Multi-objective thermal optimization method, device and system for siC power electronic converter heat dissipation system

Abstract

The application belongs to the technical field of transformer heat dissipation, and discloses a SiC power electronic converter heat dissipation system multi-objective heat optimization method, which can add temperature uniformity as an optimization target to multi-objective optimization design, and form a three-dimensional multi-objective optimization design method together with pressure drop and thermal resistance. The final optimization result is that the Pin-Fin heat dissipation structure (which is different channel width, depth and spacing when applied to microchannels, and other heat dissipation structures can be analogized in this way) with different diameters, lateral spacings and longitudinal spacings along the heat accumulation direction, and such a structure can make the corresponding chip have different junction flow thermal resistances, so as to compensate for the chip temperature non-uniformity caused by fluid heat accumulation. The working efficiency of the heat dissipation system can be effectively improved, and the temperature non-uniformity of the power electronic converter can be effectively reduced.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of converter heat dissipation, and particularly relates to a SiC power electronic converter heat dissipation system multi-objective heat optimization method, device and system. BACKGROUND

[0002] With the rapid development of new energy, electric transportation, smart grid and other technologies, the demand for efficient, reliable and miniaturized power electronic converters is increasing. Traditional silicon (Si) based power electronic systems gradually show bottlenecks in high temperature, high frequency, high voltage and high power density application scenarios. In recent years, the rapid development of wide bandgap semiconductor materials, especially silicon carbide (SiC) devices, has brought revolutionary progress to power electronic technology. Compared with traditional Si-based materials, SiC has become a new generation of power electronic switching device material to replace silicon-based devices due to its high breakdown voltage, high thermal conductivity, wide bandgap and high electron mobility. Power modules constructed with SiC MOSFET or SiC SBD have obvious advantages in improving efficiency, reducing system size and weight, and promoting the development of high power density of power electronic systems.

[0003] However, the direct challenge brought by high power density is heat management. Although SiC devices can withstand higher temperatures, their continuous reliable operation still needs efficient heat dissipation structures to support. Traditional heat dissipation schemes (such as air cooling, water cooling plate) are difficult to meet the demand, and heat management has become a core bottleneck restricting system reliability and life.

[0004] The commonly used integrated Pin-Fin heat sink (pin-shaped heat sink) has the advantages of compact structure, large heat exchange area, good disturbance effect, etc. Compared with the traditional water cooling plate structure, it can effectively reduce the junction-to-flow thermal resistance by about 40%, and has been widely used in vehicle power modules. At the same time, micro-channel heat sinks are also well known in the industry because of their extremely high heat exchange efficiency. However, whether it is the pin design of Pin-Fin heat sink or the microstructure design of micro-channel heat sink, it relies on artificial experience trial and error, which is low in efficiency and difficult to globally optimize. At the same time, although the dense pin array and microstructure can improve the heat exchange area, it leads to a significant increase in cooling fluid flow resistance (pressure drop), and the heat resistance and energy consumption need to be balanced.

[0005] In view of the problems existing in the Pin-Fin radiator and the micro-channel radiator, many kinds of multi-objective optimization methods have been applied to the optimization design of them, among which the common ones are: genetic algorithm (GA), response surface method (RSM), multi-objective particle swarm optimization (MOPSO), etc. In the process of applying these methods, they generally optimize the key parameters of the radiator as a whole, such as: optimizing the first distance (the distance between the same row of needle columns), the second distance (the distance between adjacent needle column rows), and the needle column radius R of the Pin-Fin radiator; and optimizing the channel width, depth and spacing of the micro-channel radiator. Through the optimization design of these key parameters of the radiator, they can seek an optimal solution between the thermal resistance and the pressure drop.

[0006] Although these multi-objective optimization design methods currently used have made significant progress in the collaborative optimization of thermal resistance and pressure drop, they generally ignore a key problem: the thermal uniformity of the chips inside the SiC device. With the increase of power density, when the multiple chips inside the SiC device are connected in parallel or in series, due to the difference in packaging process, uneven current distribution and fluctuation of local heat transfer efficiency of the heat dissipation structure, the temperature gradient between the chips can reach more than ten to twenty degrees Celsius, which seriously threatens the service life and reliability of the system.

[0007] Since the current multi-objective optimization method of the radiator has significant defects in the control of thermal uniformity, it is urgent to develop a three-dimensional multi-objective heat dissipation optimization design method integrating thermal resistance, pressure drop and thermal uniformity to solve the heat management problem of high-power-density SiC power electronic converters and improve the service life and reliability of the system.

[0008] Through the above analysis, the problems and defects of the prior art are:

[0009] Whether it is the needle column design of the Pin-Fin radiator or the microstructure design of the micro-channel radiator, it depends on artificial experience and trial and error, which is low in efficiency and difficult to globally optimize. At the same time, although the dense needle column array and the microstructure can improve the heat exchange area, they significantly increase the flow resistance (pressure drop) of the cooling liquid, and it is also necessary to balance the thermal resistance and energy consumption. SUMMARY

[0010] In view of the problems existing in the prior art, the present application provides a multi-objective heat optimization method for a SiC power electronic converter heat dissipation system.

[0011] The present application is implemented as follows: a multi-objective heat optimization method for a SiC power electronic converter heat dissipation system comprises:

[0012] Step 1: According to the distribution of the chips on the three half-bridges (1), (2) and (3), the needle column part of the Pin-Fin radiator is divided into three parts (6), (7) and (8), which correspond to three optimization regions 1, 2 and 3, respectively;

[0013] Step 2, the key structural parameters of the Pin-Fin radiator and the liquid cooling channel are extracted;

[0014] Step 3, the optimization range of each key parameter is determined, considering the minimum and maximum limits of the preparation process on these parameters, combining simulation and experience, and eliminating the obviously ineffective value range;

[0015] Step 4, a multi-objective optimization algorithm combined with finite element simulation is used to solve the above multi-objective problem.

[0016] Further, the key structural parameters are:

[0017] The height Hpi of the pin, the vertical distance Dvi between the pins, the horizontal distance Dhi between the pins, the radius Ri of the pin, and the depth Hci of the channel.

[0018] Further, the multi-objective optimization algorithm is:

[0019]

[0020] The multi-objective optimization algorithm is responsible for generating and iterating the key parameters, and the finite element simulation is performed on the half-bridge module without using the same heat dissipation structure according to the key structural parameter group generated by the multi-objective optimization algorithm, and the thermal resistance (Rthjf), pressure drop (△P), and thermal uniformity (ψ) obtained by simulation are transmitted to the multi-objective optimization algorithm for iteration of the structural parameter group. After a certain number of iterations, the Pareto frontier is obtained and an optimal structural parameter group is found.

[0021] Another object of the present application is to provide a SiC power electronic converter heat dissipation system multi-objective thermal optimization system, which comprises:

[0022] The dividing module is used to divide the pin part of the Pin-Fin radiator into three parts (6), (7), and (8) according to the distribution of the chips on the three half-bridges (1), (2), and (3), corresponding to three optimization regions 1, 2, and 3, respectively.

[0023] The extraction module is used to extract the key structural parameters of the Pin-Fin radiator and the liquid cooling channel;

[0024] The determination module is used to determine the optimization range of each key parameter, considering the minimum and maximum limits of the preparation process on these parameters, combining simulation and experience, and eliminating the obviously ineffective value range;

[0025] The solving module is used to solve the above multi-objective problem by using a multi-objective optimization algorithm combined with finite element simulation.

[0026] In combination with the above technical solutions and the technical problems solved, the technical solutions to be protected by the present application have the following advantages and positive effects:

[0027] First, whether it is a traditional Pin-Fin radiator and a micro-channel radiator or a Pin-Fin radiator and a micro-channel radiator after the above multi-objective optimization design, due to the same chip layout and the same heat dissipation structure, the junction flow thermal resistance (chip to fluid thermal resistance) is certain. Such consistent structure will produce an increasing temperature gradient along the flow direction due to the thermal accumulation of the cooling fluid under the direct water cooling heat dissipation form, which is temperature non-uniformity on the power electronic converter.

[0028] In view of the problem of thermal uniformity of the existing multi-objective optimization method, the multi-objective optimization method in the present technology makes the following improvements. The target radiator to be optimized is divided into regions. Taking the Pin-Fin radiator as an example, the Pin-Fin radiator bottom heat dissipation needle column region is divided into several separate optimization units according to the distribution of the chip (heat source) above the radiator along the direction of cooling fluid thermal accumulation. The purpose of this is to obtain different junction flow thermal resistances through the separate optimization of the heat dissipation needle column below the corresponding chip. In this way, even if the temperature of the downstream fluid rises, the temperature of the chip can still be kept consistent with other chips through the lower junction flow thermal resistance.

[0029] Through such design, the present application can add temperature uniformity as an optimization target to the multi-objective optimization design, together with pressure drop and thermal resistance to form a three-dimensional multi-objective optimization design method. The final optimization result is that the Pin-Fin heat dissipation structure along the thermal accumulation direction has different diameters, transverse spacings and longitudinal spacings (applied to micro-channels is different channel width, depth and spacing, other heat dissipation structures can be analogized in this way). Such structure can make the corresponding chip have different junction flow thermal resistances, thereby compensating for the chip temperature non-uniformity caused by fluid thermal accumulation.

[0030] Through the above design, the present technology can effectively improve the working efficiency of the heat dissipation system, and effectively reduce the temperature non-uniformity of the power electronic converter. Under the same working conditions, the power electronic converter using the present design can improve the temperature non-uniformity problem, reduce the junction flow thermal resistance and fluid pressure drop, so that the reliability and service life of the power electronic converter are greatly improved.

[0031] Second, with the increase of voltage and current level, multi-chip parallel is usually used in medium and high voltage power modules. With the increase of the number of chips, the problem of uneven heat distribution in the power module is more and more obvious, which will greatly affect the reliability and life of the power module. The method and heat dissipation system in the application can significantly improve the problem of uneven heat distribution in the power module, and also reduce the junction heat resistance of the system, greatly improve the reliability and life of the system, thereby reducing the frequency and cost of system maintenance and updating in the long run.

[0032] Current domestic and foreign heat dissipation design focuses on reducing the highest junction temperature or average temperature, but ignores the key indicator of temperature distribution uniformity. The scheme adds temperature uniformity as an optimization target to the multi-objective optimization system of the heat dissipation system through ingenious design. This technology fills the technical gap in this field at home and abroad, and has important innovative significance.

[0033] Most of today's power modules use liquid cooling structure, but as the temperature of the cooling liquid rises, a certain temperature gradient will be generated inside the module. In extreme working conditions, the lowest chip temperature and the highest chip temperature will have a temperature difference of more than 20℃. Since the module life and reliability are often constrained by the chip with the highest junction temperature, when the module fails, it is likely that only the chip with the highest junction temperature has a problem, so the problem of uneven temperature is a problem that the industry has been trying to solve. The application solves the problem of uneven temperature distribution of chips in the power module to some extent, and can improve the service life and reliability of the power module. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1 is a multi-objective thermal optimization method flow chart of a SiC power electronic converter heat dissipation system provided by an embodiment of the application.

[0035] Figure 2 is a multi-objective thermal optimization system structure block diagram of a SiC power electronic converter heat dissipation system provided by an embodiment of the application.

[0036] Figure 3 is an optimized area division diagram of a heat sink according to the distribution of three half-bridge chips of a three-phase half-bridge module provided by an embodiment of the application.

[0037] Figure 4 is a key structural parameter diagram of a Pin-Fin heat dissipation system provided by an embodiment of the application.

[0038] Figure 5 is a three-dimensional multi-objective thermal optimization method flow chart of thermal resistance, pressure drop and thermal uniformity provided by an embodiment of the application.

[0039] Figure 6 is a vehicle HybridPACK TMDrive inverter module and its water cooling system diagram.

[0040] Figure 7 Pin-Fin heat dissipation base plate obtained through multi-objective optimization provided by the embodiment of the application.

[0041] Figure 8 Water cooling plate obtained through multi-objective optimization provided by the embodiment of the application.

[0042] Figure 9 Pin-Fin heat dissipation base plate and water cooling plate matched with the Pin-Fin heat dissipation base plate obtained through optimization provided by the embodiment of the application.

[0043] Figure 10 HybridPACK using a traditional Pin-Fin heat sink TM Thermal simulation distribution diagram of the Drive module.

[0044] Figure 11 HybridPACK using a Pin-Fin heat sink optimized by the method of the application TM Thermal simulation distribution diagram of the Drive module. DETAILED DESCRIPTION

[0045] In order to make the objectives, technical solutions and advantages of the application clearer and more apparent, the application will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are only used to explain the application and do not limit the application.

[0046] As shown in Figure 1 , the SiC power electronic converter heat dissipation system multi-objective thermal optimization method provided by the embodiment of the application includes the following steps:

[0047] S101, according to the distribution of chips on the three half bridges (1), (2) and (3), the pin column part of the Pin-Fin heat sink is divided into three parts (6), (7) and (8), which correspond to three optimization regions 1, 2 and 3, respectively;

[0048] S102, key structural parameters of the Pin-Fin heat sink and the liquid cooling channel are extracted;

[0049] S103, the optimization range of each key parameter is determined, the minimum and maximum limits of the parameters by the preparation process need to be considered, and the obviously ineffective value range is removed in combination with simulation and experience;

[0050] S104, a multi-objective optimization algorithm combined with a finite element simulation method is used to solve the above multi-objective problem.

[0051] The key structural parameters provided by the embodiment of the application:

[0052] The height of the needle column Hpi, the vertical distance between the needle columns Dvi, the horizontal distance between the needle columns Dhi, the radius of the needle column Ri, and the depth of the flow channel Hci.

[0053] The multi-objective optimization algorithm provided in this embodiment of the invention:

[0054]

[0055] The multi-objective optimization algorithm is responsible for generating and iterating the key parameters mentioned above. The finite element simulation performs thermal-fluid coupling simulation on half-bridge modules that do not use the same heat dissipation structure based on the key structural parameter group generated by the multi-objective optimization algorithm. The thermal resistance (Rthjf), voltage drop (ΔP), and thermal uniformity (ψ) obtained from the simulation are passed to the multi-objective optimization algorithm for iterating the structural parameter group. After a certain number of iterations, the Pareto front is obtained and a set of optimal structural parameters is found from it.

[0056] like Figure 2 As shown, the multi-objective thermal optimization system for a SiC power electronic converter heat dissipation system provided in this embodiment of the invention includes:

[0057] The partitioning module is used to divide the pin column part of the Pin-Fin heat sink into three parts (6), (7), and (8) according to the distribution of chips on the three half-bridges (1), (2), and (3), which correspond to three optimization regions 1, 2, and 3 respectively;

[0058] The extraction module is used to extract key structural parameters of the Pin-Fin heat sink and the liquid cooling channel;

[0059] The determination module is used to determine the optimization range of the above key parameters. It needs to consider the minimum and maximum limits of the manufacturing process on these parameters, and combine simulation and experience to eliminate obviously invalid value ranges.

[0060] The solver module is used to solve the above-mentioned multi-objective problem by combining multi-objective optimization algorithms with finite element simulation.

[0061] To address the heat dissipation efficiency bottleneck caused by uneven heat distribution and structural parameter coupling in existing SiC power electronic converter systems, this invention constructs a multi-objective thermal optimization system integrating a "partition-extraction-determination-solution" process. Traditional thermal design schemes are mostly based on single-objective or empirical parameter tuning methods, which are ill-suited to the non-uniform thermal stress concentration and cooling system load mismatch issues caused by SiC devices under high power density and high frequency operating conditions. Especially under multi-chip integrated packaging conditions, the spatial coupling between the cooling substructure and the heat source distribution renders conventional optimization methods ineffective.

[0062] In this invention, the partitioning module establishes a mapping relationship between the chip's heat source and the heat sink structure. Based on the spatial layout of the power devices on the three half-bridges, it divides the pin column area of ​​the pin-Fin heat sink into three independent optimized sub-regions. This partitioning method breaks away from the traditional uniform distribution design approach, allowing the heat sink design to move beyond symmetry and balanced distribution, and instead enhance cooling capacity locally as needed, thus enabling the possibility of structure-heat source synergistic optimization.

[0063] The extraction module establishes a set of feature vectors based on the geometric parameters of the Pin-Fin structure (such as pin height, spacing, and arrangement) and the dimensional parameters of the liquid cooling channel (such as channel width and inlet / outlet angle). Considering the nonlinear coupling between the thermal resistance network and the flow resistance, this module does not adopt a full parameter traversal strategy. Instead, it prioritizes extracting the parameter dimensions that have the greatest impact on thermal performance through sensitivity analysis, thereby significantly compressing the optimization space and improving the efficiency of subsequent solutions.

[0064] The module is responsible for establishing the effective optimization range for each parameter. This range is not bounded by theoretical maximum or minimum values, but rather dynamically tailored based on fabrication feasibility (such as minimum processing radius and maximum pressure drop limit) and simulation experience. For example, in certain high-density regions, to prevent excessive liquid cooling rates from causing drastic changes in the Reynolds number, the channel gap needs to be limited to a critical value. Furthermore, invalid or low-performance-sensitive parameters are eliminated to further improve the optimization quality.

[0065] The solution module integrates a multi-objective genetic algorithm and a finite element simulation engine, employing non-dominated sorting and Pareto front extraction mechanisms to evaluate the optimal structural combination under multiple objective functions (such as maximum junction temperature, pressure loss, and heat dissipation uniformity). By constructing a coupled thermal-fluid finite element model, it can evaluate the steady-state thermal distribution response of each candidate solution based on the initial structural design and dynamically feed it back to the population evolution process.

[0066] The entire optimization process is adaptive and has global optimization capabilities, making it particularly suitable for cooling scenarios of highly integrated power modules with complex structures and obvious conflicting objectives. Compared to traditional fixed-structure simulation and evaluation, this system can achieve "demand-oriented" customized structural configuration, improving both thermal design accuracy and manufacturing feasibility.

[0067] Therefore, this invention not only overcomes the limitations of existing thermal optimization methods in terms of structural flexibility and parameter dimension management, but also provides a systematic thermal management solution for SiC converters in high-reliability scenarios such as rail transit, electric drives, and aerospace power supplies, demonstrating the advanced nature of integrated material-structure-simulation thermal design. SiC is a wide-bandgap semiconductor material characterized by high temperature, high frequency, high voltage, and high power. Due to its superior thermal conductivity and high-temperature resistance, SiC is commonly used in high-power-density applications. Power electronic converters: Power electronic converters are devices that convert electrical energy from one form to another, typically involving adjustments to voltage, current, and frequency. Common types include AC / DC rectifiers, DC / DC converters, and DC / AC inverters. Heat dissipation systems: Heat dissipation systems are important systems designed to effectively dissipate the heat generated during device operation, thereby preventing overheating damage. Multi-objective thermal optimization methods: In heat dissipation system design, this involves simultaneously optimizing multiple conflicting objective parameters to find the optimal compromise.

[0068] This invention provides a multi-objective thermal optimization method for heat dissipation systems of SiC power electronic converters. This method can perform three-dimensional multi-objective optimization of the thermal resistance, voltage drop, and thermal uniformity of the heat dissipation system. The specific implementation principle and method are as follows: Figures 3-5 As shown.

[0069] Taking a three-phase half-bridge module with a Pin-Fin heat dissipation structure as an example (other heat dissipation structures can be compared similarly). First, based on the distribution of chips on the three half-bridges 1, 2, and 3, the pin column portion of the Pin-Fin heat sink 5 is divided into three parts 6, 7, and 8, corresponding to three optimization regions 1, 2, and 3 respectively (here, three regions are used as an example; it can also be divided into six or twelve regions). Then, the key structural parameters of the Pin-Fin heat sink 5 and the liquid cooling channel 11 are extracted. The main structural parameters extracted here are the pin column height Hpi, the vertical distance between pin columns Dvi, the horizontal distance between pin columns Dhi, the pin column radius Ri, and the channel depth Hci (it is worth mentioning that the synergistic optimization of the pin column height and channel depth is also an innovation of this invention). Next, the optimization range of the above key parameters is determined, especially considering the minimum and maximum limitations of the manufacturing process on these parameters. In addition, simulation and experience can be combined to eliminate obviously invalid value ranges. Finally, a multi-objective optimization algorithm (NSGA-II is used as an example here; other multi-objective optimization algorithms are also applicable, only the flowchart of the multi-objective optimization algorithm needs to be changed) combined with finite element simulation (COMSOL, ANSYS, etc.) is used to solve the above multi-objective problem. The multi-objective optimization algorithm is responsible for generating and iterating the above key parameters, while the finite element simulation performs thermal-fluid coupling simulation on half-bridge modules that do not use the same heat dissipation structure based on the key structural parameter group generated by the multi-objective optimization algorithm. The thermal resistance (Rthjf), voltage drop (ΔP), and thermal uniformity (ψ) obtained from the simulation are passed to the multi-objective optimization algorithm for iterating the structural parameter group. After a certain number of iterations, the Pareto front is obtained and a set of optimal structural parameters is found from it.

[0070] Formula for multi-objective optimization problems:

[0071]

[0072] Compared to existing multi-objective optimization methods, the multi-objective optimization method in this invention does not increase the complexity of the algorithm and does not make significant changes to the calculation and process, but it takes thermal uniformity into account. The heat dissipation structure obtained by this multi-objective optimization algorithm has superior thermal management performance, achieving a good balance between thermal resistance (Rthjf), voltage drop (ΔP), and thermal uniformity (ψ), and can be well applied to the application requirements of Si power electronic converters.

[0073] like Figure 6 The diagram illustrates a multi-objective thermal optimization method for the heat dissipation system of a SiC (or other semiconductor materials) power electronic converter, applied to an automotive hybrid pack using direct water cooling (or other cooling methods such as air cooling). TMApplication example of a Drive inverter module. From top to bottom, the inverter module consists of: SiC power electronic converter power substrates 13, 14, and 15; a pin-fin heat sink base 16; and a corresponding water-cooled plate 17. Both the heat sink base 16 and the corresponding water-cooled plate 17 are CNC machined and can be made of common heat dissipation materials such as Cu and AlSiC. An O-ring gasket is added between them in a sealing groove 19, and a bolt is used for pressure sealing. The cooling fluid is an ethylene glycol solution (water, deionized water, or other commonly used cooling fluids are also acceptable). The fluid inlet 18 and fluid outlet 29 are connected to external pipes via connectors and are connected to an external water cooler to provide fluid circulation and cooling.

[0074] The multi-objective optimization algorithm used in this case study is NSGA-II (but not limited to it). The optimization algorithm and data processing are implemented in MATLAB, while the finite element thermal flow simulation and geometric modeling are implemented in COMSOL Multiphysics. The two interact through COMSOL Multiphysics with Simulink.

[0075] Figures 7-9 The optimized Pin-Fin heat dissipation structure is shown. It can be seen that the three pin columns corresponding to the three-phase half-bridge have independent designs. They have different sizes, spacing and different pin column heights. In addition, the fluid channels inside the water-cooling plate 17 are not completely flat, but have different heights and are distributed in a stepped manner, just like the corresponding pin columns.

[0076] Numerical simulation analysis shows that the optimized Pin-Fin heat dissipation structure has better heat dissipation efficiency than the traditional Pin-Fin heat dissipation structure. Under the same chip loss conditions, the proposed scheme can reduce junction thermal resistance by about 5%, reduce fluid pressure drop by about 10%, and improve thermal uniformity by about 50%.

[0077] The heat source distribution law of power modules based on the three-phase half-bridge topology in SiC power electronic converters (e.g.) Figure 3 As shown, the Pin-Fin heat sink is divided into three independent optimization areas. This division closely follows the geometric layout of the three half-bridge chips 1, 2, and 3 within the package, which helps to decouple the thermal load in space, allowing each area to have its parameters set independently, thereby improving local thermal matching capability and overall heat dissipation performance.

[0078] like Figure 4 As shown, geometric parameters closely related to heat transfer performance are extracted from each optimization region, mainly including pin diameter, spacing, pin height, substrate thickness, base plate thickness, and channel width. Parameter extraction is not limited to the microstructure but also covers the overall heat transfer channel scale, ensuring the accuracy of multi-scale heat flow linkage modeling.

[0079] During the design space determination phase, CNC machining processes are introduced to impose feasibility constraints on geometric dimensions, such as minimum machining radius, structural stability limits for the pin height / diameter ratio, and minimum base plate wall thickness. At the same time, combined with empirical values ​​from fluid mechanics and heat transfer, parameter combinations that result in low heat flux or excessive pressure drop are eliminated to construct a convergent and physically reasonable design domain.

[0080] A complete three-dimensional finite element model was constructed using the COMSOL Multiphysics platform. Based on the fluid-structure interaction theory framework, the velocity and temperature field distributions of the coolant within the needle-like region were solved. The model includes the energy conservation equation, the Navier-Stokes equation, and boundary layer analysis. Key response indicators such as thermal resistance, pressure drop, and maximum temperature difference were evaluated through steady-state simulation.

[0081] like Figure 5 As shown, NSGA-II (Non-dominated sorting genetic algorithm II) is used for multi-objective optimization. The design variables are the aforementioned structural parameters, and the objective function is to minimize thermal resistance, pressure drop, and temperature uniformity factor. Through population initialization, elite retention, and crossover mutation operations, the Pareto front solution set is iteratively evolved, reflecting the nonlinear trade-off between structure and performance.

[0082] Figures 6 to 8 The results demonstrate the mapping of the optimization results to the actual radiator structure. It can be seen that the pin-fin arrays in different optimized regions exhibit differences in height, diameter, and arrangement density. Especially in areas with high heat load, the pins are denser and taller, enhancing local heat transfer capacity; while in areas with lower heat load, the arrangement tends to be sparser, reducing pressure drop.

[0083] Compared with traditional flat flow channels, Figure 8 The water-cooled plate's internal channels employ a stepped partition design that matches the three-zone pin columns, ensuring uniform coolant distribution across the pin-fin regions at different heights. This structure utilizes flow resistance adjustment to enhance the fluid flow's responsiveness to heat source locations, effectively suppressing cold spots and overheating.

[0084] like Figure 9 As shown, the optimized Pin-Fin base plate and water-cooled plate are bolted together, and O-rings are embedded in the sealing groove 19 at the edge to form a closed fluid chamber. Coolant is injected through inlet 18, and after heat exchange in each optimized area, it is discharged from outlet 29. An external circulation pump and cooler form a closed-loop cooling circuit to ensure the thermal control stability and engineering feasibility of the entire power module.

[0085] The specific application areas or related products of this invention.

[0086] 1. Automotive Hybrid PackTM Drive three-phase half-bridge inverter module

[0087] It is a power module used in electric vehicles (EVs) or hybrid electric vehicles (HEVs), primarily for driving the motor. It integrates a three-phase inverter function, efficiently converting DC power to three-phase AC power to control motor speed and torque. This module requires high power density, low loss, and high reliability to meet the stringent thermal management and vibration requirements of automobiles.

[0088] 2. Medium- and high-voltage, high-current multi-chip parallel SiC power modules

[0089] This type of module is suitable for industrial power supplies, electric vehicle charging stations, rail transit, and new energy systems. Based on silicon carbide (SiC) devices, it features high voltage withstand capability, high temperature, high frequency, and high efficiency. Through a multi-chip parallel design, the module can output large currents at medium to high voltage levels (such as 1700V and 3300V), improving system power density and reducing energy consumption.

[0090] Figure 10 It is a HybridPACK that uses a traditional Pin-Fin heatsink. TM Thermal simulation distribution diagram of the Drive module.

[0091] Figure 11 This is a HybridPACK for a Pin-Fin Heatsink optimized using the method of this invention. TM Thermal simulation distribution diagram of the Drive module.

[0092] Traditional pin-fin heat sinks exhibit a pronounced "island-like" temperature rise under the thermal load of a three-and-a-half bridge – the junction temperature peaks in the central areas of each power submodule are discrete, and the heat flow decays too rapidly towards the edge of the heat sink, resulting in a horizontal temperature gradient exceeding 35K. Heat must diffuse laterally through high thermal resistance paths, inducing circumferential shear stress concentration in the chip solder layer and further amplifying the cyclic fatigue risk caused by thermal coupling mismatch between the IGBT and SiC-Diode.

[0093] After dividing the needle column region into three parts according to the heat source topology, the column height-to-diameter ratio can be increased and the column spacing reduced in areas with insufficient peak convective heat transfer coefficient, while a sparser channel cross-section is retained in the weak heat zone at the edge. This maintains the overall fluid resistance at an acceptable level and enables the coolant to generate secondary flow and micro-vortices in the critical superheated zone, accelerating boundary layer regeneration. Finite element-CFD co-simulation shows that the corrected local Nusselt number peak-to-valley ratio decreased from 2.4 to 1.1, and the mean squared error of the temperature field decreased by nearly half.

[0094] On the heat sink substrate side, the planar diffusion resistance and interfacial contact thermal resistance in the heat conduction path are synergistically tuned by thinning the copper layer in the high heat flux density area and locally thickening the ceramic layer. This "flexible substrate-pin coupling" layout, together with the optimization of the upstream liquid cooling channel cross-section, reduces the overall thermal resistance from 0.128K / W to 0.121K / W. The reduction of approximately 5.4% translates into a 4–5K decrease in the peak junction temperature, and the pressure drop is reduced by nearly 10% compared to the original design at a flow rate of 400L h⁻¹.

[0095] Final thermal imaging and steady-state simulation comparison, for example Figure 10 and Figure 11 As shown, the peak temperature difference converged from 6.5K to 3.4K, and the junction temperature distribution surface became smoother; the thermal uniformity index improved by about 48%, while the fluid power consumption and radiator weight did not show significant increases, verifying the effectiveness of the partitioned-multi-objective coupling strategy in automotive-grade HybridPACK. TM Engineering effectiveness on the Drive module.

[0096] It should be noted that embodiments of the present invention can be implemented in hardware, software, or a combination of both. The hardware portion can be implemented using dedicated logic; the software portion can be stored in memory and executed by a suitable instruction execution system, such as a microprocessor or dedicated-design hardware. Those skilled in the art will understand that the above-described devices and methods can be implemented using computer-executable instructions and / or included in processor control code, for example, such code provided on a carrier medium such as a disk, CD, or DVD-ROM, a programmable memory such as read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The devices and modules of the present invention can be implemented by hardware circuitry such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field-programmable gate arrays, programmable logic devices, etc., or by software executed by various types of processors, or by a combination of the above-described hardware circuitry and software, such as firmware.

[0097] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions, and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention, and within the spirit and principles of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A Pin-Fin heat dissipation structure for a SiC power electronic converter power module, comprising: A power semiconductor module, wherein the power semiconductor module is a medium-voltage high-current module; A multiphase power semiconductor module, wherein the multiphase power semiconductor module is provided with multiple half-bridge chips; Pin-Fin heat dissipation base plate is disposed below the multiphase power semiconductor module. The Pin-Fin heat dissipation base plate includes multiple pin column arrays divided into multiple regions. Each pin column array includes multiple uniformly or non-uniformly arranged columnar heat dissipation structures. A water-cooled plate is disposed below the Pin-Fin heat dissipation base plate. The water-cooled plate has a flow channel step section that corresponds one-to-one with the pin column array. The height of the step section matches the height of the corresponding pin column of the pin column array. A sealing structure is provided in the sealing groove between the Pin-Fin heat dissipation base plate and the water-cooling plate; Interface components, used to connect fluid inlets and outlets; The multi-objective thermal optimization method for the heat dissipation system of the SiC power electronic converter, applying the aforementioned Pin-Fin heat dissipation structure of the power module, includes the following steps: Step 1: Based on the distribution of chips on the three half-bridges (1), (2), and (3), the pin column part of the Pin-Fin heat sink is divided into three parts (6), (7), and (8), which correspond to three optimization regions 1, 2, and 3, respectively. Step 2: Extract the key structural parameters of the Pin-Fin heat sink and the liquid cooling channel; Step 3: Determine the optimization range of the above key structural parameters. It is necessary to consider the minimum and maximum limits of the manufacturing process on these parameters, and eliminate obviously invalid value ranges by combining simulation and experience. Step 4: Solve the above multi-objective problem using a multi-objective optimization algorithm combined with finite element simulation. The key structural parameters are: The height of the needle column Hpi, the vertical distance between the needle columns Dvi, the horizontal distance between the needle columns Dhi, the radius of the needle column Ri, and the depth of the flow channel Hci; Needle height and flow channel depth are optimized in tandem.

2. The Pin-Fin heat dissipation structure for the SiC power electronic converter power module according to claim 1, wherein, The pin array is divided into three independent optimization regions, each corresponding to one of the three half-bridge chips of the three-phase half-bridge module.

3. The Pin-Fin heat dissipation structure for the SiC power electronic converter power module according to claim 1, wherein, The structural parameters of the needle array include needle diameter, needle spacing, and needle height, and these parameters are set with different values ​​in different optimization regions.

4. The Pin-Fin heat dissipation structure for the SiC power electronic converter power module according to claim 1, wherein, The internal flow channel of the water-cooled plate adopts a stepped structure, containing three flow regions of different heights, which correspond to the needle column array in the three optimized regions.

5. The Pin-Fin heat dissipation structure for the SiC power electronic converter power module according to claim 1, wherein, The sealing structure is an O-ring, which is embedded in a closed annular sealing groove below the Pin-Fin heat dissipation base plate.

6. The Pin-Fin heat dissipation structure for the SiC power electronic converter power module according to claim 1, wherein, Both the Pin-Fin heat dissipation base plate and the water-cooling plate are made of metal materials with a thermal conductivity higher than 150W / (m·K) and are integrally formed by CNC machining.

7. The Pin-Fin heat dissipation structure of the SiC power electronic converter power module as described in claim 1, characterized in that, The multi-objective optimization algorithm: ； The multi-objective optimization algorithm is responsible for generating and iterating the key structural parameters mentioned above. The finite element simulation performs thermal-fluid coupling simulation on half-bridge modules that do not use the same heat dissipation structure based on the key structural parameter group generated by the multi-objective optimization algorithm. The thermal resistance Rthjf, voltage drop ΔP, and thermal uniformity ψ obtained from the simulation are passed to the multi-objective optimization algorithm for iterating the structural parameter group. After a certain number of iterations, the Pareto front is obtained and a set of optimal structural parameters is found from it.

8. A multi-objective thermal optimization system for a SiC power electronic converter heat dissipation system implementing the Pin-Fin heat dissipation structure of the SiC power module as described in any one of claims 1-6, characterized in that, The multi-objective thermal optimization system for the SiC power electronic converter heat dissipation system includes: The module division is used to divide any heat dissipation system into two or more parts based on the distribution of chips on the module, each corresponding to multiple optimization areas; The extraction module is used to extract key structural parameters of any optimized heat dissipation system. The determination module is used to determine the optimization range of the above key structural parameters. It needs to consider the minimum and maximum limits of the manufacturing process on these parameters, and combine simulation and experience to eliminate obviously invalid value ranges. The solver module is used to solve the above multi-objective problem by combining multi-objective optimization algorithm with finite element simulation, so as to obtain a new heat dissipation system.

Images