Parameter identification method and system for J-C constitutive model and storage medium

Abstract

The invention provides a parameter identification method for a J-C constitutive model, a parameter identification system for the J-C constitutive model, and a computer readable storage medium. The parameter identification method for the J-C constitutive model comprises the following steps: acquiring a plurality of strain field images of different plate surfaces under different loads to generate an image sequence; and taking the image sequence as input, and identifying material constitutive parameters of a J-C constitutive model according to a neural network model, the neural network model being constructed by combining a three-dimensional convolutional neural network with a two-dimensional convolutional neural network. According to the parameter identification method for the J-C constitutive model, material constitutive parameters of the J-C constitutive model can be accurately and efficiently identified, meanwhile, test materials, time and cost are remarkably saved, and the working efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of material engineering applications, and in particular to a parameter identification method for a JC constitutive model, a parameter identification system for a JC constitutive model, and a computer-readable storage medium. Background Art

[0002] The constitutive model of a material can be used to describe the dynamic response of a material during deformation. The parameters of the material constitutive model largely affect the accuracy of numerical simulation and are very important for engineering numerical simulation. In addition, the material constitutive parameters are also important for other engineering scenarios such as metal material forming, structural design, and safety assessment.

[0003] The JC (Johnson-Cook) constitutive model in the material constitutive model is:

[0004]

[0005] Among them, A, B, n, C, and m are the five major material constitutive parameters. The JC constitutive model takes into account the strain hardening and strain rate strengthening of the material. It is simple in form and can describe the strength limit of the material at high strain rates and high temperatures. It is widely used in the field of impact and load.

[0006] In the prior art (for example, Chinese invention patent 202111134816.X), the cutting parameters and chip thickness values ​​of the titanium alloy specimens are selected for the right-angle cutting test to calculate the specimen parameters of the titanium alloy specimens, and the model parameters of the JC constitutive model of the titanium alloy are solved accordingly. However, the identification and correction method of the model parameters is less universal and requires a large number of cutting tests, which is not only costly but also inefficient.

[0007] In order to overcome the above-mentioned defects of the prior art, the art urgently needs a parameter identification method for the JC constitutive model, which can accurately and efficiently identify the material constitutive parameters of the JC constitutive model while significantly saving test materials, time and costs, and improving work efficiency. Summary of the invention

[0008] A brief summary of one or more aspects is given below to provide a basic understanding of these aspects. This summary is not an exhaustive overview of all conceived aspects, and is neither intended to identify the key or critical elements of all aspects nor to define the scope of any or all aspects. Its only purpose is to give some concepts of one or more aspects in a simplified form as a prelude to a more detailed description that will be given later.

[0009] In order to overcome the above-mentioned defects in the prior art, the present invention provides a parameter identification method for a JC constitutive model, a parameter identification system for a JC constitutive model, and a computer-readable storage medium, which can accurately and efficiently identify the material constitutive parameters of the JC constitutive model while significantly saving test materials, time and costs, and improving work efficiency.

[0010] Specifically, the parameter identification method for the JC constitutive model provided according to the first aspect of the present invention includes the steps of: acquiring multiple strain field images of different plate surfaces under different loads to generate an image sequence; and using the image sequence as input to identify the material constitutive parameters of the JC constitutive model according to a neural network model, wherein the neural network model is constructed via a three-dimensional convolutional neural network combined with a two-dimensional convolutional neural network.

[0011] Preferably, in an embodiment of the present invention, the multiple strain field images of the surfaces of the different plate materials under different loads are acquired by digital image correlation technology and / or numerical simulation.

[0012] Preferably, in an embodiment of the present invention, the image sequence includes the plurality of strain field images arranged according to magnitudes of their deformation amounts.

[0013] Preferably, in one embodiment of the present invention, the convolutional neural network includes one or more of VGG, ResNet and AlexNet.

[0014] Preferably, in one embodiment of the present invention, the neural network model includes a convolutional layer, a three-dimensional global adaptive pooling layer, a flattening layer and a fully connected layer.

[0015] Preferably, in one embodiment of the present invention, the neural network model includes multiple neural network models.

[0016] Preferably, in one embodiment of the present invention, the step of constructing the neural network model further includes: searching for a hyperparameter combination of the neural network model according to the number of iterations and the learning rate.

[0017] Preferably, in an embodiment of the present invention, the step of constructing a neural network model further comprises: quantitatively analyzing the recognition performance of the neural network model according to a determination coefficient.

[0018] In addition, the parameter identification system for the JC constitutive model provided according to the second aspect of the present invention includes a memory and a processor. The memory stores computer instructions. The processor is connected to the memory and is configured to execute the computer instructions stored in the memory to implement the parameter identification method for the JC constitutive model provided in any one of the above embodiments.

[0019] In addition, the computer-readable storage medium provided according to the third aspect of the present invention stores computer instructions. When the computer instructions are executed by a processor, the parameter identification method for the JC constitutive model provided in any one of the above embodiments is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The above features and advantages of the present invention can be better understood after reading the detailed description of the embodiments of the present disclosure in conjunction with the following drawings. In the drawings, the components are not necessarily drawn to scale, and components with similar related properties or features may have the same or similar reference numerals.

[0021] Figure 1 A flow chart of a parameter identification method for a JC constitutive model provided according to some embodiments of the present invention is shown;

[0022] Figure 2 A schematic diagram of a strain field image provided according to an embodiment of the present invention is shown;

[0023] Figure 3 A schematic diagram of an image sequence provided according to an embodiment of the present invention is shown;

[0024] Figure 4 A schematic diagram of constructing a neural network model according to an embodiment of the present invention is shown; and

[0025] Figures 5A to 5D A schematic diagram showing various parameter identification results provided according to some embodiments of the present invention is shown.

[0026] Reference numerals:

[0027] 100: Parameter identification method for JC constitutive model; and

[0028] S110~S130: steps. DETAILED DESCRIPTION

[0029] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. Note that the aspects described below in conjunction with the accompanying drawings and specific embodiments are only exemplary and should not be construed as limiting the scope of protection of the present invention in any way.

[0030] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0031] In addition, the terms "upper", "lower", "left", "right", "top", "bottom", "horizontal" and "vertical" used in the following description should be understood as the directions shown in the paragraph and the related drawings. Such relative terms are only used for the convenience of description and do not mean that the device described therein must be manufactured or operated in a specific direction, and therefore should not be understood as limiting the present invention.

[0032] It is understood that although the terms "first", "second", "third", etc. may be used herein to describe various components, regions, layers and / or parts, these components, regions, layers and / or parts should not be limited by these terms, and these terms are only used to distinguish different components, regions, layers and / or parts. Therefore, the first component, region, layer and / or part discussed below may be referred to as a second component, region, layer and / or part without departing from some embodiments of the present invention.

[0033] As mentioned above, the existing identification and correction methods of model parameters have poor versatility and need to rely on a large number of cutting tests, which are costly and inefficient.

[0034] In order to overcome the above-mentioned defects in the prior art, the present invention provides a parameter identification method for a JC constitutive model, a parameter identification system for a JC constitutive model, and a computer-readable storage medium, which can accurately and efficiently identify the material constitutive parameters of the JC constitutive model while significantly saving test materials, time and costs, and improving work efficiency.

[0035] The parameter identification method for the JC constitutive model provided by the present invention combines a data-driven method and a convolutional neural network to realize accurate and efficient identification of the material constitutive parameters of the JC constitutive model. The change in the strain field distribution as the deformation increases during the material stretching process is characterized as the dynamic response of the material. The convolutional neural network is used to extract two-dimensional and three-dimensional input features and classify or regress the target to realize the extraction and processing of the strain field evolution features, and the material constitutive parameters of the material constitutive model can be obtained through the dynamic response information.

[0036] In some non-limiting embodiments, the parameter identification method for the JC constitutive model provided in the first aspect of the present invention can be implemented via the parameter identification system for the JC constitutive model provided in the second aspect of the present invention. Specifically, the parameter identification system for the JC constitutive model can be configured with a memory and a processor. The memory includes but is not limited to the computer-readable storage medium provided in the third aspect of the present invention, on which computer instructions are stored. The processor is connected to the memory and is configured to execute the computer instructions stored on the memory to implement the parameter identification method for the JC constitutive model provided in the first aspect of the present invention.

[0037] The following will first describe the working principle of the parameter identification system for the JC constitutive model in conjunction with some embodiments of the parameter identification method for the JC constitutive model. Those skilled in the art will understand that these embodiments of the parameter identification method for the JC constitutive model are only some non-limiting implementation methods provided by the present invention, and are intended to clearly demonstrate the main concept of the present invention and provide some specific solutions that are convenient for the public to implement, rather than to limit all functions or all working modes of the parameter identification system for the JC constitutive model. Similarly, the parameter identification system for the JC constitutive model is also only a non-limiting implementation method provided by the present invention, and does not constitute a limitation on the execution subject and execution order of each step in these parameter identification methods for the JC constitutive model.

[0038] Please refer to Figure 1 , Figure 1 A flow chart of a parameter identification method for a JC constitutive model provided according to some embodiments of the present invention is shown.

[0039] like Figure 1 As shown, the parameter identification method 100 for the JC constitutive model may include step S110: acquiring a plurality of strain field images of different plate surfaces under different loads to generate an image sequence.

[0040] The parameter identification system for the JC constitutive model can firstly obtain multiple strain field images of different plate surfaces under different loads through digital image correlation technology (DIC, Digital Image Correlation) or numerical simulation.

[0041] Specifically, the parameter identification system obtains multiple strain field images of different plate surfaces under different loads through DIC, and first uses the displacement control method to implement the material tensile test of the actual sample. After that, the surface strain field image of the sample plate under each displacement step is obtained through DIC. The parameter identification system can test and obtain the corresponding actual material constitutive parameters based on the strain results of DIC.

[0042] In addition, when the parameter identification system obtains multiple strain field images of different plate surfaces under different loads through numerical simulation, a finite element analysis (FEA) model for numerical simulation of tensile deformation can be used to form strain field images of deformation with different actual material constitutive parameters.

[0043] Specifically, the parameter identification system can first specify the interval of the actual material constitutive parameters and perform random sampling of the actual material constitutive parameters through the super Latin sampling method to determine the corresponding actual material constitutive parameters. After that, the tensile simulation is performed through finite element software (for example, ABAQUS, ANSYS). In the tensile simulation, one end of the sample plate is fixed and the other end is subjected to a tensile load. Each time the displacement value of the sample plate is arbitrarily increased, the strain field image under the target area can also be captured accordingly.

[0044] After acquiring multiple strain field images of different plate surfaces under different loads, the parameter identification system can generate an image sequence superimposed with multiple strain field images. Preferably, the parameter identification system can stack multiple strain field images of each sample plate under different loads in an ascending order according to the magnitude of the deformation to generate an image sequence.

[0045] The parameter identification method 100 for the JC constitutive model may include step S120: taking the image sequence as input, and identifying the material constitutive parameters of the JC constitutive model according to the neural network model, wherein the neural network model is constructed by combining a three-dimensional convolutional neural network with a two-dimensional convolutional neural network.

[0046] Here, when training the neural network model, the parameter identification system can use the image sequence as input and the actual material constitutive parameters corresponding to multiple sample plates as output labels. The two-dimensional convolutional neural network can include one or more of VGG (Visual Geometry Group), ResNet (a residual neural network) and AlexNet (a deep convolutional neural network). The three-dimensional convolutional neural network can realize the processing of three-dimensional data. By combining the three-dimensional convolutional neural network with the two-dimensional neural network, the constructed neural network model will be more accurate in identifying the material constitutive parameters of the JC constitutive model.

[0047] In a preferred embodiment, the constructed neural network model can be constructed by combining one or more of VGG, ResNet and AlexNet with a three-dimensional convolutional neural network. For example, the neural network model can include a neural network model constructed using a ResNet convolutional neural network and a three-dimensional convolutional neural network and a neural network model constructed using an AlexNet convolutional neural network and a three-dimensional convolutional neural network. Therefore, there can be multiple neural network models constructed.

[0048] The neural network model can include a convolution layer, a three-dimensional global adaptive pooling layer, a flattening layer and a fully connected layer. Through the neural network model, the parameter identification system can realize the extraction of the strain field evolution characteristics.

[0049] Preferably, after building the neural network model, the parameter identification system can search for hyperparameter combinations of the neural network model, thereby predetermining the number of iterations, learning rate and other hyperparameter combinations to achieve the best recognition effect of the neural network model. In addition, the parameter identification system can evaluate the recognition performance of the neural network model through the test set. When there are multiple neural network models, the parameter identification system can compare the performance of multiple neural network models through the test set.

[0050] In this way, the parameter identification system can obtain a neural network model that can identify the material constitutive parameters of the JC constitutive model of various materials through one training. In practice, the parameter identification system can generate a picture sequence through multiple strain field images of different plate surfaces of the obtained material under different loads. After inputting the picture sequence into the neural network model, the material constitutive parameters of the JC constitutive model of the corresponding material can be obtained.

[0051] The following is a specific non-limiting preferred embodiment, based on which the parameter identification method for the JC constitutive model proposed in the present invention is described in detail.

[0052] When training the neural network model, the parameter identification system can first specify the interval of the actual material constitutive parameters, and perform random sampling of the actual material constitutive parameters by the super Latin sampling method to generate the actual material constitutive parameters for training and testing as the output labels of the neural network model. Preferably, for the actual material constitutive parameters determined as the output labels of the neural network model, the parameter identification system can use a normalization method for processing.

[0053] In this preferred embodiment, the samples generated for training may be 200 and the test samples may be 10. The material constitutive parameters of the identified JC constitutive model may include parameter A, parameter B, parameter C and parameter n. The sampling range of parameter A may be 200-900, the sampling range of parameter B may be 200-900, the sampling range of parameter C may be 0.005-0.15, and the sampling range of parameter n may be 0.1-0.95.

[0054] Please refer to Figure 2 , Figure 2 A schematic diagram of a strain field image provided according to an embodiment of the present invention is shown.

[0055] Afterwards, the stretching simulation is performed through the finite element software. The parameter identification system can select hexahedron as the mesh type, and set the actual material constitutive parameters according to the random sampling results of the aforementioned Super Latin sampling method. The mesh of the target area is refined accordingly to improve the simulation accuracy and the reliability of the simulation results. In the stretching simulation, one end of the sample plate is fixed and the other end is subjected to a tensile load. Each time the displacement value of the sample plate is increased, preferably, the increased deformation amount each time can be 1mm. The parameter identification system gradually increases the deformation amount of the sample plate to capture the strain field image of the target area at each deformation. The captured strain field image can be as shown Figure 2 In a non-limiting embodiment, the reading of actual material constitutive parameters, stretching simulation, and storage of strain field images can be implemented by Python.

[0056] Please refer to Figure 3 , Figure 3 A schematic diagram of an image sequence provided according to an embodiment of the present invention is shown.

[0057] like Figure 3 As shown, the parameter identification system can generate an image sequence as an input of the neural network model based on the captured multiple strain field images, that is, construct a data set that characterizes the evolution of the strain image. Specifically, the parameter identification system can first adjust the size of the strain field image. For example, the size of a single strain field image data is 3×64×192, where 3 is the number of channels. In this preferred embodiment, the parameter identification system can stack the strain field images of each sample plate under different displacements in the order of loading magnitude to generate an image sequence, and the size of the final input image sequence is 3×10×64×192. Here, the parameter identification system can further divide the generated data set for training into a training set and a validation set, and the ratio of the training set to the validation set is 4:1, that is, 160 samples are used for training and 40 samples are used for validation.

[0058] Thus, the parameter identification system constructs the image sequence as input and the actual material constitutive parameters as output labels. Afterwards, the parameter identification system can continue to use the machine learning framework (pytorch, keras, etc.) to construct the three-dimensional convolutional neural network C3D, and respectively combine the classic architectures of the two-dimensional convolutional neural network VGG, ResNet and AlexNet to construct the neural network models VGG+C3D, ResNet+C3D, and AlexNet+C3D. Thus, the neural network model constructed by the parameter identification system expands the convolutional layers, pooling layers, and other parts of the two-dimensional convolutional neural network that were originally used to process two-dimensional tensors to form convolutional layers, pooling layers, and other parts for processing three-dimensional tensors. For example, the original two-dimensional convolutional layer Conv2d is expanded to a three-dimensional convolutional layer Conv3d.

[0059] Please refer to Figure 4 , Figure 4 A schematic diagram of constructing a neural network model according to an embodiment of the present invention is shown.

[0060] like Figure 4 As shown, the input of the neural network model is an image sequence. In the convolution and pooling layers of the constructed neural network model, the convolution kernel size used can be 3×3×3, and the number of channels can be increased from the initial 3 to 64. After that, it doubles as the number of convolution layers of the neural network model increases, wherein the activation function of the convolution layer can be the Relu function. The convolution layer is followed by a three-dimensional global adaptive pooling layer and a flattening layer. Each feature map is compressed to a size of 1X1X1 through a three-dimensional global adaptive pooling layer, and each feature value is flattened and connected to a fully connected layer through a flattening layer. Specifically, the fully connected layer can be composed of 1 input layer, 2 hidden layers, and 1 output layer, wherein each input layer and hidden layer can include 1024 neurons, and the output layer can be one neuron. Furthermore, a random deactivation method of neurons can be used in the fully connected layer to optimize the neural network model, reduce the interdependence between nodes, thereby realizing the regularization of the neural network model and reducing the structural risk of the neural network model. Preferably, the proportion of deactivated neurons can be 50%. Therefore, after the fully connected layer, the neural network model obtains the output material constitutive parameters A, B, C, and n.

[0061] In this preferred embodiment, the parameter identification system can further perform hyperparameter optimization on the constructed neural network models VGG+C3D, ResNet+C3D, and AlexNet+C3D. Specifically, the parameter identification system can search the learning rate of the neural network model through a deep learning hyperparameter optimization tool (e.g., Optuna) to determine the learning rates suitable for parameters A, B, C, and n. In this preferred embodiment, the parameter identification system can determine that the learning rate suitable for parameters A and B is 0.001, and the learning rate suitable for parameters C and n is 0.000008. Afterwards, hyperparameters such as optimizer, batch size, and number of iterations can also be determined by adopting this method. In this preferred embodiment, the parameter identification system can select optimizer Adam and determine that the batch size is 8 and the number of iterations is 200 according to the determined optimization result, that is, the hyperparameter with the best parameter identification effect. After the parameter identification system performs hyperparameter optimization on the neural network model, the neural network model is trained. Preferably, the loss function used for model training can be mean square error (MSE).

[0062] Furthermore, the parameter identification system can analyze the recognition performance of the constructed neural network models VGG+C3D, ResNet+C3D, and AlexNet+C3D, and compare the constructed neural network models VGG+C3D, ResNet+C3D, and AlexNet+C3D. Preferably, the parameter identification system can quantitatively analyze the recognition performance of the neural network model based on the determination coefficient (i.e., the square of the correlation coefficient R). The determination coefficient is determined as follows:

[0063]

[0064] Among them, y i is the true value, is the predicted value, is the mean of the predicted values, and m is the sample size.

[0065] Please refer to Figures 5A to 5D , Figures 5A to 5D A schematic diagram showing various parameter identification results provided according to some embodiments of the present invention is shown.

[0066] The image sequence generated by the above parameter recognition system based on the test sample is used as the input of the constructed neural network model VGG+C3D, ResNet+C3D, and AlexNet+C3D to predict the parameters A, B, C, and n of the JC constitutive model. The recognition effect of the predicted parameters A, B, C, and n is shown in Figures 5A to 5DAs shown in Table 1, the determination coefficients of the neural network models VGG+C3D, ResNet+C3D, and AlexNet+C3D for the recognition of each parameter are shown in Table 1:

[0067] A B C n ResNet+C3D 0.9324 0.9071 0.7864 0.7396 VGG+C3D 0.8362 0.7351 0.7846 0.6801 AlexNet+C3D 0.8160 0.8291 0.8754 0.7305

[0068] Table 1

[0069] Combination Figures 5A to 5D As shown in Table 1, the ResNet+C3D model outperforms the other two models in identifying parameters A, B, and n, especially the determination coefficient of parameter A can reach 0.9324. For the identification of parameter C, AlexNet+C3D shows good performance, and the determination coefficient can reach 0.8754.

[0070] Thus, the material constitutive parameters of the JC constitutive model identified by the neural network model provided by the present invention have good consistency with the actual material constitutive parameters.

[0071] In summary, the parameter identification method for the JC constitutive model provided by the present invention can solve the problems of complicated procedures and high workload in the prior art when determining material constitutive parameters. It can be applied to various materials, and only requires a tensile test of the plate of the corresponding material to identify multiple material constitutive parameters based on the strain results, which greatly saves test materials, time and costs.

[0072] Although the above methods are illustrated and described as a series of actions for simplicity of explanation, it should be understood and appreciated that these methods are not limited by the order of the actions, because according to one or more embodiments, some actions may occur in a different order and / or concurrently with other actions from those illustrated and described herein or not illustrated and described herein but understandable to those skilled in the art.

[0073] Those skilled in the art will appreciate that information, signals, and data may be represented using any of a variety of different technologies and techniques. For example, data, instructions, commands, information, signals, bits, symbols, and chips cited throughout the above description may be represented by voltage, current, electromagnetic waves, magnetic fields or magnetic particles, optical fields or optical particles, or any combination thereof.

[0074] Those skilled in the art will further appreciate that the various illustrative logic blocks, modules, circuits, and algorithm steps described in conjunction with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or a combination of the two. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps are generally described above in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the specific application and the design constraints imposed on the overall system. The technician may implement the described functionality in different ways for each specific application, but such implementation decisions should not be interpreted as resulting in a departure from the scope of the present invention.

[0075] The various illustrative logic modules and circuits described in conjunction with the embodiments disclosed herein may be implemented or executed with a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general purpose processor may be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in cooperation with a DSP core, or any other such configuration.

[0076] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. The software module may reside in a RAM memory, a flash memory, a ROM memory, an EPROM memory, an EEPROM memory, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to a processor so that the processor can read and write information from / to the storage medium. In an alternative, a storage medium may be integrated into a processor. The processor and the storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In an alternative, the processor and the storage medium may reside in a user terminal as discrete components.

[0077] In one or more exemplary embodiments, the functions described may be implemented in hardware, software, firmware, or any combination thereof. If implemented as a computer program product in software, each function may be stored on or transmitted by a computer-readable medium as one or more instructions or codes. Computer-readable media include both computer storage media and communication media, including any medium that facilitates the transfer of a computer program from one place to another. Storage media may be any available medium that can be accessed by a computer. As an example and not limitation, such a computer-readable medium may include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer. Any connection is also properly referred to as a computer-readable medium. For example, if the software is transmitted from a website, a server, or other remote source using a coaxial cable, a fiber optic cable, a twisted pair, a digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwaves, the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwaves are included in the definition of the medium. Disk and disc as used herein include compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk and Blu-ray disc, wherein disk often reproduces data magnetically, while disc reproduces data optically with lasers. Combinations of the above should also be included within the scope of computer-readable media.

[0078] The previous description of the disclosure is provided to enable any person skilled in the art to make or use the disclosure. Various modifications to the disclosure will be apparent to those skilled in the art, and the general principles defined herein may be applied to other variations without departing from the spirit or scope of the disclosure. Thus, the disclosure is not intended to be limited to the examples and designs described herein, but should be granted the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A parameter identification method for a JC constitutive model, characterized in that: Includes steps: Acquire multiple strain field images of different plate surfaces under different loads to generate an image sequence; and The image sequence is used as input, and material constitutive parameters of the JC constitutive model are identified according to a neural network model, wherein the neural network model is constructed by combining a three-dimensional convolutional neural network with a two-dimensional convolutional neural network.

2. The parameter identification method according to claim 1, characterized in that: The multiple strain field images of the surfaces of different plates under different loads are obtained by digital image correlation technology and / or numerical simulation.

3. The parameter identification method according to claim 1, characterized in that: The image sequence includes the plurality of strain field images arranged according to magnitudes of their deformation amounts.

4. The parameter identification method according to claim 1, characterized in that: The two-dimensional convolutional neural network includes one or more of VGG, ResNet and AlexNet.

5. The parameter identification method according to claim 1, characterized in that: The neural network model includes a convolution layer, a three-dimensional global adaptive pooling layer, a flattening layer and a fully connected layer.

6. The parameter identification method according to claim 1, characterized in that: The neural network model includes multiple neural network models.

7. The parameter identification method according to claim 1, characterized in that: The step of constructing the neural network model also includes: According to the number of iterations and the learning rate, a hyperparameter combination of the neural network model is searched.

8. The parameter identification method according to claim 1, characterized in that: The step of constructing the neural network model also includes: The recognition performance of the neural network model is quantitatively analyzed based on the determination coefficient.

9. A parameter identification system for a JC constitutive model, characterized in that: include: a memory having computer instructions stored thereon; and A processor is connected to the memory and is configured to execute computer instructions stored in the memory to implement the parameter identification method for the JC constitutive model according to any one of claims 1 to 8.

10. A computer-readable storage medium having computer instructions stored thereon, characterized in that: When the computer instructions are executed by a processor, the parameter identification method for the JC constitutive model according to any one of claims 1 to 8 is implemented.

Images