Response surface model correction-based turnout beam support optimization design method
By fitting the response surface model of the turnout beam support structure using multivariate adaptive regression splines, and combining it with central composite experimental design and sequential quadratic programming algorithm, the design of the turnout beam support structure is optimized. This addresses the shortcomings of traditional response surface methods and achieves higher accuracy and more economical design results.
Patent Information
- Application Number
- CN202411072818.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-06
- Publication Date
- 2024-12-06
AI Technical Summary
Traditional response surface methodology lacks adaptability in data processing, leading to overfitting or underfitting of the model and failing to accurately predict the optimal design scheme for turnout beam support structures.
A response surface model is constructed by fitting the relationship between the structural design parameters and the response using a multivariate adaptive regression spline method combined with a central composite experimental design. The design parameters are then optimized using a sequential quadratic programming algorithm to obtain the optimal solution.
It improves the design accuracy and economy of turnout beam support structures, reduces material usage, lowers engineering costs, and ensures the safety and reliability of the structure.
Smart Images

Figure CN119089537A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of turnout beam optimization, and in particular to a turnout beam support optimization design method based on response surface model correction. Background Art
[0002] Formwork and support systems are temporary structures commonly used in engineering structure construction. The reliability of their design is related to the quality of concrete pouring, the accuracy of bridge alignment, and also has a great impact on construction safety. In order to effectively avoid accidents such as collapse of the formwork frame under load, it is necessary to carry out structural design of the formwork frame to ensure that its strength, rigidity and stability meet the requirements of the specification. Considering that the cost of formwork frame accounts for a high proportion of the construction cost, if the structure can be optimized and designed under the premise of ensuring structural reliability and safety, some potential hidden dangers of the structure can be prevented and reduced, and the use of construction materials can be reduced, it will be of great significance to further reduce the cost.
[0003] Structural optimization has been widely used in the fields of bridge design, material mix ratio, etc. For example, "Design optimization of wind resistance performance of super high-rise buildings based on optimization criterion method and particle swarm algorithm [J]. Journal of Building Structures, 2023, 44(5): 58-67" can obtain the optimal design scheme of super high-rise buildings with the optimal structural cost under the premise of meeting the wind resistance design conditions, saving 3.3% of the structural cost.
[0004] For example, "Multi-objective optimization of large-span steel structures of airport terminals based on NSGA-Ⅱ and scheme optimization [J]. Journal of Hefei University of Technology (Natural Science Edition), 2023, 46(7): 941-949", through finite element simulation technology, a multi-objective optimization large-span steel system was constructed, and combined with a non-dominated genetic sorting algorithm, the optimal construction scheme improved the quality, reduced the construction period, and reduced the cost compared with the previous one.
[0005] For example, "Research on Optimal Design of Bridge Construction Brackets Driven by Parametrics [J]. Highway, 2024, (5): 247-250", optimizes the design of brackets through parametric analysis of rods, while ensuring structural performance, reducing material usage by 45.9%, improving construction efficiency, and significantly reducing the use of construction materials.
[0006] Multivariate adaptive regression spline is a regression method with strong generalization ability specifically for high-dimensional data. Its basic idea is to use a set of piecewise continuous functions to approximate the complex functional relationship between the dependent variable y and the variable x. In the regression process, two processes, "forward stepwise search" and "backward pruning" are used to gradually screen factors. The determination of the basis function and the number of basis functions do not require manual selection. By fitting the entire area piecewise, a flexible prediction model is summarized and combined.
[0007] For example, the Chinese patent with the patent publication number CN115688924A discloses "a multi-sample joint multivariate adaptive regression spline model", which establishes a unified model for multiple sample data and obtains the display expression of the fitting results, which is convenient for the subsequent unified analysis of the data characteristics of multiple samples. It has greater flexibility, can fit complex curves, allows the curve to have high curvature, multiple inflection points, nonlinear phenomena, and has fast processing speed and high accuracy.
[0008] The traditional response surface method usually pre-sets a response model form and determines the model parameters through data calculation. This method has the following problems: when processing data, it is necessary to have a priori understanding of the model form of the data. Since the distribution of the data is unknown, it is impossible to predict which response surface model form is the best choice. Therefore, artificially pre-selecting the response surface model form will have a certain impact on the accuracy of the final fitting;
[0009] This comparison document is used in the intelligent pressure sensor mainboard circuit current drift test data, and the usage scenarios are quite different from the turnout beam support; and the purpose of the comparison document is to establish a current drift model for multiple sensor mainboards. The lack of a correction process will result in problems such as over-fitting or under-fitting and low model accuracy. Summary of the invention
[0010] In view of the deficiencies in the prior art, the present invention provides a turnout beam support optimization design method based on response surface model correction to solve the above problems.
[0011] The present invention provides the following technical solutions:
[0012] A method for optimizing the design of a turnout beam support based on a response surface model correction comprises the following steps:
[0013] S1. Determine the structural design parameters and structural response range of the turnout beam support, and use central composite experimental design for experimental design;
[0014] S2. According to the experimental design plan, the initial finite element model of the turnout beam support is established to perform mechanical performance analysis and obtain the structural response corresponding to each group of structural design parameters;
[0015] S3, fitting the functional relationship between the structural design parameters and the structural response by multivariate adaptive regression spline to obtain the response surface model of each response feature;
[0016] S4. Perform accuracy tests on each response surface model to obtain several adaptive response surface models that pass the accuracy tests. Combine the structural response limits to construct the optimization objective function, and use the sequential quadratic programming algorithm to optimize the inversion and obtain the optimal design values of the structural design parameters.
[0017] Preferably, the turnout beam support includes cross-sectional dimensions and material property ranges of square timber, I-beams, Bailey beams, double I-beams, and steel pipe piles, and a central composite design is used for experimental design to obtain multiple sets of structural design parameter combinations.
[0018] Preferably, a finite element model of the turnout beam support is established using finite element software, and each set of structural design parameters in the experimental design is substituted into the finite element model for calculation to calculate the corresponding deflection and stress of the Bailey beam, the deflection and stress of the distribution beam, the deflection and stress of the double-jointed I-beam, the stress of the steel pipe pile, and the stress and deflection response of the square timber.
[0019] Preferably, the structural design parameters of the key dimensions and material properties of the main components are used as variables, the structural response characteristics are used as dependent variables, and a multivariate adaptive regression spline is used to fit a response surface model reflecting the relationship between the structural design parameters and the structural response. The expression of the multivariate adaptive regression spline is as follows:
[0020]
[0021] In the formula, is a constant; B m (x) is the mth basis function, which is the tensor product of the spline function; is the coefficient of the mth basis function; M is the number of model terms;
[0022] B m (x) is the basis function; S km The value is ±1; X v(k,m) represents the v(k,m)th component of the prediction factor x, t km is a point, [(x v(k,m) -t km )] + is a semi-cut polynomial.
[0023] Preferably, the expression of the precision test in S4 is as follows:
[0024]
[0025] Where: y RS (j) represents the jth response surface model calculation result; y(j) represents the jth finite element calculation result; N is the number of test points in the design space; y represents the average value of the finite element calculation result;
[0026] The mean square error (MSE) reflects the degree of dispersion of the test data points to the model, that is, the average value of the error of each point; the relative root mean square error (RRMSE) reflects the degree of fit of the model to the test data. The closer the relative root mean square error (RRMSE) is to 0, the better the fit of the model. A value close to 1 means that the model is almost the same as the mean of y. A high relative root mean square error (RRMSE) means that the model may have overfitted the test data. 2 The closer it is to 1, the higher the accuracy of the model.
[0027] Preferably, the optimization objective function is constructed in combination with the structural response limit, as shown in the following formula:
[0028]
[0029] x lk ≤x k ≤x uk
[0030] Where: x k is the design parameter to be modified; x lk and x uk is the design space range of structural parameters; f RSM is the test value; f 响应限值 The allowable value of each component response is determined in combination with the specifications.
[0031] Preferably, a quadratic sequence programming algorithm is used to solve the objective function. When the optimization objective function satisfies a given convergence condition, the optimal solution of the structural design parameters under the structural safety condition is obtained. If the given convergence condition is not met, the next round of finite element calculation and iterative solution is performed until the convergence condition is met.
[0032] A computer program product, comprising a non-transitory computer-readable storage medium storing a computer program, wherein the computer program is operable to enable a computer to execute the method for optimizing the design of a turnout beam support based on response surface model correction.
[0033] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the program, the method for optimizing the design of a turnout beam support based on response surface model correction is implemented.
[0034] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements a method for optimizing the design of a turnout beam support based on response surface model correction.
[0035] Preferably, the turnout beam support structure includes square timber, I-beams, Bailey beams, double I-beams, steel pipe piles and foundations;
[0036] The steel pipe pile is vertically arranged on the foundation, the double-jointed I-beam is arranged on the upper end of the steel pipe, the Bailey beam is fixedly connected to the double-jointed I-beam, distribution beams are laid at intervals above the Bailey beam, square timbers are laid at intervals above the distribution beams, and bamboo plywood is laid above the square timbers. The distribution beam adopts I-beam.
[0037] Preferably, the double-piece I-beams are provided with stiffening plates at the stress node positions, and the stiffening plates are firmly fixed to the frame columns and steel pipe columns by using steel wire ropes in combination with basket bolts.
[0038] Preferably, a wing plate supporting truss is provided between the turnout beam support structure and the bridge.
[0039] The present invention has the following beneficial technical effects:
[0040] By taking the structural design parameters as independent variables and the structural response as dependent variables, and using multivariate adaptive regression splines to automatically determine the optimal form of the function based on the data, the modeling problem when the data distribution is unknown is solved. At the same time, the optimal response surface model is fitted through the three processes of forward search, backward pruning, and optimal model selection. In the fitting process, the independent variables can be selected independently according to their importance.
[0041] After obtaining the response surface model, the structural optimization method of the support was determined based on the response limit and multi-objective optimization method. After the optimized design, the turnout beam formwork system effectively reduced the number of Bailey beams, steel pipe wall thickness and other materials compared to the original design, significantly reduced the amount of steel used, and reduced the project cost as much as possible while ensuring structural safety. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] Figure 1 It is a schematic diagram of the test point distribution of the two-factor central composite experimental design of the present invention;
[0043] Figure 2 Schematic diagram of the optimized front turnout beam support structure of the present invention;
[0044] Figure 3 It is the central composite design representation of the present invention;
[0045] Figure 4 It is a schematic diagram of the test value of the response surface model of the present invention;
[0046] Figure 5 This is a schematic diagram of the optimized turnout beam support structure of the present invention;
[0047] Figure 6 It is a schematic diagram of parameter values before and after optimization of the present invention;
[0048] Figure 7 It is a schematic diagram of the change of response value before and after optimization of the present invention;
[0049] Figure 8 It is a flow chart of the present invention.
[0050] The reference numerals in the figure are:
[0051] 1. Bridge; 2. Wing support truss; 3. Square timber; 4. I-beam; 5. Bailey beam; 6. Double I-beam; 7. Steel pipe pile; 8. Foundation. DETAILED DESCRIPTION
[0052] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0053] A method for optimizing the design of turnout beam supports based on response surface model modification, such as Figure 1-7 As shown:
[0054] The support used in the concrete pouring construction of the turnout beam is optimized to improve the safety and economy of the support. First, the key dimensions and material properties of the main components and other structural design parameters are selected as variables, and the structural response characteristics are selected as dependent variables. The central composite experimental design method is used for experimental design; the finite element model of the turnout beam support is established, and the mechanical characteristics are analyzed according to the experimental design scheme, and the structural response characteristics corresponding to each group of variables are calculated; then, the multivariate adaptive regression spline is used to fit the complex nonlinear functional relationship between the variables and the dependent variables, and the optimization objective function is constructed in combination with the response limit, and the multi-objective optimization method is used to calculate and determine the optimal design values of each structural design parameter of the turnout beam support.
[0055] Central Composite Design (CCD) is a commonly used method in experimental design. It combines the advantages of full factorial design and fractional factorial design. It evaluates the curvature of the response surface by adding central points and axial points, thereby predicting the optimal solution more accurately.
[0056] The central composite design is composed of 2 k Factorial design or fractional factorial design (normalized to the usual ±1 notation) adds 2k axis points and n c The experimental points of the two-factor central composite experimental design are distributed as follows: Figure 1 shown.
[0057] Multivariate adaptive regression spline is a regression method with strong generalization ability specifically for high-dimensional data. It uses a set of piecewise continuous functions to approximate the dependent variable y and the variable x = (x 1 ,x 2 ,...,x n ), the regression process uses the “forward stepwise search” and “backward pruning” processes to gradually screen factors. Neither the determination of basis functions nor the number of basis functions requires manual selection. The entire area is piecewise fitted and then summarized and combined to form a flexible prediction model.
[0058] The expression of the multivariate adaptive regression spline model is shown as follows:
[0059]
[0060] In the formula, is a constant; B m (x) is the mth basis function, which is the tensor product of the spline function; is the coefficient of the mth basis function; M is the number of model terms;
[0061] B m (x) is the basis function; S km The value is ±1; X v(k,m) represents the v(k,m)th component of the prediction factor x, t km is a point, [(x v(k,m) -t km )] + is a semi-cut polynomial.
[0062] Given N samples of variable x and dependent variable y, we first need to determine the basis function B m (x), a forward stepwise process is used to incorporate pairs of basis functions into the model. The determination of basis functions (the number of variables in the product and the cut-off points of the variables) and the number of basis functions are automatically completed by the data, without the need for manual selection. After this process, the model obtained will contain more basis functions than the appropriate size, resulting in overfitting. In the case of overfitting, although the fit to the training samples is high, the overfitting is greatly affected by the noise value, so the prediction ability of the model is poor.
[0063] Therefore, backward pruning is required, using the backward deletion method, deleting a basis function that contributes the least to the model in each step, that is, the basis function that increases the MSE the least (based on the least squares criterion), and deleting the basis function that causes the model to overfit. The pruning process can avoid the generation of overly complex and overfitted models. At the end of this step, a series of models of different complexity and size will be generated.
[0064] Finally, the GCV criterion is used to determine whether the obtained model is the best model.
[0065] The response surface model accuracy test is used to predict the modified accuracy of the response surface model. The expression is as follows:
[0066]
[0067] Where: yRS(j) represents the jth response surface model calculation result; y(j) represents the jth finite element calculation result; N is the number of test points in the design space; y represents the average value of the finite element calculation result;
[0068] The mean square error (MSE) reflects the degree of dispersion of the test data points to the model, that is, the average value of the error of each point; the relative root mean square error (RRMSE) reflects the degree of fit of the model to the test data. The closer the relative root mean square error (RRMSE) is to 0, the better the fit of the model. A value close to 1 means that the model is almost the same as the mean of y. A high relative root mean square error (RRMSE) means that the model may have overfitted the test data. 2 The closer it is to 1, the higher the accuracy of the model.
[0069] The sequential quadratic programming algorithm (SQP) is one of the most effective methods for solving smooth constrained nonlinear optimization problems. The algorithm approximates the solution of the nonlinear programming problem by solving a series of quadratic programming (QP) sub-problems. The basic processing of the SQP algorithm includes a series of problems such as KKT conditions, conversion to QP problems, application of Lagrangian functions, quadratic approximation, handling inequality constraints, initial value insensitivity, etc. It can effectively handle problems with equality and inequality constraints, determine the direction of descent through step-by-step quadratic programming problems, gradually reduce the step size, and repeat these steps to obtain the optimal solution that approximates the original problem. The quadratic sequential programming algorithm is used to solve the objective function, thereby obtaining the optimal solution for using materials while meeting structural safety conditions.
[0070] Example:
[0071] The turnout beams consist of 4 sections with 19 spans in total, with two spans of 32.7 and 24.7m, and the total length of the box beam is 604.3m. The support adopts a ground-type steel pipe column + Bailey beam structure, which includes, from top to bottom, a transverse distribution beam (I20a I-beam), a longitudinal Bailey beam 5, a double-piece I-beam 6 (double-piece I53 model), and a 630×8mm steel pipe pile 7. A steel unloading block is provided between the double-piece I-beam 6 and the steel pipe pile 7 for unloading the support, and a 2cm thick rubber pad is laid at the bottom of the unloading block to improve the coordination of the support settlement. 20 channel steel flat joints and scissor braces are provided between the steel pipe piles 7, and 20 channel steel walls are provided at the nodes of the steel pipes on the pier side; 4 rows of steel pipe piles 7 are provided for each span of the support, some of which fall directly on the pedestal, and some of which are equipped with PHC pipe piles or reinforced concrete expanded foundations according to different geological conditions. The original design support scheme is as follows Figure 2 shown.
[0072] Steel pipe piles 7 are vertically arranged on foundation 8, double I-beams 6 are arranged on the upper end of steel pipes 7, Bailey beams 5 are fixedly connected to the double I-beams 6, distribution beams are laid at intervals above the Bailey beams 5, square timbers 3 are laid at intervals above the distribution beams, and bamboo plywood is laid above the square timbers 3.
[0073] An initial finite element model was established based on the original design bracket scheme.
[0074] In this revision, the number of Bailey beams 5, the cross-sectional dimensions of square timber 3, the wall thickness of steel pipe piles 7, the distribution beam, and the model of double-jointed I-beams 6 were selected as independent variables, and the deflection and stress of Bailey beams 5, the deflection and stress of distribution beams, the deflection and stress of double-jointed I-beams 6, the stress of steel pipe piles 7, and the stress and deflection of square timber 3 at the key cross-sectional positions were selected as dependent variables. The central composite design was used for experimental design, and 41 samples were obtained. The sample data were substituted into the initial finite element model for 41 operations to obtain the sample data. Some experimental designs are as follows: Figure 3 shown.
[0075] The corresponding program was written using MATLAB software. According to the correction process introduced above, the multivariate adaptive spline regression algorithm was used for regression to obtain 9 response surface models. Figure 4 The precision test values of the nine response surface models obtained by fitting are given. Figure 4 It can be seen that the R2 values of the three fitted response surface models are all close to 1, indicating that the fitting accuracy is high, the relative root mean square error RRMSE is close to 0, and the model fitting degree is high. Therefore, the fitted adaptive response surface model can be used to replace the initial finite element model for correction (see Figure 4 ).
[0076] According to the Bailey beam support design specifications, the double-jointed I-beam 6 before optimization is I53 model I-beam, the cross-sectional size of the square wood 3 is 100mm, the distribution beam is I20a, the diameter of the steel pipe pile 7 is 630mm, the wall thickness is 8mm, the number of Bailey beams 5 is 18 evenly distributed, and the stress on the accessories of each component before optimization is less than 0.8 times the standard value of the component stress strength, and the deflection γ of the component is ≤0.8×L / 400 (L is the component length), which meets the requirements of the specification.
[0077] Depend on Figure 6 It can be seen that the multi-objective optimization function is established through the response surface method, and the optimization objective function is constructed by combining the structural response limit and the response surface model, and the optimization objective function is constructed by combining the structural response limit, as shown in the following formula:
[0078]
[0079] x lk ≤x k ≤x uk
[0080] Where: x k is the design parameter to be modified; x lk and x uk is the design space range of structural parameters; f RSM is the test value; f 响应限值 The allowable value of each component response is determined in combination with the specifications.
[0081] If the optimization objective function meets the given convergence condition: When the objective function is less than 0.001, the iterative operation is terminated, and the objective function is solved by the quadratic sequence programming algorithm to obtain the optimal solution for using materials under the condition of structural safety; if the given convergence condition is not met, the test value is modified, and the next round of finite element calculation and iterative solution is performed until the convergence condition is met;
[0082] The sequential quadratic programming algorithm is used to optimize the inversion and select the best solution, such as Figure 5 As shown in the figure, the number of Bailey beams 5 is reduced from 18 to 10 and arranged evenly, the diameter of the steel pipe pile 7 remains unchanged, the wall thickness is reduced from 8mm to 5mm, the distribution beam model is changed from I20a to I18, and the double I-beam 6 model is changed from I53 to I40c. The optimized finite element model is established using Midas finite element software for force analysis. The results are shown in the figure. Figure 7 As shown. Figure 7It can be seen that the maximum stress value of the optimized support structure is less than 0.8 times the standard value of the component material stress strength, and the deflection γ of the component is also less than 0.8×L / 400, which meets the requirements of structural design optimization. In summary, the use of the proposed algorithm for structural optimization design reduces construction costs and shortens construction period while ensuring structural reliability and safety.
[0083] The above-mentioned embodiments only express the specific implementation of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.
Claims
1. A method for optimizing the design of turnout beam supports based on response surface model correction, characterized in that: The following steps are involved: S1. Determine the structural design parameters and structural response range of the turnout beam support, and use central composite experimental design for experimental design; S2. According to the experimental design plan, the initial finite element model of the turnout beam support is established to perform mechanical performance analysis and obtain the structural response corresponding to each group of structural design parameters; S3, fitting the functional relationship between the structural design parameters and the structural response by multivariate adaptive regression spline to obtain the response surface model of each response feature; S4. Perform accuracy tests on each response surface model to obtain several adaptive response surface models that pass the accuracy tests. Combine the structural response limits to construct the optimization objective function, and use the sequential quadratic programming algorithm to optimize the inversion and obtain the optimal design values of the structural design parameters.
2. The method for optimizing the design of a turnout beam support based on response surface model correction according to claim 1, characterized in that: The turnout beam support includes cross-sectional dimensions and material performance ranges of square timber (3), I-beam (4), Bailey beam (5), double-jointed I-beam (6), and steel pipe pile (7), and a central composite design is used to perform experimental design to obtain multiple groups of structural design parameter combinations.
3. The method for optimizing the design of a turnout beam support based on response surface model correction according to claim 2 is characterized in that: Finite element software is used to establish a finite element model of the turnout beam support. Each set of structural design parameters in the experimental design is substituted into the finite element model for calculation to calculate the corresponding deflection and stress of the Bailey beam, the deflection and stress of the distribution beam, the deflection and stress of the double-jointed I-beam, the stress of the steel pipe pile, and the stress and deflection response of the square timber.
4. The method for optimizing the design of a turnout beam support based on response surface model correction according to claim 1, characterized in that: Taking the key dimensions and material characteristics of the main components as variables and the structural response characteristics as dependent variables, a response surface model reflecting the relationship between the structural design parameters and the structural response is fitted using multivariate adaptive regression splines. The expression of the multivariate adaptive regression splines is as follows: In the formula, is a constant; B m (x) is the mth basis function, which is the tensor product of the spline function; is the coefficient of the mth basis function; M is the number of model terms; B m (x) is the basis function; S km The value is ±1; X v(k,m) represents the v(k,m)th component of the prediction factor x, t km is a point, [(x v(k,m) -t km )] + is a semi-cut polynomial.
5. The method for optimizing the design of a turnout beam support based on response surface model correction according to claim 1, characterized in that: The expression of the precision test in S4 is as follows: Where: y RS (j) represents the jth response surface model calculation result; y(j) represents the jth finite element calculation result; N is the number of test points in the design space; Represents the average value of the finite element calculation results; The mean square error (MSE) reflects the degree of dispersion of the test data points to the model, that is, the average value of the error of each point; The relative root mean square error RRMSE reflects the degree of fit of the model to the test data. The closer the relative root mean square error RRMSE is to 0, the better the fit of the model. A value close to 1 means that the model is almost the same as the mean of y. A high relative root mean square error (RRMSE) means that the model may have overfit the test data. 2 The closer it is to 1, the higher the accuracy of the model.
6. The method for optimizing the design of a turnout beam support based on response surface model correction according to claim 1, characterized in that: The optimization objective function is constructed in combination with the structural response limit, as shown in the following formula: s.t x lk ≤x k ≤x uk Where: x k is the design parameter to be modified; x lk and x uk is the design space range of structural parameters; f RSM is the test value; f 响应限值 The allowable value of each component response is determined in combination with the specifications.
7. The method for optimizing the design of a turnout beam support based on response surface model correction according to claim 6, characterized in that: The quadratic sequence programming algorithm is used to solve the objective function. When the optimization objective function meets the given convergence conditions, the optimal solution of the structural design parameters under the structural safety conditions is obtained. If the given convergence conditions are not met, the next round of finite element calculation and iterative solution is performed until the convergence conditions are met.
8. A computer program product, characterized in that The computer program product includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute a method for optimizing the design of a turnout beam support based on response surface model correction as described in any one of claims 1 to 7.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method for optimizing the design of a turnout beam support based on response surface model correction as described in 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 program is executed by a processor, the method for optimizing the design of a turnout beam support based on response surface model correction as described in any one of claims 1 to 7 is implemented.
