Mathematical model-based vehicle-mounted charger component parameter optimization method and system

Abstract

The invention relates to the technical field of electric energy storage systems, in particular to a mathematical model-based vehicle-mounted charger component parameter optimization method and system. The method comprises the following steps: performing multi-physical field joint modeling according to a vehicle-mounted charger switching device to obtain a vehicle-mounted charger component simulation model; performing dynamic sensitivity analysis according to the vehicle-mounted charger component simulation model to obtain dynamic sensitivity matrix data; performing layered MINLP solution on the dynamic sensitivity matrix data to obtain battery impedance change curve data; and on-line parameter reconstruction is carried out according to the battery impedance change curve data to obtain vehicle-mounted charger component parameter optimization data. The adaptive capacity of the system to different working conditions is enhanced while the charging efficiency is improved, and the overall performance, stability and reliability of the vehicle-mounted charger are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric energy storage systems, and in particular to a method and system for optimizing vehicle charger component parameters based on a mathematical model. Background Art

[0002] With the rapid development of the electric vehicle (EV) and hybrid electric vehicle (HEV) markets, on-board chargers, as a core component of electric vehicle charging systems, have gradually become the key to improving the energy efficiency and safety of electric vehicles. The main function of the on-board charger is to convert the external alternating current (AC) into the direct current (DC) required by the on-board battery and charge the battery. In order to meet the needs of efficient charging in different environments, the design of the on-board charger must have high efficiency, good thermal management, and small size and weight.

[0003] At present, the thermal management and electrical characteristics of the switching devices (such as MOSFET, diodes, etc.) and PCB (printed circuit board) used in the car charger are still facing challenges. The traditional design method focuses on the electrical performance of the device and ignores its coupling effect with thermal behavior. This leads to large heat loss and poor heat dissipation in the car charger system under high power density and high temperature environment, which in turn affects the working stability and service life of the system. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a method and system for optimizing vehicle charger component parameters based on a mathematical model to solve at least one of the above technical problems.

[0005] The present application provides a method for optimizing parameters of vehicle charger components based on a mathematical model, the method comprising: Based on the on-board charger switch device, multi-physics field joint modeling is performed to obtain the on-board charger component simulation model; Perform dynamic sensitivity analysis based on the on-board charger component simulation model to obtain dynamic sensitivity matrix data; Perform hierarchical MINLP on the dynamic sensitivity matrix data to obtain the battery impedance change curve data; Online parameter reconstruction is performed based on the battery impedance change curve data to obtain the parameter optimization data of the on-board charger components.

[0006] In the present invention, the electrical, thermal, magnetic and structural characteristics are highly integrated through multi-physics field joint modeling, which can accurately describe the performance of the on-board charger switch device in a complex working environment. Through dynamic sensitivity analysis, the influence of the key device parameters of the on-board charger (such as MOSFET on-resistance, switching frequency, inductance value, etc.) on the system performance (such as charging efficiency, heat loss, battery health status, etc.) can be quantitatively evaluated. The mixed integer nonlinear programming (MINLP) method is used to optimize and solve the dynamic sensitivity matrix, realizing the coordinated optimization of discrete variables (such as the selection of the switching device type) and continuous variables (such as inductance value and switching frequency). Through the hierarchical solution strategy, the global optimality of the optimization solution can be ensured while ensuring the computational efficiency. Unlike traditional single-objective optimization, this method combines the battery impedance change curve to achieve refined control of the battery charging process, reduce overcharge and loss during the charging process, improve charging efficiency and extend battery life. Through online parameter reconstruction, the charger operating parameters (such as PWM control strategy, switching frequency, inductance value, etc.) can be dynamically adjusted according to the real-time changes in the battery impedance to ensure that the system can adapt to different battery health states, temperature environments and load conditions.

[0007] Optionally, the multi-physics field joint modeling is performed according to the on-board charger switch device to obtain the on-board charger component simulation model, including: Collecting basic data of the on-board charger for the on-board charger switch device to obtain basic data of the on-board charger, wherein the basic data of the on-board charger includes electrical data of the on-board charger and thermal data of the on-board charger; The electrical thermal model of the device is constructed according to the electrical data and thermal data of the on-board charger to obtain the electrical thermal model of the device; Obtain the physical parameter data of the on-board charger PCB and build a multi-physics field model based on the physical parameter data; The overall simulation model of the on-board charger is obtained by performing simulation calculations based on the multi-physics field model and the device electrical and thermal model.

[0008] The present invention collects electrical and thermal data of the on-board charger switch components, combines multi-physics field modeling and electrical thermal model simulation calculation, and effectively realizes the accurate prediction of the overall performance of the on-board charger. By building a highly integrated electrical-thermal simulation model, it is possible to identify and optimize electrical losses, thermal management issues, and component coupling effects in advance during the design phase, thereby significantly improving the energy efficiency and reliability of the charger. Based on a comprehensive analysis method of dynamic simulation and multi-physics field interaction, it is possible to quickly evaluate the thermal stability and electrical performance of the charger under different working conditions, providing strong data support for the design of efficient and low-loss on-board chargers.

[0009] Optionally, performing dynamic sensitivity analysis according to the vehicle charger component simulation model to obtain dynamic sensitivity matrix data includes: Perform sensitivity analysis based on the on-board charger component simulation model to obtain dynamic sensitivity data; Thermal network modeling is performed based on the vehicle charger component simulation model to obtain dynamic temperature response data; The constraint matrix is ​​constructed according to the dynamic sensitivity data and the dynamic temperature response data to obtain complete dynamic constraint data.

[0010] In the present invention, based on the simulation model, the influence of input parameters (such as electrical properties, thermal properties, working environment, etc. of components) on system output (such as power loss, heat change, performance indicators, etc.) can be evaluated through sensitivity analysis. Using the thermal network modeling method, a temperature response model of the charger under different working conditions can be established. The thermal network will consider all heat sources (such as power loss, switching loss, etc.) and heat conduction paths (such as PCB, packaging, etc.). By modeling the thermal behavior of the on-board charger, the temperature distribution and thermal response data under different working conditions are obtained, which is helpful for analyzing the thermal stability of the charger. Dynamic sensitivity data can help quantify the impact of different parameters on the performance of the charger, while dynamic temperature response data characterizes the impact of heat transfer on the performance of components. By combining these two types of data, a matrix containing electrical and thermal property constraints can be established to describe the performance boundary and feasible solution range of the system under different operating conditions.

[0011] Optionally, performing hierarchical MINLP solution on the dynamic sensitivity matrix data to obtain battery impedance change curve data includes: Perform top-level MINLP on the dynamic sensitivity matrix data to obtain top-level device combination data; Perform the underlying MINLP solution on the dynamic sensitivity matrix data to obtain the continuous optimization data; Generate time-varying constraints according to dynamic sensitivity matrix data to obtain time-varying constraint data; The top-level device combination data and the continuous optimization data are solved according to the time-varying constraint data to obtain the battery impedance change curve data.

[0012] In the present invention, by performing hierarchical MINLP (mixed integer nonlinear programming) solution on the dynamic sensitivity matrix data, the parameters of the on-board charger components can be systematically optimized to meet the multiple requirements of battery performance and thermal management. The top-level MINLP solution can ensure the selection of the best device combination, thereby optimizing the performance of the entire system; the bottom-level MINLP solution further improves the accuracy and efficiency of the system through continuous optimization. The time-varying constraint generation and solution process provides accurate modeling for the dynamic response of the battery, ensuring that the on-board charger can finely control the impedance changes of the battery at different charging stages. The obtained battery impedance change curve data helps to improve the stability, efficiency and battery life of the charging process, and improves the reliability and performance of the charger in actual use.

[0013] Optionally, performing online parameter reconstruction according to the battery impedance change curve data to obtain vehicle charger component parameter optimization data includes: Acquire real-time electrical data, and perform optimal parameter matching based on the real-time electrical data and battery impedance change curve data to obtain primary optimization parameter data; Perform real-time accelerated calculation and parameter adjustment based on primary optimization parameter data to obtain secondary optimization parameter data; Continuous feedback adjustment is performed based on the secondary optimization parameter data to obtain the vehicle charger component parameter optimization data.

[0014] In the present invention, by performing online parameter reconstruction based on the battery impedance change curve data, the working state of the vehicle charger components can be optimized in real time. The acquisition of real-time electrical data and the matching of battery impedance change curve data ensure the optimal performance of the charger under various working conditions. The primary optimization parameter data is further adjusted through real-time accelerated calculation to improve the system's response speed and optimization accuracy; the secondary optimization is adjusted through continuous feedback to ensure that the charger maintains a high degree of stability and adaptability during actual operation.

[0015] Optionally, the obtaining of physical parameter data of the on-board charger PCB and constructing a multi-physics field model according to the physical parameter data includes: Obtaining physical parameter data of the on-board charger PCB, wherein the physical parameter data includes IV characteristic curve data, on-resistance data, BH curve data of magnetic components, and winding AC resistance; Perform temperature segment fitting according to physical parameter data to obtain electrical characteristic data; Power consumption and junction temperature mapping are performed based on physical parameter data to obtain thermal characteristic data; The electrical property model of the MOSFET device is constructed according to the electrical characteristic data to obtain the electrical property model: The thermal model is constructed according to the thermal characteristic data to obtain the thermal model: Solve the thermal network model based on the power loss data in the thermal model and electrical model to obtain the junction temperature data; The electrical property model is modified according to the junction temperature data to obtain an electrical property modification model; Perform intrinsic orthogonal decomposition on the thermal model to obtain a reduced-order model; The electrical correction model and the reduced-order model are packaged to obtain a multi-physics field model.

[0016] In the present invention, by acquiring the physical parameter data of the on-board charger PCB and constructing a multi-physics field model, the performance of the charger under different working environments can be accurately simulated and optimized. Combined with the temperature segmentation fitting and the power consumption junction temperature mapping, the electrical and thermal characteristics are effectively coupled, providing a more accurate basis for the construction of electrical and thermal models. By correcting and reducing the junction temperature data, the efficiency and accuracy of the model are ensured, making the thermal management and electrical performance of the charger more stable and reliable during long-term operation.

[0017] Optionally, the electrical characteristic data includes first electrical characteristic data, second electrical characteristic data, third electrical characteristic data, fourth electrical characteristic data and fifth electrical characteristic data, and the electrical characteristic data is obtained by performing temperature segmented fitting according to the physical parameter data, including: Extract thermal characteristics based on the thermal data of the on-board charger to obtain thermal characteristic data; Perform on-state voltage extraction, saturation current extraction and switching time extraction according to IV characteristic curve data, and obtain on-state voltage data, saturation current data and switching time data respectively; Performing characteristic curve fitting on the on-voltage data, the saturation current data, and the switching time data according to the thermal characteristic data to obtain first electrical characteristic data; Performing temperature fitting on the on-resistance data according to the thermal characteristic data to obtain second electrical characteristic data; Calculate the power loss according to the on-voltage data, the saturation current data and the switching time data to obtain the power loss data; Perform switching loss fitting according to the thermal characteristic data and the power loss data to obtain third electrical characteristic data; Performing segmented exponential fitting on the BH curve data of the magnetic element according to the thermal characteristic data to obtain fourth electrical characteristic data; The fifth electrical characteristic data is obtained by fitting the temperature-frequency variation according to the thermal characteristic data and the winding AC resistance data.

[0018] In the present invention, by performing multi-dimensional fitting and optimization on the electrical characteristic data of the on-board charger, the performance changes of the device under different operating temperatures and environmental conditions can be accurately grasped. Based on the extraction and fitting of data such as IV characteristic curves, on-voltage, saturation current, switching time, etc., the electrical performance of the switching device can be reflected, especially providing accurate simulation and prediction on key indicators such as power loss and on-resistance. The introduction of thermal characteristics enables a more detailed description of the changes in power loss and switching loss, ensuring the rationality of the thermal management system. By further fitting the BH curve of the magnetic component and the winding AC resistance data, the overall electrical characteristic data is optimized, and the stability and reliability of the charger in complex environments are improved.

[0019] Optionally, performing power consumption junction temperature mapping according to the physical parameter data to obtain thermal characteristic data includes: Decompose transient power consumption based on physical parameter data and electrical characteristic data to obtain time-varying power consumption distribution data; A three-dimensional thermal impedance network model is constructed according to the physical parameter data and the time-varying power consumption distribution data to obtain a three-dimensional thermal impedance network model; Perform transient thermal field simulation based on the three-dimensional thermal impedance network model to obtain transient thermal field data; Transfer learning calibration is performed based on transient thermal field data to obtain thermal characteristic data.

[0020] The present invention introduces transient power consumption decomposition and the construction of a three-dimensional thermal impedance network model to accurately describe the thermal dynamic behavior of the on-board charger under different working conditions. The consideration of time-varying power consumption distribution makes thermal management more accurate, avoids the simplified assumptions in traditional thermal models, and improves the accuracy of thermal prediction. Through transient thermal field simulation, real-time thermal response data can be obtained, and the thermal behavior of the device under instantaneous load changes can be accurately evaluated to ensure that heat can still be effectively managed under high load. Transfer learning calibration is further adopted to correct the thermal model in combination with actual data, which further improves the reliability and accuracy of thermal characteristics.

[0021] Optionally, constructing a three-dimensional thermal impedance network model according to the physical parameter data and the time-varying power consumption distribution data to obtain the three-dimensional thermal impedance network model includes: Modeling the thermal behavior of transistors based on physical parameter data and time-varying power consumption distribution data to obtain a thermal behavior model of transistors; The thermal behavior model of the PCB layer is modeled according to the physical parameter data and the time-varying power consumption distribution data to obtain the thermal behavior model of the PCB layer; Structural thermal behavior modeling is performed according to physical parameter data and time-varying power consumption distribution data to obtain a structural thermal behavior model; According to the transistor thermal behavior model, PCB layer thermal behavior model and structural thermal behavior model, heat flow coupling simulation is performed to obtain a three-dimensional thermal impedance network model.

[0022] In the present invention, by modeling the thermal behaviors of transistors, PCB layers, and structures separately and performing heat flow coupling simulation, it is possible to achieve a detailed characterization of the thermal behaviors of each level in the on-board charger and accurately describe the heat transfer path and thermal coupling effect. Unlike traditional methods that only rely on simplified thermal network models, multi-scale thermal modeling methods can more realistically simulate the thermal coupling between different components, improving the accuracy and reliability of thermal management. Based on time-varying power consumption distribution data, the model can dynamically reflect the thermal response of the device under different working conditions, avoiding the limitations of static models.

[0023] Optionally, the present application further provides a vehicle charger component parameter optimization system based on a mathematical model, which is used to execute the vehicle charger component parameter optimization method based on a mathematical model as described above, and the vehicle charger component parameter optimization system based on a mathematical model includes: A multi-physics joint modeling module is used to perform multi-physics joint modeling based on the on-board charger switch device to obtain a simulation model of the on-board charger components; A dynamic sensitivity analysis module is used to perform dynamic sensitivity analysis based on the vehicle charger component simulation model to obtain dynamic sensitivity matrix data; A hierarchical MINLP solution module is used to perform hierarchical MINLP solution on the dynamic sensitivity matrix data to obtain the battery impedance change curve data; The online parameter reconstruction module is used to perform online parameter reconstruction according to the battery impedance change curve data to obtain the parameter optimization data of the vehicle charger components.

[0024] The purpose of the present invention is to integrate electrical, thermal and structural characteristics into a unified simulation model through multi-physics field joint modeling, so as to comprehensively and accurately describe the working behavior of on-board charger components. By performing dynamic sensitivity analysis on the simulation model of on-board charger components, the present invention can accurately identify the degree of influence of each parameter on system performance, and generate a dynamic sensitivity matrix, providing scientific and reliable basic data for the optimization step. When performing hierarchical MINLP solution, the hierarchical solution strategy effectively avoids the problem of excessive computational complexity in traditional single optimization algorithms. This optimization process can not only dynamically adjust the battery impedance change, but also provide real-time feedback on the changes in system performance. Through hierarchical optimization, the method can better adapt to complex scenarios with multiple objectives and multiple constraints, and improve the solution efficiency and system stability. By combining the battery impedance change curve data for online parameter reconstruction, the method realizes real-time dynamic optimization of on-board charger components. The optimization process can adjust the device parameters in real time as the environment and load change, ensuring the optimal performance of the charger under different working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting implementations made with reference to the following drawings: Figure 1 A flowchart showing a method for optimizing vehicle charger component parameters based on a mathematical model according to an embodiment of the present invention is provided; Figure 2 A flowchart of the steps of a multi-physics field joint modeling method according to an embodiment is shown; Figure 3 A flow chart showing the steps of a dynamic sensitivity analysis method according to an embodiment is shown; Figure 4 A flowchart of the steps of a hierarchical MINLP solution method according to an embodiment is shown; Figure 5 A flowchart of the steps of an online parameter reconstruction method according to an embodiment is shown; The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0026] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0027] In addition, the drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0028] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0029] See also Figures 1 to 5 The present application provides a method for optimizing parameters of vehicle charger components based on a mathematical model, the method comprising: S1. Perform multi-physics field joint modeling based on the on-board charger switch device to obtain a simulation model of the on-board charger components; In one embodiment, COMSOL Multiphysics, ANSYS, or other simulation software that supports multi-physics coupling is used. Clarify key parameters such as material properties, current carrying capacity, and operating frequency of switching components such as MOSFET and diodes. The electric field, thermal field, and magnetic field should be coupled for modeling. According to the structure of the charger, establish its physical geometry model, including the power input terminal, output terminal, heat dissipation structure, etc. Set electrical boundary conditions (such as voltage and current), thermal boundary conditions (such as the heat dissipation effect of the cooling system), etc. according to the actual working environment.

[0030] S2. Perform dynamic sensitivity analysis based on the vehicle charger component simulation model to obtain dynamic sensitivity matrix data; In one embodiment, the key performance indicators (KPIs) of the on-board charger are selected, such as conversion efficiency, voltage fluctuation, thermal runaway, etc. The simulation model is used to perform dynamic simulation under different working conditions (such as load changes, ambient temperature fluctuations, etc.) to obtain the response data of the charger during operation. The influence of each component (such as power switch, filter capacitor, etc.) on the target performance indicator is calculated. A partial derivative-based method or a sensitivity matrix construction method (such as global sensitivity analysis) is used. A dynamic sensitivity matrix is ​​generated, which reflects the influence of changes in different component parameters on the performance of the on-board charger.

[0031] In one embodiment, a key performance indicator (KPI) is selected: charging efficiency , junction temperature , output voltage fluctuation . Local sensitivity analysis: ,in are the component parameters, such as MOSFET switching frequency, transformer turns, The KPIs of the charger, such as charging efficiency, junction temperature, and output voltage fluctuation, are component parameters, such as MOSFET switching frequency, transformer turns, filter capacitor value, etc. Global sensitivity analysis: ,in For parameters The global sensitivity index, for, Cause The variance of the KPI expected value caused by the change, For system response, such as charging efficiency, junction temperature, output voltage fluctuation, is an input parameter, such as MOSFET switching frequency, transformer turns, filter capacitor value, etc. is the total variance of the KPI. Organize the sensitivity values ​​of all parameters to form a dynamic sensitivity matrix.

[0032] S3, performing hierarchical MINLP solution on the dynamic sensitivity matrix data to obtain the battery impedance change curve data; In one embodiment, an optimization objective function is set, which is usually system efficiency, power loss, thermal stability, etc. Technical and physical constraints are set, such as the maximum current carrying capacity of components, operating temperature limit, cost limit, etc. The optimized variables are defined, including the switching frequency of the power switch, the capacity of the filter capacitor, the structure of the heat sink, etc. Layer 1: Large-scale parameter optimization, rough optimization is performed to determine which component parameters have a greater impact on the performance, and these components are optimized first. Global optimization methods such as genetic algorithms and particle swarm optimization are used. Layer 2: Fine optimization, for key parameters, fine optimization is performed, and local optimization algorithms such as the Newton-Raphson method are used to improve the solution accuracy. Solve the above mixed integer nonlinear programming problem to obtain the optimal parameter configuration of the components. A MINLP solver (such as GAMS, XPRESS) can be used for this step, and processed by an optimization algorithm (such as branch and bound method, interior point method).

[0033] Use an equivalent circuit model or electrochemical model to model the battery. The battery model includes the internal resistance, capacitance, and inductance of the battery. Test the impedance change curve of the battery under different charging conditions through simulation or experiment. Analyze the relationship between battery impedance and charger output: Analyze the impact of the output current and voltage of the on-board charger on the battery impedance. Draw the battery impedance change curve based on the impedance change data of the battery under different charging conditions, and evaluate its correlation with the charger parameters.

[0034] In one embodiment, the sensitivity matrix data is optimized and solved to obtain the optimal battery impedance model. Set the objective function: , so that ,in To minimize, is the total power loss of the system, including MOSFET loss, transformer loss, filter loss, etc. Constraints: . Layered solution, first layer: coarse-grained optimization (genetic algorithm), second layer: fine-grained optimization (gradient descent method). Through optimization, calculate the impedance curve of the battery equivalent circuit model: ,in For batteries at different frequencies The impedance under is the static ohmic resistance (DC internal resistance) of the battery, is the dynamic resistance of the battery, which is used to describe the polarization effect, is an imaginary unit, is the angular frequency, is the time constant, describing the time scale of the battery polarization effect.

[0035] S4. Perform online parameter reconstruction according to the battery impedance change curve data to obtain vehicle charger component parameter optimization data.

[0036] In one embodiment, during the actual operation of the on-board charger, the battery impedance and the charger output characteristics are measured in real time. Based on the battery impedance change curve and the sensitivity analysis results, the control parameters of the on-board charger are adjusted in real time (for example, the switching frequency, output voltage, etc.). The real-time data is combined with the optimization model to achieve closed-loop control of the parameters, ensuring that the charger is always in the optimal working state.

[0037] In one embodiment, a high-precision ADC is used to measure the input voltage in real time. , charging current , battery terminal voltage . Perform parameter estimation: ,in is the state variable at time k+1 (parameters to be estimated, such as battery state of charge SOC, battery internal resistance, etc.), is the state transfer matrix, which describes the state change of the system from time k to time k+1. is the state variable at time k (battery status at the current time), is the control matrix, which represents the influence of the control input on the system state. is the input variable at time k, such as charging current and voltage, is the process noise (model error or measurement noise). Match the optimal switching frequency, PWM duty cycle, etc. Calculate the optimal value of MOSFET switching frequency: ,in, is the optimal switching frequency of MOSFET, is the constant term of pi, is the inductance value, is the filter capacitor value. Combined with the fuzzy control algorithm, the charging mode is automatically adjusted, such as constant current mode (CC), constant voltage mode (CV), and trickle charge mode (Trickle Charge), so as to adjust the optimization parameters in real time.

[0038] Optionally, the multi-physics field joint modeling is performed according to the on-board charger switch device to obtain the on-board charger component simulation model, including: S11, collecting basic data of the on-board charger from the on-board charger switch device to obtain basic data of the on-board charger, wherein the basic data of the on-board charger includes electrical data of the on-board charger and thermal data of the on-board charger; In one embodiment, standard test equipment (such as a power analyzer, a digital oscilloscope) is used to measure the input voltage and output current of the charger. The main parameters of the electrical data include the input voltage range (such as DC 12V-36V), the output voltage (such as 12V, 24V), the output current (such as 5A-20A), etc. Measure the switching frequency of the switching element (usually tens of kHz to hundreds of kHz) and evaluate the switching loss. Use an oscilloscope and a high-frequency probe to collect the switching waveform and analyze the switching loss, conduction loss, reverse recovery loss, etc. Measure the parameters of key electrical components, such as capacitors, inductors, internal resistance of the switch tube, quiescent current, etc.

[0039] Use a thermal imager or temperature sensor to locate the heat source in the charger and obtain the temperature distribution data inside and outside the charger. Heat sources include power switch devices, filter capacitors, transformers, etc. Measure the thermal resistance of the radiator and evaluate the heat dissipation efficiency. Use temperature sensors and fluid temperature probes to test the temperature change on the surface of the radiator and the cooling efficiency of the cooling system. Obtain the heat conduction data of the charger and establish a heat conduction model. Obtain the temperature gradient inside the charger through experiments, or use numerical simulation to simulate the thermal convection characteristics.

[0040] In one embodiment, a power analyzer is used to calculate the conversion efficiency: ,in is the conversion efficiency, is the output power, is the input power. Use an infrared thermal imager to measure the operating temperature of key components such as MOSFET, diodes, and transformers: ,in Thermal resistance is a measure of the thermal conductivity of components. is the junction temperature, the maximum temperature of the MOSFET or other components (°C), is the ambient temperature, For power loss.

[0041] S12, constructing a device electrical thermal model according to the electrical data and thermal data of the on-board charger to obtain the device electrical thermal model; In one embodiment, the electrical model includes an equivalent circuit model, an inductor and capacitor model, and a switching loss model. Equivalent circuit model: Construct an equivalent circuit model based on the electrical data of the switching device. For switching devices such as MOSFET or IGBT, parameters such as their on-resistance, switching frequency, and current carrying capacity must be included. Inductor and capacitor model: Construct an inductor and capacitor network model of the charger power supply part, mainly considering the inductor and capacitor parameters of the power supply filter part. Switching loss model: Consider the switching loss, reverse recovery loss, conduction loss, etc. of the switching tube, and construct the corresponding loss model through the collected waveform data. Obtain the loss function of the switching tube through experiments and introduce it into the circuit simulation.

[0042] The thermal model includes the thermal convection and radiation model and the component thermal model. According to the thermal data, a thermal conduction model inside the charger is established. Considering the heat transfer process between components (such as switching devices, capacitors, inductors, etc.), the Fourier heat conduction equation or a more complex non-steady-state heat conduction model is used. The thermal convection and radiation model obtains the convection thermal resistance between the charger surface and the environment, as well as the radiation heat loss through experiments. For natural convection and forced convection (such as fan cooling), separate modeling is required. Build an independent thermal model for each component (such as MOSFET, diode, transformer, etc.) to obtain the component thermal model, considering its heat generation, heat transfer and heat diffusion characteristics.

[0043] On the basis of the electrical model, the influence of temperature on electrical performance is considered, and the thermal-electrical coupling effect is constructed through a mathematical model. For example, the on-resistance of a switching device usually increases with increasing temperature, thus affecting the switching efficiency. A nonlinear model is used to describe the relationship between temperature and electrical parameters. Combined with the electrical-thermal coupling model, the thermal management strategy of the charger is designed, such as selecting appropriate radiators, fans, coolants, etc., to ensure the stability of the charger in a high temperature environment.

[0044] In one embodiment, MOSFET electrical-thermal modeling includes a model of on-resistance variation with temperature: ,in, is the temperature-dependent on-resistance, is the on-resistance at the reference temperature, is the temperature coefficient, the temperature change rate of the MOSFET on-resistance (1 / °C), is the current operating temperature of the MOSFET, is the reference temperature; switching loss model: ,in is the switching loss, the switching loss of MOSFET when working (W), is the drain-source voltage, the voltage between the drain and source of the MOSFET (V), is the drain current, MOSFET drain current (A), is the turn-on time, the time (s) for MOSFET to go from off to fully on. is the turn-off time, the time (s) from when the MOSFET is turned on to when it is completely turned off. is the switching frequency, the switching frequency of the MOSFET (Hz). Transformer electrical-thermal modeling includes BH curve measurement: ,in is the transformer core loss, the transformer core loss (W), is the loss factor, a constant related to transformer material and design, is the switching frequency, is the frequency index, reflecting the index of loss changing with frequency ( ), is the maximum magnetic flux density, the maximum magnetic flux density of the transformer core (T), is the flux density index, which reflects the index of loss changing with flux density ( ); Winding loss calculation: ,in is the AC winding resistance, the AC resistance of the transformer winding (Ω), is the DC winding resistance, the DC resistance of the transformer winding (Ω), is the frequency correlation coefficient, the frequency influence coefficient, which is related to the skin effect. is the switching frequency, the operating switching frequency of the transformer (Hz).

[0045] S13, obtaining physical parameter data of the on-board charger PCB, and constructing a multi-physics field model according to the physical parameter data; In one embodiment, the physical parameter data of the PCB is obtained, and the material properties of the PCB board are measured and obtained, such as the resistivity of the conductive material (copper), the dielectric constant of the dielectric material, etc. The layout information of each component on the PCB is collected, including geometric dimension data such as layer thickness, line width, spacing, etc. These data can be exported through CAD software. The thermal conductivity characteristics of the PCB are analyzed, and the path of the current flowing through the PCB is drawn, considering the current carrying capacity and heat dissipation capacity of the wire. Based on the physical parameters of the PCB, an electrical-thermal coupling model is established. Simulation software (such as ANSYS Maxwell, COMSOL, etc.) is used to deal with problems such as current density distribution, temperature distribution, and heat source generation. For the high-frequency current transmission part, the coupling between the electromagnetic field and the thermal field cannot be ignored. Through multi-physics field modeling, the electromagnetic effect caused by the current and the thermal effect caused by the current loss are considered.

[0046] In one embodiment, the physical parameters of the PCB material, the thickness of the copper layer (35µm, 70µm, etc.), the thermal conductivity of the FR4 dielectric layer (k=0.3W / m\K), and the resistance of the PCB copper wire are calculated: ,in is the resistance of the PCB copper wire, the resistance of the PCB copper wire (Ω), is the resistivity of copper, the resistivity of copper ( ), , is the length of the copper wire, the length of the PCB copper wire (m), is the cross-sectional area of ​​the copper wire, the cross-sectional area of ​​the copper wire ( ), calculation method: , is the copper wire width (m), is the thickness of the copper layer (m). Establish the electric field-thermal field-mechanical field coupling model: ,in is the current density, is the current density through the copper wire ( ), is the cross-sectional area of ​​the copper wire, the cross-sectional area of ​​the copper wire ( ). Heat conduction calculation: ,in is the thermal conductivity, the rate of heat transfer through the PCB material (W), is the thermal conductivity, the thermal conductivity of the material , is the temperature gradient, the spatial rate of change of temperature (K / m).

[0047] S14. Perform simulation calculations based on the multi-physics field model and the device electrical thermal model to obtain an overall simulation model of the on-board charger.

[0048] In one embodiment, a simulation tool that supports multi-physics field coupling is used, such as COMSOL Multiphysics, ANSYS, Simulink (for dynamic electrical analysis), etc. Electrical simulation is performed based on the electrical model to calculate the current, voltage and other parameters of the charger when it is working. Thermal simulation is performed using a thermal model to calculate the temperature distribution and thermal effect of the charger. Electrical-thermal coupling simulation is performed to evaluate the impact of temperature changes on electrical performance and the feedback effect of electrical losses on the temperature field. The simulation results are analyzed to evaluate the working efficiency, thermal management effect, temperature distribution of components, switching losses, etc. of the charger. According to the simulation results, the selection, layout, heat dissipation scheme, etc. of components are optimized to improve the overall performance and stability of the charger.

[0049] In one embodiment, the boundary condition is set to room temperature. . Load Setting Run the simulation and perform the following calculation steps to calculate the charger output voltage and current, calculate the power loss of MOSFET, diode, and transformer, and calculate the change of MOSFET junction temperature. , is the junction temperature change of MOSFET. Junction temperature (°C), is the power loss, MOSFET power loss (W), is the thermal resistance, is the base of natural logarithm, For time, is the thermal time constant. Calculate the electromagnetic field distribution, analyze the high-frequency current effect, calculate the heat transfer path, optimize the heat dissipation design, and based on the above calculations, obtain the overall simulation model of the on-board charger.

[0050] Optionally, performing dynamic sensitivity analysis according to the vehicle charger component simulation model to obtain dynamic sensitivity matrix data includes: S21, performing sensitivity analysis according to the vehicle charger component simulation model to obtain dynamic sensitivity data; In one embodiment, the purpose of sensitivity analysis is to evaluate the impact of different component parameters or operating conditions on the overall performance of the on-board charger. Dynamic sensitivity refers to how the performance indicators of the on-board charger (such as efficiency, output fluctuation, thermal response, etc.) change over time when the component parameters change. Select the performance indicators to be optimized or controlled, such as output voltage fluctuation, conversion efficiency, switching loss, current ripple, etc. According to the goal, the sensitivity analysis will evaluate the impact of component changes on these indicators. The parameters of the analysis object (component) are clarified, including switching frequency, on-resistance, switching loss of power tube, value of filter capacitor, flux density of transformer, etc. Based on the simulation model of the on-board charger (including electrical and thermal models), the dynamic behavior of the charger under actual working conditions is simulated. Use simulation tools (such as Simulink, ANSYS Maxwell, COMSOL, etc.) to perform time domain simulation to simulate the response of the charger under dynamic conditions such as load changes and temperature fluctuations. The impact of changes in each component parameter (such as the on-resistance of the power switch) on the performance of the charger is calculated by partial differential method. That is, for the target performance indicator f(x), the sensitivity of each component parameter is solved. Use the Monte Carlo method or the Sobol index method to perform global sensitivity analysis to evaluate the overall impact of all components and operating parameters on system performance. The impact of all parameters on performance indicators is summarized in a matrix form to obtain a dynamic sensitivity matrix. Through the above steps, the dynamic sensitivity data obtained contains the immediate response of each component parameter change to system performance. The result is a sensitivity matrix (such as a multidimensional matrix), where rows represent component parameters and columns represent performance indicators.

[0051] In one embodiment, local sensitivity analysis: ,in are the component parameters, such as MOSFET switching frequency, transformer turns, For system performance indicators, such as conversion efficiency, output voltage fluctuation, MOSFET junction temperature, Component parameters are design parameters that affect charger performance, such as switching frequency, MOSFET on-resistance, and transformer turns ratio. Global sensitivity analysis: ,in is the global sensitivity index, indicating the parameter For output variables The global impact of is the variance of the conditional expectation, is the conditional expectation, is the system output, are component parameters, is the total variance of the system output. Organize the sensitivity values ​​of all parameters to form a dynamic sensitivity matrix: ; in is the sensitivity matrix, is the sensitivity of charging efficiency to switching frequency, is the sensitivity of charging efficiency to MOSFET on-resistance, is the sensitivity of charging efficiency to transformer turns ratio, is the sensitivity of output voltage fluctuation to switching frequency, is the sensitivity of output voltage fluctuation to MOSFET on-resistance, is the sensitivity of output voltage fluctuation to transformer turns ratio, is the sensitivity of MOSFET junction temperature to switching frequency, is the sensitivity of MOSFET junction temperature to MOSFET on-resistance, is the sensitivity of the MOSFET junction temperature to the transformer turns ratio.

[0052] S22, performing thermal network modeling according to the vehicle charger component simulation model to obtain dynamic temperature response data; In one embodiment, the on-board charger generates heat during operation, which affects the temperature distribution of components and thus changes their electrical performance. A thermal model is established for each key component (such as MOSFET, filter capacitor, transformer, etc.). These components can usually be regarded as heat sources, and the heat generated by them (for example, heat generated by switching losses) is defined. For each component, its thermal resistance (indicating the heat flow conduction capability) and thermal capacitance (indicating its heat storage capability) are defined. For example, for metal materials, the thermal resistance is calculated using its thermal conductivity. A thermal network model of the on-board charger is constructed, taking into account the heat conduction path between the components. The thermal network can be modeled using an electrothermal coupling simulation tool (such as the COMSOL Heat Transfer module). Based on the power loss data in the electrical model, the heat source intensity of each component is defined. Considering the thermal coupling characteristics of the charger and the air convection cooling effect, the thermal resistance of the cooling system (such as a radiator, fan, etc.) is defined. Using the thermal network model, a time-domain thermal simulation is performed to simulate the dynamic response of the charger temperature under different working conditions (such as switching frequency changes, load fluctuations, etc.). Extract dynamic temperature data from the simulation results to observe the temperature changes of each component in different time periods and how temperature fluctuations affect the performance of the charger. The dynamic temperature response data of each component is obtained through simulation, which is usually output in the form of a time series, such as the temperature value of each component at different time points. These data are combined with the sensitivity data to establish a complete constraint matrix.

[0053] In one embodiment, a thermal resistance-capacitance network is established: , is the change of MOSFET junction temperature over time, is the power loss, is the thermal resistance, is the base of natural logarithm, For time, is the thermal time constant. The finite element method (FEM) is used to calculate the heat flow path: MOSFET junction temperature → PCB → heat sink, transformer core loss → insulation layer, inductor copper loss → air convection. Calculate heat conduction: ,in is the heat conduction power, is the thermal conductivity, is the temperature gradient. Finite difference method (FDM) or thermal network simulation (ANSYS Icepak) is used to calculate the steady-state temperature and transient response curves, including MOSFET temperature rise curve, PCB copper layer temperature distribution, and transformer thermal response time.

[0054] S23. Construct a constraint matrix according to the dynamic sensitivity data and the dynamic temperature response data to obtain complete dynamic constraint data.

[0055] In one embodiment, the dynamic constraint matrix is ​​a matrix that describes how various constraints (such as temperature limits, power loss, working efficiency, etc.) interact with each other during the dynamic change of the system. By combining sensitivity data with temperature response data, a constraint matrix can be constructed and adjusted according to the performance requirements of the on-board charger. Based on the thermal response data of the components, the operating temperature range of each component is defined (for example, the maximum operating temperature of MOSFET is 150°C), and the conditions beyond the temperature range are constrained. According to the results of the dynamic sensitivity analysis, the maximum tolerance value of the power loss is set. For example, under high load conditions, the switching loss cannot exceed a certain value to avoid overheating. The conversion efficiency of the charger is set to be kept within a certain range, for example, the system efficiency is required to be not less than 90%. Constraints such as temperature, power loss, and efficiency are constructed in a matrix form. The constraint matrix is ​​usually a multidimensional matrix, in which rows represent different constraints (such as temperature constraints, power constraints), and columns represent component parameters or performance indicators. The temperature limit of each component is listed and combined with the temperature response data. For example, when the temperature of a component exceeds 150°C, the constraint condition is not met. The maximum power loss limit of each component is listed and combined with the sensitivity data. For example, when the power loss of a switching device is too high, its operating frequency needs to be adjusted. The constraint matrix will include all dynamic constraints (such as temperature, power loss, efficiency, etc.) and the relationship between component parameters or performance indicators to form complete dynamic constraint data. This data can be used to further optimize or adjust the design of the on-board charger.

[0056] In one embodiment, the objective function is set: , so that ,in To minimize, is the power loss. Constraints: . Construct the optimization constraint matrix: ,in To optimize the constraint matrix, To optimize the variables, For the constraint vector, form the matrix:

[0057] in is the switching frequency, is the on-resistance of the MOSFET, is the number of turns of the transformer primary winding, is the number of turns of the transformer secondary winding. The Lagrange multiplier method is used to solve the optimal parameters ,in is the gradient change of the Lagrangian function, is the objective function gradient, is the Lagrange multiplier, is the constraint gradient. A genetic algorithm is used to perform a global search to find the optimal switching frequency, MOSFET parameters, and transformer parameters.

[0058] Optionally, performing hierarchical MINLP solution on the dynamic sensitivity matrix data to obtain battery impedance change curve data includes: S31, performing top-level MINLP solution on the dynamic sensitivity matrix data to obtain top-level device combination data; In one embodiment, the charger conversion efficiency is maximized: ,in To maximize, is the conversion efficiency, is the output power, is the input power. Minimize the power loss of components such as MOSFET, transformer, filter capacitor, etc.: ,in To minimize, is the total power loss, is the MOSFET conduction loss, is the MOSFET switching loss, is the transformer core loss, is the AC loss of the filter capacitor. The MOSFET junction temperature does not exceed 150°C: , transformer core temperature does not exceed 120°C: . Select different specifications of available MOSFETs, diodes, transformers, and filter capacitors to form a set of discrete variables: , ,in Select the variant for the MOSFET, For silicon-based MOSFET, For SiC MOSFET, For GaN MOSFET, Select the variable for the number of transformer turns, For winding scheme 1, For winding scheme 2, The winding scheme 3 is a preset winding parameter data, which can be obtained by querying existing data and solved by MINLP method. ,in To minimize, are the variables to be optimized, including MOSFET, transformer, filter capacitor, etc. is the objective function, the charger power loss model, is an inequality constraint, is an equality constraint.

[0059] S32, performing underlying MINLP solution on the dynamic sensitivity matrix data to obtain continuous optimization data; In one embodiment, continuous parameter optimization (such as switching frequency, PWM duty cycle, inductance value, etc.) is performed on the selected components. Optimization variables are defined, such as switching frequency , optimized between 50kHz-500kHz: ; PWM duty cycle D, adjusted between 0.2-0.8: ; The inductance value L satisfies the CCM (continuous conduction mode) condition: ,in is the inductance value, is the output voltage, is the PWM duty cycle, is the switching frequency, is the current ripple. Establish a continuous optimization mathematical model: ,in To minimize, is the power loss, is the switching frequency, is the PWM duty cycle, is the inductance value; define constraints, such as voltage ripple constraint: ,in is the output voltage ripple; the maximum current constraint of the battery terminal is: battery charging current Maximum allowable current . Using the MINLP method to solve, ,in, To optimize the variable iteration value, is the current optimization variable value, is the learning rate, is the objective function gradient.

[0060] S33, generating time-varying constraints according to the dynamic sensitivity matrix data to obtain time-varying constraint data; In one embodiment, a time-varying constraint is imposed on the optimization variable based on the battery SOC (state of charge) change. SOC-related battery impedance: ,in is the battery impedance, is the DC equivalent resistance, is the AC equivalent resistance, is an imaginary unit, is the angular frequency, is the battery equivalent capacitance; ,in is the SOC-related battery internal resistance, is the DC equivalent resistance, is the exponential decay coefficient, is the natural index, is the SOC influence coefficient, is the battery charge state. Dynamically adjust the constraints, such as maintaining the battery voltage within the allowable range during charging: ,in is the battery terminal voltage, is the minimum voltage, the lowest allowed charging voltage (V), is the maximum voltage, the highest allowed charging voltage (V); dynamically adjust the charging current: ,in is the charging current, the current (A) supplied by the charger to the battery, It is a SOC-related charging current function and a dynamic adjustment strategy for the charging current.

[0061] S34. Solve the top-level device combination data and the continuous optimization data according to the time-varying constraint data to obtain the battery impedance change curve data.

[0062] In one embodiment, the battery impedance change curve is calculated using the optimized device combination and continuous optimization parameters. Combining all the optimization parameters, the optimization problem is constructed: ,constraint: ,in is the battery terminal voltage, is the minimum voltage, the lowest allowed charging voltage (V), is the maximum voltage, the highest allowed charging voltage (V). Using MATLAB Simulink / EIS (electrochemical impedance spectroscopy) testing, draw the SOC change vs. impedance curve, calculate the impedance change under different charging parameters, and obtain the battery impedance change curve data.

[0063] Optionally, performing online parameter reconstruction according to the battery impedance change curve data to obtain vehicle charger component parameter optimization data includes: S41, acquiring real-time electrical data, and performing optimal parameter matching according to the real-time electrical data and battery impedance change curve data to obtain primary optimization parameter data; In one embodiment, the voltage and current data of the battery terminal of the on-board charger during the charging process are collected in real time by high-precision voltage and current sensors. A high-frequency sampling module (such as an ADC module) is used to sample at a rate of more than one thousand times per second to ensure that the dynamic signals that change rapidly during the charging process can be captured. According to the internal resistance change characteristics of the battery, a known battery electrochemical model (such as the Randles circuit model) is used to describe the change of battery impedance. The internal resistance of the battery usually changes with temperature, current and state of charge (SOC). By monitoring the charging and discharging process of the battery, real-time impedance change data is obtained. The impedance of the battery is measured in real time using a pulse signal or a frequency response analysis method (such as EIS, electrochemical impedance spectroscopy) to obtain a curve of the change of the battery impedance over time. The real-time battery impedance change curve is combined with the voltage and current data at the battery terminal. According to the change law of the battery impedance, it is used as a constraint condition to match the electrical data of the on-board charger to calculate the charger parameters suitable for the current working state. For example, according to the change of the internal resistance of the battery, the output voltage and current of the charger are adjusted to improve the charging efficiency and avoid overcharging. Use optimization algorithms (such as genetic algorithms, particle swarm optimization, least squares method, etc.) to match parameters. Input variables such as battery impedance change, output current, voltage, etc. into the optimization algorithm to obtain preliminary optimization parameters of charger components (such as switching frequency, capacitance, inductance, etc.). Through optimal parameter matching, primary optimization parameter data is obtained. For example, set the charging current, charging voltage, and parameters of other components (such as MOSFET switching frequency) suitable for the current battery state. These primary parameters can be used as the basis for preliminary adjustments during real-time charging.

[0064] S42, performing real-time accelerated calculation and parameter adjustment according to the primary optimization parameter data to obtain secondary optimization parameter data; In one embodiment, the primary optimization parameter data is calculated in real time through the simulation model of the on-board charger (including electrical and thermal models) to simulate various dynamic situations that may occur in the actual operation of the charger, such as battery voltage fluctuations, current peaks, etc. The simulation is accelerated using a real-time computing framework (such as the real-time simulation module of MATLAB Simulink) to ensure that accurate results can be obtained in a short time. The optimization parameters of the charger components are quickly calculated through parallel computing and GPU acceleration technology. For example, GPU parallel computing is performed using frameworks such as CUDA and OpenCL, thereby greatly improving the calculation speed. According to the results of the accelerated calculation, the key parameters of the on-board charger (such as switching frequency, current peak, filter capacitor, etc.) are adjusted to ensure that the working state of the charger components during the charging process is within the optimization range. For example, if the simulation results show that a certain parameter will cause excessive heating, its value is adjusted in time. The charging strategy is adjusted in real time according to the calculation results. For example, during the charging process, as the battery impedance changes, the charging current or charging voltage can be automatically adjusted, thereby extending the battery life and improving the charging efficiency. Based on the results of real-time accelerated calculation and adjustment, more accurate secondary optimization parameters are obtained. For example, further fine-tuned battery charging curves, current waveforms, charging time, etc. are included.

[0065] S43: Perform continuous feedback adjustment based on the secondary optimization parameter data to obtain vehicle charger component parameter optimization data.

[0066] In one embodiment, by real-time monitoring of the operating status of the battery and the charger, combined with the above-mentioned optimization parameters, the various parameters in the charging process are continuously fed back and adjusted. This is achieved through an embedded control system, which feeds back the battery status information (such as SOC, impedance, etc.) in real time during each charging cycle, and automatically adjusts the output of the charger. The charger can automatically adjust the working parameters according to the real-time status of the battery and the charging demand through a predefined control algorithm. For example, using algorithms such as fuzzy control and PID control, the charging current and voltage are adjusted according to the battery status. According to the performance of the battery during multiple charging processes, the working parameters of the charger are gradually adjusted. For example, as the number of charging times increases, the internal resistance of the battery will change, and the charger dynamically adjusts the working voltage, current, etc. to ensure the best charging effect. The stability of the system is analyzed to ensure that the working state of the charger after each adjustment can run smoothly and does not cause excessive load or damage to the battery or charger components. Through continuous feedback adjustment, the optimal parameter data of the on-board charger components is finally obtained. Including the final charging current, voltage, switching frequency, filter capacitor, battery terminal voltage and other parameters, improve charging efficiency, extend battery life, and ensure the safety and stability of the charging process.

[0067] Optionally, the obtaining of physical parameter data of the on-board charger PCB and constructing a multi-physics field model according to the physical parameter data includes: Obtaining physical parameter data of the on-board charger PCB, wherein the physical parameter data includes IV characteristic curve data, on-resistance data, BH curve data of magnetic components, and winding AC resistance; In one embodiment, a source measurement unit (SMU) or a power analyzer is used to perform a current-voltage (IV) characteristic test on the power semiconductor device (MOSFET, diode) of the on-board charger. The conduction characteristics, current leakage characteristics, avalanche voltage characteristics, etc. of the MOSFET or diode at different temperatures are recorded. A certain gate drive voltage is applied under different temperature conditions, and the source-drain on-resistance of the MOSFET is measured. A temperature control platform (such as a hot stage) is used to measure different temperature points, usually ranging from -40°C to 150°C, for electrical model construction to determine the relationship between temperature and conduction loss. The hysteresis loop (BH curve) of the core material is measured using a magnetic analyzer (such as an LCR meter, a vector network analyzer). It is used to describe the nonlinear permeability characteristics of transformers and inductors in order to accurately simulate inductance, current ripple and core loss. A spectrum analyzer or a vector network analyzer is used to measure the AC resistance (R_AC) of the transformer and inductor windings at different frequencies. It is used to analyze the skin effect and proximity effect under high-frequency current, and optimize the winding structure to reduce AC loss.

[0068] Perform temperature segment fitting according to physical parameter data to obtain electrical characteristic data; In one embodiment, since the electrical characteristics of devices such as MOSFET and inductor vary with temperature, segmented fitting is required to describe their temperature dependence. Polynomial fitting or exponential fitting is used to fit the curve of on-resistance varying with temperature, and curve fitting is performed on the BH curve data to extract the magnetic permeability (μ) at different temperatures. The winding AC resistance data is fitted to construct a frequency-temperature dependency relationship. A set of temperature-dependent electrical characteristic data for electrical simulation is obtained.

[0069] Power consumption and junction temperature mapping are performed based on physical parameter data to obtain thermal characteristic data; In one embodiment, the relationship between power consumption and junction temperature is established for thermal modeling. Finite element analysis (FEM) tools (such as ANSYS Icepak) are used to calculate the effect of power consumption of each component on junction temperature, and power consumption-junction temperature mapping relationship data is obtained for subsequent thermal model construction.

[0070] The electrical property model of the MOSFET device is constructed according to the electrical characteristic data to obtain the electrical property model: In one embodiment, an electrical model of MOSFET is established for electromagnetic simulation. Based on the data of temperature segment fitting, a SPICE equivalent model of MOSFET is created, and the parameters are adjusted to match the measured IV characteristics. The equivalent capacitance and equivalent inductance are established using the Cauer model or the Foster model. The parasitic parameters of MOSFET (such as gate charge, gate-drain capacitance, etc.) are calculated.

[0071] The thermal model is constructed according to the thermal characteristic data to obtain the thermal model: In one embodiment, a thermal network model is established to describe the thermal conduction and thermal radiation characteristics of components such as MOSFET. The temperature distribution is calculated using the finite difference method (FDM) or the finite element method (FEM). ,in is the MOSFET junction temperature, is the total power loss, is the thermal resistance, is the base of natural logarithm, For time, is the thermal time constant, the heat transfer path of junction temperature-PCB-heat sink-external air.

[0072] Solve the thermal network model based on the power loss data in the thermal model and electrical model to obtain the junction temperature data; In one embodiment, the junction temperature of different components of the on-board charger during operation is calculated. The thermal model and the electrical model are combined to perform thermal simulation using software such as COMSOL and ANSYS to solve the temperature distribution. The junction temperature data is output for optimizing the electrical model.

[0073] The electrical property model is modified according to the junction temperature data to obtain an electrical property modification model; In one embodiment, based on the calculated junction temperature data, a weighted calculation is performed. ,in is the temperature-dependent on-resistance, is the on-resistance at the reference temperature, is the temperature coefficient, Adjust the IV characteristic parameters of MOSFET for the MOSFET junction temperature.

[0074] Perform intrinsic orthogonal decomposition on the thermal model to obtain a reduced-order model; In one embodiment, POD decomposition is used: ,in is the distribution of thermal field over time and space, is the mode index, representing the index of the eigenmode, To retain the number of modes, the number of reduced eigenmodes is selected. is the time correlation coefficient, reflecting the weight coefficient of the mode over time, is the spatial characteristic function, representing the spatial distribution of the main heat transfer mode. The high-dimensional thermal model is reduced in dimension. The main characteristic modes are calculated: ,in is the left singular matrix, representing the main spatial modes of the system, is a singular value diagonal matrix, representing energy distribution. The larger the singular value, the more important the mode is. is a right singular matrix, representing time-related information, is the input matrix, representing the temperature field data matrix of the thermal model, For the singular value decomposition, calculate the matrix Singular value decomposition of , extracting the most important modes. Build a reduced-order state-space model: ,in is the state variable derivative, is the reduced-order system matrix, and the reduced-order system dynamics matrix, is the state variable, the reduced state variable vector, is the input matrix, the reduced input matrix, is the input variable of the system, such as the heat source power, is the output variable of the system, such as the temperature measurement value, is the output matrix after reduction.

[0075] The electrical correction model and the reduced-order model are packaged to obtain a multi-physics field model.

[0076] In one embodiment, a ROM framework is used to integrate different sub-models into a simulation framework, integrating electrical and thermal models to form a complete multi-physics simulation framework. Use MatlabSimulink+Simscape to integrate multi-physics models, such as electrical models (SPICE / Simulink), thermal models (COMSOL / Icepak), and magnetic models (JMAG / Maxwell). Establish simulation processes, such as input power, calculate electrical losses, calculate heat flux distribution, update electrical parameters, and output system simulation results. Build data transmission interfaces, such as electrical-thermal coupling: , thermal-electric feedback: ,in is the power loss, such as the total power loss of MOSFET, inductor, and transformer (W), is the MOSFET junction temperature, the MOSFET junction temperature over time (°C), is the on-resistance, the temperature-dependent on-resistance of the MOSFET (Ω), is the saturation current, the saturation current of the MOSFET (A).

[0077] Optionally, the electrical characteristic data includes first electrical characteristic data, second electrical characteristic data, third electrical characteristic data, fourth electrical characteristic data and fifth electrical characteristic data, and the electrical characteristic data is obtained by performing temperature segmented fitting according to the physical parameter data, including: Extract thermal characteristics based on the thermal data of the on-board charger to obtain thermal characteristic data; In one embodiment, an infrared thermal imager or a thermocouple is used to measure the temperature field of the main components of the on-board charger (MOSFET, inductor, transformer, etc.) under different working conditions. The measurement points include MOSFET junction temperature, PCB copper wire temperature, magnetic component temperature, and air convection temperature. The thermal resistance ( ): ,in is the thermal resistance, is the MOSFET junction temperature, is the ambient temperature, the air temperature around the on-board charger (°C), is the total power loss, the total power loss caused by the MOSFET (W).

[0078] Perform on-state voltage extraction, saturation current extraction and switching time extraction according to IV characteristic curve data, and obtain on-state voltage data, saturation current data and switching time data respectively; In one embodiment, key electrical characteristic data including on-state voltage, saturation current and switching time are extracted from the IV curve. Use a source measurement unit (SMU) to perform IV testing on the MOSFET: Extract the on-state voltage ( )、Extract saturation current( ). To measure the switching time, use an oscilloscope to capture the MOSFET turn-on / turn-off waveforms: Get the rise time ( ), get the fall time ( ).

[0079] Performing characteristic curve fitting on the on-voltage data, the saturation current data, and the switching time data according to the thermal characteristic data to obtain first electrical characteristic data; In one embodiment, a linear regression method is used to fit the relationship between the on-state voltage and the temperature: ,in is the temperature-dependent threshold voltage, the MOSFET Threshold voltage (V) is the threshold voltage at the reference temperature, Threshold voltage (V) at (e.g. 25°C), is the temperature coefficient, which reflects the effect of temperature on the threshold voltage. is the device temperature, As the reference temperature, the exponential decay model is used to fit the saturation current temperature: ,in is the temperature-dependent saturation current, is the saturation current at the reference temperature, is the natural index, is the exponential decay coefficient, is the device temperature, As the reference temperature, the second-order polynomial fitting is used to fit the switching time temperature. ,in is the temperature-dependent switching time, is the second-order coefficient, controlling the coefficient of the square term of temperature, is the device temperature, is the first-order coefficient, controlling the coefficient of the linear temperature term, is a constant term that controls the temperature independent offset.

[0080] Performing temperature fitting on the on-resistance data according to the thermal characteristic data to obtain second electrical characteristic data; In one embodiment, a model of MOSFET on-resistance changing with temperature is established, and the on-resistance of MOSFET at different temperatures is measured by a four-probe method. ,in is the temperature-dependent on-resistance, the on-resistance of the MOSFET at temperature T (Ω), is the drain-source voltage, MOSFET drain-source voltage (V), is the drain current, MOSFET drain current (A). A second-order fit to the temperature dependence is used: , is the temperature-dependent on-resistance, is the on-resistance at the reference temperature, is the first-order temperature coefficient, used for fitting in the medium and low temperature range. is the device temperature, It is the second-order temperature coefficient, which mainly affects the resistance change in the high temperature zone. The coefficient is obtained by fitting empirical data or current data.

[0081] Calculate the power loss according to the on-voltage data, the saturation current data and the switching time data to obtain the power loss data; In one embodiment, the power loss is calculated based on the on-state voltage, saturation current and switching time to calculate the loss of the MOSFET in the on-state: ,in, is the conduction loss, the power loss of the MOSFET in the on state (W), is the square of the drain current, the square of the MOSFET drain current ( ), is the on-resistance, the on-resistance of the MOSFET (Ω). The switching loss is calculated using the standard loss mathematical model: ,in is the switching loss, is the drain-source voltage, MOSFET drain-source voltage (V), is the drain current, the drain current of MOSFET (A), is the MOSFET turn-on time (s), is the MOSFET turn-off time (s), is the switching frequency of the MOSFET (Hz).

[0082] Perform switching loss fitting according to the thermal characteristic data and the power loss data to obtain third electrical characteristic data; In one embodiment, a mathematical model of MOSFET switching loss changing with temperature is established, and an exponential regression model is used to fit the switching loss temperature relationship. ,in is the temperature dependent switching loss, is the switching loss at reference temperature, is the natural index, is the temperature gain coefficient, is the device temperature, is the reference temperature.

[0083] Performing segmented exponential fitting on the BH curve data of the magnetic element according to the thermal characteristic data to obtain fourth electrical characteristic data; In one embodiment, a temperature-dependent magnetic model of the transformer and the inductor is established, and the BH curve is used to fit the core loss by segmented exponential fitting. ,in is the temperature-dependent magnetic induction intensity, is the magnetic induction intensity at the reference temperature, is the natural index, is the temperature attenuation coefficient, is the device temperature, is the reference temperature.

[0084] The fifth electrical characteristic data is obtained by fitting the temperature-frequency variation according to the thermal characteristic data and the winding AC resistance data.

[0085] In one embodiment, power law regression is used to fit the AC resistance temperature frequency relationship. ,in, is the temperature-frequency dependent AC resistance, is the AC resistance at reference temperature and frequency, is the frequency gain factor, is the frequency power law term, the frequency The power law relationship reflects the influence of skin effect and proximity effect. is the natural index, is the temperature exponential growth coefficient, is the device temperature, is the reference temperature.

[0086] Optionally, performing power consumption junction temperature mapping according to the physical parameter data to obtain thermal characteristic data includes: Decompose transient power consumption based on physical parameter data and electrical characteristic data to obtain time-varying power consumption distribution data; In one embodiment, electrical loss data, such as MOSFET conduction loss, is extracted: ,in is the MOSFET conduction loss, is the square of the MOSFET drain current, is the on-resistance of MOSFET; MOSFET switching loss: ,in is the MOSFET switching loss, is the MOSFET drain-source voltage, is the MOSFET drain current, is the MOSFET turn-on time, is the MOSFET turn-off time, is the MOSFET switching frequency; transformer core loss: , is the core loss, is the loss coefficient, is the switching frequency exponential term, reflecting the power law relationship between core loss and switching frequency. is the exponential coefficient, the frequency's influence on the core loss exponentially, is the magnetic induction intensity index term, reflecting the change of core loss with magnetic induction intensity. is the exponential coefficient, the influence index of magnetic induction intensity on core loss. Decomposition of high frequency switching loss: , is the high frequency switching loss, is the harmonic component index, is the harmonic order, is the harmonic amplitude, is the resonant frequency, For time, For the The phase angle (rad) of the first-order harmonic is used to analyze the power consumption at different time points using time-varying signal analysis.

[0087] A three-dimensional thermal impedance network model is constructed according to the physical parameter data and the time-varying power consumption distribution data to obtain a three-dimensional thermal impedance network model; In one embodiment, transistor thermal network modeling: ,in is the transient thermal impedance, is the thermal resistance component index, is the number of thermal network layers, For the Layer thermal resistance, is the natural index, For time, For the Layer thermal time constant. The horizontal heat diffusion of the PCB copper layer is calculated using the finite element method (FEM): ,in is the heat conduction power, is the thermal conductivity, For the temperature gradient, calculate the heat flow path from PCB to heat sink, and analyze the air cooling / liquid cooling solution.

[0088] Perform transient thermal field simulation based on the three-dimensional thermal impedance network model to obtain transient thermal field data; In one embodiment, a thermal field simulation is performed based on a three-dimensional thermal impedance network model to analyze the temperature changes during the operation of the charger. Simulation boundary conditions are set, power consumption data (from transient power consumption decomposition) is input, and the ambient temperature setting ( ), air convection heat dissipation parameters (wind speed, heat dissipation coefficient). Solve the steady-state heat distribution: ,in is the steady-state temperature distribution, is the thermal power input, is the thermal conductivity, is the heat dissipation area, is the ambient temperature. Solve for the transient temperature change: ,in is the material density ( ), is the specific heat capacity of the material ( ), is the heat diffusion operator, i.e. the heat conduction term, which describes the spatial diffusion of heat flow. is the thermal conductivity of the material, is the spatial rate of change of temperature, To calculate the power consumption of the device, COMSOL Multiphysics / ANSYS Icepak is used for transient simulation. Extract the MOSFET junction temperature curve: ,in is the MOSFET junction temperature, is the MOSFET junction temperature function, junction temperature and power loss Thermal resistance relationship function.

[0089] Transfer learning calibration is performed based on transient thermal field data to obtain thermal characteristic data.

[0090] In one embodiment, experimental measurement data is compared with simulation data, and transfer learning is used to optimize the thermal model. A transfer learning model is constructed using a neural network (DNN) + Gaussian process regression (GPR). The simulated temperature data is input and the experimental temperature measurement data is output. The loss function is used to ,in To minimize the loss function, are the thermal model parameters to be optimized, is the temperature data obtained through transient thermal field simulation (°C), is the real temperature data (°C) measured by an infrared thermal imager or thermocouple. Update the thermal resistance parameters after training: ,in is the thermal resistance optimized by transfer learning, is the thermal resistance (°C / W) in the initial simulation model, is the thermal resistance correction value (°C / W) updated after transfer learning.

[0091] Optionally, constructing a three-dimensional thermal impedance network model according to the physical parameter data and the time-varying power consumption distribution data to obtain the three-dimensional thermal impedance network model includes: Modeling the thermal behavior of transistors based on physical parameter data and time-varying power consumption distribution data to obtain a thermal behavior model of transistors; In one embodiment, based on physical parameter data and time-varying power consumption distribution, a thermal behavior model of power semiconductors such as MOSFET and IGBT is established to describe the relationship between power loss and junction temperature change. The thermal conductivity of MOSFET chips (such as Si, SiC, GaN) is obtained using a material database, such as =148W / (m\K). GaN: =230W / (m\K), SiC: =350W / (m\K). Specific heat capacity ( ): Calculate the specific heat capacity of chip materials, such as the specific heat capacity of Si: =700J / (kg\K). Obtain the junction-to-case thermal resistance through experiments or data sheets: ,in is the junction-to-case thermal resistance, the thermal resistance from the transistor junction to the case, is the junction-to-case temperature difference, the temperature difference between the transistor junction temperature and the case temperature (°C), is the power loss of the device.

[0092] Build a transistor thermal network model: ,in is the transient thermal impedance, the transistor is Thermal impedance under (°C / W), is the thermal resistance component index, is the number of thermal network layers, For the Layer thermal resistance, is the natural index, For time, For the Layer thermal time constant. Steady-state thermal distribution: , is the steady-state junction temperature, the junction temperature of the MOSFET / IGBT in steady state (°C), is the power loss, is the junction-to-case thermal resistance, is the case-air thermal resistance.

[0093] The thermal behavior model of the PCB layer is modeled according to the physical parameter data and the time-varying power consumption distribution data to obtain the thermal behavior model of the PCB layer; In one embodiment, the thermal conductivity of the copper layer includes the copper layer: =385W / (m\K), FR4 insulation layer: =0.3W / (m\K).

[0094] The finite difference method (FDM) is used to calculate the internal temperature distribution of the PCB. The PCB can be decomposed into multiple layers, and the temperature of each layer can be calculated by the thermal balance equation: ,in is the thermal conductivity, the thermal conductivity of the material ( ), which determines the heat conduction capacity, For the Temperature at time / sampling point, For the Temperature at time / sampling point, is the interlayer thickness, The heat flow path is calculated as the horizontal heat diffusion of the copper layer, the thermal resistance of the heat flow path from the pad to the MOSFET, and the conduction path from the PCB to the heat sink.

[0095] Structural thermal behavior modeling is performed according to physical parameter data and time-varying power consumption distribution data to obtain a structural thermal behavior model; In one embodiment, the thermal behavior of the heat sink, air cooling convection, and package shell is calculated. The thermal conductivity of the aluminum heat sink is obtained ( =237W / (m\K). Calculate the heat transfer coefficient of the heat sink (h): , is the convective heat transfer coefficient, is the thermal conductivity of the fluid, is the characteristic length, is the Nusselt number (calculated from the Reynolds number and Prandtl number). Calculate the cooling capacity of the fan , The fan cooling power, is the air mass flow rate, is the specific heat capacity of air, is the air temperature rise. Calculate the thermal resistance of the housing: , is the case thermal resistance, is the package thickness, is the thermal conductivity of the packaging material, is the package surface area.

[0096] According to the transistor thermal behavior model, PCB layer thermal behavior model and structural thermal behavior model, heat flow coupling simulation is performed to obtain a three-dimensional thermal impedance network model.

[0097] In one embodiment, a three-dimensional thermal impedance network is constructed by combining the thermal behavior model of transistors, PCBs, and structures, and a heat flow coupling simulation is performed to establish a three-dimensional thermal network equation: ,in is the thermal conductivity matrix, is the temperature vector, is the heat source vector. Finite element analysis (FEM) or finite difference method (FDM) is used to simulate using COMSOL, ANSYS, and MATLAB PDEToolbox. Optionally, the present application further provides a vehicle charger component parameter optimization system based on a mathematical model, which is used to execute the vehicle charger component parameter optimization method based on a mathematical model as described above, and the vehicle charger component parameter optimization system based on a mathematical model includes: A multi-physics joint modeling module is used to perform multi-physics joint modeling based on the on-board charger switch device to obtain a simulation model of the on-board charger components; A dynamic sensitivity analysis module is used to perform dynamic sensitivity analysis based on the vehicle charger component simulation model to obtain dynamic sensitivity matrix data; A hierarchical MINLP solution module is used to perform hierarchical MINLP solution on the dynamic sensitivity matrix data to obtain the battery impedance change curve data; The online parameter reconstruction module is used to perform online parameter reconstruction according to the battery impedance change curve data to obtain the parameter optimization data of the vehicle charger components.

[0098] Therefore, from any point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the attached application documents rather than the above description, and it is intended that all changes falling within the meaning and scope of equivalent elements of the application documents are included in the present invention.

[0099] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for optimizing the parameters of vehicle charger components based on a mathematical model, characterized in that: The method comprises: Carry out multi-physics joint modeling based on the on-board charger switch device to obtain the on-board charger component simulation model; Perform dynamic sensitivity analysis based on the on-board charger component simulation model to obtain dynamic sensitivity matrix data; Perform hierarchical MINLP on the dynamic sensitivity matrix data to obtain the battery impedance change curve data; Online parameter reconstruction is performed based on the battery impedance change curve data to obtain the parameter optimization data of the on-board charger components.

2. The method according to claim 1, characterized in that The multi-physics field joint modeling is performed according to the on-board charger switch device to obtain the on-board charger component simulation model, including: Collecting basic data of the on-board charger for the on-board charger switch device to obtain basic data of the on-board charger, wherein the basic data of the on-board charger includes electrical data of the on-board charger and thermal data of the on-board charger; The electrical thermal model of the device is constructed according to the electrical data and thermal data of the on-board charger to obtain the electrical thermal model of the device; Obtain the physical parameter data of the on-board charger PCB and build a multi-physics field model based on the physical parameter data; The overall simulation model of the on-board charger is obtained by performing simulation calculations based on the multi-physics field model and the device electrical and thermal model.

3. The method according to claim 1, characterized in that The dynamic sensitivity analysis is performed according to the vehicle charger component simulation model to obtain dynamic sensitivity matrix data, including: Perform sensitivity analysis based on the on-board charger component simulation model to obtain dynamic sensitivity data; Thermal network modeling is performed based on the vehicle charger component simulation model to obtain dynamic temperature response data; The constraint matrix is ​​constructed according to the dynamic sensitivity data and the dynamic temperature response data to obtain complete dynamic constraint data.

4. The method according to claim 1, characterized in that: The layered MINLP solution is performed on the dynamic sensitivity matrix data to obtain the battery impedance change curve data, including: Perform top-level MINLP on the dynamic sensitivity matrix data to obtain top-level device combination data; Perform the underlying MINLP solution on the dynamic sensitivity matrix data to obtain the continuous optimization data; Generate time-varying constraints according to dynamic sensitivity matrix data to obtain time-varying constraint data; The top-level device combination data and the continuous optimization data are solved according to the time-varying constraint data to obtain the battery impedance change curve data.

5. The method according to claim 1, characterized in that The online parameter reconstruction is performed according to the battery impedance change curve data to obtain the vehicle charger component parameter optimization data, including: Acquire real-time electrical data, and perform optimal parameter matching based on the real-time electrical data and battery impedance change curve data to obtain primary optimization parameter data; Perform real-time accelerated calculation and parameter adjustment based on primary optimization parameter data to obtain secondary optimization parameter data; Continuous feedback adjustment is performed based on the secondary optimization parameter data to obtain the vehicle charger component parameter optimization data.

6. The method according to claim 2, characterized in that The obtaining of physical parameter data of the on-board charger PCB and constructing a multi-physics field model according to the physical parameter data includes: Obtaining physical parameter data of the on-board charger PCB, wherein the physical parameter data includes IV characteristic curve data, on-resistance data, BH curve data of magnetic components, and winding AC resistance; Perform temperature segment fitting according to physical parameter data to obtain electrical characteristic data; Power consumption and junction temperature mapping are performed based on physical parameter data to obtain thermal characteristic data; The electrical property model of the MOSFET device is constructed according to the electrical characteristic data to obtain the electrical property model: The thermal model is constructed according to the thermal characteristic data to obtain the thermal model: Solve the thermal network model based on the power loss data in the thermal model and electrical model to obtain the junction temperature data; The electrical property model is modified according to the junction temperature data to obtain an electrical property modification model; Perform intrinsic orthogonal decomposition on the thermal model to obtain a reduced-order model; The electrical correction model and the reduced-order model are packaged to obtain a multi-physics field model.

7. The method according to claim 6, characterized in that The electrical characteristic data includes first electrical characteristic data, second electrical characteristic data, third electrical characteristic data, fourth electrical characteristic data and fifth electrical characteristic data, and the electrical characteristic data is obtained by performing temperature segmented fitting according to the physical parameter data, including: Extract thermal characteristics based on the thermal data of the on-board charger to obtain thermal characteristic data; Perform on-state voltage extraction, saturation current extraction and switching time extraction according to IV characteristic curve data, and obtain on-state voltage data, saturation current data and switching time data respectively; Performing characteristic curve fitting on the on-voltage data, the saturation current data, and the switching time data according to the thermal characteristic data to obtain first electrical characteristic data; Performing temperature fitting on the on-resistance data according to the thermal characteristic data to obtain second electrical characteristic data; Calculate the power loss according to the on-voltage data, the saturation current data and the switching time data to obtain the power loss data; Perform switching loss fitting according to the thermal characteristic data and the power loss data to obtain third electrical characteristic data; Performing segmented exponential fitting on the BH curve data of the magnetic element according to the thermal characteristic data to obtain fourth electrical characteristic data; The fifth electrical characteristic data is obtained by fitting the temperature-frequency variation according to the thermal characteristic data and the winding AC resistance data.

8. The method according to claim 6, characterized in that The power consumption junction temperature mapping is performed according to the physical parameter data to obtain thermal characteristic data, including: Decompose transient power consumption based on physical parameter data and electrical characteristic data to obtain time-varying power consumption distribution data; A three-dimensional thermal impedance network model is constructed according to the physical parameter data and the time-varying power consumption distribution data to obtain a three-dimensional thermal impedance network model; Perform transient thermal field simulation based on the three-dimensional thermal impedance network model to obtain transient thermal field data; Transfer learning calibration is performed based on transient thermal field data to obtain thermal characteristic data.

9. The method according to claim 8, characterized in that The three-dimensional thermal impedance network model is constructed according to the physical parameter data and the time-varying power consumption distribution data to obtain the three-dimensional thermal impedance network model, including: Modeling the thermal behavior of transistors based on physical parameter data and time-varying power consumption distribution data to obtain a thermal behavior model of transistors; The thermal behavior model of the PCB layer is modeled according to the physical parameter data and the time-varying power consumption distribution data to obtain the thermal behavior model of the PCB layer; Structural thermal behavior modeling is performed according to physical parameter data and time-varying power consumption distribution data to obtain a structural thermal behavior model; According to the transistor thermal behavior model, PCB layer thermal behavior model and structural thermal behavior model, heat flow coupling simulation is performed to obtain a three-dimensional thermal impedance network model.

10. A vehicle charger component parameter optimization system based on mathematical model, characterized in that: Used to execute the vehicle charger component parameter optimization method based on mathematical model as claimed in claim 1, the vehicle charger component parameter optimization system based on mathematical model comprises: A multi-physics joint modeling module is used to perform multi-physics joint modeling based on the on-board charger switch device to obtain a simulation model of the on-board charger components; A dynamic sensitivity analysis module is used to perform dynamic sensitivity analysis based on the vehicle charger component simulation model to obtain dynamic sensitivity matrix data; A hierarchical MINLP solution module is used to perform hierarchical MINLP solution on the dynamic sensitivity matrix data to obtain the battery impedance change curve data; The online parameter reconstruction module is used to perform online parameter reconstruction according to the battery impedance change curve data to obtain the parameter optimization data of the vehicle charger components.

Images