Disclosure of Invention
The invention provides a multi-mode power management method and device for a gallium nitride adapter, which solve the technical problems of low heat management efficiency and unstable power control of the miniaturized gallium nitride adapter in the related technology.
The invention provides a multimode power management method and device of a gallium nitride adapter, comprising the following steps:
dividing the interior of the adapter into a plurality of cooling priority areas according to the element heat sensitivity and integrating the three-dimensional layered micro-channel structure;
According to the division of the cooling priority areas and the three-dimensional layered micro-channel structure, a sensor network system is constructed, and the temperature of each area and the state of a cooling system are monitored in real time;
based on the monitoring data of the temperature of each area and the state of the cooling system, constructing a self-adaptive flow distribution algorithm, and dynamically adjusting the flow distribution of the cooling liquid according to the real-time temperature states of the divided cooling priority areas;
Calculating the maximum allowable power under the current cooling capacity according to the monitoring data of the temperature of each area and the state of the cooling system and the cooling liquid flow distribution result, and dynamically adjusting the power parameters;
And dynamically switching the working mode of the adapter according to the result of calculating the maximum allowable power under the current cooling capacity, the dynamically adjusted power parameter, the load state and the temperature condition.
Further, the dividing into a plurality of cooling priority areas includes:
Dividing the interior of the adapter into a high priority cooling zone, a medium priority cooling zone and a low priority cooling zone;
the high-priority cooling area comprises a GaN power device and a driving IC, the medium-priority cooling area comprises a transformer and a high-power resistor, and the low-priority cooling area comprises a filter capacitor and an interface circuit.
Further, the integrated three-dimensional hierarchical micro-channel structure comprises:
The method comprises the steps that microchannels with different densities and trends are configured for different priority areas, a high-density serpentine microchannel is adopted in a high priority area, a medium-density linear microchannel is adopted in a medium priority area, and a low-density radiation type microchannel is adopted in a low priority area;
dielectric liquids having high thermal conductivity, low electrical conductivity and high dielectric strength characteristics are selected as cooling medium.
Further, the flow distribution of the self-adaptive flow distribution algorithm is to distribute the total flow of the cooling liquid according to the proportion of the cooling demand index of each area to the total demand;
The cooling demand index of the divided cooling priority areas is calculated according to the difference value between the current temperature and the reference temperature of the divided cooling priority areas and the weight coefficient of the areas, the weight coefficient of the high priority area is 1.5 to 2.0, the weight coefficient of the medium priority area is 0.8 to 1.2, and the weight coefficient of the low priority area is 0.3 to 0.6.
Further, the maximum allowable power under the current cooling capacity is calculated, the power upper limit is determined according to the thermal resistance value measured in real time and the target maximum temperature, and the maximum allowable power is equal to the difference between the target maximum temperature and the environment temperature divided by the thermal resistance value.
Further, the dynamically adjusting the power parameter includes:
When the predicted temperature rising trend is obvious, the switching frequency and the driving voltage are reduced;
when the predicted temperature tends to stabilize or decrease, the switching frequency and the driving voltage are increased;
The switching frequency is adjusted in the range of 65kHz to 1MHz, and the driving voltage is adjusted in the range of 3.3V to 6.5V.
Further, the four working modes of the dynamic switching adapter include:
normal mode, temperature is within a safe range, i.eWith full power output as the primary goal;
low temperature optimization mode, i.e. when the temperature is lowOptimizing energy efficiency by increasing switching frequency;
High temperature derating mode, i.e. when the temperature approaches the thresholdThe power output is reduced, and the heat dissipation is preferentially ensured;
critical protection mode, i.e. when the temperature is close to the limitThe power is reduced rapidly, the cooling force is increased, and the system safety is ensured.
Furthermore, the working mode of the dynamic switching adapter avoids power fluctuation caused by switching between modes, and power and cooling parameters gradually change according to a set time constant in the switching process, and the method comprises the following specific steps:
Defining a state transition function, so that the power and the cooling parameters in the switching process are gradually changed according to a set time constant;
the power parameter uses a shorter time constant, and the cooling parameter uses a longer time constant;
The parameter state is updated every 5ms during the mode switching process, ensuring a smooth transition.
Further, the dynamically adjusting the power parameter further includes implementing a power management policy based on heat redistribution, and specifically includes the steps of:
forming a heat transmission path from the heat sensitive area to the heat resistant area through the topological structure of the micro-channel network;
establishing a direct connection between a microchannel outlet of the heat sensitive zone and a microchannel inlet of the heat resistant zone;
By controlling the flow distribution proportion of each micro-channel section, the directional transmission of heat is realized, the temperature of the GaN power device is reduced, the temperature of the transformer and the capacitor is increased, and more uniform temperature distribution is realized.
The invention provides a multimode power management device of a gallium nitride adapter, which comprises a memory and one or more processors, wherein executable codes are stored in the memory, and the one or more processors are used for executing the multimode power management method of the gallium nitride adapter when executing the executable codes.
The invention has the beneficial effects that through the synergistic effect of the regional differential micro-channel cooling structure and the multi-mode power management, the technical problems of the heat management and the power control of the miniaturized gallium nitride adapter are solved, and the technical effects of improving the temperature management effect, prolonging the reliability and the service life and enhancing the power output capability are obtained;
The adapter has the advantages of reducing the highest temperature of the adapter, reducing the temperature difference of hot spots, realizing high-efficiency operation in a full temperature range, reducing the thermal protection triggering rate, prolonging the service life, improving the continuous power density and the peak power due to the improvement of the heat dissipation efficiency under the condition of unchanged volume, realizing the self-adaptive adjustment of the working mode according to the load and the temperature without manual intervention, reducing the energy consumption of a cooling system through accurate temperature management and a dual-mode heat dissipation system, realizing the improvement of the efficiency in the full load range, smoothly transiting various working modes, avoiding the abrupt frequency reduction or shutdown of the traditional adapter and providing stable and continuous power supply.
Detailed Description
The subject matter described herein will now be discussed with reference to example embodiments. It is to be understood that these embodiments are merely discussed so that those skilled in the art may better understand and implement the subject matter described herein and that changes may be made in the function and arrangement of the elements discussed without departing from the scope of the disclosure herein. Various examples may omit, replace, or add various procedures or components as desired. In addition, features described in some examples may be combined in other examples as well.
In at least one embodiment of the present invention, a method for multi-mode power management of a gallium nitride adapter is disclosed, as shown in fig. 1 to 6, comprising:
Step 1, dividing the interior of an adapter into a plurality of cooling priority areas according to the element thermal sensitivity and integrating a three-dimensional layered micro-channel structure;
The method specifically comprises the following steps:
step 1.1, carrying out regional classification according to the heat sensitivity and the heating characteristic of the element;
dividing the interior of the adapter into a high-priority cooling area (comprising key elements such as GaN power devices and drive ICs), a medium-priority cooling area (comprising transformers, high-power resistors and the like) and a low-priority cooling area (comprising high-temperature resistant elements such as filter capacitors and interface circuits);
Optionally, in some embodiments, the high priority cooling region may be further subdivided into an ultra-high priority region (e.g., gaN power tube chip) and a high priority region (e.g., driver IC), respectively employing different microchannel densities and cooling strategies;
Step 1.2, integrating a three-dimensional layered micro-channel structure in the PCB;
The method comprises the steps that microchannels with different densities and trends are configured for different priority areas, a high-density serpentine microchannel is adopted in a high priority area, a medium-density linear microchannel is adopted in a medium priority area, and a low-density radiation type microchannel is adopted in a low priority area;
In some embodiments, the micro-channels can adopt a multi-layer structure, such as constructing independent micro-channel networks at different layers of the PCB, and realizing three-dimensional cooling through vertical connecting channels, or adopting micro-channels with variable cross sections, wherein the cross sections at the inlet are larger, and the cross sections at the outlet are gradually reduced so as to enhance the flow rate of the cooling liquid;
Step 1.3, selecting dielectric liquid as a cooling medium;
the dielectric liquid has the characteristics of high heat conductivity coefficient, low conductivity and high dielectric strength, and ensures the safety of use around electronic elements;
Optionally, the cooling medium can be fluorinated liquid (such as FC-72, HFE-7100, etc.) or mineral insulating oil, and is selected according to specific temperature range and insulation requirement;
Step 1.4, integrating a miniature temperature sensor array;
At least 5 miniature temperature sensors are arranged at key positions inside the adapter, and the temperature state of each area is monitored in real time.
In certain applications, such as medical device chargers or server power supplies where more accurate temperature monitoring is required, the number of temperature sensors may be increased to 8-12, or hybrid applications employing different types of sensors, such as thermocouples, thermistors, and bandgap temperature sensors, to improve measurement accuracy and reliability.
Step 2, according to the division of the cooling priority areas and the three-dimensional layered micro-channel structure, a sensor network system is constructed, and the temperature of each area and the state of a cooling system are monitored in real time;
On the basis of a region differential micro-channel cooling structure constructed by the region differential micro-channel cooling structure, a complete sensor network system is constructed, the internal temperature distribution of an adapter and the state of a cooling system are monitored in real time, and data support is provided for subsequent flow distribution and temperature prediction, and the method specifically comprises the following steps:
step 2.1, integrating miniature pressure and flow sensors;
the miniature pressure and flow sensors are respectively arranged at the inlet and the outlet of the cooling liquid and used for monitoring the pressure difference and the flow change of the cooling system;
Step 2.2, calculating cooling efficiency according to the monitoring data;
Cooling efficiencyThe calculation formula is as follows:
;
Wherein the method comprises the steps ofIndicating the cooling efficiency of the cooling device,Indicating the removal of heat and,Representing the pump power and,Represents the specific heat capacity of the cooling liquid,Indicating the mass flow of the cooling liquid,Indicating the temperature difference between the inlet and the outlet of the cooling liquid,Representing the pump driving voltage, the pump driving voltage is,Representing the pump drive current.
Step 2.3, constructing a temperature monitoring network data processing unit;
Filtering, normalizing and anomaly detecting are carried out on the collected temperature, pressure and flow data, so that the data accuracy is ensured;
step 2.4, establishing a thermal property database;
the temperature distribution characteristics under different working conditions are stored, and data support is provided for the subsequent power management strategies.
Step 3, constructing a self-adaptive flow distribution algorithm based on the monitoring data of the temperature of each area and the state of the cooling system, and dynamically adjusting the flow distribution of the cooling liquid according to the real-time temperature states of the divided cooling priority areas;
based on real-time temperature data acquired by a multi-layer distributed temperature monitoring network established by the multi-layer distributed temperature monitoring network, constructing a self-adaptive flow distribution algorithm and a temperature prediction model, and dynamically optimizing cooling resource distribution of each region, wherein the method specifically comprises the following steps of:
step 3.1, configuring a piezoelectric micropump and a micro valve array system;
the micropump is responsible for providing cooling liquid circulation power, and the micro valve array controls flow distribution of micro channels in each area;
In a high-end application scene, for example, a data center server power supply, a backup micro pump system is optionally configured, and the backup micro pump system is automatically switched when a main pump fails to ensure the reliability of a cooling system;
Step 3.2, constructing a self-adaptive flow distribution algorithm;
According to the real-time temperature state of each area, the cooling liquid distribution is dynamically regulated, and the flow distribution calculation formula is as follows:
;
Wherein the method comprises the steps ofRepresentation allocation to the firstThe flow rate of the cooling liquid in each region,Indicating the total flow rate and,AndRespectively represent the firstAnd (b)The weighting coefficients of the individual zone cooling priorities,Represent the firstThe temperature of the individual zones is set to be,The reference temperature is indicated as such,Represent the firstThe temperature of the individual zones is set to be,Representation of allThe individual regions are summed up and,Representing the total number of regions;
the implementation process of the self-adaptive flow distribution algorithm comprises the steps of firstly carrying out normalization processing on temperature data of each cooling area, calculating a difference value between the temperature of each area and a reference temperature, then calculating a cooling demand index of each area according to a preset weight coefficient (the weight coefficient of a high-priority area is 1.5-2.0, the weight coefficient of a medium-priority area is 0.8-1.2, and the weight coefficient of a low-priority area is 0.3-0.6), and finally distributing flow according to the proportion of the cooling demand index of each area to the total demand. In practical application, the algorithm effectively avoids the problem of overhigh temperature of a hot spot area by updating flow distribution once every 100 ms;
Optionally, for different use scenarios, different flow distribution strategy variants can be adopted, for example, for game notebook chargers with frequent sudden loads, a predictive factor can be added, and flow distribution can be adjusted in advance according to a historical load mode;
Step 3.3, constructing a thermal diffusion prediction model based on a finite difference method;
predicting the temperature change of 10-300ms in the adapter, wherein the discrete form of the thermal diffusion equation is as follows:
;
Wherein the method comprises the steps ofRepresenting the positionAt a time stepI.e. the temperature at the next moment; representing the positionAt a time stepI.e. the temperature at the current moment; Representing the coefficient of thermal diffusion, describing the rate at which heat diffuses in the material; representing a time step, namely the interval between two adjacent time points in calculation; representing the size of the space grid in the x direction, namely calculating the spacing of the grid in the x direction; representing the size of the space grid in the y direction, namely calculating the spacing of the grid in the y direction; representing the positionAt a time stepI.e. the temperature of the current point in the x direction of the forward adjacent point; representing the positionAt a time stepI.e. the temperature of the current point in the negative direction of the x direction; representing the positionAt a time stepI.e. the temperature of the current point in the y direction in the forward direction of the adjacent point; representing the positionAt a time stepI.e. the temperature of the current point in the negative direction of the y direction of the adjacent point; representing the positionHeat source term, i.e. heat generated per unit volume; representing the density of the material, describing the mass to volume ratio of the material; Representing the specific heat capacity of the material, describing the amount of heat required to raise the unit temperature per unit mass of material;
The thermal diffusion prediction model is concretely realized by dividing the internal space of an adapter into 10 multiplied by 3 three-dimensional grids, establishing a thermal network comprising temperature nodes, heat source nodes and boundary nodes, applying the discretization thermal diffusion equation to each node, solving a temperature field by adopting an explicit iteration method, setting the time step to be 1ms, setting the space grid size to be 1mm multiplied by 0.5mm, and calibrating the thermal diffusion coefficient of each region by recording the historical power data and the corresponding temperature response of each power element. In practical application, the model can accurately predict the temperature change trend caused by sudden load change, predict the temperature change 10-300ms in advance and provide enough response time for power parameter adjustment;
In some embodiments, the thermal diffusion prediction model can be coupled with a simplified fluid dynamics model to calculate fluid flow and heat transfer characteristics in the micro-channel, or for low-cost application scenarios with limited computational resources, a simplified centralized parameter thermal model can be adopted to divide the interior of the adapter into a small number of thermal areas, represented by a thermal resistance-thermal capacitance network, so that computational complexity is reduced;
step 3.4, the current power data is used as a heat source item to be input into a prediction model;
And carrying out temperature change trend prediction by combining past temperature data. In the present embodiment, the measured power data of the main heating element such as GaN power device, driving IC, transformer is converted into the heat generation rate per unit volume, and the heat generation rate is input as the heat source term to the prediction model, thereby realizing the prediction of the future temperature.
For example, in an application scenario where a high performance notebook computer is charged, when a user starts a graphics rendering intensive application, the power demand of the notebook computer may increase from 45W to 95W in a short period of time. The predictive model of this embodiment is able to predict that the temperature of the high priority region will rise by about 15 ℃ in the initial phase of the power demand rise (within about 50 ms), thus initiating preventive power management measures in advance, ensuring that a safe temperature range is maintained in the event of a sudden load increase.
Step 4, calculating the maximum allowable power under the current cooling capacity according to the monitoring data of the temperature of each area and the state of the cooling system and the cooling liquid flow distribution result, and dynamically adjusting the power parameters;
the intelligent flow distribution and temperature prediction unit is used for realizing the cooperative control of cooling capacity and power parameters by using temperature monitoring data and prediction results provided by the intelligent flow distribution and temperature prediction system, ensuring that the system can ensure heat dissipation and simultaneously maximizing power output capacity, and specifically comprises the following steps:
step 4.1, calculating the maximum allowable power under the current cooling capacity;
determining the upper power limit according to the thermal resistance value measured in real time and the target highest temperature, wherein the calculation formula is as follows:
;
Wherein the method comprises the steps ofIndicating the maximum allowable power to be applied,Indicating the maximum temperature allowed to be reached,Indicating the temperature of the environment and,Representing the currently measured thermal resistance;
Optionally, in an application scene with larger fluctuation of the ambient temperature (such as charging of outdoor portable equipment), an ambient temperature sensor can be integrated into an adapter, the ambient temperature change is monitored in real time, and the maximum allowable power is dynamically adjusted;
Step 4.2, dynamically adjusting the power parameters according to the temperature prediction result;
The device comprises a switching frequency and a driving voltage, wherein the frequency adjustment range is 65kHz-1MHz, and the driving voltage adjustment range is 3.3V-6.5V, and the adjustment strategy is that the switching frequency and the driving voltage are reduced when the rising trend of the predicted temperature is obvious, and the switching frequency and the driving voltage are improved to improve the performance when the predicted temperature tends to be stable or reduced;
The method comprises the steps of realizing dynamic adjustment of power parameters through a digitally controlled frequency synthesizer and a programmable drive voltage regulator, establishing a mapping relation between a temperature change rate and a power parameter adjustment amplitude, reducing the switching frequency by 5% each time and the drive voltage by 0.2V each time when a temperature prediction result shows that the temperature rise rate exceeds 1 ℃ per second in the future 100ms, increasing the switching frequency by 3% each time and the drive voltage by 0.1V each time when the temperature change rate is lower than 0.2 ℃ per second and the temperature is lower than a threshold value, and carrying out parameter adjustment once every 50ms to ensure that the system can respond to temperature change in time;
For example, in a fast charging scenario, when a high-power mobile phone (such as a flagship mobile phone supporting 65W fast charging) is connected to the adapter, the embodiment can intelligently adjust the switching frequency according to temperature prediction, namely, the switching frequency is kept high (about 800 kHz) at the initial stage of charging to improve efficiency;
Step 4.3, realizing a power management strategy based on heat redistribution;
and by actively controlling the flow direction of the cooling liquid, heat is guided to transfer from a heat sensitive area (such as a GaN power device) to a heat resistant area (such as a transformer and a capacitor), and the whole power capacity of the system is expanded.
The specific implementation mode of the heat redistribution strategy comprises the steps of forming a heat transmission path from a heat sensitive area to a heat resistant area through a topological structure of a micro-channel network, establishing direct connection between a micro-channel outlet of the heat sensitive area and a micro-channel inlet of the heat resistant area, and realizing directional transmission of heat by controlling flow distribution proportion of each micro-channel section. In practical application, the strategy reduces the highest temperature of the GaN power device by 8 ℃, and simultaneously the temperature of the transformer and the capacitor rises by 5-7 ℃, so that more uniform temperature distribution is realized as a whole.
In some embodiments, the heat redistribution strategy may be used in conjunction with dynamic adjustment of power parameters, such as balancing the partial temperatures without reducing the overall power output by increasing the power load sharing of the heat resistant region while reducing the power load of the GaN device when GaN power device temperatures are detected near the threshold but the overall heat capacity of the system remains a margin.
Step 5, dynamically switching the working mode of the adapter according to the result of calculating the maximum allowable power under the current cooling capacity, the dynamically adjusted power parameter, the load state and the temperature condition;
according to the power thermal management result executed by the multi-mode collaborative power thermal management unit, combining the load state and the temperature condition, dynamically switching the working mode of the adapter to realize the balance of energy efficiency and heat dissipation, and optimizing the system performance under different working conditions, the method specifically comprises the following steps:
Step 5.1, constructing a multistage temperature response strategy;
four modes of operation are divided:
normal mode, temperature is within a safe range, i.eWith full power output as the primary goal;
low temperature optimization mode, i.e. when the temperature is lowOptimizing energy efficiency by increasing switching frequency;
High temperature derating mode, i.e. when the temperature approaches the thresholdThe power output is reduced, and the heat dissipation is preferentially ensured;
critical protection mode, i.e. when the temperature is close to the limitThe power is reduced rapidly, the cooling force is increased, and the system safety is ensured;
Temperature (temperature)Refers to real-time monitoring of temperature of key heating elements (such as GaN power devices, driver ICs, etc.) inside the adapter, i.e., representative temperature nodes of the high-priority cooling region. The temperature is acquired in real time by a distributed temperature sensor network, reflects the actual working temperature of the part of the system most susceptible to heat, and is used as a criterion for switching the working modes.
Optionally, the temperature threshold can be adjusted according to different application scenes, for example, for scenes with high reliability requirements such as avionic device chargers, the temperature threshold of each mode can be reduced by 5-10 ℃ to provide larger safety margin, and for scenes such as server power supplies needing to maximize power output, the threshold can be increased by 3-5 ℃ on the premise of ensuring safety, and the working range is expanded;
Step 5.2, realizing a bimodal heat dissipation system in cooperation with a load state;
under the light load mode (load <30% rated power), the active cooling system is deactivated only by heat conduction and natural convection to save energy consumption;
activating active microfluid to circularly cool in a heavy load mode (the load is more than or equal to 30% of rated power), and ensuring temperature control during high power output;
In some embodiments, an intermediate transition mode (load is within 20-40% of rated power) can be added, a pulse type active cooling strategy is adopted, namely, a micro-fluid circulation system is intermittently started to balance cooling effect and energy consumption;
step 5.3, realizing a mode smooth switching algorithm;
The power fluctuation caused by the switching between modes is avoided, the power and the cooling parameters are gradually changed according to the set time constant in the switching process, and the output stability is ensured.
The mode smooth switching algorithm is concretely realized by defining a state transition function:
;
Wherein the method comprises the steps ofTime of presentationThe value of the state parameter at the time,As a value of the initial state parameter,As a value of the target state parameter,Is a time constant (typically set to 50-200 ms),Indicating a switch from the start to the current time,An exponential factor representing the decay over time,A base representing natural logarithms;
The method selects proper time constants for different parameter types, uses shorter time constants for power parameters (such as switching frequency and driving voltage), uses longer time constants for cooling parameters (such as flow rate and pump speed), and updates parameter states every 5ms in the mode switching process to ensure smooth transition. In practical application, the algorithm effectively eliminates output voltage fluctuation during mode switching, and the fluctuation range is reduced from 5-8% of the traditional method to 0.5-1%.
For example, in a portable medical device charging scenario, when the device suddenly enters a high-power consumption diagnosis mode from a standby state, the mode smooth switching algorithm of the embodiment can finish a smooth transition from a low-temperature optimization mode to a normal mode within 200ms, and simultaneously starts an active cooling system, and output voltage fluctuation is controlled within +/-0.8% in the whole process, so that stable power supply of the medical device is ensured, and diagnosis data distortion or device restarting problems caused by power supply fluctuation are avoided.
A multi-mode power management apparatus for a gallium nitride adapter includes a memory and one or more processors, the memory storing executable code, the one or more processors executing the executable code to perform a multi-mode power management method for a gallium nitride adapter as described above.
Here, the present invention provides an implementation example:
This embodiment has been applied and validated in the product development of 65W high power density gallium nitride adapters. The adapter has a volume of only 40X 35X 30mm, adopts a single-stage LLC topological structure, has an output voltage range of 5-20V and has a maximum output current of 5A. The specific application of this embodiment to this product will be described below.
The gallium nitride adapter is mainly applied to fast charging scenes of high-end notebook computers, tablet computers and smart phones. Such scenarios present significant challenges to the thermal management and power control capabilities of the adapter, particularly in extreme use environments, such as high load long term operation and high ambient temperatures. Taking a certain high-performance game notebook as an example, when the high-performance game notebook runs at full load (high-load game and rendering software are run at the same time), 60-65W of power supply is continuously required, and the working mode can last for a plurality of hours, so that the traditional adapter is extremely easy to overheat and protect, and power output is reduced or shutdown is easy to occur.
In this adapter, the specific implementation procedure of this embodiment is as follows:
the micro-channel cooling structure is implemented inside the four-layer PCB of the adapter, integrating a micro-channel structure with a total length of about 210 mm. GaN power device and its drive IC area (high priority area) adopts 0.8mm wide serpentine microchannel with density of 3.5mm2/mm2, LLC transformer and rectifier diode area (medium priority area) adopts 1.2mm wide linear microchannel with density of 2.2mm2/mm2, and input/output filter capacitor area (low priority area) adopts 1.5mm wide radial microchannel with density of 1.0mm2/mm2. The selected cooling medium is modified fluorinated liquid HFE-7000, and has a heat conductivity coefficient of 0.075W/(m.K) and a dielectric strength of 35kV/mm. Micro piezoelectric pump (size 8X 3 mm) provides maximum flow of 15ml/min, maximum pressure of 25kPa.
The temperature monitoring system realizes that 7 micro-thermistor temperature sensors are arranged inside the adapter, wherein the number of GaN power device areas is 3, the number of transformer areas is 2, and the number of control ICs and output interface areas is 1. The sampling frequency is 200Hz, and the temperature measurement precision is +/-0.5 ℃. The miniature pressure sensor and the flow sensor are respectively arranged at the inlet and the outlet of the cooling liquid to monitor the pressure difference and the flow of the system. The data were collected by a 12-bit ADC and processed by a built-in 32-bit microcontroller (STM 32F042, 48MHz main frequency).
The implementation of the adaptive flow allocation is based on real-time temperature data, the system adjusts the flow allocation once every 100 ms. For example, when the notebook computer suddenly enters a high-load working state from a standby state, the temperature rising rate of the GaN power device area reaches 2.5 ℃ per second, the flow distribution algorithm immediately adjusts the distribution proportion, the flow of the high-priority area is increased from the initial 40% to 65%, the medium-priority area is kept 35%, and the low-priority area is reduced to only 10% of the total flow. With this dynamic adjustment, the temperature rise of the GaN region is limited to within 8 ℃ instead of 15 ℃ rise when a fixed allocation scheme is employed.
The multi-mode cooperative power thermal management is realized in the adapter, the dynamic power parameter adjustment is realized, the switching frequency range is 75kHz-850kHz, and the driving voltage range is 3.8V-6.2V. Taking an actual use case as an example, when the adapter charges a game notebook and the notebook runs a game and a video derivation task at the same time, the adapter detects a high load demand, and meanwhile, the temperature of the GaN power device area rises rapidly. The system first calculates the maximum allowable power at the current cooling capacity to be 68W (when the temperature of the GaN region is 65 ℃ and the thermal resistance is 0.32 ℃ per W), then reduces the switching frequency from the initial 800kHz to 600kHz, reduces the driving voltage from 6.0V to 5.2V, simultaneously starts the active cooling of the maximum flow, stabilizes the temperature of the GaN region at about 70 ℃ and maintains the continuous output power of 60W.
The self-adaptive working mode switching is realized in different use scenes, and the system can be automatically and smoothly switched among four working modes. For example, when a user connects a cell phone for quick charging (20W load), the system initially operates in a low temperature optimized mode (temperature about 45 ℃) and the switching frequency is maintained at a higher level (about 820 kHz) to improve energy efficiency, when the user subsequently connects a notebook computer and initiates a large file download, the load rapidly rises to 50W and the system predicts that the temperature will exceed 65 ℃ in a short time, thus switching smoothly to normal mode in advance while gradually starting the active cooling system. The whole switching process lasts about 180ms, the fluctuation of the output voltage is controlled within +/-0.6%, and the stability of the equipment operation is ensured.
The temperature management effect and the power output capability are fully verified for two key technical effects of the embodiment.
The highest temperature comparisons under different load conditions at ambient temperature 25 ℃ are shown in table 1:
TABLE 1 comparison of maximum temperatures under different load conditions (ambient temperature 25 ℃)
The continuous output power comparisons at different ambient temperatures are shown in table 2:
TABLE 2 continuous output Power comparison at different ambient temperatures
From the above data, it can be seen that the present embodiment achieves improvement in temperature management, average reduction of about 21% in maximum temperature under various load conditions, and continuous output power improvement in power output capability, especially under high ambient temperature conditions, up to 47.4% -62.5%, which fully verifies the technical effects of the present embodiment.
While the embodiments of the present invention have been described above, the embodiments are not limited to the above-described embodiments, which are intended to be illustrative only and not limiting, and many equivalents thereof may be made by those of ordinary skill in the art in light of the present disclosure, which fall within the scope of the embodiments.