A forward design method for machine tool structural parts based on triangular element structure

Abstract

The present invention discloses a forward design method for machine tool structural components based on a trident structure. The method comprises: using the trident structure as the basic structural unit for the internal ribs of the structural component to be designed; constructing the internal rib structure of the structural component to be designed according to the gradual growth pattern of the trident structure, establishing a mathematical model and flow chart for the growth, establishing a parametric model of the structural component to be designed using three-dimensional software, and generating a number of model samples; randomly sampling the model samples using a Latin hypercube algorithm, and analyzing the static and dynamic characteristics of the model samples using ANSYS software; training the static and dynamic characteristic analysis results of the model samples using a BP neural network to construct a BP neural network prediction model; and cyclically approximating the BP neural network prediction model using a genetic algorithm, comparing it with a traditional element structure form, and determining the optimal size within the design size range of the structural component to be designed. The present invention can improve the static and dynamic characteristics of the structure, save a large amount of manpower and material resources, and facilitate processing and manufacturing.

Description

Technical Field

[0001] The present invention belongs to the field of mechanical engineering, and in particular relates to a forward design method for machine tool structural parts based on a trident structure, in particular to the forward design of a bed structural part. Background Art

[0002] At present, my country's machine tool industry has a low independent innovation and design capability, and lacks a complete and efficient design method for the forward design of machine tools. The earliest machine tool structure design mainly relied on traditional experience and imitation of foreign machine tools. The quality of machine tool performance mainly depends on the experience and level of the designer. The cycle is long, the efficiency is low, and resources are wasted. There are many simplifications in the design and the accuracy is low.

[0003] Most existing machine tools use traditional rib plate styles, mainly concentrated in the shape of a field, a well, and a rice, making it difficult to obtain the optimal design solution. The rib plate design method mainly relies on simplified mechanical model analysis and the experience of designers. The final solution is obtained through repeated modifications. The entire design process cannot be automated, resulting in low design efficiency and difficulty in handling design problems of complex structures. Alternatively, the existing structure can be improved and topology optimization software can be used to complete the topology optimization of the internal rib plate structure, remove pseudo-density materials, and realize model reconstruction. However, the optimized model obtained by this method has many irregular shapes, which is not convenient for actual processing and production, and requires subsequent simplification before it can be used in the manufacture of actual machine tools.

[0004] To address the above problems, an efficient structural component forward design method is needed to design the bed structure in order to improve the static and dynamic characteristics of the bed. Summary of the Invention

[0005] The purpose of the present invention is to overcome the deficiencies in the prior art and provide a new forward design method for machine tool structural parts. A three-pronged element structure is used to grow and fill the entire structural space. While improving the static and dynamic characteristics of the structural parts, it reduces the processing difficulty, improves the design efficiency, and provides a basis for the forward design of machine tool structural parts.

[0006] The purpose of the present invention is achieved through the following technical solutions:

[0007] A forward design method for machine tool structural components based on a trident element structure specifically includes the following steps:

[0008] S1. Use the trident structure as the basic structural unit of the internal rib plate of the structural part to be designed;

[0009] S2. Using the trident structure to gradually grow, construct an internal rib plate structure for the structural component to be designed, where the internal rib plate structure is constrained by the external dimensions of the structural component to be designed; establishing a mathematical growth model and a growth flow chart based on the trident structure's gradual growth pattern; creating a parametric model of the structural component to be designed using 3D software, and generating several model samples based on the size range of the structural component to be designed;

[0010] S3. Randomly sample the model samples using the Latin hypercube algorithm to extract N model samples; generate .x_t format files for ANSYS analysis using 3D software for the N model samples; and perform static and dynamic characteristic analysis of the N model samples using ANSYS to obtain the maximum deformation value, maximum deformation value, minimum deformation value, and first six natural frequencies of the N model samples;

[0011] S4. Take the dimensional design variables of the structural parts to be designed as the neurons of the input layer, and take the maximum deformation value of the structural parts to be designed, the maximum deformation value of the guide rail base surface, the minimum deformation value of the guide rail base surface and the first six natural frequencies as the neurons of the output layer. After several repeated experiments, the number of neurons in the hidden layer is determined. The maximum deformation value, the maximum deformation value, the minimum deformation value and the first six natural frequencies of N model samples are trained using a BP neural network to obtain a BP neural network prediction model that meets the error requirements.

[0012] S5. Based on the genetic algorithm, the BP neural network prediction model is cyclically approximated to obtain the optimal size within the size range of the structural part to be designed and complete the final design.

[0013] Furthermore, in step S2, an ordered set A consisting of the growth angles of the trident structure, an ordered set L consisting of the growth lengths of the trident structure, and an ordered set H consisting of the heights of the trident structure are set; and a growth mathematical model is established as follows:

[0014]

[0015] Among them, α in set A i , β i , γ i are the three direction angles at the i-th node of the internal ribbed plate structure; m represents the total number of nodes in the internal ribbed plate structure; l g , t r 、h a are the growth length, thickness, and height of the trident structure respectively; n represents the total number of trident structures; T t Contains the set of internal rib structure thickness; T e Contains a set of shell thicknesses of the structural parts to be designed; the shell thicknesses of the structural parts to be designed include t b , t s and tf :t b is the side thickness of the shell of the structural component to be designed, t s is the thickness of the supporting surface of the shell of the structural part to be designed, t f is the thickness of the functional surface of the shell of the structural part to be designed; f(A, L, T t ,T e ,H) is the design target: the static and dynamic characteristics of the structural part to be designed; W(A,L,T t ,T e ,H) is the weight function related to the size variable elements of the structural part to be designed, W min The lower limit of the weight of the structural component to be designed, W max The upper limit of the weight of the structural parts to be designed; D(A, L, T t ,T e ,H) represents the design set composed of the external contour dimensions of the structural parts to be designed; U,D0 is the set composed of all external contour dimensions of the structural parts to be designed within the design requirements.

[0016] The growth thickness and length of the trifurcated element structure, as well as the functional surface, support surface, and side thickness of the shell of the structural part to be designed, are determined as driving dimensional parameters. The internal rib structure of the structural part to be designed is considered to be gradually grown from a number of bifurcation points and bifurcation units, and the growth is completed through a growth mathematical model. A growth flow chart is established, and secondary development is performed using SolidWorks to simulate the growth process and establish a parametric model of the structural part to be designed.

[0017] The bifurcation nodes of the internal rib structure have two different growth modes, namely synchronous growth and asynchronous growth; the angle vector of each growth point is:

[0018]

[0019] Where α1, β1, and γ1 are the initial bifurcation point angles, α i , β i , γ i are the three direction angles at the i-th node of the internal rib plate structure; is the angle increment of each growth point; SolidWorks is used for secondary development to establish a parametric model of the structural part to be designed; the size variables of the structural part to be designed are arbitrarily changed and a model sample of the structural part to be designed is quickly generated, and several model samples are generated.

[0020] Furthermore, in step S3, random sampling is implemented using MATLAB to extract N = 1000 model samples; in the static and dynamic characteristic analysis, a new material with the same density, elastic modulus, and Poisson's ratio as the material of the structural part to be designed is added to the workbench; the 1000 model samples are imported into the workbench and the new material properties are assigned to the model samples, the grid is divided, and according to the actual working conditions, fixed constraints are applied to the supports of the structural part to be designed, and gravity acceleration is added to realize the gravity application and the load at the guide rail base surface position, and then the static and dynamic solutions are performed, and finally the static deformation results and modal analysis results are extracted.

[0021] Furthermore, in step S4, the 1000 model samples are divided into a calculation prediction set and a verification set in proportion; three groups of BP neural network prediction models are established, the first group is the three-element structure growth length, growth thickness and the shell thickness of the structural part to be designed as the input layer neurons, and the maximum deformation value of the structural part to be designed is the output layer neurons; the second group is the three-element structure growth length, growth thickness and the shell thickness of the structural part to be designed as the input layer neurons, and the maximum and minimum deformation values of the guide rail base surface are the output layer neurons; the third group is the three-element structure growth length, growth thickness and the shell thickness of the structural part to be designed as the input layer neurons, and the first six-order natural frequency values of the structural part to be designed are the output layer neurons; finally, the BP neural network prediction model results are compared with the ANSYS calculation results; and a BP neural network prediction model that meets the error requirements is obtained.

[0022] Furthermore, in step S5, a cyclic approximation is performed on the BP neural network prediction model based on a genetic algorithm. When using a genetic algorithm, a fitness function needs to be selected. The fitness function is taken as y = abs(sim(net)). Under the constraint of material consumption, the size optimization mathematical model of the structural part to be designed is shown as follows, with the maximum deformation value of the structural part to be designed, the deformation difference of the guide rail base surface, and the first six frequencies as targets:

[0023]

[0024] Where, sim is the simulation function of the BP neural network prediction model; net is the created BP neural network prediction model; Φ1(X) and Φ2(X) are the objective functions; ω1 and ω2 are the weighted coefficients of the maximum deformation value of the structural part to be designed and the deformation difference of the guide rail base surface, respectively; Δd(X) is the maximum deformation value of the structural part to be designed, δ(X) is the deformation difference of the guide rail base surface of the structural part to be designed; W(X) is the weight function related to the dimensional design variables of the structural part to be designed; f j is the jth natural frequency value of the structural part to be designed; X is the set of dimensional variables of the structural part to be designed; l g is the growth length of the trifurcated structure, t ris the growth thickness of the trident structure, t b is the side thickness of the shell of the structural component to be designed; t s is the thickness of the supporting surface of the shell of the structural component to be designed; t f is the thickness of the functional surface of the shell of the structural part to be designed; the BP neural network prediction model is cyclically approximated using a genetic algorithm; under the same quality, the structural parts with a three-pronged element structure for the internal rib plate structure are compared and evaluated with the structural parts with a traditional element structure to obtain the optimal size within the size range of the structural part to be designed and complete the final design.

[0025] Furthermore, the growth process of the trident structure in step S2 is as follows:

[0026] (201) Initializing the parameters of the size variables of the structural component to be designed and selecting the initial growth point;

[0027] (202) An initial trifurcated structure is formed at the initial growth point, and the initial trifurcated structure is defined as forming three primary branches;

[0028] (203) The ends of the three primary branches grow in a step-by-step bifurcation manner, and each branch end point is judged in turn to see if it is a bifurcation point. If so, two next-level branches are generated under the corresponding branch end direction angle; if not, it is a junction point formed by the combination of the two branch end points, and the next-level branch is generated under the junction point direction angle; in the step-by-step bifurcation growth process, both the bifurcation point and the junction point are defined as the sub-step end point;

[0029] (204) Determine whether the end point of the sub-step is a boundary point, i.e., a point on the shell of the structural part to be designed; if so, end the growth; if not, return to step (203).

[0030] The present invention also provides an electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the forward design method for machine tool structural parts based on a trident structure are implemented.

[0031] The present invention also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the forward design method of a machine tool structural component based on a trident structure are implemented.

[0032] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0033] 1. The present invention adopts a three-pronged element structure as the basic structural unit of the internal rib plate of the structural part to be designed. The three-pronged element structure has improved deformation resistance caused by external forces and dynamic characteristics compared with the traditional element structure.

[0034] 2. The present invention constructs the internal rib plate structure of the structural part to be designed in a gradually growing mode using a trident structure, and establishes a mathematical model of growth and a growth flow chart. In traditional designs, the internal rib plate structure of the structural part is mostly designed based on experience. The present technical invention provides a mathematical theoretical basis for the design of the internal rib plate of the structural part to be designed, and establishes a parametric model of the structural part to be designed. It can quickly generate corresponding model samples within the required size range of the structural part to be designed, thereby greatly improving the modeling efficiency; the model samples are analyzed using ANSYS to obtain a large number of reliable static and dynamic characteristic results of the structural part to be designed, thereby improving the credibility of the results.

[0035] 3. The present invention adopts the Latin hypercube algorithm to perform sampling analysis on the model samples, which has the characteristic of uniform stratification and can obtain the tail model samples with less sampling. This makes the Latin hypercube sampling more efficient and reliable than the ordinary sampling method, and at the same time ensures that the structure of the model samples is relatively close to the overall structure, thereby indirectly improving the prediction accuracy of the BP neural network prediction model.

[0036] 4. This invention utilizes a BP neural network capable of implementing complex nonlinear mappings, making it particularly suitable for solving the mapping relationship between the dimensional variables of the structural component to be designed and the static and dynamic characteristics of the model sample. It also employs a genetic algorithm that directly uses the objective function value as search information. This method measures the quality of the model sample solely using the fitness function value, without requiring the derivation or differentiation of the objective function value, thus improving computational efficiency.

[0037] 5. When evaluating the final results, a comprehensive evaluation method is adopted, which uses the weighted difference between the maximum static deformation and the guide rail base deformation and the first six natural frequencies. In existing technical inventions, the dynamic characteristics of the design structure are evaluated only from the perspective of whether the first-order natural frequency has been improved. This evaluation method is rather one-sided. Others use technology and software to perform topological optimization to select and discard structures, resulting in uneven structures, while the design results of the present invention are easy to process. Under the condition of the same quality, the present invention is compared with the traditional internal rib plate structure, and finally the static and dynamic characteristics of the structural parts are verified by comparing the results, and the optimal size of the structural parts to be designed is obtained. This invention method saves a lot of manpower and material resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a flow chart of the machine tool structural component design method of the present invention;

[0039] Figure 2 Schematic diagram of the three-pronged element structure used in the present invention;

[0040] Figure 3a It is a schematic diagram of the process of synchronous growth of the trident structure; Figure 3b It is a diagram of the asynchronous growth process of the trifurcated structure;

[0041] Figure 4 It is a flow chart of the growth process of the triad structure;

[0042] Figure 5 It is a schematic diagram showing the design structure divided into two parts: the outer shell and the internal rib structure, as well as the symbol settings for the dimensions of each part;

[0043] Figures 6a to 6d These are the external drawings of the rectangular bed with four internal rib structures, among which Figure 6a It is an internal rib plate structure composed of orthogonal quadrilateral structures. Figure 6b It is a rib-plate structure composed of a mixture of four-pronged and three-pronged element structures. Figure 6c and Figure 6d There are two main forms generated by the trident structure under different main vector directions;

[0044] Figure 7 It is the deformation curve of the critical path in the length direction under four internal ribbed plate structures;

[0045] Figure 8 This is the deformation difference diagram of the guide rail base surface of four internal rib plate structures;

[0046] Figure 9a It is a distribution diagram of model samples extracted using the Latin hypercube algorithm within the size range of the structural parts to be designed; Figure 9b This is a diagram of the internal rib structure of some model samples;

[0047] Figure 10 It is the guide rail base surface deformation curve diagram of the structural component with internal rib plate structure designed with three-pronged element structure and the structural component with traditional element structure design;

[0048] Figure 11 It is a drawing of the machine tool structure;

[0049] Figure 12a and Figure 12b This is the internal rib structure diagram of the T-type bed, where Figure 12a It is a traditional internal rib plate structure style. Figure 12b It is the three-pronged element structure style adopted by the present invention;

[0050] Figure 13 It is the deformation curve diagram of the guide rail base surface of the T-type bed adopting the traditional element structure and the three-pronged element structure;

[0051] Figure markings: 1-joining point, 2-bifurcation point, 3-structural part shell, 4-internal rib structure, 5-T-type bed, 6-column, 7-slide, 8-spindle box, 9-turntable box, 10-turntable table. DETAILED DESCRIPTION

[0052] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0053] As attached Figure 1 As shown, the present invention is a method for designing machine tool structural parts based on a trident structure, which includes the following main steps:

[0054] Step 1

[0055] First, the internal ribs of machine tool components are considered to be composed of a single element structure. Traditional internal ribs are mostly in a crisscross pattern. Inspired by nature, this invention uses a triangular element structure. Numerous natural hexagonal structures exist in nature. The smallest indivisible structure within a hexagonal structure is a triangular element structure. A hexagonal structure can be considered a stack of triangular elements, which serve as the basic structural unit for the internal ribs of the designed component.

[0056] The mechanical model of the trifurcated structure was extracted and the Young's modulus in the x and y directions was compared to obtain formula (1):

[0057]

[0058] Among them, E x 、E y are the equivalent Young's modulus in x and y directions respectively; v a 、v b are the equivalent Poisson's ratios in the two cases when subjected to uniform loads in the x and y directions, respectively, and E is the Young's modulus of the material; l g is the growth length of the trifurcated structure; t r is the growth thickness of the trident structure. From formula (1), it can be concluded that when a uniform trident structure is used, the difference in the properties of the materials in the x and y directions is very small, so this structure can be used to simplify the extraction of the equivalent mechanical model of the structural component. The trident structure used is as follows: Figure 2 As shown in the figure, the angle between adjacent forks is 120°, and the length and thickness of the forks are consistent. The internal rib plate structure of the structural part to be designed is constructed by growing the three-pronged element structure.

[0059] Step 2

[0060] The internal rib structure of the structural part to be designed is constructed using a three-pronged element structure in a gradual growth pattern. The internal rib structure is limited by the external contour size of the structural part to be designed. Based on the gradual growth pattern of the three-pronged element structure, a growth mathematical model and growth flow chart are established. A parametric model of the structural part to be designed is created using 3D software, and several model samples are generated based on the size range of the structural part to be designed.

[0061] In the process of constructing a growth mathematical model using a trident structure, it is necessary to find the optimal growth vector angle and optimal thickness. The following key topological elements are selected to form a set, assuming that the ordered set A is composed of the growth angles of the trident structure, the ordered set L is composed of the growth lengths of the trident structure, and the ordered set H is composed of the heights of the trident structure. The growth mathematical model is established as follows:

[0062]

[0063] Among them, α in set A i , β i , γ i are the three direction angles at the i-th node of the internal stiffener plate structure; m represents the total number of nodes in the internal stiffener plate structure; l g , t r 、h a are the growth length, thickness, and height of the trident structure respectively; n represents the total number of trident structures; T t Contains the set of internal rib structure thickness; T e Contains a set of shell thicknesses of the structural parts to be designed; the shell thicknesses of the structural parts to be designed include t b , t s and t f :t b is the side thickness of the shell of the structural component to be designed, t s is the thickness of the supporting surface of the shell of the structural part to be designed, t f is the thickness of the functional surface of the shell of the structural part to be designed; f(A, L, T t ,T e ,H) is the design target: the static and dynamic characteristics of the structural part to be designed; W(A,L,T t ,T e ,H) is the weight function related to the size variable elements of the structural part to be designed, W min The lower limit of the weight of the structural component to be designed, W max The upper limit of the weight of the structural parts to be designed; D(A, L, T t ,T e ,H) is the design set composed of the external contour dimensions of the structural parts to be designed generated by this technical method; U,D0 is the set composed of all external contour dimensions of the structural parts to be designed within the design requirements.

[0064] The growth thickness, growth length, functional surface, support surface and side thickness of the shell of the structural part to be designed are determined as the main driving dimensional parameters. The internal rib plate structure of the structural part to be designed is regarded as gradually growing from multiple bifurcation points and bifurcation units, and a growth model is established to complete the growth. A growth flow chart is established, and SolidWorks is used for secondary development to simulate the growth process and establish a parametric model of the structure of the structural part to be designed. There are two different growth modes for the bifurcation nodes of the internal rib plate structure, which are divided into synchronous growth and asynchronous growth. The angle vector of each growth point is:

[0065]

[0066] Where α1, β1, and γ1 are the initial bifurcation point angles, α i , β i , γ i are the three direction angles at the i-th node of the internal rib plate structure; The angle increment for each growth point. Use SolidWorks for secondary development to establish a parametric model of the structural part to be designed. Arbitrarily change the dimensional variables of the structural part to be designed and quickly generate model samples of the structural part to be designed, generating several model samples.

[0067] Specifically, a three-branch element structure is used to determine the growth mode of the internal rib structure, which is divided into synchronous growth and asynchronous growth. Synchronous growth means that each bifurcation unit grows once in each growth sub-step, and the effect is as follows: Figure 3a As shown; asynchronous growth is when the local bifurcation unit grows once and is not synchronized with other bifurcation units, the effect is as follows Figure 3b As shown. Figure 3a The middle junction point 1 is the end point where the two forks of the previous growth sub-step meet, and only one fork is generated when the junction point grows. Figure 3b The next growth sub-step at bifurcation point 2 will produce two bifurcations. The point to determine whether the growth stops is the boundary point. When the trifurcated structure bifurcation reaches the boundary area, the growth stops. The flow chart for setting the trifurcated structure growth is as follows: Figure 4 shown. Figure 5 In order to define the dimensions of each structural part, a complete structural part can be divided into two parts: Figure 5 The structural shell 3 and the internal rib plate structure 4 grown from the trident structure can be seen.

[0068] The external dimensions are driven by the design factors, and the geometric relationship is used to obtain the dimensions of the structural parts to be designed:

[0069]

[0070] The relationship between the structural dimensions of the structural member to be designed and the structural dimensions of the internal ribs:

[0071]

[0072] The size symbols involved in the formula are Figure 5 The corresponding can be found in, where l g is the growth length of the trifurcated structure, t r is the growth thickness of the trident structure, l a is the length of the internal rib structure, w a is the width of the internal rib structure, N h is the number of rows of internal stiffener structure units, h a is the height of the internal rib structure; l t is the length of the shell of the structural part to be designed, w t h is the width of the shell of the structural part to be designed, t is the height of the shell of the structural part to be designed, t f is the thickness of the functional surface (the upper surface of the functional components such as the mounting rail), t s is the thickness of the support surface (the bottom surface of the installation support), t b is the side thickness of the shell of the structural part to be designed, and the number of horizontal rows of ribs (i.e., a hexagonal structure composed of three trident structures) is N. h , k = 1, 2, 3…; k + 1 equals the total number of vertical rows of reinforcement. Based on the above growth flow chart, SolidWorks secondary development was used to establish a parametric model. New 3D models could be generated by inputting any dimensions, eliminating the need for repetitive design. Several model samples were generated based on the size range of the structural component to be designed.

[0073] Step 3

[0074] Using the Latin Hypercube algorithm in MATLAB, a random sampling of 1000 model samples was performed on the parametric model. Three-dimensional software was used to generate .x_t files for ANSYS analysis from these 1000 model samples. ANSYS was used to analyze the model samples' static and dynamic characteristics, obtaining the maximum deformation of the model samples, the maximum and minimum deformation of the guide rail base surface, and the first six natural frequencies. A new material was added to the workbench, which should have the same density, elastic modulus, and Poisson's ratio as the material of the structural component to be designed. The model samples were imported into the workbench and assigned material properties. The mesh was then created. Based on the actual operating conditions, fixed constraints were applied to the supports of the structural component to be designed. Gravitational acceleration was also added to achieve gravity and loads at locations such as the guide rail base surface. Static and dynamic analysis was then performed, and finally, static deformation results and modal analysis results were extracted.

[0075] This embodiment takes a rectangular parallelepiped bed as an example for specific implementation, and compares the traditional structural form of a T-shaped bed of a certain type of machine tool with the design form of this embodiment.

[0076] The triangular element structure is preliminarily compared with other traditional element structures. The traditional stiffened plate structure is regarded as a structure composed of orthogonal quadrilateral elements. Figure 6a As shown, the internal rib plate structure composed of a mixture of orthogonal quadrilateral structure and triangular element structure is as follows Figure 6b As shown, the two main forms generated by the trident structure under different main vector directions are as follows Figure 6c and Figure 6d shown. Figures 6a to 6d The four structures are of the same quality. ANSYS Workbench module is used to compare and analyze the four structures. Figures 6a to 6d The middle structure simulates the shape of a rectangular bed, with a length of 2500mm, a width of 1500mm, and a height of 800mm. When implementing the solution of the present invention on the bed, the bed height is set to 800mm. The .x_t format files of the four structures are imported into the workbench, and then a material with the same properties as the bed material is set, with an elastic modulus of 130GPa and a density of 7.30×10 -9 The weight of the model is 30 mm, and the Poisson's ratio is 0.25. The model sample is then meshed with a grid size of 30 mm. The four corner bosses at the bottom of the bed are set as fixed constraints. The bed is subjected to vertical downward gravity acceleration. The static deformation of the structure under gravity is analyzed, and the key paths of different structures are extracted to simulate the deformation of the base surface of the guide rail. Curve interpolation and function fitting are performed in MATLAB to obtain the following: Figure 7 The deformation trend of the guide rail base surface can be observed by the critical path deformation curve. The maximum deformation value of the structural member with the trident structure is the smallest, and then the critical path deformation difference is compared. Figure 8 As shown in the figure, the overall deformation diagrams under four internal rib plate structures and the deformation diagram of the upper surface of the extracted installation rail are included. It can be concluded that the critical path deformation difference is the smallest when the internal rib plate structure is a three-pronged element structure.

[0077] Using the established parametric model, Latin hypercube sampling was performed within the size range of the structural parts to be designed. The size range of the structural parts to be designed is shown in Table 1. 1000 model samples were extracted. The distribution of the model samples and the schematic diagram of the internal stiffener structure of some model samples are shown in Figure 9.

[0078] Table 1 Size range of the structural parts to be designed

[0079]

[0080] The extracted model samples are subjected to finite element analysis to obtain the static and dynamic characteristics of the model samples.

[0081] Step 4

[0082] A BP neural network prediction model is constructed based on the static and dynamic characteristic analysis results of the model samples selected in step 3. The dimensional design variables of the structural part to be designed are used as neurons in the input layer, and the maximum deformation value of the structural part to be designed, the maximum and minimum deformation values of the guide rail base surface, and the first six natural frequencies of the structural part are used as neurons in the output layer. The number of neurons in the hidden layer is determined through repeated experiments. The maximum deformation value, the maximum deformation value, the minimum deformation value, and the first six natural frequencies of N model samples are used to train the BP neural network to obtain a BP neural network prediction model that meets the error requirements.

[0083] In this example, 1,000 model samples were divided proportionally into a prediction set and a validation set. Three sets of prediction models were established: the first set used the growth length, growth thickness, and shell thickness of the trident structure as input neurons, and the maximum deformation value of the structure as output neurons; the second set used the growth length, growth thickness, and shell thickness of the trident structure as input neurons, and the maximum and minimum deformation values of the guide rail base as output neurons; the third set used the growth length, growth thickness, and shell thickness of the trident structure as input neurons, and the first six natural frequencies of the structure as output neurons. Finally, the prediction results of the BP neural network prediction model were compared with the ANSYS calculation results to obtain a BP neural network prediction model that met the error requirements.

[0084] In this embodiment, a three-layer BP neural network prediction model was used for model prediction. After repeated experiments, the number of hidden layer neurons was determined to be 10. The training errors of the three corresponding BP neural network prediction models were 1.9283E-6, 3.7277E-7, and 3.0947E-2, respectively; the correlation coefficients R were 0.986, 0.9386, and 0.98661, respectively. The BP neural network prediction model is convincing. The predictions of some model samples using the obtained BP neural network prediction model were compared with finite element calculations as shown in Tables 2 and 3 below:

[0085] Table 2. BP neural network prediction model test of the maximum deformation value of randomly selected model samples

[0086]

[0087] Table 3 BP neural network prediction model test of the first six natural frequencies of randomly selected model samples

[0088]

[0089] Step 5

[0090] Based on the genetic algorithm, the BP neural network prediction model is cyclically approximated to obtain the optimal size within the size range of the structural part to be designed, and the final design is completed.

[0091] The fitness function is y = abs(sim(net)). Under the constraint of material consumption, the optimization mathematical model with the maximum deformation value of the structural component to be designed, the deformation difference of the guide rail base surface, and the first six frequencies as the targets is shown in the following formula:

[0092]

[0093] Where, sim is the simulation function of the BP neural network prediction model; net is the created BP neural network prediction model; Φ1(X) and Φ2(X) are the objective functions; ω1 and ω2 are the weighted coefficients of the maximum deformation value of the structural part to be designed and the deformation difference of the guide rail base surface, respectively; Δd(X) is the maximum deformation value of the structural part to be designed, δ(X) is the deformation difference of the guide rail base surface of the structural part to be designed; W(X) is the weight function related to the dimensional design variables of the structural part to be designed; f j is the jth natural frequency value of the structural part to be designed; X is the set of dimensional variables of the structural part to be designed; l g is the growth length of the trifurcated structure, t r is the growth thickness of the trident structure, t p is the shell thickness of the structural part to be designed; the BP neural network prediction model is cyclically approximated using a genetic algorithm; under the same quality, the structural parts with a three-pronged element structure for the internal rib plate structure are compared and evaluated with the structural parts with a traditional element structure to obtain the optimal size within the size range of the structural part to be designed and complete the final design.

[0094] On the basis of the BP neural network prediction model, the bed size was optimized using the cyclic approximation optimization technology of the genetic algorithm. The crossover probability was set to 0.4, the mutation probability was 0.05, and the population size for each iteration was 400. After 5 iterative operations, the optimal fitness value was obtained: 250mm-30mm-30mm. The obtained size was used for design and simulation comparison, and compared with the traditional internal rib plate structure under the same quality. In the case of a rectangular bed, the maximum deformation of the traditional internal rib plate structure is 4.6240μm, the maximum deformation of the guide rail base is 4.4393μm, and the minimum deformation is 2.1856μm. The maximum deformation of the optimized internal rib plate structure is 3.8140μm, the maximum deformation of the guide rail base is 3.8135μm, and the minimum deformation is 1.5811μm. The deformation curve comparison of the critical path of the guide rail base is as follows. Figure 10 The comparison of the first six natural frequencies of the bed before and after optimization is shown in Table 4. From the results in the table, it can be seen that the natural frequencies of each order are improved compared with the traditional structure.

[0095] Table 4 Comparison of the first six natural frequencies of the bed before and after optimization

[0096]

[0097] This method can be used to design a certain type of horizontal machining center structure. Figure 11 As shown, it includes a T-shaped bed 5, a column 6, a slide 7, a spindle box 8, a turntable box 9, and a turntable table 10. The internal rib structure of the bed is redesigned by the method of the present invention. Figure 12b As shown in the figure, the length of the trident structure is 240mm, and the thickness of the trident structure and the machine tool outer wall is 25mm. Figure 12a As shown, under the same three-point support conditions, the guide rail base deformation curves of the two structures are compared as shown in Figure 13 As shown in the figure, it can be clearly concluded that the static characteristics of the design method of the present invention are improved by the maximum deformation value and the deformation curve trend of the guide rail base surface. At the same time, the comparison of the first six natural frequency values is shown in Table 5 below. Under the same order, the natural frequency value of the model implemented by the method of the present invention is improved.

[0098] Table 5 Comparison of low-order natural frequencies of the T-type bed before and after optimization

[0099]

[0100] In summary, under the same quality conditions, the static and dynamic characteristics of the structural parts designed using the method of the present invention are improved. Moreover, the design method is easy to operate, saves manpower and material resources, does not require repeated design improvements, and reduces design costs.

[0101] Finally, it should be noted that the above examples are intended only to illustrate the calculation process of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the above examples, those skilled in the art will appreciate that the calculation process described in the above examples may be modified or some parameters may be replaced with equivalents. Such modifications or replacements do not deviate from the spirit and scope of the calculation method of the present invention.

[0102] The present invention is not limited to the embodiments described above. The above description of the specific embodiments is intended to describe and illustrate the technical solutions of the present invention. The above specific embodiments are merely illustrative and not restrictive. Without departing from the scope of the present invention and the scope of protection of the claims, those skilled in the art may make various specific modifications based on the teachings of the present invention, all of which fall within the scope of protection of the present invention.

Claims

1. A forward design method for machine tool structural parts based on a trident element structure, characterized in that: The specific steps are as follows: S1. Use the trident structure as the basic structural unit of the internal rib plate of the structural part to be designed; S2. The trident structure is constructed in a gradually growing pattern to form an internal rib plate structure of the structural member to be designed, wherein the internal rib plate structure is limited by the external contour size of the structural member to be designed; Based on the gradual growth pattern of the trifurcated element structure, a growth mathematical model and a growth flow chart are established; a parametric model of the structural component to be designed is established using 3D software, and several model samples are generated based on the size range of the structural component to be designed; S3. Randomly sample the model samples using the Latin hypercube algorithm to extract N model samples; generate .x_t format files for ANSYS analysis using 3D software for the N model samples; and perform static and dynamic characteristic analysis of the N model samples using ANSYS to obtain the maximum deformation value, maximum deformation value, minimum deformation value, and first six natural frequencies of the N model samples; S4. The dimensional design variables of the structural component to be designed are used as the input layer neurons, and the maximum deformation value of the structural component to be designed, the maximum deformation value of the guide rail base surface, the minimum deformation value of the guide rail base surface, and the first six natural frequencies are used as the output layer neurons. After repeated experiments, the number of neurons in the hidden layer is determined. The maximum deformation value, the maximum deformation value, the minimum deformation value, and the first six natural frequencies of N model samples are used to train the BP neural network to obtain a BP neural network prediction model that meets the error requirements. S5. Based on the genetic algorithm, the BP neural network prediction model is cyclically approximated to obtain the optimal size within the size range of the structural part to be designed and complete the final design.

2. The forward design method for machine tool structural parts based on the trident structure according to claim 1 is characterized in that: In step S2, an ordered set A consisting of the growth angles of the trident structure, an ordered set L consisting of the growth lengths of the trident structure, and an ordered set H consisting of the heights of the trident structure are set; and a growth mathematical model is established as follows: ; Among them, the set A 、 、 are the three direction angles at the i-th node of the internal stiffener structure; m represents the total number of nodes in the internal stiffener structure; 、 、 are the growth length, thickness, and height of the trifurcated structure, respectively; n represents the total number of trifurcated structures; A collection containing the thickness of the internal rib structure; A collection containing the shell thicknesses of the structural members to be designed; The shell thickness of the structural part to be designed includes 、 and : is the side thickness of the shell of the structural component to be designed, is the thickness of the supporting surface of the shell of the structural component to be designed, The thickness of the functional surface of the shell of the structural component to be designed; Design objectives: static and dynamic characteristics of the structural parts to be designed; is a weight function related to the size variable elements of the structural part to be designed, The lower limit of the weight of the structural parts to be designed, The upper limit of the weight of the structural parts to be designed; Represents a design set consisting of the external contour dimensions of the structural part to be designed; It is a set of all external contour dimensions of the structural part to be designed that are within the design requirements. The growth thickness and length of the trifurcated element structure, as well as the functional surface, support surface, and side thickness of the shell of the structural part to be designed, are determined as driving dimensional parameters. The internal rib structure of the structural part to be designed is considered to be gradually grown from a number of bifurcation points and bifurcation units, and the growth is completed through a growth mathematical model. A growth flow chart is established, and secondary development is performed using SolidWorks to simulate the growth process and establish a parametric model of the structural part to be designed. The bifurcation nodes of the internal rib structure have two different growth modes, namely synchronous growth and asynchronous growth; the angle vector of each growth point is: ; In the formula 、 、 is the initial bifurcation point angle, 、 、 are the three direction angles at the i-th node of the internal rib plate structure; 、 、 is the angle increment of each growth point; SolidWorks is used for secondary development to establish a parametric model of the structural part to be designed; the size variables of the structural part to be designed are arbitrarily changed and a model sample of the structural part to be designed is quickly generated, and several model samples are generated.

3. The forward design method for machine tool structural parts based on the trident structure according to claim 1 is characterized in that: In step S3, random sampling is implemented using MATLAB to extract N=1000 model samples. In the static and dynamic characteristic analysis, a new material with the same density, elastic modulus, and Poisson's ratio as the material of the structural part to be designed is added to the workbench. The 1000 model samples are imported into the workbench and the new material properties are assigned to the model samples. The mesh is divided, and according to the actual working conditions, fixed constraints are applied to the supports of the structural part to be designed. At the same time, gravity acceleration is added to realize the application of gravity and the load at the guide rail base surface position. Then, the static and dynamic solutions are solved, and finally the static deformation results and modal analysis results are extracted.

4. The forward design method for machine tool structural parts based on the trident structure according to claim 1 is characterized in that: In step S4, the 1000 model samples are divided into a calculation prediction set and a validation set according to the proportion; Three groups of BP neural network prediction models were established. The first group used the growth length, growth thickness and shell thickness of the trident structure to be designed as input layer neurons, and the maximum deformation value of the structural part to be designed as output layer neurons. The second group is the neurons in the input layer, which are the growth length, growth thickness and shell thickness of the structure to be designed of the trident structure, and the neurons in the output layer, which are the maximum and minimum deformation values of the guide rail base surface; The third group is the neurons in the input layer, which take the growth length, growth thickness and shell thickness of the structural part to be designed as the three-pronged element structure, and the neurons in the output layer, which take the first six natural frequency values of the structural part to be designed as the output layer. Finally, the results of the BP neural network prediction model are compared with the results of ANSYS calculation. The BP neural network prediction model that meets the error requirements is obtained.

5. The forward design method for machine tool structural parts based on the trident structure according to claim 1 is characterized in that: In step S5, the BP neural network prediction model is cyclically approximated based on the genetic algorithm. When the genetic algorithm is used, a fitness function needs to be selected. The fitness function is taken as y=abs(sim(net)). Under the constraint of material consumption, the size optimization mathematical model of the structural part to be designed is shown as follows, with the maximum deformation value of the structural part to be designed, the deformation difference of the guide rail base surface, and the first six frequencies as the targets: ; Where, sim is the simulation function of the BP neural network prediction model; net is the created BP neural network prediction model; 、 is the objective function; 、 are the weighted coefficients of the maximum deformation value of the structural component to be designed and the deformation difference of the guide rail base surface; is the maximum deformation value of the structural member to be designed, is the guide rail base surface deformation difference of the structural component to be designed; is a weight function related to the dimensional design variables of the structural component to be designed; is the jth natural frequency value of the structural part to be designed; X is the set of dimensional variables of the structural part to be designed; is the growth length of the trifurcated structure, is the growth thickness of the trident structure, is the side thickness of the shell of the structural component to be designed; is the thickness of the supporting surface of the shell of the structural component to be designed; is the thickness of the functional surface of the shell of the structural part to be designed; the BP neural network prediction model is cyclically approximated using a genetic algorithm; under the same quality, the structural parts with a three-pronged element structure for the internal rib plate structure are compared and evaluated with the structural parts with a traditional element structure to obtain the optimal size within the size range of the structural part to be designed and complete the final design.

6. The forward design method for machine tool structural parts based on the trident structure according to claim 1 is characterized in that: The growth process of the trident structure in step S2 is as follows: (201) Initialize the parameters of the size variables of the structural parts to be designed and select the initial growth point; (202) An initial trifurcated structure is formed at the initial growth point, and the initial trifurcated structure is defined as forming three primary branches; (203) The ends of the three primary branches grow in a step-by-step bifurcation manner, and each branch end point is judged in turn to see if it is a bifurcation point. If so, two next-level branches are generated at the corresponding branch end direction angle; if not, it is a junction point formed by the combination of the two branch end points, and the next-level branch is generated at the junction point direction angle. In the step-by-step bifurcation growth process, both the bifurcation point and the junction point are defined as the sub-step end point. (204) Determine whether the end point of the sub-step is a boundary point, i.e., a point on the shell of the structural part to be designed; if so, end the growth; if not, return to step (203).

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the forward design method of machine tool structural parts based on the trident structure according to any one of claims 1 to 6 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the forward design method of a machine tool structural component based on a trident structure as described in any one of claims 1 to 6 are implemented.

Images