Topological optimization method and system for liquid cooling plate of lithium ion battery pack

Abstract

The invention belongs to the technical field of advanced manufacturing and intelligent design, and provides a topological optimization method and system for a liquid cooling plate of a lithium ion battery pack, and the method comprises the steps: building a parameterized simulation model of a battery pack liquid cooling system, constructing a multi-dimensional design variable space of the liquid cooling plate, and formally defining a multi-objective optimization problem; on the basis of the initial training data set, independently constructing a probabilistic agent model capable of predicting a target value and quantifying uncertainty for each optimization target function; performing quantum behavior enhanced multi-universe optimizer iterative optimization on the constructed agent model, selecting new sample points from a multi-dimensional design variable space of the liquid cooling plate according to a set criterion after a set number of iterations is completed, performing high-fidelity CFD simulation to obtain real data, supplementing the real data to a data set, and updating or reconstructing the agent model; and judging whether an iterative optimization condition is met or not, and if so, screening and outputting a final Pareto optimal solution set from all samples subjected to high-fidelity simulation verification. And efficient multi-objective optimal design of the liquid cooling plate is realized.

Description

Technical Field

[0001] This invention belongs to the field of advanced manufacturing and intelligent design technology, and in particular relates to a method and system for topology optimization of liquid cooling plates for lithium-ion battery packs. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Lithium-ion batteries generate heat during charging and discharging due to internal resistance and polarization. If heat dissipation is not timely, high temperatures (above 50ºC) can accelerate irreversible side reactions and shorten lifespan, and may even trigger thermal runaway (fire or explosion). Furthermore, excessive temperature differences between cells within the battery pack can lead to performance degradation, creating a "weakest link" effect. Therefore, an ideal Battery Thermal Management System (BTMS) needs to control the battery temperature between 20-40ºC, with a temperature difference within the pack of less than 5ºC.

[0004] In existing thermal management solutions, liquid cooling has become the mainstream technology for large-capacity battery packs due to its high specific heat capacity and uniform heat exchange. The liquid cooling plate is the core heat exchange component of the liquid cooling system—its internal flow channel structure directly determines the cooling capacity. ), temperature uniformity ( ) and system energy consumption (corresponding to liquid cooling plate pressure drop) The Pareto optimal balance needs to be found among the three.

[0005] Liquid cooling plate optimization is a high-dimensional, nonlinear, multi-objective problem. Existing approaches fall into two categories: one is parametric optimization (pre-defined flow channel configuration and optimization of geometric parameters), but it is limited by the pre-defined configuration and it is difficult to achieve breakthrough designs; the other is topology optimization (automatic generation of flow channels within the design domain), which has a wider design space. However, both methods rely on computational fluid dynamics (CFD) simulation to evaluate performance.

[0006] CFD simulation (demand-based solution of fluid-structure interaction equations) is computationally extremely expensive (requiring hours to days per simulation), while the optimization process requires hundreds to thousands of simulations, resulting in development cycles that can last for weeks or even months, becoming a bottleneck for technology implementation. Existing mitigation solutions (such as orthogonal experiments to screen key factors and the improved multiverse optimizer (MVO) algorithm) all have limitations: the former uses static sampling and cannot be dynamically adjusted, while the latter still relies on a large number of CFD simulations, failing to fundamentally solve the computational cost problem. Summary of the Invention

[0007] To address at least one of the technical problems mentioned above, this invention provides a method and system for optimizing the topology of liquid cooling plates in lithium-ion battery packs. This method achieves efficient, accurate, and robust automated optimization of the liquid cooling plate topology under multiple objectives, significantly reducing the number of CFD calls and improving the quality and coverage of the obtained Pareto solution set.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: The first aspect of the present invention provides a method for topology optimization of liquid cooling plates in lithium-ion battery packs, comprising the following steps: A multidimensional design variable space for the liquid cooling plate is constructed, a parameterized simulation model of the battery pack liquid cooling system is established, and a multi-objective optimization problem is formally defined. Construct an initial training dataset, and based on the initial training dataset, independently construct a probabilistic surrogate model that can predict target values ​​and quantify uncertainty for each optimization objective function; The constructed surrogate model is subjected to iterative optimization using a quantum behavior-enhanced multiverse optimizer. After completing a set number of iterations, new sample points are selected from the multidimensional design variable space of the liquid-cooled plate according to the set criteria. High-fidelity CFD simulation is performed to obtain real data and supplement it to the dataset to update or reconstruct the surrogate model. Determine whether the iterative optimization conditions are met. If they are met, select and output the final Pareto optimal solution set from all high-fidelity simulation verification samples; otherwise, continue iterative optimization.

[0009] Furthermore, the establishment of a parameterized simulation model of the battery pack liquid cooling system and the formal definition of a multi-objective optimization problem include: Define the composition and key attributes of the integrated model of the battery module liquid cooling plate; A parameterized integrated model of the liquid cooling plate for the battery module was developed, and a multi-dimensional design variable space for the liquid cooling plate was constructed. Based on the constructed multidimensional design variable space of the liquid cooling plate, a parametric simulation model of the battery pack liquid cooling system is established. The parameterized simulation model of the battery pack liquid cooling system defines a multi-objective optimization function and constraints.

[0010] Furthermore, the construction of the initial training dataset, and the independent construction of a probabilistic proxy model that can predict target values ​​and quantify uncertainty for each optimization objective function based on the initial training dataset, includes: Multiple liquid cooling plate design schemes were generated based on the sampling of the constructed multidimensional design variable space of the liquid cooling plate as the initial training dataset. CFD full-process high-fidelity simulation was performed on each of the multiple liquid cooling plate design schemes to obtain the corresponding real objective function value set; Based on the obtained set of real objective function values, independent proxy models are constructed for each objective.

[0011] Furthermore, the construction of independent surrogate models for each objective based on the acquired set of real objective function values ​​includes: the independent surrogate model adopts the Kriging model, which represents the objective response as the superposition of a global trend term and a local deviation term, and uses the Gaussian correlation function as its kernel function to describe the spatial correlation between design variables.

[0012] Furthermore, the iterative optimization of the constructed proxy model using a multi-universe optimizer with enhanced quantum behavior includes: Initialize a quantum universe population, where the position of each universe is represented by a phase angle vector of a qubit, and each element of the vector corresponds to the quantum state of a design variable; The measurement operation is performed to collapse the quantum universe population from the quantum state to the classical state, that is, to convert the phase angle vector of each quantum bit of the universe into a set of classical design variable solutions through a preset mapping function; To perform rapid fitness assessment, the classical design variable solution is input into the surrogate model constructed for each objective, and the corresponding multiple optimization objective values ​​are quickly predicted, and the fitness of each universe is calculated. Perform quantum cosmological evolution operations, sort the universes according to their fitness, and implement a quantum behavior-enhanced cosmological renewal mechanism.

[0013] Furthermore, the established criteria are the expected hypervolume improvement criteria, which include comprehensively evaluating the potential improvement value and model prediction uncertainty of a point by calculating the expected hypervolume increment that a new sample point can contribute to the current Pareto front, and selecting the point with the largest EHVI value as the next sample point for high-fidelity numerical simulation.

[0014] Furthermore, the parameterized battery module liquid cooling plate integration model includes geometric and topological parameterization, operating condition parameterization, and boundary position parameterization; wherein, the geometric and topological parameters include: the number of first-level branches on one side. N Main channel width W m First-level branch width W b Channel height H Angle between primary branch and main channel α and the chamfer radius of the channel wall R Operating parameters include: inlet mass flow rate. M Boundary position parameters include: entry position S in and export location S out The above design variables together constitute the multidimensional design variable space D.

[0015] A second aspect of the present invention provides a topology optimization system for a liquid cooling plate in a lithium-ion battery pack, comprising: The objective definition module is used to construct the multi-dimensional design variable space of the liquid cooling plate, establish a parameterized simulation model of the battery pack liquid cooling system, and formally define the multi-objective optimization problem. The initial model building module is used to build the initial training dataset. Based on the initial training dataset, it independently builds a probabilistic surrogate model that can predict the target value and quantify the uncertainty for each optimization objective function. The iterative optimization module is used to perform quantum behavior-enhanced multiverse optimizer iterative optimization on the constructed surrogate model. After completing a set number of iterations, it selects new sample points from the liquid-cooled plate multidimensional design variable space according to the set criteria, performs high-fidelity CFD simulation to obtain real data and adds it to the dataset, updates or reconstructs the surrogate model, and determines whether the iterative optimization conditions are met. If they are met, it selects and outputs the final Pareto optimal solution set from all samples verified by high-fidelity simulation; otherwise, it continues iterative optimization.

[0016] A third aspect of the present invention provides a computer-readable storage medium.

[0017] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above-described method for topology optimization of a liquid cooling plate for a lithium-ion battery pack.

[0018] A fourth aspect of the present invention provides a computer device.

[0019] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps in the above-described method for topology optimization of liquid cooling plates for lithium-ion battery packs.

[0020] Compared with the prior art, the beneficial effects of the present invention are: This invention constructs a closed-loop optimization framework for the parametric topology design of battery pack liquid cooling plates. It formally defines a multi-objective optimization problem and performs a fast global search on a low-cost surrogate model. Simultaneously, the EHVI criterion guides the limited CFD simulation budget towards the most valuable sample points. This invention achieves efficient, accurate, and robust automated optimization of liquid cooling plate topology under multi-objective conditions, significantly reducing the number of CFD calls and improving the quality and coverage of the obtained Pareto solution set.

[0021] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0022] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0023] Figure 1 This is a flowchart of a method for topology optimization of liquid cooling plates for lithium-ion battery packs provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the parametric geometric model of the battery pack and the biomimetic leaf vein structure liquid cooling plate provided in the embodiment of the present invention, wherein (a) is a schematic diagram of the overall geometric model and (b) is a schematic diagram of the local geometric model; Figure 3 This is a schematic diagram of the construction process of the proxy model provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the quantum rotation gate update mechanism, the core of the QE-MVO algorithm provided in this embodiment of the invention; Figure 5 This is a schematic diagram of the quantum bit encoding and measurement mapping mechanism for design variables provided in an embodiment of the present invention; Figure 6 This is a two-dimensional target space schematic diagram of the expected hypervolume improvement (EHVI) criterion for dynamic point addition provided in an embodiment of the present invention; Figure 7 This is a convergence curve of the hypervolume of the real Pareto front during the optimization process provided in the embodiments of the present invention, which shows the increase of the number of CFD simulations. Figure 8 This is a distribution map of the final Pareto optimal solution set in the three-dimensional target space provided by the embodiments of the present invention. Detailed Implementation

[0024] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0025] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0026] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0027] This invention addresses the technical problems in existing battery pack liquid cooling plate topology optimization processes, such as high dependence on high-fidelity CFD simulation, low optimization efficiency, susceptibility to local optima, and inefficient resource allocation. It proposes a multi-universe optimizer based on a dynamic surrogate model and quantum behavior enhancement: A three-dimensional parameterized model is established, including channel topology type, size parameters, and operating parameters; multiple objective optimization problems such as maximum temperature, temperature difference, and pressure drop are defined; initial samples are obtained using a space-filling experimental design, and high-fidelity CFD simulations are performed to construct an initial training dataset; based on the training dataset, a probabilistic surrogate model with uncertainty quantification capability is independently constructed for each optimization objective. A quantum behavior-enhanced multiverse optimization is performed on the surrogate model. Through quantum phase angle encoding, measurement mapping, and adaptive quantum rotation gates, a collaborative optimization of global exploration and local development is achieved to obtain a candidate Pareto solution set. After several rounds of surrogate optimization, a dynamic addition strategy based on uncertainty is adopted to select new samples with the greatest improvement potential from the global design space, perform high-fidelity CFD verification, and add them to the training set. The surrogate model is dynamically updated or reconstructed to form a closed-loop iteration of "surrogate optimization - active addition - high-fidelity calibration". A global termination condition is set. If the condition is not met, the iteration continues. If the condition is met, the final Pareto optimal solution set is selected from the high-fidelity verified samples and output.

[0028] The method of this invention replaces most expensive simulations with a surrogate model and introduces a quantum-inspired adaptive search and uncertainty-driven resource allocation mechanism, which significantly improves optimization efficiency and global optimization capability, ensures the physical authenticity and engineering reliability of the results, has good versatility and scalability, and can be extended to other high-cost simulation-driven multi-objective engineering optimization scenarios.

[0029] The method of this invention can be deployed on the cloud / server side for high-fidelity numerical simulation and model training, or it can be performed on the edge / terminal side for data acquisition and result visualization, or a cloud-edge collaborative framework can be adopted to meet real-time and data security requirements. The computing and interaction nodes can be physical servers, clusters, and containerized platforms, or terminal devices such as smartphones, tablets, industrial gateways, and engineering workstations; the nodes can be directly or indirectly connected via Ethernet, Wi-Fi, cellular networks, Bluetooth, UWB, and their evolved forms. The components and processes of the method of this invention can be flexibly tailored according to the application scale and resource conditions, and this application does not limit them.

[0030] Example 1 like Figure 1 As shown, this embodiment provides a topology optimization method for liquid cooling plates in lithium-ion battery packs, including the following steps: S1: Establish a parameterized simulation model of the battery pack liquid cooling system and formally define the multi-objective optimization problem; Specifically, the steps include the following: S11. Define the composition and key attributes of the integrated model of the battery module liquid cooling plate; In this embodiment, the composition and key attributes of the battery module liquid cooling plate integrated model are described by decomposition diagram and parameterized definition; like Figure 2 As shown, Figure 2 (a) is a schematic diagram of the overall geometric model. Figure 2 (b) is a schematic diagram of the local geometric model. The integrated model consists of: cell array 201, thermal pad 202, liquid cooling plate entity 203, flow channel topology and fluid boundary; Specifically, the key properties of the ensemble model include: The cell array parameters include: 24 cells, each cell measuring 148 mm × 91 mm × 12 mm, arranged in a square shape.

[0031] The thermal pad parameters include: 1mm thickness, 3.0 W / (m·K), located between the bottom of the battery cell and the upper surface of the liquid cooling plate, filling the contact gap and enhancing heat transfer.

[0032] The liquid cooling plate has the following physical dimensions: 300 mm × 300 mm × 10 mm, made of aluminum alloy with a thermal conductivity of 202.4 W / (m·K).

[0033] The flow channel topology includes: a biomimetic leaf vein structure, comprising a main channel running through the center and symmetrical multi-level branch channels 205 on both sides; The fluid boundary includes: coolant inlet 204 and outlet 207. Coolant 206 flows in from the inlet, exchanges heat with the battery cell through a complex flow channel network, and then flows out from the outlet.

[0034] S12, Parametric battery module liquid cooling plate integration model, constructing a multi-dimensional design variable space for the liquid cooling plate; In this embodiment, the parameterized battery module liquid cooling plate integrated model includes geometric and topological parameterization, operating condition parameterization, and boundary position parameterization; wherein, the geometric and topological parameters include: the number of first-level branches on one side. N Main channel width W m First-level branch width W b Channel height H Angle between primary branch and main channel α and the chamfer radius of the channel wall R Operating parameters include: inlet mass flow rate. M Boundary position parameters include: entry position S in and export location S out The aforementioned design variables collectively constitute the multidimensional design variable space D; The range of values ​​for each design variable is limited based on engineering experience and manufacturing constraints to meet the following requirements. Interval constraints; Specifically, the geometric and topological parameters include: Number of first-level branches on one side N : X 1. An integer, with a feasible range of [3, 4, 5, 6, 7]; Main channel width W m : X 2. Continuous variable, feasible region is [8 mm, 15 mm]; First-level branch width W b : X 3. Continuous variable, feasible region is [4 mm, 8 mm]; Channel height H : X 4. Continuous variable, feasible region is [2 mm, 5 mm]; Angle between primary branch and main channel α : X 5. For continuous variables, the feasible region is... ; Channel wall chamfer radius R : X 6. Use continuous variables to reduce flow losses.

[0035] Operating parameters include: Inlet quality flow M : X7. Continuous variable, feasible region is [0.01 kg / s, 0.05 kg / s].

[0036] Boundary position parameters include: Entrance location S in : X 8. Discrete variable, with two possible locations: center and corner. The center is denoted as 0, and the corner as 1. Export location S out : X 9. Discrete variable, with two possible locations: center and corner. The center is denoted as 0, and the corner as 1.

[0037] S13. Based on the multidimensional design variable space of the constructed liquid cooling plate, establish a parameterized simulation model of the battery pack liquid cooling system. In this embodiment, the heat transfer and flow characteristics of the coupling between the liquid cooling plate and the battery cell are obtained through high-fidelity numerical simulation, and the simulation conditions, physical properties and solution settings are defined.

[0038] The specific simulation configuration includes: simulation platform parameters, heating model parameters, coolant property parameters, mesh and near-wall treatment, and solution control parameters.

[0039] The heating model parameters include: the cell operating condition is 3C constant current discharge; the volumetric heat generation rate is calculated based on the experimentally calibrated Bernardi model. Q gen , set as W / m 3 .

[0040] The Bernardi model calculation formula is: , In the formula: The heating power is expressed in W. Let be the current, in A; The open-circuit voltage is V; The terminal voltage is V; For absolute temperature, K; first term Representing irreversible Joule heating and polarization heating, the second term This represents the entropy heat of a reversible reaction. These parameters can usually be determined experimentally.

[0041] The coolant physical properties include: the working fluid is a 50% ethylene glycol-water solution; density, specific heat, thermal conductivity and viscosity are expressed as polynomial functions with temperature; and the inlet temperature is constant at 25ºC.

[0042] S14. Define the multi-objective optimization function and constraints for the parameterized simulation model based on the battery pack liquid cooling system; In this embodiment, the multi-objective optimization function and constraints are expressed as follows: , in, This is used to minimize the highest temperature of the battery cell, thus characterizing the system's ultimate heat dissipation capability. This is used to minimize the maximum temperature difference between battery cells to characterize the temperature uniformity of the system. This is used to minimize the pressure drop at the inlet and outlet of the liquid cooling plate, characterize the pumping power required to drive the coolant circulation, and is directly related to the system energy efficiency. Represent decision variables; Represents a multi-objective optimization function; This indicates the highest temperature of all battery cells used to evaluate the system's ultimate heat dissipation capability. This indicates the lowest temperature of all battery cells used to evaluate the system's ultimate heat dissipation capability; This represents the maximum temperature difference between all cells used to evaluate the temperature uniformity of the system. The total pressure drop between the inlet and outlet of the liquid-cooled plate, used to evaluate the pumping energy consumption of the system, represents the decision variables that satisfy the feasible region constraint. .

[0043] S2: Construct an initial training dataset. Based on the initial training dataset, independently construct a probabilistic surrogate model that can predict target values ​​and quantify uncertainty for each optimization objective function. Specifically, the steps include the following: S21. Based on the constructed multidimensional design variable space of the liquid cooling plate, generate multiple liquid cooling plate design schemes as the initial training dataset; In this embodiment, a space-filling experimental design is used to generate initial sample points within the feasible region of the design variables. Specifically, the Latin Hypercube Sampling (LHS) function is used to generate initial sample points within the aforementioned multidimensional design space. One initial sample point, corresponding to 50 liquid cooling plate design schemes; to ensure that the initial sample points achieve projection uniformity and space filling in the multi-dimensional design space.

[0044] S23. Perform CFD simulations on multiple liquid cooling plate design schemes to obtain the corresponding real objective function value sets, and construct independent proxy models for each objective based on the obtained real objective function value sets. Specifically, the steps include the following: S231, Design schemes for multiple liquid cooling plates respectively n Execute high-fidelity CFD simulations one by one with zero initial sample points, and record the results of each scheme. T max Δ T Δ PThe true values ​​are obtained by retrieving the corresponding set of true objective function values. F ( X 1), F ( X 2), ..., F ( X n0 )}; S232, Based on the obtained true objective function value set { F ( X 1), F ( X 2), ..., F ( X n0 )} are the respective objectives Build an independent agent model; In this embodiment, when constructing independent proxy models for each target based on the obtained set of real objective function values, the independent proxy models employ the Kriging model, which represents the target response. Represented as a global trend item With a local deviation term The superposition of variables is used, and the Gaussian correlation function is used as its kernel function to describe the spatial correlation between design variables; The independent agent model adopts the Kriging model, which is represented as: , in, The target response value to be predicted; The vector represents the regression function. These are the regression coefficients; Zero mean and variance σ 2 A Gaussian random process.

[0045] The Gaussian correlation function is used as its kernel function to describe the spatial correlation between design variables, expressed as: , in, The covariance function is used to describe the spatial correlation of response values ​​between any two design points. and For any two design points; A set of hyperparameters for the covariance function determines the "length scale" of the correlation; Dimensions of design variables Dimension index representing the design variable; The hyperparameters representing the variance function are used to determine the first... The length scale of correlation across design variable dimensions; Indicates the first The design point at the first Values ​​in each dimension; Indicates the first The design point at the first The values ​​can be taken in each dimension.

[0046] Finally, the probabilistic surrogate model that predicts the target value and quantifies uncertainty is expressed as: , , in, As the best predictor of the objective function value, variance Quantify the uncertainty or error in forecasting. Let be any unknown point; R This is the correlation matrix between samples; This is the correlation vector between the point to be tested and all sample points; G This is the regression matrix; ,variance σ 2 Model parameters are typically obtained through the Maximum Likelihood Estimation (MLE) method; The vector represents the regression function. This is the maximum likelihood estimate of the variance; Let be the residual vector of the regression function.

[0047] S3: Set the global optimization parameters, execute the Quantum-behaved Enhanced Multi-Verse Optimizer (QE-MVO) based on the surrogate model for iterative optimization, and determine whether the termination condition is met based on the set global optimization parameters. If it is met, the process terminates; otherwise, it continues iterative execution. like Figure 3 As shown, the specific steps include the following: S31. Set global optimization parameters; set the total CFD simulation budget (termination condition) to... And initialize the number of CFD simulations that have been executed so far. .

[0048] S32. Perform QE-MVO iterative optimization, specifically including: S321. Initialize the quantum universe population, where the position of each universe is represented by a quantum bit phase angle vector, and each element of the vector corresponds to the quantum state of a design variable; Specifically, each d The position of the dimensional universe is determined by one d Phase angle vector representation of a 1D quantum bit.

[0049] S322. Perform a measurement operation to collapse the quantum universe population from a quantum state to a classical state, that is, to convert the phase angle vector of each universe's qubit into a set of classical design variable solutions through a preset mapping function; Specifically, in terms of measurement operation, the first i The first universe j Phase angle of each qubit Mapped to the j Classic values ​​of design variables .

[0050] S323. Perform rapid fitness assessment by inputting the classical design variable solutions into the surrogate models constructed for each objective, quickly predicting the corresponding multiple optimization objective values, and calculating the fitness of each universe based on the multi-objective optimization ranking strategy. In this embodiment, the multi-objective optimization sorting strategy specifically adopts a sorting strategy that combines non-dominated sorting with crowding distance; S324. Perform quantum cosmic evolution operations, sorting the universes according to their fitness, and execute a quantum behavior-enhanced cosmic renewal mechanism, including quantum information exchange based on white hole / black hole mechanisms and wormhole jumps based on quantum rotating gates, to generate a new generation of quantum cosmic populations. The quantum rotating gate renewal mechanism is as follows: for the first... i The first universe j Each dimension, its phase angle Phase angle toward the current globally optimal universe To perform the rotation, the specific steps include the following: S3241. Sort the universes by fitness and designate a set number (e.g., the top 20%) of the universes as "white holes", corresponding to the high-quality liquid cooling plate design scheme; designate the rest as "black holes", corresponding to the scheme to be improved. In this embodiment, when sorting by universe fitness, sorting is prioritized by non-dominance level: first, the non-dominance level of each universe is compared, and the universe with the lower non-dominance level (i.e., in a higher non-dominance layer) is sorted higher; for universes of the same level, they are sorted by crowding distance: for universes with the same non-dominance level, their crowding distance is further compared, and the universe with the larger crowding distance (indicating that the distribution in the target space is more sparse and the diversity is better) is sorted higher.

[0051] This sorting method can prioritize the selection of universes with superior overall performance and diverse distribution, providing a priority basis for subsequent "quantum information exchange based on white hole / black hole mechanism" and "wormhole jump based on quantum rotating gate"—universes ranked higher (optimal solutions) will be more likely to retain high-quality information in the evolution, guiding the population to converge toward the global optimum.

[0052] S3242. For each "black hole", a "white hole" is randomly selected as the information source. By replacing part of the phase angle dimension of the black hole with the corresponding dimension of the white hole, the high-quality design features are absorbed, and the performance is initially improved. S3243. Check if the phase angle of the black hole is within [0, 2π). If it is outside this range, truncate and adjust to ensure the quantum state is valid. S3244. Select the universe with the highest fitness from the current population as the globally optimal universe, and record its phase angle in each dimension. θ best,j ; S3245, For each non-optimal universe, the... j Dimension, based on its current phase angle θ i,j and θ best,j Calculate the difference in rotation angle Δ θ i,j Ensure that the rotation amplitude is appropriate; S3246, The non-optimal universe number j The phase angle is updated to complete the "wormhole jump" and approach the global optimum. The update method is expressed as follows: , in, For the first t +1 times i The first universe j Phase angle in each dimension, For the first t sequence i The first universe j Phase angle in each dimension, t For the number of times.

[0053] S3247, the updated new generation of population, enters the next round of QE-MVO iteration; S33. After completing a predetermined number of QE-MVO iterations, new sample points are selected from the global design space according to the set criteria. High-fidelity CFD simulation is performed to obtain real data and supplement it to the training set, thereby updating or reconstructing the surrogate model. In this embodiment, the criterion is the Expected Hypervolume Improvement (EHVI) criterion. This criterion comprehensively evaluates the potential improvement value and model prediction uncertainty of a point by calculating the expected hypervolume increment that a new sample point can contribute to the current Pareto front, and selects the point with the largest EHVI value as the next sample point for high-fidelity numerical simulation.

[0054] S34. Determine whether the termination condition is met based on the set global optimization parameters. If the termination condition is not met, return to continue iterating based on the updated proxy model. If the condition is met, select and output the final Pareto optimal solution set from the samples verified by high-fidelity simulation.

[0055] If the condition is met, the process terminates; otherwise, iterative execution continues. In this embodiment, if the following conditions are met... If the termination condition is met, the loop will not be met when it first enters the loop because 50 < 150. The loop body will then continue to perform iterative optimization.

[0056] like Figure 4 As shown, in the quantum classical mapping and evolution process of this invention, the phase angle variable of each quantum individual is first... θ 401 via nonlinear mapping function f map 402 converted to the corresponding classic design variable x 403, for all populations N p The mapping is performed sequentially on 100 universes and their respective dimensions, resulting in 100 classical design schemes. These classical solutions are then input into three pre-trained Kriging surrogate models to obtain instantaneously predicted multi-objective function values. A comprehensive fitness is calculated based on non-dominated ranking and crowding distance to measure the quality and distribution diversity of the solutions. Based on this, the optimal quantum individual Θ for the current generation is determined according to the fitness results. best This is used to guide subsequent quantum evolution updates.

[0057] like Figure 5 As shown, in the quantum rotation gate update mechanism of this invention, the qubits of the universe to be updated have a unit circular phase angle. This indicates that, at the current optimal cosmic phase angle θ best,j Guided by [the principle], the rotation angle was calculated. Perform a rotation and update to the new state. The above operations are performed sequentially on each dimension of all non-optimal universes. After setting the number of iterations in the inner loop to 200, the output is an approximate Pareto optimal solution set based on the existing surrogate model.

[0058] like Figure 6 As shown, in this invention, EHVI is used to guide the adaptive addition of points and model updates in the outer loop: in the two-dimensional target space ( f 1, fIn the illustration of 2), known real CFD simulation points form the Pareto front, and a reference point is set. The shaded area enclosed by the two is the current hypervolume. For any candidate design, its Kriging prediction is a distribution with mean and variance. EHVI measures the expected increment of the hypervolume for that candidate point after integrating all possible real results. This embodiment uses particle swarm optimization (PSO) to maximize EHVI in the seven-dimensional design space to obtain new sample points. X new Subsequently, regarding X new Perform a complete high-fidelity CFD simulation to obtain the true target vector. F ( X new );Will( X new , F ( X new Add the data to the training dataset and retrain the three Kriging models, while updating the counters. N CFD,current = N CFD,current +1; The process returns to the termination condition and continues the outer loop. The outer loop executes 100 times in total, each time selecting a single high-value sample for high-cost evaluation using EHVI, in order to continuously improve the global accuracy of the surrogate model and its approximation to the Pareto front.

[0059] like Figure 7 As shown in the convergence diagram provided by this invention, the curves depict the trend of the Pareto front hypervolume, composed of completed real CFD simulation points, changing with the number of simulations: the horizontal axis represents the number of CFD simulations, and the vertical axis represents the normalized hypervolume value. The results show that in the early stages of optimization, the hypervolume increases rapidly, indicating that the proposed method can efficiently utilize information and quickly obtain a new scheme significantly superior to the initial design; subsequently, the slope of the curve gradually decreases and flattens, indicating that the optimization process is approaching convergence and it is difficult to continue obtaining a new design that significantly improves the current Pareto front.

[0060] like Figure 8 As shown, this invention obtains a three-dimensional target space after performing a non-dominated sort on 150 candidate points. T max Δ T Δ P The Pareto optimal solution set for region A. The solution for region A has low... T max With low Δ T However, it has a high voltage drop, making it suitable for high-performance applications; Region B has a low voltage drop but... T max With Δ TThe upper limit is suitable for energy-saving scenarios; region C (knee area) achieves a better balance among the three objectives, resulting in high cost-effectiveness for engineering applications. This solution set can be used to select the appropriate solution based on vehicle model positioning, cost, and weights. Compared to NSGA-II's direct optimization under the same 150 CFD budgets, this invention achieves approximately a 35% improvement in Pareto front hypervolume, and at the lowest... T max With the lowest Δ T It is superior in terms of extreme value indicators, and the proof method has higher efficiency and stronger optimization performance.

[0061] Example 2 This embodiment provides a topology optimization system for liquid cooling plates in lithium-ion battery packs, including: The objective definition module is used to construct the multi-dimensional design variable space of the liquid cooling plate, establish a parameterized simulation model of the battery pack liquid cooling system, and formally define the multi-objective optimization problem. The initial model building module is used to build the initial training dataset. Based on the initial training dataset, it independently builds a probabilistic surrogate model that can predict the target value and quantify the uncertainty for each optimization objective function. The iterative optimization module is used to perform quantum behavior-enhanced multiverse optimizer iterative optimization on the constructed surrogate model. After completing a set number of iterations, it selects new sample points from the liquid-cooled plate multidimensional design variable space according to the set criteria, performs high-fidelity CFD simulation to obtain real data and adds it to the dataset, updates or reconstructs the surrogate model, and determines whether the iterative optimization conditions are met. If they are met, it selects and outputs the final Pareto optimal solution set from all samples verified by high-fidelity simulation; otherwise, it continues iterative optimization.

[0062] It should be noted that the specific implementation of the topology optimization system for liquid cooling plates of lithium-ion battery packs in this embodiment of the invention is similar to the specific implementation of the topology optimization method for liquid cooling plates of lithium-ion battery packs in this embodiment of the invention. For details, please refer to the description in the method section. To reduce redundancy, it will not be repeated here.

[0063] Example 3 This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in the above-described method for optimizing the topology of a liquid cooling plate for a lithium-ion battery pack.

[0064] Example 4 This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the above-described method for optimizing the topology of a liquid cooling plate for a lithium-ion battery pack.

[0065] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0066] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0067] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0068] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0069] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0070] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for topology optimization of a lithium-ion battery pack liquid cooling plate, characterized in that, The method comprises the following steps: constructing a multi-dimensional design variable space of the liquid cooling plate, establishing a parameterized simulation model of the battery pack liquid cooling system, and formally defining a multi-objective optimization problem; constructing an initial training data set, and based on the initial training data set, independently constructing a probabilistic proxy model for each optimization objective function to predict the target value and quantify the uncertainty; performing quantum behavior enhanced multi-universe optimizer iterative optimization on the constructed proxy model, after completing a set number of iterations, selecting a new sample point from the liquid cooling plate multi-dimensional design variable space according to a set criterion, performing high-fidelity CFD simulation to obtain real data and supplementing the data set, and updating or reconstructing the proxy model; determining whether the iterative optimization condition is met, if yes, selecting and outputting the final Pareto optimal solution set from all high-fidelity simulation verified samples, otherwise, continuing the iterative optimization.

2. The topology optimization method for lithium-ion battery pack liquid cooling plate according to claim 1, wherein, The method of establishing a parameterized simulation model of the battery pack liquid cooling system and formally defining a multi-objective optimization problem comprises: defining the composition and key attributes of the battery module liquid cooling plate integrated model; parameterizing the battery module liquid cooling plate integrated model to construct a multi-dimensional design variable space of the liquid cooling plate; based on the constructed multi-dimensional design variable space of the liquid cooling plate, establishing a parameterized simulation model of the battery pack liquid cooling system; based on the parameterized simulation model of the battery pack liquid cooling system, defining a multi-objective optimization function and constraint.

3. The topology optimization method for lithium-ion battery pack liquid cooling plate according to claim 1, wherein, The method of constructing an initial training data set, and based on the initial training data set, independently constructing a probabilistic proxy model for each optimization objective function to predict the target value and quantify the uncertainty comprises: sampling a plurality of liquid cooling plate design schemes based on the constructed multi-dimensional design variable space of the liquid cooling plate to generate an initial training data set; respectively performing CFD full-process high-fidelity simulation on each of the plurality of liquid cooling plate design schemes to obtain a corresponding real target function value set; based on the obtained real target function value set, independently constructing a proxy model for each target.

4. The topology optimization method for lithium-ion battery pack liquid cooling plate according to claim 3, wherein, The method of independently constructing a proxy model for each target based on the obtained real target function value set comprises: the independent proxy model adopts a Kriging model, representing the target response as a superposition of a global trend item and a local deviation item, and adopting a Gaussian correlation function as its kernel function to describe the spatial correlation between design variables.

5. The topology optimization method for lithium-ion battery pack liquid cooling plate of claim 1, wherein, The method of performing quantum behavior enhanced multi-universe optimizer iterative optimization on the constructed proxy model comprises: initializing a quantum universe population, wherein the position of each universe is represented by a quantum bit phase angle vector, and each element of the vector corresponds to the quantum state of a design variable; performing a measurement operation to collapse the quantum universe population from a quantum state to a classical state, i.e., converting the quantum bit phase angle vector of each universe into a set of classical design variable solutions through a pre-set mapping function; performing a fast fitness evaluation, inputting the classical design variable solutions into the proxy model constructed for each target to quickly predict the corresponding multiple optimization target values, and calculating the fitness of each universe; performing a quantum universe evolution operation, sorting the universes according to their fitness, and performing a quantum behavior enhanced universe updating mechanism.

6. The topology optimization method for lithium-ion battery pack liquid cooling plate of claim 1, wherein, The set criterion is an expected hypervolume improvement criterion, which comprehensively evaluates the potential improvement value and the uncertainty of model prediction of a new sample point by calculating the expected hypervolume increment that the point can contribute to the current Pareto front, and selects the point with the largest EHVI value as the next sample point for high-fidelity numerical simulation.

7. The topology optimization method for lithium-ion battery pack liquid cooling plate of claim 1, wherein, The parameterized battery module liquid cooling plate integrated model includes geometric and topological parameterization, operating condition parameterization, and boundary position parameterization; among which, geometric and topological parameters include: the number of first-level branches on one side. N Main channel width W m First-level branch width W b Channel height H Angle between primary branch and main channel α and the chamfer radius of the channel wall R Operating parameters include: inlet mass flow rate. M Boundary position parameters include: Inlet position S in and outlet position S out The above design variables together form a multi-dimensional design variable space D.

8. A lithium-ion battery pack liquid cooling plate topology optimization system, comprising: Comprise: The target definition module is used to construct a multi-dimensional design variable space of the liquid cooling plate, establish a parameterized simulation model of the battery pack liquid cooling system, and formally define a multi-objective optimization problem; The initial model construction module is used to construct an initial training data set, and based on the initial training data set, independently construct a probabilistic proxy model for each optimization objective function, and quantify the uncertainty of the predicted target value; The iterative optimization module is used to perform quantum behavior enhanced multi-universe optimizer iterative optimization on the constructed proxy model, after completing a set number of iterations, select a new sample point from the multi-dimensional design variable space of the liquid cooling plate according to the set criterion, perform high-fidelity CFD simulation to obtain real data and supplement to the data set, update or reconstruct the proxy model; judge whether the iterative optimization condition is met, if met, select and output the final Pareto optimal solution set from all high-fidelity simulation verified samples, otherwise continue iterative optimization.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to realize the steps in the lithium ion battery pack liquid cooling plate topology optimization method of any one of claims 1-7.

10. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to realize the steps in the lithium ion battery pack liquid cooling plate topology optimization method of any one of claims 1-7.

Images