determining the turn-on margin correction factor, includes the following sub-steps: Step S31, calculate the peak value of the transformer excitation current based on the real-time switching cycle and the current asymmetric conduction duty cycle, and calculate the inductive reactive power stored in the transformer excitation inductance at the moment of turn-off; Step S32, calculate the minimum pumping energy required for the parasitic capacitance of the half-bridge node to complete charging and discharging; Step S33, calculate the ratio of the inductive reactive power to the minimum pumping energy; Step S34, in response to the ratio being greater than the preset energy margin threshold, configure the turn-on margin correction factor to the preset upper limit saturation value. In response to the ratio not being greater than a preset energy margin threshold, the turn-on margin correction factor is linearly reduced to a preset lower cutoff value as the ratio decreases.
6. The dynamic power allocation method with multiple charging interfaces according to claim 1, characterized in that, Step S4, which applies dynamic feedforward compensation to the drive-off phase of the secondary synchronous rectifier based on the real-time switching cycle, includes the following sub-steps: Step S41, during the operating frequency transition, calculate the turn-off advance angle based on the real-time switching frequency, drive propagation delay time, and body diode reverse recovery time margin; Step S42, clamp the drive pulse conduction width of the secondary synchronous rectifier to the difference electrical angle obtained by subtracting the turn-off advance angle from the reference electrical angle corresponding to the secondary conduction, and drive the secondary synchronous rectifier to turn off in advance before the secondary current crosses zero in the reverse direction.
7. The dynamic power allocation method with multiple charging interfaces according to claim 1, characterized in that, The method also includes: in response to detecting that the temperature difference between the heat sink base temperature and the outer surface temperature of the housing decreases to 0°C and the temperature rise rate of the heat sink base temperature is 0°C / s, determining that the transient thermal conductivity of the second layer is 0W; locking the maximum allowable change slope of the switching frequency to 0Hz / s, and stopping the jump adjustment of the operating frequency to the target switching frequency, maintaining the current power distribution state.
8. The dynamic power allocation method with multiple charging interfaces according to claim 4, characterized in that, The method also includes: in response to a sudden drop in total power demand caused by disconnection of any output charging interface, shifting the target switching frequency to a higher frequency end; reducing the turn-on margin correction factor based on the reduced excitation energy storage to pump energy ratio to narrow the maximum allowable change slope of the switching frequency; and linearly increasing the driving operating frequency to the target high-frequency operating point by a preset step size according to the narrowed maximum allowable change slope of the switching frequency.
9. The dynamic power allocation method with multiple charging interfaces according to claim 1, characterized in that, Step S4, which drives the operating frequency to transition to the target switching frequency within the switching frequency dynamic tuning window according to the maximum permissible change slope of the switching frequency, includes the following sub-steps: Step S43, calculate the absolute difference between the real-time operating frequency and the target switching frequency to be adjusted; Step S44, divide the absolute difference between the frequency to be adjusted by the maximum permissible change slope of the switching frequency to calculate the total transition time of smooth frequency modulation. Step S45: During the total transition time of smooth frequency modulation, divide multiple interpolation control intervals with a preset control cycle as the step size. Within each control step size, update the operating frequency according to the product of the maximum allowable change slope of the switching frequency and the control cycle until the operating frequency reaches the target switching frequency.
10. The dynamic power allocation method with multiple charging interfaces according to claim 1, characterized in that, The method also includes: during the transition from the operating frequency to the target switching frequency, simultaneously acquiring the real-time output sampling values of the step-down converter circuits after each output charging interface; calculating the target duty cycle of each step-down converter circuit based on the target allocated power of each output charging interface and the real-time output voltage of the bus; adjusting the duty cycle of each step-down converter circuit according to a timing gradient synchronized with the operating frequency change step size, smoothing transient fluctuations in the bus voltage, and completing the power redistribution of multiple interfaces.
Citation Information
Patent Citations
Dynamic adjustment-based llc intelligent charger
CN113394861A
