Targeted crust temperature cycle wind-cooled power cycle parameter optimization method

Abstract

The application discloses a forced air-cooled power cycle parameter optimization method for target junction and shell temperature cycle, comprising the following steps: obtaining forced air-cooled power cycle platform parameters; setting target temperature cycle indexes; establishing a mapping relationship between heating current and power loss; establishing a periodic power input model; establishing a power module junction temperature / shell temperature joint thermal response model; establishing a forced air-cooled radiator dynamic model; calculating actual junction temperature / shell temperature cycle output; constructing a target temperature cycle reachable domain; establishing a multi-objective optimization function; solving the platform parameters by using a multi-objective particle swarm optimization algorithm; and selecting a platform parameter scheme according to actual requirements and setting the platform parameters. The forced air-cooled power cycle parameter optimization method for target junction and shell temperature cycle improves power cycle platform parameter design efficiency, improves target temperature cycle achievement accuracy, can judge whether the target temperature cycle is reachable before testing, improves platform operation safety, takes into account test efficiency and platform burden, and has a wide range of applications.

Description

Technical Field

[0001] This invention belongs to the field of air-cooled platform parameter optimization technology, and relates to an air-cooled power cycle parameter optimization method oriented towards target junction temperature cycle. Background Technology

[0002] Power semiconductor modules are core components in power electronic systems such as new energy power generation, electric vehicles, rail transportation, industrial frequency converters, energy storage converters, and flexible DC transmission. During actual operation, power modules typically withstand cyclic current and thermal loads, and their junction temperature, case temperature, and internal temperature distribution continuously change with operating conditions. To assess the reliability of power modules under thermal cycling conditions, a power cycling test platform is usually used to periodically heat and cool the power modules, creating controlled temperature cycles.

[0003] In power cycling tests, commonly used temperature cycling metrics include maximum junction temperature, minimum junction temperature, junction temperature swing, maximum case temperature, minimum case temperature, case temperature swing, turn-on time, turn-off time, and cycle time. Junction temperature reflects the thermal response of the power chip, while case temperature reflects the thermal response of the module bottom, copper substrate, or heat dissipation interface. For different testing requirements, researchers may need to set different upper and lower limits for junction temperature, case temperature, and temperature swing. Therefore, the key function of a power cycling platform is not only to achieve heating and cooling, but also to stably generate the target junction / case temperature cycles under safe constraints.

[0004] Existing power cycling platforms typically employ water cooling, liquid cooling, or forced air cooling for heat dissipation. Water-cooled and liquid-cooled platforms offer strong cooling capabilities, but their complex structures require coolant, pumps, piping, heat exchangers, flow control units, and temperature control systems, resulting in higher costs and more complex maintenance. Forced air-cooled power cycling platforms, on the other hand, offer advantages such as simple structure, lower cost, convenient maintenance, and flexible deployment, making them suitable for building laboratory power cycling test platforms, verifying modular prototypes, and conducting multi-module comparative tests.

[0005] However, the target temperature cycling capability of a forced air-cooled power cycling platform is influenced by multiple factors, including heating current, power device conduction losses, turn-on time, turn-off time, fan speed, heat sink heat capacity, effective heat transfer area of ​​the heat sink, heat sink-to-air thermal resistance, and ambient temperature. Increasing the heating current can improve heat input, but it will increase the maximum junction temperature and maximum case temperature, leading to overheating risks; extending the turn-on time can increase the upper temperature limit, but it will increase the cycle time; extending the turn-off time can lower the minimum temperature, but it will reduce test efficiency; increasing the fan speed can enhance cooling capacity, but it may cause the maximum temperature to fail to reach the target; increasing the heat sink size can increase the heat transfer area, but it will also increase the heat sink heat capacity, which may weaken the cycle temperature swing.

[0006] Currently, parameter settings for forced air-cooled power cycling platforms largely rely on experience-based debugging. Experimenters typically select the heatsink and fan first, then adjust the heating current, on-time, off-time, and fan speed through multiple experiments to gradually bring the junction or case temperature curves closer to the target. This approach has the following drawbacks: First, it's difficult to determine before the experiment whether the target junction / case temperature cycle can be achieved by the current platform; second, when the target temperature cycle cannot be achieved, it's difficult to determine whether the limiting factor is insufficient current capability, excessive heat capacity of the heatsink, excessive or insufficient air cooling, or an unreasonable switching time setting; third, it's difficult to simultaneously consider the target temperature cycle error, cycle period, safety margin, power supply load, fan load, and heatsink size; and fourth, it involves numerous debugging attempts, resulting in low efficiency in platform parameter design.

[0007] Therefore, there is a need to provide a forced air-cooled power cycle platform parameter optimization method oriented towards target junction / shell temperature cycling, which can be used to establish a platform thermal response model before the test, determine whether the target temperature cycle is achievable, and automatically solve for the platform parameter combination that satisfies the target temperature cycle and safety constraints. Summary of the Invention

[0008] The purpose of this invention is to provide a method for optimizing air-cooled power cycle parameters for target junction temperature cycling, which solves the problems in the prior art such as forced air-cooled power cycle platform relying on experience-based debugging, difficulty in judging the achievability of target temperature cycling, low parameter design efficiency, and difficulty in balancing temperature cycling accuracy, cycle efficiency, safety margin, and platform load.

[0009] The technical solution adopted in this invention is a method for optimizing air-cooled power cycle parameters for target junction temperature cycling, comprising: Step 1: Obtain the parameters of the forced air-cooled power circulation platform; Step 2: Set the target temperature cycle parameters; Step 3: Establish the mapping relationship between heating current and power loss; Step 4: Establish a periodic power input model; Step 5: Establish a joint thermal response model for the junction temperature and case temperature of the power module; Step 6: Establish a dynamic model of the forced air-cooled radiator; Step 7: Calculate the actual junction temperature / case temperature cycle output; Step 8: Construct the target temperature cyclic reachable domain; Step 9: Establish a multi-objective optimization function; Step 10: Solve for the platform parameters using a multi-objective particle swarm optimization algorithm; Step 11: Select a platform parameter scheme and set the platform parameters according to actual needs.

[0010] The invention is further characterized by: Step 1 includes: Forced air-cooled power cycling platform parameters including power module parameters, constant current power supply current range, conduction time range, turn-off time range, fan speed range, heat sink thermal parameters, heat sink geometric parameters, air-cooling channel parameters, and ambient temperature; power module parameters including module rated voltage and rated current, conduction voltage drop characteristics, junction-to-case transient thermal resistance parameters, and case-to-heat sink equivalent thermal resistance parameters. Step two includes: target temperature cycle parameters including target maximum junction temperature, target minimum junction temperature, target junction temperature swing, target maximum case temperature, target minimum case temperature and target case temperature swing.

[0011] Step three includes: The power loss during the power module's conduction phase is expressed as: , in, To reduce power loss, For heating current, For the junction temperature, This refers to the saturation voltage drop of power devices. Step four includes: The power cycle periodic power input model is as follows: , in, For input power loss, For power loss during the conduction phase, This refers to the residual power during the turn-off phase. For conduction time, For shutdown time, It is a cycle period.

[0012] Step five includes: The combined thermal response model of the power module junction temperature and case temperature is expressed as follows: , , in, For the junction temperature, Shell temperature, For input power loss, This is the equivalent temperature of the radiator. For the transient thermal impedance of the junction to the shell, The equivalent thermal resistance from the shell to the heat sink, symbol This represents the convolution operation.

[0013] Step six includes: The dynamic model of the forced air-cooled radiator is as follows: , in, The equivalent heat capacity of the radiator. This is the equivalent temperature of the radiator. For input power loss, The equivalent thermal resistance from the heatsink to the air. The ambient temperature; With the heatsink and airflow conditions fixed, the relationship between the equivalent thermal resistance of the heatsink to the air and the fan speed is expressed as follows: ,or , in, The equivalent thermal resistance from the heatsink to the air. This refers to fan speed or PWM duty cycle. , , These are parameters obtained through experimental calibration; When the radiator width is used as an optimization variable, the equivalent heat capacity of the radiator is expressed as: , in, The equivalent heat capacity of the radiator. For the width of the heatsink, For the effective heat capacity coefficient, The density of the radiator material, The specific heat capacity of the heat sink material. This refers to the cross-sectional area of ​​the heat sink. The equivalent thermal resistance from the heatsink to the air is expressed as: , in, The equivalent thermal resistance from the heatsink to the air. For the width of the heatsink, This refers to fan speed or PWM duty cycle. The convective heat transfer coefficient is related to the radiator width and fan speed. The effective heat exchange area is related to the width of the radiator.

[0014] Step seven includes: Based on the model established in steps three through six, calculate the junction temperature, case temperature, and radiator temperature under steady-state conditions during the power cycle, and extract the actual temperature cycle output: , in, The output vector is the actual temperature cycle. Design parameter combinations for the platform; This is the highest junction temperature within one cycle. The lowest junction temperature within one cycle; The junction temperature swing within one cycle and ; The highest shell temperature within one cycle; The lowest shell temperature within one cycle; The shell temperature swing within one cycle and ; Platform design parameter combination , in, For the width of the heatsink, For heating current, For conduction time, For shutdown time, This refers to either the fan speed or the PWM duty cycle.

[0015] Step eight includes: constructing the target temperature cycle reachability domain and determining whether the target temperature cycle is reachable. If the target temperature cycle error meets the set requirements, the target temperature cycle is reachable; if the target temperature cycle error does not meet the set requirements, the target temperature cycle is not reachable. Construct the target temperature cyclic reachable domain based on the platform design parameter combination and model parameter set: , in, The target temperature cycling reachable range; This is the highest junction temperature within one cycle. The lowest junction temperature within one cycle; This refers to the junction temperature swing within one cycle. The highest shell temperature within one cycle; The lowest shell temperature within one cycle; This refers to the shell temperature fluctuation within one cycle. Design parameter combinations for the platform; For the set of model parameters; Model parameter set , in, It is the current-loss mapping function; The transient thermal resistance from junction to shell; The equivalent transient thermal resistance from the shell to the heat sink; This is the equivalent heat capacity of the radiator. The equivalent thermal resistance from the radiator to the air; The ambient temperature; When the target temperature cycle adopts the precise target form, the target temperature cycle error is: , in, For the target temperature cycle error, These are the weighting coefficients. The first in the actual temperature cycle output One indicator, The first of the target temperature cycle index One target value; When the target temperature cycle adopts the threshold target form, the target temperature cycle error is: , in, For the target temperature cycle error, and These are the weighting coefficients. For junction temperature swing amplitude, For shell temperature fluctuation, This is the required value for junction temperature swing. This is the required value for shell temperature swing.

[0016] Step nine includes: establishing a multi-objective optimization function with target temperature cycling error, cycle time, safety risk, and platform load as optimization objectives. , in, For multi-objective optimization, the objective function vector, This is the first functional sub-objective corresponding to the target temperature cyclic error. The second sub-objective function corresponding to the periodic time. The third sub-objective of the function corresponding to security risks, The fourth sub-objective of the corresponding platform's workload; First function sub-objective ,in, The target temperature cycling error; Second function sub-objective ,in, For conduction time, This refers to the shutdown time; Third function sub-objective ,in, The highest junction temperature within one cycle. To the upper limit of junction temperature safety, The highest shell temperature within one cycle. This is the upper limit of the shell temperature safety range; Fourth function sub-objective ,in, , , These are the weighting coefficients. For heating current, For the maximum allowable heating current, This refers to fan speed or PWM duty cycle. For the maximum allowable fan speed or PWM duty cycle, For the width of the heatsink, This represents the maximum permissible width of the heatsink.

[0017] Step ten includes: Step 10.1: Initialize the particle swarm optimization and external non-dominated solution archive; Step 10.2: Perform electrothermal response calculations for the platform parameter combinations corresponding to each particle, and calculate the corresponding actual temperature cyclic output according to steps three to seven. Step 10.3: Calculate the multi-objective function value for each particle according to Step 9; Step 10.4: Update the external non-dominated solution file based on the non-dominated relationship; Step 10.5: Select a global guided solution from the external non-dominated solution archive based on the crowding distance or mesh density; Step 10.6: Update particle velocity and position; Step 10.7: Repeat steps 10.2 to 10.6 until the iteration termination condition is met; Step 10.8: Output the optimal combination of Pareto platform parameters.

[0018] Step eleven includes: Based on actual needs, different preferred platform parameter schemes are selected from the Pareto optimal platform parameter combinations, and the parameters of the selected schemes are set on the forced air-cooled power cycling platform for power cycling tests. The platform parameter scheme should include at least heating current, on-time, off-time, and fan speed.

[0019] The beneficial effects of this invention are: Improve the efficiency of power cycling platform parameter design: This invention can automatically solve the platform parameters according to the target junction temperature / case temperature cycling requirements, reducing the number of manual adjustments. Improving the accuracy of achieving the target temperature cycle: This invention uses a junction temperature / shell temperature joint thermal response model and a multi-objective optimization algorithm to make the temperature cycle output by the platform closer to the target requirements; The invention can determine whether the target temperature cycle is achievable before the test: it determines whether there is a parameter combination that meets the target requirements within the current platform parameter range by using the target temperature cycle achievable domain. To improve platform operational safety, this invention incorporates the maximum junction temperature and maximum case temperature safety limits into the optimization objectives or constraints, thereby avoiding the risk of overheating due to parameter selection. Balancing experimental efficiency and platform load: This invention considers cycle time, power load, fan load and heat sink size during the optimization process, making the parameter scheme more suitable for actual platform applications; Wide range of applications: This invention is applicable to the optimization of operating parameters under fixed heat sinks and air-cooling channels, and also to the optimization of design parameters of new platforms involving heat sink width. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the method flow in an embodiment of the present invention; Figure 2 This is a schematic diagram of the forced air-cooled power circulation platform structure in an embodiment of the present invention; Figure 3 This is a schematic diagram of the combined thermal network model of the power module junction temperature, case temperature, and heat sink in an embodiment of the present invention; Figure 4 This is a schematic diagram illustrating the process of constructing the target temperature cyclic reachable domain in an embodiment of the present invention; Figure 5 This is a schematic diagram of the multi-objective particle swarm optimization algorithm in an embodiment of the present invention; Figure 6 This is a schematic diagram illustrating the optimization results of operating parameters under fixed cooling hardware in an embodiment of the present invention; Figure 7 This is a schematic diagram illustrating the results of the coordinated optimization of heat sink width and operating parameters in an embodiment of the present invention. Detailed Implementation

[0021] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0022] Example 1 This embodiment proposes a method for optimizing air-cooled power cycle parameters for target junction temperature cycling, including: Step 1: Obtain the parameters of the forced air-cooled power circulation platform; Step 2: Set the target temperature cycle parameters; Step 3: Establish the mapping relationship between heating current and power loss; Step 4: Establish a periodic power input model; Step 5: Establish a joint thermal response model for the junction temperature and case temperature of the power module; Step 6: Establish a dynamic model of the forced air-cooled radiator; Step 7: Calculate the actual junction temperature / case temperature cycle output; Step 8: Construct the target temperature cyclic reachable domain; Step 9: Establish a multi-objective optimization function; Step 10: Solve for the platform parameters using a multi-objective particle swarm optimization algorithm; Step 11: Select a platform parameter scheme and set the platform parameters according to actual needs.

[0023] Example 2 Based on Example 1, this example proposes that the forced air-cooled power cycling platform parameters in step 1 include power module parameters, constant current power supply current range, conduction time range, turn-off time range, fan speed range, radiator thermal parameters, radiator geometric parameters, air-cooling channel parameters, and ambient temperature. Power module parameters include module rated voltage and rated current, on-state voltage drop characteristics, junction-to-case transient thermal impedance parameters, and case-to-heat sink equivalent thermal impedance parameters. The constant current power supply range includes both the minimum heating current and the maximum heating current; The conduction time range includes the minimum conduction time and the maximum conduction time; The shutdown time range includes the minimum shutdown time and the maximum shutdown time; The fan speed range includes the minimum fan speed and the maximum fan speed, or the minimum PWM duty cycle and the maximum PWM duty cycle; When the size of the heat sink is fixed, the heat sink thermal parameters include the heat sink equivalent heat capacity, the heat sink to air equivalent thermal resistance, and the heat dissipation parameters related to fan speed. When the radiator width is used as the parameter to be optimized, the radiator geometric parameters include the radiator cross-sectional area, effective heat transfer area per unit width, material density, material specific heat capacity, effective heat capacity coefficient, and the range of values ​​for the radiator width.

[0024] Example 3 Based on Example 1, this example proposes a second step including: setting the target temperature cycling index according to the power cycling test requirements; The target temperature cycling parameters include the target maximum junction temperature, the target minimum junction temperature, the target junction temperature swing, the target maximum case temperature, the target minimum case temperature, and the target case temperature swing. When the target temperature cycle is used in a precise target form, the platform output temperature cycle should be as close as possible to the target temperature cycle specification; the target temperature cycle specification is expressed as: , in, To achieve the target maximum junction temperature, To achieve the target minimum junction temperature, For the target junction temperature swing, To achieve the target maximum shell temperature, To achieve the target minimum shell temperature, The target shell temperature swing amplitude; When the target temperature cycle adopts the threshold target form, the platform output temperature swing must not be lower than the given requirement value: , , in, For junction temperature swing amplitude, For shell temperature fluctuation, This is the required value for junction temperature swing. This is the required value for shell temperature swing.

[0025] Example 4 Based on Example 1, this example proposes step three, which includes: Since the variable directly set by the forced air cooling power cycle platform is the heating current, and the temperature rise of the power module is determined by the power loss, a mapping relationship between the heating current and the power loss is established, and the power loss of the power module during the conduction phase is calculated based on the input heating current. When the power cycling platform can measure the on-state voltage drop of the power module in real time, it calculates the instantaneous power loss based on the real-time current and on-state voltage drop: , in, This refers to the instantaneous power loss of the power module. For real-time current, For the on-state voltage drop of the power module; The power loss during the power module's conduction phase is expressed as: , in, To reduce power loss, For heating current, For the junction temperature, This refers to the saturation voltage drop of power devices. The saturation pressure drop is obtained through experimental calibration and can be expressed as: , in, For power device saturation voltage drop, For heating current, For the junction temperature, , , , represents the fitting coefficient.

[0026] Example 5 Based on Example 1, this example proposes step four, which includes: establishing a power cycle periodic power input model based on the on-time and off-time; The power cycle periodic power input model is as follows: , in, For input power loss, For power loss during the conduction phase, This refers to the residual power during the turn-off phase. For conduction time, For shutdown time, It is a cycle period, and If the power module is not turned on during the shutdown phase: .

[0027] Example 6 Based on Example 1, this example proposes step five, which includes: The thermal response relationship between the power module junction temperature, case temperature, and heat sink equivalent temperature is established. The joint thermal response model of the power module junction temperature / case temperature is expressed as follows: , , in, For the junction temperature, Shell temperature, For input power loss, This is the equivalent temperature of the radiator. For the transient thermal impedance of the junction to the shell, The equivalent thermal resistance from the shell to the heat sink, symbol This represents the convolution operation; The junction-to-case transient thermal impedance and the case-to-heatsink equivalent thermal impedance are represented using the Foster thermal network method: , , in, For the junction-to-shell thermal resistance first thermal resistance, For the junction-to-shell thermal resistance first First thermal time constant, The equivalent thermal resistance from the shell to the heat sink is the first thermal resistance, The equivalent thermal resistance from the shell to the heat sink is the first The thermal time constant.

[0028] Example 7 Based on Example 1, this example proposes step six, which includes: The dynamic model of the forced air-cooled radiator is used to describe the heat storage and heat dissipation process of the radiator under cyclic power input. The change of radiator temperature over time is calculated based on fan speed, radiator heat capacity and equivalent thermal resistance from radiator to air. The dynamic model of the forced air-cooled radiator is as follows: , in, The equivalent heat capacity of the radiator. This is the equivalent temperature of the radiator. For input power loss, The equivalent thermal resistance from the heatsink to the air. Ambient temperature; Excessive temperature rise of radiator ,but Radiator thermal time constant ,in, This is the equivalent temperature of the radiator. For ambient temperature, The equivalent thermal resistance from the heatsink to the air. For excessive heat rise of the radiator, The equivalent heat capacity of the radiator. Input power loss; Under the condition of a fixed heat sink and air cooling channel, the relationship between the equivalent thermal resistance of the heat sink to the air and the fan speed is expressed as follows: , , in, The equivalent thermal resistance from the heatsink to the air. This refers to fan speed or PWM duty cycle. , , These are parameters obtained through experimental calibration; When the radiator width is used as an optimization variable, the equivalent heat capacity of the radiator is expressed as: , in, The equivalent heat capacity of the radiator. For the width of the heatsink, For the effective heat capacity coefficient, Density of the radiator material The specific heat capacity of the heat sink material. This refers to the cross-sectional area of ​​the heat sink. The equivalent thermal resistance from the heatsink to the air is expressed as: , in, The equivalent thermal resistance from the heatsink to the air. For the width of the heatsink, This refers to fan speed or PWM duty cycle. The convective heat transfer coefficient is related to the radiator width and fan speed. The effective heat exchange area is related to the width of the radiator.

[0029] Example 8 Based on Example 1, this example proposes step seven, which includes: Based on the model established in steps three through six, calculate the junction temperature, case temperature, and radiator temperature under steady-state conditions during the power cycle, and extract the actual temperature cycle output: , in, The output vector is the actual temperature cycle. Design parameter combinations for the platform; This is the highest junction temperature within one cycle. The lowest junction temperature within one cycle; The junction temperature swing within one cycle and ; The highest shell temperature within one cycle; The lowest shell temperature within one cycle; The shell temperature swing within one cycle and ; Platform design parameter combination , in, For the width of the heatsink, For heating current, For conduction time, For shutdown time, This refers to either the fan speed or the PWM duty cycle.

[0030] Example 9 Based on Example 1, this example proposes step eight, which includes: constructing the target temperature cycle reachability domain and determining whether the target temperature cycle is reachable. If the target temperature cycle error meets the set requirements, the target temperature cycle is reachable; if the target temperature cycle error does not meet the set requirements, the target temperature cycle is not reachable. The target temperature cycle reachability domain is used to determine whether a set of platform design parameters exists under given current range, conduction time range, turn-off time range, fan speed range, heatsink parameters, and safe temperature constraints. This ensures that the actual temperature cycle output meets the target temperature cycle requirements. Construct the target temperature cyclic reachable domain based on the platform design parameter combination and model parameter set: , in, The target temperature cycling reachable range; This is the highest junction temperature within one cycle. The lowest junction temperature within one cycle; This refers to the junction temperature swing within one cycle. The highest shell temperature within one cycle; The lowest shell temperature within one cycle; This refers to the shell temperature fluctuation within one cycle. Design parameter combinations for the platform; For the set of model parameters; Model parameter set , in, It is the current-loss mapping function; The transient thermal resistance from junction to shell; The equivalent transient thermal resistance from the shell to the heat sink; This is the equivalent heat capacity of the radiator. The equivalent thermal resistance from the radiator to the air; The ambient temperature; When the target temperature cycle adopts the precise target form, the target temperature cycle error is: , in, For the target temperature cycle error, These are the weighting coefficients. The first in the actual temperature cycle output One indicator, The first of the target temperature cycle index One target value; When the target temperature cycle adopts the threshold target form, the target temperature cycle error is: , in, For the target temperature cycle error, and These are the weighting coefficients. For junction temperature swing amplitude, For shell temperature fluctuation, This is the required value for junction temperature swing. This is the required value for shell temperature swing.

[0031] Example 10 Based on Example 1, this example proposes step nine, which includes: establishing a multi-objective optimization function with target temperature cycling error, cycle time, safety risk, and platform burden as optimization objectives. , in, For multi-objective optimization, the objective function vector, This is the first functional sub-objective corresponding to the target temperature cyclic error. The second sub-objective function corresponding to the periodic time. The third sub-objective of the function corresponding to security risks, The fourth sub-objective of the corresponding platform's workload; First function sub-objective ,in, The target temperature cycling error; Second function sub-objective ,in, For conduction time, This refers to the shutdown time; Third function sub-objective ,in, The highest junction temperature within one cycle. To the upper limit of junction temperature safety, The highest shell temperature within one cycle. This is the upper limit of the shell temperature safety range; Fourth function sub-objective ,in, , , These are the weighting coefficients. For heating current, For the maximum allowable heating current, This refers to fan speed or PWM duty cycle. For the maximum allowable fan speed or PWM duty cycle, For the width of the heatsink, This represents the maximum permissible width of the heatsink.

[0032] Example 11 Based on Example 1, this example proposes step ten, which includes: The platform parameter combination is solved using a multi-objective particle swarm optimization algorithm, where each particle represents a set of candidate platform parameters: , in, Design parameter combinations for the i-th platform group. Let be the width of the i-th heatsink group. Let i be the heating current of the i-th group. Let i be the conduction time of the i-th group. Let i be the shutdown time of the i-th group. For the i-th fan speed or PWM duty cycle; Step 10.1: Initialize the particle swarm optimization and external non-dominated solution archive; Step 10.2: Perform electrothermal response calculations for the platform parameter combinations corresponding to each particle, and calculate the corresponding actual temperature cyclic output according to steps three to seven. Step 10.3: Calculate the multi-objective function value for each particle according to Step 9; Step 10.4: Update the external non-dominated solution file based on the non-dominated relationship; Step 10.5: Select a global guided solution from the external non-dominated solution archive based on the crowding distance or mesh density; Step 10.6: Update particle velocity and position; The formulas for updating particle velocity and position are: , , in, Let be the velocity vector of the i-th particle in the (k+1)-th iteration. For inertia rights; Let be the velocity vector of the i-th particle in the k-th iteration; For individual learning factors; A uniformly distributed random number between 0 and 1; Let i be the optimal position of the i-th particle in the k-th generation; Let be the position vector of the i-th particle in the k-th iteration; As a global learning factor; A uniformly distributed random number between 0 and 1; This is the global guiding solution for the i-th particle in the k-th generation; Let be the position vector of the i-th particle in the (k+1)-th generation iteration; Let be the current position vector of the i-th particle in the k-th iteration; This is the particle velocity vector obtained from the (k+1)th generation iteration calculation; Step 10.7: Repeat steps 10.2 to 10.6 until the iteration termination condition is met; Step 10.8: Output the optimal combination of Pareto platform parameters; The platform parameter combination includes one or more of the following: heating current, turn-on time, turn-off time, fan speed, and heatsink width.

[0033] Example 12 Based on Example 1, this example proposes step eleven, which includes: Based on actual needs, select different preferred platform parameter schemes from the Pareto optimal platform parameter combinations, including fast cycle priority scheme, temperature matching priority scheme, safety margin priority scheme, and low platform load scheme; The heating current, conduction time, turn-off time, and fan speed of the selected scheme are set in the forced air-cooled power cycle platform. When the radiator width is involved in the optimization, the corresponding radiator is processed or selected according to the output radiator width, and the power cycle test is carried out according to the optimized operating parameters.

[0034] This invention is used to optimize the solution of parameters such as heating current, turn-on time, turn-off time, fan speed, and heat sink width under given constraints of power module, power supply current range, fan speed range, heat sink structure, and air cooling channel, based on the target junction temperature cycle and target case temperature cycle requirements, thereby improving the parameter design efficiency and target temperature cycle achievement accuracy of the forced air cooling power cycle platform.

[0035] This invention simultaneously considers the highest junction temperature, lowest junction temperature, junction temperature swing, highest case temperature, lowest case temperature, and case temperature swing, ensuring that power cycling platform parameter optimization is not limited to a single temperature index. This invention unifies constant current heating, current-related losses, module transient thermal impedance, and radiator air-cooling boundaries into a single model framework, enabling rapid prediction from platform parameters to temperature cycling output. This invention constructs a target temperature cycling reachable domain, determining before testing whether the target junction / case temperature cycling can be achieved with the current hardware and parameter range. This invention simultaneously considers target temperature cycling error, cycle time, safety risks, and platform load, and can output multiple Pareto optimal parameter schemes. This invention is applicable to operating parameter optimization under existing radiator and airflow conditions, as well as hardware-operating parameter co-optimization involving radiator width. The parameters output by this invention include heating current, turn-on time, turn-off time, fan speed, and radiator width, directly guiding the setting of forced air-cooled power cycling platforms.

[0036] Figure 1 This is a flowchart of the overall optimization method in one embodiment of the present invention. The flowchart shows eleven core steps in the order of execution. Starting from S1, which obtains platform hardware constraint parameters and environmental parameters, the steps are as follows: S2, which sets the target temperature cycle index; S3, which establishes the current-loss mapping model; S4, which establishes the cycle power input model; S5, which establishes the junction temperature / case temperature joint thermal response model; S6, which establishes the dynamic model of the forced air-cooled radiator; S7, which extracts the actual temperature cycle output; S8, which constructs the target temperature cycle reachability domain and determines its reachability; S9, which establishes the multi-objective optimization function; S10, which uses the multi-objective particle swarm optimization algorithm to solve the Pareto parameter combination; and finally, S11, which outputs and applies the platform parameters. The flowchart also marks the constraints and model parameters corresponding to each step, including parameter range constraints such as current and conduction time, as well as model parameters such as thermal impedance, thermal capacity, and thermal resistance, clearly presenting the complete logical link of the entire optimization method from input to output.

[0037] Figure 2This is a schematic diagram of the forced air-cooled power cycling platform in one embodiment of the present invention. The diagram fully presents the hardware composition and signal and energy transfer relationships of the platform. The host computer is responsible for parameter setting and data recording, and outputs control signals to the constant current power supply and the gate drive circuit. The constant current power supply provides a large current heating input to the power module under test. The gate drive circuit controls the power module to periodically turn on and off according to the set on / off time. The power module under test is placed in the air-cooling channel, with its bottom in contact with the heat sink. The axial fan provides a forced convection cooling boundary for the air-cooling channel, forming a heat exchange between the heat sink and the air. The temperature acquisition system collects the junction temperature, case temperature, heat sink temperature and ambient temperature in real time and feeds them back to the host computer. At the same time, the diagram marks the physical quantities and model parameters corresponding to each link, including the periodic power input, temperature parameters at each stage, and parameters such as heat capacity, thermal resistance and width of the heat sink, which intuitively shows the hardware architecture and working principle of the forced air-cooled power cycling platform.

[0038] Figure 3 This is a schematic diagram of a combined thermal network model of the power module junction temperature-case temperature-heat sink in one embodiment of the present invention. It presents the complete heat transfer path from the chip junction region to the ambient air in the form of a thermal resistance-thermal capacity equivalent network. The top layer is the chip junction temperature node, which is connected to the case temperature node through a multi-level Foster thermal impedance network from the junction to the case. Each level of thermal impedance is composed of thermal resistance and thermal time constant units of the corresponding order. The case temperature node is further connected to the heat sink equivalent temperature node through a multi-level Foster thermal impedance network from the case to the heat sink. The bottom layer is a first-order thermal resistance-thermal capacity network from the heat sink to the air, which includes the heat sink equivalent thermal capacity and the heat sink equivalent thermal resistance from the air, and is finally connected to the ambient temperature node. This model clearly shows the thermal coupling relationship between the junction temperature, case temperature, and heat sink temperature, and is the core model foundation for calculating the dynamic temperature response under periodic power input.

[0039] Figure 4 This diagram illustrates the process of constructing the target temperature cycle reachability domain in one embodiment of the present invention, showcasing the complete logic from platform parameter input to reachability judgment. The left side shows the platform parameter space and model parameter set. The platform design parameter combination includes heatsink width, heating current, conduction time, turn-off time, and fan speed. The model parameter set includes the current-loss mapping function, thermal impedance at each stage, heatsink heat capacity, thermal resistance, and ambient temperature. After parameter input, the actual temperature cycle output is calculated via the electrothermal response model, thereby constructing the target temperature cycle reachability domain. Subsequently, the actual output is compared with the target temperature cycle or threshold requirement, and the safety temperature constraint is used to determine whether the target temperature cycle is reachable. If reachable, the corresponding parameter result is output; if not reachable, the limiting factors are further analyzed, providing a basis for hardware parameter adjustment. This fully presents the logic for constructing the reachability domain and the reachability judgment process.

[0040] Figure 5This is a flowchart illustrating a multi-objective particle swarm optimization algorithm in one embodiment of the present invention. It shows the complete iterative process of the algorithm from initialization to outputting the optimal solution. At the beginning of the algorithm, the target temperature cycle, variable range, and safety constraints are input. First, the position and velocity of the particle swarm are initialized, with each particle corresponding to a set of candidate platform parameters. Then, the current-loss model is called sequentially to calculate the conduction loss, the thermal response model is called to calculate the dynamic changes of junction temperature, case temperature, and radiator temperature, the temperature cycle output index is extracted, and the multi-objective function value including temperature error, cycle time, safety risk, and platform burden is calculated. Next, out-of-bounds particles are penalized or subjected to boundary projection processing according to the constraints. Then, the external non-dominated solution file is updated according to the non-dominated relationship, and a global guiding solution is selected from the file by crowding distance or grid density to update the velocity and position of the particles. After completing one iteration, it is determined whether the termination condition is met. If not, the above calculation process is repeated. If it is met, the final Pareto optimal platform parameter combination is output.

[0041] Figure 6 This diagram illustrates the optimization results of operating parameters under fixed cooling hardware in a certain embodiment of the present invention. The diagram corresponds to an application scenario where the heatsink size, air-cooling channel, and fan model are all fixed, showcasing the output characteristics of multi-objective particle swarm optimization. In this scenario, the optimization variables are heating current, on-time, off-time, and fan PWM duty cycle. The constraints are the minimum requirements for junction temperature swing and case temperature swing. The optimization objectives include temperature error, cycle time, safety risk, and platform load. The diagram marks three typical Pareto optimal schemes: the fast cycle priority scheme has the shortest cycle and the temperature is close to the safety boundary; the temperature matching priority scheme has a smaller temperature error and a comprehensive compromise across all indicators; and the safety margin priority scheme has a higher safety margin and a higher fan speed but a relatively longer cycle. This visually presents the performance trade-offs between different preferred parameter schemes under fixed hardware conditions.

[0042] Figure 7 This diagram illustrates the results of the coordinated optimization of radiator width and operating parameters in one embodiment of the present invention. It corresponds to a design scenario where the radiator width is adjustable, demonstrating the distribution of results after simultaneously incorporating the radiator width and operating parameters into the optimization. In this scenario, the radiator width, as a new optimization variable, simultaneously affects the radiator's equivalent heat capacity and effective heat transfer area, thus having a coupled impact on temperature swing, maximum temperature, and cycle time. The diagram illustrates the performance trade-offs under different radiator widths: a smaller radiator width reduces heat capacity and more easily meets temperature swing requirements, but the maximum temperature is closer to the safety limit; a larger radiator width reduces the maximum temperature and increases the safety margin, but increases the platform size and prolongs the cycle time. Co-optimization can achieve a balance between radiator size, operating parameters, temperature cycle indicators, and safety constraints, outputting optimal parameter schemes with different preferences.

[0043] In one embodiment of the present invention, the operating parameters under fixed cooling hardware are optimized. The forced air-cooled power circulation platform includes an IGBT power module, a constant current heating power supply, a gate drive circuit, a fixed heat sink, an axial fan, an air-cooling channel, and a temperature acquisition system. The heat sink and air-cooling channel are already determined, and the fan speed can be adjusted via PWM duty cycle.

[0044] The platform parameter constraints are as follows: ; ; ; ; The target temperature cycling requirement is: ; ; The safe temperature constraint is: ; ; Based on experimental calibration, the current-loss mapping relationship was obtained. Then, the junction-to-case thermal resistance, the case-to-heat sink thermal resistance, and the heat sink-to-air thermal resistance at different fan speeds were identified. Subsequently, the heating current, conduction time, turn-off time, and fan PWM duty cycle were used as optimization variables, and the Pareto optimal parameter combination was solved using a multi-objective particle swarm optimization algorithm. The representative optimized schemes are shown in Table 1.

[0045] Table 1 Pareto parameter scheme under fixed cooling hardware

[0046] As shown in Table 1, with the heat sink and air cooling channel fixed, the present invention can output multiple sets of parameter schemes that meet the target junction temperature / case temperature cycling and safety constraints. The fast cycling priority, temperature matching priority, or safety margin priority schemes can be selected according to actual needs.

[0047] In one embodiment of the present invention, the radiator width and operating parameters are synergistically optimized. The power module, power supply range, fan model, and radiator cross-sectional shape are known, and the radiator width is... Undetermined; The width constraint of the heat sink is: ; The operating parameters are constrained as follows: ; ; ; ; The target temperature cycling requirement is: ; ; Based on the radiator cross-sectional parameters, the relationship between radiator width and heat capacity, effective heat transfer area and radiator-to-air thermal resistance is established. The radiator width, heating current, conduction time, turn-off time and fan PWM duty cycle are simultaneously used as optimization variables. The platform parameters are solved using a multi-objective particle swarm optimization algorithm. The representative optimized schemes are shown in Table 2.

[0048] Table 2. Co-optimization scheme of radiator width and operating parameters

[0049] As shown in Table 2, a larger heatsink width is not necessarily better. A smaller width helps reduce heat capacity and enhance temperature swing, but the maximum temperature approaches the safety limit. A larger width helps reduce the maximum temperature, but increases the platform size and extends the cycle time. This invention achieves comprehensive optimization between heatsink width, operating parameters, target temperature cycle, and safety constraints.

[0050] In one embodiment of the present invention, the achievability of a target temperature cycle is determined by setting target junction temperature swing and case temperature swing requirements, and searching for feasible solutions within a given current range, fan speed range, and switching time range. When a certain combination of parameters exists such that: , , , If the target temperature cycle is reached, the corresponding parameter combination is output. If no parameter combination satisfies the above conditions, the target temperature cycle is determined to be unreachable, and limiting factors are analyzed based on the optimization results. If all candidate solutions exhibit the following characteristics... and If the output is insufficient, it indicates that the current platform has insufficient heat input, the heat sink has excessive heat capacity, or the cycle time range is insufficient. If all candidate solutions show... or Exceeding the limits indicates insufficient cooling capacity or an excessively high current setting. This assessment provides a basis for adjusting the heatsink, fan, power supply capacity, or switching time range.

Claims

1. A method for optimizing air-cooled power cycle parameters for target junction temperature cycling, characterized in that, include: Step 1: Obtain the parameters of the forced air-cooled power circulation platform; Step 2: Set the target temperature cycle parameters; Step 3: Establish the mapping relationship between heating current and power loss; Step 4: Establish a periodic power input model; Step 5: Establish a joint thermal response model for the junction temperature and case temperature of the power module; Step 6: Establish a dynamic model of the forced air-cooled radiator; Step 7: Calculate the actual junction temperature / case temperature cycle output; Step 8: Construct the target temperature cyclic reachable domain; Step 9: Establish a multi-objective optimization function; Step 10: Solve for the platform parameters using a multi-objective particle swarm optimization algorithm; Step 11: Select a platform parameter scheme and set the platform parameters according to actual needs.

2. The method for optimizing air-cooled power cycle parameters for target junction temperature cycling according to claim 1, characterized in that, Step one includes: the forced air-cooled power cycling platform parameters include power module parameters, constant current power supply current range, conduction time range, turn-off time range, fan speed range, heat sink thermal parameters, heat sink geometric parameters, air-cooling channel parameters, and ambient temperature; the power module parameters include module rated voltage and rated current, conduction voltage drop characteristics, junction-to-case transient thermal impedance parameters, and case-to-heat sink equivalent thermal impedance parameters. Step two includes: target temperature cycle parameters including target maximum junction temperature, target minimum junction temperature, target junction temperature swing, target maximum case temperature, target minimum case temperature and target case temperature swing.

3. The method for optimizing air-cooled power cycle parameters for target junction temperature cycling according to claim 1, characterized in that, Step three includes: The power loss during the power module's conduction phase is expressed as: ， in, To reduce power loss, For heating current, For the junction temperature, This refers to the saturation voltage drop of power devices. Step four includes: The power cycle periodic power input model is as follows: ， in, For input power loss, For power loss during the conduction phase, This refers to the residual power during the turn-off phase. For conduction time, For shutdown time, It is a cycle period.

4. The method for optimizing air-cooled power cycle parameters for target junction temperature cycling according to claim 1, characterized in that, Step five includes: The combined thermal response model of the power module junction temperature and case temperature is expressed as follows: ， ， in, For the junction temperature, Shell temperature, For input power loss, This is the equivalent temperature of the radiator. For the transient thermal impedance of the junction to the shell, The equivalent thermal resistance from the shell to the heat sink, symbol This represents the convolution operation.

5. The method for optimizing air-cooled power cycle parameters for target junction temperature cycling according to claim 1, characterized in that, Step six includes: The dynamic model of the forced air-cooled radiator is as follows: ， in, The equivalent heat capacity of the radiator. This is the equivalent temperature of the radiator. For input power loss, The equivalent thermal resistance from the heatsink to the air. Ambient temperature; With the heatsink and airflow conditions fixed, the relationship between the equivalent thermal resistance of the heatsink to the air and the fan speed is expressed as follows: ,or , in, The equivalent thermal resistance from the heatsink to the air. This refers to fan speed or PWM duty cycle. , , These are parameters obtained through experimental calibration; When the radiator width is used as an optimization variable, the equivalent heat capacity of the radiator is expressed as: ， in, The equivalent heat capacity of the radiator. For the width of the heatsink, For the effective heat capacity coefficient, Density of the radiator material The specific heat capacity of the heat sink material. This refers to the cross-sectional area of ​​the heat sink. The equivalent thermal resistance from the heatsink to the air is expressed as: ， in, The equivalent thermal resistance from the heatsink to the air. For the width of the heatsink, This refers to fan speed or PWM duty cycle. The convective heat transfer coefficient is related to the radiator width and fan speed. The effective heat exchange area is related to the width of the radiator.

6. The method for optimizing air-cooled power cycle parameters for target junction temperature cycling according to claim 1, characterized in that, Step seven includes: Based on the model established in steps three through six, calculate the junction temperature, case temperature, and radiator temperature under steady-state conditions during the power cycle, and extract the actual temperature cycle output: ， in, The output vector is the actual temperature cycle. Design parameter combinations for the platform; This is the highest junction temperature within one cycle. The lowest junction temperature within one cycle; The junction temperature swing within one cycle and ; The highest shell temperature within one cycle; The lowest shell temperature within one cycle; The shell temperature swing within one cycle and ; Platform design parameter combination , in, For the width of the heatsink, For heating current, For conduction time, For shutdown time, This refers to either the fan speed or the PWM duty cycle.

7. The method for optimizing air-cooled power cycle parameters for target junction temperature cycling according to claim 1, characterized in that, Step eight includes: constructing the target temperature cycle reachability domain and determining whether the target temperature cycle is reachable. If the target temperature cycle error meets the set requirements, the target temperature cycle is reachable; if the target temperature cycle error does not meet the set requirements, the target temperature cycle is not reachable. Construct the target temperature cyclic reachable domain based on the platform design parameter combination and model parameter set: ， in, The target temperature cycling reachable range; This is the highest junction temperature within one cycle. The lowest junction temperature within one cycle; This refers to the junction temperature swing within one cycle. The highest shell temperature within one cycle; The lowest shell temperature within one cycle; This refers to the shell temperature fluctuation within one cycle. Design parameter combinations for the platform; For the set of model parameters; Model parameter set , in, It is the current-loss mapping function; The transient thermal resistance from junction to shell; The equivalent transient thermal resistance from the shell to the heat sink; This is the equivalent heat capacity of the radiator. The equivalent thermal resistance from the radiator to the air; Ambient temperature; When the target temperature cycle adopts the precise target form, the target temperature cycle error is: ， in, For the target temperature cycle error, These are the weighting coefficients. The first in the actual temperature cycle output One indicator, The first of the target temperature cycle index One target value; When the target temperature cycle adopts the threshold target form, the target temperature cycle error is: ， in, For the target temperature cycle error, and These are the weighting coefficients. For junction temperature swing amplitude, For shell temperature fluctuation, This is the required value for junction temperature swing. This is the required value for shell temperature swing.

8. The method for optimizing air-cooled power cycle parameters for target junction temperature cycling according to claim 1, characterized in that, Step nine includes: establishing a multi-objective optimization function with target temperature cycling error, cycle time, safety risk, and platform load as optimization objectives. ， in, For multi-objective optimization, the objective function vector, This is the first functional sub-objective corresponding to the target temperature cyclic error. The second sub-objective function corresponding to the periodic time. The third sub-objective of the function corresponding to security risks, The fourth sub-objective of the corresponding platform's workload; First function sub-objective ,in, The target temperature cycling error; Second function sub-objective ,in, For conduction time, This refers to the shutdown time; Third function sub-objective ,in, The highest junction temperature within one cycle. To the upper limit of junction temperature safety, The highest shell temperature within one cycle. This is the upper limit of the shell temperature safety range; Fourth function sub-objective ,in, , , These are the weighting coefficients. For heating current, For the maximum allowable heating current, This refers to fan speed or PWM duty cycle. For the maximum allowable fan speed or PWM duty cycle, For the width of the heatsink, This represents the maximum permissible width of the heatsink.

9. The method for optimizing air-cooled power cycle parameters for target junction temperature cycling according to claim 1, characterized in that, Step ten includes: Step 10.1: Initialize the particle swarm optimization and external non-dominated solution archive; Step 10.2: Perform electrothermal response calculations for the platform parameter combination corresponding to each particle, and calculate the corresponding actual temperature cyclic output according to Steps 3 to 7. Step 10.3: Calculate the multi-objective function value for each particle according to Step 9; Step 10.4: Update the external non-dominated solution file based on the non-dominated relationship; Step 10.5: Select a global guided solution from the external non-dominated solution archive based on the crowding distance or mesh density; Step 10.6: Update particle velocity and position; Step 10.7: Repeat steps 10.2 to 10.6 until the iteration termination condition is met; Step 10.8: Output the optimal combination of Pareto platform parameters.

10. The method for optimizing air-cooled power cycle parameters for target junction temperature cycling according to claim 1, characterized in that, Step eleven includes: Based on actual needs, different preferred platform parameter schemes are selected from the Pareto optimal platform parameter combinations, and the parameters of the selected schemes are set on the forced air-cooled power cycling platform for power cycling tests. The platform parameter scheme should include at least heating current, on-time, off-time, and fan speed.

Images