A large-size composite material fuselage wallboard assembly positioning and clamping layout optimization method

Abstract

The application provides a large-size composite material fuselage wallboard assembly positioning and clamping layout optimization method, and the steps are as follows: 1: selecting a region with good rigidity on the wallboard as a potential clamping point; 2: using a binary integer variable to encode the stringer position and frame position where the potential clamping point is located; 3: establishing a finite element model; 4: using hypercube random sampling and finite element analysis to generate training and test data sets; 5: establishing a BPNN prediction model of the maximum gravity deformation and the maximum Mises stress of the wallboard; 6: comprehensively considering the deformation, stress and tooling cost of the large-size composite material fuselage wallboard in the assembly clamping positioning process, taking the allowable deformation of the composite material fuselage wallboard as a constraint condition, taking the minimum stress and the least number of clamping points as optimization objectives, and adopting the NSGA-II method to find an optimal solution set of the clamping layout; the large-size composite material fuselage wallboard assembly clamping positioning is high in precision, high in performance and low in cost, and has a good application prospect in the field of aircraft assembly.

Description

Technical Field

[0001] This invention provides a method for optimizing the assembly positioning and clamping layout of large-size composite fuselage panels. It is a method applicable to optimizing the clamping layout of large-size composite fuselage panels and belongs to the field of mechanical engineering / aircraft assembly. Background Technology

[0002] Due to their high specific strength and specific stiffness, excellent fatigue resistance and resistance to media corrosion, as well as unique designability of mechanical properties, composite material parts are increasingly used in aircraft structures. Among them, the fuselages of new-generation large passenger aircraft such as the B787, A350XWB, and CR929 all use ultra-large-sized integrally molded composite material panels. To ensure the flight performance of the aircraft, the shape accuracy of the composite fuselage after manufacturing must meet the stringent requirements of aircraft aerodynamic shape precision. However, because these large-sized thin-walled composite fuselage panels have relatively weak stiffness, they are prone to significant deformation under external loads such as gravity during assembly, affecting the final quality of the aircraft product. Flexible process equipment technology is an effective way to solve this problem. As a widely used flexible tooling, multi-point flexible fixtures are widely used in the manufacturing process of thin-walled parts, and excessive deformation of thin-walled parts is suppressed by optimizing the layout of clamping points.

[0003] Currently, the method of fixture layout optimization for thin-walled parts using coupled optimization algorithms based on finite element analysis has been extensively studied. However, this method requires numerous calls to the finite element solver, resulting in high computational costs and low optimization efficiency. Therefore, in recent years, scholars both domestically and internationally have begun to use surrogate models with lower computational costs, such as artificial neural networks (ANNs), to replace the finite element model and approximate the objective function value, significantly reducing computational costs and improving optimization efficiency. The fixture layout optimization method combining surrogate models and optimization squares has become a new approach to solving the problem of clamping deformation control for thin-walled parts.

[0004] Currently, most fixture layout optimization methods based on surrogate models are geared towards simple, regularly shaped thin-walled metal parts, aiming to minimize workpiece deformation and optimizing fixture layout at clamping point locations. During optimization, the number of clamping points is fixed, and their feasible regions are mostly continuous spaces. However, large composite fuselage panels are complex assemblies with horizontal and vertical stiffening and geometric features such as door and window openings. Because the panels are discontinuous stiffness structures, the distribution of clamping points is restricted to discrete spaces. Furthermore, composite panels are brittle and easily damaged under stress, making it difficult to suppress excessive deformation. In addition to the shape, the internal stress level must also be strictly controlled; furthermore, the number of clamping points is also an important consideration in fixture layout design; increasing the number of clamping points can reduce workpiece deformation; however, in practical engineering applications, the more clamping points there are, the higher the tooling manufacturing and maintenance costs will be; therefore, the number and position of clamping points need to be considered together during the optimization process, resulting in a high dimensionality of design variables; in summary, existing fixture layout optimization methods based on surrogate models cannot well handle this type of multi-objective optimization problem of discrete clamping layout with coupling of number and position. Summary of the Invention:

[0005] (I) Purpose of the Invention

[0006] To address the aforementioned technical problems in existing technologies, the present invention aims to propose an optimization method for the assembly, positioning, and clamping layout of large-size composite fuselage panels. This method is a discrete multi-objective optimization method that simultaneously considers panel deformation, internal stress levels, and tooling costs. It can simultaneously determine the optimal number and position of clamping points. First, a binary integer variable is used to characterize the clamping layout, achieving coupled characterization of the number and position of clamping points. Based on this, sampling is performed using hypercube random sampling and finite element analysis. Then, a BPNN prediction model is constructed based on the samples to describe the nonlinear mapping relationship between the composite fuselage panel clamping layout and the maximum gravitational deformation / maximum Mises stress. Next, constraints are established using the allowable deformation of the composite fuselage panel, and optimization objectives are established with minimum stress and minimum number of clamping points. Finally, the NSGA-II method is used to find the Pareto optimal solution set of the clamping layout.

[0007] The "BPNN" mentioned above refers to the Backpropagation Neural Network Model, which is an algorithmic mathematical model that simulates the workings of the human brain's neural network to process information.

[0008] The "BPNN prediction model" refers to a model that calculates the relationship between input and output based on the BPNN algorithm and using samples.

[0009] The "maximum Mises stress" refers to an equivalent stress used to evaluate whether a material has yielded, and its calculation formula is as follows:

[0010]

[0011] In the formula, σ Mises Let σ1 be the first principal stress, σ2 be the second principal stress, and σ3 be the third principal stress.

[0012] The “NSGA-II method” mentioned above refers to the second-generation genetic method; the genetic algorithm is a method for searching for the optimal solution based on the evolutionary laws of organisms in nature; the second-generation genetic method has improved the traditional genetic algorithm and increased the computational efficiency.

[0013] (II) Technical Solution

[0014] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0015] This invention discloses an optimization method for the assembly, positioning, and clamping layout of large-size composite material fuselage panels. It is a multi-objective optimization method for discrete clamping layout that combines BPNN and NSGA-II methods, specifically comprising the following steps:

[0016] Step 1: Select areas with good rigidity on the panel as potential clamping points; since the composite fuselage panel is reinforced by stringers and bulkheads along the circumferential and yaw directions respectively, the cross-shaped area where the stringers and bulkheads intersect is selected as a potential clamping point; the panel has m stringers and n bulkheads, so a total of m·n potential clamping points can be obtained.

[0017] Step 2: Encode the girder position and frame position where the potential clamping point is located using binary integer variables;

[0018] The specific representations of its truss position vector and frame position vector are as follows:

[0019] String position vector: x = [x1, x2, ..., x m ] T x i ∈(0,1)

[0020] Frame vector: y = [y1, y2, ..., y m ] T y i ∈(0,1)

[0021] In the formula: x i Indicates whether the i-th truss has a clamping point, with a value of 0 or 1, x i =0 indicates that there is no clamping point in the i-th truss, x i =1 indicates that there is a clamping point in the i-th truss; y j Indicates whether the j-th frame has a clamping point, with a value of 0 or 1, y j =0 indicates that there is no clamping point in the j-th frame, yj =1 indicates that there is a clamping point in the j-th frame; x is the sum of all x i The vector formed by all y's j The vector formed;

[0022] x×y T This represents the potential clamping point matrix; when x i =1 and y j =1 indicates that there is a clamping point at the intersection of frame i / frame j; otherwise, there is no clamping point. Therefore, the matrix x×y T The position of the non-zero term is the location of the clamping point, and ||x||0×||y||0 is the number of clamping points; the stringer vector x and the frame vector y are used as design variables to characterize the clamping layout of the composite fuselage panel;

[0023] Based on the actual requirements of the assembly process, x and y should satisfy the following constraints:

[0024] (1) The clamping point must avoid geometric features such as doors and windows, i.e.

[0025] x i ≠1, y j ≠1, i∈Ω1,j∈ω1

[0026] In the formula: Ω1 represents the range of values ​​for i corresponding to the area where the door or window is located. When i∈Ω1, x i The value of cannot be 1; ω1 represents the range of values ​​for j corresponding to the area where the door or window is located. When j∈ω1, y j The value of cannot be 1;

[0027] (2) The distance between the outer clamping point and the edge of the wall panel, and the directional and circumferential distances between adjacent clamping points should be greater than the safe distance of the locator. If the safe distance of the wall panel in the directional direction is N str For each frame location, the circumferential safety distance of the wall panel is N. cir Each truss position

[0028] x i ≠1, y j ≠1, i∈Ω2,j∈ω2

[0029] In the formula: Ω2 represents the range of values ​​for i corresponding to the area less than the safe distance from the edge of the wall panel. When i∈Ω2, x i The value of cannot be 1; ω2 represents the range of values ​​for j corresponding to the area less than the safety distance from the edge of the wall panel. When j∈ω2, y j The value of cannot be 1;

[0030] When x i When x = 1, j ≠1, (|ji|≤N) str)

[0031] When y i When y = 1, j ≠1, (|ji|≤N) cir )

[0032] The "N" mentioned str "This refers to the safe distance in the direction of the wall panel's flight;

[0033] The "N" mentioned cir "This refers to the safe distance around the perimeter of the wall panel;

[0034] In the formula: x i Let x represent any heading gripping point variable that takes the value 1. j Represents any distance x i Less than N str The heading gripping point variable, x j The value of y cannot be 1; i Let y represent any circumferential gripping point variable that takes the value 1. j Represents any distance y i Less than N cir The circumferential clamping point variable, y j The value of cannot be 1;

[0035] (3) To ensure stable clamping of the panel, and considering the tooling cost and the potential for large coupling errors due to excessive clamping points, the number of clamping points in the heading and circumferential directions should be within a certain range. Based on the preliminary analysis results of finite element simulation, we assume that the minimum number of clamping points in the heading direction of the panel is N. inf There are a maximum of N clamping points. sup One clamping point; minimum M circumferential direction. inf There are a maximum of M clamping points. sup One clamping point, namely

[0036] N inf ≤||x||0≤N sup M inf ≤||y||0≤M sup

[0037] The "N" mentioned inf "This refers to the minimum number of clamping points along the yaw direction of the panel;

[0038] The "N" mentioned sup "This refers to the maximum number of clamping points along the yaw direction of the panel;

[0039] The so-called "M" inf "This refers to the minimum number of clamping points around the perimeter of the wall panel;

[0040] The so-called "M" sup "This refers to the maximum number of clamping points around the perimeter of the wall panel;

[0041] In the formula: ||x||0 represents the non-zero dimension of vector x, i.e., the number of clamping points in the yaw direction of the wall panel, N inf ≤||x||0≤N sup This indicates that the number of clamping points along the y-direction of the panel is not less than the minimum number of clamping points along the y-direction of the panel, and not greater than the maximum number of clamping points along the y-direction of the panel; ||y||0 represents the non-zero dimension of vector y, i.e., the number of clamping points along the circumferential direction of the panel, M. inf ≤||y||0≤M sup This indicates that the number of circumferential clamping points of the wall panel is not less than the minimum number of circumferential clamping points of the wall panel, and not greater than the maximum number of circumferential clamping points of the wall panel;

[0042] Step 3: Establish a finite element model; Based on the dimensional parameters of the composite fuselage panel, establish a three-dimensional geometric model, then discretize it into a mesh model and assign material and mechanical properties, apply loads and boundary conditions, and establish a finite element model;

[0043] Step 4: Generate training and test datasets using Hypercube Random Sampling (MLHS) and Finite Element Analysis (FEA);

[0044] Step 5: Establish a BPNN prediction model for the maximum gravitational deformation and maximum Mises stress of the wall panel; the input layer of the BPNN has N neurons, corresponding to each element in the stringer position vector x and the frame position vector y, respectively, and the output layer has 1 neuron, which is the maximum gravitational deformation / maximum Mises stress of the wall panel.

[0045] The term "neuron" refers to the basic unit of a neural network, which performs one or more functions in the input, processing, and output of data in a neural network model.

[0046] Step 6: Taking into account the deformation, stress, and tooling cost of the large composite fuselage panel during assembly, clamping, and positioning, and using the allowable deformation of the composite fuselage panel as a constraint, with the minimum stress and the minimum number of clamping points as optimization objectives, the NSGA-II method is used to find the Pareto optimal solution set for the clamping layout. This problem is a binary integer programming problem with high-dimensional variables, multiple constraints, and multiple objectives. Its mathematical model can be expressed as:

[0047] The term "Pareto" refers to Pareto analysis, also known as principal-subject analysis, which is a method of classifying and ranking things based on their main technical and economic characteristics. A Pareto solution is also called a non-dominated solution. Because there are conflicts and incomparability between objectives when there are multiple objectives, the optimization problem in this case will inevitably weaken at least one other objective function while improving any objective function; these solutions are called non-dominated solutions or Pareto solutions. A set of optimal solutions to a set of objective functions is called the Pareto optimal set.

[0048] Find: x = [x1, x2, ..., x m ] T x i ∈(0,1)

[0049] y = [y1, y2, ..., y n ] T y i ∈(0,1)

[0050] Minimize: F = [F1, F2]

[0051] F1(x,y)=||x||0×||y||0

[0052] F2(x,y)=max{σ j (x,y)}

[0053] Subject to: max{ω j (x,y)}≤ω lim

[0054] N inf ≤||x||0≤N sup M inf ≤||y||0≤M sup

[0055] In the formula, x and y represent the truss position vector and frame position vector of each clamping layout scheme, respectively; f1 and f2 represent the two objective functions of maximum Mises stress and number of clamping points, respectively; σ j ω represents the Mises stress after deformation at the j-th finite element node, and its value is derived from the BPNN prediction model; j ω represents the displacement of the j-th finite element node after deformation; lim This indicates the allowable deformation of the composite fuselage panel under multi-point clamping.

[0056] (III) Advantages and Efficacy of the Invention

[0057] This invention provides an optimization method for the assembly, positioning, and clamping layout of large-size composite fuselage panels. It is a method applicable to optimizing the clamping layout of ultra-large composite fuselage panels, and its advantages and effects are: Compared with existing technologies, this invention targets large-size, horizontally and vertically stiffened, and discontinuously stiff composite panels, employing a binary integer variable to characterize the clamping layout, achieving coupled characterization of the number and position of clamping points; based on finite samples obtained through hypercube random sampling and finite element analysis, a BPNN prediction model is constructed to describe the nonlinear mapping relationship between the composite fuselage panel clamping layout and the maximum gravitational deformation / maximum Mises stress; and the allowable deformation of the composite fuselage panel is used as approximately... The present invention optimizes the clamping layout by minimizing stress and the number of clamping points under the constraint conditions, and uses the NSGA-II method to find the Pareto optimal solution set. This invention can optimize the clamping layout for assembling large-size composite fuselage panels of aircraft, and can simultaneously optimize the number and position of tooling clamping points during assembly. During the optimization process, factors such as deformation, stress, and tooling costs of large composite fuselage panels during assembly and clamping can be considered simultaneously. This reduces deformation and stress during clamping and assembly while decreasing the number of clamping tooling, achieving high-precision, high-performance, and low-cost assembly and clamping positioning of large composite fuselage panels, and has promising application prospects in the field of aircraft assembly. Attached Figure Description

[0058] Figure 1 This is a flowchart of the method of the present invention.

[0059] Figure 2 This is a schematic diagram showing the potential clamping point locations.

[0060] Figure 3 A schematic diagram of the variables in the design.

[0061] Figure 4 This is a finite element model diagram of a large composite wall panel.

[0062] Figure 5 This is a schematic diagram of the BPNN method.

[0063] Figure 6 The Pareto optimal solution set found by the NSGA-II method.

[0064] Figure 7 The diagram shows the optimal clamping layout.

[0065] The symbols and codes in the diagram are explained as follows:

[0066] BPNN: Backpropagation Neural Network Model

[0067] NSGA-II: Non-dominated sequencing genetic method 2

[0068] x: Truss vector of the clamping layout scheme

[0069] y: The frame vector of the clamping layout scheme

[0070] x i The i-th input variable

[0071] D: The total number of input variables, equal to the sum of the dimensions of x and y.

[0072] y i The i-th output variable

[0073] Q: The total number of output variables; in this example, Q = 1. Detailed Implementation

[0074] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0075] To address the aforementioned technical problems in existing technologies, the present invention aims to propose an optimization method for the assembly, positioning, and clamping layout of large-size composite fuselage panels. This method is a discrete multi-objective optimization method that simultaneously considers panel deformation, internal stress levels, and tooling costs. It can simultaneously determine the optimal number and location of clamping points. First, sampling is performed based on hypercube random sampling and finite element analysis. Then, a BPNN prediction model is constructed based on the samples to describe the nonlinear mapping relationship between the composite fuselage panel clamping layout and the maximum gravity deformation / maximum Mises stress. Next, constraints are established using the allowable deformation of the composite fuselage panel, and optimization objectives are established with minimum stress and minimum number of clamping points. Finally, the NSGA-II algorithm is used to find the Pareto optimal solution set of the clamping layout.

[0076] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0077] This invention relates to an optimization method for the assembly, positioning, and clamping layout of large-size composite material fuselage panels. It is a multi-objective optimization method for discrete clamping layout that combines BPNN and NSGA-II methods, specifically including the following steps: Figure 1 As shown:

[0078] Step 1: Select areas with good rigidity on the panel as potential clamping points; since the composite fuselage panel is reinforced by stringers and bulkheads along the circumferential and azimuth directions respectively, the cross-shaped area where the stringers and bulkheads intersect is selected as a potential clamping point. The shape of a typical large composite panel and its potential clamping points are as follows: Figure 2 As shown; in this example, the wall panel has 26 long stringers and 22 partition frames, so a total of 26×22 potential clamping points can be obtained;

[0079] Step 2: Encode the girder position and frame position where the potential clamping point is located using binary integer variables;

[0080] The specific representations of its truss position vector and frame position vector are as follows:

[0081] String position vector: x = [x1, x2, ..., x 22 ] T x i ∈(0,1)

[0082] Frame vector: y = [y1, y2, ..., y 26 ] T y i ∈(0,1)

[0083] x×y T This represents the potential clamping point matrix; when x i =1 and y j =1 indicates that there is a clamping point in frame i / j, otherwise there is no clamping point; therefore, the matrix x×y T The position of the non-zero term indicates the location of the clamping point, and ||x||0×||y||0 represents the number of clamping points. Using the truss vector x and frame vector y as design variables, the clamping layout of the composite fuselage panel is characterized. The graphical display of the coding results is shown below. Figure 3 As shown;

[0084] Based on the actual requirements of the assembly process, x and y should satisfy the following constraints:

[0085] (1) The clamping point must avoid geometric features such as doors and windows, i.e.

[0086] x 15 ,x 16 ,x 17 ,x 18 ,x 19 ,x 20 ,x 21 ≠1, y 15 ,y 16 ,y 17 ,y 18 ,y 19 ≠1

[0087] (2) The distance between the outer clamping point and the edge of the wall panel, as well as the directional and circumferential distances between adjacent clamping points, should be greater than the safety distance of the locator. If the safety distance of the wall panel in the directional direction is 1 frame position, and the safety distance of the wall panel in the circumferential direction is 3 stringer positions, then...

[0088] x1,x2,x3,x 24 ,x 25 ,x 26 ≠1, y1, y 22 ≠1

[0089] When xi When x = 1, i-3 ,x i-2 ,x i-1 ,x i+1 ,x i+2 ,x i+3 ≠1

[0090] When y i When y = 1, i-1 ,y i+1 ≠1

[0091] (3) To ensure stable clamping of the panel, and considering tooling costs and the potential for significant coupling errors from excessive clamping points, the number of clamping points in both the yaw and circumferential directions should be within a certain range. Based on the preliminary analysis results of finite element simulation, we set the minimum number of clamping points and the maximum number of clamping points in the yaw direction to be 3 and 6, and the minimum number of clamping points and the maximum number of clamping points in the circumferential direction to be 2 and 4, respectively.

[0092] 3≤||x||0≤6, 2≤||y||0≤4

[0093] Step 3: Establish the finite element model. Based on the dimensional parameters of the composite fuselage panel, establish a three-dimensional geometric model, then discretize it into a mesh model and assign material and mechanical properties. Apply loads and boundary conditions to establish the finite element model. A typical finite element model of the panel is shown below. Figure 4 As shown;

[0094] Step 4: Generate training and test datasets using Hypercube Random Sampling (MLHS) and Finite Element Analysis (FEA);

[0095] Step 5: Establish a BPNN prediction model for the maximum gravitational deformation and maximum Mises stress of the wall panel; the input layer of the BPNN has 28 neurons, corresponding to each element in the stringer position vector x and the frame position vector y, respectively, and the output layer has 1 neuron, which is the maximum gravitational deformation / maximum Mises stress of the wall panel; a typical structure of the BPNN model is as follows. Figure 5 As shown;

[0096] Step 6: Taking into account the deformation, stress, and tooling cost of the large composite fuselage panel during assembly, clamping, and positioning, and using the allowable deformation of the composite fuselage panel as a constraint, and minimizing stress and the number of clamping points as optimization objectives, a multi-objective optimization problem is established. This problem is a binary integer programming problem with high-dimensional variables, multiple constraints, and multiple objectives, and its mathematical model can be expressed as:

[0097] Find: x = [x1, x2, ..., x 13 ] T x i ∈(0,1)

[0098] y = [y1, y2, ..., y 15 ] T y i ∈(0,1)

[0099] Minimize: F = [F1, F2]

[0100] F1(x,y)=||x||0×||y||0

[0101] F2(x,y)=max{σ j (x,y)}

[0102] Subject to: max{ω j (x,y)}≤ω lim

[0103] 2≤||x||0≤4, 3≤||y||0≤6

[0104] In the formula, x and y represent the truss position vector and frame position vector of each clamping layout scheme, respectively; f1 and f2 represent the two objective functions of maximum Mises stress and number of clamping points, respectively; σ j ω represents the Mises stress after deformation at the j-th finite element node, and its value is derived from the BPNN prediction model; j ω represents the displacement of the j-th finite element node after deformation; lim This represents the allowable deformation of the composite fuselage panel under multi-point clamping. Solving this problem using the NSGA-II algorithm yields the Pareto optimal solution set, the distribution of which is shown below. Figure 6 As shown, translating the solution set into the corresponding variables, the distribution of the clamping layout is as follows: Figure 7 As shown;

[0105] This invention provides an optimization method for the assembly, positioning, and clamping layout of large-size composite fuselage panels. It is a method applicable to optimizing the clamping layout of ultra-large composite fuselage panels, and its advantages and effects are: Compared with existing technologies, this invention targets large-size, horizontally and vertically stiffened, and discontinuously stiff composite panels, employing a binary integer variable to characterize the clamping layout, achieving coupled characterization of the number and position of clamping points; based on finite samples obtained through hypercube random sampling and finite element analysis, a BPNN prediction model is constructed to describe the nonlinear mapping relationship between the composite fuselage panel clamping layout and the maximum gravitational deformation / maximum Mises stress; and the allowable deformation of the composite fuselage panel is used as approximately... The present invention optimizes the clamping layout by minimizing stress and the number of clamping points under the constraint conditions, and uses the NSGA-II method to find the Pareto optimal solution set of the clamping layout. This invention can optimize the clamping layout for the assembly of large-size composite fuselage panels for aircraft, and can simultaneously optimize the number and position of tooling clamping points during the assembly process. During the optimization process, factors such as deformation, stress, and tooling cost of large composite fuselage panels during assembly and clamping can be considered simultaneously. This reduces deformation and stress during clamping and assembly while decreasing the number of clamping tooling, achieving high-precision, high-performance, and low-cost assembly and clamping positioning of large composite fuselage panels, and has good application prospects in the field of aircraft assembly.

[0106] The content not described in detail belongs to the prior art known to those skilled in the art.

Claims

1. A method for optimizing the assembly, positioning, and clamping layout of large-size composite material fuselage panels, characterized in that, Includes the following steps: Step 1: Select areas with good rigidity on the fuselage panel as potential clamping points; since the composite fuselage panel is reinforced by stringers and bulkheads along the circumferential and azimuth directions respectively, the cross-shaped area where the stringers and bulkheads intersect is selected as a potential clamping point; the panel has Long truss Root partition, a total of One potential clamping point; Step 2: Encode the girder position and frame position where the potential clamping point is located using binary integer variables; The specific representations of its truss position vector and frame position vector are as follows: String position vector: , ; frame vector: , ; In the formula: Indicates the first i Does the truss have clamping points? The value is either 0 or 1. Indicates the first i The truss has no clamping points. Indicates the first i The truss has clamping points; Indicates the first j Does the box have gripping points? The value is 0 or 1. Indicates the first j The frame has no clamping points. Indicates the first j The frame has clamping points; For all The vector formed For all The vector formed; This represents the potential clamping point matrix; when and ,represent girder / There are clamping points at the intersection of the boxes; otherwise, there are none. Therefore, the matrix... The location of the zero term in the non-zero term is the location of the clamping point. This represents the number of clamping points; expressed as a stringer vector. and frame vector As a design variable, it characterizes the clamping layout of the composite fuselage panel; Based on the actual assembly requirements, and The following constraints must be met: (1) The clamping point must avoid the geometric features of doors and windows, i.e. ， ， ； In the formula: This indicates the area corresponding to the door and window. i The range of values ​​for , when hour, The value of cannot be 1; This indicates the area corresponding to the door and window. The range of values ​​for , when hour, The value of cannot be 1; (2) The distance between the outer clamping point and the edge of the wall panel, and the directional and circumferential distances between adjacent clamping points should be greater than the safe distance of the locator. If the safe distance of the wall panel in the directional direction is... For each frame location, the circumferential safety distance of the wall panel is... Each truss position ， ， ； In the formula: This indicates the area corresponding to the distance from the edge of the wall panel that is less than the safe distance. i The range of values ​​for , when hour, The value of cannot be 1; This indicates the area corresponding to the distance from the edge of the wall panel that is less than the safe distance. The range of values ​​for , when hour, The value of cannot be 1; when hour, ; when hour, ; In the formula: This represents the heading grip point variable when its value is 1. Represents any distance Less than The heading clamp point variable, The value of cannot be 1; This represents the circumferential gripping point variable when its value is 1. Represents any distance Less than Circumferential clamping point variable, The value of cannot be 1; (3) To ensure stable clamping of the wall panel and avoid coupling error, the minimum yaw rate of the wall panel is assumed. At most one clamping point, One clamping point; Minimum circumference At most one clamping point, One clamping point, namely ， ； In the formula: Representing vectors The non-zero dimension, i.e., the number of yaw clamping points of the panel. Representing vectors The non-zero dimension, i.e., the number of circumferential clamping points of the panel; Step 3: Establish a finite element model; Based on the dimensional parameters of the composite fuselage panel, establish a three-dimensional geometric model, then discretize it into a mesh model and assign material and mechanical properties, apply loads and boundary conditions, and establish a finite element model; Step 4: Generate training and test datasets using Hypercube Random Sampling (MLHS) and Finite Element Analysis (FEA); Step 5: Establish a BPNN prediction model for the maximum gravitational deformation and maximum Mises stress of the wall panel; the input layer of the BPNN has N neurons, each corresponding to the stringer position vector. and frame vector Each element in the output layer has one neuron, which is the maximum gravitational deformation / maximum Mises stress of the panel; Step 6: Taking into account the deformation, stress and tooling cost of the large composite fuselage panel during the assembly, clamping and positioning process, the allowable deformation of the composite fuselage panel is used as a constraint, and the minimum stress and the minimum number of clamping points are used as optimization objectives. The NSGA-II method is used to find the Pareto optimal solution set of the clamping layout.

2. The method for optimizing the assembly, positioning, and clamping layout of large-size composite material fuselage panels according to claim 1, characterized in that: The Pareto mentioned above refers to the Pareto analysis method, and the set of optimal solutions to a set of objective functions is called the Pareto optimal set; Find: , ; ， ； Minimize: ; ； ； Subject to: ; ， ； In the formula, and These represent the truss position vector and frame position vector for each clamping layout scheme, respectively; and These represent the two objective functions: maximum Mises stress and the number of clamping points, respectively. For the first The Mises stress of each finite element node after deformation is obtained from the BPNN prediction model. For the first Displacement of each finite element node after deformation; This indicates the allowable deformation of the composite fuselage panel under multi-point clamping.

Images