A Collaborative Optimization Method for the Cross-Section of Large Heavy-Duty Rolling Guide Pairs Based on an Improved Multi-Objective Genetic Algorithm
By improving the multi-objective genetic algorithm and the circular arc modified line roller profile model, the problem of stress concentration at the roller end and static stiffness contradiction in the design of large heavy-duty rolling guide pairs was solved. The collaborative optimization of roller and raceway was realized, and multiple feasible parameter combinations were provided, which improved the performance and life of the guide pair.
Patent Information
- Application Number
- CN202611357668.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-03
- Publication Date
- 2026-11-10
Smart Images

Figure CN123069904A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a collaborative optimization method for the cross-section of large heavy-duty rolling guide pairs based on an improved multi-objective genetic algorithm, belonging to the technical field of rolling guide pair structural optimization design. Background Technology
[0002] Large, heavy-duty rolling guideways are widely used in heavy-duty CNC machine tools, gantry machining centers, precision measuring equipment, and other high-load linear motion equipment. These guideways typically need to withstand large vertical loads, cutting impact loads, and overturning moments within a small installation space. Their contact strength, static stiffness, and dynamic stability directly affect the positioning accuracy, machining stability, and service life of the equipment.
[0003] Current rolling guide pair designs typically employ empirical selection, single-parameter correction, or single-objective verification methods. For the contact problem between rollers and raceways, traditional Hertzian contact calculations can estimate the contact stress level, but they often treat the rollers as ideal line contact bodies, insufficiently considering the boundary effects at the roller ends. Under heavy-duty conditions, significant contact stress peaks easily form at the ends of straight generatrix rollers, leading to end fatigue spalling, crushing, or localized wear, thus affecting the lifespan of the guide pair.
[0004] To reduce edge effects, existing solutions modify the convexity of the roller generatrix, such as using a fully convex circular arc, logarithmic curve, or other end-shaping methods. However, if the modification length is insufficient, stress concentration at the ends still exists; if the convexity or modification length is too large, it will shorten the effective load-bearing length, increase the contact stress in the middle, and reduce the contact stiffness of the guideway pair. In other words, the roller profile parameters are not necessarily better the larger they are, but rather a balance needs to be struck between contact stress uniformity and effective load-bearing length.
[0005] On the other hand, there is a significant coupling between the vertical static stiffness, torsional static stiffness, and contact strength of large, heavy-duty rolling guide pairs. Increasing the contact angle may improve the anti-overturning capability in a certain direction, but it will weaken the load-bearing component in the vertical direction; increasing the effective length of the rollers is beneficial for reducing contact stress and increasing stiffness, but it is limited by the slider structure, ball return space, and installation dimensions; increasing static stiffness may also change the structural mass and stiffness distribution, causing changes in low-order natural frequencies and bringing the risk of resonance. It is difficult to obtain a comprehensive solution that meets the requirements of strength, stiffness, and dynamic stability by simply relying on empirical formulas or a single performance index for design.
[0006] In existing technologies, although general multi-objective evolutionary algorithms such as NSGA-II have been widely used in mechanical structure optimization, the following unresolved technical obstacles exist when directly applying them to the section optimization of large, heavy-duty rolling guide pairs: First, the constraint space height is non-uniform. There is a strong coupling between the shaping length and the convexity measure: when the shaping length... As the convexity increases, the effective contact length decreases, leading to an increase in contact stress; while the convexity... Tiny changes (at the micrometer level) can significantly alter the stress concentration at the ends. Traditional fixed-probability genetic algorithms retain high-intensity random perturbations in the later stages of evolution, easily disrupting the established methods that meet micrometer-level accuracy requirements. Feasible combinations lead to population oscillations near the Pareto front without convergence.
[0007] Second, feasible and infeasible regions are intertwined and nested. Contact stress constraints, structural dimension constraints, preload constraints, and low-order natural frequency constraints form complex boundaries in the 5-dimensional design space. The constraint handling strategies of the classic NSGA-II (such as simple penalty functions) often cause the population to gather in large numbers at the edge of the infeasible region, wasting computational resources and making it difficult to obtain solutions that meet the engineering safety margin.
[0008] Third, there is a conflict between global search and local accuracy. This relates to the engineering feasibility requirements of roller profiles. The optimization results must achieve a specific level of accuracy (e.g.) Accurate to 0.1mm, (Accurate to 1μm), but the discrete encoding precision of classical global optimization algorithms and the finite population size cannot guarantee this precision, and direct rounding may degenerate the solution into an infeasible solution or a suboptimal solution.
[0009] Therefore, existing general-purpose multi-objective optimization algorithms cannot be directly applied to solving this problem, and an improved optimization method that is deeply coupled with the physical characteristics of the cross section of large heavy-duty rolling guide pairs is urgently needed. Summary of the Invention
[0010] The technical problem this invention aims to solve is to provide a collaborative optimization method for the cross-section of large, heavy-duty rolling guide pairs based on an improved multi-objective genetic algorithm. This method establishes a circular arc-corrected roller cross-section model, using maximum Hertzian contact stress, vertical static stiffness, and torsional static stiffness as collaborative optimization objectives. It also introduces constraints on structural dimensions, contact angle, and contact stress, and utilizes an improved hybrid NSGA-II algorithm to output a combination of cross-section parameters that meets multiple performance requirements. This solves the problems of contact stress concentration at the roller ends, difficulty in determining cross-section modification parameters, and the inability to simultaneously achieve both vertical and torsional static stiffness in existing designs of large, heavy-duty rolling guide pairs.
[0011] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: A collaborative optimization method for the cross-section of large, heavy-load rolling guide pairs based on an improved multi-objective genetic algorithm includes the following steps: Step 1: Obtain the structural parameters, material parameters, load parameters, and boundary constraint parameters of the large heavy-duty rolling guide pair, based on the single-sided profile length of the roller. Roller end convexity Contact angle Vertical center distance of slider and the horizontal center distance of the slider As a design variable , And establish a circular arc correction line type roller profile model; Step 2: Based on the circular arc modified line roller profile model, establish the objective function of the maximum Hertzian contact stress between the roller and the raceway to minimize the maximum contact stress between the roller and the raceway. Step 3: Based on the roller normal load distribution relationship under vertical load, establish the objective function of vertical static stiffness of the guide rail pair to maximize the support stiffness of the guide rail pair under vertical load. Step 4: Based on the relationship between the slider rotation angle, roller row lever arm and normal contact deformation under torsional load, establish the objective function of the guide rail pair torsional static stiffness in order to maximize the anti-overturning ability of the guide rail pair under torsional load. Step 5: Establish constraints including shape modification parameter constraints, contact angle constraints, contact stress constraints, and structural dimension constraints. Use the improved hybrid NSGA-II multi-objective genetic algorithm to solve the multi-objective optimization model consisting of maximum Hertzian contact stress, vertical static stiffness, and torsional static stiffness to obtain the Pareto optimal solution set. Step 6: Perform local refinement and finite element verification on the Pareto optimal solution set, and select the rolling guide pair cross-sectional parameter combination that meets the requirements of contact strength, static stiffness and dynamic stability based on the verification results.
[0012] Compared with the prior art, the present invention, employing the above technical solution, has the following technical effects: 1. This invention parameterizes the cross-sectional profile of the arc-corrected roller, and controls the transition profile of the roller end by jointly controlling the correction length and convexity, which can reduce stress concentration at the end edge while maintaining the effective load-bearing length in the middle.
[0013] 2. This invention incorporates the maximum Hertzian contact stress, vertical static stiffness, and torsional static stiffness into a multi-objective optimization model, avoiding the problems of increased contact strength but decreased stiffness, or increased stiffness but increased contact stress caused by single-objective optimization.
[0014] 3. This invention uses the contact angle, vertical center distance of the slider, and horizontal center distance of the slider as coupling design variables, which can adjust the load-bearing component and anti-overturning lever arm according to the width-to-height ratio of the slider, so that the guide rail pair can obtain a more reasonable stiffness distribution in the vertical and torsional directions.
[0015] 4. This invention uses an improved hybrid NSGA-II algorithm for global search and performs local refinement and finite element verification on the Pareto solution set, which can provide multiple sets of optional parameter schemes for engineering design and improve the adaptability and feasibility of guide rail pair profile design. Attached Figure Description
[0016] Figure 1 This is a flowchart of the method for collaborative optimization of the cross-section of large heavy-duty rolling guide pairs according to the present invention; Figure 2 This is a schematic diagram of the cross-sectional parameters of the arc-corrected line roller of the present invention; Figure 3 This is a schematic diagram of the force exerted on the roller in contact with the raceway line according to the present invention; Figure 4 This is a schematic diagram of the force and deformation of the guide rail pair under vertical load according to the present invention; Figure 5 This is a schematic diagram of the slider rotation angle and roller row lever arm under torsional load according to the present invention; Figure 6 This is a flowchart of the improved hybrid NSGA-II multi-objective optimization process of this invention; Figure 7 This is a flowchart of the Pareto optimal solution set selection and finite element verification process of the present invention; Figure 8 This is the Pareto distribution diagram of the improved hybrid NSGA-II of this invention. Detailed Implementation
[0017] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.
[0018] To address the problems of stress concentration at the roller ends, mutual constraints between vertical static stiffness and torsional stiffness, and difficulty in coordinating static performance optimization and dynamic stability in existing large heavy-duty rolling guide pairs under heavy-load conditions, this invention proposes a cross-sectional collaborative optimization method for large heavy-duty rolling guide pairs based on an improved multi-objective genetic algorithm, such as... Figure 1 As shown, it includes the following steps: (1) Obtain the structural parameters, material parameters, load parameters and boundary constraint parameters of large heavy-duty rolling guide pairs. The structural parameters include the total length of the rollers. Roller diameter The roller raceway curvature radius and material parameters include the elastic modulus of the roller material. Poisson's ratio The load parameters include the vertical load acting on the slider. Torsional load Boundary constraint parameters include slider geometry and total roller length. The arrangement space is determined by the slider structure and the circulating ball returner.
[0019] (2) The single-sided shaping length of the roller Roller convexity measurement Contact angle Vertical center distance of slider and the horizontal center distance of the slider As a design variable ,Right now Contact angle This indicates the angle between the direction of the normal force in contact between the roller and the raceway and the horizontal or reference direction; the vertical center distance of the slider. Used to characterize the lever arm distance between the upper and lower roller rows, horizontal center distance of the slider Used to characterize the lever arm distance between the left and right rows of rollers.
[0020] effective contact length of rollers The total length L of the roller and the single-sided trimming length Derivation determination: , Reshaping Length Increasing the length can alleviate the edge stress at the roller end, but the effective contact length... Simultaneously reduce; convexity measure Too small a size will result in under-shaping and insufficient convexity. Excessive size can lead to over-shaping. Therefore, this invention achieves simultaneous coordination through multi-objective optimization. and This is to achieve lower maximum contact stress and higher static stiffness.
[0021] Based on this, a modified arc-shaped roller profile model is established. The roller profile includes a straight middle section and arc transition sections at both ends. The straight middle section ensures that the roller has a sufficient effective load-bearing length under heavy load conditions; the arc transition sections at both ends reduce the contact stress concentration at the roller ends.
[0022] like Figure 2 As shown, the origin of the coordinate system is the axial center of the roller. Let the axial coordinate of the roller be... The radial trimming amount is ,when When the roller is in the middle straight segment, the following condition is met: , when At this time, the roller is in the arc transition section, and the radial modification amount changes according to the arc relationship: , And satisfy the following boundary conditions: , , This allows the arc transition section to be continuously connected to the straight middle section, and causes the roller end to form a preset shrinkage amount relative to the straight middle section.
[0023] The radius of curvature of the arc transition section depends on the shaping length. and convex metric This ensures that the arc transition section smoothly connects to the straight middle section at the connection point. This is achieved through adjustment. and Different roller profiles can be obtained and used for subsequent calculations of contact stress, static stiffness and modal properties.
[0024] (3) Based on the circular arc modified line roller profile model, establish the objective function of the maximum Hertzian contact stress between the roller and the raceway, and take the maximum Hertzian contact stress as the objective to be minimized; When the roller and raceway form a line contact, a target function for the maximum Hertzian contact stress is established based on the equivalent elastic parameters, normal load, contact curvature, and effective contact length. The target function for the maximum Hertzian contact stress is expressed as: , in, Design variables The corresponding maximum Hertzian contact stress between the roller and the raceway.
[0025] like Figure 3 As shown, in one embodiment, the maximum Hertzian contact stress is expressed by the following formula: , in, This is a stress correction factor related to the material's Poisson's ratio and contact state; For the normal load borne by a single roller; The contact half-width is Hertz; Hertz contact half-width It can be determined according to the following formula: , in, The equivalent radius of curvature; It is the equivalent elastic modulus, which can be determined by the elastic modulus and Poisson's ratio of the roller and raceway materials.
[0026] (4) Based on the distribution relationship of the roller normal load and the contact deformation relationship under vertical load, establish the objective function of the vertical static stiffness of the guide rail pair, and take the vertical static stiffness as the objective to be maximized; The vertical static stiffness of the guide rail pair is used to characterize the slider's ability to resist vertical displacement under vertical load. The objective function for vertical static stiffness is expressed as: , in, Design variables The corresponding vertical static stiffness; like Figure 4 As shown, the vertical static stiffness can be expressed as: , in, The vertical load acting on the slider This represents the vertical displacement of the slider.
[0027] Roller normal load under vertical load With vertical load Number of bearing roller rows Number of effective load-bearing rollers per column and contact angle Related, can be expressed as: , in, To accommodate the number of rows of rollers, The number of effective load-bearing rollers per column, Contact angle; Contact deformation between roller and raceway This can be expressed as a Hertzian contact nonlinearity relationship: , In the formula, The contact deformation coefficient is... The Hertzian contact index; Vertical displacement of slider Deformation in contact with normal direction and contact angle Related can be represented as: , This allows us to establish a vertical load. With vertical displacement The mapping relationship between them is obtained by differentiation to obtain the vertical static stiffness. .
[0028] (5) Based on the relationship between the slider rotation angle, roller row lever arm and normal contact deformation under torsional load, establish the objective function of the torsional static stiffness of the guide rail pair, and take the torsional static stiffness as the objective to be maximized; The torsional static stiffness of the guide rail pair is used to characterize the slider's ability to resist angular deformation under torsional loads. The objective function for torsional static stiffness is expressed as: , in, Design variables The corresponding torsional static stiffness; Torsional static stiffness can be expressed as: , in, For the torsional load acting on the slider, This is the rotation angle produced by the slider under torsional load.
[0029] like Figure 5 As shown, under torsional load, the normal loads of different roller rows are redistributed. Let the vertical center distance of the sliders be... The horizontal center distance of the sliders is The equivalent arm formed by the roller array It can be represented as: , in, As equivalent force arm, It is a function determined by the vertical center distance of the slider, the horizontal center distance of the slider, and the contact angle; In one implementation, the following equivalent resultant lever arm relationship can be adopted: , Based on the relationship between torsional load, equivalent force arm, and slider rotation angle, establish the torsional load... With corner The mapping relationship between them is used to obtain the torsional static stiffness. .
[0030] When the slider structure is nearly symmetrical, the contact angle A balance between vertical and horizontal support can be achieved at approximately 45°. When the slider has a tall and narrow structure, a relatively small contact angle can be used to increase the vertical load-bearing component; when the slider has a short and wide structure, a relatively large contact angle can be used to increase the horizontal load-bearing component. The above contact angle adjustments are automatically completed through the optimization model.
[0031] (6) To ensure the engineering feasibility of the optimization results, constraints including structural size constraints, shape modification parameter constraints, contact stress constraints and contact angle constraints are established. Shape modification parameter constraints: To ensure the feasibility of roller machining, convexity measurement It can be set to the micrometer range; to ensure the effective load-bearing length of the roller, the profile length is adjusted. It should not exceed the total length of the rollers. Preset ratio: , , in, and These are the lower and upper limits of the shaping length, respectively; and These are the lower and upper limits of the convexity metric, respectively. , in, The minimum effective contact length to meet the load-bearing stiffness requirements.
[0032] Contact angle constraint: , in, and These represent the lower and upper limits for contact angle design, respectively.
[0033] The contact stress constraint is: , in, This represents the allowable contact stress of the material.
[0034] The structural dimensional constraints are as follows: , , in, and These are the lower and upper limits of the vertical center distance of the slider, respectively. and These are the lower and upper limits of the horizontal center distance of the slider, respectively.
[0035] The above constraints prevent optimization results from encountering problems such as inability to process or install, excessive contact stress, insufficient effective bearing length, or excessively low low-order modal frequencies. Based on these constraints, a design variable combination that simultaneously satisfies all constraints is defined as a feasible solution; if any constraint is not satisfied, it is defined as an infeasible solution. For design variable combinations... The total constraint violation degree is determined by calculating the degree of violation of each constraint. When the total constraint violation degree is zero, the combination of design variables belongs to the feasible region; when the total constraint violation degree is greater than zero, the combination of design variables belongs to the infeasible region.
[0036] (7) An improved hybrid NSGA-II multi-objective genetic algorithm is used to solve the multi-objective optimization model consisting of maximum Hertzian contact stress, vertical static stiffness and torsional static stiffness. After the global evolution reaches the preset number of generations or the change of the objective function is less than the preset threshold, the local refinement process is started. The neighborhood of each candidate solution in the Pareto solution set is searched twice to correct the local deviations of the shape length, convexity and contact angle, and to obtain the Pareto optimal solution set.
[0037] The improved hybrid NSGA-II multi-objective genetic algorithm employs adaptive crossover and mutation probabilities, dynamically adjusting the search strategy based on the number of generations in the population evolution. It introduces a feasible solution priority mechanism based on constraint violation, eliminating the need to pre-set the weights of each objective function and enabling the acquisition of a set of non-dominated solutions under multi-objective conflict conditions. For example... Figure 6 As shown, the specific solution process is as follows:
[0038] A1. Initialize the population. Randomly generate an initial population based on the upper and lower limits of the design variables. Each individual corresponds to a combination of guide rail pair profile parameters. The initial population is used as the parent population.
[0039] A2, Calculate the objective function. Calculate the maximum Hertzian contact stress for each individual in the parent population. Vertical static stiffness and torsional static stiffness .
[0040] A3, Calculate constraint violation degree. If an individual does not meet the constraints of shape modification parameters, contact stress, structural dimensions, or modal frequencies, a constraint penalty value is assigned according to the degree of violation.
[0041] A4 performs a fast non-dominated sort on the current population, dividing the population into different non-dominated levels.
[0042] A5, Calculate crowding distance. Calculate crowding distance for individuals within the same non-dominated rank to maintain the uniformity of the Pareto solution set distribution.
[0043] A6 performs selection, crossover, and mutation operations to generate the offspring population. The adaptive crossover probability is also considered. and mutation probability Adjust dynamically according to the following formula: , , in, , These represent the initial crossover probability and the initial mutation probability; The current generation number; The maximum number of generations; For individual constraint violation degree; For constraint correction factors; The constraint feedback strength coefficient; , These are the minimum crossover probability and mutation probability, respectively, to ensure that the strength of basic genetic operations is maintained in the later stages of evolution; For the first Average constraint violation rate of the generation population.
[0044] This function form causes the crossover and mutation probabilities to decrease linearly with the number of generations to a non-zero lower bound. , And dynamically compensate based on the overall constraint satisfaction of the population: when the average constraint violation rate of the population is high, The system automatically increases the intensity of genetic operations to explore new feasible regions; when the population approaches feasibility, Approaching zero, this term approaches 1, focusing the search on Pareto frontier refinement. Individual-level constraint violation. pass The term imposes additional inhibition on individuals that do not meet the constraints. As the number of generations increases, the overall probability of crossover and mutation decreases, and the population tends to stabilize in the later stages.
[0045] A7 merges the parent and offspring populations and preserves elites based on non-dominant hierarchy and crowding distance, serving as the parent population for the next generation of evolution.
[0046] A8 checks if the termination condition is met. If the maximum number of generations has been reached, output the Pareto solution set; otherwise, return to A2.
[0047] To improve solution accuracy in the later stages of convergence, this invention initiates a local refinement process after the global search is completed. The local refinement process centers on each candidate solution in the Pareto solution set, adjusting the shape length... and convex metric A secondary search is performed within the neighborhood of each candidate solution using a small-step grid search. Specifically, for each design variable of each candidate solution, a preset step size is used. and A finite neighborhood traversal is performed to evaluate the objective function value and constraint violation degree at each point within the neighborhood, retaining solutions that satisfy the constraints and have a better overall objective function value. If no improvement is obtained after the neighborhood search, the search is continued along the descent direction of the objective function with a step size halved until the step size is lower than the preset halving threshold or the maximum number of local iterations is reached. This corrects local deviations in the shaping length, convexity, and contact angle, enabling the final solution to achieve a more accurate balance between contact stress, static stiffness, and dynamic stability.
[0048] To verify the optimization performance of the improved hybrid NSGA-II algorithm proposed in this invention, it is compared with the classic NSGA-II algorithm and the MOEA / D algorithm on the same multi-objective optimization problem of guide rail cross section. All three algorithms use the same population size, maximum number of generations, and design variable boundary conditions to ensure fairness in the comparison. The unified parameters for each algorithm are: population size... Maximum number of generations The design variable dimension is 5 (roller single-sided shaping length). convex measurement Contact angle Vertical center distance of slider and horizontal center distance The specific parameters for the two comparison algorithms are set as follows: the neighborhood size of MOEA / D. The weight vectors are uniformly distributed. The parameters of the improved algorithm in this invention are set as follows: initial crossover probability... Initial mutation probability Constraint correction factor Initial step size for local fine-tuning , , step size halving threshold .
[0049] Two universally accepted comprehensive evaluation metrics in the field of multi-objective optimization are adopted: Inverse Generation Distance (IGD) and Hypervolume (HV). IGD is used to measure the degree of approximation and distribution uniformity between the obtained Pareto front and the true Pareto front, and is defined as: , in, The Pareto front is used as a reference (it is formed by taking the union of all non-dominated solutions obtained by running the classic NSGA-II and MOEA / D independently 30 times each, and then sorting and filtering them by non-dominated sorting, and is unrelated to the algorithm being evaluated). The Pareto solution set output by the algorithm to be evaluated. To solve and The Euclidean distance between them. The smaller the IGD value, the better the convergence and distribution of the solution set obtained by the algorithm.
[0050] HV is defined as the hypervolume enclosed by the Pareto solution set and the reference point in the objective function space, and its definition is: , in, For the Lebesgue measure, Use the reference point (take the worst value of each objective function). Indicates from the reference point reconciliation A defined hyperrectangle. A larger HV value indicates better convergence and diversity of the solution set.
[0051] Each algorithm was run independently 30 times, and the mean and standard deviation of the IGD and HV indices were recorded. The statistical results are shown in Table 1.
[0052] Table 1. Comparison of IGD and HV metrics for three algorithms (30 independent runs)
[0053] As shown in Table 1, regarding the IGD index, the mean of the improved hybrid NSGA-II algorithm of this invention is 1.284e-2, which is about 66.6% lower than the classic NSGA-II and about 56.6% lower than MOEA / D. Meanwhile, the standard deviation is only 1.73e-3, significantly lower than the two comparative algorithms, indicating that the improved algorithm has significant advantages in solution set convergence accuracy and distribution uniformity, and its stability after multiple runs is greatly improved. Regarding the HV index, the mean of the improved algorithm of this invention is 0.8215, which is about 13.5% higher than the classic NSGA-II and about 8.3% higher than MOEA / D. At the same time, the standard deviation is reduced to 8.45e-3, indicating that the Pareto solution set obtained by the improved algorithm covers a better and wider region, resulting in higher overall quality. The comparison of standard deviations shows that the improved algorithm has the smallest IGD and HV standard deviations among the two comparative algorithms. This proves that the adaptive crossover / mutation probability mechanism and the feasible solution priority constraint handling strategy effectively suppress random fluctuations, making the algorithm more robust in the highly constrained and multi-conflict objective problem of guide rail cross section optimization. The introduction of the local refinement process allows the Pareto solution set to be further refined in the later stage of convergence, which is an important reason why the improved algorithm's IGD and HV indices are significantly better than the two global optimization algorithms.
[0054] To evaluate algorithm performance from two dimensions—convergence speed and distribution uniformity—the number of generations required for the three algorithms to reach convergence and the Pareto solution set spacing index are supplemented. The spacing index measures the uniformity of distances between adjacent solutions on the Pareto front and is defined as follows: , in, For the first The Euclidean distance from each solution to its nearest neighbor solution. For all The mean, The value represents the size of the Pareto solution set. A smaller Spacing value indicates a more uniform distribution of the solution set. The convergence criterion is set as follows: the rate of change of IGD is less than 1% and the rate of change of HV is less than 0.5 over 20 consecutive generations. Statistical results are shown in Table 2.
[0055] Table 2 Comparison of convergence algebra and spacing metrics for the three algorithms (30 independent runs)
[0056] As shown in Table 2, regarding convergence speed, the improved algorithm of this invention requires an average of only 128 generations to meet the convergence criterion, which is about 31.6% less than the classic NSGA-II and about 17.9% less than MOEA / D. Furthermore, the standard deviation of the convergence generation is only 6.3, the smallest among the three algorithms. This demonstrates that the adaptive probability adjustment mechanism and feasible solution priority strategy enable the algorithm to focus on high-quality search regions more quickly, avoiding ineffective random walks. Regarding distribution uniformity, the improved algorithm has a mean Spacing value of 0.0247, which is about 52.8% lower than the classic NSGA-II and about 36.0% lower than MOEA / D. The standard deviation is also the smallest among the three algorithms, indicating that the Pareto solution set obtained by the improved algorithm is more uniformly and stably distributed in the target space, providing a more representative set of parameter schemes for engineering design. As can be seen from the results in Tables 1 and 2, the improved hybrid NSGA-II algorithm of this invention outperforms the classic NSGA-II and MOEA / D in four dimensions: convergence accuracy (IGD), solution set coverage (HV), convergence speed (convergence algebra), and distribution uniformity (Spacing). This verifies the technical effect of the synergistic improvement of adaptive crossover / mutation probability, feasible solution priority constraint processing, and local refinement. Figure 8 The Pareto distribution diagram of the improved hybrid NSGA-II of this invention.
[0057] (8) Perform finite element verification on the Pareto optimal solution set, and select the rolling guide pair cross section parameter combination that meets the requirements of contact strength, static stiffness and dynamic stability based on the verification results.
[0058] After obtaining the Pareto optimal solution set, this invention further performs engineering screening and finite element verification, such as... Figure 7 As shown: The screening process includes: B1, eliminate candidate solutions where the maximum contact stress is close to the allowable limit and the safety margin is insufficient; B2, eliminate candidate solutions whose vertical static stiffness or torsional static stiffness is lower than the design requirements; B3 is a candidate solution that balances contact stress, vertical static stiffness, and torsional static stiffness.
[0059] The finite element verification process includes: C1, Establish the roller-raceway contact finite element model corresponding to the selected parameter combination; C2, apply a preset normal load and extract the contact stress distribution in the direction of the roller generatrix; C3, determine whether there is a significant edge stress peak at the end of the roller; C4, apply a vertical static load, extract the vertical displacement of the slider and calculate the vertical static stiffness; C5, apply torsional load, extract slider rotation angle and calculate torsional static stiffness; C6. Establish the working modal model, extract the low-order natural frequencies and determine whether the dynamic constraints are satisfied. Once the candidate solution passes the above verification, it is determined as the combination of profile parameters for large heavy-duty rolling guide pairs.
[0060] In one specific embodiment, the total length of the roller is... Take 100 mm, roller diameter Take 30 mm, the roller material is structural steel, and the elastic modulus is... Take 210 GPa, Poisson's ratio Take 0.3. Apply a normal load of 15000 N to the roller half-section and set the friction coefficient between the roller and the contact base to 0.01.
[0061] In this embodiment, a circular arc-corrected line roller is used, and the correction length is... and convex metric As a key microscopic truncated shape variable. Finite element verification shows that when the modification length... When the length is too small, significant edge stress concentration easily occurs at the roller end; when the shaping length increases to a suitable range, the stress distribution in the roller generatrix direction tends to be flat. , For example, the peak contact stress in the direction of the roller generatrix can be reduced to about 985 MPa, and the stress concentration at the end edge is suppressed.
[0062] Regarding contact angle optimization, when the slider cross-section is nearly symmetrical, the initial range of the contact angle can be set around 45°, and the final value is determined by a multi-objective optimization model based on the combined performance of vertical static stiffness and torsional static stiffness. When the slider has a tall and narrow structure, the optimization result can be biased towards a smaller contact angle; when the slider has a short and wide structure, the optimization result can be biased towards a larger contact angle.
[0063] The above embodiments demonstrate that the present invention can achieve synergistic optimization among roller end stress control, effective contact length maintenance, vertical static stiffness improvement, torsional static stiffness improvement, and low-order modal frequency constraint, providing an implementable parametric optimization method for the cross-sectional design of large heavy-duty rolling guide pairs.
[0064] Based on the same inventive concept, embodiments of this application provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the aforementioned method for collaborative optimization of the cross section of a large heavy-duty rolling guideway based on an improved multi-objective genetic algorithm.
[0065] Based on the same inventive concept, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned steps of the collaborative optimization method for the cross-section of large heavy-duty rolling guide pairs based on an improved multi-objective genetic algorithm.
[0066] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0067] This invention is described with reference to flowchart illustrations of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each step in the flowchart, and combinations of steps in the flowchart, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing device, generate instructions for implementing the steps in the flowchart. Figure 1 A device for a function specified in one or more processes.
[0068] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 The function specified in one or more processes.
[0069] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 Steps of a specified function in one or more processes.
[0070] The above embodiments are merely illustrative of the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solutions based on the technical concept proposed in this invention shall fall within the scope of protection of this invention.
Claims
1. A collaborative optimization method for the cross-section of large, heavy-load rolling guide pairs based on an improved multi-objective genetic algorithm, characterized in that, Includes the following steps: Step 1: Obtain the structural parameters, material parameters, load parameters, and boundary constraint parameters of the large heavy-duty rolling guide pair, based on the single-side profile length of the roller. Roller end convexity Contact angle Vertical center distance of slider and the horizontal center distance of the slider As a design variable , And establish a circular arc correction line type roller profile model; Step 2: Based on the circular arc modified line roller profile model, establish the objective function of the maximum Hertzian contact stress between the roller and the raceway to minimize the maximum contact stress between the roller and the raceway. Step 3: Based on the roller normal load distribution relationship under vertical load, establish the objective function of vertical static stiffness of the guide rail pair to maximize the support stiffness of the guide rail pair under vertical load. Step 4: Based on the relationship between the slider rotation angle, roller row lever arm and normal contact deformation under torsional load, establish the objective function of the guide rail pair torsional static stiffness in order to maximize the anti-overturning ability of the guide rail pair under torsional load. Step 5: Establish constraints including shape modification parameter constraints, contact angle constraints, contact stress constraints, and structural dimension constraints. Use the improved hybrid NSGA-II multi-objective genetic algorithm to solve the multi-objective optimization model consisting of maximum Hertzian contact stress, vertical static stiffness, and torsional static stiffness to obtain the Pareto solution set. Step 6: Perform local refinement and finite element verification on the Pareto solution set, and select the rolling guide pair cross-sectional parameter combination that meets the requirements of contact strength, static stiffness and dynamic stability based on the verification results.
2. The method for collaborative optimization of the cross-section of large heavy-load rolling guide pairs based on an improved multi-objective genetic algorithm according to claim 1, characterized in that, In step 1, the roller profile includes a straight middle section and rounded transition sections at both ends. The straight middle section is used to ensure that the roller has an effective contact length under heavy load conditions, and the rounded transition sections at both ends are used to reduce the contact stress concentration at the ends of the roller. With the axial center of the roller as the origin, the direction along the roller axis as the axial direction, and the radial direction pointing outwards, the axial coordinate of the roller is: The radial trimming amount is ,when When the roller is in the middle straight segment, the following condition is met: , when At this time, the roller is in the arc transition section, and the radial modification amount is: , radius of curvature of the circular transition segment for: And satisfy , This ensures a continuous connection between the arc transition segment and the middle straight segment; in, The total length of the rollers, , This represents the effective contact length of the roller.
3. The method for collaborative optimization of the cross-section of large heavy-duty rolling guide pairs based on an improved multi-objective genetic algorithm according to claim 1, characterized in that, In step 2, the objective function for the maximum Hertzian contact stress is expressed as: , in, Design variables The corresponding maximum Hertzian contact stress between the roller and the raceway; , , in, This is a stress correction factor related to the Poisson's ratio of the roller material and the contact state. For the normal load borne by a single roller, For Hertzian contact half-width, This is the effective contact length of the roller. The equivalent radius of curvature, It is the equivalent elastic modulus.
4. The method for collaborative optimization of the cross-section of large heavy-duty rolling guide pairs based on an improved multi-objective genetic algorithm according to claim 1, characterized in that, In step 3, the objective function for vertical static stiffness is expressed as: , in, Design variables The corresponding vertical static stiffness; , in, The vertical load acting on the slider This represents the vertical displacement of the slider. , , , in, This refers to the contact deformation between the roller and the raceway. The contact deformation coefficient is... For the normal load borne by a single roller, The Hertzian contact index, To accommodate the number of rows of rollers, This represents the number of effective load-bearing rollers per column.
5. The method for collaborative optimization of the cross-section of large heavy-duty rolling guide pairs based on an improved multi-objective genetic algorithm according to claim 1, characterized in that, In step 4, the objective function for torsional static stiffness is expressed as: , in, Design variables The corresponding torsional static stiffness; , in, For the torsional load acting on the slider, This is the rotation angle generated by the slider under torsional load; Under torsional load, the normal load of different roller rows is redistributed, and the equivalent force arm formed by the roller rows... Represented as: 。 6. The method for collaborative optimization of the cross-section of large heavy-load rolling guide pairs based on an improved multi-objective genetic algorithm according to claim 1, characterized in that, In step 5, the shaping parameter constraints are as follows: , , , in, and These are the lower and upper limits of the unilateral shaping length, respectively. and These are the lower and upper bounds of the convexity metric, respectively. This is the effective contact length of the roller. Minimum effective contact length to meet load-bearing stiffness requirements; The contact angle constraint is: , in, and These are the lower and upper limits for contact angle design, respectively. The contact stress constraint is: , in, For the maximum Hertzian contact stress, This refers to the allowable contact stress of the material. The structural dimensional constraints are as follows: , , in, and These are the lower and upper limits of the vertical center distance of the slider, respectively. and These are the lower and upper limits of the horizontal center distance of the slider, respectively.
7. The method for collaborative optimization of the cross-section of large heavy-duty rolling guide pairs based on an improved multi-objective genetic algorithm according to claim 1, characterized in that, In step 5, the improved hybrid NSGA-II multi-objective genetic algorithm uses adaptive crossover probability and mutation probability to generate the offspring population. and mutation probability Update dynamically using the following formula: , , in, , These are the initial crossover probability and the initial mutation probability, respectively. For the current generation, For the maximum number of generations, For individual constraint violation degree, For constraint correction factors, , These are the minimum crossover probability and the mutation probability, respectively. To constrain the feedback strength coefficient, This represents the average constraint violation rate of the population corresponding to the current generation.
8. A computer device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the collaborative optimization method for the cross section of large heavy-load rolling guide pairs based on an improved multi-objective genetic algorithm as described in any one of claims 1 to 7.
9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the collaborative optimization method for the cross section of large heavy-duty rolling guide pairs based on an improved multi-objective genetic algorithm as described in any one of claims 1 to 7.
