Battery pack optimization design method based on improved multi-objective particle swarm optimization

Through the improved multi-objective particle swarm algorithm, combined with the objective function and protection mechanism with constraints, the battery pack design is optimized, and the problem of multi-objective optimization in the battery pack design is solved, achieving the best balance of quality, stiffness and strength and design schemes. accuracy.

CN118917173APending Publication Date: 2024-11-08JIANGSU UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410919102.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-10
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

When facing multi-objective optimization, existing battery pack designs are difficult to effectively take into account quality, stiffness and strength, and the particle search capability is limited, resulting in insufficient optimization of the design scheme.

Method used

Using the improved multi-objective particle swarm algorithm, by establishing a constrained objective function and protection mechanism, the battery pack design is optimized, ensuring that particles fly in the search space, generating non-dominant solutions, and using the difference function to find the comprehensive index to obtain the most Excellent design plan.

Benefits of technology

The optimal design scheme that takes into account the quality, stiffness and strength of the battery pack is achieved, ensuring that the design scheme meets actual requirements, and ensuring the accuracy and completeness of the optimal solution through the use of protection mechanisms and difference functions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118917173A_ABST
    Figure CN118917173A_ABST
Patent Text Reader

Abstract

The invention discloses a battery pack optimization design method based on an improved multi-objective particle swarm optimization algorithm, and the method comprises the steps: building a plurality of objective functions constrained by a battery pack, and determining the constraint condition of the battery pack; performing optimization design on the battery pack by using an improved multi-target particle swarm algorithm with a protection mechanism, and solving an external file, namely an excellent design scheme set of the battery pack; wherein each particle of the improved multi-target particle swarm algorithm represents a design scheme of the battery pack; and in the solved external file, a comprehensive index is solved by using a difference function, so that the solution with the minimum comprehensive index value in the external file is the optimal solution taking quality, rigidity and strength into consideration, namely the most excellent battery pack design scheme. According to the method, the battery pack optimization design scheme can be accurately and quickly obtained.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of battery pack optimization design, and specifically relates to a battery pack optimization design method based on an improved multi-objective particle swarm algorithm. Background Art

[0002] As the core power component of new energy vehicles, battery packs have become a key factor affecting the performance and market competitiveness of new energy vehicles. From the past to the present, the automotive field has always focused on the lightweight design route of high stiffness and light weight in product design, which has brought significant economic benefits and competitiveness such as reduced manufacturing costs and low fuel consumption. Similarly, today's battery pack design, in the face of various national guidelines and subsidy policies, also urgently needs to work hard on the energy density of the monomer mass, and maintain its market competitiveness through continuous technological innovation. That is, follow the development route of light weight and high energy density.

[0003] When facing multi-objective optimization problems, many excellent algorithms have shown good performance. However, compared with other classic algorithms, the particle swarm algorithm has stronger search capabilities and a wider search space, because the particle parallel search mode adopted by the particle swarm algorithm can obtain multiple non-dominated solutions at the same time, which is convenient for the composition of non-dominated solution sets. The particle swarm algorithm has low requirements for the optimization object and has strong universality. It can also be combined with other excellent algorithms to jointly deal with multi-objective problems, which makes the multi-objective particle swarm algorithm attract more and more scholars' attention and gradually become a hot issue explored by the academic community. Summary of the invention

[0004] In view of this, the present invention provides a battery pack optimization design method based on an improved multi-objective particle swarm algorithm.

[0005] The present invention achieves the above technical objectives through the following technical means.

[0006] A battery pack optimization design method based on improved multi-objective particle swarm optimization algorithm:

[0007] Establish multiple objective functions with constraints for the battery pack and determine the constraints of the battery pack;

[0008] An improved multi-objective particle swarm algorithm with a protection mechanism is used to optimize the design of the battery pack and obtain an external archive, i.e., a set of excellent design solutions for the battery pack; each particle of the improved multi-objective particle swarm algorithm represents a design solution for the battery pack;

[0009] In the obtained external files, the comprehensive index W is obtained using the difference function. The solution in the external file that minimizes the comprehensive index W is the optimal solution that takes into account mass, stiffness and strength, that is, the best battery pack design solution.

[0010] The improved multi-objective particle swarm algorithm includes:

[0011] Initialize the particle swarm, randomly assign values ​​to the speed and position of the particles in the swarm, and then use the polynomial function to find the strength, stiffness and mass of each particle according to the size of the particle's position. Then use the objective function to find the objective function value, and find the non-dominated solution of the particle swarm according to the objective function value of each particle, and add the non-dominated solution to the external archive;

[0012] The initial position of the particle is determined as the initial individual optimal position of the particle, and the obtained non-dominated solution is selected as the initial global optimal position of the particle;

[0013] Use protection mechanisms to ensure that particles fly within the search space, update particle speed and position, reselect individual optimal positions of particles, and update the global optimal position;

[0014] Maintain and update external archives;

[0015] If the specified maximum number of iterations is reached, the search stops.

[0016] Furthermore, the multiple objective functions of the battery pack with constraints are:

[0017] f1=-|F obj1 |=K max -K

[0018] f2=-|F obj2 |=σ max -σ

[0019] f3=m

[0020] Where f1 is the stiffness objective function value of the battery pack, K is the stiffness of the battery pack, and K max Indicates the maximum stiffness required by the battery pack under various working conditions, -|F obj1 | indicates that the result on the right side of the equation must be less than or equal to 0; f2 is the strength objective function value of the battery pack, σ is the strength of the battery pack, σ max Indicates the maximum strength of the battery pack under various working conditions, -|F obj2 | indicates that the result on the right side of the equation must be less than or equal to 0; f3 is the quality objective function value of the battery pack, and m represents the quality of the battery pack.

[0021] Furthermore, the constraints of the battery pack include: the thickness constraint of the battery pack body: h1≤h≤h m , battery pack width constraint: s1≤s≤s m , where h1 and h m are the upper and lower boundaries of the thickness, s1 and s mThey are the upper and lower boundaries of the width respectively.

[0022] Furthermore, the fast dominance strategy is used to find the non-dominated solution of the particle swarm, which is:

[0023] A particle x is randomly selected from the population, and the objective function values ​​of particle x are compared with those of other particles in the population. If particle x is dominated by a certain particle, x is updated to the certain particle. After the update, the particle continues to be compared with the objective function values ​​of other particles until it is no longer dominated by other particles in the population and is placed in an external archive. If other particles in the population are dominated by x, particle x does not change, and other particles are updated with x, and particle x is placed in an external archive. At this time, all particles in the external archive are the final non-dominated solution.

[0024] Furthermore, the protection mechanism is as follows: when a particle flies out of the boundary, the position of the particle becomes the previous position of the particle, and the speed of the particle becomes half of the original speed. If the previous position of the particle also flies out of the boundary, the position of the particle becomes the previous position of the previous position of the particle, and the speed is multiplied by 1 / 2 until the position of the particle is within the boundary.

[0025] Furthermore, the reselection of the optimal individual position of the particle is specifically as follows: when particle i obtains a new position x i When the new position x i The optimal position of the dominant individual p i , then the individual optimal position of particle i is x i If p i Dominate x i , p i is still the individual optimal position of particle i; if p i With x i If they are not dominated by each other, then at the individual optimal position p i A new particle j is generated on the surface, the speed of particle j is equal to the speed of particle i, and the individual optimal position of particle j is p i , and the individual optimal position of particle i is x i .

[0026] Furthermore, the maintenance and updating of external archives is specifically as follows:

[0027] After the update, for each newly generated non-dominated solution y, if y dominates some solutions in the external archive P, these dominated partial solutions are deleted and y is saved in the external archive;

[0028] For the remaining newly generated non-dominated solutions y' that cannot dominate some solutions in the external archive, if the scale of P does not reach the set maximum value, the non-dominated solution y' is directly added to the external archive; if the scale of P reaches the set maximum value, the fast domination strategy is executed in the external archive and y', and the dominated particles are directly deleted; if the total number of solutions after deletion is less than or equal to the maximum scale of the external archive, all of them are retained in the external archive; if the total number is greater than the maximum scale of the external archive, a solution d is randomly selected from the external archive, its difference function value is calculated, and compared with the difference function value of y'. If all y' can be eliminated, the randomly selected solution will still enter the external archive; the newly generated non-dominated solution whose difference function value in y' is less than the difference function value of the randomly selected solution is recorded as z. If the number of z exceeds the number of solutions vacated in the external archive, another solution is randomly selected from the external archive for comparison, and d is also eliminated, until the number of newly generated non-dominated solutions is less than or equal to the number of solutions vacated in the external archive.

[0029] Furthermore, the difference function is:

[0030] W=(|f 1.i -f 1.min |+1)*(|f 2.i -f 2.min |+1)*(|f 3.i -f 3.min |+1)

[0031] Where: f 1.i 、f 2.i 、f 3.i represents the objective function of particle i, f 1.min represents the minimum objective function related to stiffness in the optimal solution of the external archive, f 2.min represents the minimum objective function related to strength in the optimal solution of the external archive, f 3.min It represents the minimum objective function regarding quality in the optimal solution of the external archive.

[0032] The beneficial effects of the present invention are:

[0033] (1) The present invention adopts a constrained objective function so that each design scheme of the battery pack in the obtained external file can meet the actual requirements;

[0034] (2) A protection mechanism is used to ensure that particles fly within the search space, so that particles will not lose the comparison qualification due to flying out of the search space, and the non-dominated solution in the external archive is finally obtained by comparing all solutions;

[0035] (3) When the particle updates the individual optimal position, if the individual optimal position and the new individual optimal position do not dominate each other, the present invention retains the non-dominated solution by generating a new particle to inherit one of the individual optimal positions, thereby ensuring that the global optimal position is selected by comparing the optimal positions of all individuals;

[0036] (4) When the present invention maintains and updates the external file, when the size of the external file reaches the maximum, the difference function is used in advance for comparison to ensure that each design scheme of the battery pack is compared;

[0037] (5) The present invention uses a difference function to calculate a comprehensive index, and takes the solution in the external file with the smallest comprehensive index value as the best battery pack design solution, and obtains the optimal design solution through simple calculations. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 The overall flow chart of the battery pack optimization design based on the improved multi-objective particle swarm algorithm of the present invention;

[0039] Figure 2 This is a flow chart of the improved multi-objective particle swarm algorithm described in the present invention. DETAILED DESCRIPTION

[0040] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments, but the protection scope of the present invention is not limited thereto.

[0041] like Figure 1 As shown, the present invention provides a battery pack optimization design method based on an improved multi-objective particle swarm algorithm, which specifically includes the following steps:

[0042] Step (1), collecting data of multiple battery packs of the same model, including the size, mass, stiffness and strength of the battery packs, and the stiffness and strength required under various working conditions; using the obtained data to fit a polynomial function relationship between strength, stiffness and mass and the thickness of the battery pack case and the width of the battery pack;

[0043] The polynomial function has the following form:

[0044] K=a1*h 4 +a2*h 3 +…+a5+a6*s 4 +...+a9*s 1

[0045] σ=b1*h 4 +b2*h 3 +…+b5+b6*s 4 +...+b9*s 1

[0046] m=c1*(h*s)4 +c2*(h*s) 3 +...+c5

[0047] Among them, h is the thickness of the battery pack box, s is the width of the battery pack, K is the stiffness, m is the mass, σ is the strength, a1…a9, b1…b9, c1…c9 are all coefficients of the polynomial, which are obtained through data fitting.

[0048] Based on the collected data, multiple objective functions with constraints for the battery pack are established, and the constraints of the battery pack are determined.

[0049] The multi-objective optimization of the battery pack takes the thickness of the battery pack as the design variable, and takes the minimum mass, maximum strength and maximum stiffness of the battery pack as the goal. When setting the objective function, the objective function is constrained, that is, the stiffness of the battery pack must be greater than or equal to the required maximum stiffness, and the strength of the battery pack must be greater than or equal to the required maximum strength. The corresponding objective function is as follows:

[0050] f1=-|F obj1 |=K max -K

[0051] f2=-|F obj2 |=σ max -σ

[0052] f3=m

[0053] Where f1 is the stiffness objective function value of the battery pack, K is the stiffness of the battery pack, and K max Indicates the maximum stiffness required by the battery pack under various working conditions, -|F obj1 | indicates that the result on the right side of the equation must be less than or equal to 0. The greater the stiffness of the battery pack, the smaller f1. f2 is the strength objective function value of the battery pack, σ is the strength of the battery pack, and σ max Indicates the maximum strength of the battery pack under various working conditions, -|F obj2 | indicates that the result on the right side of the equation must be less than or equal to 0. The greater the strength of the battery pack, the smaller f2. f3 is the quality objective function value of the battery pack. m represents the quality of the battery pack. The smaller the quality of the battery pack, the smaller f3.

[0054] Determine the constraints of the battery pack: Battery pack box thickness constraint: h1≤h≤h m , where h1 and h m The upper and lower boundaries of the thickness respectively; the width constraint of the battery pack: s1≤s≤s m , where s1 and s m They are the upper and lower boundaries of the width respectively.

[0055] Step (2), using an improved multi-objective particle swarm algorithm with a protection mechanism to optimize the design of the battery pack and obtain the external file; wherein each particle of the improved multi-objective particle swarm algorithm represents a design scheme of the battery pack.

[0056] like Figure 2 As shown, the improved multi-objective particle swarm algorithm includes:

[0057] S1, initialize the particle swarm, randomly assign values ​​to the speed and position of the particles in the swarm, and then use the polynomial function to find the strength, stiffness and mass of each particle according to the size of the particle position, and then use the objective function to find the objective function value, and find the non-dominated solution of the particle swarm according to the objective function value of each particle, and add the non-dominated solution to the external archive;

[0058] The definition of particle dominance is as follows: x1,x2∈X, the solution x1 dominates x2 if and only if the following conditions hold:

[0059]

[0060] Among them, x1,x2∈X means x1 and x2 are particles in the particle group X, f i (x1),f i (x2), f j (x1),f j (x2) represents the objective function value of the particle;

[0061] The above conditions indicate that if all objective function values ​​of x1 are not greater than the objective function values ​​of x2, and there is at least one objective function value of x1 that is smaller than one objective function value of x2, then x1 dominates x2 and x1 is a non-dominated solution.

[0062] In order to improve the overall performance of the algorithm and reduce the convergence time of the algorithm as much as possible, a fast domination strategy is used to find the non-dominated solution of the particle swarm.

[0063] The description of the fast dominance strategy is as follows:

[0064] (1) Randomly select a particle x from the population;

[0065] (2) Compare the objective function values ​​of particle x with those of other particles in the population. If particle x is dominated by a certain particle, then update x to that particle. After the update, the particle continues to compare its objective function values ​​with those of other particles until it is no longer dominated by other particles in the population and is placed in an external archive. If all other particles in the population are dominated by x, then particle x remains unchanged, and all other particles are updated with x, and particle x is placed in an external archive. At this point, all particles in the external archive are the final non-dominated solution.

[0066] S2, the initial position of the particle is determined as the initial individual optimal position of the particle, and the non-dominated solution obtained in S1 is selected as the initial global optimal position of the particle.

[0067] The individual optimal position is the position of each particle when it finds a non-dominated solution during its iteration. Each particle records its current position and the non-dominated solution position it has reached previously.

[0068] The global optimal position is the position of the entire population when it finds a non-dominated solution during the entire search process, which is obtained by comparing the objective function value of the individual optimal position of each particle.

[0069] S3, use the protection mechanism to ensure that the particle flies within the search space. Under this premise, update the particle's speed and position, reselect the individual optimal position of the particle, and update the global optimal position;

[0070] The protection mechanism to ensure that particles do not fly out of the boundary is as follows: when a particle flies out of the boundary, the particle's position becomes the particle's previous position, and the particle's speed becomes half of the original speed. If the particle's previous position also flies out of the boundary, the particle's position becomes the particle's previous position, and the speed is multiplied by 1 / 2 until the particle's position is within the boundary. The particle's position and speed are updated according to the following formula:

[0071]

[0072] x i (t+1)=x i (t)+v i (t+1)

[0073] Among them, v i (t) is the velocity of particle i at time t, v i (t+1) is the velocity of particle i at time t+1, x i (t+1) is the position of particle i at time t+1, x i (t) is the position of particle i at time t, w is the inertia weight, which is used to balance local search and global search, c1 and c2 are acceleration constants. The existence of c1 can make the particle fly to its own optimal position, and the existence of c2 can make the particle fly to the global optimal position. r1 and r2 are acceleration factors, which are random numbers generated by the (0, 1) normal distribution. and x gbest They are the individual historical optimal value and the global historical optimal solution respectively.

[0074] When particle i obtains a new position x i When the new position x i The optimal position of the dominant individual p i , then the individual optimal position of particle i is xi , if p i Dominate x i , p i It is still the individual optimal position of particle i. If p i With x i If they are not dominated by each other, then at the individual optimal position p i A new particle j is generated on the surface, the speed of particle j is equal to the speed of particle i, and the individual optimal position of particle j is p i , and the individual optimal position of particle i is x i .

[0075] It is obtained by comparing the objective function value of the individual optimal position of each particle after update.

[0076] S4, maintain and update external archives. The specific process is as follows:

[0077] (1) After S3 update, for each newly generated non-dominated solution y, if y dominates some solutions in the external archive P, delete these dominated partial solutions and save y in the external archive;

[0078] (2) For the remaining newly generated non-dominated solutions y' that cannot dominate some solutions in the external archive, if the size of P does not reach the set maximum value, the non-dominated solution y' is directly added to the external archive. If the size of P reaches the set maximum value, the fast domination strategy is executed in the external archive and y', and the dominated particles are directly deleted. If the total number of solutions after deletion is less than or equal to the maximum size of the external archive, all of them are retained in the external archive. If the total number is greater than the maximum size of the external archive, a solution d is randomly selected from the external archive, and the difference function of step (3) is executed in advance. The number of solutions is compared with the difference function value of y'. If all y' can be eliminated, the randomly selected solution will still enter the external archive, and the newly generated non-dominated solution whose difference function value in y' is less than the difference function value of the randomly selected solution is recorded as z. If the number of z exceeds the number of solutions vacated in the external archive, a solution e is randomly selected from the external archive for comparison, and d is also eliminated, until the number of new non-dominated solutions screened out is less than or equal to the number of solutions vacated in the external archive; here, the minimum objective function value in the difference function is selected from the external archive and the newly generated non-dominated solution;

[0079] S5: If the specified maximum number of iterations is reached, stop searching; otherwise, go to S3.

[0080] The final external archive is a collection of excellent design solutions for battery packs.

[0081] Step (3), in the obtained external file, the difference function is used to obtain the comprehensive index W. The solution in the external file that minimizes the W value is the optimal solution that takes into account mass, stiffness and strength, that is, the best battery pack design. The difference function formula is as follows:

[0082] W=(|f 1.i -f 1.min |+1)*(|f 2.i -f 2.min |+1)*(|f 3.i -f 3.min |+1)

[0083] Where: f 1.i 、f 2.i 、f 3.i represents the objective function of particle i, f 1.min represents the minimum objective function related to stiffness in the optimal solution of the external archive, f 2.min represents the minimum objective function related to strength in the optimal solution of the external archive, f 3.min It represents the minimum objective function regarding quality in the optimal solution of the external archive.

[0084] The embodiments are preferred implementations of the present invention, but the present invention is not limited to the above-mentioned implementations. Any obvious improvements, substitutions or modifications that can be made by those skilled in the art without departing from the essential content of the present invention belong to the protection scope of the present invention.

Claims

1. A battery pack optimization design method based on an improved multi-objective particle swarm algorithm, characterized in that: Establish multiple objective functions with constraints for the battery pack and determine the constraints of the battery pack; An improved multi-objective particle swarm algorithm with a protection mechanism is used to optimize the design of the battery pack and obtain an external archive, i.e., a set of excellent design solutions for the battery pack; each particle of the improved multi-objective particle swarm algorithm represents a design solution for the battery pack; In the obtained external files, the comprehensive index W is obtained using the difference function. The solution in the external file that minimizes the comprehensive index W is the optimal solution that takes into account mass, stiffness and strength, that is, the best battery pack design solution. The improved multi-objective particle swarm algorithm includes: Initialize the particle swarm, randomly assign values ​​to the speed and position of the particles in the swarm, and then use the polynomial function to find the strength, stiffness and mass of each particle according to the size of the particle's position. Then use the objective function to find the objective function value, and find the non-dominated solution of the particle swarm according to the objective function value of each particle, and add the non-dominated solution to the external archive; The initial position of the particle is determined as the initial individual optimal position of the particle, and the obtained non-dominated solution is selected as the initial global optimal position of the particle; Use protection mechanisms to ensure that particles fly within the search space, update particle speed and position, reselect individual optimal positions of particles, and update the global optimal position; Maintain and update external archives; If the specified maximum number of iterations is reached, the search stops.

2. The battery pack optimization design method according to claim 1, characterized in that: The multiple objective functions of the battery pack with constraints are: f1=-|F obj1 |=K max -K f2=-|F obj2 |=s max -s f3=m Where f1 is the stiffness objective function value of the battery pack, K is the stiffness of the battery pack, and K max Indicates the maximum stiffness required by the battery pack under various working conditions, -|F obj1 | indicates that the result on the right side of the equation must be less than or equal to 0; f2 is the strength objective function value of the battery pack, σ is the strength of the battery pack, σ max Indicates the maximum strength of the battery pack under various working conditions, -|F obj2 | indicates that the result on the right side of the equation must be less than or equal to 0; f3 is the quality objective function value of the battery pack, and m represents the quality of the battery pack.

3. The battery pack optimization design method according to claim 1, characterized in that: The constraints of the battery pack include: the thickness constraint of the battery pack body: h1≤h≤h m , battery pack width constraint: s1≤s≤s m , where h1 and h m are the upper and lower boundaries of the thickness, s1 and s m They are the upper and lower boundaries of the width respectively.

4. The battery pack optimization design method according to claim 1, characterized in that: The non-dominated solution of the particle swarm is obtained by using the fast dominance strategy, which is as follows: A particle x is randomly selected from the population, and the objective function values ​​of particle x are compared with those of other particles in the population. If particle x is dominated by a certain particle, x is updated to the certain particle. After the update, the particle continues to be compared with the objective function values ​​of other particles until it is no longer dominated by other particles in the population and is placed in an external archive. If other particles in the population are dominated by x, particle x does not change, and other particles are updated with x, and particle x is placed in an external archive. At this time, all particles in the external archive are the final non-dominated solution.

5. The battery pack optimization design method according to claim 1, characterized in that: The protection mechanism is as follows: when a particle flies out of the boundary, the position of the particle becomes the previous position of the particle, and the speed of the particle becomes half of the original speed. If the previous position of the particle also flies out of the boundary, the position of the particle becomes the previous position of the previous position of the particle, and the speed is multiplied by 1 / 2 until the position of the particle is within the boundary.

6. The battery pack optimization design method according to claim 1, characterized in that: The reselection of the optimal individual position of the particle is specifically as follows: when particle i obtains a new position x i When the new position x i The optimal position of the dominant individual p i , then the individual optimal position of particle i is x i If p i Dominate x i , p i is still the individual optimal position of particle i; if p i With x i If they are not dominated by each other, then at the individual optimal position p i A new particle j is generated on the surface, the speed of particle j is equal to the speed of particle i, and the individual optimal position of particle j is p i , and the individual optimal position of particle i is x i .

7. The battery pack optimization design method according to claim 1, characterized in that: The maintenance and updating of external archives are as follows: After the update, for each newly generated non-dominated solution y, if y dominates some solutions in the external archive P, these dominated partial solutions are deleted and y is saved in the external archive; For the remaining newly generated non-dominated solutions y' that cannot dominate some solutions in the external archive, if the size of P does not reach the set maximum value, the non-dominated solution y' is directly added to the external archive; if the size of P reaches the set maximum value, the fast domination strategy is executed in the external archive and y', and the dominated particles are directly deleted; if the total number of solutions after deletion is less than or equal to the maximum size of the external archive, all of them are retained in the external archive; If the total number is greater than the maximum size of the external archive, a solution d is randomly selected from the external archive, its difference function value is calculated, and compared with the difference function value of y'. If all y' can be eliminated, the randomly selected solution will still enter the external archive; the newly generated non-dominated solution whose difference function value in y' is less than the difference function value of the randomly selected solution is recorded as z. If the number of z exceeds the number of solutions vacated in the external archive, another solution is randomly selected from the external archive for comparison, and d is also eliminated until the number of newly generated non-dominated solutions is less than or equal to the number of solutions vacated in the external archive.

8. The battery pack optimization design method according to claim 7, characterized in that: The difference function is: W=(|f 1.i -f 1.min |+1)*(|f 2.i -f 2.min |+1)*(|f 3.i -f 3.min |+1) Where: f 1.i 、f 2.i 、f 3.i represents the objective function of particle i, f 1.min represents the minimum objective function related to stiffness in the optimal solution of the external archive, f 2.min represents the minimum objective function related to strength in the optimal solution of the external archive, f 3.min It represents the minimum objective function regarding quality in the optimal solution of the external archive.