A method for calculating the thickness of a shape memory polymer core mold for composite material forming

Abstract

The present invention discloses a method for calculating the thickness of a shape memory polymer core mold for composite material forming. The process includes: establishing a geometric model of a composite material part; establishing a shape memory polymer core mold model of the composite material part, where the outer surface of the shape memory polymer core mold model is consistent with the inner shape of the geometric model of the composite material part; establishing finite element models of shape memory polymer core molds with different thicknesses; calculating the deformation characteristics of the finite element models with different thicknesses under various working states; establishing a core mold thickness prediction neural network model based on the deformation characteristic data of the finite element models; and determining the thickness of the shape memory polymer core mold under specified deformation characteristic parameters through the neural network model. The present invention can shorten the design cycle of the core mold, ensure the design reliability of the core mold, reduce the test and time costs in engineering, and is a fast and effective method for calculating the thickness of a shape memory polymer core mold.

Description

Technical Field

[0001] The present invention belongs to the field of composite molds in aircraft assembly production, and particularly relates to a method for optimizing the calculation of the thickness of a shape memory polymer core mold for composite forming. Background Art

[0002] With the development of advanced composite materials, composites are increasingly used in aircraft structures, such as air intakes, fuselage sections, pressure bulkheads, etc. Therefore, the requirements for the manufacturing and forming technology of composites are also getting higher and higher. Among them, the precise manufacturing of composites in special-shaped structures is the most challenging. Taking the composite air intake as an example, to ensure its good aerodynamic and stealth characteristics, it is necessary to form as a whole as much as possible. However, the cross-section of the air intake often has a changing irregular shape, which makes it difficult to demold the composite air intake, difficult to ensure the forming accuracy, low reuse rate of the forming mold, high cost, and long manufacturing cycle. To solve the problem of difficult high-precision forming of complex composite structures, shape memory polymer (SMP), with its superior large deformation characteristics and shape memory recovery characteristics, has become a potential recyclable mold material for composite air intake forming.

[0003] At present, the use process of the shape memory polymer core mold in engineering is as follows: (1) Design the core mold structure and manufacture it; (2) Use the core mold to complete the material paving of the composite air intake; (3) Put the composite air intake with the core mold into the autoclave, heat up and pressurize to form; (4) After cooling and depressurizing, take out the composite air intake; (5) Raise the temperature, and after the core mold softens, demold the core mold from the composite air intake; (6) Raise the temperature again to make the core mold return to its initial shape and enter the next composite air intake forming. Compared with traditional metal molds, the shape memory polymer core mold has a smaller stiffness. Therefore, when the external dimensions are determined, it is necessary to design a sufficiently large thickness to ensure that enough pressure can be borne during the composite forming process. At the same time, to ensure the reuse rate of the shape memory polymer core mold, its thickness cannot be too large to prevent the core mold from deforming and unable to return to its initial core mold shape. Therefore, the thickness parameter has become an important parameter determining the feasibility of the shape memory polymer core mold. In actual engineering, the final core mold size is generally determined by experimental verification methods. This method takes a long time, has high costs, and has low accuracy. To ensure the stiffness requirements of the shape memory polymer core mold in composite manufacturing applications, improve the forming accuracy of the composite air intake, and shorten the design cycle of the core mold thickness parameter, a method for calculating the thickness of the shape memory polymer core mold for composite forming is proposed. Summary of the Invention

[0004] To address the deficiencies of existing technologies, the present invention provides a method for calculating the thickness of a shape memory polymer core mold for composite material forming. A deformation characteristic database of shape memory polymer core molds with different thicknesses is obtained through finite element calculations. Based on this database, a neural network prediction model for the thickness of the shape memory polymer core mold is established, and then the thickness of the shape memory polymer core mold under specified performance requirements is calculated and determined.

[0005] To achieve the above objectives, the present invention adopts the following technical solutions:

[0006] A method for calculating the thickness of a shape memory polymer core mold for composite material forming, comprising the following steps:

[0007] Step 1, establish a geometric model of the composite part;

[0008] Step 2, establish a shape memory polymer core mold model for the composite part, where the outer surface of the shape memory polymer core mold model is consistent with the inner shape of the geometric model of the composite part;

[0009] Step 3, establish finite element models of shape memory polymer core molds with different thicknesses;

[0010] Step 4, calculate the deformation characteristics of the finite element models with different thicknesses under various working conditions;

[0011] 4-1 Perform state simulation on the finite element model. First, raise the temperature above the glass transition temperature and apply internal pressure to the shape memory polymer core mold; keep the pressure constant and lower the temperature below the glass transition temperature; keep the temperature constant and unload the pressure; then raise the temperature above the glass transition temperature and keep it for a period of time;

[0012] 4-2 After completing the entire 4-1 simulation process, obtain a deformation characteristic database of the shape memory polymer core mold, and the data parameters include mass, displacement, shape fixation rate, shape recovery rate, specific deformation, and specific stress. Among them, the shape fixation rate R of the shape memory polymer core mold f is defined as

[0013]

[0014] U f is the displacement of the shape memory polymer core mold after unloading, and U l is the displacement after the end of the first-stage loading,

[0015] The shape recovery rate R of the shape memory polymer core mold r is defined as

[0016]

[0017] U ris the displacement of the shape memory polymer core mold after the temperature rise in the fourth stage ends.

[0018] The specific deformation R of the shape memory polymer core mold ε is defined as:

[0019]

[0020] P is the pressure borne by the shape memory polymer core mold, ε l is the maximum strain during the deformation of the shape memory polymer core mold, and t is the thickness of the shape memory polymer core mold.

[0021] The specific stress R of the shape memory polymer core mold str is defined as:

[0022]

[0023] P is the pressure borne by the shape memory polymer core mold, S l is the maximum stress during the deformation of the shape memory polymer core mold.

[0024] Step 5: Establish a neural network model for predicting the core mold thickness based on the deformation characteristic data of the finite element model;

[0025] The established neural network model is a learning rate optimization neural network model, and the learning rate optimization formula is as follows:

[0026] f l =(l max -l min ) / (1 + k * t)+l min

[0027] where l max is the maximum learning rate, l min is the minimum learning rate, t is the number of iterations, and k is a hyperparameter used to adjust the influence of the number of iterations on the learning rate.

[0028] The learning rate optimization model includes three layers: an input layer, a hidden layer, and an output layer. The input layer of the neural network model is the shape recovery rate, shape fixation rate, displacement, specific deformation, specific stress, and weight of the core mold, and the output layer is the core mold thickness. The number of neuron nodes N in the hidden layer of the neural network h is

[0029]

[0030] where, n inp is the number of nodes in the input layer, m out is the number of nodes in the output layer, and C is an integer from 0 to 10.

[0031] Step 6: Determine the thickness of the shape memory polymer core mold under specified deformation characteristic parameters through the neural network model.

[0032] Advantages of the invention: Aiming at the current situation of the long design cycle and high cost of determining the structure of the shape memory polymer core mold, the present invention proposes a method for calculating the thickness of the shape memory polymer core mold for composite material forming. By establishing a finite element model of the core mold, calculating the deformation characteristics of the core mold with different thicknesses, and establishing a database of core mold deformation characteristics. Based on this database, a neural network core mold thickness prediction model based on learning rate optimization is proposed. Through this model, the thickness of the core mold under specified parameter requirements can be determined, which can shorten the design cycle of the core mold, ensure the reliability of the core mold design, reduce the test and time costs in engineering, and is a fast and effective method for calculating the thickness of the shape memory polymer core mold. Description of the drawings

[0033] Figure 1 It is a schematic structural diagram of the shape memory polymer core mold;

[0034] Figure 2 It is a schematic diagram of temperature and displacement during the deformation process of the shape memory polymer core mold;

[0035] Figure 3 It is a neural network model for predicting the thickness of the shape memory polymer core mold;

[0036] Figure 4 It is the accuracy of the neural network model. Detailed implementation manners

[0037] The following further describes in detail the implementation manners of the present invention in conjunction with the drawings and embodiments. The following embodiments are used to illustrate the present invention, but cannot be used to limit the scope of the present invention.

[0038] A method for calculating the thickness of the shape memory polymer core mold for composite material forming. In this embodiment, a certain aircraft air-conditioning composite intake duct is taken as the research object, as Figure 1 shown, and it includes the following steps:

[0039] Step 1: Establish a geometric model of the air-conditioning intake duct;

[0040] Step 2: Generate the outer surface of the shape memory polymer core mold according to the inner surface of the geometric model of the intake duct;

[0041] Step 3: Establish finite element models of the shape memory polymer core mold with different thicknesses (the thicknesses are 3, 4, 5, 6, 8, 10, 12, 15, 20, 25 mm), and the corresponding core mold masses are between 3.89 Kg and 32.4 Kg. The thickness range should make the required weight of 15 Kg fall within this interval;

[0042] Step 4: As Figure 2As shown in the figure, the calculation process of the deformation characteristics of the core mold is as follows:

[0043] (1) Raise the temperature to 60°C and apply an internal pressure of 0.1 MPa to the shape memory polymer core mold within 1 minute;

[0044] (2) Keep the pressure unchanged and lower the temperature to room temperature of 25°C within 2 minutes;

[0045] (3) Keep the temperature at 25°C unchanged and unload the pressure within 10 seconds;

[0046] (4) Raise the temperature to 80°C and keep it for 2 minutes.

[0047] The displacement of the shape memory polymer core mold in the loading stage, the displacement after the unloading stage, the maximum stress and maximum strain during the deformation process are calculated. The shape fixation rate, shape recovery rate, specific deformation, and specific stress of the core mold are determined based on the above parameters.

[0048] Step Five: Based on the finite element calculation data of the core mold, establish a neural network model for the core mold thickness. The established neural network model is a learning rate optimization neural network model, and the learning rate optimization formula is as follows:

[0049] f l =(l max -l min ) / (1 + k * t)+l min

[0050] where l max is the maximum learning rate, which is 0.1, l min is the minimum learning rate, which is 0.005, t is the number of iterations, the maximum value of which is 10000, and k is a hyperparameter, which is 0.2 and is used to adjust the influence of the number of iterations on the learning rate.

[0051] The learning rate optimization model includes three layers: the input layer, the hidden layer, and the output layer. The input layer of the neural network model is the shape recovery rate, shape fixation rate, displacement, specific deformation, specific stress, and weight of the core mold, and the output layer is the core mold thickness. As Figure 3 shown, the number of neuron nodes N h in the hidden layer of the neural network

[0052]

[0053] where n inp is the number of nodes in the input layer, which is 6, m out is the number of nodes in the output layer, which is 1, C is an integer from 0 to 10, and 9 is taken this time, resulting in 14 hidden layers. The neural network model is trained, and the training results are as Figure 4As shown, it can be seen that the absolute error of the training is within ±0.23 mm, and the relative error is controlled within ±3.1%, that is, the model can effectively predict the core mold thickness.

[0054] Step six, through the established neural network model, determine the shape memory polymer core mold thickness under specified deformation characteristic parameters.

[0055] In this embodiment, a certain aircraft air-conditioning composite air intake duct is used as the research object. It is required that the shape memory polymer core mold used to form this air intake duct can have a displacement of not less than 20 mm when subjected to a pressure of 0.02 MPa at 60 °C. During the deformation process, the maximum stress of the material does not exceed 4.2 MPa, the specific stress is not greater than 10, the specific deformation is not less than 3, the weight is not greater than 15 Kg. After being reused 10 times, the shape fixation rate is not less than 92%, and the shape recovery rate is not less than 90%.

[0056] From the requirements, it can be known that within 10 times, the shape fixation rate is not less than 92%, and the shape recovery rate is not less than 90%. Then, the shape fixation rate in each use process is not less than 0.92 1 / 10 = 0.9917, and the shape recovery rate is not less than 0.90 1 / 10 = 0.9895. Therefore, the input of the neural network model is [20 15 0.9917 0.9895 10 3], the prediction result is 15.536 mm, the actual thickness that can be taken is 16 mm, and the maximum stress of the material is S l = R str *P*t = 10 * 0.02 * 16 = 3.2 MPa < 4.2 MPa, which meets the requirements.

Claims

1. A method for calculating the thickness of a shape memory polymer core mold for composite material forming, characterized in that, Including: Step 1: Establish the geometric model of the composite part; Step 2: Establish the shape memory polymer core mold model of the composite part, where the outer surface of the shape memory polymer core mold model is consistent with the inner shape of the geometric model of the composite part; Step 3: Establish the finite element models of shape memory polymer core molds with different thicknesses; Step 4: Calculate the deformation characteristics of the finite element models with different thicknesses under various working conditions: a. Conduct state simulation on the finite element model. First, raise the temperature above the glass transition temperature, and apply internal pressure to the shape memory polymer core mold; keep the pressure unchanged, and lower the temperature below the glass transition temperature; keep the temperature unchanged and unload the pressure; then raise the temperature above the glass transition temperature and keep it for a period of time; b. After completing the entire simulation process in a, obtain the deformation characteristic database of the shape memory polymer core mold. The data parameters include mass, displacement, shape fixation rate, shape recovery rate, specific deformation, and specific stress. Among them, the shape fixation rate R of the shape memory polymer core mold f is defined as U f is the displacement of the shape memory polymer core mold after unloading, U l is the displacement after the end of the first stage of loading Shape recovery rate R of the shape memory polymer core mold r It is defined as U r is the displacement of the shape memory polymer core mold after the temperature rise in the fourth stage ends. Specific deformation R of the shape memory polymer core mold ε It is defined as: P is the pressure borne by the shape memory polymer core mold, and ε l is the maximum strain during the deformation process of the shape memory polymer core mold, and t is the thickness of the shape memory polymer core mold. Specific stress R of the shape memory polymer core mold str It is defined as: P is the pressure borne by the shape memory polymer core mold, and S l is the maximum stress during the deformation process of the shape memory polymer core mold; Step 5: Establish a core mold thickness prediction neural network model based on the deformation characteristic data of the finite element model; Step 6: Determine the thickness of the shape memory polymer core mold under specified deformation characteristic parameters through the neural network model.

2. The method for calculating the thickness of a shape memory polymer core mold for composite material forming according to claim 1, characterized in that In the above Step 5, the established neural network model is a learning rate optimized neural network model, and the learning rate optimization formula is as follows: f l = (l max - l min ) / (1 + k * t) + l min where l max is the maximum learning rate, l min is the minimum learning rate, t is the number of iterations, and k is a hyperparameter used to adjust the influence of the number of iterations on the learning rate.

3. The method for calculating the thickness of a shape memory polymer core mold for composite material forming according to claim 2, characterized in that The described learning rate optimization model includes three layers: an input layer, a hidden layer, and an output layer. The input layer of the neural network model is the shape recovery rate, shape fixation rate, displacement, specific deformation, specific stress, and weight of the core mold, and the output layer is the core mold thickness. The number of neuron nodes N in the hidden layer of the neural network h is where n inp is the number of nodes in the input layer, m out is the number of nodes in the output layer, and C is an integer from 0 to 10.

Images