A battery pack thermal management method based on multi-objective optimization for multi-element dynamic thermal load

Abstract

The application discloses a battery pack thermal management method for multi-element dynamic thermal load based on multi-objective optimization, calibrates a high-precision heat generation model, optimizes a composite phase change material with high enthalpy and high thermal conductivity and a lotus leaf configuration liquid cooling plate, balances a prediction accuracy and a calculation efficiency of an agent model, introduces a core optimization objective function set of a worst activation criterion, adopts an initialization strategy and a cross variation mechanism for adaptively locking a parameter reasonable search space, solves a multi-objective optimization model based on an improved adaptive genetic algorithm, obtains a corresponding Pareto front solution set, and further combines a TOPSIS decision method to determine optimal configurations of key parameters such as a phase change layer thickness, a liquid cooling flow channel width, an inlet flow velocity and a system delay time, so that the battery pack can be controlled in a complex thermal environment to realize thermal safety and temperature uniformity control with the lowest energy consumption.

Description

Technical Field

[0001] This invention relates to the technical field of lithium-ion batteries, and in particular to a battery pack thermal management method based on multi-objective optimization and oriented towards multi-variable dynamic thermal loads. Background Technology

[0002] Against the backdrop of continuous energy structure adjustment and rapid development of energy storage technology, lithium-ion batteries have become key energy storage units in fields such as electric vehicles and energy storage power stations. However, these batteries are highly sensitive to operating temperature, and their ideal operating range is usually limited to 20-45℃. Under conditions of frequent charging and discharging of electric vehicles and drastic changes in operating conditions, or long-term high-density operation of battery packs in energy storage power stations, heat accumulates continuously inside the battery, which can easily exceed the safe temperature threshold, resulting in a decrease in power output and usable capacity, as well as accelerating the electrochemical aging process and significantly shortening service life. Therefore, building efficient and reliable battery thermal management strategies for different application scenarios is a key prerequisite for improving the safety, stability, and economy of lithium-ion battery systems.

[0003] Chinese patent application number CN201910787335.5 discloses a battery pack cooling pipeline and a battery pack cooling system, including a main pipeline and multiple branch pipelines. Each branch pipeline is connected to the main pipeline and to each sub-battery pack, allowing the coolant in the main pipeline to flow into each branch pipeline and cool each sub-battery pack through each branch pipeline. This enables convenient and rapid heat dissipation of the battery pack, with high cooling efficiency and good cooling uniformity. However, the method still has shortcomings. It only adds branch cooling pipelines for unidirectional cooling of the battery pack. This unidirectional cooling method can easily cause heat-generating components located in the latter half of the unidirectional cooling pipeline to not be properly cooled. Its performance improvement depends on increasing the flow rate or optimizing the flow channel structure to enhance turbulence, which inevitably leads to a significant increase in pump power consumption.

[0004] Chinese patent application number CN201810744959.4 discloses an automatic control thermal management system for power batteries based on phase change energy storage and thermoelectric effect. Under normal operating conditions, phase change material-assisted heat pipes are used for rapid heat dissipation, ensuring battery temperature stability without consuming any other form of energy. By changing the current direction, the semiconductor thermoelectric element can simultaneously perform cooling and heating functions, enabling rapid cooling at extreme high temperatures and rapid heating at extreme low temperatures. However, this method still has shortcomings, such as high difficulty in operation and control, low reliability, inability to solve the problem of sudden high-rate discharge and timely heat dissipation in a short period of time, and the limited phase change enthalpy of the phase change material, making it difficult to cope with extreme heat load conditions such as continuous high-rate discharge.

[0005] Therefore, synergistic thermal management strategies that integrate the advantages of active and passive cooling have attracted widespread attention. However, existing technologies are limited to multi-objective optimization under a single fixed operating condition, failing to fully consider the adaptive design requirements under diverse dynamic scenarios such as ambient temperature and discharge rate. They also struggle to clarify the optimization direction of design parameters when internal heat sources or external heat dissipation conditions change. To achieve an effective balance between peak temperature, temperature uniformity, and system energy consumption when coupling BTMS, a precise control strategy needs to be established to fully leverage the energy-saving and temperature uniformity potential of phase change materials, while avoiding premature intervention of the liquid cooling system to reduce operating energy consumption and improve system economy. Therefore, it is necessary to study a synergistic thermal management strategy that integrates the advantages of active and passive cooling. Summary of the Invention

[0006] The purpose of this invention is to solve the problems in the prior art and propose a battery pack thermal management method based on multi-objective optimization for multi-variable dynamic thermal loads, which can achieve thermal safety and temperature uniformity control of the battery pack with the lowest energy consumption in complex thermal environments.

[0007] To achieve the above objectives, this invention proposes a battery pack thermal management method based on multi-objective optimization and oriented towards multi-element dynamic heat load, comprising the following steps: Step 1: Based on the heat generation rate model of a uniform heat source, the internal resistance and temperature entropy coefficient are inverted by fitting the linear relationship between the temperature rise rate and the current during the discharge process under near-adiabatic conditions, and a high-precision heat generation model is obtained by calibration. Step II: Based on the obtained high-precision heat generation model, use computational fluid dynamics to simulate the working condition samples of the design space and obtain the numerical solution set of the entire working condition. Step III: Select a surrogate model for multi-objective optimization analysis that balances prediction accuracy and computational efficiency; Step IV: For objective functions with similar physical magnitudes and significant impact on system consistency, introduce the worst activation criterion to create a set of core optimization objective functions; Step V: Based on the actual cooling requirements of the working conditions, an initialization strategy for the search space of the adaptive locking parameters and a crossover mutation mechanism are introduced to obtain an improved adaptive genetic algorithm. Step VI: Solve the multi-objective optimization model based on the improved adaptive genetic algorithm to obtain the corresponding Pareto front solution set, and then combine it with the TOPSIS decision method to obtain the optimal configuration of key parameters; Step VII: Verify the optimization results through a simulation platform, simulate the system thermal state under different configurations after optimization, and monitor and verify the temperature control effect and design feasibility.

[0008] Preferably, the full-condition numerical solution set in step II includes the battery module peak temperature, lateral temperature difference, longitudinal temperature difference, liquid cooling system operating pump power, convective heat transfer intensity, and temperature standard deviation.

[0009] Preferably, the surrogate model in step III includes a backpropagation neural network, a radial basis function neural network, and a Gaussian process regression.

[0010] Preferably, the objective functions in step IV that are of similar physical magnitude and have a significant impact on system consistency include the lateral temperature difference and longitudinal temperature difference of the battery module, with three core optimization objective functions.

[0011] Preferably, the temperature difference adaptive activation method for the battery module in step IV, which addresses the lateral and longitudinal temperature differences, is set as follows: ; in, This indicates the activation of temperature difference. Indicates the lateral temperature difference. This indicates the longitudinal temperature difference.

[0012] Preferably, the adaptive genetic algorithm in step V is based on an improved NSGA-III algorithm, the crossover and mutation mechanism is executed for each generation of the population, and the initialization search space of the adaptive locking parameter reasonable search space initialization strategy includes the delay time as a parameter.

[0013] Preferably, the initialization strategy for the adaptive locking parameter reasonable search space in step V is based on the following initialization method using the delay time: ; ; ; in, The number of individuals in the population. The allowed range of values ​​for the delay time is: , and These are Boolean functions used to determine the initial population; The dynamic mutation rate is set according to the evolutionary stage as follows: ; in, The total number of generations of evolution, For the current algebra, Let early termination algebra be defined as the mutation rate. Midterm End Algebra .

[0014] The beneficial effects of this invention are as follows: By calibrating a high-precision heat generation model; selecting high-enthalpy, high-thermal-conductivity composite phase change materials and lotus leaf-shaped liquid cooling plates; employing a surrogate model that comprehensively balances prediction accuracy and computational efficiency; introducing a core optimization objective function set based on the worst-case activation criterion; and using an adaptive initialization strategy and crossover mutation mechanism to lock the parameter reasonable search space, this invention solves the multi-objective optimization model based on an improved adaptive genetic algorithm to obtain the corresponding Pareto front solution set. Furthermore, by combining the TOPSIS decision method, the optimal configuration of key parameters such as phase change layer thickness, liquid cooling channel width, inlet flow velocity, and system delay time is determined. This enables the battery pack to achieve thermal safety and temperature uniformity control with minimal energy consumption under complex thermal environments, providing theoretical support and practical reference for the thermal management design of next-generation electric vehicles and energy storage systems.

[0015] The features and advantages of the present invention will be described in detail through embodiments and in conjunction with the accompanying drawings. Attached Figure Description

[0016] Figure 1 This is a flowchart of a battery pack thermal management method based on multi-objective optimization and oriented towards multi-variable dynamic thermal loads, according to the present invention. Figure 2 This is a scatter matrix diagram of the design space sample distribution of a battery pack thermal management method based on multi-objective optimization and oriented towards multi-variable dynamic thermal loads according to the present invention. Figure 3 This invention relates to a battery pack thermal management method based on multi-objective optimization and oriented towards multi-element dynamic heat load, specifically targeting the peak temperature of the battery module. T max Scatter plot of test set prediction results for the proxy model; Figure 4 This invention relates to a battery pack thermal management method based on multi-objective optimization and oriented towards multi-element dynamic heat load, specifically addressing the lateral temperature difference of the battery module. Scatter plot of test set prediction results for the proxy model; Figure 5 This invention relates to a battery pack thermal management method based on multi-objective optimization and oriented towards multi-element dynamic heat load, specifically addressing the lateral temperature difference of the battery module. Scatter plot of test set prediction results for the proxy model; Figure 6 This invention relates to a battery pack thermal management method based on multi-objective optimization for multi-variable dynamic heat loads, specifically addressing the operation of pump power in liquid cooling systems. E pump Scatter plot of test set prediction results for the proxy model; Figure 7 This invention relates to a battery pack thermal management method based on multi-objective optimization for multi-element dynamic heat loads, specifically targeting the convective heat transfer intensity of liquid cooling systems.h Scatter plot of test set prediction results for the proxy model; Figure 8 This invention relates to a battery pack thermal management method based on multi-objective optimization for multi-variable dynamic heat loads, specifically addressing the temperature standard deviation of liquid cooling systems. T σ Scatter plot of test set prediction results for the proxy model; Figure 9 This is the Pareto front diagram based on the improved NSGA-III algorithm under the 25℃-1C condition of a battery pack thermal management method for multi-objective optimization oriented to multi-variable dynamic thermal loads according to the present invention. Figure 10 This is the Pareto front diagram based on the improved NSGA-III algorithm under the 25℃-2C condition of a battery pack thermal management method for multi-objective optimization oriented to multi-variable dynamic thermal load, according to the present invention. Figure 11 This is the Pareto front diagram based on the improved NSGA-III algorithm under the 25℃-3C condition of a battery pack thermal management method for multi-objective optimization oriented to multi-variable dynamic thermal load, according to the present invention. Figure 12 This is the Pareto front diagram based on the improved NSGA-III algorithm under the 25℃-4C condition of a battery pack thermal management method for multi-objective optimization oriented to multiple dynamic thermal loads according to the present invention. Figure 13 This is the optimal solution configuration diagram based on TOPSIS decision-making under normal temperature conditions for a battery pack thermal management method based on multi-objective optimization and oriented towards multi-variable dynamic heat load, according to the present invention. Figure 14 This is a comparison chart of the predicted and simulated values ​​of each objective function under normal temperature conditions for a battery pack thermal management method based on multi-objective optimization and oriented towards multi-variable dynamic heat load according to the present invention. Figure 15 This is the Pareto front diagram based on the improved NSGA-III algorithm under the 35℃-1C condition of a battery pack thermal management method for multi-objective optimization oriented to multi-variable dynamic thermal load, according to the present invention. Figure 16 This is the Pareto front diagram of a battery pack thermal management method based on multi-objective optimization and oriented towards multi-variable dynamic thermal load under the 35℃-2C condition, based on the improved NSGA-III algorithm. Figure 17 This is the Pareto front diagram of a battery pack thermal management method based on multi-objective optimization and oriented towards multi-variable dynamic thermal load under 35℃-3C conditions, based on the improved NSGA-III algorithm. Figure 18This is the Pareto front diagram based on the improved NSGA-III algorithm under the 35℃-4C condition of a battery pack thermal management method for multi-objective optimization oriented to multi-variable dynamic thermal load, according to the present invention. Figure 19 This is the optimal solution configuration diagram based on TOPSIS decision-making under high temperature conditions for a battery pack thermal management method based on multi-objective optimization and oriented towards multi-variable dynamic thermal loads according to the present invention. Figure 20 This is a comparison chart of the predicted and simulated values ​​of each objective function under high-temperature conditions in a battery pack thermal management method based on multi-objective optimization and oriented towards multi-variable dynamic heat loads according to the present invention. Figure 21 This is a graph showing the relative error between the predicted and simulated values ​​of each objective function under all operating conditions for a battery pack thermal management method based on multi-objective optimization and oriented towards multi-variable dynamic thermal loads, according to the present invention. Detailed Implementation

[0017] Example 1 like Figure 1 As shown, this invention provides a battery pack thermal management method based on multi-objective optimization for multi-variable dynamic heat loads, which independently configures the BTMS for normal temperature operating conditions, including a low-rate scenario (25℃-1C / 2C) corresponding to normal temperature endurance cruising and a high-rate scenario (25℃-3C / 4C) corresponding to intense driving in cool weather. The method includes: Step 1: Based on the heat generation rate model of a uniform heat source, the internal resistance and temperature entropy coefficient are inverted by fitting the linear relationship between the temperature rise rate and the current during the discharge process under near-adiabatic conditions, and a high-precision heat generation model is obtained by calibration. Step II: Based on the obtained high-precision heat generation model, use computational fluid dynamics methods to simulate the working condition samples of the design space to obtain a comprehensive numerical solution set covering all working conditions; Step III: Select a surrogate model for multi-objective optimization analysis that balances prediction accuracy and computational efficiency; Step IV: For objective functions with similar physical magnitudes and significant impact on system consistency, introduce the worst activation criterion to create an optimal set of core objective functions; Step V: Based on the actual cooling requirements of the working conditions, an initialization strategy for the search space of the adaptive locking parameters and a crossover mutation mechanism are introduced to obtain an improved adaptive genetic algorithm. Step VI: Solve the multi-objective optimization model based on the improved adaptive genetic algorithm to obtain the corresponding Pareto front solution set, and then combine it with the TOPSIS decision method to obtain the optimal configuration of key parameters; Step VII: Verify the optimization results through a simulation platform, simulate the system thermal state under different configurations after optimization, and monitor and verify the temperature control effect and design feasibility. In one example, the numerical solution set for the entire operating condition includes the battery module peak temperature, lateral temperature difference, longitudinal temperature difference, pump work of the liquid cooling system, convective heat transfer intensity, and temperature standard deviation. In one example, the surrogate model includes a backpropagation neural network, a radial basis function neural network, and a Gaussian process regression; In one example, the objective functions that are similar in physical magnitude and have a significant impact on system consistency include the lateral temperature difference and longitudinal temperature difference of the battery module, and there are three core optimization objective functions. In one example, the adaptive genetic algorithm is based on an improvement of the NSGA-III algorithm. The crossover and mutation mechanism is executed for each generation of the population. The parameter that is used to initialize the search space is the delay time. In one example, the temperature difference adaptive activation method is set as follows: ; in, , and These represent the activation temperature difference, lateral temperature difference, and longitudinal temperature difference, respectively. In one example, the adaptive bounded search initialization method based on delay time is as follows: ; ; ; in, The number of individuals in the population. The allowed range of values ​​for the delay time is: , and These are Boolean functions used to determine the initial population; The dynamic mutation rate is set according to the evolutionary stage as follows: ; in, The total number of generations of evolution, For the current algebra, Let early termination algebra be defined as the mutation rate. Midterm End Algebra ; Specifically, such as Figure 2As shown, computational fluid dynamics simulation was performed using commercial ANSYS 2022R1 software to simulate the temperature control effect under sample operating conditions. The heat dissipation parameters of the battery module coupled with the BTMS system were collected under different phase change layer thicknesses, liquid cooling channel widths, inlet velocities, system delay times, ambient temperatures, and discharge rates. Data processing was then performed to generate a training set, which served as training data for the backpropagation neural network. Data processing included denoising and standardization to ensure the quality of the neural network training data. Figure 3 , Figure 4 , Figure 5 , Figure 6 , Figure 7 and Figure 8 As shown in the scatter plot of the neural network's prediction results on the test set, the prediction results of the objective function (battery module peak temperature, lateral temperature difference, longitudinal temperature difference, cooling pump work, convective heat transfer intensity, and temperature standard deviation) are in good agreement with the actual values. Based on the prediction results of the backpropagation neural network, multi-objective optimization of design parameters (phase change layer thickness, liquid cooling channel width, inlet flow rate, and system delay time) is performed. The NSGA-III algorithm can optimize multiple objectives simultaneously, ensuring that the system provides the best temperature control effect with the least energy consumption under variable heat conditions. The improved NSGA-III algorithm can quickly focus on the core optimal objective function set and adaptively lock the reasonable search space of parameters according to the cooling requirements of specific operating conditions, avoiding wasting computational resources in invalid intervals. Through the non-dominated sorting strategy, multiple objectives including temperature difference, peak temperature, and pump power are optimized simultaneously, avoiding the shortcomings of single-objective optimization methods in multi-condition environments, thereby improving the overall performance of the system under variable heat conditions. During the optimization process, the algorithm generates new solution sets through crossover, adaptive mutation, and other operations, and continuously iterates and updates them to finally obtain the Pareto front solution. The TOPSIS decision method is used to select the optimal solution, ensuring efficient temperature control under variable operating conditions. Finally, the optimization results were verified through a simulation platform. The temperature changes under optimized parameters such as phase change layer thickness, liquid cooling channel width, inlet flow rate and system delay time were simulated to detect the system temperature control effect and ensure that the optimized system can reliably improve the safety and service life of the battery under various operating conditions. like Figure 9 , Figure 10 , Figure 11 and Figure 12As shown, the Pareto front solution under ambient temperature conditions indicates that liquid cooling is not a necessary component under such mild operating conditions. All non-dominated solutions point to a purely passive cooling strategy. Specifically, under extremely light load conditions with a discharge rate of 1C, the phase change layer thickness (TH) in the optimal TOPSIS solution is only 3.7 mm. Even so, this thin layer exhibits excellent thermal buffering and temperature homogenization performance: such as Figure 14 As shown, the battery pack , and T max The temperatures were controlled at excellent levels of 0.17°C, 0.27°C, and 34.2°C, respectively. This is mainly due to the high thermal conductivity of the phase change material, which can quickly diffuse the heat generated by the battery laterally, effectively homogenizing the temperature between batteries within the module. Simultaneously, its solid-state sensible heat capacity acts as a transient heat sink, delaying the overall temperature rise of the system. When the discharge rate increases to 2C and above, the battery heat generation rate increases significantly. Figure 13 As shown, the optimized framework adaptively adjusts the TH to 10.0 mm. The increase in TH brings two positive effects: first, the overall sensible heat capacity of the phase change material increases linearly, providing greater instantaneous heat absorption capacity; second, a larger latent heat reserve provides a safety margin to cope with possible short-term load fluctuations, such as... Figure 14 As shown, under this configuration, BTMS coupling was successfully achieved. , and T max Constrained within 1.0°C, 1.5°C, and 43°C respectively, the optimization results clearly reveal that the normal temperature operating condition is the efficient applicable range for pure passive cooling. By adjusting TH, it can effectively adapt to the heat load increase within a certain range, further confirming the economy and operational reliability of pure passive cooling under normal operating conditions.

[0018] Example 2 like Figure 1 As shown, the present invention provides a battery pack thermal management method based on multi-objective optimization for multi-variable dynamic heat loads, which performs independent BTMS configuration for high-temperature conditions, including high-temperature low-rate scenarios (35℃-1C / 2C) corresponding to summer congested road conditions, and high-temperature high-rate scenarios (35℃-3C / 4C) corresponding to long-term uphill climbing in summer. Compared to ambient temperature conditions, high-temperature environments significantly exacerbate the challenges of thermal management and trigger adaptive shifts in optimization strategies. A prominent feature is that the rise in ambient temperature severely weakens the system's passive heat dissipation capacity, causing the phase change buffer of the phase change material to be consumed prematurely, thereby compressing the effective window of passive cooling. Figure 15 , Figure 16 , Figure 17 and Figure 18The Pareto front solution for high-temperature conditions is shown below. Figure 19 The optimal design parameter configuration based on TOPSIS decision-making clearly reflects this change. The system strategy shifts from whether to activate liquid cooling to how to best coordinate phase change and liquid cooling. Specifically, under moderate heat load conditions (1C / 2C), passive cooling (10mm thick phase change layer) remains the primary heat dissipation method. Even though there is a tendency to activate liquid cooling as the heat load increases, a delay of 1435s is still achieved under 2C discharge conditions, resulting in a pump power saving of up to 79.7%. When operating conditions continue to deteriorate to 3C and 4C discharge, the battery heat load increases significantly, and the capacity of single passive cooling reaches its limit. The system design then shifts to a robust active-passive hybrid mode. It is worth noting that the optimal configuration parameters of TOPSIS have... The high degree of consistency, namely the combination of narrow flow channels (6mm), a thick phase change layer (10mm), and low flow velocity (0.01m / s), reflects a robust design approach to cope with high-intensity thermal shocks. The narrow flow channels help concentrate the coolant in the core hot zone, while the extremely low flow velocity, although reducing the convective heat transfer coefficient, significantly reduces pump work and makes the flow smoother and minimizes eddy current losses. Simultaneously, the lateral thermal diffusion capacity of the high thermal conductivity cold plate and the phase change layer compensates for the weakening of local heat transfer. At this point, the main role of the phase change material shifts from the main radiator to a transient load buffer, complementing the liquid cooling system. Under 3C conditions, the optimal solution (870.1s delay on start-up) exhibits excellent comprehensive thermal management performance: , and T max The temperature was suppressed to extremely low levels of 0.45℃, 0.22℃, and 43.6℃ respectively, while the pump power was only 11.3 mJ; under the more stringent 4C discharge conditions (liquid cooling delayed by 334.3 s to start), the system was still able to... T max The temperature was stably controlled at 43.8℃, with a maximum temperature difference of only 0.74℃. Figure 20 As shown, the above performance indicators are comprehensively superior to the thermal safety thresholds commonly used in the electric vehicle industry. T max <45℃ <5℃), fully demonstrating that even in high-temperature summer environments, this coupled BTMS still possesses excellent thermal safety assurance capabilities and sufficient safety redundancy when dealing with extreme operating scenarios such as continuous vehicle climbing and aggressive driving. Furthermore, as Figure 21 As shown, the error between the predicted results and the actual values ​​is less than 6% under all operating conditions, proving the reliability of this optimization framework and conclusions.

[0019] The heat generation model established in this invention has high accuracy, and the simulated battery surface points fit well with experimental measurements. The maximum temperature deviation under all operating conditions is still controlled within 3%, providing a reliable computational basis for the optimization of the thermal management system. This invention takes "lowest battery peak temperature, optimal temperature uniformity, and low system energy consumption" as its core design orientation, creates a core objective function set, and optimizes configuration parameters based on an improved adaptive genetic algorithm: phase change layer thickness, liquid cooling channel width, inlet flow rate, and system delay time. This achieves the multi-objective goal of simultaneously optimizing cooling system energy consumption, battery pack temperature control accuracy, and battery module temperature uniformity. The intelligent optimization framework constructed in this invention has scenario adaptability and can achieve reliable configuration decisions for typical automotive scenarios such as normal / high temperature and low / high rate. For extreme discharge conditions of high temperature and high rate, through a robust parameter combination of narrow channel, low flow rate, and thick CPCM, it can achieve ultimate control of thermal safety and temperature uniformity with minimal energy consumption, with the battery pack peak temperature and temperature difference not exceeding 44℃ and 0.8℃, respectively.

[0020] The above embodiments are illustrative of the present invention and are not intended to limit the present invention. Any simple modifications to the present invention are within the scope of protection of the present invention.

Claims

1. A battery pack thermal management method based on multi-objective optimization for multi-variable dynamic heat load, characterized in that: Includes the following steps: Step 1: Based on the heat generation rate model of a uniform heat source, the internal resistance and temperature entropy coefficient are inverted by fitting the linear relationship between the temperature rise rate and the current during the discharge process under near-adiabatic conditions, and a high-precision heat generation model is obtained by calibration. Step II: Based on the obtained high-precision heat generation model, use computational fluid dynamics to simulate the working condition samples of the design space and obtain the numerical solution set of the entire working condition. Step III: Select a surrogate model for multi-objective optimization analysis that balances prediction accuracy and computational efficiency; Step IV: For objective functions with similar physical magnitudes and significant impact on system consistency, introduce the worst activation criterion to create a set of core optimization objective functions; Step V: Based on the actual cooling requirements of the working conditions, an initialization strategy for the search space of the adaptive locking parameters and a crossover mutation mechanism are introduced to obtain an improved adaptive genetic algorithm. Step VI: Solve the multi-objective optimization model based on the improved adaptive genetic algorithm to obtain the corresponding Pareto front solution set, and then combine it with the TOPSIS decision method to obtain the optimal configuration of key parameters; Step VII: Verify the optimization results through a simulation platform, simulate the system thermal state under different configurations after optimization, and monitor and verify the temperature control effect and design feasibility.

2. The battery pack thermal management method based on multi-objective optimization for multi-element dynamic heat load as described in claim 1, characterized in that: The full-condition numerical solution set in step II includes the battery module peak temperature, lateral temperature difference, longitudinal temperature difference, liquid cooling system operating pump power, convective heat transfer intensity, and temperature standard deviation.

3. The battery pack thermal management method based on multi-objective optimization for multi-element dynamic thermal load as described in claim 1, characterized in that: The surrogate models in step III include backpropagation neural networks, radial basis function neural networks, and Gaussian process regression.

4. The battery pack thermal management method based on multi-objective optimization for multi-element dynamic thermal load as described in claim 1, characterized in that: The objective functions in step IV that are of similar physical magnitude and have a significant impact on system consistency include the lateral temperature difference and longitudinal temperature difference of the battery module. There are three core optimization objective functions.

5. A battery pack thermal management method based on multi-objective optimization for multi-element dynamic thermal load as described in claim 4, characterized in that: The temperature difference adaptive activation method for the battery module in step IV, which addresses the lateral and longitudinal temperature differences, is set as follows: ； in, This indicates the activation of temperature difference. Indicates the lateral temperature difference. This indicates the longitudinal temperature difference.

6. The battery pack thermal management method based on multi-objective optimization for multi-element dynamic thermal load as described in claim 1, characterized in that: The adaptive genetic algorithm in step V is an improvement on the NSGA-III algorithm. The crossover and mutation mechanism is executed according to the determination of the population in each generation. The initialization strategy of the adaptive locking parameter reasonable search space is used to initialize the search space. The parameters locked include the delay time.

7. A battery pack thermal management method based on multi-objective optimization for multi-element dynamic thermal load as described in claim 6, characterized in that: The initialization strategy for the adaptive locking parameter reasonable search space in step V, based on the initialization method of the delay time, is as follows: ； ； ； in, The number of individuals in the population. The allowed range of values ​​for the delay time is: , and These are Boolean functions used to determine the initial population; The dynamic mutation rate is set according to the evolutionary stage as follows: ； in, The total number of generations of evolution, For the current algebra, Let early termination algebra be defined as the mutation rate. Midterm End Algebra .

Images