Multi-objective collaborative optimization method for battery thermal management system

Abstract

The invention relates to a multi-objective collaborative optimization method for a battery thermal management system. The method comprises the following steps: comprehensively collecting key parameters influencing the performance of the battery thermal management system; based on the collected key parameters, an improved genetic programming model is adopted to construct a coupling relation model between the key parameters and the multiple optimization targets; based on the coupling relation model, adopting an improved non-dominated sorting genetic algorithm to carry out multi-objective collaborative optimization solution; and performing simulation verification on the optimal parameter combination. By introducing multi-task learning, a dynamic tree structure and a regularization item, a coupling relation model among key parameters, temperature uniformity and flow pressure drop is accurately constructed, and the problem that complex coupling of parameters and multiple targets is difficult to process in the prior art is effectively solved; in combination with an improved non-dominated sorting genetic algorithm, through simulation verification and a closed-loop adjustment mechanism, the reliability of an optimization result is ensured, the optimization efficiency is improved, the comprehensive performance of a battery thermal management system is remarkably improved, and the stability and safety of lithium battery operation are ensured.

Description

Technical Field

[0001] This invention relates to a multi-objective collaborative optimization method for a battery thermal management system, and relates to the field of battery thermal management technology. Background Technology

[0002] The core power source of electric vehicles relies on lithium batteries, which continuously generate heat during charging and discharging. Excessive or uneven temperature distribution directly affects the performance stability and lifespan of lithium batteries, and in severe cases, may even pose safety risks. Therefore, the battery thermal management system is a crucial component for ensuring the reliable operation of electric vehicles. Currently, liquid cooling is widely used in battery thermal management systems due to its advantages such as high heat dissipation efficiency and compact structure. Its core principle is to remove the heat generated by the battery through the flow of coolant within cooling channels, while simultaneously considering multiple optimization objectives such as temperature control and flow pressure drop to improve the overall system performance.

[0003] Existing multi-objective optimization methods for battery thermal management systems still suffer from the following problems. Specifically, these methods often struggle to accurately handle the complex coupling relationships between various design parameters and multiple optimization objectives, and fail to fully identify the interactive influence of different parameters on objectives such as temperature uniformity and flow pressure drop. This results in a lack of effective coordination among multiple objectives during the optimization process, with optimization results often biased towards a single objective at the expense of other objectives. This fails to guarantee a consistently uniform battery temperature distribution, effectively reduce pressure drop losses during coolant flow, and also suffers from low efficiency, making it unsuitable for meeting the stringent requirements of comprehensive system performance in practical applications.

[0004] Based on this, in order to solve the above technical problems, this paper proposes a multi-objective collaborative optimization method for battery thermal management system. Summary of the Invention

[0005] In view of the shortcomings of the prior art, the technical problem to be solved by the present invention is to provide a multi-objective collaborative optimization method for battery thermal management system, which solves the problems of difficulty in accurately handling parameter coupling with multiple objectives, lack of collaboration among multiple objectives and low optimization efficiency in the prior art.

[0006] To solve the above-mentioned technical problems, the technical solution of the present invention is: a multi-objective collaborative optimization method for a battery thermal management system, comprising the following steps:

[0007] S1. Collect all key parameters that affect the performance of the battery thermal management system. These key parameters include the structural parameters of the cooling channels and the operating parameters of the coolant.

[0008] S2. Based on the collected key parameters, an improved genetic programming model is used to construct a coupling relationship model between the key parameters and multiple optimization objectives. The coupling relationship model describes the correspondence between parameter changes and optimization objective performance changes. The optimization objectives include temperature uniformity and flow pressure drop.

[0009] S3. Based on the coupling relationship model, an improved non-dominated sorting genetic algorithm is used to solve the multi-objective collaborative optimization problem. The solution process includes constructing a collaborative optimization objective function that takes into account both the improvement of temperature uniformity and the reduction of flow pressure drop, and selecting the optimal parameter combination that simultaneously meets the performance requirements of multiple optimization objectives within the preset constraint range.

[0010] S4. Perform simulation verification on the optimal parameter combination. If the simulation results do not meet the preset comprehensive performance requirements, adjust the parameter weights in the coupling relationship model and the iterative parameters of the improved non-dominated sorting genetic algorithm according to the deviation direction of the simulation results, and re-execute the multi-objective collaborative optimization solution steps.

[0011] Preferably, in step S1, the comprehensive collection of key parameters affecting the performance of the battery thermal management system includes:

[0012] Collect the structural parameters of the cooling channel, including the geometric parameters of the cooling channel and the structural parameters of the cooling plate;

[0013] Collect the operating parameters of the coolant, including the inlet temperature, inlet flow rate, and flow pressure.

[0014] Parameters are collected using a combination of real-time monitoring and static measurement.

[0015] Preferably, in step S2, the construction of a coupling relationship model between the key parameters and multiple optimization objectives based on the collected key parameters using an improved genetic programming model includes:

[0016] Key parameters are classified and preprocessed to eliminate redundant parameters;

[0017] Through multiple rounds of iterative calculations, the influence of key parameters on the optimization objectives of temperature uniformity and flow pressure drop is deeply explored. The optimization objective of temperature uniformity is quantified by the standard deviation of the cell temperature, and the formula for standard deviation quantification is as follows:

[0018]

[0019] Where, σ T T represents the standard deviation of the temperature of a single battery cell. k Let K be the temperature of the k-th cell. The average temperature of all battery cells is N, where N is the total number of battery cells.

[0020] A coupled model is formed to describe the relationship between parameter changes and target performance changes.

[0021] Preferably, in step S2, the improved genetic programming model automatically discovers the nonlinear coupling relationship between key parameters and optimization objectives by generating and evolving a set of programs; the evolution process includes selection, crossover, and mutation operations, and the program performance is evaluated based on the fitness function, which aims to minimize the deviation between the model's predicted values ​​and the actual values; the model structure has adaptive adjustment capabilities.

[0022] Preferably, in step S3, the multi-objective collaborative optimization solution based on the coupling relationship model and using an improved non-dominated sorting genetic algorithm includes:

[0023] Based on the physical value range of key parameters and the minimum performance requirements of the optimization objective;

[0024] A collaborative optimization objective function is constructed, which balances the improvement of temperature uniformity and the reduction of flow pressure drop. The collaborative optimization objective function is as follows:

[0025] F(X) = w1·f1(X) + w2·f2(X);

[0026] Where X is the design parameter combination vector, f1(X) is the objective function for optimizing temperature uniformity, f2(X) is the objective function for optimizing flow pressure drop, and w1 and w2 are dynamic weighting coefficients, satisfying w1+w2=1;

[0027] Within the constraints, the parameter combination is iteratively optimized through multiple rounds using an improved non-dominated sorting genetic algorithm to select the optimal parameter combination that simultaneously satisfies the performance requirements of multiple optimization objectives.

[0028] Preferably, the parameter combination is iteratively optimized through multiple rounds using an improved non-dominated sorting genetic algorithm, including:

[0029] Perform non-dominated sorting and divide the current population into multiple non-dominated levels;

[0030] Crowding degree is calculated for individuals at the same non-dominated level to maintain solution diversity;

[0031] Perform a selection operation, filtering individuals based on non-dominated hierarchy and crowding;

[0032] Perform crossover and mutation operations to generate a new population with different parameter combinations.

[0033] Preferably, in step S4, the simulation verification of the optimal parameter combination includes:

[0034] The optimal parameter combination is substituted into the simulation model of the battery thermal management system to simulate the system's operating state under these parameter settings.

[0035] The temperature distribution data and flow pressure drop loss data output by the detection simulation model were examined.

[0036] The detected data is compared with preset temperature uniformity requirements and flow pressure drop requirements.

[0037] Preferably, in step S4, if the simulation results do not meet the preset comprehensive performance requirements, the parameter weights in the coupling relationship model and the iterative parameters of the improved non-dominated sorting genetic algorithm are adjusted according to the deviation direction of the simulation results, including:

[0038] If the temperature uniformity performance is poor, increase the dynamic weight coefficient corresponding to the temperature uniformity optimization objective in the collaborative optimization objective function;

[0039] If the flow pressure drop performance is poor, increase the dynamic weight coefficient corresponding to the flow pressure drop optimization objective in the collaborative optimization objective function;

[0040] Based on the difference between the simulation results and the preset performance requirements, the number of iterations, population size, or mutation rate of the improved non-dominated sorting genetic algorithm are adaptively adjusted.

[0041] Preferably, the improved genetic programming model is improved in that:

[0042] A multi-task learning mechanism is introduced to simultaneously learn the coupling relationship between key parameters and multiple optimization objectives;

[0043] A dynamic tree structure is adopted to adaptively adjust the complexity and depth of the model in order to describe the correspondence between parameter changes and target performance changes;

[0044] By introducing a regularization term, the prediction accuracy and stability of the model can be improved.

[0045] Preferably, the improved non-dominated sorting genetic algorithm is improved in that:

[0046] An elite retention strategy is introduced to preserve and pass on the optimal solution;

[0047] An adaptive crossover rate and mutation rate mechanism is adopted to dynamically adjust the operation parameters according to the population evolutionary state;

[0048] It integrates multimodal optimization capabilities to search for and identify multiple optimal solution clusters on the Pareto front.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] This invention utilizes an improved genetic programming model, incorporating multi-task learning, a dynamic tree structure, and regularization terms to accurately construct a model of the coupling relationship between key parameters and temperature uniformity and flow pressure drop. This effectively addresses the challenge of handling complex coupling between parameters and multiple objectives in existing technologies. Furthermore, an improved non-dominated sorting genetic algorithm, employing elite retention, adaptive crossover mutation rate, and multimodal optimization capabilities, coupled with a dynamically weighted collaborative optimization objective function, achieves multi-objective collaborative optimization, preventing optimization results from favoring a single objective at the expense of other performance aspects. Simultaneously, simulation verification and a closed-loop adjustment mechanism optimize model parameter weights and algorithm iteration parameters based on deviations, ensuring reliable optimization results and improving optimization efficiency. Ultimately, this significantly enhances the overall performance of the battery thermal management system, guaranteeing the stability and safety of lithium battery operation.

[0051] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0052] Figure 1 This is a flowchart of the method of the present invention;

[0053] Figure 2 This is a flowchart of the key parameter acquisition and processing of the present invention;

[0054] Figure 3 This is a flowchart illustrating the improved genetic programming modeling process of the present invention.

[0055] Figure 4 This is a flowchart of the multi-objective optimization and verification process of the present invention. Detailed Implementation

[0056] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0057] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of this application. 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 application pertains.

[0058] 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 exemplary embodiments according to this application. 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.

[0059] like Figures 1-4 As shown, this embodiment provides a multi-objective collaborative optimization method for a battery thermal management system, including the following steps:

[0060] S1. Collect all key parameters that affect the performance of the battery thermal management system. These key parameters include the structural parameters of the cooling channels and the operating parameters of the coolant.

[0061] S2. Based on the collected key parameters, an improved genetic programming model is used to construct a coupling relationship model between the key parameters and multiple optimization objectives. The coupling relationship model describes the correspondence between parameter changes and optimization objective performance changes. The optimization objectives include temperature uniformity and flow pressure drop.

[0062] S3. Based on the coupling relationship model, an improved non-dominated sorting genetic algorithm is used to solve the multi-objective collaborative optimization problem. The solution process includes constructing a collaborative optimization objective function that takes into account both the improvement of temperature uniformity and the reduction of flow pressure drop, and selecting the optimal parameter combination that simultaneously meets the performance requirements of multiple optimization objectives within the preset constraint range.

[0063] S4. Perform simulation verification on the optimal parameter combination. If the simulation results do not meet the preset comprehensive performance requirements, adjust the parameter weights in the coupling relationship model and the iterative parameters of the improved non-dominated sorting genetic algorithm according to the deviation direction of the simulation results, and re-execute the multi-objective collaborative optimization solution steps.

[0064] In this embodiment of the invention, step S1, which involves comprehensively collecting key parameters affecting the performance of the battery thermal management system, includes:

[0065] Collect the structural parameters of the cooling channel, including the geometric parameters of the cooling channel and the structural parameters of the cooling plate;

[0066] Collect the operating parameters of the coolant, including the inlet temperature, inlet flow rate, and flow pressure.

[0067] Parameters are collected using a combination of real-time monitoring and static measurement.

[0068] In this embodiment of the invention, step S2, which involves constructing a coupling relationship model between the key parameters and multiple optimization objectives based on the collected key parameters using an improved genetic programming model, includes:

[0069] Key parameters are classified and preprocessed to eliminate redundant parameters;

[0070] Through multiple rounds of iterative calculations, the influence of key parameters on the optimization objectives of temperature uniformity and flow pressure drop is deeply explored. The optimization objective of temperature uniformity is quantified by the standard deviation of the cell temperature, and the formula for standard deviation quantification is as follows:

[0071]

[0072] Where, σ TT represents the standard deviation of the temperature of a single battery cell. k Let K be the temperature of the k-th cell. The average temperature of all battery cells is N, where N is the total number of battery cells.

[0073] A coupled model is formed to describe the relationship between parameter changes and target performance changes.

[0074] In this embodiment of the invention, in step S2, the improved genetic programming model automatically discovers the nonlinear coupling relationship between key parameters and optimization objectives by generating and evolving a set of programs; the evolution process includes selection, crossover, and mutation operations, and the program performance is evaluated based on the fitness function, which aims to minimize the deviation between the model's predicted values ​​and the actual values; the model structure has adaptive adjustment capabilities.

[0075] In this embodiment of the invention, step S3, which involves using an improved non-dominated sorting genetic algorithm to solve a multi-objective collaborative optimization problem based on a coupling relationship model, includes:

[0076] Based on the physical value range of key parameters and the minimum performance requirements of the optimization objective;

[0077] A collaborative optimization objective function is constructed, which balances the improvement of temperature uniformity and the reduction of flow pressure drop. The collaborative optimization objective function is as follows:

[0078] F(X) = w1·f1(X) + w2·f2(X);

[0079] Where X is the design parameter combination vector, f1(X) is the objective function for optimizing temperature uniformity, f2(X) is the objective function for optimizing flow pressure drop, and w1 and w2 are dynamic weighting coefficients, satisfying w1+w2=1;

[0080] Within the constraints, the parameter combination is iteratively optimized through multiple rounds using an improved non-dominated sorting genetic algorithm to select the optimal parameter combination that simultaneously satisfies the performance requirements of multiple optimization objectives.

[0081] In this embodiment of the invention, the parameter combination is iteratively optimized through multiple rounds using an improved non-dominated sorting genetic algorithm, including:

[0082] Perform non-dominated sorting and divide the current population into multiple non-dominated levels;

[0083] Crowding degree is calculated for individuals at the same non-dominated level to maintain solution diversity;

[0084] Perform a selection operation, filtering individuals based on non-dominated hierarchy and crowding;

[0085] Perform crossover and mutation operations to generate a new population with different parameter combinations.

[0086] In this embodiment of the invention, step S4, which involves simulating and verifying the optimal parameter combination, includes:

[0087] The optimal parameter combination is substituted into the simulation model of the battery thermal management system to simulate the system's operating state under these parameter settings.

[0088] The temperature distribution data and flow pressure drop loss data output by the detection simulation model were examined.

[0089] The detected data is compared with preset temperature uniformity requirements and flow pressure drop requirements.

[0090] In this embodiment of the invention, step S4, where the step of adjusting the parameter weights in the coupling relationship model and the iterative parameters of the improved non-dominated sorting genetic algorithm according to the deviation direction of the simulation results if the simulation results do not meet the preset comprehensive performance requirements, includes:

[0091] If the temperature uniformity performance is poor, increase the dynamic weight coefficient corresponding to the temperature uniformity optimization objective in the collaborative optimization objective function;

[0092] If the flow pressure drop performance is poor, increase the dynamic weight coefficient corresponding to the flow pressure drop optimization objective in the collaborative optimization objective function;

[0093] Based on the difference between the simulation results and the preset performance requirements, the number of iterations, population size, or mutation rate of the improved non-dominated sorting genetic algorithm are adaptively adjusted.

[0094] In this embodiment of the invention, the improved genetic programming model is improved in that:

[0095] A multi-task learning mechanism is introduced to simultaneously learn the coupling relationship between key parameters and multiple optimization objectives;

[0096] A dynamic tree structure is adopted to adaptively adjust the complexity and depth of the model in order to describe the correspondence between parameter changes and target performance changes;

[0097] By introducing a regularization term, the prediction accuracy and stability of the model can be improved.

[0098] In this embodiment of the invention, the improved non-dominated sorting genetic algorithm is improved in that:

[0099] An elite retention strategy is introduced to preserve and pass on the optimal solution;

[0100] An adaptive crossover rate and mutation rate mechanism is adopted to dynamically adjust the operation parameters according to the population evolutionary state;

[0101] It integrates multimodal optimization capabilities to search for and identify multiple optimal solution clusters on the Pareto front.

[0102] Specific implementation process:

[0103] Please see Figure 1-4 This embodiment provides a multi-objective collaborative optimization method for a battery thermal management system, and details it in conjunction with the thermal management system of a ternary lithium battery pack in a compact electric vehicle. This lithium battery pack contains 100 individual cells and employs a liquid-cooled thermal management structure, using a 50% ethylene glycol aqueous solution as the coolant. The aim is to achieve a comprehensive improvement in temperature uniformity and flow pressure drop through multi-objective collaborative optimization. The specific implementation includes the following steps:

[0104] First, a comprehensive collection of key parameters affecting the performance of the battery thermal management system is performed. These key parameters include the structural parameters of the cooling channels and the operating parameters of the coolant. The completeness and accuracy of the parameters must be ensured during the collection process to provide reliable data support for subsequent model building and optimization. The structural parameters of the cooling channels cover both the geometric shape of the cooling channels and the structural parameters of the cooling plates. The geometric parameters of the cooling channels include a rectangular cross-section, a width ranging from 8-12 mm, a height ranging from 4-6 mm, and a channel spacing ranging from 15-25 mm. The structural parameters of the cooling plates include a thickness of 3-5 mm, the use of 6061 aluminum alloy, and 20 channels on the cooling plates matching the arrangement of the battery cells. The operating parameters of the coolant include an inlet temperature ranging from 20-30℃, an inlet flow rate ranging from 5-15 L / min, and a flow pressure ranging from 0.1-0.3 MPa. Parameter acquisition employs a combination of real-time monitoring and static measurement. Static measurement is used to obtain fixed structural parameters such as the geometry of the cooling channel and the thickness of the cooling plate. Multiple measurements are taken using high-precision vernier calipers, laser rangefinders, and other equipment, and the average value is calculated. Real-time monitoring uses temperature sensors, flow sensors, pressure sensors, and other equipment to continuously collect dynamic operating parameters such as coolant inlet temperature, inlet flow rate, and flow pressure during the charging and discharging process of the battery pack. The sampling frequency is set to 1Hz, and the monitoring duration covers a complete charge and discharge cycle.

[0105] Furthermore, based on the collected key parameters, an improved genetic programming model was used to construct a coupling relationship model between the key parameters and multiple optimization objectives. This model describes the correspondence between parameter changes and changes in the performance of the optimization objectives, including temperature uniformity and flow pressure drop. This model can accurately uncover the complex correlation between parameters and objectives, providing a solid foundation for subsequent optimization. First, the collected key parameters were preprocessed by classification. Redundant parameters were eliminated through correlation analysis, specifically using the Pearson correlation coefficient method to calculate the correlation coefficients between each parameter. A correlation coefficient threshold of 0.85 was set. When the correlation coefficient between two parameters exceeded this threshold, the parameters with a more significant impact on the optimization objectives were retained. Finally, seven key parameters were selected: cooling channel width, channel height, channel spacing, cooling plate thickness, coolant inlet temperature, inlet flow rate, and flow pressure. Subsequently, through multiple rounds of iterative calculations, the influence of these key parameters on the temperature uniformity and flow pressure drop optimization objectives was deeply explored. The temperature uniformity optimization objective was quantified by the standard deviation of the battery cell temperature. The standard deviation quantification formula is:

[0106]

[0107] In the formula, σ T T represents the standard deviation of the temperature of a single battery cell; a smaller value indicates better temperature uniformity. k The temperature of the kth battery cell is obtained by temperature sensors arranged on the surface of each battery cell. The average temperature of all battery cells is calculated from the arithmetic mean of all battery cell temperatures; N is the total number of battery cells, which is 100 in this embodiment. The flow pressure drop optimization target is quantified by calculating the pressure loss of the coolant flowing in the cooling channel, using the Darcy-Weisbach formula:

[0108]

[0109] In the formula, ΔP is the flow pressure drop; λ is the friction coefficient, which is related to the coolant flow state and the roughness of the channel wall; L is the total length of the cooling channel; d e ρ is the equivalent diameter of the cooling channel; v is the density of the coolant; and v is the average flow velocity of the coolant within the channel.

[0110] Furthermore, the improved genetic programming model automatically discovers the nonlinear coupling relationship between key parameters and optimization objectives by generating and evolving a set of procedures. Its improvements are reflected in three aspects: First, it introduces a multi-task learning mechanism to simultaneously learn the coupling relationship between key parameters and two optimization objectives: temperature uniformity and flow pressure drop. This improves learning efficiency and model generalization ability by sharing the parameter space. Second, it adopts a dynamic tree structure, adaptively adjusting the model's complexity and depth based on data characteristics. When the relationship between parameters and objectives is complex, the model automatically increases the tree depth and number of branches; when the relationship is relatively simple, it simplifies the model structure to avoid overfitting or underfitting. Third, it introduces an L2 regularization term to constrain parameter weights during model training, improving the model's prediction accuracy and stability. The regularization term expression is: Where λ is the regularization coefficient, with a value of 0.001, w i These represent the weights of the model parameters. The model's evolution process includes selection, crossover, and mutation operations. The program's performance is evaluated based on the fitness function, which aims to minimize the deviation between the model's predicted and actual values. The mean squared error (MSE) is used as the fitness function, and its expression is:

[0111]

[0112] In the formula, K is the sample size, and y i These are actual measured values. The fitness function value is used as the model prediction value. Through multiple rounds of iterative training, the fitness function value converges to below the preset threshold of 0.01, and finally a coupled model describing the relationship between parameter changes and target performance changes is formed.

[0113] Furthermore, based on the constructed coupling relationship model, an improved non-dominated sorting genetic algorithm is used for multi-objective collaborative optimization. The solution process includes constructing a collaborative optimization objective function that considers both improving temperature uniformity and reducing flow pressure drop, and selecting the optimal parameter combination that simultaneously satisfies the performance requirements of multiple optimization objectives within a preset constraint range. This process can achieve synergistic improvement of multiple objectives and avoid performance imbalance caused by single-objective optimization. First, the constraint range is defined. Based on the physical value range of key parameters and the minimum performance requirements of the optimization objectives, the following constraints are set: cooling channel width 8-12mm, channel height 4-6mm, channel spacing 15-25mm, cooling plate thickness 3-5mm, coolant inlet temperature 20-30℃, inlet flow rate 5-15L / min, flow pressure 0.1-0.3MPa; temperature uniformity standard deviation σ T ≤2℃; flow pressure drop ΔP≤5kPa.

[0114] Furthermore, a collaborative optimization objective function is constructed, which balances the improvement of temperature uniformity and the reduction of flow pressure drop. The collaborative optimization objective function is as follows:

[0115] F(X) = w1·f1(X) + w2·f2(X);

[0116] In the formula, X is the design parameter combination vector, X = [x1, x2, x3, x4, x5, x6, x7], where x1 is the cooling channel width, x2 is the channel height, x3 is the channel spacing, x4 is the cooling plate thickness, x5 is the coolant inlet temperature, x6 is the inlet flow rate, and x7 is the flow pressure; f1(X) is the objective function for optimizing temperature uniformity, f1(X) = σ T The objective is to minimize the value of this function; f2(X) is the objective function for optimizing the flow pressure drop, f2(X) = ΔP, and the objective is also to minimize the value of this function; w1 and w2 are dynamic weight coefficients, and they satisfy w1 + w2 = 1. The weight coefficients are set using the analytic hierarchy process (AHP) combined with a dynamic adjustment mechanism. First, the initial weights are determined using the AHP, a judgment matrix is ​​constructed, and the weight vector is calculated. If the initial system temperature uniformity problem is more prominent, w1 = 0.6 and w2 = 0.4 are set. If the flow pressure drop loss is large, w1 = 0.4 and w2 = 0.6 are set. During the optimization process, the weights are dynamically adjusted according to the target performance of each generation of the population. When the improvement in temperature uniformity is significantly lower than that in flow pressure drop, w1 is appropriately increased, and vice versa.

[0117] Furthermore, within the constraints, the parameter combination is iteratively optimized through multiple rounds using an improved non-dominated sorting genetic algorithm. The improvements include three aspects: First, an elite retention strategy is introduced, directly retaining the top 10% of non-dominated solutions from each generation to the next, ensuring the inheritance of optimal solutions and avoiding the loss of superior genes. Second, an adaptive crossover and mutation rate mechanism is adopted, dynamically adjusting the operating parameters according to the population's evolutionary state. The initial crossover rate is set to 0.8, and the initial mutation rate to 0.1. When the population similarity is high, the crossover rate is increased to 0.9, and the mutation rate is decreased to 0.05. When the population diversity is insufficient, the crossover rate is decreased to 0.7, and the mutation rate is increased to 0.15 to balance the population convergence speed and diversity. Third, multimodal optimization capabilities are integrated. By setting a niche radius, multiple optimal solution clusters are searched and identified on the Pareto front, ensuring the richness and practicality of the optimization results.

[0118] Further, the specific optimization process of the algorithm is as follows: First, the population is initialized with a size of 200. Each individual corresponds to a set of parameter combinations, and the parameter values ​​are randomly generated within the constraints. Next, non-dominated sorting is performed. Based on the collaborative optimization objective function value, the current population is divided into multiple non-dominated levels, with the first level of non-dominated solutions being the current optimal solution. Then, the crowding degree is calculated for individuals in the same non-dominated level. The crowding degree reflects the density of an individual in the solution space and is determined by calculating the number of other individuals within a certain range around the individual. The higher the crowding degree, the denser the solution in that area, and vice versa. The purpose of calculating the crowding degree is to maintain... To ensure the diversity of solutions and prevent the algorithm from converging to local optima, a selection operation is then performed. Individuals are selected based on their non-dominated level and crowding, prioritizing those with high non-dominated levels and low crowding to enter the next generation. A roulette wheel selection method combined with an elite retention strategy is used to ensure fairness in selection and the preservation of the optimal solution. Finally, crossover and mutation operations are performed. The crossover operation uses a single-point crossover method, randomly selecting the crossover point of two individuals, exchanging the parameter fragments after the crossover point, and generating a new individual. The mutation operation uses a random mutation method, randomly perturbing some parameters of individuals within the constraints to generate new parameter combinations, forming a new parameter combination population. The above iterative process is repeated for a preset number of iterations of 100 generations. The iteration stops when the preset number of iterations is reached or the change in the optimal objective function value of the population for 10 consecutive generations is less than 0.001. The optimal parameter combination that simultaneously satisfies multiple optimization objective performance requirements is then selected.

[0119] Furthermore, simulation verification was performed on the optimal parameter combination. If the simulation results did not meet the preset comprehensive performance requirements, the parameter weights in the coupling relationship model and the iterative parameters of the improved non-dominated sorting genetic algorithm were adjusted according to the deviation direction of the simulation results. The multi-objective collaborative optimization solution steps were then re-executed, and the reliability and effectiveness of the optimization results were ensured through closed-loop optimization. The selected optimal parameter combination was substituted into the simulation model of the battery thermal management system. The simulation model was constructed using ANSYS Icepak software, modeled according to the actual battery pack structure, cooling channel layout, and material parameters. The system's operating state under these parameter settings was simulated, with the simulation condition set as the battery pack charging and discharging at a 1C rate under normal temperature conditions. During the simulation, the temperature distribution data and flow pressure drop loss data output by the simulation model were detected. The temperature distribution data included the surface temperature of each battery cell and the overall temperature standard deviation of the battery pack. The flow pressure drop loss data included the pressure difference between the inlet and outlet of the cooling channel. The detected data were compared with the preset temperature uniformity requirements and flow pressure drop requirements. The preset temperature uniformity standard deviation σ T ≤2℃, flow pressure drop ΔP≤5kPa.

[0120] If the simulation results meet the preset comprehensive performance requirements, then the parameter combination is the final optimal parameter combination; if the simulation results do not meet the preset requirements, the relevant parameters need to be adjusted according to the direction of the deviation in the simulation results. If the temperature uniformity is poor (i.e., the simulated temperature standard deviation is greater than a preset threshold), increase the dynamic weight coefficient w1 corresponding to the temperature uniformity optimization objective in the co-optimization objective function by 0.1-0.2, and simultaneously increase the iteration number of the improved non-dominated sorting genetic algorithm by 20-50 generations. If the flow pressure drop is poor (i.e., the simulated flow pressure drop is greater than a preset threshold), increase the dynamic weight coefficient w2 corresponding to the flow pressure drop optimization objective in the co-optimization objective function by 0.1-0.2, and appropriately increase the population size of the algorithm by 50-100 to improve the search range. If neither of these requirements is met, adjust w1 and w2 proportionally according to the degree of deviation, and adjust the mutation rate of the algorithm by 0.02-0.05 to enhance population diversity. After adjustment, re-execute the multi-objective co-optimization solution steps until the simulation results meet the preset comprehensive performance requirements.

[0121] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A multi-objective collaborative optimization method for a battery thermal management system, characterized in that: Includes the following steps: S1. Collect all key parameters that affect the performance of the battery thermal management system. These key parameters include the structural parameters of the cooling channels and the operating parameters of the coolant. S2. Based on the collected key parameters, an improved genetic programming model is used to construct a coupling relationship model between the key parameters and multiple optimization objectives. The coupling relationship model describes the correspondence between parameter changes and optimization objective performance changes. The optimization objectives include temperature uniformity and flow pressure drop. S3. Based on the coupling relationship model, an improved non-dominated sorting genetic algorithm is used to solve the multi-objective collaborative optimization problem. The solution process includes constructing a collaborative optimization objective function that takes into account both the improvement of temperature uniformity and the reduction of flow pressure drop, and selecting the optimal parameter combination that simultaneously meets the performance requirements of multiple optimization objectives within the preset constraint range. S4. Perform simulation verification on the optimal parameter combination. If the simulation results do not meet the preset comprehensive performance requirements, adjust the parameter weights in the coupling relationship model and the iterative parameters of the improved non-dominated sorting genetic algorithm according to the deviation direction of the simulation results, and re-execute the multi-objective collaborative optimization solution steps.

2. The multi-objective collaborative optimization method for the battery thermal management system according to claim 1, characterized in that: In step S1, the comprehensive collection of key parameters affecting the performance of the battery thermal management system includes: Collect the structural parameters of the cooling channel, including the geometric parameters of the cooling channel and the structural parameters of the cooling plate; Collect the operating parameters of the coolant, including the inlet temperature, inlet flow rate, and flow pressure. Parameters are collected using a combination of real-time monitoring and static measurement.

3. The multi-objective collaborative optimization method for the battery thermal management system according to claim 1, characterized in that: In step S2, based on the collected key parameters, an improved genetic programming model is used to construct a coupling relationship model between the key parameters and multiple optimization objectives, including: Key parameters are classified and preprocessed to eliminate redundant parameters; Through multiple rounds of iterative calculations, the influence of key parameters on the optimization objectives of temperature uniformity and flow pressure drop is deeply explored. The optimization objective of temperature uniformity is quantified by the standard deviation of the cell temperature, and the formula for standard deviation quantification is as follows: Where, σ T T represents the standard deviation of the temperature of a single battery cell. k Let K be the temperature of the k-th cell. The average temperature of all battery cells is N, where N is the total number of battery cells. A coupled model is formed to describe the relationship between parameter changes and target performance changes.

4. The multi-objective collaborative optimization method for the battery thermal management system according to claim 1, characterized in that: In step S2, the improved genetic programming model automatically discovers the nonlinear coupling relationship between key parameters and optimization objectives by generating and evolving a set of programs; the evolution process includes selection, crossover, and mutation operations, and the program performance is evaluated based on the fitness function, which aims to minimize the deviation between the model's predicted values ​​and the actual values; The model structure has adaptive adjustment capabilities.

5. The multi-objective collaborative optimization method for the battery thermal management system according to claim 1, characterized in that: In step S3, the multi-objective collaborative optimization solution based on the coupling relationship model and using an improved non-dominated sorting genetic algorithm includes: Based on the physical value range of key parameters and the minimum performance requirements of the optimization objective; A collaborative optimization objective function is constructed, which balances the improvement of temperature uniformity and the reduction of flow pressure drop. The collaborative optimization objective function is as follows: F(X) = w1·f1(X) + w2·f2(X); Where X is the design parameter combination vector, f1(X) is the objective function for optimizing temperature uniformity, f2(X) is the objective function for optimizing flow pressure drop, and w1 and w2 are dynamic weighting coefficients, satisfying w1+w2=1; Within the constraints, the parameter combination is iteratively optimized through multiple rounds using an improved non-dominated sorting genetic algorithm to select the optimal parameter combination that simultaneously satisfies the performance requirements of multiple optimization objectives.

6. The multi-objective collaborative optimization method for the battery thermal management system according to claim 1, characterized in that: The parameter combination is optimized through multiple rounds of iterative optimization using an improved non-dominated sorting genetic algorithm, including: Perform non-dominated sorting and divide the current population into multiple non-dominated levels; Crowding degree is calculated for individuals at the same non-dominated level to maintain solution diversity; Perform a selection operation, filtering individuals based on non-dominated hierarchy and crowding; Perform crossover and mutation operations to generate a new population with different parameter combinations.

7. The multi-objective collaborative optimization method for the battery thermal management system according to claim 1, characterized in that: In step S4, the simulation verification of the optimal parameter combination includes: The optimal parameter combination is substituted into the simulation model of the battery thermal management system to simulate the system's operating state under these parameter settings. The temperature distribution data and flow pressure drop loss data output by the detection simulation model were examined. The detected data is compared with preset temperature uniformity requirements and flow pressure drop requirements.

8. The multi-objective collaborative optimization method for the battery thermal management system according to claim 1, characterized in that: In step S4, if the simulation results do not meet the preset comprehensive performance requirements, the parameter weights in the coupling relationship model and the iterative parameters of the improved non-dominated sorting genetic algorithm are adjusted according to the deviation direction of the simulation results, including: If the temperature uniformity performance is poor, increase the dynamic weight coefficient corresponding to the temperature uniformity optimization objective in the collaborative optimization objective function; If the flow pressure drop performance is poor, increase the dynamic weight coefficient corresponding to the flow pressure drop optimization objective in the collaborative optimization objective function; Based on the difference between the simulation results and the preset performance requirements, the number of iterations, population size, or mutation rate of the improved non-dominated sorting genetic algorithm are adaptively adjusted.

9. The multi-objective collaborative optimization method for the battery thermal management system according to claim 1, characterized in that: The improvement in the genetic programming model lies in: A multi-task learning mechanism is introduced to simultaneously learn the coupling relationship between key parameters and multiple optimization objectives; A dynamic tree structure is adopted to adaptively adjust the complexity and depth of the model in order to describe the correspondence between parameter changes and target performance changes; By introducing a regularization term, the prediction accuracy and stability of the model can be improved.

10. The multi-objective collaborative optimization method for the battery thermal management system according to claim 1, characterized in that: The improved non-dominated sorting genetic algorithm is improved in the following aspects: An elite retention strategy is introduced to preserve and pass on the optimal solution; An adaptive crossover rate and mutation rate mechanism is adopted to dynamically adjust the operation parameters according to the population evolutionary state; It integrates multimodal optimization capabilities to search for and identify multiple optimal solution clusters on the Pareto front.

Images