Method for optimizing aluminum material roll forming springback

Abstract

To provide a method for optimizing aluminum material roll forming springback that improves accuracy of roll forming technique, improves production efficiency, and increases production flexibility.SOLUTION: This method includes the processes in which: process of aluminum material roll forming is analyzed; determination variable of aluminum material springback and thickness decrease is confirmed; an aluminum material roll forming springback optimization model based on a multi-target optimization algorithm is constructed; constraint conditions in an aluminum material roll forming process are clarified, a corresponding algorithm is submitted on the basis of features of the model, and optimization is performed for aluminum material springback and thickness decrease; renewal is performed for each determination variable affecting aluminum material springback and thickness decrease; and means for determining repetition stop condition and improving roll formed aluminum material spring back is outputted.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to the technical field of rebound during roll bending of aluminum materials, and more particularly to a method for optimizing the amount of rebound during roll bending of aluminum materials. [Background technology]

[0002] In the high-speed railway equipment manufacturing industry, the requirements for train nose structural strength, dimensional accuracy, and product quality are becoming increasingly stringent. Due to its lightweight, high strength, and corrosion resistance, aluminum is widely used for large structural components such as the front car of a train. However, due to the complexity of the interface of the side member of a train nose, as well as the characteristics of its large bending modulus, length, and curvature changes, conventional bending equipment and processes cannot meet the demand for high-quality forming when dealing with aluminum materials with such complex cross sections and low deformation resistance. Therefore, how to improve the processing accuracy of aluminum materials is particularly important at this time.

[0003] In the forming process of large structural components such as the lead car of a train, a roll bending machine generates a large amount of rebound when bending aluminum material, which affects various aspects such as the processed shape and processing accuracy. To reduce the rebound, it is necessary to change the thickness of the material. How to ensure that the aluminum material can not only achieve the expected bent shape during the bending process but also maintain a sufficient thickness, that is, how to precisely control the process parameters during processing, is of great significance.

[0004] Multi-objective optimization algorithms are swarm intelligence algorithms that solve optimization problems with multiple conflicting or interrelated objective functions. Such problems require simultaneous consideration of multiple optimization objectives, and there may be inter-constraint relationships between these objectives, making it difficult to simultaneously achieve single-objective optimization. The goal is to find a set of solutions that balance and satisfy each objective, rather than a single optimal solution. Multi-objective optimization algorithms have greater flexibility and adaptability. In practical applications, multi-objective optimization algorithms have been widely applied in various fields, such as construction design and logistics planning, to find the optimal balance between multiple objectives and maximize overall profits. Summary of the Invention

[0005] The objective of the present invention is to provide a method for optimizing the bending rebound of aluminum material roll bending, which solves the problem of relatively large bending rebound in roll bending technology, improves the precision of roll bending technology, and has the advantages of improving production efficiency and enhancing production flexibility.

[0006] To achieve the above object, the present invention provides a method for optimizing the bending rebound amount during roll bending of aluminum material, which includes the following steps: In step S1, the aluminum material roll bending process is analyzed, and the determining variables for the rebound amount and thickness reduction amount of the aluminum material are determined. In S2, a rebound optimization model for roll bending of aluminum materials is constructed based on a multi-objective optimization algorithm, and the constraint conditions in the roll bending process of aluminum materials are clarified. In step S3, a corresponding algorithm is proposed based on the characteristics of the model in step S2, and the rebound amount and thickness reduction amount of the aluminum material are optimized. In S4, each decision variable that affects the rebound amount and the thickness reduction amount of the aluminum material is updated. In S5, a condition for stopping the repetition is determined, and a means for improving the bending rebound of the roll-bent aluminum material is output.

[0007] Preferably, in step S1, the problem of rebound amount during roll bending in the process of roll bending of aluminum material is digitized and the current problem is expressed using genetic sequences, i.e., the coding process. When coding the process, it is divided into four parts: friction coefficient, die gap, angle between roll wheels, and speed difference between upper and lower rolls.

number

[0008] Preferably, the specific steps of step S2 are as follows: In step S201, a bending rebound optimization model for roll bending of aluminum material is constructed based on a multi-objective optimization algorithm, and the rebound amount and thickness reduction amount during roll bending of aluminum material are used as two objective functions. In S202, a constraint condition is established in a rebound amount optimization model for roll bending of aluminum material based on a multi-objective optimization algorithm, and the following constraint condition is included when roll bending of aluminum material:

number

[0009] Preferably, the specific steps of step S3 are as follows: In step S301, a species group is initialized, and the rebound amount and thickness reduction amount in the aluminum material forming process, the species population number, selection probability, and mutation probability of the multi-objective optimization model are initialized, and the maximum number of iterations is set. In S302, a fitness function is established, and specifically, the formula F is written as follows: F=min(f1,f2)(7) Here, f1 represents the rebound amount during roll bending of the aluminum material, and f2 represents the thickness reduction amount during roll bending of the aluminum material.

[0010] Preferably, the specific steps of step S4 are as follows: S401: Encoding and decoding are performed. S402, a crossover operation of chromosomal sequences. S403, which is a chromosomal mutation manipulation. S404 is an environment selection mechanism.

[0011] Preferably, the specific steps of step S401 are as follows: S4011, in which the encoding adopts a real number encoding method, and four real numbers are randomly generated in each chromosome in the species group. These four real numbers respectively represent the values ​​of four decision variables: the friction coefficient u of the contact surface between the aluminum material and the upper roll of the bending machine, the size G of the gap between the aluminum material and the roll during the bending process, the size ω of the angle between the upper roll and the lower roll, and the speed difference Δv between the upper roll and the lower roll. These real numbers directly correspond to the parameter values ​​in the target space. In S4012, when generating the four real numbers, it is necessary to ensure that the randomly generated values ​​in the chromosome satisfy the constraints in step S202. The four real numbers (μ, G, ω, Δv) are combined to form the real number encoding of one individual. In S4013, if there are n individuals in the species group, n chromosomes with a combination of (μ, G, ω, Δv) are generated. At S4014, after the encoding is completed, the decoding is also completed because each randomly generated value in the chromosome at that point corresponds directly to a parameter value in the space.

[0012] Preferably, the specific steps of step S402 are as follows: In S4021, two intersections C1 and C2 are randomly selected in the code strings of the two parent individuals, and a new archive set S is established. In S4022, after the crossing point is determined, some genes at the crossing point of the two parent individuals p1 and p2 are exchanged to generate a new child individual. In S4023, the newly generated child individuals are evaluated according to the target value. In S4024, if a newly generated child individual is superior to an individual in the archive set S during the iteration process, the new child individual is added to the archive set S, and the inferior individual is removed from the archive set S.

[0013] Preferably, the chromosome sequence mutation method in step S403 adopts a species group evolution method, and the specific steps are as follows: In S4031, after each chromosome crossover, the two target values ​​f1 and f2 of the species population individuals are recalculated, and the magnitude of the index value I of each individual is calculated. The calculation formula is as follows:

number

[0014] Preferably, in step S404, the environment selection is divided into convergent selection and divergent selection of individuals, and the specific steps are as follows: In S4041, a set of uniformly distributed reference vectors is generated in the target space using the NBI method, which requires the number of known optimization targets M and the number of partitions p for each target, with the partitioning step size being 1 / p. The formula for the number of generated reference vectors H is as follows:

number

number

[0015] Preferably, step S5 specifically comprises: A maximum number of iterations is set, and if the individuals in the species group reach the maximum number of iterations after multiple evolution iterations, a feasible solution that satisfies the constraints is obtained. Then, the multi-attribute decision TOPSIS method is used to select the optimal solution from the set of solutions, and the distance between each individual in the set of solutions and the best point and the distance between each individual and the worst point are calculated, and all the distances obtained are ordered, and the solution that is closest to the best point and farthest from the worst point is considered to be the optimal solution.

[0016] Therefore, the present invention adopts the method for optimizing the bending rebound amount during roll bending of aluminum material, which has the following beneficial effects: 1) Improve the precision of the roll bending process. The multi-objective optimization algorithm handles the two conflicting optimization objectives of bending rebound and thickness reduction, finds the optimal balance between the objectives, and simultaneously optimizes these two important indicators to ensure that the aluminum material not only achieves the expected bending shape during the bending process, but also maintains sufficient thickness, thereby meeting the requirements for product precision in actual production. 2) Improve production efficiency: The multi-objective optimization algorithm finds the optimal process combination according to the specific process parameters of the roll bending machine and the characteristics of the aluminum material, and accurately controls the operating parameters of the roll bending machine so that the aluminum material can quickly and accurately reach the desired shape during the bending process. 3) Increase production flexibility. It can be customized and optimized for different types, specifications, and materials of aluminum. This means that the roll bending machine can adapt to more production tasks and improve the flexibility and versatility of the production line. Enterprises can quickly adjust the process parameters of the roll bending machine according to requirements to meet the production needs of different products.

[0017] To sum up, the present invention uses a multi-objective optimization algorithm to optimize the bending rebound and thickness reduction of the roll bending machine, which has the advantages of improving the bending process accuracy, improving production efficiency, and increasing production flexibility. It can avoid the problem of reduced production accuracy caused by errors due to manual experience, reduce the time spent on manually adjusting process parameters, and improve processing accuracy and processing efficiency. The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and examples. [Brief explanation of the drawings]

[0018] [Figure 1] FIG. 1 is an overall flow diagram of an embodiment of a method for optimizing the bending rebound amount during roll bending of an aluminum material according to the present invention. [Figure 2] FIG. 1 is a flow chart of a multi-objective optimization algorithm of an embodiment of a method for optimizing the bending repulsion amount during roll bending of an aluminum material according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0019] The technical solution of the present invention will be further explained below with reference to the accompanying drawings and examples. Unless otherwise defined, technical or scientific terms used herein shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention belongs.

[0020] Example 1 As shown in FIG. 1, the present invention provides a method for optimizing the bending rebound amount of roll bending of aluminum material, which includes the following steps:

[0021] In S1, the aluminum roll bending process is analyzed to determine the determining variables for the rebound amount and thickness reduction amount of the aluminum material. The rebound amount problem in the aluminum roll bending process is digitized and the current problem is expressed using genetic sequences, that is, the coding process. When coding the process, it is divided into four parts: friction coefficient, die gap, angle between roll wheels, and speed difference between the upper and lower rolls.

number

[0022] In S2, a rebound optimization model for aluminum roll bending is established based on a multi-objective optimization algorithm, and constraints in the aluminum roll bending process are clarified. The flowchart of the multi-objective optimization algorithm is shown in Figure 2, and the specific steps are as follows: In step S201, a bending rebound optimization model for roll bending of aluminum material is constructed based on a multi-objective optimization algorithm, and the rebound amount and thickness reduction amount during roll bending of aluminum material are used as two objective functions. In S202, a constraint condition is established in a rebound amount optimization model for roll bending of aluminum material based on a multi-objective optimization algorithm, and the following constraint condition is included when roll bending of aluminum material:

number

[0023] In step S3, an algorithm framework is established to reduce the rebound and thickness loss of aluminum material during the processing process, and the model in step S2 is optimized. The specific steps are as follows: In step S301, a species group is initialized, and the rebound amount and thickness reduction amount of the aluminum material during the aluminum material forming process, the species population number, selection probability, and mutation probability of the multi-objective optimization model are initialized, and the maximum number of iterations is set. In S302, a fitness function is established, and specifically, the formula F is written as follows: F=min(f,f2)(7) Here, f1 represents the amount of rebound during roll bending of aluminum material, and f2 represents the amount of thickness reduction during roll bending of aluminum material. Both f1 and f2 are functions of the coefficient of friction u of the contact surface between the aluminum material and the upper roll of the roll bending machine, the size of the gap G between the aluminum material and the roll during the roll bending process, the size of the angle ω between the upper roll and the lower roll, and the speed difference Δv between the upper roll and the lower roll.

[0024] In S4, each decision variable that affects the rebound amount and thickness reduction amount of the aluminum material is updated, and the specific steps are as follows: S401: Encoding and decoding are performed, and the specific steps are as follows: In S4011, binary code is prone to loss of precision when encoding two decision variables. In order to improve the precision of chromosome description, the code adopts a real code method, and four real numbers are randomly generated for each chromosome in the species group. These four real numbers respectively represent the values ​​of four decision variables: the friction coefficient μ of the contact surface between the aluminum material and the upper roll of the bending machine, the size G of the gap between the aluminum material and the roll during the bending process, the size ω of the angle between the upper roll and the lower roll, and the speed difference Δv between the upper roll and the lower roll. These real numbers directly correspond to the parameter values ​​in the target space. In step S4012, to generate the four real numbers, it is necessary to ensure that the randomly generated values ​​in the chromosome satisfy the constraints in step S202. The four real numbers (μ, G, ω, Δv) are combined to form a real number code for one individual. In S4013, when there are n individuals in the species group, n chromosomes with a combination of (μ, G, ω, Δv) are generated. At S4014, after the code is completed, the decoding is already complete because each randomly generated value in the chromosome at that point already corresponds directly to a parameter value in space. S402 is a chromosome sequence crossover operation, which adopts a two-point crossover policy of introducing an external archive set in chromosome sequence crossover, that is, randomly selecting two points in a chromosome for crossover, and introducing an external archive to preserve superior individuals or gene segments during the crossover process, and the specific steps are as follows: In S4021, two crossover points c1 and c2 are randomly selected in the code strings of the two parent individuals, and the positions of these two crossover points are random, and a new archive set S is created. In S4022, after determining the crossover point, partial genes between the crossover points of the two parent individuals p1 and p2 are exchanged to generate a new child individual. In S4023, the newly generated child individual is evaluated according to a target value, and the smaller the target value, the better. In S4024, if a newly generated child individual is superior to an individual in the external archive set during the iteration process, the inferior individual is removed from the archive set, and the new child individual is added to the archive set S. S403 is chromosome mutation operation, and the chromosome sequence mutation method adopts the species group evolution method, and the specific steps are as follows: In S4031, after each chromosome crossover, the two target values ​​f1 and f2 of the species group individuals are recalculated, and the magnitude of the index value I for each individual is calculated, and the calculation formula is as follows:

number

number

number

number

number

[0025] In S5, the repetition stopping condition is determined, and a solution for improving the bending rebound amount of the roll-bent aluminum material is output, specifically as follows: A maximum number of iterations is set, and if the individuals in the species group reach the maximum number of iterations after multiple evolution iterations, a feasible solution that satisfies the constraints is obtained. Then, the multi-attribute decision TOPSIS method is used to select the optimal solution from the set of solutions, and the distance between each individual in the set of solutions and the best point and the distance between each individual and the worst point are calculated, and all the distances obtained are ranked, and the solution that is closest to the best point and farthest from the worst point is considered to be the optimal solution. Specifically, it includes the following steps: In S501, the attribute values ​​are normalized by equation (13), and then the attribute weighted sum of each solution is calculated by equation (14), and the weight value of each attribute is set to be the same.

number

number

number

[0026] Therefore, the present invention adopts the above-mentioned method for optimizing the bending rebound amount during roll bending of aluminum material, analyzes the aluminum material roll bending process, and combines a multi-objective optimization algorithm to obtain decision variables that result in a relatively small bending rebound amount and a relatively small thickness reduction amount, namely, the appropriate contact surface friction parameters and clearance parameters between the aluminum material and the upper roll of the roll bending machine, thereby realizing high-precision parameter setting for roll bending technology.

[0027] Each chromosome in the algorithm corresponds to a decision variable array that affects the bending rebound and thickness reduction, and the decision variable array is coded and decoded, and operations such as crossover, mutation, and selection are performed between each array to continuously evolve and iterate the chromosome. Multi-attribute decision is then used to determine a relatively good solution set, and optimal technical setting parameters with minimum bending rebound and minimum thickness reduction are obtained. This solves the problem of relatively large bending rebound in roll bending technology, and has the advantages of improving the precision of roll bending technology, improving production efficiency, and enhancing production flexibility.

[0028] Finally, it should be noted that the above embodiments only illustrate the technical solutions of the present invention, and are not intended to limit the same, and although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to depart from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for optimizing the bending rebound amount during roll bending of an aluminum material, comprising the following steps: S1: Analyzing the process of roll bending of the aluminum material, and determining the determining variables of the rebound amount and thickness reduction amount of the aluminum material; S2: Build an aluminum roll bending rebound optimization model based on a multi-objective optimization algorithm, and clarify the constraints in the aluminum roll bending process. The specific steps are as follows: S201, constructing a bending rebound amount optimization model for roll bending of aluminum material based on a multi-objective optimization algorithm, in which the rebound amount and the thickness reduction amount during roll bending of aluminum material are two objective functions; S202: Constraint conditions are established in a rebound amount optimization model for roll bending of an aluminum material based on a multi-objective optimization algorithm, and the following constraint conditions are included in the roll bending of the aluminum material: [Equation 1] Equation (1) is the constraint on the friction coefficient of the contact surface between the aluminum material and the upper roll of the roll bending machine, where μ ｍｉｎ is the minimum allowable friction coefficient during bending, and μ ｍａｘ is the maximum allowable friction coefficient during bending, and equation (2) is the constraint on the gap between the aluminum material and the roll wheel during the roll bending process, G ｍｉｎ is the minimum gap allowed during bending, and G ｍａｘ is the maximum allowable gap during bending, and in equation (3), ω ｍaｘ represents the maximum angle between the upper and lower rolls, and ω ｍｉｎ represents the minimum value of the angle between the upper roll and the lower roll, Δv in equation (4) represents the constraint on the speed difference between the upper roll and the lower roll, equation (5) is the constraint on the thickness reduction amount required by the aluminum material when processing a certain part, and T ｍａｘ is the maximum thickness reduction required, S3: Provide a corresponding algorithm based on the characteristics of the model in S2 to optimize the rebound amount and thickness reduction amount of the aluminum material, the specific steps of which are as follows: S301: Initializing a species group; initializing the aluminum material rebound amount and thickness reduction amount during the aluminum material forming process; the species group population number, selection probability, and mutation probability of a multi-objective optimization model; and setting the maximum number of iterations; S302: Establish a fitness function, specifically, write the formula F as follows: F=min(f,f ２ )(7) where f １ indicates the rebound amount during roll bending of aluminum material, and f ２ indicates the thickness reduction amount during roll bending of aluminum material, S4: Update each decision variable that affects the rebound amount and thickness reduction amount of the aluminum material, and the specific steps are as follows: S401, performing encoding and decoding; S402 is a chromosome sequence crossover operation, the specific steps of which are as follows: S4021, two crossover points c are randomly selected in the code strings of the two parent individuals. １ , c ２ and create a new archive set S, S4022, after determining the crossover point, exchange partial genes between the crossover points of the two parent individuals p1 and p2 to generate new child individuals; S4023: Evaluating the newly generated child individual according to the target value; S4024: In the process of iteration, if a newly generated child individual is superior to an individual in the archive set S, add the new child individual to the archive set S, and remove the inferior individual from the archive set S; S403 is a chromosome mutation operation, in which the chromosome sequence mutation method adopts a species group evolution method, and the specific steps thereof are as follows: S4031: After each chromosome crossover, two target values ​​f １ , f ２ and calculate the magnitude of the index value I for each individual. The calculation formula is as follows: [Equation 2] S4032: Ranking the index values ​​I of 2N individuals in the species group, and the individual with the smallest index value I is ranked as the first species group P １ and the individuals with the highest index value I are placed in the second species group P ２ Put it in S4033, and the second species group P ２ All individuals in are mutated, and the mutation method is differential mutation. S404 is an environmental selection mechanism, in which the environmental selection is divided into convergent selection and divergent selection of individuals, and its specific steps are as follows: S4041, generating a set of uniformly distributed reference vectors in the target space using the NBI method, and the formula for the number H of generated reference vectors is as follows: [Equation 3] where C represents the number of combinations in the sequence combination, M represents the number of optimization targets, p represents the number of divisions per target, and 1 / p represents the division step size. In S4042, the parent individuals and the child individuals are integrated. At this time, the number of species groups is set to 2N. The target values ​​of all the individuals are calculated, and the target values ​​of the individuals are normalized as follows: [Equation 4] S4043: ordering the individuals in the regions in a non-dominated manner, and dividing all individuals in the species group into different frontiers or strata according to two target normalization values; S4044, taking the first K frontier surfaces in each region, and making the number of individuals n of the species group greater than N; S4045, in order to reduce the number of individuals of the species group to N, cluster all currently held individuals using a hierarchical clustering method, and remove a individuals from each class so that the number of individuals of the species group at this time is just N; S5: A method for optimizing the bending rebound amount during roll bending of aluminum material, characterized by determining a repetition stop condition and outputting a plan for improving the bending rebound amount of the roll-bent aluminum material.

2. 2. The method for optimizing the bending rebound amount during roll bending of aluminum material according to claim 1, wherein in step S1, the problem of the bending rebound amount during roll bending of the aluminum material is digitized, and the current problem is expressed using genetic sequences, i.e., the coding process, and the coding for the process is divided into four parts, namely, the friction coefficient, the die gap, the angle between the roll wheels, and the speed difference between the upper and lower rolls. [Equation 5] where μ represents the coefficient of friction of the contact surface between the aluminum material and the upper roll of the roll bending machine, G represents the size of the gap between the aluminum material and the roll during the roll bending process, ω represents the size of the angle between the upper roll and the lower roll, Δv represents the speed difference between the upper roll and the lower roll, n represents the number of individuals in the species group, and k = {1, 2, 3, ..., n}.

3. S5, A maximum number of iterations is set, and when the individuals in the species group reach the maximum number of iterations after multiple evolution iterations, a feasible solution that satisfies the constraints is obtained. The method for optimizing the bending rebound amount in roll bending of aluminum material described in claim 1, characterized in that the optimal solution is then selected from the set of solutions using the multi-attribute decision TOPSIS method, the distance between each individual in the set of solutions and the best point, and the distance between each individual and the worst point are calculated, and the solution closest to the best point and farthest from the worst point is considered to be the optimal solution.

Images