Cold roll forming multi-stage roll cooperative optimization method and system based on simulation data

Abstract

This invention discloses a method and system for collaborative optimization of multi-stage rolls in cold bending forming based on simulation data, relating to the field of production line optimization technology for cold bending forming. First, the sheet metal properties and roll geometric parameters are acquired to construct a virtual cold bending forming simulation model. Second, based on a preset set of driving parameters, several modified sets of driving parameters are generated by repeatedly changing each parameter. Then, these parameter sets are used to drive the simulation model, constructing a feature library containing multiple virtual verification formed parts. Finally, by acquiring actual formed part data from the production line and comparing it with the virtual feature library, the most similar virtual finished product is identified. Based on the differences between the modified parameters corresponding to this virtual finished product and the preset parameters, the parameter optimization strategy for the actual production line is determined. This invention, by constructing a virtual deviation mapping library, achieves precise tracing and reverse compensation of production deviations, effectively improving the efficiency of multi-stage roll optimization and the accuracy of finished products.

Description

Technical Field

[0001] This invention relates to the field of production line optimization technology for cold bending forming, and in particular to a method and system for collaborative optimization of multi-stage rolls in cold bending forming based on simulation data. Background Technology

[0002] Cold roll forming, as a highly efficient and energy-saving metal sheet and strip processing technology, uses sequentially arranged multi-pass forming rolls to induce continuous plastic deformation in metal, thereby obtaining products with specific cross-sectional shapes. It has been widely used in key fields such as automobile manufacturing, building structures, and photovoltaic brackets. However, due to the extremely complex physical processes involved in cold roll forming, including large nonlinear material deformation, dynamic contact friction, and residual stress rebound after forming, process debugging and precision control have always been highly challenging tasks. In current production practices, the optimization of multi-stage cold roll forming units mainly relies on manual trial and error. Experienced technicians make incremental adjustments to the reduction, gap, or speed of each roll based on experience, based on defects such as edge waviness, longitudinal distortion, or dimensional deviations in the finished product. This not only leads to lengthy debugging cycles and huge raw material losses but also makes it difficult to cope with the severe inter-stage coupling effects present in the forming of complex cross-sections. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system that can provide auxiliary guidance for the coordinated optimization of multi-stage rolls.

[0004] This invention discloses a method for collaborative optimization of multi-stage rolls in cold bending forming based on simulation data, including: Step S100: Obtain the property parameters of the sheet material to be processed and the geometric parameters of the rolls, which are used to construct a virtual cold bending simulation model to simulate the cold bending forming process. Step S200: Preset a set of roll drive parameters. The set of roll drive parameters can drive the virtual cold bending forming simulation model to generate a virtual standard cold bending forming part. The roll drive parameters in the preset set of roll drive parameters are changed several times to obtain a set of roll drive parameters after several changes. Step S300: Based on the changed roll drive parameter group, drive the virtual cold bending forming simulation model to obtain the virtual verification cold bending forming part; Step S400: Obtain the actual cold-bent forming part of the actual production line, compare the actual cold-bent forming part with different virtual verification cold-bent forming parts, determine the matching virtual verification cold-bent forming part, and determine the parameter optimization strategy of the rolls on the actual production line based on the difference characteristics between the modified roll drive parameter group and the preset roll drive parameter group corresponding to the virtual verification cold-bent forming part.

[0005] In some embodiments disclosed in this invention, the method for constructing a virtual cold bending simulation model simulating the cold bending process includes: Step S101: Obtain the elastic-plastic property parameters of the plate, define its rheological stress curve and work hardening criterion, and construct the physical constitutive relationship of the plate element. Step S102: Import the geometric data of the multi-stage rolls and configure their spatial arrangement; perform adaptive mesh refinement on the forming deformation zone of the virtual sheet material; and establish an initial state discretization model. Step S103: Establish a multi-body contact pair between the rigid roll and the deformable sheet metal, set the contact friction coefficient and boundary constraints, and simulate the load transfer path. Step S104: According to the cold bending process sequence, rotational driving force is applied to each level of virtual rolls in sequence to simulate the dynamic extrusion and bending deformation process of the sheet material continuously passing through each level of rolls, and to realize the inter-stage coupling transfer of deformation energy and residual stress. Step S105: Perform nonlinear solution calculation to obtain the formed virtual cold-bent part, extract the three-dimensional coordinates, stress and strain data of its key points, and generate digital feature vectors.

[0006] In some embodiments disclosed in this invention, the method for making several changes to each roll drive parameter in a preset roll drive parameter group includes: Step S201: Randomly generate random parameter change strategies for several roll drive parameter groups. Each random parameter change strategy includes the parameter change amount of each roll drive parameter. Each parameter change amount is randomly selected within a preset parameter change range. Step S202: Based on the random parameter change strategy, the preset roll drive parameter group is changed to obtain the changed roll drive parameter group. The changed roll drive parameter group corresponding to the virtual verification cold bending forming part matched later is mapped, and the corresponding random parameter change strategy is mapped out, which is denoted as the mapped random parameter change strategy. Step S203: Classify the mapped random parameter change strategies based on similarity to obtain several similar mapped random parameter change strategy sets. Based on the data volume ratio between similar mapped random parameter change strategy sets, expand the content of different similar mapped random parameter change strategy sets in an equal proportion to obtain expanded random parameter change strategy sets. Based on the expanded random parameter change strategy sets, change the preset roll drive parameter group.

[0007] In some embodiments disclosed in this invention, the method for classifying the similarity of mapped random parameter variation strategies includes: Step S2031: Analyze the relative parameter change difference of each parameter change among the mapped random parameter change strategies. If all of them are less than or equal to the preset value, then the mapped random parameter change strategies are considered to be similar to each other.

[0008] In some embodiments disclosed in this invention, a method for expanding the content of similarly mapped random parameter variation strategy sets includes: Step S2032: Construct a change reference axis for each parameter change in the mapped random parameter change strategy, set several change segments for the change reference axis, and determine the selected probability corresponding to each change segment based on the mapping ratio characteristics of each change segment. Step S2033: Based on the selected probability corresponding to each change range, amplify the mapping random parameter change strategy.

[0009] In some embodiments disclosed in this invention, the method for determining the selected probability corresponding to each change segment based on the mapping ratio characteristics of each change segment includes: Step S20321: Analyze each change in all random parameter change strategies, determine the change range corresponding to each change, and count the number of times the change range is mapped by the change, which is recorded as the number of regular mappings. Step S20322: Analyze each change in the mapped random parameter change strategy, determine the change range corresponding to each change, and count the number of times the change range is mapped by the change, which is recorded as the number of high-value mappings. Step S20323: Configure weight coefficients for the number of regular mappings and the number of high-value mappings corresponding to each change range, calculate the product of the weight coefficient and the number of regular mappings, and record it as the regular mapping influence parameter. Calculate the product of the weight coefficient and the number of high-value mappings, and record it as the high-value mapping influence parameter. Calculate the sum of the regular mapping influence parameter and the high-value mapping influence parameter, and record it as the mapping weight. Step S20324: Determine the selected probability corresponding to different change segments based on the ratio between the mapping weights corresponding to different change segments.

[0010] In some embodiments disclosed in this invention, the method for determining the matching virtual verification cold-bent part includes: Step S401: Define the bent portions in the cold-formed part and determine the bending sequence of each bent portion. Step S402: Compare the cross sections of the bent portions with the same bending sequence between the cold-formed parts, calculate the overlap parameters between them, and if all overlap parameters are greater than or equal to the preset values, then the cold-formed parts are considered to be matched with each other.

[0011] In some embodiments disclosed in this invention, a method for defining the bent portion in a cold-formed part includes: Step S4011: Determine the trend line of the side section of the cold-formed part. If the trend line has a non-straight starting point, start cutting the bending part. If the trend line has a straight starting point, then the starting point is identified as the end position of the bending part cutting.

[0012] In some embodiments disclosed in this invention, the method for comparing and aligning the cross-sections of the bent portion includes: Step S4021: Determine the trend line of the cross section of the bent part, and make the respective trend lines coincide. The coincidence method includes determining the center of the trend line, aligning the center, dynamically rotating the trend line, and analyzing the coincidence ratio between the two in real time. Step S4022: The largest overlap ratio is identified as the overlap parameter.

[0013] In some embodiments disclosed in this invention, a multi-stage roll collaborative optimization system for cold bending forming based on simulation data includes: The first module is used to obtain the property parameters of the sheet material to be processed and the geometric parameters of the rolls, and to build a virtual cold bending simulation model to simulate the cold bending forming process. The second module is used to preset the roll drive parameter group. The roll drive parameter group can drive the virtual cold bending forming simulation model to generate a virtual standard cold bending forming part. The roll drive parameter group of each roll in the preset roll drive parameter group is changed several times to obtain several changed roll drive parameter groups. The third module is used to drive the virtual cold bending forming simulation model based on the changed roll drive parameter group to obtain a virtual verification cold bending forming part. The fourth module is used to acquire actual cold-bent parts from the actual production line, compare the actual cold-bent parts with different virtual verification cold-bent parts, determine the matching virtual verification cold-bent parts, and determine the parameter optimization strategy for the rolls on the actual production line based on the differences between the modified roll drive parameter group and the preset roll drive parameter group corresponding to the virtual verification cold-bent parts.

[0014] This invention discloses a method and system for collaborative optimization of multi-stage rolls in cold bending forming based on simulation data, relating to the field of production line optimization technology for cold bending forming. First, the sheet metal properties and roll geometric parameters are acquired to construct a virtual cold bending forming simulation model. Second, based on a preset set of driving parameters, several modified sets of driving parameters are generated by repeatedly changing each parameter. Then, these parameter sets are used to drive the simulation model, constructing a feature library containing multiple virtual verification formed parts. Finally, by acquiring actual formed part data from the production line and comparing it with the virtual feature library, the most similar virtual finished product is identified. Based on the differences between the modified parameters corresponding to this virtual finished product and the preset parameters, the parameter optimization strategy for the actual production line is determined. This invention, by constructing a virtual deviation mapping library, achieves precise tracing and reverse compensation of production deviations, effectively improving the efficiency of multi-stage roll optimization and the accuracy of finished products.

[0015] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the method steps of the cold bending forming multi-stage roll collaborative optimization method based on simulation data disclosed in this embodiment of the invention. Detailed Implementation

[0017] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0018] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings and specific embodiments. It should be understood that the preferred embodiments described herein are only for illustration and explanation of the present invention and should not be construed as limiting the scope of protection of the present invention. Those skilled in the art can make some non-essential improvements and adjustments based on the following content of the present invention. In the present invention, unless otherwise expressly specified and limited, the technical terms used in the present invention should have the ordinary meaning understood by those skilled in the art.

[0019] Example: This invention discloses a method for collaborative optimization of multi-stage rolls in cold bending forming based on simulation data. (See reference...) Figure 1 ,include: Step S100: Obtain the property parameters of the sheet material to be processed and the geometric parameters of the rolls, which are used to construct a virtual cold bending simulation model to simulate the cold bending forming process.

[0020] The core principle of this step lies in the "digital mapping of physical processes." By combining the microscopic elastoplastic mechanical properties of the metal sheet (such as rheological stress curves and isotropic hardening criteria) with the macroscopic geometric topology of the rolls, a highly realistic virtual processing scenario is constructed within the finite element analysis (FEA) environment. This step utilizes nonlinear numerical calculation methods to simulate the actual extrusion and bending processes, transforming complex continuous plastic deformation into a solvable mathematical model, thus providing a low-cost, repeatable "digital laboratory" for the entire optimization system. It not only simulates the external shape of the sheet but also restores the internal residual stress distribution, enabling all subsequent parameter explorations to be conducted without consuming physical materials, forming the fundamental basis for achieving "virtual-real fusion."

[0021] Step S200: Preset a set of roll drive parameters. The set of roll drive parameters can drive the virtual cold bending forming simulation model to generate a virtual standard cold bending forming part. The roll drive parameters in the preset set of roll drive parameters are changed several times to obtain a set of roll drive parameters after several changes.

[0022] The principle behind this step is based on "sensitivity analysis and exhaustive enumeration of all operating conditions." The system doesn't blindly experiment with parameters; instead, it uses ideal design conditions as a baseline and actively introduces random disturbances that conform to real-world patterns (such as gap changes caused by simulated roll wear, or reduction fluctuations due to insufficient frame stiffness). This controlled "parameter variation" simulates the unavoidable random fluctuations in real production. The essence of this approach is to explore the robustness boundaries of the process system by artificially creating a large number of "variable samples," rehearsing various possible production anomalies in a virtual space beforehand. This "purposeful randomness" establishes a comprehensive process characteristic reference system, ensuring that the system has sufficient "insight" to identify and address the myriad deviations encountered in reality, thus eliminating the over-reliance on expert experience in traditional debugging.

[0023] Step S300: Based on the changed roll drive parameter group, drive the virtual cold bending forming simulation model to obtain the virtual verification cold bending forming part.

[0024] The principle behind this step embodies "forward decoding from process input to quality response." It utilizes a high-performance computing engine to drive a virtual model, transforming abstract combinations of driving parameters into concrete geometric deformation results—essentially performing a complex physical evolution mapping. Each generated virtual verification cold-bent part is actually a "geometric response fingerprint" under a corresponding parameter set, recording the product's final shape, stress state, and potential defects under that specific perturbation. Through this large-scale computation, the system constructs a massive "virtual defect feature library," achieving a deep correlation between process parameters and forming quality. This library is like a comprehensive "encyclopedia of errors," recording thousands of causal evidences that "because the parameters were adjusted this way, the product changed that way."

[0025] Step S400: Obtain the actual cold-bent forming part of the actual production line, compare the actual cold-bent forming part with different virtual verification cold-bent forming parts, determine the matching virtual verification cold-bent forming part, and determine the parameter optimization strategy of the rolls on the actual production line based on the difference characteristics between the modified roll drive parameter group and the preset roll drive parameter group corresponding to the virtual verification cold-bent forming part.

[0026] The principle behind this step lies in "pattern recognition and error compensation based on cause-and-effect analysis." When dimensional deviations occur in products produced on the real production line, the system collects the digital features of the physical product and performs a high-dimensional similarity search in a virtual library. Its core logic is to find the virtual sample that most closely resembles the actual finished product. Once a match is found, because the underlying parameter variations of the virtual sample are known, the system can instantly deduce the root cause parameters leading to the real-world deviation—for example, whether the third-stage roll is too shallow or the eighth-stage roll has excessive springback. Based on this "difference characteristic," the system can calculate precise reverse compensation amounts, transforming the previously labored debugging process into an automated, collaborative optimization strategy that can offset real-world errors and achieve global optimality.

[0027] In some embodiments disclosed in this invention, the method for constructing a virtual cold bending simulation model simulating the cold bending process includes: Step S101: Obtain the elastic-plastic property parameters of the plate, define its rheological stress curve and work hardening criterion, and construct the physical constitutive relationship of the plate element.

[0028] The core principle of this step lies in "mathematical modeling of material mechanical response." By acquiring the elastoplastic parameters of the sheet metal (such as Young's modulus and Poisson's ratio) and combining them with measured rheological stress curves, the system reconstructs the plastic flow and hardening laws of metallic materials under stress in a digital environment. Essentially, it utilizes the constitutive equations of materials to accurately describe the slip and strengthening behavior of metal atomic layers under macroscopic stress, thereby ensuring that the simulation model can accurately predict the yielding timing and springback degree of the sheet metal under complex extrusion conditions, providing the most fundamental physical basis for the entire forming simulation.

[0029] Step S102: Import the geometric data of the multi-stage rolls and configure their spatial arrangement. Perform adaptive mesh refinement on the forming deformation zone of the virtual sheet and establish an initial state discretization model.

[0030] The principle of this step is based on "spatial discretization technology in finite element analysis". By decomposing the continuous roll geometry and the solid sheet into finite elements composed of a large number of nodes, the system transforms the complex partial differential mechanics problem into a solvable algebraic operation. The key to adopting the adaptive mesh refinement technology lies in the "dynamic balance between efficiency and accuracy": automatically increasing the mesh density in areas of severe bending and large strain gradients in the sheet, while maintaining a sparse mesh in flat areas. This ensures that extremely subtle deformation features are captured while maximizing the allocation of computational resources, constructing an initial state model that combines high fidelity and computational feasibility.

[0031] Step S103: Establish a multi-body contact pair between the rigid roll and the deformable sheet metal, set the contact friction coefficient and boundary constraints, and simulate the load transfer path.

[0032] In step S104, according to the cold bending process sequence, rotational driving force is applied to each level of virtual rolls in sequence to simulate the dynamic extrusion and bending deformation process of the sheet material continuously passing through each level of rolls, and to realize the inter-stage coupling transfer of deformation energy and residual stress.

[0033] Step S105: Perform nonlinear solution calculation to obtain the formed virtual cold-bent part, extract the three-dimensional coordinates, stress and strain data of its key points, and generate digital feature vectors.

[0034] The principle behind this step lies in "convergent solution and data dimensionality reduction for complex nonlinear systems." Using explicit or implicit solution algorithms, the system iterates continuously within a time step until it meets the mechanical equilibrium conditions under triple nonlinearity of geometry, materials, and contact. The extraction process after molding is a form of "data dimensionality reduction and feature engineering": the massive simulation results containing millions of nodes are condensed into core features reflecting the quality of the finished product, such as the three-dimensional coordinates of key points, stress distribution, and residual strain. This generation of digital feature vectors transforms the complex physical entity into a standardized dataset that can be recognized and compared by computers, laying the data foundation for subsequent error matching with physical products.

[0035] In some embodiments disclosed in this invention, the method for making several changes to each roll drive parameter in a preset roll drive parameter group includes: Step S201: Randomly generate random parameter change strategies for several roll drive parameter groups. Each random parameter change strategy includes the parameter change amount of each roll drive parameter, wherein each parameter change amount is randomly selected within a preset parameter change range.

[0036] Because cold bending forming units involve dozens of rolls, their driving parameters constitute an extremely high-dimensional search space, and real-world disturbances (such as mechanical vibrations) possess unpredictable randomness. By randomly generating varying step sizes within a preset physical safety range, the system essentially establishes a "stress test set" for the machine. This approach aims to break the limitations of idealized simulation, utilizing randomness to exhaustively enumerate various combinations of microwave disturbances that may occur on the production floor. This ensures that the simulation model can not only simulate "correct" scenarios but also rehearse thousands of possible "incorrect" scenarios in the virtual world, providing the broadest possible source material for subsequent fault diagnosis.

[0037] Step S202: Based on the random parameter change strategy, the preset roll drive parameter group is changed to obtain the changed roll drive parameter group. The changed roll drive parameter group corresponding to the virtual verification cold bending forming part matched later is mapped, and the corresponding random parameter change strategy is mapped out, which is denoted as the mapped random parameter change strategy.

[0038] The principle behind this step lies in "bidirectional mapping and high-value data annotation." During the forward simulation, each set of perturbation parameters produces a specific physical result, and the core task of S202 is to establish "identity cards" for these causal relationships. By associating the successfully matched virtual product with its underlying driving strategy, the system achieves a precise lock from "outcome characteristics" to "initial causes." Essentially, this process extracts "realistically meaningful" deviation fingerprints from massive amounts of random data, eliminates invalid simulations that are detached from reality, and transforms the chaotic random data into a structured diagnostic dictionary. This mapping mechanism enables the system to "trace the cause from the effect," providing clear navigation coordinates for subsequent targeted local precision enhancement.

[0039] Step S203: Classify the mapped random parameter change strategies based on similarity to obtain several similar mapped random parameter change strategy sets. Based on the data volume ratio between similar mapped random parameter change strategy sets, expand the content of different similar mapped random parameter change strategy sets in an equal proportion to obtain expanded random parameter change strategy sets. Based on the expanded random parameter change strategy sets, change the preset roll drive parameter group.

[0040] The core principle of this step is "adaptive local encrypted sampling based on probability distribution." Given limited computing resources, the system cannot simulate the entire space with equal precision. Therefore, S203 adopts a "fill in the gaps where they are important" approach. By comparing and classifying successful "mapped random parameters" based on similarity and proportionally expanding them, the system identifies the "high-risk parameter intervals" that best approximate the actual error patterns of the production line. This proportional expansion essentially performs a higher-density secondary exploration in high-value areas, equivalent to "high-magnification observation" of the most problematic local spaces. This principle, through dynamically optimizing the sample distribution, resolves the contradiction between simulation efficiency and search accuracy, ensuring the highest sample density in the virtual library within key deviation areas, greatly improving the system's approximation of real production fluctuations.

[0041] In some embodiments disclosed in this invention, the method for classifying the similarity of mapped random parameter variation strategies includes: Step S2031: Analyze the relative parameter change difference of each parameter change among the mapped random parameter change strategies. If all of them are less than or equal to the preset value, then the mapped random parameter change strategies are considered to be similar to each other.

[0042] In some embodiments disclosed in this invention, a method for expanding the content of similarly mapped random parameter variation strategy sets includes: Step S2032: Construct a change reference axis for each parameter change in the mapped random parameter change strategy. Set several change ranges for the change reference axis. Based on the mapping ratio characteristics of each change range, determine the selected probability corresponding to each change range.

[0043] The core principle of this step lies in "discretization statistics and hotspot region identification in multidimensional feature space". By constructing a reference axis for each parameter change and dividing it into several segments, the system transforms the originally continuous and infinite parameter search space into finite, quantifiable statistical units. Determining the selection probability based on "mapping ratio characteristics" is essentially performing discretization fitting of the probability density function (PDF): if a certain segment is frequently hit in previous comparisons (i.e., high mapping ratio), it indicates that the parameter range is highly correlated with the actual deviation fluctuation of the current production line. This principle is similar to drawing a "value heat map" in the parameter space, providing a scientific weighting basis for subsequent simulation resource allocation by quantifying the contribution of different segments, thereby identifying which parameter ranges are the "high-frequency zones" leading to molding defects.

[0044] Step S2033: Based on the selected probability corresponding to each change range, amplify the mapping random parameter change strategy.

[0045] The principle behind this step is "knowledge amplification and precision focusing based on importance sampling." After obtaining the selected probabilities for each segment, the amplification process is no longer a blind, random generation, but a biased "enhanced search." The logic is to use verified prior knowledge to guide simulation calculations, filling the core areas (i.e., the areas closest to real-world conditions) with higher probability samples, while reducing input in the peripheral areas with lower probability samples. This principle simulates the thought process of human experts' "targeted review"—since certain parameter combinations have shown a high degree of similarity to reality, more subtle and intensive simulations are performed near these combinations. In this way, the system significantly improves the resolution and coverage of the virtual verification library in key deviation areas without significantly increasing the total computational load, evolving the optimization strategy from "fuzzy matching" to "precise focusing."

[0046] In some embodiments disclosed in this invention, the method for determining the selected probability corresponding to each change segment based on the mapping ratio characteristics of each change segment includes: Step S20321: Analyze each change in all random parameter change strategies, determine the change range corresponding to each change, and count the number of times the change range is mapped by the change, which is recorded as the number of regular mappings.

[0047] Step S20322: Analyze each change in the mapped random parameter change strategy, determine the change range corresponding to each change, and count the number of times the change range is mapped by the change, which is recorded as the number of high-value mappings.

[0048] Step S20323: Configure weight coefficients for the number of regular mappings and the number of high-value mappings corresponding to each variable segment, calculate the product of the weight coefficient and the number of regular mappings, and record it as the regular mapping influence parameter. Calculate the product of the weight coefficient and the number of high-value mappings, and record it as the high-value mapping influence parameter. Calculate the sum of the regular mapping influence parameter and the high-value mapping influence parameter, and record it as the mapping weight.

[0049] Step S20324: Determine the selected probability corresponding to different change segments based on the ratio between the mapping weights corresponding to different change segments.

[0050] In some embodiments disclosed in this invention, the method for determining the matching virtual verification cold-bent part includes: Step S401: Define the bent portions in the cold-formed part and determine the bending sequence of each bent portion.

[0051] Step S402: Compare the cross sections of the bent portions with the same bending sequence between the cold-formed parts, calculate the overlap parameters between them, and if all overlap parameters are greater than or equal to the preset values, then the cold-formed parts are considered to be matched with each other.

[0052] In some embodiments disclosed in this invention, a method for defining the bent portion in a cold-formed part includes: Step S4011: Determine the trend line of the side section of the cold-formed part. If the trend line has a non-straight starting point, start cutting the bending part. If the trend line has a straight starting point, then the starting point is identified as the end position of the bending part cutting.

[0053] In some embodiments disclosed in this invention, the method for comparing and aligning the cross-sections of the bent portion includes: Step S4021: Determine the trend line of the cross section of the bent part, and make the respective trend lines coincide. The coincidence method includes determining the center of the trend line, aligning the center, dynamically rotating the trend line, and analyzing the coincidence ratio between the two in real time.

[0054] Step S4022: The largest overlap ratio is identified as the overlap parameter.

[0055] In some embodiments disclosed in this invention, a multi-stage roll collaborative optimization system for cold bending forming based on simulation data includes: The first module is used to obtain the property parameters of the sheet material to be processed and the geometric parameters of the rolls, and to build a virtual cold bending simulation model to simulate the cold bending forming process. The second module is used to preset the roll drive parameter group. The roll drive parameter group can drive the virtual cold bending forming simulation model to generate a virtual standard cold bending forming part. The roll drive parameter group of each roll in the preset roll drive parameter group is changed several times to obtain several changed roll drive parameter groups. The third module is used to drive the virtual cold bending forming simulation model based on the changed roll drive parameter group to obtain a virtual verification cold bending forming part. The fourth module is used to acquire actual cold-bent parts from the actual production line, compare the actual cold-bent parts with different virtual verification cold-bent parts, determine the matching virtual verification cold-bent parts, and determine the parameter optimization strategy for the rolls on the actual production line based on the differences between the modified roll drive parameter group and the preset roll drive parameter group corresponding to the virtual verification cold-bent parts.

[0056] This invention discloses a method and system for collaborative optimization of multi-stage rolls in cold bending forming based on simulation data, relating to the field of production line optimization technology for cold bending forming. First, the sheet metal properties and roll geometric parameters are acquired to construct a virtual cold bending forming simulation model. Second, based on a preset set of driving parameters, several modified sets of driving parameters are generated by repeatedly changing each parameter. Then, these parameter sets are used to drive the simulation model, constructing a feature library containing multiple virtual verification formed parts. Finally, by acquiring actual formed part data from the production line and comparing it with the virtual feature library, the most similar virtual finished product is identified. Based on the differences between the modified parameters corresponding to this virtual finished product and the preset parameters, the parameter optimization strategy for the actual production line is determined. This invention, by constructing a virtual deviation mapping library, achieves precise tracing and reverse compensation of production deviations, effectively improving the efficiency of multi-stage roll optimization and the accuracy of finished products.

[0057] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented in hardware or by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) and includes several instructions to cause a computer device (such as a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0058] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to 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 deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for collaborative optimization of multi-stage rolls in cold bending forming based on simulation data, characterized in that, include: Step S100: Obtain the property parameters of the sheet material to be processed and the geometric parameters of the rolls, which are used to construct a virtual cold bending simulation model to simulate the cold bending forming process. Step S200: Preset a set of roll drive parameters. The set of roll drive parameters can drive the virtual cold bending forming simulation model to generate a virtual standard cold bending forming part. The roll drive parameters in the preset set of roll drive parameters are changed several times to obtain a set of roll drive parameters after several changes. Step S300: Based on the changed roll drive parameter group, drive the virtual cold bending forming simulation model to obtain the virtual verification cold bending forming part; Step S400: Obtain the actual cold-bent forming part of the actual production line, compare the actual cold-bent forming part with different virtual verification cold-bent forming parts, determine the matching virtual verification cold-bent forming part, and determine the parameter optimization strategy of the rolls on the actual production line based on the difference characteristics between the modified roll drive parameter group and the preset roll drive parameter group corresponding to the virtual verification cold-bent forming part.

2. The method for collaborative optimization of multi-stage rolls in cold bending forming based on simulation data according to claim 1, characterized in that, Methods for constructing virtual cold bending simulation models that simulate the cold bending process include: Step S101: Obtain the elastic-plastic property parameters of the plate, define its rheological stress curve and work hardening criterion, and construct the physical constitutive relationship of the plate element. Step S102: Import the geometric data of the multi-stage rolls and configure their spatial arrangement; perform adaptive mesh refinement on the forming deformation zone of the virtual sheet material; and establish an initial state discretization model. Step S103: Establish a multi-body contact pair between the rigid roll and the deformable sheet metal, set the contact friction coefficient and boundary constraints, and simulate the load transfer path. Step S104: According to the cold bending process sequence, rotational driving force is applied to each level of virtual rolls in sequence to simulate the dynamic extrusion and bending deformation process of the sheet material continuously passing through each level of rolls, and to realize the inter-stage coupling transfer of deformation energy and residual stress. Step S105: Perform nonlinear solution calculation to obtain the formed virtual cold-bent part, extract the three-dimensional coordinates, stress and strain data of its key points, and generate digital feature vectors.

3. The method for collaborative optimization of multi-stage rolls in cold bending forming based on simulation data according to claim 1, characterized in that, The method for making several changes to each roll drive parameter in the preset roll drive parameter group includes: Step S201: Randomly generate random parameter change strategies for several roll drive parameter groups. Each random parameter change strategy includes the parameter change amount of each roll drive parameter. Each parameter change amount is randomly selected within a preset parameter change range. Step S202: Based on the random parameter change strategy, the preset roll drive parameter group is changed to obtain the changed roll drive parameter group. The changed roll drive parameter group corresponding to the virtual verification cold bending forming part matched later is mapped, and the corresponding random parameter change strategy is mapped out, which is denoted as the mapped random parameter change strategy. Step S203: Classify the mapped random parameter change strategies based on similarity to obtain several similar mapped random parameter change strategy sets. Based on the data volume ratio between similar mapped random parameter change strategy sets, expand the content of different similar mapped random parameter change strategy sets in an equal proportion to obtain expanded random parameter change strategy sets. Based on the expanded random parameter change strategy sets, change the preset roll drive parameter group.

4. The method for collaborative optimization of multi-stage rolls in cold bending forming based on simulation data according to claim 3, characterized in that, Methods for classifying the similarity of mapped random parameter variation strategies include: Step S2031: Analyze the relative parameter change difference of each parameter change among the mapped random parameter change strategies. If all of them are less than or equal to the preset value, then the mapped random parameter change strategies are considered to be similar to each other.

5. The method for collaborative optimization of multi-stage rolls in cold bending forming based on simulation data according to claim 3, characterized in that, Methods for expanding the content of similarly mapped random parameter variation strategy sets include: Step S2032: Construct a change reference axis for each parameter change in the mapped random parameter change strategy, set several change segments for the change reference axis, and determine the selected probability corresponding to each change segment based on the mapping ratio characteristics of each change segment. Step S2033: Based on the selected probability corresponding to each change range, amplify the mapping random parameter change strategy.

6. The method for collaborative optimization of multi-stage rolls in cold bending forming based on simulation data according to claim 5, characterized in that, Based on the mapping proportion characteristics of each change segment, methods for determining the selected probability corresponding to each change segment include: Step S20321: Analyze each change in all random parameter change strategies, determine the change range corresponding to each change, and count the number of times the change range is mapped by the change, which is recorded as the number of regular mappings. Step S20322: Analyze each change in the mapped random parameter change strategy, determine the change range corresponding to each change, and count the number of times the change range is mapped by the change, which is recorded as the number of high-value mappings. Step S20323: Configure weight coefficients for the number of regular mappings and the number of high-value mappings corresponding to each change range, calculate the product of the weight coefficient and the number of regular mappings, and record it as the regular mapping influence parameter. Calculate the product of the weight coefficient and the number of high-value mappings, and record it as the high-value mapping influence parameter. Calculate the sum of the regular mapping influence parameter and the high-value mapping influence parameter, and record it as the mapping weight. Step S20324: Determine the selected probability corresponding to different change segments based on the ratio between the mapping weights corresponding to different change segments.

7. The method for collaborative optimization of multi-stage rolls in cold bending forming based on simulation data according to claim 1, characterized in that, Methods for determining matching virtual verification cold-formed parts include: Step S401: Define the bent portions in the cold-formed part and determine the bending sequence of each bent portion. Step S402: Compare the cross sections of the bent portions with the same bending sequence between the cold-formed parts, calculate the overlap parameters between them, and if all overlap parameters are greater than or equal to the preset values, then the cold-formed parts are considered to be matched with each other.

8. The method for collaborative optimization of multi-stage rolls in cold bending forming based on simulation data according to claim 7, characterized in that, Methods for defining the bent portion in a cold-formed part include: Step S4011: Determine the trend line of the side section of the cold-formed part. If the trend line has a non-straight starting point, start cutting the bending part. If the trend line has a straight starting point, then the starting point is identified as the end position of the bending part cutting.

9. The method for collaborative optimization of multi-stage rolls in cold bending forming based on simulation data according to claim 7, characterized in that, Methods for comparing and aligning the cross-sections of the bent portion include: Step S4021: Determine the trend line of the cross section of the bent part, and make the respective trend lines coincide. The coincidence method includes determining the center of the trend line, aligning the center, dynamically rotating the trend line, and analyzing the coincidence ratio between the two in real time. Step S4022: The largest overlap ratio is identified as the overlap parameter.

10. A multi-stage roll collaborative optimization system for cold bending forming based on simulation data, characterized in that, include: The first module is used to obtain the property parameters of the sheet material to be processed and the geometric parameters of the rolls, and to build a virtual cold bending simulation model to simulate the cold bending forming process. The second module is used to preset the roll drive parameter group. The roll drive parameter group can drive the virtual cold bending forming simulation model to generate a virtual standard cold bending forming part. The roll drive parameter group of each roll in the preset roll drive parameter group is changed several times to obtain several changed roll drive parameter groups. The third module is used to drive the virtual cold bending forming simulation model based on the changed roll drive parameter group to obtain a virtual verification cold bending forming part. The fourth module is used to acquire actual cold-bent parts from the actual production line, compare the actual cold-bent parts with different virtual verification cold-bent parts, determine the matching virtual verification cold-bent parts, and determine the parameter optimization strategy for the rolls on the actual production line based on the differences between the modified roll drive parameter group and the preset roll drive parameter group corresponding to the virtual verification cold-bent parts.

Images