Wing tooling flag thickness prediction method based on neural network

Abstract

The invention discloses a wing tooling flag thickness prediction method based on a neural network, and the method comprises the steps: 1, building a three-dimensional geometric model of a wing tooling flag, and obtaining a plurality of simplified models with different thicknesses and corresponding neutral format models through the simplification processing and parameter setting of the three-dimensional geometric model; step 2, performing finite element calculation on the neutral format models with different thicknesses, and performing statics solution in a mode of applying concentrated load to the neutral format models with different thicknesses so as to calculate displacement results and stress results of the wing tooling flags with different thicknesses under the concentrated load; step 3, according to the displacement result and the stress result calculated in the step 2, establishing a neural network learning rate optimization model, and obtaining an output wing tooling flag thickness through model training; and step 4, through a neural network learning rate optimization model obtained after training, obtaining a tool flag thickness prediction value under a specified displacement. In the technical scheme provided by the invention, the tool flag thickness is quickly predicted by considering factors such as material attributes of the wing tool flag, the height of the tool flag rib, the width of the tool flag rib, displacement, load and weight, so that the design period of the tool flag is shortened, the design reliability of the tool flag is ensured, and the engineering design cost is reduced; the method is a rapid and effective wing tool flag thickness prediction method.

Description

Technical Field

[0001] The invention relates to the field of tooling structure design technology, but is not limited to it, and in particular to a method for predicting the thickness of a wing tooling flag based on a neural network.

[0002] Specifically, it is a method to quickly predict the thickness of wing tooling flags through artificial neural networks in order to optimize the design. Background Art

[0003] During the wing assembly process, insufficient thickness of the tooling flag on the assembly tooling will cause the wing to deform beyond tolerance and reduce the assembly accuracy. In order to improve the assembly quality, the thickness of the tooling flag needs to be optimized.

[0004] Traditionally, the thickness of the tooling flag is determined by manual experience, lacking analysis of the thickness of the tooling flag and wing deformation, resulting in inaccurate design values. Using finite element method to calculate the thickness of the wing tooling flag takes a long time and is difficult to operate, requiring a large number of iterative calculations for different design schemes. Summary of the invention

[0005] Purpose of the invention: In order to solve the above-mentioned technical problems, an embodiment of the present invention provides a wing tooling flag thickness prediction method based on a neural network, so as to solve the problem of inaccurate design values ​​obtained by artificial experience design method in the existing design of wing tooling flag thickness, as well as the time cost of finite element calculation method, high operation difficulty, and the need for a large number of iterative calculations.

[0006] Technical solution of the present invention: In a first aspect, an embodiment of the present invention provides a method for predicting the thickness of a wing tooling flag based on a neural network, comprising:

[0007] Step 1: Establish a three-dimensional geometric model of the wing tooling flag, and obtain multiple simplified models with different thicknesses and their corresponding neutral format models by simplifying and setting parameters of the three-dimensional geometric model;

[0008] Step 2: Perform finite element calculation on neutral format models of different thicknesses, and perform statics solution by applying concentrated loads in neutral format models of different thicknesses to calculate the displacement results and stress results of wing tooling flags of different thicknesses under concentrated loads;

[0009] Step 3: According to the displacement results and stress results calculated in step 2, a neural network learning rate optimization model is established, and the output wing tooling flag thickness is obtained through model training;

[0010] Step 4: The predicted value of the thickness of the tooling flag under the specified displacement is obtained by optimizing the neural network learning rate model obtained after training.

[0011] Optionally, in the above-mentioned method for predicting the thickness of a wing tooling flag based on a neural network, the step 1 comprises:

[0012] Step 1-1, using CAD software to establish a three-dimensional geometric model of the wing tooling flag, and obtaining a simplified model of the wing tooling flag by deleting all parts except bolts and gaskets in the three-dimensional geometric model;

[0013] Step 1-2, taking the thickness of the end plate of the wing tooling flag in the simplified model as a variable parameter to establish multiple simplified models with different thicknesses of the end plate of the tooling flag;

[0014] Step 1-3, saving the established multiple simplified models with different tooling flag end plate thicknesses as corresponding multiple neutral format models.

[0015] Optionally, in the above-mentioned method for predicting the thickness of a wing tooling flag based on a neural network, the simplified model of the wing tooling flag obtained in step 1-1 has the following requirements:

[0016] The simplified model is required to have the same length, width and height as the actual wing tooling flag, and the tooling flag end plate, tooling flag ribs, tooling flag front end face and bolt hole positions in the simplified model are required to be consistent with the actual wing tooling flag.

[0017] Optionally, in the above-mentioned method for predicting the thickness of a wing tooling flag based on a neural network, the step 2 comprises:

[0018] Step 2-1, import each neutral format model into the finite element calculation software, and set the Young's modulus, Poisson's ratio, and material density of the wing tooling flag according to the material type of the wing tooling flag;

[0019] Step 2-2, dividing the imported neutral format model into grid units, wherein the maximum length, width and height of the grid units in the neutral format model shall not exceed 10% of the length, width and height of the actual wing tooling flag;

[0020] Step 2-3, applying a concentrated load to the front end surface of the tooling flag of the neutral format model, and solving it by a statics solution method to calculate the displacement results and stress results of the wing tooling flag represented by the neutral format model under the concentrated load.

[0021] Optionally, in the above-mentioned method for predicting the thickness of a wing tooling flag based on a neural network, the step three comprises:

[0022] Step 3-1, number each grid unit in the neutral format model, and store the displacement component corresponding to the number in the following displacement matrix [U],

[0023]

[0024] The first column is the number of grid cells 1 to n in the neutral format model; the second to fourth columns are the displacement components corresponding to the grid cell numbers; u i,1 、u i,2 、u i,3 Respectively represent the displacement components of the i-th grid unit in the heading, span, and gravity directions;

[0025] Step 3-2: Build a neural network learning rate optimization model and set the neural network learning rate to:

[0026]

[0027] Among them, δ is the learning rate in the neural network, δ max ,δ min are the maximum learning rate and the minimum learning rate respectively; n is the number of iterations, ξ is the coefficient used to adjust the impact of the number of iterations on the learning rate; the neural network learning rate optimization model includes three layers: input layer, hidden layer, and output layer;

[0028] Step 3-3, determine the number of neuron nodes N in the hidden layer n , with the input layer as the displacement matrix [U] obtained in step 3-1 and the output layer as the thickness of the wing tooling flag, neural network training is performed to obtain the output thickness of the wing tooling flag.

[0029] Optionally, in the above-mentioned method for predicting the thickness of a wing tooling flag based on a neural network, the number of neuron nodes N of the hidden layer in step 3-1 is n for:

[0030]

[0031] Among them, N in is the number of input layer nodes in the neural network; N out is the number of nodes in the output layer of the neural network, and C is the coefficient.

[0032] Optionally, in the above-mentioned method for predicting the thickness of a wing tooling flag based on a neural network, the step 4 comprises:

[0033] In the neural network learning rate optimization model obtained after training, the load size, tooling flag rib height, tooling flag rib width, material properties, and displacement and weight of the wing tooling flag to be designed are input as input parameters to obtain the output wing tooling flag thickness, which is the thickness prediction value.

[0034] Optionally, the wing tooling flag thickness prediction method based on a neural network as described above further includes:

[0035] Step five, using the predicted thickness value of the tooling flag to design the wing tooling flag, and performing strength verification on the designed wing tooling flag.

[0036] Optionally, in the above-mentioned method for predicting the thickness of a wing tooling flag based on a neural network, the step five comprises:

[0037] The predicted thickness of the tooling flag is used to carry out the tooling flag design, and the designed wing tooling flag model is imported into the finite element software for strength verification to calculate the equivalent stress value;

[0038] If the equivalent stress value does not exceed the material strength design value, the final tooling flag thickness result is obtained; if the equivalent stress value exceeds the strength design value, the tooling flag thickness is increased at a rate of 1%; and steps four and five are repeated until the calculated equivalent stress value meets the strength design requirements.

[0039] In a second aspect, an embodiment of the present invention further provides a computer-readable storage medium, including: a memory and a processor;

[0040] Wherein, the memory is used to store a computer-readable program;

[0041] The processor is used to implement the wing tooling flag thickness prediction method based on neural network as described in any one of the above when executing the computer readable program.

[0042] Beneficial effects of the present invention: The embodiment of the present invention provides a wing tooling flag thickness prediction method based on a neural network, which takes into account factors such as the material properties of the wing tooling flag, the tooling flag rib height, the tooling flag rib width, displacement, load, weight, etc., and establishes a neural network model according to the finite element calculation results of the wing tooling flag. By establishing a good neural network model, the tooling flag thickness under a specified displacement is determined, thereby achieving the prediction of the tooling flag thickness before formal design, selecting the best thickness scheme in the design, avoiding the repeated iteration process of design and stiffness calculation, shortening the tooling flag design cycle, ensuring the tooling flag design reliability, and reducing engineering design costs; that is, the wing tooling flag thickness prediction method provided by the present invention can provide a theoretical basis for optimizing the design, and is a fast and effective wing tooling flag thickness prediction method. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The accompanying drawings are used to provide a further understanding of the technical solution of the present invention and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the technical solution of the present invention and do not constitute a limitation on the technical solution of the present invention.

[0044] Figure 1 A flowchart of a method for predicting the thickness of a wing tooling flag based on a neural network proposed by the present invention;

[0045] Figure 2 A schematic diagram of a geometric model of a wing tooling flag in a wing tooling flag thickness prediction method based on a neural network provided in an embodiment of the present invention;

[0046] Figure 3 A schematic diagram of a mesh unit model of a finite element in a wing tooling flag thickness prediction method based on a neural network provided in an embodiment of the present invention;

[0047] Figure 4 A schematic diagram of a neural network model calculation process in a wing tooling flag thickness prediction method based on a neural network provided in an embodiment of the present invention;

[0048] Figure 5 A schematic diagram of the prediction accuracy of the neural network model in the neural network-based wing tooling flag thickness prediction method provided in an embodiment of the present invention.

[0049] Description of reference numerals:

[0050] 1. Wing tooling flag; 2. Tooling flag end plate; 3. Tooling flag rib; 4. Tooling flag front end face; 5. Bolt hole; 6. Grid unit. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the present invention more clear, the embodiments of the present invention will be described in detail with reference to the accompanying drawings. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other arbitrarily without conflict.

[0052] The above background technology has explained the important role of wing tooling flags in wing assembly and the necessity of designing the thickness of wing tooling flags. However, the existing design methods for the thickness of wing tooling flags have the problem of inaccurate design values ​​obtained by manual experience design methods, and the time cost of finite element calculation methods, high difficulty in operation, and the need for a large number of iterative calculations.

[0053] Based on the above problems, it is urgent to provide a method for quickly obtaining the prediction of the thickness of the wing tooling flag. With the development of artificial intelligence, neural network technology has become a widely used model prediction technology. In the present invention, neural networks are considered to be applied in the prediction of the thickness of the tooling flag. By establishing a neural network model that considers factors such as the material properties of the wing tooling flag, the height of the tooling flag rib, the width of the tooling flag rib, displacement, load, weight, etc., the thickness of the tooling flag can be quickly predicted. The thickness of the tooling flag is predicted before the formal design, and the best thickness scheme is selected in the design, which avoids the repeated iteration process of the design and stiffness calculation, shortens the tooling flag design cycle, ensures the reliability of the tooling flag design, and reduces the engineering design cost. It is a fast and effective method for predicting the thickness of the wing tooling flag.

[0054] The present invention provides the following specific embodiments which can be combined with each other. The same or similar concepts or processes may not be described in detail in some embodiments.

[0055] Figure 1 The flowchart of the wing tooling flag thickness prediction method based on a neural network proposed by the present invention. The wing tooling flag thickness prediction method based on a neural network provided by the embodiment of the present invention comprises the following steps:

[0056] Step 1: Establish a three-dimensional geometric model of the wing tooling flag, and obtain multiple simplified models with different thicknesses and their corresponding neutral format models by simplifying and setting parameters of the three-dimensional geometric model;

[0057] Step 2: Perform finite element calculation on neutral format models of different thicknesses, and perform statics solution by applying concentrated loads in neutral format models of different thicknesses to calculate the displacement results and stress results of wing tooling flags of different thicknesses under concentrated loads;

[0058] Step 3: According to the displacement results and stress results calculated in step 2, a neural network learning rate optimization model is established, and the output wing tooling flag thickness is obtained through model training;

[0059] Step 4: The predicted value of the thickness of the tooling flag under the specified displacement is obtained by optimizing the neural network learning rate model obtained after training.

[0060] In one implementation of the embodiment of the present invention, the implementation process of the above step 1 may include:

[0061] Step 1-1, using CAD software to establish a three-dimensional geometric model of the wing tooling flag, and obtaining a simplified model of the wing tooling flag by deleting all parts except bolts and gaskets in the three-dimensional geometric model;

[0062] Step 1-2, taking the thickness of the end plate of the wing tooling flag in the simplified model as a variable parameter to establish multiple simplified models with different thicknesses of the end plate of the tooling flag;

[0063] Step 1-3, saving the established multiple simplified models with different tooling flag end plate thicknesses as corresponding multiple neutral format models.

[0064] In the specific implementation of this implementation, the simplified model of the wing tooling flag obtained in the above step 1-1 has the following requirements:

[0065] The simplified model is required to have the same length, width and height as the actual wing tooling flag, and the tooling flag end plate, tooling flag ribs, tooling flag front end face and bolt hole positions in the simplified model are required to be consistent with the actual wing tooling flag.

[0066] In one implementation of the embodiment of the present invention, the implementation process of the above step 2 may include:

[0067] Step 2-1, import each neutral format model into the finite element calculation software, and set the Young's modulus, Poisson's ratio, and material density of the wing tooling flag according to the material type of the wing tooling flag;

[0068] Step 2-2, dividing the imported neutral format model into grid units, wherein the maximum length, width and height of the grid units in the neutral format model shall not exceed 10% of the length, width and height of the actual wing tooling flag;

[0069] Step 2-3, applying a concentrated load to the front end surface of the tooling flag of the neutral format model, and solving it by a statics solution method to calculate the displacement results and stress results of the wing tooling flag represented by the neutral format model under the concentrated load.

[0070] In one implementation of the embodiment of the present invention, the implementation process of the above step three may include:

[0071] Step 3-1, number each grid unit in the neutral format model, and store the displacement component corresponding to the number in the following displacement matrix [U],

[0072]

[0073] The first column is the number of grid cells 1 to n in the neutral format model; the second to fourth columns are the displacement components corresponding to the grid cell numbers; u i,1 、u i,2 、u i,3 Respectively represent the displacement components of the i-th grid unit in the heading, span, and gravity directions;

[0074] Step 3-2: Build a neural network learning rate optimization model and set the neural network learning rate to:

[0075]

[0076] Among them, δ is the learning rate in the neural network, δ max ,δ min are the maximum learning rate and the minimum learning rate respectively; n is the number of iterations, ξ is the coefficient used to adjust the impact of the number of iterations on the learning rate; the neural network learning rate optimization model includes three layers: input layer, hidden layer, and output layer;

[0077] Step 3-3, determine the number of neuron nodes N in the hidden layer n, with the input layer as the displacement matrix [U] obtained in step 3-1 and the output layer as the thickness of the wing tooling flag, neural network training is performed to obtain the output thickness of the wing tooling flag.

[0078] In the specific implementation of this implementation, the number of neuron nodes N in the hidden layer in the above step 3-1 is n for:

[0079]

[0080] Among them, N in is the number of input layer nodes in the neural network; N out is the number of nodes in the output layer of the neural network, and C is the coefficient.

[0081] In one implementation of the embodiment of the present invention, the implementation method of the above step 4 is:

[0082] In the neural network learning rate optimization model obtained after training, the load size, tooling flag rib height, tooling flag rib width, material properties, and displacement and weight of the wing tooling flag to be designed are input as input parameters to obtain the output wing tooling flag thickness, which is the thickness prediction value.

[0083] Furthermore, the wing tooling flag thickness prediction method based on a neural network provided in an embodiment of the present invention may further include, after calculating and obtaining the output thickness prediction value:

[0084] Step five, using the predicted value of the tooling flag thickness to design the wing tooling flag, and performing strength verification on the designed wing tooling flag.

[0085] The specific implementation of step five includes:

[0086] The predicted thickness of the tooling flag is used to carry out the tooling flag design, and the designed wing tooling flag model is imported into the finite element software for strength verification to calculate the equivalent stress value;

[0087] If the equivalent stress value does not exceed the material strength design value, the final tooling flag thickness result is obtained; if the equivalent stress value exceeds the strength design value, the tooling flag thickness is increased by 1%; and steps 4 and 5 are repeated until the calculated equivalent stress value meets the strength design requirements.

[0088] The neural network-based wing tooling flag thickness prediction method provided in the embodiment of the present invention takes into account factors such as the material properties of the wing tooling flag, the tooling flag rib height, the tooling flag rib width, displacement, load, weight, etc., and establishes a neural network model based on the finite element calculation results of the wing tooling flag. Through the established neural network model, the tooling flag thickness under the specified displacement is determined, thereby achieving the prediction of the tooling flag thickness before the formal design, selecting the best thickness scheme in the design, avoiding the repeated iteration process of the design and stiffness calculation, shortening the tooling flag design cycle, ensuring the reliability of the tooling flag design, and reducing the engineering design cost; that is, the wing tooling flag thickness prediction method provided by the present invention can provide a theoretical basis for the optimized design, and is a fast and effective wing tooling flag thickness prediction method.

[0089] Based on the neural network-based wing tooling flag thickness prediction method provided by the above-mentioned embodiment of the present invention, the embodiment of the present invention further provides a computer-readable storage medium, characterized in that it includes: a memory and a processor;

[0090] Wherein, the memory is used to store a computer-readable program;

[0091] The processor is used to implement the wing tooling flag thickness prediction method based on neural network as provided in any of the above embodiments when executing the computer readable program.

[0092] The following is a schematic illustration of an implementation example of the neural network-based wing tooling flag thickness prediction method provided in an embodiment of the present invention.

[0093] Implementation Example

[0094] The neural network-based wing tooling flag thickness prediction method provided in this implementation example is referenced Figure 1 The process shown includes the following steps:

[0095] Step 1: Establish a three-dimensional geometric model of the wing tooling flag, and obtain multiple simplified models with different thicknesses and their corresponding neutral format models by simplifying and setting parameters of the three-dimensional geometric model;

[0096] Step 2: Perform finite element calculation on neutral format models of different thicknesses, and perform statics solution by applying concentrated loads in neutral format models of different thicknesses to calculate the displacement results and stress results of wing tooling flags of different thicknesses under concentrated loads;

[0097] Step 3: According to the displacement results and stress results calculated in step 2, a neural network learning rate optimization model is established, and the output wing tooling flag thickness is obtained through model training;

[0098] Step 4: The predicted value of the thickness of the tooling flag under the specified displacement is obtained by optimizing the neural network learning rate model obtained after training.

[0099] In this implementation example, the specific process of establishing multiple simplified models with different thicknesses and their corresponding neutral format models in the above step 1 is as follows:

[0100] Step 1-1, using CATIA V5 2018 software to establish a three-dimensional geometric model of the wing tooling flag, wherein all parts except bolts and gaskets in the wing tooling flag model are deleted as a simplified model.

[0101] It should be noted that in this step, the simplified model is required to have the same length, width and height as the original wing tooling flag. The tooling flag end plate 2, tooling flag rib 3, tooling flag front end face 4 and bolt hole position 5 in the simplified model are consistent with the wing tooling flag. Figure 2 As shown, a schematic diagram of the geometric model of the wing tooling flag in the wing tooling flag thickness prediction method based on a neural network provided by an embodiment of the present invention.

[0102] Step 1-2, taking the thickness of the end plate of the tooling flag as a variable parameter, the input parameter is 30 mm to 50 mm, and establishing multiple simplified models with different thicknesses of the end plate of the tooling flag at intervals of 5 mm;

[0103] Step 1-3, save the established multiple simplified models with different tooling flag end plate thicknesses as corresponding multiple neutral format models. In this implementation example, the models are all selected in .STP format files.

[0104] In this implementation example, the implementation method of performing finite element calculation on neutral format models of different thicknesses in the above step 2 includes the following steps:

[0105] Step 2-1, import each neutral format model into the finite element calculation software, set the Young's modulus and Poisson's ratio of the wing tooling flag to 210 GPa and 0.30 respectively, and the material density to 7.8×10 -9 t / mm 3 .

[0106] Step 2-2, divide the imported neutral format model into grid units. The maximum length, width and height of the grid unit in the neutral format model shall not exceed 10% of the actual length, width and height of the wing tooling flag, such as Figure 3 , which is a schematic diagram of a finite element grid unit model in a wing tooling flag thickness prediction method based on a neural network provided in an embodiment of the present invention.

[0107] Step 2-3, apply a concentrated load of 150N to the front end surface of the tooling flag of the neutral format model, and solve it using the statics solution method to calculate the displacement and stress results of the wing tooling flag under the concentrated load.

[0108] In this implementation example, the implementation method of establishing a neural network learning rate optimization model according to the calculated displacement results and stress results in the above step 3 includes the following steps:

[0109] Step 3-1, number each grid unit in the neutral format model, and store the displacement component corresponding to the number in the following displacement matrix [U],

[0110]

[0111] The first column is the number of grid cells 1 to n in the neutral format model; the second to fourth columns are the displacement components corresponding to the grid cell numbers; u i,1 、u i,2 、u i,3 Respectively represent the displacement components of the i-th grid unit in the heading, span, and gravity directions;

[0112] Step 3-1, build a neural network learning rate optimization model, and set the neural network learning rate for the model to:

[0113]

[0114] Among them, δ is the learning rate in the neural network, δ max ,δ min are the maximum learning rate and the minimum learning rate, which are 0.1 and 0.005 respectively; n is the number of iterations, which is 10000; ξ is the coefficient used to adjust the impact of the number of iterations on the learning rate, which is 0.2 to 0.5. The learning rate optimization model consists of three layers: input layer, hidden layer, and output layer. Figure 4 As shown, it is a schematic diagram of the process of neural network model calculation in the wing tooling flag thickness prediction method based on neural network provided in an embodiment of the present invention; the input layer is the matrix [U] obtained in step 3-1, and the output layer is the tooling flag thickness.

[0115] Step 3-3, determine the number of neuron nodes N in the hidden layer n , with the input layer as the displacement matrix [U] obtained in step 3-1, and the output layer as the thickness of the wing tooling flag, neural network training is performed to obtain the output thickness of the wing tooling flag, and neural network training is started. The number of neuron nodes in the hidden layer is N n for:

[0116]

[0117] Among them, Nin N is the number of input layer nodes in the neural network. in =6; N out is the number of nodes in the output layer of the neural network, N out =1, C is the adjustment coefficient, and C=9. The neural network model is trained, and the training results are as follows: Figure 5 FIG. 1 is a schematic diagram showing the prediction accuracy of the neural network model in the wing tooling flag thickness prediction method based on a neural network provided by an embodiment of the present invention. Figure 5 It can be seen that the absolute error of training is within ±0.01mm, and the relative error is controlled within ±10%, that is, the model can effectively predict the thickness of tooling flags.

[0118] In this implementation example, the implementation method of obtaining the predicted value of the thickness of the tooling flag under the specified displacement by the neural network learning rate optimization model obtained after training in the above step 4 includes the following steps:

[0119] Step 4-1: Input the structural parameters of the tooling flag to be designed as the input layer for thickness prediction. For the steel tooling flag, the Young's modulus and Poisson's ratio are 210 GPa and 0.3 respectively, and the material density is 7.8×10 -9 t / mm 3 , the rib width and height are 10mm and 16mm, the load size is 150N, the maximum deformation is required to be 0.2mm, and the weight does not exceed 100kg. The input conditions are [0.065 10 16 150 210 0.3 7.8×10 -9 ], the output thickness is 35.2mm, which means that the requirement can be met as long as the thickness is greater than 35.2mm, so it is rounded to 40mm.

[0120] Furthermore, the method provided by the implementation example also includes the following steps:

[0121] Step 5, perform strength check on the predicted thickness of the tooling flag, and obtain the maximum stress of 115MPa through strength calculation, which does not exceed the design allowable value, and then obtain the final thickness of the tooling flag.

[0122] The technical solution provided by the present invention quickly predicts the thickness of the wing tooling flag by considering factors such as the material properties of the wing tooling flag, the height of the tooling flag rib, the width of the tooling flag rib, displacement, load, weight, etc., thereby shortening the tooling flag design cycle, ensuring the reliability of the tooling flag design, and reducing engineering design costs. The technical solution provided by the present invention is a fast and effective method for predicting the thickness of the wing tooling flag.

[0123] Although the embodiments disclosed in the present invention are as above, the contents are only embodiments adopted to facilitate understanding of the present invention and are not intended to limit the present invention. Any technician in the field to which the present invention belongs can make any modifications and changes in the form and details of implementation without departing from the spirit and scope disclosed in the present invention, but the patent protection scope of the present invention shall still be subject to the scope defined in the attached claims.

Claims

1. A method for predicting the thickness of wing tooling flag based on neural network, characterized in that: include: Step 1: Establish a three-dimensional geometric model of the wing tooling flag, and obtain multiple simplified models with different thicknesses and their corresponding neutral format models by simplifying and setting parameters of the three-dimensional geometric model; Step 2: Perform finite element calculation on neutral format models of different thicknesses, and perform statics solution by applying concentrated loads in neutral format models of different thicknesses to calculate the displacement results and stress results of wing tooling flags of different thicknesses under concentrated loads; Step 3: According to the displacement results and stress results calculated in step 2, a neural network learning rate optimization model is established, and the output wing tooling flag thickness is obtained through model training; Step 4: The predicted value of the thickness of the tooling flag under the specified displacement is obtained by optimizing the neural network learning rate model obtained after training.

2. A method for predicting the thickness of a wing tooling flag based on a neural network according to claim 1, characterized in that: The step one comprises: Step 1-1, using CAD software to establish a three-dimensional geometric model of the wing tooling flag, and obtaining a simplified model of the wing tooling flag by deleting all parts except bolts and gaskets in the three-dimensional geometric model; Step 1-2, taking the thickness of the end plate of the wing tooling flag in the simplified model as a variable parameter to establish multiple simplified models with different thicknesses of the end plate of the tooling flag; Step 1-3, saving the established multiple simplified models with different tooling flag end plate thicknesses as corresponding multiple neutral format models.

3. The method for predicting the thickness of a wing tooling flag based on a neural network according to claim 2, characterized in that: The simplified model of the wing tooling flag obtained in step 1-1 has the following requirements: The simplified model is required to have the same length, width and height as the actual wing tooling flag, and the tooling flag end plate, tooling flag ribs, tooling flag front end face and bolt hole positions in the simplified model are required to be consistent with the actual wing tooling flag.

4. The method for predicting the thickness of a wing tooling flag based on a neural network according to claim 2, characterized in that: The second step comprises: Step 2-1, import each neutral format model into the finite element calculation software, and set the Young's modulus, Poisson's ratio, and material density of the wing tooling flag according to the material type of the wing tooling flag; Step 2-2, dividing the imported neutral format model into grid units, wherein the maximum length, width and height of the grid units in the neutral format model shall not exceed 10% of the length, width and height of the actual wing tooling flag; Step 2-3, applying a concentrated load to the front end surface of the tooling flag of the neutral format model, and solving it by a statics solution method to calculate the displacement results and stress results of the wing tooling flag represented by the neutral format model under the concentrated load.

5. The method for predicting the thickness of a wing tooling flag based on a neural network according to claim 4, characterized in that: The step three comprises: Step 3-1, number each grid unit in the neutral format model, and store the displacement component corresponding to the number in the following displacement matrix [U], The first column is the number of grid cells 1 to n in the neutral format model; the second to fourth columns are the displacement components corresponding to the grid cell numbers; u i,1 、u i,2 、u i,3 Respectively represent the displacement components of the i-th grid unit in the heading, span, and gravity directions; Step 3-2: Build a neural network learning rate optimization model and set the neural network learning rate to: Among them, δ is the learning rate in the neural network, δ max , δ min are the maximum learning rate and the minimum learning rate respectively; n is the number of iterations, ξ is the coefficient used to adjust the impact of the number of iterations on the learning rate; the neural network learning rate optimization model includes three layers: input layer, hidden layer, and output layer; Step 3-3, determine the number of neuron nodes N in the hidden layer n , with the input layer as the displacement matrix [U] obtained in step 3-1 and the output layer as the thickness of the wing tooling flag, neural network training is performed to obtain the output thickness of the wing tooling flag.

6. A method for predicting the thickness of a wing tooling flag based on a neural network according to claim 5, characterized in that: The number of neuron nodes N in the hidden layer in step 3-1 n for: Among them, N in is the number of input layer nodes in the neural network; N out is the number of nodes in the output layer of the neural network, and C is the coefficient.

7. The method for predicting the thickness of a wing tooling flag based on a neural network according to claim 1, characterized in that: The fourth step comprises: In the neural network learning rate optimization model obtained after training, the load size, tooling flag rib height, tooling flag rib width, material properties, and displacement and weight of the wing tooling flag to be designed are input as input parameters to obtain the output wing tooling flag thickness, which is the thickness prediction value.

8. A method for predicting the thickness of a wing tooling flag based on a neural network according to any one of claims 1 to 7, characterized in that: Also includes: Step five, using the predicted value of the tooling flag thickness to design the wing tooling flag, and performing strength verification on the designed wing tooling flag.

9. A method for predicting the thickness of a wing tooling flag based on a neural network according to claim 8, characterized in that: The step five comprises: The predicted thickness of the tooling flag is used to carry out the tooling flag design, and the designed wing tooling flag model is imported into the finite element software for strength verification to calculate the equivalent stress value; If the equivalent stress value does not exceed the material strength design value, the final tooling flag thickness result is obtained; if the equivalent stress value exceeds the strength design value, the tooling flag thickness is increased at a rate of 1%; and steps four and five are repeated until the calculated equivalent stress value meets the strength design requirements.

10. A computer-readable storage medium, characterized in that: include: Memory and processor; Wherein, the memory is used to store a computer-readable program; The processor is used to implement the wing tooling flag thickness prediction method based on neural network as described in any one of claims 1 to 9 when executing a computer-readable program.

Images