Liquid cooling plate flow channel design optimization method and device based on topological optimization and liquid cooling plate

Abstract

The invention provides a liquid cooling plate flow channel design optimization method and device based on topological optimization and a liquid cooling plate, and relates to the technical field of thermal management, and the method comprises the steps: S1, determining the geometric parameters of the liquid cooling plate according to the size of a battery module, constructing a double-layer two-dimensional design domain, and setting material, heat source and boundary conditions; s2, a topological optimization algorithm is adopted, material distribution is represented with a relative density variable, a multi-objective function giving consideration to the maximum heat exchange amount and the minimum pressure drop is constructed, a clear fluid-solid interface is obtained through density filtering and a projection function, and a target double-layer two-dimensional flow channel contour is obtained through iterative solution under volume fraction constraint; and S3, taking the target double-layer two-dimensional flow channel contour as a cross section to carry out spatial stretching bending deformation, and generating a double-layer special-shaped flow channel structure with a diffusion layer and a collection layer. According to the method, global optimization of two-dimensional topological optimization and three-dimensional geometric reconstruction are fused, the contradiction between the heat dissipation performance and the flow resistance is effectively solved, and a high-performance heat dissipation solution is provided for battery thermal management.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of battery thermal management, and particularly relates to a liquid cooling plate flow channel design optimization method and device based on topological optimization and a liquid cooling plate. BACKGROUND

[0002] With the continuous consumption of non-renewable energy and the increasing environmental problems, lithium batteries, as representatives of efficient and clean energy, have become the main power source in the field of new energy vehicles. A large amount of heat is inevitably generated in the charging and discharging process of lithium batteries, and abnormal increase of working temperature not only significantly shortens the service life of the battery, but also may cause safety hazards such as thermal runaway. Therefore, an efficient and reliable thermal management system has become a key technical support for ensuring the safety of lithium battery energy storage and improving the performance of the battery.

[0003] Among various battery thermal management technologies, the cold plate liquid cooling technology has become the mainstream choice of the current battery thermal management system due to its high thermal conductivity and uniform heat dissipation characteristics. The basic principle of this technology is that the heat generated by the battery is transferred to the surface of the cold plate through conduction, the cooling liquid flows into the internal flow channel of the cold plate from the inlet, and the heat is taken away through convective heat exchange, and finally discharged from the outlet. At present, the research on the cold plate liquid cooling technology mainly focuses on the selection of cooling liquid type, the improvement of cold plate material and the design of flow channel structure, among which the design of flow channel structure is particularly important, directly affecting the heat dissipation efficiency and system energy consumption.

[0004] However, there are still some technical bottlenecks in the existing design of liquid cooling plate flow channel. Traditional flow channels mostly adopt single-layer, straight or regular arrangement structure, such as rectangular or H-shaped flow channels, which have single structure and large flow resistance, resulting in high system pressure drop, increased pump power consumption and limited overall heat dissipation efficiency. In addition, the cooling liquid in the single-layer flow channel often flows in one direction, and the convective heat exchange direction is fixed, which is difficult to flexibly adapt to the actual distribution of high-temperature areas in the battery module, and may cause uneven temperature distribution on the surface of the cold plate, thereby affecting the temperature consistency of the battery pack. Especially in high-energy-density battery modules, the problem of local heat accumulation is more prominent, and the traditional flow channel structure is difficult to achieve precise heat dissipation of high-temperature areas, limiting the further improvement of the overall performance of the liquid cooling system. SUMMARY

[0005] Therefore, the present application proposes a liquid cooling plate flow channel design optimization method and device based on topological optimization to solve the technical problems of low heat dissipation efficiency, poor temperature uniformity and high system energy consumption caused by the structural limitations of the existing liquid cooling plate flow channel.

[0006] The technical scheme of the present application is implemented as follows: In a first aspect, the present application discloses a liquid cooling plate flow channel design optimization method based on topological optimization, comprising the following steps: S1, obtain the geometric parameters of the liquid cooling plate, construct a double-layer two-dimensional design domain of the liquid cooling plate in a simulation environment, and determine the inlet and outlet boundary conditions of the double-layer two-dimensional design domain; S2, define a relative density variable of the filling material, fill the double-layer two-dimensional design domain, obtain a two-dimensional initial flow channel profile of the liquid cooling plate, construct a multi-objective function, and perform topology optimization on the two-dimensional initial flow channel profile to balance the maximization of heat exchange and the minimization of flow channel pressure drop; S3, perform density filtering on the relative density variable through a density filtering function, and process the filtered relative density variable through a projection function to obtain a clear relative density variable, so as to eliminate intermediate density values and obtain a clear fluid-solid interface; S4, introduce a flow channel volume fraction constraint condition, bring the clear relative density variable into the multi-objective function for iterative solution, and obtain a target double-layer two-dimensional topology flow channel profile; S5, based on the target double-layer two-dimensional topology flow channel profile, generate a collection flow channel layer entity and a diffusion flow channel layer entity through a parameterized stretching operation; perform scanning deformation on the entity thickness section based on a pre-defined bending trajectory to generate a special-shaped collection flow channel layer entity and a special-shaped diffusion flow channel layer entity; and generate a connection flow channel body entity based on a pre-defined connection rule, integrate the special-shaped collection flow channel layer entity and the special-shaped diffusion flow channel layer entity to form a double-layer special-shaped flow channel structure entity; S6, based on the generated double-layer special-shaped flow channel structure entity, construct a solid liquid cooling plate three-dimensional model, remove the overlapping part by using a Boolean subtraction operation, and generate a liquid cooling plate with a double-layer special-shaped flow channel structure.

[0007] On the basis of the above technical scheme, preferably, the boundary condition of one of the two-dimensional design domains adopts a collection flow mode of a four-corner inlet and a center outlet, and the boundary condition of the other of the two-dimensional design domains adopts a diffusion flow mode of a center inlet and a four-corner outlet.

[0008] On the basis of the above technical scheme, preferably, the step of defining the relative density variable γ of the filling material comprises: determining a material permeability attribute based on the relative density variable γ through an interpolation function as follows: wherein, α ( gamma ) is an actual permeability, α f is a fluid permeability, α s is a solid permeability.

[0009] On the basis of the above technical scheme, preferably, the multi-objective function is a weighted sum of a heat exchange single-objective function and a fluid power consumption single-objective function, and depends on the defined relative density variable γ, and an expression of the multi-objective function is: wherein, J is a multi-objective function, represents a heat exchange amount single objective function value at time t, represents a fluid power consumption single objective function value at time t, w 1 is a heat exchange amount objective function weight factor, w 2 is an energy consumption objective function weight factor, t is an iteration time variable, Y is a main objective value coefficient, X is an error compensation coefficient; wherein, is a heat exchange amount single objective function, Omega is a two-dimensional design domain, dOmega is an area integral infinitesimal on the design domain, T is a reference temperature, T in is an inlet temperature, H n is a wall heat transfer coefficient; wherein, is an energy consumption single objective function, gamma_i is a clear relative density variable, Re is a Reynolds number, u is an inlet flow velocity, rho f is a fluid density, D h is a hydraulic diameter of a liquid flow inlet, 2 is a gradient operator, ∇²u² represents a square term of a Laplacian operator of a velocity field, α ( gamma_i ) is a permeability of the clear relative density variable gamma_i , represents viscous power consumption due to a velocity gradient, represents local pressure drop power consumption due to a flow channel structure.

[0010] On the basis of the above technical solution, preferably, the flow channel volume fraction constraint condition expression is: wherein, v f is an upper limit of the fluid volume fraction, V Ω is a double-layer two-dimensional design domain area.

[0011] On the basis of the above technical solution, preferably, the step of performing density filtering on the relative density variable γ by using the density filtering function comprises: The relative density variable γ is filtered by solving the Helmholtz equation to obtain the filtered relative density variable. The expression for the Helmholtz equation is: ,in, R The filter radius; The filtered relative density variable is processed through a projection function. The processing steps include: The filtered relative density variable The input hyperbolic tangent projection function is processed to obtain the sharp relative density variable after projection, which is used for subsequent iterative updates. gamma_i The expression for the projection function is: ,in, β The projection slope, gamma β This is the projection threshold.

[0012] Based on the above technical solution, preferably, the bending trajectory is a parabola, and the expression for the parabola is: ,in, y Let the vertical axis be the coordinate origin at the center of the thickness section. x Let the x-axis be the coordinate system with the center of the thickness section as the origin. b This is the distance from the center of the thickness section to the wall of the parabolic flow channel; p Let be the direction of the parabola's opening, when p When >0, it deforms upwards; when p When <0, it deforms downwards; This represents the coefficient for the size of the parabola's opening.

[0013] Based on the above technical solution, preferably, the double-layer irregular flow channel structure includes a converging flow channel layer and a diffusion flow channel layer, wherein the converging flow channel layer adopts a radial main channel and the diffusion flow channel layer adopts a mesh-like straight branch.

[0014] Secondly, this application discloses a liquid-cooled plate flow channel design optimization device based on topology optimization, comprising: The design domain construction module is used to obtain the geometric parameters of the liquid cooling plate, construct a two-dimensional design domain of the liquid cooling plate in the simulation environment, and determine the inlet and outlet boundary conditions of the two-dimensional design domain. The topology optimization module is used to define the relative density variable of the filling material, fill the double-layer two-dimensional design domain, obtain the two-dimensional initial flow channel profile of the liquid cooling plate, construct a multi-objective function, and perform topology optimization on the two-dimensional initial flow channel profile to achieve a balance between maximizing heat transfer and minimizing flow channel pressure drop. The density processing module is used to filter the relative density variable through a density filtering function and process the filtered relative density variable through a projection function to obtain a clear relative density variable, so as to eliminate intermediate density values ​​and obtain a clear fluid-solid interface. The iterative solution module is used to introduce the flow channel volume fraction constraint condition, and to substitute the clear relative density variable into the multi-objective function for iterative solution to obtain the target two-layer two-dimensional topological flow channel profile; The 3D reconstruction module, based on the target double-layer 2D topological flow channel contour, generates a converging flow channel layer entity and a diffusion flow channel layer entity through parametric stretching operations; it performs scanning deformation on the entity thickness section based on a predefined bending trajectory to generate an irregularly shaped converging flow channel layer entity and an irregularly shaped diffusion flow channel layer entity; and it generates a connecting flow channel body entity based on predefined connection rules, integrating the irregularly shaped converging flow channel layer entity and the irregularly shaped diffusion flow channel layer entity to form a double-layer irregular flow channel structure entity; The model generation module is used to construct a three-dimensional model of a solid liquid cooling plate based on the generated double-layer irregular flow channel structure entity, and to remove overlapping parts by using Boolean subtraction operation to generate a liquid cooling plate with a double-layer irregular flow channel structure.

[0015] Thirdly, this application discloses a liquid cooling plate whose internal flow channel is implemented using the liquid cooling plate flow channel design optimization method based on topology optimization as described in the first aspect.

[0016] This application has the following advantages over the prior art: (1) The liquid cooling plate flow channel design optimization method disclosed in this application combines the global optimization capability of two-dimensional topology optimization with the geometric innovation of three-dimensional reconstruction to produce a high-performance double-layer irregular flow channel structure. This method not only fundamentally solves the contradiction between heat dissipation and flow resistance in traditional design and achieves global optimal performance, but also significantly improves design efficiency and quality through automated processes, providing a breakthrough heat dissipation solution for battery thermal management systems.

[0017] (2) The diffusion flow mode adopts a symmetrical arrangement of a single central inlet and multiple corner outlets. After the coolant is injected from the center, it diffuses radially to the surrounding area. Forced convection is generated through the inlet jet effect. The flow velocity is reduced by gradually expanding the flow cross section, realizing the conversion of dynamic pressure to static pressure. The convergence flow mode adopts a symmetrical arrangement of multiple corner inlets and a single central outlet. After the coolant is injected from the four corners, it converges to the central area. Centripetal flow is formed by balanced liquid distribution through multiple inlets, and turbulent mixing effect is generated in the central area. Both modes eliminate the problem of uneven velocity distribution in traditional parallel channels through symmetrical arrangement.

[0018] (3) By limiting the material property interpolation method in the topology optimization process, the smooth transition of material properties is realized. The numerical instability that may occur in the optimization process is avoided by using a continuously differentiable interpolation function, which significantly improves the convergence and computational stability of the optimization algorithm.

[0019] (4) By defining a multi-objective function, the two conflicting objectives of maximizing heat transfer and minimizing fluid power consumption are intelligently integrated into a weighted sum form, thereby achieving an optimized balance in the design of liquid cooling plate flow channels. This function depends on the relative density variable and enhances numerical stability and error handling by dynamically adjusting the weighting factors and coefficients. It guides the topology optimization algorithm to automatically generate high-performance flow channel layouts, significantly improves heat dissipation efficiency, reduces flow resistance, and ensures temperature uniformity. At the same time, it improves design flexibility and automation, shortens the design cycle, and ensures clear fluid-solid interface and manufacturability.

[0020] (5) By processing the relative density variable through density filtering and projection function, the numerical instability in topology optimization, such as checkerboard phenomenon and intermediate density value, is effectively eliminated, and a clear fluid-solid interface is generated, thereby significantly improving the convergence speed, numerical stability and engineering feasibility of the algorithm; ensuring that the optimization results can be directly used for manufacturing, reducing the need for post-processing, and enhancing the performance reliability and manufacturing efficiency of the flow channel design, providing key digital foundation support for the overall liquid cooling plate flow channel design method.

[0021] (6) The equation of the parabola is obtained by dynamically adjusting the opening size coefficient. This mechanism enables intelligent response of the curved trajectory to changes in multiple objective functions, achieving automatic optimization of the flow channel shape: when the opening decreases, heat transfer is enhanced while energy consumption is reduced to achieve a performance balance; when the opening increases, flow resistance is reduced first, but heat dissipation is moderately sacrificed to ensure design adaptability. This mechanism significantly improves the heat dissipation efficiency, flow performance, and manufacturing feasibility of the liquid cooling plate, supports the efficient and reliable operation of the battery thermal management system, and shortens the design cycle while enhancing the overall automation level of the approach.

[0022] (7) By limiting the diffusion channel layer to a mesh-like straight branch and the converging channel layer to a radial main channel in the double-layer irregular flow channel structure, a systematic improvement in thermal management efficiency is achieved. This design innovatively utilizes the high coverage characteristics of the mesh-like straight branch to effectively eliminate heat dissipation blind spots at the end and edge areas of the battery module, ensuring uniform distribution of coolant and thus significantly enhancing temperature uniformity. At the same time, the radial main channel optimizes the heat collection efficiency and reduces the flow path length through radial convergence flow, avoiding the problem of temperature difference increase under low-speed flow conditions. Compared with traditional rectangular or I-shaped flow channels, this structure works in synergy, not only solving the problem of increased energy consumption caused by branch pushing under high-speed flow conditions, but also reducing the overall flow resistance and improving heat dissipation efficiency through a multi-level branch synergistic heat dissipation mechanism. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart of the liquid cooling plate flow channel design optimization method disclosed in this application; Figure 2 This is a schematic diagram of the two-layer two-dimensional topology design domain disclosed in this application; Figure 3 This is a schematic diagram of the outline of the topology-optimized two-dimensional flow channel structure. Figure 4 To generate the converging flow channel layer entity and the diffusion flow channel layer entity through the stretching operation; Figure 5 A schematic diagram of a solid model for a double-layer irregular flow channel structure; Figure 6 Schematic diagram of a double-layer irregular flow channel structure in a liquid cooling plate; Figure 7 This is a cloud map showing the battery temperature distribution corresponding to different flow channel structures; among them... Figure 7 Figure a in the diagram is a cloud map of the battery temperature distribution corresponding to the double-layer rectangular flow channel; Figure 7 Figure b in the figure is a cloud map of the battery temperature distribution corresponding to the double-layer I-shaped flow channel; Figure 7 Figure c in the figure is a cloud map of the battery temperature distribution corresponding to the double-layer anisotropic flow channel; Figure label: 1. Irregularly shaped collecting channel layer entity; 2. Connecting channel body entity; 3. Irregularly shaped diffusion channel layer entity; 4. Collecting channel layer entity; 5. Collecting channel layer; 6. Diffusion channel layer. Detailed Implementation

[0025] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0026] See attached document Figure 1 As shown in the figure, this application discloses a liquid cooling plate flow channel design optimization method based on topology optimization, including the following steps: Step S1: Obtain the geometric parameters of the liquid cooling plate, construct a two-dimensional design domain of the liquid cooling plate in the simulation environment, and determine the inlet and outlet boundary conditions of the two-dimensional design domain.

[0027] This step is the initialization phase of the optimization process, providing basic geometric and physical constraints for topology optimization. Based on the actual geometric dimensions of the battery module (e.g., length, width, and height), the geometric parameters of the liquid cooling plate are obtained. For example, a liquid cooling plate with the same length and width as the bottom surface of a battery module with dimensions of 200mm*300mm is placed to completely cover the bottom surface of the battery module. A precise two-layer two-dimensional planar design domain is constructed in a simulation environment (e.g., finite element software such as COMSOL or ANSYS). This design domain represents the effective heat dissipation area of ​​the liquid cooling plate. The design domain is discretized into a finite number of mesh elements to facilitate subsequent topology optimization. Simultaneously, this step also needs to determine the inlet and outlet boundary conditions of the two-layer two-dimensional design domain. This is to define the flow path of the coolant, thereby simulating a real thermal management scenario and avoiding blind optimization. The construction of the two-dimensional design domains facilitates the subsequent design of two-dimensional channels in the liquid cooling plate. In this embodiment, the inlet and outlet positions of the two-dimensional design domains are set differently. This is to enable the subsequent design of the two-dimensional channels to establish a coolant flow relationship. For example, the coolant enters from the inlet of one channel and the outlet is connected to the inlet of the other channel, so that the cooling medium enters from one channel of the liquid cooling plate and flows out from the other channel, achieving heat dissipation and temperature uniformity.

[0028] Step S1 involves transforming a complex engineering problem into a computable mathematical model. By constructing a two-dimensional design domain, computational complexity is significantly reduced, avoiding the enormous computational resource consumption associated with direct three-dimensional optimization. Furthermore, precise boundary condition settings ensure that the optimization process is conducted under real-world conditions, guaranteeing the physical feasibility and engineering practicality of the final design from the outset, and laying a reliable foundation for subsequent optimizations.

[0029] Step S2: Define the relative density variable of the filling material, fill the two-dimensional design domain to obtain the two-dimensional initial flow channel profile of the liquid cooling plate, construct a multi-objective function, and perform topology optimization on the two-dimensional initial flow channel profile to achieve a balance between maximizing heat transfer and minimizing flow channel pressure drop.

[0030] A relative density variable γ is defined to characterize the material distribution, with a value range of [0, 1], where 0 represents solid material and 1 represents fluid material. A "filling" operation is performed on the two-dimensional design domain to generate an initial two-dimensional flow channel profile. A multi-objective function dependent on the relative density variable γ is constructed to achieve a balance between maximizing heat transfer and minimizing flow channel pressure drop. This function simultaneously pursues two often conflicting objectives: maximizing heat dissipation efficiency (maximizing heat transfer) and minimizing flow resistance (minimizing pressure drop). The algorithm automatically adjusts the design variable γ to find the optimal material distribution scheme that simultaneously satisfies both objectives throughout the entire design domain. This method overcomes the limitations of traditional empirical design, automatically and systematically finding the globally optimal balance point between heat dissipation performance and flow resistance characteristics, resolving the contradiction between heat dissipation and flow resistance in traditional design without relying on empirical trial and error.

[0031] S3. The relative density variable is filtered by the density filtering function, and the filtered relative density variable is processed by the projection function to obtain a clear relative density variable, so as to eliminate intermediate density values ​​and obtain a clear fluid-solid interface.

[0032] The raw results directly generated by topology optimization often exhibit numerical instability, such as grayscale cells (intermediate densities that are neither 0 nor 1) and checkerboard patterns, making them unsuitable for direct manufacturing. A density filtering function is used to smooth the relative density variable γ, eliminating common checkerboard patterns and intermediate density values ​​encountered during optimization. Then, a projection function (such as the hyperbolic tangent function) is used to binarize the filtered variable (to 0 or 1), generating a clear relative density variable. gamma_i The purpose of this step is to obtain a clear fluid-structure interface, ensuring the engineering feasibility of the optimization results and directly outputting a flow channel profile with clear boundaries that can be used for subsequent CAD modeling and manufacturing. It avoids performance uncertainties caused by grayscale cells, demonstrating the crucial role of mathematical regularization in topology optimization and enhancing the convergence and reliability of the algorithm.

[0033] S4. Introduce flow channel volume fraction constraints to define the clear relative density variable. gamma_i The solution is substituted into the multi-objective function for iterative solving (e.g., using gradient descent). Constraints ensure that the flow channel volume does not exceed a preset limit, and the iterative process automatically adjusts... gamma_i The objective function is optimized to ultimately output the target two-layer two-dimensional topology flow channel profile.

[0034] The purpose of this step is to obtain feasible and high-performance optimization results under physical constraints, ensuring that the flow channel layout meets both heat dissipation requirements and engineering realities (such as material usage control), and to reduce manual intervention through automated iteration.

[0035] Steps S1-S4 automate and optimize the flow channel layout. This breaks away from the traditional trial-and-error design model that relies on engineer experience, automatically exploring countless possible layouts within the design space to find the optimal global balance between heat dissipation performance and flow resistance. Simultaneously, through filtering and projection processing, it ensures the mathematical stability and manufacturability of the optimization results, directly outputting a clear-boundary flow channel geometry that requires no post-processing, significantly improving design efficiency and quality.

[0036] Step S5: Based on the target double-layer two-dimensional topological flow channel contour, generate a converging flow channel layer entity and a diffusion flow channel layer entity through parametric stretching operation; perform scanning deformation on the thickness section of the entity based on a predefined bending trajectory to generate an irregular converging flow channel layer entity and an irregular diffusion flow channel layer entity; and generate a connecting flow channel body entity based on a predefined connection rule, and integrate the irregular converging flow channel layer entity and the irregular diffusion flow channel layer entity to form a double-layer irregular flow channel structure entity.

[0037] Step S5 aims to transform the two-dimensional topological flow channel profile obtained in Step S4 (from topology optimization) into a three-dimensional solid model, thereby creating a high-performance, two-layer irregular flow channel structure. Its core objective is to achieve automated conversion from two-dimensional to three-dimensional, generating a solid model representing the flow channel cavity (i.e., the geometry of the "fluid portion"), providing input for subsequent Boolean subtraction operations. Through parametric stretching, sweep deformation, and integrated connectors, S5 ensures the geometric complexity of the flow channel structure (such as irregular bends) to enhance flow uniformity and heat dissipation efficiency.

[0038] For details, please refer to the appendix. Figure 4 and 5 As shown, the two-dimensional topological flow channel profile obtained in step S4 is used as a two-dimensional sketch. This sketch is then stretched in three-dimensional space. Since it is a double-layered two-dimensional topological flow channel profile, each layer is stretched separately to obtain the collecting flow channel layer entity and the diffusion flow channel layer entity. Because the flow channel profile is a closed curve, after stretching to a certain thickness in three dimensions, a flow channel layer entity is formed. This flow channel layer entity can be understood as a solid entity formed by fluid filling the flow channel space. The stretching thickness is usually based on flow channel design parameters (such as hydraulic diameter) to ensure that the entity represents the geometric boundary of the flow channel cavity. The collecting flow channel layer entity and the diffusion flow channel layer entity are generated to achieve efficient collection (collecting layer) and uniform distribution of coolant (diffusion layer). By scanning and deforming the solid cross-section based on a predefined bending trajectory, irregularly shaped converging and diffusing flow channel layers are generated. These irregularly shaped converging and diffusing flow channel layers, obtained in this step, can reduce flow resistance, eliminate hot spots, and increase heat transfer area after corresponding flow channel cavities are generated in the subsequent liquid cooling plate. Simultaneously, the irregular design improves the convective heat transfer coefficient by disturbing the flow.

[0039] The irregularly shaped collecting channel layer and the irregularly shaped diffusing channel layer are a double-layer solid structure. These two solid structures are not connected; they serve as the basis for subsequently forming a double-layer flow channel cavity in the solid liquid cooling plate. However, within the liquid cooling plate, these two flow channel cavities need to be connected to allow coolant flow. Therefore, this embodiment generates a connecting channel body entity based on predefined connection rules, integrating the irregularly shaped collecting channel layer entity and the irregularly shaped diffusing channel layer entity to form a double-layer irregularly shaped flow channel structure entity. In this way, the connecting channel body entity in the double-layer irregularly shaped flow channel structure entity can form a connecting cavity in the liquid cooling plate, thereby connecting the two flow channel cavities and ensuring coolant flow between the two layers.

[0040] Specifically, the irregularly shaped collecting flow channel layer and the irregularly shaped diffusing flow channel layer can be connected by stretching to form a cylindrical connecting flow channel body, thereby forming a double-layer irregularly shaped flow channel structure.

[0041] S6. Based on the generated double-layer irregular flow channel structure entity, a solid liquid cooling plate 3D model is constructed using 3D software (such as SolidWorks). Boolean subtraction operation is used to remove overlapping parts (i.e., subtract the flow channel entity from the solid plate) to generate a liquid cooling plate with a double-layer irregular flow channel structure.

[0042] This step is the final stage of manufacturing preparation, designed to integrate the double-layer irregular flow channel structure into the solid liquid cooling plate, generating a model ready for production. Through Boolean subtraction, the flow channel portion is removed from the solid plate, forming a flow channel cavity and ensuring the liquid cooling plate possesses the required internal flow channel structure.

[0043] Step S6 generates a dual-layer irregular flow channel structure comprising a diffusion channel layer and a collection channel layer. This structure enables multi-directional and coordinated flow of coolant within the flow channel, greatly increasing the heat exchange area and reducing disturbances. It effectively eliminates local hot spots and heat dissipation blind spots in the battery module, thereby significantly improving the temperature uniformity and heat dissipation efficiency of the entire battery pack.

[0044] In this embodiment, step S5 is set to bridge the two-dimensional optimization results and the generation of the three-dimensional entity, solving the following key problems: 2D to 3D Conversion: Topology optimization (S1-S4) outputs a 2D contour, but the actual liquid cooling plate is a 3D structure. Step S5 transforms the 2D data into a 3D solid through stretching and deformation, enabling the optimization results to be applied in 3D space and ensuring that performance is not lost due to changes in dimensionality.

[0045] Creation of irregular structures: S5 introduces scan deformation based on predefined trajectories (such as parabolas), which allows the generation of complex irregular flow channels (such as bends or twists). These structures cannot be fully optimized in 2D, but can significantly improve performance in 3D (such as increasing turbulence and heat transfer).

[0046] Channel integration requirements: The collecting layer and the diffusion layer, as separate entities, need to be integrated through a connector to form a connected channel. The integrated operation of S5 ensures that the coolant can flow in these two layers, avoiding dead zones and achieving complete thermal management functions.

[0047] Completeness of Automated Processes: S5 is a key link in the automation chain, ensuring that the entire process from optimization to manufacturing preparation requires no human intervention, improving design efficiency and consistency. Without S5, 3D reconstruction would rely on manual CAD operations, introducing errors and delays.

[0048] In this embodiment, steps S5 and S6 have a close dependency and synergy, which together achieve a seamless transition from design to manufacturing: Data flow dependency: The output of S5 (the dual-layer irregular flow channel structure entity) is directly used as the input of S6. The Boolean subtraction operation of S6 depends on the entity generated by S5 to define the region to be removed. Without S5, S6 cannot be executed; conversely, without S6, the entity of S5 cannot be integrated into the liquid cooling plate.

[0049] Complementary functions: S5 is responsible for creating the flow channel geometry (representing the cavity), while S6 is responsible for integrating the cavity into the solid plate. S5 focuses on flow channel performance optimization (e.g., through irregular designs), while S6 focuses on manufacturing feasibility (e.g., through Boolean subtraction to form the cavity). Together, they ensure the design is both high-performance and manufacturable.

[0050] The liquid cooling plate flow channel design optimization method disclosed in this application combines the global optimization capability of two-dimensional topology optimization with the geometric innovation of three-dimensional reconstruction to produce a high-performance double-layer irregular flow channel structure. This method not only fundamentally solves the contradiction between heat dissipation and flow resistance in traditional designs, achieving global optimal performance, but also significantly improves design efficiency and quality through automated processes, providing a breakthrough heat dissipation solution for battery thermal management systems.

[0051] As some implementation methods, see the appendix. Figure 2 As shown, the boundary conditions of one layer of the two-dimensional design domain adopt a converging flow pattern with four corner inlets and a central outlet, while the boundary conditions of the other layer of the two-dimensional design domain adopt a diffusion flow pattern with a central inlet and four corner outlets.

[0052] Among them, the convergence flow mode adopts a symmetrical arrangement of multiple inlets at the four corners and a single outlet at the center. After the coolant is injected from the four corners, it converges and flows towards the central area. The centripetal flow is formed by the balanced distribution of liquid through multiple inlets, and a turbulent mixing effect is generated in the central area. The diffusion flow mode adopts a symmetrical arrangement of a single inlet at the center and multiple outlets at the four corners. After the coolant is injected from the center, it diffuses and flows radially to the surrounding areas. Forced convection is generated through the inlet jet effect. The flow velocity is reduced by the gradually expanding flow cross section, and the dynamic pressure is converted into static pressure.

[0053] Both of the above modes eliminate the problem of uneven velocity distribution in traditional parallel channels through symmetrical arrangement. The diffusion mode reduces dynamic pressure loss by reducing velocity through gradually expanding cross section, while the convergence mode reduces local pressure drop by splitting the flow through multiple inlets. In addition, the radial / centripetal flow paths completely eliminate the flow dead zones at the corners of rectangular channels.

[0054] In liquid cooling plate design, the collecting channel layer and the diffusion channel layer perform different functions: the collecting channel layer is responsible for collecting hot liquid (through a flow pattern from the four corner inlets to the center outlet), while the diffusion channel layer is responsible for uniformly distributing coolant (through a flow pattern from the center inlet to the four corner outlets). If these two layers are isolated, the coolant cannot flow between the layers, leading to flow dead zones, heat accumulation, and reduced efficiency. The outlet of the collecting channel layer needs to be connected to the inlet of the diffusion channel layer to ensure continuous coolant flow, thereby optimizing heat dissipation performance and flow efficiency. This connectivity provides a direct basis for setting the connecting channel body entity in step S5, ultimately achieving seamless integration of the two channel structures. The center outlet of the collecting layer naturally corresponds to the center inlet of the diffusion layer, and the connecting channel body entity is automatically generated based on these coordinates without manual intervention.

[0055] The specific boundary conditions set in this embodiment provide a clear optimization direction for subsequent topology optimization algorithms, control the flow characteristics from the source, and ensure that the flow channel layout generated by topology optimization is engineering feasible. At the same time, the boundary conditions in this embodiment significantly improve the convergence speed and stability of the topology optimization process.

[0056] This application also achieves a continuous transition of material permeability through a specific mathematical interpolation function, providing a key physical property calculation basis for topology optimization algorithms.

[0057] Specifically, based on the relative density variable γ, the material permeability property is determined using the following interpolation function: ,in, α ( gamma (This represents the actual penetration rate.) α f For fluid permeability, α s This represents the solid permeability.

[0058] The above mathematical expression establishes a quantitative mapping relationship between relative density and material physical properties. The relative density γ, as a design variable, changes dynamically during the optimization process, taking values ​​in the range [0,1], where 0 represents the pure solid region and 1 represents the pure fluid region. Fluid permeability... α f and solid permeability α s The physical properties of the two extreme states are represented respectively, and a continuous transition to the intermediate state is achieved through linear interpolation. This interpolation method essentially builds a bridge between material properties and design variables, enabling the optimization algorithm to continuously change the physical properties of local regions by adjusting the γ value, thus providing a mathematical basis for the synergistic optimization of fluid-structure interaction fields.

[0059] By restricting the material property interpolation method in the topology optimization process, a smooth transition of material properties is achieved. The use of a continuously differentiable interpolation function avoids potential numerical instability during optimization, significantly improving the convergence and computational stability of the optimization algorithm. Secondly, the interpolation method has clear physical meaning: when γ approaches 0, permeability approaches solid properties; when γ approaches 1, permeability approaches fluid properties. This clear physical correspondence ensures the engineering feasibility of the optimization results. More importantly, this method provides an effective regularization tool for topology optimization. By controlling the permeability variation in the intermediate density region, it guides the optimization process towards a 0-1 binary result, ultimately generating a manufacturable flow channel profile with a clear fluid-solid interface. This mathematical rigor lays a reliable foundation for subsequent 3D reconstruction, ensuring the coherence and feasibility of the entire technical chain from optimization algorithm to solid fabrication.

[0060] As one of the implementation methods, this embodiment defines a multi-objective function in the topology optimization process, providing a rigorous mathematical framework for achieving synergistic optimization of heat dissipation performance and flow resistance characteristics.

[0061] Specifically, the multi-objective function is a weighted sum of the heat exchange single-objective function and the fluid power consumption single-objective function, and depends on the defined relative density variable γ. The expression of the multi-objective function is as follows: ,in, J It is a multi-objective function. Y The coefficients for the primary objective value are used for weighting. and The synergistic component enhances the weight of key objectives (such as heat exchange), typically taking positive values; the larger the value, the more emphasis is placed on heat dissipation performance.

[0062] w 1 The weighting factor for the heat exchange objective function is used to adjust... The higher the value, the greater the priority in maximizing heat exchange.

[0063] w 2 These are weighting factors for the energy consumption objective function, used to adjust... The importance of the objective is indicated by a higher value, which means that minimizing fluid power consumption is given greater priority.

[0064] This represents the single-objective function value of heat transfer at time t, quantifying heat dissipation performance; This represents the single-objective function value of fluid power consumption at time t, quantifying the flow resistance. t This is the iteration time variable, corresponding to the iteration step size or simulation time of the optimization algorithm, used to capture the dynamic optimization effect.

[0065] This expression is expressed in summation integral form (from...) t =0 to t ) integrated and The weighted combination. Xiang emphasized and Synergistic enhancement (additive part). X The terms handle conflict balancing (subtraction) between them, thereby achieving global optimization during the iteration process. Weights w 1 and w 2 Adjusted automatically or by preset algorithms to explore the Pareto frontier.

[0066] ,in, For heat exchange, a single objective function, Omega For a two-dimensional design domain, dOmega Let the area be a differential element over the design domain. T For reference temperature, take the average temperature of the flow channel wall or area, in K or °C; T in Inlet temperature, the temperature of the coolant when it enters the flow channel, in K or °C; H n γ is the wall heat transfer coefficient, expressed in W / (m²·K), which characterizes the convective heat transfer efficiency between the coolant and the channel wall. γ∈[0,1], where 0 represents solid material and 1 represents the channel region. (1-γ) represents the proportion of solid material, because the solid region participates in heat transfer, while the channel region (γ=1) contributes less. This indicates the temperature difference.

[0067] single objective function for heat exchange The convective heat transfer between the coolant and the flow channel wall was calculated to maximize... It can improve heat dissipation efficiency.

[0068] Specifically, in the single objective function for heat exchange, The quantification of heat dissipation performance relies on a simulation environment. The computer calculates the integral over the meshed design domain Ω based on the current γ distribution. For each mesh cell, (1-γ) determines the solid contribution. H n and Derived from boundary conditions and heat source settings. Time integrals represent the cumulative effect, achieved through numerical integration methods (such as the trapezoidal rule). The value was used J In the function, the driving optimization moves towards maximizing heat exchange.

[0069] ,in, For energy consumption, gamma_i To clarify the relative density variable, Re Let Reynolds number be 1. u For inlet flow rate, rho f For fluid density, D h The hydraulic diameter of the liquid inlet. 2 This is the gradient operator. α ( gamma_i (a clear relative density variable) gamma_i The permeability is given by the formula. The first term involves viscous dissipation, where ∇²u² represents the square of the velocity Laplace operator, capturing flow resistance. The second term is based on a clear relative density variable. gamma_i and permeability α ( gamma_i ), which represents the local pressure drop within the flow channel.

[0070] Fluid power consumption single objective function The working mechanism is: Quantifying flow resistance involves CFD simulations. The computer first solves the flow field (e.g., using the Navier-Stokes equations) to obtain the velocity. u And pressure distribution, then calculate ∇²u² (discrete using finite difference or finite element method). Permeability α ( gamma_i Based on projection gamma_i Ensure a clear fluid-structure interface. Time integration is similar. Cumulative power consumption. The value was used J Within the function, the driver optimization shifts towards minimizing power consumption. Re and D h Used for dimensionless calculations, making calculations more general.

[0071] The working mechanism of multi-objective functions is as follows: J The function, as the core objective of the optimization algorithm, guides the iterative process. In steps S2 and S4, the computer (e.g., using COMSOL or ANSYS software) calculates based on the current distribution of the relative density variable γ. and The value, then substitute it into J Expression. Algorithms (such as gradient descent or MMA methods) automatically adjust weight factors. w 1 and w 2 and coefficient Y and X To minimize or maximize J The summation ensures that historical performance is taken into account during optimization, achieving a dynamic balance. For example, in each iteration, the summation is calculated. J After the value is obtained, the algorithm updates the γ distribution to improve it. J Until convergence. Y and X The coefficients are adjusted via scripts or adaptive strategies. Y Strengthen synergy, X Handle conflicts and errors to ensure stable convergence.

[0072] The above function J、 and There are close dependencies and collaborative relationships between them: J It is a top-level multi-objective function that depends on sub-functions. and . and It is a single objective function that quantifies heat dissipation and power consumption performance separately. J Provide input. During the optimization process, the computer first calculates... and Based on the current γ distribution, then substitute them into J expression. J The value of γ drives the update of γ, and the change of γ in turn affects and The calculation forms a closed-loop feedback. For example, γ affects (1-γ) , through α( gamma_i )Influence .

[0073] and There is often a conflict (increasing heat exchange may increase power consumption). J Through weighted sum form ( Y Items) and conflict balance forms ( X(Item) to deal with this trade-off. Y Emphasizing collaboration, X Handle the discrepancies to ensure that optimization finds the Pareto optimal solution.

[0074] By defining a multi-objective function, the conflicting objectives of maximizing heat transfer and minimizing fluid power consumption are intelligently integrated into a weighted sum, thereby achieving an optimal balance in the design of liquid cooling plate flow channels. This function relies on the relative density variable and enhances numerical stability and error handling by dynamically adjusting the weighting factors and coefficients. It guides the topology optimization algorithm to automatically generate high-performance flow channel layouts, significantly improving heat dissipation efficiency, reducing flow resistance, and ensuring temperature uniformity. At the same time, it improves design flexibility and automation, shortens the design cycle, and ensures a clear fluid-solid interface and manufacturability.

[0075] In some embodiments, the expression for the flow channel volume fraction constraint is: ,in, v f This represents the upper limit of fluid volume fraction. V Ω The area of ​​the two-dimensional design domain is defined as a double-layered two-dimensional design.

[0076] Introducing a channel volume fraction constraint is a key physical constraint into the topology optimization process, ensuring that the generated channel layout is both engineering-feasible and cost-effective. By defining the channel volume fraction constraint, this method limits the maximum allowable volume percentage of the channels within the liquid cooling plate, thereby avoiding unrealistic channel sizes (such as excessively large channels leading to material waste, or excessively small channels causing a sharp increase in flow resistance). Specifically, The constraints are designed to control the total volume of the flow channel section, where γ is a relative density variable (range [0,1], 1 represents the flow channel region). v f This is the preset upper limit of fluid volume fraction (usually 0.3-0.5, representing the maximum proportion of the flow channel). V Ω This represents the total area of ​​the design domain. This constraint ensures that the optimization process is conducted under real-world manufacturing limitations. For example, in battery thermal management systems, it helps balance heat dissipation performance with material costs, preventing designs from becoming overly complex or difficult to manufacture. The aim is to enhance the practicality and reliability of the method, making the optimization results directly applicable to real-world engineering applications.

[0077] Specifically, expressions The entire two-dimensional design domain was calculated. Omega The integral of the relative density variable γ represents the total area (in the two-dimensional case) or volume (in the three-dimensional generalization) of the flow channel region. Constraints require that this integral value does not exceed... This means the flow channel volume cannot exceed a proportion of the total design domain volume. In step S4, the computer integrates this constraint as a hard constraint into the optimization algorithm (e.g., using the Lagrange multiplier method or penalty function method). For example, in each iteration, the algorithm adjusts the distribution of γ to optimize the multi-objective function. J Simultaneously, it checks whether the constraints are met. If a constraint is violated, the algorithm automatically penalizes or adjusts the value to ensure convergence to a feasible solution. The working mechanism is fully automated, implemented through the optimization module of simulation software (such as COMSOL), requiring no manual intervention and ensuring computational efficiency and numerical stability.

[0078] By defining the flow channel volume fraction constraint, the flow channel layout generated by topology optimization is ensured to conform to engineering reality in terms of volume ratio, thereby improving the feasibility, economy and performance of the design. As a hard constraint in the optimization process, it guides the algorithm to balance heat dissipation and flow resistance under limited resources, avoids unrealistic structures, enhances numerical stability and manufacturing adaptability, and ultimately provides an efficient, reliable and directly producible design solution for liquid cooling plates.

[0079] Figure 3 The diagram shows the 2D flow channel profile obtained through topology optimization with a flow channel volume fraction constraint of 0.5. The black areas represent solid material (γ=0), and the white areas represent the flow channel space (γ=1), with clear boundaries and no grayscale cells. The optimized flow channel exhibits a multi-branched, interconnected network structure (such as a mesh or radial rudimentary shape), ensuring uniform coolant coverage and efficient flow, and avoiding the dead zone problem of traditional parallel flow channels.

[0080] In topology optimization based on the relative density method, directly iteratively calculating the design variable γ is prone to producing checkerboard patterns and intermediate density values. These numerical artifacts can lead to non-physical oscillations and fuzzy fluid-structure boundaries in the optimization results, making them unsuitable for direct use in CAD modeling and actual manufacturing.

[0081] To address the aforementioned problems, this embodiment employs the following technical solution.

[0082] By introducing a density filtering function (based on the Helmholtz equation) and a projection function (based on the hyperbolic tangent function), this method aims to smooth the distribution of the relative density variable γ and force it to be binarized to 0 or 1, thereby generating a clear fluid-structure interface. Specifically, density filtering eliminates local oscillations through a diffusion mechanism, ensuring that the optimization result is mesh-independent; the projection function pushes the filtered variable to extreme values ​​through an S-shaped curve, avoiding non-physical intermediate states. The goal is to enhance the engineering feasibility of the optimization results, making the flow channel profile directly usable for CAD modeling and manufacturing, while improving the convergence and stability of the algorithm, providing reliable input for subsequent steps (such as iterative solving and 3D reconstruction).

[0083] In this embodiment, the step of performing density filtering on the relative density variable γ using a density filtering function includes: The relative density variable γ is filtered by solving the Helmholtz equation to obtain the filtered relative density variable. The expression for the Helmholtz equation is: ,in, R The filter radius controls the smoothness. 2 It is the Laplace operator (representing diffusion). The computer discretizes the equation in a simulation environment (such as using the finite element method) and obtains a smooth solution through iterative solving. The filter is distributed to eliminate high-frequency noise and checkerboard patterns. The filter radius R is typically set as a multiple of the grid size to ensure scale control.

[0084] The filtered relative density variable is processed through a projection function. The processing steps include: The filtered relative density variable The input hyperbolic tangent projection function is processed to obtain the sharp relative density variable after projection, which is used for subsequent iterative updates. gamma_i The expression for the projection function is: ,in, β The projection slope controls the binarization steepness. gamma β The projection threshold, typically set to 0.5, utilizes the S-shaped property of the hyperbolic tangent to push the value towards 0 (solid) or 1 (flow channel), generating a clear relative density variable. gamma_i The computer automatically calculates this function, ensuring the output is a binary distribution for easy subsequent iterative updates. Binarization yields a 0-1 distributed relative density variable, ensuring a clear boundary between the flow channel and the solid, avoiding manufacturing difficulties caused by grayscale units, and improving design feasibility.

[0085] By processing relative density variables through density filtering and projection functions, numerical instabilities in topology optimization, such as checkerboard patterns and intermediate density values, are effectively eliminated, generating a clear fluid-structure interface. This significantly improves the algorithm's convergence speed, numerical stability, and engineering feasibility. It ensures that the optimization results can be directly used in manufacturing, reducing post-processing requirements, while enhancing the performance reliability and manufacturing efficiency of flow channel design. This provides crucial digital foundation support for the overall liquid cooling plate flow channel design methodology.

[0086] In some embodiments, the curvature trajectory is a parabola, and the expression for the parabolic mode is: ,in, y Let the vertical axis be the coordinate origin at the center of the thickness section. xLet the x-axis be the coordinate system with the center of the thickness section as the origin. b This is the distance from the center of the thickness section to the wall of the parabolic flow channel; p Let be the direction of the parabola's opening, when p When >0, it deforms upwards; when p When <0, it deforms downwards; 1 / J ( J h ,J f ) represents the coefficient for the size of the parabola's opening.

[0087] in, As a coefficient for opening size. It is a multi-objective function value that integrates heat exchange. and fluid power consumption , Controlling the size of the parabola's opening: When J A decrease in the value indicates that the optimization objective is more inclined towards performance improvement. 1 / J The value increases, leading to an increase in the coefficient of the quadratic term. As the diameter increases, the parabolic opening becomes smaller, and the flow channel wall contracts further; when J When the value increases, it indicates that the optimization objective focuses more on reducing energy consumption, 1 / J As the value decreases, the coefficient of the quadratic term decreases, the opening of the parabola becomes larger, and the flow channel wall expands further.

[0088] When 1 / J As the value increases (the opening becomes smaller), the flow channel cross-section contracts, the coolant velocity increases, enhancing the convective heat transfer coefficient and turbulence effect, thereby increasing the heat transfer capacity. At the same time, the shrinkage of the flow channel wall reduces the flow cross-sectional area, lowers flow resistance and viscous dissipation, leading to increased energy consumption. Decrease.

[0089] When 1 / J As the value decreases (the opening becomes larger), the flow channel cross-section expands, the flow velocity decreases, and the heat transfer efficiency is reduced. The flow resistance decreases, but the flow resistance also decreases. Increase.

[0090] The curved trajectory is mathematically linked to a multi-objective function. Dynamic correlation enables intelligent optimization of the flow channel shape. Definition The opening size coefficient is used to ensure that the opening size of the parabola directly responds to changes in the optimization objective: when the multi-objective function... J When the value changes, the opening size automatically adjusts to balance maximizing heat exchange and minimizing energy consumption. Specifically, when... The value increases (i.e.) JAs the value decreases, the parabolic opening shrinks, causing the channel wall to contract, thus enhancing flow velocity and heat transfer efficiency; when The value decreases (i.e.) J As the value increases, the parabolic opening becomes larger, reducing flow resistance but potentially sacrificing some heat dissipation performance. This design ensures that the bending trajectory is not fixed but dynamically adapts according to the optimization process, thereby generating a high-performance irregular flow channel structure in three-dimensional reconstruction.

[0091] Through response J As the value changes, the parabolic opening size automatically optimizes the flow channel shape; when the opening becomes smaller, heat exchange... Increased energy consumption This reduces heat exchange, thereby achieving a balance between maximizing heat exchange and minimizing energy consumption.

[0092] In this embodiment, the opening size coefficient directly depends on the multi-objective function. This ensures that the bending trajectory is dynamically synchronized with the optimization target, guaranteeing that the three-dimensional reconstruction and topology optimization results in step S5 are consistent.

[0093] The equation of the parabola is obtained by dynamically adjusting the opening size coefficient. This mechanism enables intelligent response of the curved trajectory to changes in multiple objective functions, achieving automatic optimization of the flow channel shape: when the opening decreases, heat transfer is enhanced while energy consumption is reduced to achieve a performance balance; when the opening increases, flow resistance is reduced first, but heat dissipation is moderately sacrificed to ensure design adaptability. This mechanism significantly improves the heat dissipation efficiency, flow performance, and manufacturing feasibility of the liquid cooling plate, supports the efficient and reliable operation of the battery thermal management system, and shortens the design cycle while enhancing the overall automation level of the approach.

[0094] See attached document Figure 6 As shown, in the above-mentioned double-layer irregular flow channel structure, the diffusion flow channel layer adopts a mesh-like straight branch, and the converging flow channel layer adopts a radial main channel.

[0095] Specifically, the mesh structure consists of multiple parallel or nearly parallel straight branches, which are connected by transverse channels to form a grid, increasing the flow channel coverage density. The design of the straight branches reduces abrupt changes in flow direction and lowers local resistance; the mesh layout ensures that the coolant can be evenly distributed across the entire heat dissipation surface, avoiding flow dead zones.

[0096] The radial main channel extends outward from the outlet, with branch channels distributed radially and gradually converging into the main channel. This structure utilizes the centripetal flow principle to reduce the flow path length and accelerate hydrothermal discharge. The radial distribution matches the heat distribution characteristics of the battery module, increasing channel density in high-temperature regions and improving collection efficiency. The radial channels maintain geometric continuity during three-dimensional bending deformation, preventing flow separation.

[0097] By defining a dual-layer irregular flow channel structure with a mesh-like straight branch channel in the diffusion layer and a radial main channel in the converging layer, a systematic improvement in thermal management efficiency is achieved. This design innovatively utilizes the high coverage of the mesh-like straight branch channels to effectively eliminate heat dissipation blind spots at the ends and edges of the battery module, ensuring uniform coolant distribution and significantly enhancing temperature uniformity. Simultaneously, the radial main channel optimizes heat collection efficiency and reduces flow path length through radial convergence, avoiding temperature rise issues at low flow rates. Compared to traditional rectangular or I-shaped flow channels, this structure works collaboratively, not only solving the increased energy consumption caused by branch pushing at high flow rates but also reducing overall flow resistance and improving heat dissipation efficiency through a multi-level branch collaborative heat dissipation mechanism.

[0098] See attached document Figure 7 As shown, the cloud map uses a color gradient to represent the temperature distribution, with red representing high temperature and blue representing low temperature. The rectangular flow channel ( Figure 7 a) It shows obvious hot spots and temperature gradients, confirming its poor uniformity; I-shaped flow channel ( Figure 7 b) Some improvement, but temperature rise still exists at the end; while irregular flow channels ( Figure 7 c) The battery module exhibits a uniform low-temperature distribution, with a maximum temperature of 31.03℃ and a minimum temperature of 27.24℃, a temperature difference of 3.79℃. This temperature difference is controlled within 3.79℃, meeting the company's standard requirements for durable operating conditions. The flow channel of this application ( Figure 7 c) The maximum temperature was significantly reduced, and local heat accumulation was eliminated, verifying the overall effectiveness of multi-objective optimization and three-dimensional reconstruction in the design method.

[0099] This application also discloses a liquid-cooled plate flow channel design optimization device based on topology optimization, characterized in that it includes: The design domain construction module is used to obtain the geometric parameters of the liquid cooling plate, construct a two-dimensional design domain of the liquid cooling plate in the simulation environment, and determine the inlet and outlet boundary conditions of the two-dimensional design domain. The topology optimization module is used to define the relative density variable of the filling material, fill the double-layer two-dimensional design domain, obtain the two-dimensional initial flow channel profile of the liquid cooling plate, construct a multi-objective function, and perform topology optimization on the two-dimensional initial flow channel profile to achieve a balance between maximizing heat transfer and minimizing flow channel pressure drop. The density processing module is used to filter the relative density variable through a density filtering function and process the filtered relative density variable through a projection function to obtain a clear relative density variable, so as to eliminate intermediate density values ​​and obtain a clear fluid-solid interface. The iterative solution module is used to introduce the flow channel volume fraction constraint condition, and to substitute the clear relative density variable into the multi-objective function for iterative solution to obtain the target two-layer two-dimensional topological flow channel profile; The 3D reconstruction module, based on the target double-layer 2D topological flow channel contour, generates a converging flow channel layer entity and a diffusion flow channel layer entity through parametric stretching operations; it performs scanning deformation on the entity thickness section based on a predefined bending trajectory to generate an irregularly shaped converging flow channel layer entity and an irregularly shaped diffusion flow channel layer entity; and it generates a connecting flow channel body entity based on predefined connection rules, integrating the irregularly shaped converging flow channel layer entity and the irregularly shaped diffusion flow channel layer entity to form a double-layer irregular flow channel structure entity; The model generation module is used to construct a three-dimensional model of a solid liquid cooling plate based on the generated double-layer irregular flow channel structure entity, and to remove overlapping parts by using Boolean subtraction operation to generate a liquid cooling plate with a double-layer irregular flow channel structure.

[0100] This application also discloses a liquid cooling plate whose internal flow channels are implemented using the above-described design optimization method.

[0101] The liquid cooling plate is manufactured using a topology optimization-based design method. Its internal double-layer irregular flow channel structure works in concert with the mesh straight branches of the diffusion layer and the radial main channels of the collection layer, which significantly improves heat dissipation efficiency and temperature uniformity, and effectively eliminates heat dissipation blind spots at the end and edge of the battery module. At the same time, the optimized flow channel layout reduces flow resistance and pressure drop, and reduces pump power consumption.

[0102] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A liquid cooling plate flow channel design optimization method based on topology optimization, characterized in that, The steps include the following: S1. Obtain the geometric parameters of the liquid cooling plate, construct the two-dimensional design domain of the liquid cooling plate in the simulation environment, and determine the inlet and outlet boundary conditions of the two-dimensional design domain. S2. Define the relative density variable of the filling material, fill the two-dimensional design domain to obtain the two-dimensional initial flow channel profile of the liquid cooling plate, construct a multi-objective function, and perform topology optimization on the two-dimensional initial flow channel profile to achieve a balance between maximizing heat transfer and minimizing flow channel pressure drop. S3. The relative density variable is filtered by the density filtering function, and the filtered relative density variable is processed by the projection function to obtain a clear relative density variable, so as to eliminate intermediate density values ​​and obtain a clear fluid-solid interface. S4. Introduce the flow channel volume fraction constraint condition, and substitute the clear relative density variable into the multi-objective function for iterative solution to obtain the target two-layer two-dimensional topological flow channel profile; S5. Based on the target two-dimensional topological flow channel contour, generate a converging flow channel layer entity and a diffusion flow channel layer entity through parametric stretching operation; perform scanning deformation on the thickness section of the entity based on a predefined bending trajectory to generate an irregular converging flow channel layer entity and an irregular diffusion flow channel layer entity; and generate a connecting flow channel body entity based on a predefined connection rule, and integrate the irregular converging flow channel layer entity and the irregular diffusion flow channel layer entity to form a two-layer irregular flow channel structure entity; S6. Based on the generated double-layer irregular flow channel structure entity, construct a solid liquid cooling plate three-dimensional model, and use Boolean subtraction operation to remove overlapping parts to generate a liquid cooling plate with a double-layer irregular flow channel structure.

2. The liquid cooling plate flow channel design optimization method based on topology optimization as described in claim 1, characterized in that: The boundary conditions of one layer of the two-dimensional design domain adopt a converging flow pattern with four corner inlets and a central outlet, while the boundary conditions of the other layer of the two-dimensional design domain adopt a diffusion flow pattern with a central inlet and four corner outlets.

3. The liquid cooling plate flow channel design optimization method based on topology optimization as described in claim 1, characterized in that: The step of defining the relative density variable γ of the filling material includes: determining the material permeability property based on the relative density variable γ using the following interpolation function: ,in, α ( γ (This represents the actual penetration rate.) α f For fluid permeability, α s This represents the solid permeability.

4. The liquid cooling plate flow channel design optimization method based on topology optimization as described in claim 3, characterized in that: The multi-objective function is a weighted sum of the single-objective functions for heat transfer and fluid power consumption, and depends on the defined relative density variable γ. The expression of the multi-objective function is as follows: ,in, J It is a multi-objective function. This represents the single-objective function value of heat exchange at time t. This represents the single-objective function value of fluid power consumption at time t. w 1 The weighting factors are the objective function weights for heat exchange. w 2 These are the weighting factors for the energy consumption objective function. t For the iteration time variable, Y The coefficient of the main objective value, X This is the error compensation coefficient; ,in, For heat exchange, a single objective function, Ω For a two-dimensional design domain, dΩ Let the area be a differential element over the design domain. T For reference temperature, T in Inlet temperature, H n The wall heat transfer coefficient; ,in, For energy consumption as a single objective function, γi To clarify the relative density variable, Re The Reynolds number is... u For inlet flow rate, ρ f For fluid density, D h The hydraulic diameter of the liquid inlet. For gradient operators, This represents the squared term of the Laplace operator for the velocity field. α ( γi (a clear relative density variable) γi penetration rate This represents the viscous power consumption due to the velocity gradient. This indicates the power consumption due to local voltage drop caused by the flow channel structure.

5. The liquid cooling plate flow channel design optimization method based on topology optimization as described in claim 4, characterized in that: The expression for the flow channel volume fraction constraint condition is as follows: ,in, v f This represents the upper limit of fluid volume fraction. V Ω The area of ​​the two-dimensional design domain is defined as a double-layered two-dimensional design.

6. The liquid cooling plate flow channel design optimization method based on topology optimization as described in claim 4, characterized in that, The step of performing density filtering on the relative density variable γ using a density filtering function includes: The relative density variable γ is filtered by solving the Helmholtz equation to obtain the filtered relative density variable. The expression for the Helmholtz equation is: :in, R The filter radius; The filtered relative density variable is processed through a projection function. The processing steps include: The filtered relative density variable The input hyperbolic tangent projection function is processed to obtain the sharp relative density variable after projection, which is used for subsequent iterative updates. γi The expression for the projection function is: ,in, β The projection slope, γ β This is the projection threshold.

7. The liquid cooling plate flow channel design optimization method based on topology optimization as described in claim 4, characterized in that: The curved trajectory is a parabola, and the expression for the parabola is: ,in, y Let the vertical axis be the coordinate origin at the center of the thickness section. x Let the x-axis be the coordinate system with the center of the thickness section as the origin. b This is the distance from the center of the thickness section to the wall of the parabolic flow channel; p Let be the direction of the parabola's opening, when p When >0, it deforms upwards; when p When <0, it deforms downwards; This represents the coefficient for the size of the parabola's opening.

8. The liquid cooling plate flow channel design optimization method based on topology optimization as described in claim 1, characterized in that: The dual-layer irregular flow channel structure includes a converging flow channel layer and a diffusion flow channel layer. The converging flow channel layer adopts a radial main channel, and the diffusion flow channel layer adopts a mesh-like straight branch.

9. A device for optimizing the design of liquid-cooled plate flow channels based on topology optimization, characterized in that, include: The design domain construction module is used to obtain the geometric parameters of the liquid cooling plate, construct a two-dimensional design domain of the liquid cooling plate in the simulation environment, and determine the inlet and outlet boundary conditions of the two-dimensional design domain. The topology optimization module is used to define the relative density variable of the filling material, fill the double-layer two-dimensional design domain, obtain the two-dimensional initial flow channel profile of the liquid cooling plate, construct a multi-objective function, and perform topology optimization on the two-dimensional initial flow channel profile to achieve a balance between maximizing heat transfer and minimizing flow channel pressure drop. The density processing module is used to filter the relative density variable through a density filtering function and process the filtered relative density variable through a projection function to obtain a clear relative density variable, so as to eliminate intermediate density values ​​and obtain a clear fluid-solid interface. The iterative solution module is used to introduce the flow channel volume fraction constraint condition, and to substitute the clear relative density variable into the multi-objective function for iterative solution to obtain the target two-layer two-dimensional topological flow channel profile; The 3D reconstruction module, based on the target double-layer 2D topological flow channel contour, generates a converging flow channel layer entity and a diffusion flow channel layer entity through parametric stretching operations; it performs scanning deformation on the entity thickness section based on a predefined bending trajectory to generate an irregularly shaped converging flow channel layer entity and an irregularly shaped diffusion flow channel layer entity; and it generates a connecting flow channel body entity based on predefined connection rules, integrating the irregularly shaped converging flow channel layer entity and the irregularly shaped diffusion flow channel layer entity to form a double-layer irregular flow channel structure entity; The model generation module is used to construct a three-dimensional model of a solid liquid cooling plate based on the generated double-layer irregular flow channel structure entity, and to remove overlapping parts by using Boolean subtraction operation to generate a liquid cooling plate with a double-layer irregular flow channel structure.

10. A liquid-cooled plate, characterized in that, Its internal flow channel is implemented using the topology optimization-based liquid cooling plate flow channel design optimization method as described in any one of claims 1 to 8.

Images