Elastic sheet structure generation method and device, computer equipment and storage medium

Abstract

The invention relates to an elastic sheet structure generation method and device, computer equipment and a storage medium. The method comprises the following steps: generating an initial structure control parameter of an elastic sheet based on a preset spatial constraint parameter; fitting is carried out based on the initial structure control parameters, and a three-dimensional elastic piece model is obtained; the elastic sheet three-dimensional model is an elastic sheet three-dimensional model conforming to a preset geometric verification rule; performing performance simulation on the elastic sheet three-dimensional model to obtain a performance simulation result; and based on the performance simulation result, performing iterative optimization on the initial structure control parameter to obtain a target structure control parameter and a target elastic sheet three-dimensional model corresponding to the target structure control parameter. By adopting the method, the elastic sheet design efficiency can be improved, the labor cost is reduced, and the elastic sheet performance is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and in particular to a spring structure generation method and device, computer equipment and a storage medium. BACKGROUND

[0002] With the development of electronic device integration and miniaturization, the spring structure plays an important role in the connection of flexible circuit board antennas, laser direct forming antennas and circuit boards. However, the traditional spring design is highly dependent on manual experience, and the design efficiency is low and difficult to meet complex geometric constraints and performance requirements, which requires manual adjustment of parameters and re-simulation, and the degree of automation is low, time-consuming and labor-intensive.

[0003] Therefore, there is still a problem of low spring structure design efficiency and high labor cost in the prior art. SUMMARY

[0004] Therefore, it is necessary to provide a spring structure generation method, device, computer equipment and storage medium capable of improving the spring design efficiency, reducing the labor cost and improving the spring performance in view of the above technical problems.

[0005] In a first aspect, the present application provides a spring structure generation method, which comprises:

[0006] generating initial structure control parameters of a spring based on preset spatial constraint parameters;

[0007] fitting based on the initial structure control parameters to obtain a spring three-dimensional model; the spring three-dimensional model is a spring three-dimensional model meeting preset geometric verification rules;

[0008] performing performance simulation on the spring three-dimensional model to obtain a performance simulation result;

[0009] based on the performance simulation result, iteratively optimizing the initial structure control parameters to obtain target structure control parameters and a target spring three-dimensional model corresponding to the target structure control parameters.

[0010] In one embodiment, the initial structure control parameters include a control point set; and the generating initial structure control parameters of a spring based on preset spatial constraint parameters comprises:

[0011] generating a plurality of seed control points meeting the preset spatial constraint parameters in a preset simulation space;

[0012] iteratively generating a plurality of candidate control points based on a plurality of the seed control points;

[0013] determining the control point set based on the candidate control points meeting the preset spatial constraint parameters.

[0014] In one of the embodiments, the fitting based on the initial structure control parameters to obtain the 3D model of the shell includes:

[0015] Based on the B-spline curve fitting algorithm, the initial structure control parameters are curve fitted to obtain a shell medial axis curve satisfying a preset curvature condition;

[0016] The shell medial axis curve is subjected to normal plane offsetting and thickness stretching processing to obtain the 3D model of the shell.

[0017] In one of the embodiments, the preset geometric verification rule includes one or more of boundary constraint verification, self-intersection constraint verification, curvature constraint verification, and spacing constraint verification;

[0018] The boundary constraint verification includes verifying, based on a preset polygonal region and a preset safety distance, whether the 3D model of the shell is located within the preset simulation space and maintains a safety distance from the boundary of the preset simulation space; the self-intersection constraint verification includes detecting, based on a preset self-intersection judgment algorithm, whether the medial axis curve has self-intersection other than the end points; the curvature constraint verification includes verifying, based on discrete points of the curve, whether the curvature radius of the medial axis curve is not less than a preset lower limit of curvature; and the spacing constraint verification includes verifying, based on the coordinates of the control points, whether the distance between any two non-adjacent control points is not less than a preset minimum spacing.

[0019] In one of the embodiments, the performance simulation of the 3D model of the shell includes:

[0020] Based on the material parameters of the 3D model of the shell and preset grid parameters, the 3D model of the shell is subjected to grid division to obtain a grid model of the 3D model of the shell;

[0021] Based on the grid model, a load or displacement is applied to solve the stiffness value and the maximum stress value of the shell in multiple directions to obtain a performance simulation result.

[0022] In one of the embodiments, the number of initial structure control parameters is multiple; and the iterative optimization of the initial structure control parameters based on the performance simulation result includes:

[0023] Based on the performance simulation results of the 3D models of the shells corresponding to the multiple initial structure control parameters and a preset performance evaluation function, the initial structure control parameter corresponding to the 3D model of the shell with the optimal simulation result is taken as the current optimal control parameter;

[0024] Based on a weighted reorganization operator, the current optimal control parameter is subjected to cross operation to generate multiple candidate control parameters as the initial structure control parameters of the next round of optimization process.

[0025] In one embodiment, the performance simulation on the shell three-dimensional model to obtain the performance simulation result comprises:

[0026] Based on each initial structure control parameter, the initial structure control parameter is assigned to a parallel computing task queue;

[0027] Based on multiple independent solver instances, the performance simulation on each shell three-dimensional model to obtain the performance simulation result is synchronously executed.

[0028] In a second aspect, the present application provides a shell structure generation device, the device comprising:

[0029] A parameter generation module is configured to generate an initial structure control parameter of a shell based on a preset spatial constraint parameter;

[0030] A model generation module is configured to obtain a shell three-dimensional model based on fitting of the initial structure control parameter; the shell three-dimensional model is a shell three-dimensional model that meets a preset geometric verification rule;

[0031] A performance simulation module is configured to perform performance simulation on the shell three-dimensional model to obtain a performance simulation result;

[0032] A parameter optimization module is configured to perform iterative optimization on the initial structure control parameter based on the performance simulation result to obtain a target structure control parameter and a target shell three-dimensional model corresponding to the target structure control parameter.

[0033] In a third aspect, the present application provides a computer device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the method as described above when executing the computer program.

[0034] In a fourth aspect, the present application provides a computer readable storage medium having a computer program stored thereon, and the computer program is executed by a processor to implement the method as described above.

[0035] The above-mentioned shrapnel structure generation method, device, computer equipment and storage medium generate initial structural control parameters of the shrapnel based on preset spatial constraint parameters; perform fitting based on the initial structural control parameters to obtain a three-dimensional model of the shrapnel; the three-dimensional model of the shrapnel is a three-dimensional model of the shrapnel that meets preset geometric verification rules; perform performance simulation on the three-dimensional model of the shrapnel to obtain performance simulation results; based on the performance simulation results, iteratively optimize the initial structural control parameters to obtain target structural control parameters and a target shrapnel three-dimensional model corresponding to the target structural control parameters, reduce dependence on manual experience through automated parameter configuration, achieve rapid modeling with the help of geometric fitting, effectively avoid obvious geometric defects in the automatically designed three-dimensional model of the shrapnel through preset geometric verification rules, form a closed-loop optimization mechanism in combination with simulation feedback, and output model results that meet the requirements of spatial assembly, manufacturing process and electrical performance, thereby achieving the technical effects of improving design efficiency, reducing labor costs and optimizing shrapnel performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A diagram showing an application environment of a method for generating a spring structure in one embodiment;

[0037] Figure 2 Schematic diagram of a flow chart of a method for generating a spring structure in one embodiment;

[0038] Figure 3 It is a structural block diagram of a spring-flake structure generating device in one embodiment;

[0039] Figure 4 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0041] The spring structure generation method provided in the embodiment of the present application can be applied to Figure 1The application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data required by the server 104 to process. The data storage system can be integrated on the server 104, or placed on the cloud or other network servers. The terminal 102 generates the initial structure control parameter of the shell based on the preset spatial constraint parameter; based on the initial structure control parameter, the shell three-dimensional model is obtained by fitting; the shell three-dimensional model is a shell three-dimensional model that meets the preset geometric verification rule; the performance simulation result is obtained by performing performance simulation on the shell three-dimensional model; based on the performance simulation result, the initial structure control parameter is iteratively optimized to obtain the target structure control parameter and the target shell three-dimensional model corresponding to the target structure control parameter. Among them, the terminal 102 can be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0042] In one embodiment, as Figure 2 shown, a shell structure generation method is provided, and the shell structure generation method comprises:

[0043] Step S100, generating the initial structure control parameter of the shell based on the preset spatial constraint parameter.

[0044] Among them, the preset spatial constraint parameter can be a parameter set for constraining the space range in the shell design process, which is used to provide basic space limitation for the design of the shell structure, and ensure that the generated structure meets the physical assembly requirements. Exemplarily, the preset spatial constraint parameter can include one or more of the parameters such as space boundary, length limit, width limit, height limit, etc. In this embodiment, the preset spatial constraint parameter can be input by the user, and can also be obtained by extracting from the configuration file pre-stored in the electronic device.

[0045] The initial structure control parameter can be a basic parameter for controlling the generation process of the shell three-dimensional model, which can cooperate with the generation algorithm of the shell three-dimensional model, so that the generation algorithm can generate a determined shell three-dimensional model based on the initial structure control parameter. In one exemplary embodiment, the initial structure control parameter can be a parameter set composed of one or more control points, and the shell three-dimensional model is generated based on the above control points as the shape basis. In another exemplary embodiment, the initial structure control parameter can also be a control curve describing the overall direction of the shell three-dimensional model, and the shell three-dimensional model can be obtained by transforming and extending based on the control curve. The initial structure control parameter can also be other parameters for controlling the generation process of the shell three-dimensional model, which is not limited in this embodiment.

[0046] The initial structure control parameter of the shell fragment can be determined by determining a design boundary based on the preset spatial constraint parameter and generating a parameter for controlling the modeling constraint of the three-dimensional model of the shell fragment. For example, the corresponding parameter can be extracted from the pre-stored parameter mapping table by a lookup table method, and the initial value can also be generated by a mathematical function calculation, so as to realize the automatic initial configuration of the shell structure parameter and reduce the manual setting workload.

[0047] In step S200, the three-dimensional model of the shell fragment is obtained by fitting based on the initial structure control parameter. The three-dimensional model of the shell fragment is a three-dimensional model of the shell fragment that meets the preset geometric verification rule.

[0048] The three-dimensional model of the shell fragment can be a three-dimensional digital form of the geometric structure of the shell fragment. In this embodiment, the three-dimensional model of the shell fragment can be obtained by constructing according to the initial structure control parameter. It can be understood that, since the initial structure control parameter determines the basic shape and / or extension direction of the three-dimensional model of the shell fragment, the three-dimensional model of the shell fragment can be constructed by fitting according to the basic shape and / or extension direction. Further, the three-dimensional model of the shell fragment can include but is not limited to a linear shell model, a U-shaped shell model, a spiral shell model, etc.

[0049] The preset geometric verification rule can be a rule set for judging whether the three-dimensional model of the shell fragment meets the basic geometric structure feasibility, for ensuring that the generated three-dimensional model of the shell fragment has manufacturability and structural integrity. In one specific embodiment, the preset geometric verification rule can be verified by the preset structure control parameter and the three-dimensional model of the shell fragment according to a preset algorithm, for example, judging whether the generated three-dimensional model of the shell fragment meets the requirements of the preset spatial constraint parameter, judging whether the three-dimensional model of the shell fragment has problems such as too small curvature and curve intersection, which obviously have manufacturing difficulty and / or low performance, to verify whether it meets the preset geometric condition. For example, the preset geometric verification rule can include but is not limited to one or more of the minimum curvature radius rule, the structural continuity rule, the non-self-intersection rule, etc.

[0050] The three-dimensional model of the shell fragment can be obtained by fitting based on the initial structure control parameter, that is, the fitting three-dimensional model is obtained after fitting based on the initial structure control parameter, and the fitting three-dimensional model is verified by the preset geometric verification rule, so that the model meeting the preset geometric verification rule is output as the three-dimensional model of the shell fragment, to ensure that the generated three-dimensional model of the shell fragment basically meets the requirements of the geometric rule, and to avoid performance simulation of the model that obviously does not have physical feasibility, thereby avoiding waste of computing resources.

[0051] In step S300, the performance simulation result of the three-dimensional model of the shell fragment is obtained.

[0052] The performance simulation result can be a data result output by performing physical performance analysis on the three-dimensional model of the shell, and is used to evaluate the performance of the shell under actual working conditions. In this embodiment, the performance simulation result can be calculated by a finite element analysis software to perform mechanical performance calculation, and by solving the mechanical response of the shell, performance indexes such as stiffness value and maximum stress value are extracted, so that the shell performance is evaluated without physical testing, and the verification period is shortened. Further, the performance simulation result can include but is not limited to one or more of stress distribution result, stiffness simulation result, springback amount simulation result, and contact resistance simulation result.

[0053] In step S400, the initial structure control parameters are iteratively optimized based on the performance simulation result to obtain target structure control parameters and a target three-dimensional model of the shell corresponding to the target structure control parameters.

[0054] The target structure control parameters can be a set of shell structure parameters that meet the performance requirements after iterative optimization, and are used to generate a final available three-dimensional model of the shell. In an exemplary embodiment, the target structure control parameters can be obtained by using an evolutionary strategy algorithm and other iterative optimization algorithms to continuously adjust the structure control parameters based on the performance simulation result by using an optimization control module, so as to obtain an optimal scheme that meets the multi-objective performance requirements. Further, the target structure control parameters can include but are not limited to one or more of target length parameter, target width parameter, target height parameter, and target curvature distribution.

[0055] The target three-dimensional model of the shell can be a final three-dimensional structure of the shell that meets the spatial constraints and performance requirements, and is used as a design output that can be put into manufacturing. In this embodiment, the target three-dimensional model of the shell can be a three-dimensional solid model regenerated by calling a geometric modeling method based on the target structure control parameters.

[0056] Based on the performance simulation result, the initial structure control parameters are iteratively optimized. For example, a more optimal design scheme can be searched and a model can be regenerated by using a simulation result feedback mode and a driving parameter adjustment mode. For example, a gradient descent method, a genetic algorithm, or the like can be used to search for a more optimal parameter combination, and the parameters are updated and the model is regenerated each time, so that automatic optimization of the parameters is realized, and manual repeated debugging is reduced.

[0057] Taking the antenna shrapnel design scenario of smart wearable devices as an example, when designing the connecting shrapnel between the flexible circuit board antenna and the laser direct forming antenna of a smart watch, the spatial boundary inside the watch case can be input as the preset spatial constraint parameter. The terminal 102 generates the initial structural control parameters within the spatial boundary and fits the shrapnel three-dimensional model; after the model passes the geometric verification rule check, performance simulation is performed to obtain performance simulation results; the terminal 102 adjusts the initial structural control parameters such as curvature based on the simulation results and key performance indicators, and outputs the target shrapnel three-dimensional model that meets the key performance indicators after multiple iterations.

[0058] This embodiment provides a method for generating a shrapnel structure, which generates initial structural control parameters of the shrapnel based on preset spatial constraint parameters; performs fitting based on the initial structural control parameters to obtain a three-dimensional model of the shrapnel; the three-dimensional model of the shrapnel is a three-dimensional model of the shrapnel that meets preset geometric verification rules; performs performance simulation on the three-dimensional model of the shrapnel to obtain performance simulation results; based on the performance simulation results, iteratively optimizes the initial structural control parameters to obtain target structural control parameters and a target three-dimensional model of the shrapnel corresponding to the target structural control parameters, reduces dependence on manual experience through automated parameter configuration, achieves rapid modeling with the help of geometric fitting, effectively avoids obvious geometric defects in the automatically designed three-dimensional model of the shrapnel through preset geometric verification rules, forms a closed-loop optimization mechanism in combination with simulation feedback, and outputs model results that meet the requirements of spatial assembly, manufacturing process and electrical performance, thereby achieving the technical effects of improving design efficiency, reducing labor costs and optimizing shrapnel performance.

[0059] In one embodiment, the initial structural control parameters include a set of control points; based on the preset spatial constraint parameters, the initial structural control parameters of the spring fragment are generated including:

[0060] In a preset simulation space, multiple seed control points that meet preset spatial constraint parameters are generated;

[0061] Based on multiple seed control points, iteratively generate multiple candidate control points;

[0062] A set of control points is determined based on candidate control points that meet preset spatial constraint parameters.

[0063] The seed control points can be basic control points that meet spatial constraints and are initialized within a preset simulation space, and can serve as the distribution starting point for generating candidate control points. In this embodiment, the seed control points can be distributed and generated within the allowed design area based on preset spatial constraint parameters to ensure that the positions of the control points do not violate boundary restrictions or avoidance area requirements. Furthermore, the seed control points can be generated using random distribution, regular grid distribution, or other methods to achieve spatial coverage of the initial point positions.

[0064] The plurality of seed control points meeting the preset spatial constraint parameters can be a feasible region defined according to the preset spatial constraint parameters, and initial point positions are distributed in the simulation space to ensure that all seed control points are located within the allowed distribution range and within the specified boundary range. Further, the plurality of seed control points meeting the minimum point spacing and the boundary safety distance can be generated in the polygon design region specified by the preset spatial constraint parameters by using the Poisson disc sampling algorithm. Further, the fixed starting point of the key connection region can be included as an anchor point in the sampling result, thereby improving the rationality and space utilization of the initial point position distribution.

[0065] The candidate control points can be potential control points generated by expanding the seed control points to enrich the geometric expression capability of the control point set. Further, the candidate control points can be new point positions generated based on the seed control points by using local perturbation, directional expansion, topological derivation strategy and the like, and the control points obtained after geometric verification. For example, the directional expansion can be a vector offset in the normal direction or the tangential direction, and the local perturbation can be a random direction perturbation introduced to increase the morphological diversity.

[0066] The control point set can be an ordered or unordered set composed of candidate control points meeting the preset spatial constraint parameters, and can be used as a basic data structure for describing the geometric shape of the shell. In this embodiment, the control point set is formed by selecting point positions that completely meet the spatial constraint conditions from the generated candidate control points and organizing them in geometric order or topological relationship, for determining the direction and shape of the axis of the shell.

[0067] Based on the candidate control points meeting the preset spatial constraint parameters, the control point set can be determined by performing spatial compliance verification on all generated candidate control points, selecting point positions that meet the boundary, avoidance area and curvature limit, and organizing them into the final control point set according to geometric continuity or topological order. Further, it can be determined by Boolean operation whether the point is located outside the region specified by the boundary, and the isolated point group can be removed by combining connectivity analysis, thereby ensuring the geometric feasibility of the control point set.

[0068] The shell structure generation method provided in this embodiment generates seed control points meeting the preset spatial constraint parameters in a preset simulation space, iteratively generates a plurality of candidate control points based on a plurality of seed control points, determines a control point set based on candidate control points meeting the preset spatial constraint parameters, and gradually constructs a legal and diversified control point structure through a hierarchical generation mechanism. The generation process of the initial structure control parameters of the shell has spatial compliance guarantee and shape exploration ability, which can achieve the technical effects of improving the degree of design automation, enhancing the structure adaptability and design flexibility, reducing manual intervention, and improving the effectiveness of generating a shell model in a complex geometric environment.

[0069] In one embodiment, fitting based on the initial structure control parameters to obtain the three-dimensional model of the shell includes:

[0070] Based on the B-spline curve fitting algorithm, the initial structure control parameters are fitted to obtain a shell center axis curve that satisfies the preset curvature condition;

[0071] The shell center axis curve is subjected to normal plane offsetting and thickness stretching processing to obtain a three-dimensional model of the shell.

[0072] In this embodiment, the construction of the three-dimensional model of the shell can be to generate a smooth center axis parameter curve by fitting the initial structure control parameters, and then offsetting in the normal direction of the center axis curve to construct a three-dimensional shell entity model with thickness and width.

[0073] The B-spline curve fitting algorithm can be to generate a parameter curve of the shell center axis by fitting control points, calculate the node vector of the spline curve according to the number and distribution of the control points, and ensure smooth transition of the curve segment, which can be used to generate a smooth and manufacturing requirement-compliant center axis path to avoid performance degradation caused by sudden curvature. Exemplarily, the B-spline curve fitting algorithm can include but is not limited to a cubic B-spline algorithm, a non-uniform B-spline algorithm, or a rational B-spline algorithm, etc.

[0074] The shell center axis curve can be a one-dimensional parameterized curve describing the geometric center path of the shell. By offsetting the center axis curve in the normal plane direction thereof, the upper and lower boundary curves of the shell can be generated, and further, the boundary curves can be stretched into upper and lower surfaces according to the set thickness to generate a three-dimensional entity model with closed thickness. In this embodiment, the shell center axis curve generates an initial center axis curve by calling the B-spline formula and performs geometric legality verification to ensure that the model satisfies the curvature constraint.

[0075] The preset curvature condition can be a limit condition for the local bending degree of the shell center axis curve, which is used to avoid stress concentration or manufacturing difficulty caused by excessive bending, and can be used to ensure that the generated center axis curve has good manufacturability and mechanical stability. In one specific embodiment, the preset curvature condition can pre-set a minimum curvature radius or a maximum curvature change rate threshold according to material properties, processing technology, and reliability requirements. Exemplarily, the preset curvature condition can include one or more of a minimum curvature radius threshold, a curvature continuity level, a curvature change smoothness requirement, etc.

[0076] Further, the normal plane offsetting and thickness stretching processing of the shell center axis curve can also change the local width of the shell by controlling the function of the offset distance in the normal direction of the curve to realize dynamic adjustment of the width.

[0077] The method for generating a shell structure provided in the embodiment comprises the following steps: performing curve fitting on initial structure control parameters based on a B-spline curve fitting algorithm to obtain a shell central axis curve meeting a preset curvature condition; performing normal plane offsetting and thickness stretching on the shell central axis curve to obtain a shell three-dimensional model; generating a continuous and smooth central axis path by fitting control points using the B-spline curve fitting algorithm and verifying the geometric rationality by combining curvature constraint verification; and performing normal offsetting and thickness stretching to realize accurate modeling from a parameterized curve to a three-dimensional entity, so that a high-performance and manufacturable shell structure can be automatically generated in a complex and compact space, the modeling automation degree and design quality are improved, and the demand for manual intervention is reduced.

[0078] In one of the embodiments, the preset geometric verification rules include one or more of boundary constraint verification, self-intersection constraint verification, curvature constraint verification, and spacing constraint verification.

[0079] The boundary constraint verification comprises verifying whether the shell three-dimensional model is located in a preset simulation space and maintains a safe distance from the boundary of the preset simulation space based on a preset polygonal region and a preset safe distance. The preset polygonal region can be a closed geometric region used to limit the space in which the shell three-dimensional model can exist, and is used to provide a spatial reference for the boundary constraint verification to ensure that the shell structure does not exceed the allowed physical range. For example, the preset polygonal region can extract the contour boundary of the installation region from an electronic device structure design file or draw a specified region by a user through a graphical interface. The preset polygonal region can participate in the boundary constraint verification together with the preset safe distance. The preset safe distance can be a minimum interval distance that needs to be maintained between the shell structure and the boundary of the simulation space, and is used to prevent the shell from touching the surrounding devices or structural members due to manufacturing or assembly deviations. In one exemplary embodiment, the preset safe distance can be obtained according to empirical values or calculated values of manufacturing tolerances, material thermal expansion coefficients, or assembly errors. The boundary constraint verification can be quickly verified by point-polygon inclusion relationship judgment combined with distance field calculation, so that the shell structure can be ensured to be in the specified space and have sufficient assembly allowance, and the design feasibility is improved.

[0080] The self-intersection constraint verification comprises detecting whether the central axis curve has a self-intersection other than the end points based on a preset self-intersection judgment algorithm. In the embodiment, the self-intersection constraint verification can be based on the parameter expression or the discrete line segment set of the central axis curve, and the preset self-intersection judgment algorithm is used to detect whether there is a non-end point intersection. For example, the preset self-intersection judgment algorithm can use one or more of a line segment intersection detection algorithm, a bounding box pruning algorithm, and a parameter equation solving algorithm, so that a self-intersection structure can be avoided, and the geometric rationality and manufacturability of the shell are ensured.

[0081] The curvature constraint verification includes verifying whether the curvature radius of the medial axis curve is not less than a preset curvature lower limit based on discrete points of the curve. The discrete points of the curve can be a point set sampled along the medial axis curve at a fixed step or in an adaptive manner, or can be control points based on which the medial axis curve is fitted, and are used for numerically evaluating the local geometric characteristics of the medial axis curve, such as the curvature radius. Further, the discrete points of the curve can be used for calculating the curvature radius in the curvature constraint verification. The preset curvature lower limit can be a minimum curvature radius threshold allowed by the shell medial axis curve, and is used to prevent stress concentration or processing fracture of the shell due to excessive bending. For example, the preset curvature lower limit can be set according to the material yield strength, the minimum bending process capability, the fatigue life requirement, and the like. Further, the preset curvature lower limit can participate in the curvature constraint verification together with the discrete points of the curve. It can be understood that the curvature constraint verification can calculate the curvature distribution of the curve by differential geometry, verify whether the curvature radius is greater than the specified lower limit everywhere, and thus can prevent the shell from being excessively bent to cause material damage or signal transmission distortion.

[0082] The distance constraint verification includes verifying whether the distance between any two non-adjacent control points is not less than a preset minimum distance based on the control point coordinates. The preset minimum distance can be a minimum distance requirement that must be met between any two non-adjacent control points, and is used to avoid local overlapping of the structure or manufacturing difficulty caused by excessively dense distribution of the control points. In an exemplary embodiment, the preset minimum distance can be set based on the shell material thickness, the etching process resolution, the structure stability index, and the like. Further, the preset minimum distance can participate in the distance constraint verification together with the control point coordinates. Further, the distance constraint verification can use KD-tree to accelerate the nearest point pair search, and reduce the computational complexity. Through the distance constraint verification, the unreasonable distribution of the control points can be effectively avoided to cause structural abnormalities, and the stability of the parameterized modeling is improved.

[0083] The shell structure generation method provided in this embodiment introduces a multi-dimensional geometric verification mechanism by one or more of the boundary constraint verification, the self-intersection constraint verification, the curvature constraint verification, and the distance constraint verification, automatically checks the spatial adaptability, the structure continuity, the bending reasonableness, and the parameter stability of the shell model, effectively intercepts unfeasible designs, reduces invalid simulation iterations, further improves the robustness and output quality of the automatic design, and achieves the technical effects of enhancing the reliability and efficiency of the shell structure design.

[0084] In one of the embodiments, the performance simulation on the shell three-dimensional model includes:

[0085] Based on the material parameters of the shell three-dimensional model and the preset grid parameters, the shell three-dimensional model is meshed to obtain a grid model of the shell three-dimensional model;

[0086] The stiffness values and maximum stress values of the spring in multiple directions are solved based on the load or displacement applied to the grid model, and performance simulation results are obtained.

[0087] The material parameters can be a set of parameters of physical properties of materials used by the elastic three-dimensional model, and are applied to performance simulation calculation of the spring three-dimensional model. For example, the material parameters can include, but are not limited to, one or more of elastic modulus, Poisson's ratio, yield strength, etc.

[0088] The preset grid parameters can be configuration parameters for controlling the density and quality of finite element grid division, and can be used to affect the fineness and simulation calculation efficiency of the grid model. In an exemplary embodiment, the preset grid parameters can set the unit type, size, etc. of the grid division according to the simulation stage, and further can select a shell unit or a solid unit for configuration. For example, the preset grid parameters can include, but are not limited to, one or more of the unit size parameter, the grid growth rate parameter, the boundary layer number parameter, etc.

[0089] The grid model can be a numerical calculation model that discretizes the electronic component contact piece three-dimensional model into a finite number of units and nodes, and can be used as a calculation carrier for performance simulation and applied to numerical solution of mechanical equations.

[0090] Based on the material parameters and the preset grid parameters, the electronic component contact piece three-dimensional model is divided into a grid, and in an exemplary embodiment, the material properties and grid division settings can be combined to call a finite element analysis interface tool to generate a grid model corresponding to the spring three-dimensional model, including node and unit structures, thereby ensuring that the simulation model has sufficient geometric fidelity and calculation stability, and improving the credibility of the simulation results.

[0091] The load and displacement can be external excitation conditions simulating the mechanical action received by the spring during assembly or use. For example, according to different application methods, the load can include, but is not limited to, concentrated force load, distributed pressure load, etc., and the displacement can include forced displacement boundary conditions.

[0092] Based on the grid model, the load or displacement is applied to solve the stiffness values and maximum stress values. The stiffness values and maximum stress values of the spring in multiple directions can be calculated by applying fixed constraints and displacement or force loads on the grid model to control the solver to perform statics analysis. Further, by applying different load combinations under different working conditions in steps, multi-scenario simulation can be realized, thereby realizing multi-dimensional quantitative evaluation of the mechanical properties of the spring.

[0093] The shell structure generation method provided in the embodiment can achieve the technical effect of enhancing the reliability of the simulation result by introducing the material parameters and the preset mesh parameters to divide the mesh model, and combining the load or displacement condition to realize the quantitative analysis of the mechanical properties of the shell.

[0094] In one of the embodiments, the number of initial structure control parameters is multiple; the iterative optimization of the initial structure control parameters based on the performance simulation result includes:

[0095] Based on the performance simulation result of the shell three-dimensional model corresponding to the multiple initial structure control parameters and the preset performance evaluation function, the initial structure control parameter corresponding to the shell three-dimensional model with the optimal simulation result is taken as the current optimal control parameter.

[0096] Based on the weighted reorganization operator, the current optimal control parameter is cross-operated to generate multiple candidate control parameters as the initial structure control parameters in the next round of optimization process.

[0097] The preset performance evaluation function can be a comprehensive fitness function constructed according to multiple objective and multiple constraint requirements, which guides the search of a design scheme satisfying the constraints in the form of a penalty term, thereby providing a comparable performance score basis for different design schemes. In one exemplary embodiment, the preset performance evaluation function can be a simulation index including one or more of stiffness, stress, rebound amount, contact resistance, etc., and mapping to a single score value, and through the weighted index combination, the quantitative evaluation of the multiple objective performance is realized. In some other embodiments, the preset performance evaluation function can also be a threshold judgment function, a hierarchical evaluation function, etc., which is not limited in the embodiment.

[0098] The current optimal control parameter can be the initial structure control parameter with the highest performance score in the current optimization round, which can be used to generate the next generation of candidate control parameters. In one exemplary embodiment, the current optimal control parameter can be the candidate control parameter corresponding to the highest score from the multiple initial structure control parameters according to the score of the preset performance evaluation function.

[0099] The weighted recombination operator can be a combination manner of updating the parameter vector based on numerical weights, so as to explore new parameter combinations while maintaining good design features. In one specific embodiment, the weighted recombination operator can assign different weights to the dimensional parameters of the current optimal control parameters, and generate a new parameter vector through linear combination or nonlinear mapping. For example, the weighted recombination operator can use linear weighting operators, Gaussian perturbation weighting operators, adaptive weighting operators, etc.

[0100] In this embodiment, the candidate control parameters can be a set of potentially feasible parameters generated by weighted recombination, which can be used as the initial structural control parameters for the next round of optimization process. For example, the weighted recombination mechanism can be used to combine the control point position information of multiple elite individuals to generate new solutions, while applying random mutations conforming to a normal distribution to the current optimal control parameters and combining a mutation rejection mechanism to ensure the legality of offspring, to generate multiple candidate control parameters, so as to explore the adjacent design space while preserving high-performance design features, and improve the optimization convergence efficiency.

[0101] The method for generating a shell structure provided in this embodiment can effectively avoid invalid simulation rounds, improve the convergence speed and result quality of the optimization process under multi-objective and multi-constraint conditions, and can achieve the technical effects of improving the shell design efficiency and reducing the dependence on the experience of designers.

[0102] In one embodiment, the performance simulation of the shell three-dimensional model obtains performance simulation results including:

[0103] Based on each initial structural control parameter, the initial structural control parameter is assigned to a parallel computing task queue.

[0104] Based on multiple independent solver instances, the performance simulation of each shell three-dimensional model is synchronously performed to obtain performance simulation results.

[0105] In this embodiment, the parallel computing task queue can be a scheduling structure for managing performance simulation tasks, used to realize centralized management and resource scheduling of multiple tasks, and improve the organization efficiency of the simulation process. In this embodiment, the parallel computing task queue can receive simulation tasks corresponding to the initial structural control parameters and distribute tasks to multiple independent solver instances. Further, the parallel computing task queue can also use static task queues, dynamic load balancing queues, or priority sorting task queues, etc.

[0106] The initial structure control parameters can be assigned to the parallel computing task queue by encapsulating each set of initial structure control parameters as a standardized simulation task unit and sequentially or based on load status injecting the parallel computing task queue for execution. In a specific embodiment, the tasks can be assigned by using a polling method, and the assignment strategy can be dynamically adjusted according to the current load of each solver instance, so that the simulation tasks can be automatically submitted in batches, manual start can be avoided, and the task management efficiency can be improved.

[0107] The independent solver instance can be a simulation engine process running in an independent memory space or computing node, used to perform finite element analysis or other physical field solving. Through the processing of multiple independent solver instances, performance simulation of multiple shell three-dimensional models can be simultaneously performed, and the overall simulation time consumption can be shortened.

[0108] In an exemplary embodiment, multiple independent simulation software solver processes can be started by calling open source tool interfaces, each instance loads the grid model and performs boundary condition and load setting in performance simulation, and independently completes the mechanical performance simulation of the shell three-dimensional model.

[0109] Based on multiple independent solver instances, simultaneous performance simulation of each shell three-dimensional model can be performed by triggering multiple independent solver instances to concurrently read the shell three-dimensional model in the task queue by the task scheduling module, each instance calls the grid division function to complete grid division, sets material properties, boundary conditions and loads through the model analysis function, and calls the simulation software solver to complete statics calculation to generate performance simulation results including stiffness in each direction and maximum stress. By utilizing the operating system level parallel capability to realize true synchronous simulation, the overall time of multi-scheme simulation can be significantly shortened, and the overall response speed of the iteration optimization process can be improved.

[0110] The shell structure generation method provided in the embodiment assigns the initial structure control parameters to the parallel computing task queue based on each initial structure control parameter, and simultaneously performs performance simulation of each shell three-dimensional model based on multiple independent solver instances to obtain performance simulation results. By uniformly incorporating the simulation tasks corresponding to multiple initial structure control parameters into the parallel computing task queue for centralized management, and utilizing multiple independent solver instances to simultaneously perform performance simulation, the time bottleneck of single-process serial simulation can be effectively broken, the waiting time can be reduced, the hardware resource utilization rate can be improved, and the technical effects of improving the automation efficiency and response speed of shell structure generation can be achieved.

[0111] In order to more clearly set forth the technical solutions of the present application, a detailed embodiment is further provided.

[0112] Applicant found through research that the current mainstream VCM spring design mostly follows the empirical structure (such as four-corner double spring arms, S-shaped spring, etc.), which is relatively conservative in geometric expansion, and the design space is limited by human experience. Due to the lack of automated design generation means, engineers often fine-tune the size from existing mature schemes to meet new requirements, and it is difficult to find breakthrough new structures in a timely manner. For example, in order to balance the compliance in the vertical direction and the stability in the horizontal direction, the classic design is to add long elastic arms to the four sides of the square outer frame to form a "cross" or "frame + cantilever" structure. But this preset topology may not be the best, it is just an experienced choice. The lack of topology optimization means may lead to the solidification of the spring topology structure, and the performance is close to the bottleneck. In addition, the spring shape parameters are numerous and coupled with each other, and manual parameter adjustment is difficult to understand the whole situation. Designers often focus on ensuring that indicators such as stiffness and stroke meet the standards, but do not consider stress concentration and modal characteristics, leaving hidden dangers. For example, some springs thin local thickness to increase the stroke, which may introduce stress hotspots and reduce fatigue life. Because of the different experiences of design personnel, the performance of springs designed by different teams varies greatly. It can be seen that human design is not up to the task in exploring complex geometry, and the design scheme is limited to a narrow range.

[0113] At the same time, in the traditional spring design process, performance simulation is also a time-consuming and labor-intensive link. Whenever the geometry is modified, the mesh needs to be rebuilt and the finite element analysis needs to be run to evaluate stiffness, resonance frequency, stress, etc. If the electromagnetic driving force needs to be considered, the magnetic field simulation also needs to be coupled, and the process is complicated. The lack of automated tool chain leads to the complete reliance on manual driving, and each iteration may take several days to complete. This directly lengthens the design cycle and limits the number of schemes that can be tried. At the same time, different simulation conditions and targets (static stiffness, dynamic modal, fatigue life, etc.) are often calculated by different software or modules, and the compromise between the results needs to be manually weighed. Since spring design involves multi-objective optimization, such as increasing stiffness will increase stress, increasing elasticity will reduce stability, etc., manual adjustment is difficult to meet the optimization of multiple indicators at the same time, and often requires multiple rounds of trial and error to achieve a balance. In addition, the precision and complexity of the simulation model also bring challenges. For example, the thickness of the spring is only a few tens of microns, and too coarse meshing will affect the accuracy, and too fine meshing will increase the computational load; for example, the spring undergoes nonlinear large deformation, contact or eddy current loss after being stressed, which requires high-order simulation to solve. The engineering often simplifies the processing, but this may not be consistent with the actual situation, and the model needs to be verified and corrected through repeated trial production, further slowing down the design process. It can be seen that in the existing process where simulation analysis and design adjustment are disconnected, the efficiency is low and problems are easily missed, increasing the cost and time, which has become a bottleneck in the spring optimization process.

[0114] Due to the above reasons, the development of a new scheme of VCM spring sheet usually takes a long period of time and is heavily dependent on the experience of senior engineers. From conceptual design to finalization, it often needs to go through multiple rounds of cycles of conceptual scheme, simulation verification, sample trial production, test feedback and scheme revision. In this process, human decisions dominate, such as judging by experience which area of the spring sheet needs to be strengthened or thinned. The limitations of this experience-driven mode are: on the one hand, the human brain is difficult to globally optimize multiple parameters, so the design is often suboptimal rather than globally optimal; on the other hand, industry experience is mainly mastered by a few experts, and it is difficult to train new people and innovate. In addition, human design is easily affected by subjective preferences, and the schemes given by different engineers differ greatly, so it is difficult to standardize the design quality. Due to insufficient design capability and experience accumulation, it is necessary to catch up through repeated trial and error, and the long design period is obviously contradictory to the rapid iteration rhythm of consumer electronics, so the long development period has been difficult to meet the needs of the whole machine manufacturers. Therefore, how to reduce the dependence on individual experience, shorten the design period and improve the success rate of design at one time has become a problem to be solved in the prior art. In summary, the current VCM spring sheet design process has the problem of insufficient automation and intelligence, which seriously affects the improvement of product performance and development efficiency.

[0115] To solve the above technical problems, in one embodiment, taking the design of a VCM spring sheet as an example, a spring sheet structure generation method is provided, which is applied to a spring sheet automatic generation system. The spring sheet automatic generation system adopts a modular architecture, divides the full-automatic design process of the VCM spring sheet into a plurality of mutually decoupled sub-modules, forms a closed-loop process from parameter configuration, shape generation, simulation evaluation to optimization decision, and realizes automatic iterative optimization through clear interface interaction, sequential execution and data feedback of each module.

[0116] The overall architecture of the spring sheet automatic generation system includes: a geometric parameter configuration module for defining a design space and constraints; a control point sampling module for randomly generating initial structure parameters; a curve generation and contour construction module for generating a three-dimensional model of the spring sheet; a geometric validity verification module for ensuring the effectiveness of the structure; a finite element simulation automation module for performance simulation calculation; an optimization control module for updating the design according to the simulation results; and a parallel acceleration and data output module for improving the calculation efficiency and exporting the design results. The above modules are designed in a loose coupling manner, which can be independently debugged and are closely linked in the process: the optimization control module drives the remaining modules to work in sequence, and finally realizes the automatic generation and optimization of the spring sheet structure. Data interfaces are used for transmission between modules, such as geometric parameter objects, control point sets, curve coordinates, meshes and performance indicators, to ensure smooth and consistent information flow.

[0117] The shell structure generation method can include: determining design boundary conditions by a parameter configuration module, then a sampling module generates candidate designs (control point set), a curve / profile module constructs shell geometry, a geometry validity verification module screens, a simulation module calculates mechanical performance, and finally an optimization control module adjusts the next batch of design schemes accordingly, and iterates until the convergence condition is met. The overall process of the automatic shell generation system realizes fully automatic closed-loop optimization, and automatically from design input to optimization output without human intervention. The architecture design focuses on module decoupling and exception control, for example, the simulation module sets up error handling inside to ensure that the system can automatically retry or skip when a single design simulation fails without affecting the overall iteration process; the optimization control module checks the validity of the received data to ensure that the output of each link meets the interface requirements, so that the system has high robustness and stability.

[0118] Further, the geometry parameter configuration module (GeometryParameters) is responsible for defining the overall parameters and constraint conditions of the shell design, and provides unified design input for other modules. This module reads the design requirements provided by the user as needed, such as the spatial range of the shell (design area polygon contour), fixed starting point position, material thickness, tolerance range, and geometric / physical constraints, etc., and encapsulates it as a global parameter object for subsequent use. By centralized configuration, it ensures that each sub-module works under unified constraints and avoids parameter inconsistency.

[0119] The input parameters of the geometry parameter configuration module can include: the polygon area of the design space (such as the installation boundary of the camera module shell), the starting control point or fixed end position, material and size parameters (thickness, width upper and lower limits, etc.), and target performance index threshold (stiffness requirement in each direction, stress upper limit, etc.).

[0120] The output parameters of the geometry parameter configuration module can include: a parameter data structure containing the above design settings, such as the design area vertex set, the initial control point coordinates, the curvature lower limit value κ min , the control point spacing limit d min , the boundary distance safety distance ε, etc. By providing an output interface, other modules can query the required parameters.

[0121] The geometry parameter configuration module includes a plurality of parameter acquisition interfaces, such as an initial space acquisition get_design_domain(), a control point acquisition get_start_point(), a thickness acquisition get_thickness(), and the like, which are used to call the design boundary and initial conditions by the sampling and modeling modules. The module first reads and verifies the integrity and validity of the input parameters (such as checking the design area polygon closure and reasonable parameter range), and then generates a global parameter object at system initialization and broadcasts it to each module. Once initialized, the parameters remain unchanged or are only adjusted by the user throughout the optimization process. By concentrating the design conditions in the geometry parameter configuration module, the decoupling of parameter configuration and algorithm logic can be achieved, which facilitates subsequent maintenance and modification of design requirements without changing the core algorithm code.

[0122] Further, the control point sampling module (PoissonPolygonSampler) can be a Poisson disc sampling algorithm that randomly generates a control point set in the specified polygon design area to serve as the initial control vertices of the splint central axis curve. By random sampling, the initial topology diversity is ensured, while strictly meeting the spatial distribution constraints, thereby avoiding illegal initial geometry from the source. The Poisson disc sampling feature is to generate uniformly distributed points in the region, maintaining a certain minimum distance between any two points, thereby providing a good point set basis for subsequent curve generation.

[0123] The input parameters of the control point sampling module can include a design area polygon (which can be provided by the geometry parameter configuration module) and sampling parameters such as the number of control points or the desired density, the minimum point spacing d, and the minimum distance ε from the boundary constraint threshold. The module can also accept existing anchor points (such as a fixed starting point) as part of the input to ensure that the key points are included in the sampling results.

[0124] The output parameters of the control point sampling module can include a list of control point coordinates that meet the constraints, typically in the form of two-dimensional coordinates in the design plane. The output point set will be used for subsequent B-spline curve fitting. For the case of including a fixed starting point, the output will ensure that the point is included and that the other points are uniformly distributed relative to it.

[0125] The control point sampling module generates n points in O(n log n) complexity by efficient Poisson sampling methods such as Bridson algorithm. The execution process includes: randomly selecting a seed point inside the design polygon, then iteratively generating new points to ensure that each new point is at least d distance away from existing points and at least ε distance away from the polygon boundary. The sampling terminates once a certain number of attempts fail to find a qualified new point. Throughout the process, the module checks the boundary constraint and distance constraint in real time: using the inclusion test of points in a polygon to ensure that the points fall within the boundary and leave a safety margin ε, and using spatial indexing to quickly determine whether the nearest distance between the new point and existing points meets the d requirement. After sampling is complete, the initial control point set is output. By providing interfaces such as sample_points(polygon, n, d, ε) for generating point sets, the module can be called multiple times to obtain different random distribution schemes. During optimization, the module can also be used to generate new points in the mutation phase: for example, when the optimization control module attempts to adjust the control points, the module can be called to resample the local area to replace some invalid points, thereby ensuring that the scheme after mutation still meets the constraints. Through Poisson sampling, the system avoids topological limitations in the initial design stage, allowing the generation of control point layouts with arbitrary distribution, laying the foundation for exploring diverse shapes.

[0126] Further, the curve generation and contour construction module (SplineBuilder, StripBuilder) can include a spline curve generation module SplineBuilder and a solid contour construction module StripBuilder: the former can generate a smooth splinter medial axis curve according to the control point set, and the latter can construct a three-dimensional splinter solid contour with thickness and width along the medial axis curve. This module converts discrete control points into continuous geometry and forms a solid model that can be used for finite element analysis.

[0127] Among them, the spline curve generation module SplineBuilder can use a cubic B-spline algorithm to fit control points to generate a parametric curve of the splinter medial axis. The SplineBuilder module first calculates the node vector of the spline curve according to the number and distribution of control points to ensure smooth transition of the curve segments (satisfy certain G 2 Continuity). Then, by calling the B-spline formula, an initial medial axis curve is generated, and an interface is provided to regenerate the curve after adjusting the control points to optimize the shape. The function interface of this module, such as the build_spline(points) function for building a spline curve, can obtain the parametric equation or discrete point list representing the medial axis curve by inputting the control point list. For example, by inputting a set of control point coordinates (provided by the sampling module or the optimization control module), the output splinter center curve can be discretized into multiple line segments for subsequent modeling. SplineBuilder can ensure that the curve meets the curvature smoothness requirement by calculating the second derivative of the curve to verify that the local curvature is not lower than the threshold κ.min If there is an acute corner with too small curvature, it can be corrected by adding a control point or adjusting the position of an existing control point to ensure the smoothness of the medial axis curve. In addition, SplineBuilder supports fine-tuning of the curve as needed: if it is found during optimization that the curvature or shape of a certain segment needs to be improved, the adjustment interface provided by the module can be called to update the local control points, and then the spline is regenerated to achieve the purpose of optimizing the curve shape.

[0128] After obtaining the medial axis curve, the solid profile construction module StripBuilder can extend the medial axis curve to a three-dimensional strip solid model with thickness and width. Among them, the parameters including the parametric description of the medial axis curve (such as the discrete point column), the preset strip thickness value, and the width distribution function or parameter that may vary along the curve can be input to obtain the output result of the strip three-dimensional solid geometric model. Exemplarily, the strip three-dimensional solid geometric model can adopt the representation form of CAD surface or solid, etc., and can export IGES format file. Exemplarily, the solid profile construction module StripBuilder can offset the curve in the normal plane direction of the medial axis curve to generate the upper and lower boundary curves of the strip, and then stretch the boundary curves into upper and lower surfaces according to the set thickness, and finally generate a closed thickness solid. Further, to improve the design flexibility, StripBuilder can also support dynamic adjustment of the width parameterization, that is, by controlling the function of the offset distance in the normal direction along the curve to change the local width of the strip.

[0129] Further, in the early stage of optimization, a large amplitude of width disturbance (such as Gaussian random disturbance, but truncated to the range of [min_width, max_width]) can be applied to extensively explore possible cross-section changes; in the later stage, the disturbance amplitude is reduced to fine-tune the width, achieving more fine shape optimization. The position calculation of the width and the medial axis curve adopts a coupling mechanism, that is, each width change is recalculated by offsetting in the normal direction of the curve, which can ensure that the transition of the strip edge is smooth and has no sudden change. After generating the solid model, StripBuilder calls the geometry verification module to check the solid, such as whether the upper and lower surfaces are self-intersecting or whether they exceed the design boundary, etc. Only the model that passes the verification enters the next step of simulation. The curve generation and profile construction module provides an interface such as the build_strip (curve, thickness, width_func) function to build a spline curve, which internally calls the SplineBuilder result and superimposes the thickness and width parameters to generate the final CAD model. Through SplineBuilder and StripBuilder, the system can convert discrete design variables into continuous solids, preparing for simulation analysis, and ensuring that the model meets the geometric design requirements.

[0130] The geometric validity verification module is throughout the design generation process, which is used to check the generated control point set, curve and entity model for a series of strict constraint checks, so as to ensure that each candidate control parameter, i.e. candidate design, meets the pre-set geometric constraint conditions before entering the finite element simulation. The embodiment adopts a "strong constraint" verification strategy, that is, any design that does not meet the constraint will be immediately identified and rejected or corrected, avoiding sending unfeasible designs into simulation to cause calculation interruption, so as to ensure that the entire optimization process is always carried out in the effective feasible domain.

[0131] The checking content of the geometric validity verification module includes boundary constraint, self-intersection constraint, curvature constraint and spacing constraint, etc., wherein:

[0132] Boundary constraint: used to confirm that all control points and generated curves are within the design polygon range, and the curve and the design region boundary keep a safe distance of not less than ε. The polygon inclusion test and the nearest distance calculation can be used to ensure that the curve is completely within the allowed range.

[0133] Self-intersection constraint: used to check whether the axis curve of the shell is a simple curve, that is, there is no self-intersection except the two fixed end points. The line segment pairwise intersection detection or the ray detection method can be used to judge whether the curve is self-intersected. If the curve is found to be self-intersected, the module will mark the design as illegal.

[0134] Curvature constraint: used to calculate the curvature distribution of the curve through differential geometry, and verify whether the curvature radius is greater than the specified lower limit κ min at each place to avoid too sharp bending. For any small segment of the curve, if the curvature violates the requirement, it is determined to be illegal.

[0135] Spacing constraint: for discrete control points, further check whether the distance between non-adjacent points is not less than the set minimum value d min , so as to prevent the control points from being too dense to cause local jitter of the spline or local narrowness of the entity. Spatial index is used to accelerate the distance calculation between points and improve the detection efficiency.

[0136] The geometric validity verification module can be invoked immediately after each generation / modification of the curve to review the current design. For example, the module can be an interface through a validity verification function is_valid(design) that performs all the checks described above at once, returning a Boolean result or specific violation information. If it returns as invalid, the optimization control module will discard the design and trigger an alternative (e.g. resample control points or adjust parameters). In a parallel simulation scenario, the verification module can also be embedded in the simulation subprocess to prevent some designs from crashing midway due to the "cliff effect" that suddenly does not satisfy the constraints. This strategy of strong constraint verification is different from traditional soft constraint handling (such as the penalty function method allows violations but gives penalties). By directly rejecting infeasible solutions and resampling new solutions, it fundamentally avoids optimization stagnation caused by infeasible designs that cannot calculate the objective function, achieving precise control of the feasible region and providing a reliable foundation for subsequent optimization.

[0137] The finite element simulation automation module integrates mesh generation, physical modeling, and solution analysis functions, and can automatically complete the mechanical performance calculation of the shell structure using the Python interface of existing software. The finite element simulation automation module can include a mesh generation unit MeshBuilder and a modeling and solution unit ModelAnalysis, which can programmatically control independent solvers through open-source tools, thereby automatically evaluating the stiffness and stress simulation of each candidate design, and compressing the time-consuming step of manual simulation to seconds.

[0138] The mesh generation unit MeshBuilder is responsible for discretizing the shell geometry model into a mesh model required for finite element analysis. The input parameters of the mesh generation module can include the shell three-dimensional model output by the StripBuilder, material properties (such as Young's modulus and Poisson's ratio of the shell material), and mesh partitioning settings (element type, size, etc.), thereby obtaining a finite element mesh model (which can include node and element information) corresponding to the shell three-dimensional model. For example, a mesh can be generated through the command interface provided by an open-source tool, such as creating a geometric entity by calling a command (e.g. reconstructing the shell shape by connecting key points, stretching patches, etc.), then specifying a shell element or solid element type, and dividing the mesh to obtain a mesh model. In this embodiment, to improve efficiency, a simplified shell element model can be generated in the optimization coarse screening stage, and a solid element model can be switched to in the fine evaluation stage. The mesh generation unit MeshBuilder can provide interfaces such as the mesh model function mesh_model(geometry, element_type) to select the element type as needed and automatically complete the division. During the division process, the module can also control the mesh quality, such as checking the element distortion rate to ensure the solution accuracy.

[0139] ModelAnalysis can be a unit that sets boundary conditions, loads working conditions and solves the mechanical response of the spring after obtaining the mesh. For example, ModelAnalysis can solve the performance indicators of the spring design, including stiffness values in each direction and maximum stress values, according to the mesh model provided by the mesh generation unit, loading and constraint conditions (such as fixed support position, applied electromagnetic force or displacement amount), and the type of results to be extracted. ModelAnalysis can be controlled by an open source tool to control an independent solver, specify material properties to the mesh model, apply boundary conditions (for example, fix the mounting frame area of the spring to simulate the actual assembly constraints), apply loads or displacements to simulate the action of the VCM on the spring (such as applying displacement in the optical axis direction to calculate Z-direction stiffness, and applying lateral force in X / Y direction to calculate lateral stiffness), thereby solving the statics problem, reading the calculation results after the independent solver completes the calculation, including the equivalent stiffness (force / displacement) of the spring in each loading direction and the stress distribution cloud diagram, and extracting the maximum stress value. Through the open source tool, the result data is obtained without manual intervention into the solving interface. ModelAnalysis can encapsulate a series of preset solver software commands through the analysis function run_analysis (mesh, load_case). For each spring design, ModelAnalysis can store key results in data structures available to the optimization control module, such as X / Y / Z three-direction stiffness and stress values. In order to shorten the simulation time and meet the optimization requirements, ModelAnalysis can also run multiple simulation instances simultaneously under the scheduling of the parallel acceleration module, and after the solution is completed, the results are transmitted back to the optimization control module through shared memory or temporary files, realizing the full automation and seamless integration of the simulation process, and connecting the modeling, partitioning, loading, solving and post-processing links to be automatically executed by the program, thereby avoiding manual repeated operations, eliminating file import / export waiting, and reducing the time-consuming of structure evaluation.

[0140] Optimizer can be an evolution strategy algorithm (Evolution Strategy) for global optimization of spring structure design. Optimizer can continuously adjust design parameters according to the performance feedback provided by the simulation module, and iteratively search for the optimal solution that meets the multi-objective performance requirements. Unlike traditional local optimization methods that rely on manual experience adjustment, Optimizer can achieve constraint-driven global optimization and independently evolve new structure topologies and shapes in high-dimensional design space.

[0141] For example, the optimization control module can encode the design freedom of the shell geometry as a set of optimization variables, including the coordinate positions of the control points and possibly the width distribution parameter along the length of the shell, etc. Further, the variables can be represented in real numbers. The initial population is generated by the aforementioned Poisson sampling method or the like to generate multiple sets of random candidate solutions, ensuring that the initial solution set has diversity and feasibility. Each candidate solution, i.e., candidate control parameters, corresponds to a set of control points, which generates a shell solid model via the spline generation module SplineBuilder and the solid profile construction module StripBuilder.

[0142] Further, the optimization control module can also implement control in the overall optimization iteration by maintaining a specific population list, which can include all design individuals and their performance evaluations of the current iteration.

[0143] The optimization control module involves multiple objectives and multiple constraints, i.e., the X / Y direction stiffness ≥ preset threshold, the Z direction stiffness in a preset range, the maximum stress in each direction ≤ material tolerance, and the geometric non-intersection, curvature smoothing, etc. In this embodiment, the optimization control module converts the above constraints into a comprehensive objective function or fitness function. By setting a penalty term for the case of non-compliance of stiffness or stress exceeding the limit, it is included in the objective function evaluation, so that the algorithm automatically tends to meet the constraints when searching.

[0144] In addition, since the geometric constraints have been verified by the legality verification module, the infeasible design will not enter the target evaluation stage. In this embodiment, by using the weight summation strategy to unify multiple objectives into a single fitness value, or by alternating optimization of different objectives during the evolution process, the automatic balanced optimization of stiffness, strength and other performances can be achieved. For example, the objective function f(P) can be defined as f(P) = - (a · k Z (P) + β · k XY (P) - γ · max(σ(P) - σ max , 0)...), where a, β, γ are weights, P is a parameter set, k Z (P) represents the overall stiffness of the structure in the Z direction, k XY (P) represents the overall stiffness of the structure in the XY plane, σ(P) represents the stress in the structure, and σ max is the maximum allowable stress of the material, so as to consider the requirements of improving stiffness and reducing over-limit stress. By adjusting the weights or the form of the objective function according to the design requirements, different design preferences can be achieved.

[0145] The optimization control module can be a (μ, λ) evolutionary strategy that generates λ offspring candidate designs from μ parents in each generation. Illustratively, the evolutionary strategy and iteration can include selection, mutation, recombination, and survivor selection. In the selection, a number of elite individuals (e.g., μ) can be selected from a current population according to fitness levels for reproduction, random mutation can be applied to the design parameters of the elite individuals to generate new candidate designs. The mutation can be a small random perturbation (a random vector following a normal distribution) added to the control point coordinates, with a perturbation amplitude σ that is large in the early stages of the algorithm to encourage extensive exploration and gradually reduced in the later stages to fine search. Further, the optimization control module can also employ a mutation rejection mechanism, i.e., if a certain mutation is determined by the geometric legality verification module to no longer satisfy the geometric constraints, the mutation is rejected and resampled to ensure that 100% of the offspring are legal. In addition, the optimization control module can also employ a weighted recombination mechanism, i.e., the control point position information of multiple excellent individuals can be combined to generate a new solution to integrate the advantages of different solutions and improve population diversity and excellent gene inheritance. Each new individual is immediately evaluated for performance by the simulation module. When all λ candidate solutions in a generation are evaluated, the optimization control module selects the μ individuals with the best performance from the total set of parents and offspring to enter the next generation based on a survivor retention strategy, and eliminates inferior designs. This cycle is iterated to continuously approach the optimal solution. The entire evolutionary process continues until a termination condition is met, such as reaching a preset upper limit on the number of generations or the population fitness improvement being less than a preset condition.

[0146] In the present embodiment, the optimization control module can start the optimization function run_optimization() through a unified scheduling interface to control the entire process to start. The optimization control module can periodically call the interfaces of each module: the sampling and modeling modules are called when generating an initial population, the simulation module is called in batches to obtain results when evaluating individuals, and the sampling module or an in-built function of the optimization control module is called to generate perturbations when applying mutation operations. During the optimization process, the optimization control module monitors the satisfaction of constraints and performance indicators in real time and records the best design and fitness value of each generation. To facilitate analysis and tracing, the optimization control module can automatically save the input parameters and output results of each iteration to the population list. Through interface function calls rather than coupled code, the interaction with independent solvers is also encapsulated through an abstraction layer (e.g., a unified design evaluation function evaluate_design (design) interface), so that the underlying algorithm or simulation tool can be easily replaced without affecting the overall process. By combining evolutionary algorithms with engineering constraints, the optimization control module can automatically explore VCM slug designs, and compared to traditional manual parameter tuning methods, the global optimal or near-optimal solution can be found to avoid falling into a local suboptimal trap.

[0147] A parallel acceleration and data output module is used to improve the simulation calculation efficiency and arrange the final design data. The multi-core parallel calculation can greatly shorten the optimization iteration time, and the design results can be provided to engineering applications and displays. The parallel acceleration unit provides high-speed simulation support in the optimization evaluation stage, and the data output unit provides result saving and visualization functions at the end or intermediate stage of optimization.

[0148] Further, multi-core parallel calculation can be used for the case of a large number of candidate designs to be evaluated in each generation of optimization. Open source tools are used to realize multi-process parallel calculation. For example, at the beginning of optimization, the optimization control module sends the design individual set to be evaluated to the parallel calculation task queue, and the parallel calculation task queue automatically schedules multiple independent solver calculation kernel instances (the kernel instance can correspond to the CPU core number, such as 16 cores) to undertake different design simulation tasks. The module ensures balanced task allocation to avoid overloading of some processes. Each independent solver instance concurrently performs modeling, meshing, and solving processes, and uses shared memory or files to return the results to the master. When all parallel tasks are completed, the optimization control module aggregates the results for the next evolution calculation. Through the multi-core parallel calculation mechanism, the simulation calculation bottleneck can be greatly alleviated, and the overall optimization time can be significantly reduced.

[0149] Further, to ensure smooth parallel process, a geometry verification module can be used for pre-checking before task distribution, so as to filter out illegal designs and avoid wasting computing resources. At the same time, constraint monitoring can be embedded in each process. If an abnormality occurs in the simulation of a certain fragment design, such as material non-convergence or geometric deformation, an error handling mechanism can be used to capture and notify the main process to retry or skip. Further, flexible core number configuration and automatic process management can be used to make the simulation process have good scalability. In one specific embodiment, in a 16-core parallel environment, the actual speedup ratio can reach about 12 times, and the simulation time can be reduced from 15-30 seconds in series to 1-3 seconds. Further, the embodiment uses a hierarchical simulation strategy of coarse screening shell unit and fine adjustment entity unit, and can complete the simulation workload equivalent to 14 days in the traditional technology within 2-3 hours, fully exploiting the performance potential of multi-core hardware, and enabling the global optimization algorithm to complete high-intensity calculation tasks within an acceptable time.

[0150] After the optimization process is finished, the data output unit will organize and export the final optimized design and process data for engineering application and decision reference. The data output unit can call the spline generation module StripBuilder to regenerate the fine CAD model of the final design and provide IGES file export function to facilitate the import of the design shape into CAD / CAM system for detailed design and manufacturing. Further, the results of the simulation module (such as stiffness values in all directions, stress distribution) can be synchronized and output.

[0151] Further, for the performance simulation results, key performance indicators can be generated in reports or tables, and visualization graphs of the simulation results can be exported, such as shell stress contour maps, deformation morphological diagrams, etc.

[0152] In addition, the data output module can also organize historical data during the optimization process, including the parameters and values of each generation of population, so as to build a traceable optimization history database, thereby providing data backup for future analysis of optimization convergence trend or improvement of the algorithm. The data output module can be called by the export_results (design) and save_history (log) interfaces when the optimization control module issues an end instruction.

[0153] Further, the data output module can also attach corresponding meta information to the output data, such as coordinate unit, material assumption, constraint condition, etc., to facilitate user understanding and reproduction of the design, so that the design process data is presented in an engineering-friendly form.

[0154] A shrapnel structure generation method provided in this embodiment breaks through the limitations of traditional reliance on a fixed initial topological structure for local adjustment through global design space exploration and topological innovation. Poisson random sampling combined with B-spline parameterization can achieve exploration of the global design space. Compared with the existing technology that can only fine-tune near the initial structure, the method of this embodiment can freely generate new shrapnel shapes within a given design domain, including topological forms that were previously difficult to imagine manually (for example, evolving an "Ω"-shaped new structure from a traditional wave shape). This broadens the design freedom and covers feasible areas that traditional methods cannot reach, thereby significantly improving the diversity of design schemes and helping to discover structures with better performance. Through a strong constraint handling mechanism, including Poisson sampling and a mutation rejection strategy, all candidate designs are guaranteed to satisfy hard constraints from the outset. Unlike the penalty functions or constraint violation penalty methods commonly used in existing algorithms (which require assigning virtual inferiority scores to solutions that fail to satisfy the constraints and may result in failed individuals that cannot evaluate the objective function during the optimization process), the algorithm directly eliminates illegal solutions and resamples them, eliminating infeasible solutions from participating in the evolution. This improves the reliability and efficiency of the optimization process when handling geometric discrete constraints (such as "cliff effect" constraints such as self-intersections and boundary violations). There are no simulation interruptions due to constraint violations throughout the optimization run, ensuring the continuity of the algorithm's convergence and the credibility of the results. Furthermore, because the algorithm always searches within the feasible region, it avoids wasting computing power on infeasible directions, significantly improving effective search efficiency. Through multi-objective collaborative optimization and automatic balancing, a weighted recombination evolutionary strategy is used to achieve simultaneous optimization of multiple objectives. Unlike existing technologies that often rely on manual weight adjustment or separate optimization followed by manual compromise, by embedding a multi-objective balancing mechanism in the evolutionary algorithm and setting reasonable objective weights and recombination operators, the algorithm can simultaneously consider requirements such as improving X / Y stiffness, controlling Z stiffness within a certain range, and reducing maximum stress in a single run, thus achieving automatic trade-offs between various objectives. For example, the optimization process introduces weighted recombination of elite individuals to generate new solutions, allowing design genes that excel in both stiffness and stress to be combined to produce offspring that balance both. Compared to the traditional method of item-by-item optimization and reconciliation, the collaborative evolutionary strategy of this embodiment can find a design solution that simultaneously meets all key performance requirements, significantly improving the comprehensiveness of design performance. It can also solve the problem of traditional design thinking that blindly improving stiffness leads to excessive stress.By introducing the double acceleration strategy of hierarchical solution and multi-core parallel in the simulation link, a two-level simulation scheme of shell element coarse screening and solid element fine adjustment is adopted, that is, the stiffness / stress performance of a large number of candidate designs is approximately evaluated by using the calculation of the fast shell element, and the inferior scheme is eliminated, and then the selected design is switched to the high-precision solid element simulation verification, which is different from the existing single-precision simulation, which can greatly reduce the total calculation amount, while ensuring the accuracy of the final result, and at the hardware level, the parallel interface of the open source tool is used to realize multi-thread and multi-process parallel simulation: multiple independent solver instances are run simultaneously to evaluate different individuals in the population, which can effectively improve the simulation efficiency on multi-core CPU compared with the traditional serial simulation one by one, and the accuracy is not sacrificed, but the reliability of the result is ensured through fine simulation verification, thereby greatly improving the timeliness and practical value of the system. Through end-to-end automatic integration, each link of the fragment design (including parameter modeling, geometry generation, simulation evaluation, result processing, etc.) is seamlessly integrated in the same automatic platform, which can be compared with the discrete process in the traditional technology, such as manual sketching, manual importing into simulation software, and multiple manual adjustments. The CAD and CAE links are connected by the open source tool in the embodiment, eliminating manual intervention and file interaction delay, and the simulation input and output are all processed by the program, without the need for manual waiting, so that the overall design process is digitized and continuous, not only reducing the dependence on human experience and reducing human errors, but also allowing the design iteration to run automatically without interruption, greatly improving the research and development efficiency. At the same time, the abnormal processing mechanism of the embodiment, such as automatic retry of simulation failure and automatic saving of data, can ensure the robustness of long-time unattended operation, thereby realizing the intelligentization of the whole process of VCM fragment from design to verification, and achieving the double improvement of design efficiency and quality.

[0155] It should be understood that, although each step in the flowchart involved in each embodiment as described above is displayed in sequence according to the direction of the arrow, these steps are not necessarily executed in sequence according to the direction of the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other orders. Moreover, at least part of the steps in the flowchart involved in each embodiment as described above can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be alternately or alternately executed with at least part of other steps or steps or stages in other steps.

[0156] Based on the same inventive concept, the embodiments of the present application also provide a shell structure generation device for implementing the shell structure generation method described above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more shell structure generation device embodiments provided below can refer to the limitations of the shell structure generation method described above, which will not be repeated here.

[0157] In one embodiment, as shown in Figure 3 a shell structure generation device is provided, the device comprising:

[0158] A parameter generation module 100 is configured to generate initial structure control parameters of a shell based on preset spatial constraint parameters.

[0159] A model generation module 200 is configured to fit based on the initial structure control parameters to obtain a shell three-dimensional model. The shell three-dimensional model is a shell three-dimensional model that meets the preset geometric verification rule.

[0160] A performance simulation module 300 is configured to perform performance simulation on the shell three-dimensional model to obtain a performance simulation result.

[0161] A parameter optimization module 400 is configured to perform iterative optimization on the initial structure control parameters based on the performance simulation result to obtain target structure control parameters and a target shell three-dimensional model corresponding to the target structure control parameters.

[0162] In one embodiment, the initial structure control parameters include a control point set; the parameter generation module 100 is further configured to:

[0163] generate a plurality of seed control points that meet the preset spatial constraint parameters within a preset simulation space;

[0164] iteratively generate a plurality of candidate control points based on the plurality of seed control points;

[0165] determine the control point set based on the candidate control points that meet the preset spatial constraint parameters.

[0166] In one embodiment, the model generation module 200 is further configured to:

[0167] perform curve fitting on the initial structure control parameters based on a B-spline curve fitting algorithm to obtain a shell median axis curve that meets a preset curvature condition;

[0168] perform normal plane offsetting and thickness stretching processing on the shell median axis curve to obtain the shell three-dimensional model.

[0169] In one of the embodiments, the preset geometric verification rule comprises one or more of a boundary constraint verification, a self-intersection constraint verification, a curvature constraint verification, and a spacing constraint verification.

[0170] The boundary constraint verification comprises verifying, based on a preset polygonal region and a preset safety distance, whether the shell three-dimensional model is located within the preset simulation space and maintains a safety distance from the boundary of the preset simulation space; the self-intersection constraint verification comprises detecting, based on a preset self-intersection judgment algorithm, whether the medial axis curve has self-intersections other than endpoints; the curvature constraint verification comprises verifying, based on discrete points of the curve, whether the curvature radius of the medial axis curve is not less than a preset lower limit of curvature; and the spacing constraint verification comprises verifying, based on coordinates of the control points, whether the distance between any two non-adjacent control points is not less than a preset minimum spacing.

[0171] In one of the embodiments, the performance simulation module 300 is further configured to:

[0172] perform mesh division on the shell three-dimensional model based on material parameters of the shell three-dimensional model and preset mesh parameters, to obtain a mesh model of the shell three-dimensional model;

[0173] solve the stiffness values and the maximum stress values of the shell in multiple directions based on the mesh model applied with loads or displacements, to obtain performance simulation results.

[0174] In one of the embodiments, the number of the initial structure control parameters is multiple; and the parameter optimization module 400 is further configured to:

[0175] based on the performance simulation results of the shell three-dimensional models corresponding to the multiple initial structure control parameters and a preset performance evaluation function, take the initial structure control parameter corresponding to the shell three-dimensional model with the optimal simulation result as the current optimal control parameter;

[0176] based on a weighted recombination operator, perform cross operation on the current optimal control parameter to generate multiple candidate control parameters as initial structure control parameters for the next round of optimization process.

[0177] In one of the embodiments, the parameter optimization module 400 is further configured to:

[0178] based on each of the initial structure control parameters, distribute the initial structure control parameter to a parallel computing task queue;

[0179] based on multiple independent solver instances, synchronously perform performance simulation on each of the shell three-dimensional models to obtain performance simulation results.

[0180] Each module in the above-mentioned spring-flake structure generation device can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in the computer device in the form of software, so that the processor can call and execute the corresponding operations of each module.

[0181] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, memory, a communication interface, a display screen, and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal via wired or wireless communication. The wireless communication can be achieved via Wi-Fi, a mobile cellular network, NFC (near-field communication), or other technologies. When executed by the processor, the computer program implements a method for generating a spring structure. The display screen of the computer device can be a liquid crystal display or an electronic ink display. The input device of the computer device can be a touch layer covering the display screen, or keys, a trackball, or a touchpad provided on the computer device housing, or an external keyboard, touchpad, or mouse.

[0182] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0183] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the method for generating a spring structure according to any of the above embodiments is implemented:

[0184] Generate the initial structural control parameters of the shrapnel based on the preset spatial constraint parameters;

[0185] Fitting is performed based on the initial structural control parameters to obtain a three-dimensional model of the shrapnel; the three-dimensional model of the shrapnel is a three-dimensional model of the shrapnel that complies with preset geometric verification rules;

[0186] Performing performance simulation on the three-dimensional model of the shrapnel to obtain performance simulation results;

[0187] Based on the performance simulation result, the initial structure control parameter is iteratively optimized to obtain a target structure control parameter and a target shell three-dimensional model corresponding to the target structure control parameter.

[0188] In one embodiment, a computer readable storage medium is provided, and the computer readable storage medium stores a computer program. The computer program is executed by a processor to implement the shell structure generation method of any one of the above embodiments.

[0189] Based on the preset spatial constraint parameter, an initial structure control parameter of the shell is generated;

[0190] Based on the initial structure control parameter, fitting is performed to obtain a shell three-dimensional model; the shell three-dimensional model is a shell three-dimensional model that meets a preset geometric verification rule;

[0191] Performance simulation is performed on the shell three-dimensional model to obtain a performance simulation result;

[0192] Based on the performance simulation result, the initial structure control parameter is iteratively optimized to obtain a target structure control parameter and a target shell three-dimensional model corresponding to the target structure control parameter.

[0193] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties.

[0194] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0195] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist, it should be considered as the scope of the present application.

[0196] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method for generating a spring structure, characterized in that: The shrapnel structure generation method comprises: Generate the initial structural control parameters of the shrapnel based on the preset spatial constraint parameters; Fitting is performed based on the initial structural control parameters to obtain a three-dimensional model of the shrapnel; the three-dimensional model of the shrapnel is a three-dimensional model of the shrapnel that complies with preset geometric verification rules; Performing performance simulation on the three-dimensional model of the shrapnel to obtain performance simulation results; Based on the performance simulation results, the initial structural control parameters are iteratively optimized to obtain target structural control parameters and a target shrapnel three-dimensional model corresponding to the target structural control parameters.

2. The method for generating a spring structure according to claim 1, wherein: The initial structural control parameters include a set of control points; the initial structural control parameters for generating the spring fragment based on the preset spatial constraint parameters include: In a preset simulation space, generating a plurality of seed control points that meet the preset space constraint parameters; Iteratively generating multiple candidate control points based on the multiple seed control points; The control point set is determined based on the candidate control points that meet the preset spatial constraint parameters.

3. The method for generating a spring structure according to claim 1, wherein: The fitting based on the initial structure control parameters to obtain the three-dimensional model of the shrapnel includes: Based on the B-spline curve fitting algorithm, the initial structural control parameters are subjected to curve fitting to obtain the shrapnel central axis curve that meets the preset curvature conditions; The center axis curve of the spring piece is subjected to normal plane offset and thickness stretching processing to obtain a three-dimensional model of the spring piece.

4. The method for generating a spring structure according to claim 1, wherein: The preset geometric validation rules include one or more of boundary constraint validation, self-intersection constraint validation, curvature constraint validation, and spacing constraint validation; Among them, the boundary constraint verification includes verifying whether the shrapnel three-dimensional model is located in the preset simulation space and maintains a safe distance from the boundary of the preset simulation space based on a preset polygonal area and a preset safety distance; the self-intersection constraint verification includes detecting whether the medial axis curve has self-intersections other than the endpoints based on a preset self-intersection judgment algorithm; the curvature constraint verification includes verifying whether the curvature radius of the medial axis curve is not less than the preset curvature lower limit based on the discrete points of the curve; the spacing constraint verification includes verifying whether the distance between any two non-adjacent control points is not less than the preset minimum spacing based on the control point coordinates.

5. The method for generating a spring structure according to claim 1, wherein: The performance simulation of the shrapnel three-dimensional model includes: Based on the material parameters of the shrapnel three-dimensional model and the preset grid parameters, the shrapnel three-dimensional model is meshed to obtain a mesh model of the shrapnel three-dimensional model; Based on the grid model, load or displacement is applied to solve the stiffness value and maximum stress value of the spring in multiple directions to obtain performance simulation results.

6. The method for generating a spring structure according to claim 1, wherein: There are multiple initial structural control parameters; and the iterative optimization of the initial structural control parameters based on the performance simulation results includes: Based on the performance simulation results of the three-dimensional spring fragment models corresponding to the multiple initial structural control parameters and a preset performance evaluation function, the initial structural control parameters corresponding to the three-dimensional spring fragment model with the best simulation results are used as the current optimal control parameters; Based on the weighted recombination operator, a cross operation is performed on the current optimal control parameters to generate multiple candidate control parameters as initial structural control parameters for the next round of optimization process.

7. The method for generating a spring structure according to claim 6, wherein: The performance simulation of the three-dimensional model of the shrapnel is performed to obtain the performance simulation result, which includes: Based on each of the initial structure control parameters, allocating the initial structure control parameters to a parallel computing task queue; Based on multiple independent solver instances, performance simulation is performed on each of the spring fragment three-dimensional models simultaneously to obtain performance simulation results.

8. A device for generating a spring structure, characterized in that: The device comprises: A parameter generation module, used to generate initial structural control parameters of the shrapnel based on preset space constraint parameters; A model generation module is used to perform fitting based on the initial structural control parameters to obtain a three-dimensional model of the shrapnel; the three-dimensional model of the shrapnel is a three-dimensional model of the shrapnel that complies with preset geometric verification rules; A performance simulation module is used to perform performance simulation on the three-dimensional model of the shrapnel to obtain performance simulation results; A parameter optimization module is used to iteratively optimize the initial structural control parameters based on the performance simulation results to obtain target structural control parameters and a target shrapnel three-dimensional model corresponding to the target structural control parameters.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images