Method and device for constructing fatigue crack propagation parameter library of complex overall structure

Abstract

The invention provides a method and device for constructing a fatigue crack propagation parameter library of a complex overall structure, and belongs to the field of crack data process.The method comprises the steps that a finite element model of the complex overall structure is constructed, and fatigue crack propagation simulation analysis is conducted on the finite element model to obtain crack propagation data; simplifying a stress intensity factor calculation formula, and constructing an agent model; constructing a neural network model, and training the neural network model based on the crack propagation data to obtain a trained neural network model; and obtaining new crack propagation data of the complex overall structure, and predicting the crack propagation data of the new complex overall structure under different working conditions through the trained neural network model to obtain crack propagation prediction data of a plurality of new complex overall structures. The crack propagation prediction data comprises crack length, a load cycle number, a stress intensity factor and an agent coefficient in the agent model, so that a new parameter library of the whole crack propagation process of the overall structure is constructed.

Description

Technical Field

[0001] The present application belongs to the technical field of fatigue data processing, and in particular relates to a method and device for constructing a fatigue crack propagation parameter library of a complex integral structure. Background Art

[0002] In materials science and engineering applications, fatigue crack growth behavior is an important factor affecting the life and safety of structural parts. The stress intensity factor of a crack is an important parameter for evaluating the crack growth rate and predicting the remaining life. It is of great significance in fatigue and crack growth analysis of engineering structures and can effectively evaluate the performance degradation of materials in long-term use.

[0003] When obtaining the stress intensity factor, the traditional method generally consults the manual, selects a specific form of structural parts, obtains the corresponding comprehensive correction factor Y, and then calculates the stress intensity factor according to the stress intensity factor. The stress intensity factor is calculated by the formula, where Y is the geometric correction coefficient and σ is the stress difference. This method has poor representation and applicability when facing complex overall structures in engineering, and the empirical data is highly dispersed and highly dependent on test results. The error will gradually increase after iteration by using the engineering combination method of element multiplication through interpolation, and the final result is often far from the actual deviation. Therefore, the use of traditional methods has the disadvantages of being time-consuming, labor-intensive, costly, and having large errors.

[0004] When faced with structural parts with continuous structures and complex types in actual engineering problems, how to use existing crack growth data to analyze, calculate, predict and build a database for stress intensity factors, so as to facilitate the use of them to achieve rapid and accurate assessment and prediction of fatigue crack growth life, is an urgent problem to be solved. Summary of the invention

[0005] The purpose of the present application is to provide a method and device for constructing a fatigue crack growth parameter library of a complex integral structure, so as to solve or alleviate at least one problem in the background technology.

[0006] The technical solution of the present application is: a method for constructing a fatigue crack growth parameter library of a complex integral structure, comprising:

[0007] Constructing a finite element model of a complex overall structure, performing fatigue crack propagation simulation analysis on the finite element model of the complex overall structure, and obtaining crack propagation data, wherein the crack propagation data includes crack length a, load cycle number n, and stress intensity factor K;

[0008] Simplify the stress intensity factor calculation formula and build a proxy model In the formula, ΔK is the proxy stress intensity factor, a is the crack length, and β is the proxy coefficient;

[0009] Constructing a neural network model, and training the neural network model based on the crack extension data to obtain a trained neural network model;

[0010] The crack propagation data of the new complex overall structure is obtained, and the crack propagation data of the new complex overall structure under different working conditions is predicted by the trained neural network model to obtain a plurality of new crack propagation prediction data of the complex overall structure, wherein the crack propagation prediction data includes crack length a, number of load cycles n, stress intensity factor K and its corresponding proxy coefficient β in the proxy model, and the crack propagation data of the new complex overall structure and the crack propagation prediction data are combined to construct a parameter library of the whole process of crack propagation of the new overall structure.

[0011] Preferably, the finite element model of the complex overall structure is constructed based on finite element software, and the finite element software includes Abaqus and Ansys.

[0012] Preferably, fatigue crack propagation simulation analysis is performed on the complex integral structure finite element model by crack propagation simulation software, and the crack propagation simulation software includes Zencrack.

[0013] Preferably, during the fatigue crack propagation simulation analysis of the complex integral structure finite element model, a slit of a predetermined width is introduced to simulate the extension of the crack to match the crack propagation position.

[0014] Preferably, the width of the slit is determined by comparing the values ​​of the stress intensity factor before and after the simulation analysis, and using a regression analysis method to ensure that the error meets the requirements.

[0015] Preferably, the neural network model adopts a sequential model, which is composed of multiple layers stacked in sequence, and data is transmitted from front to back between the multiple layers.

[0016] Preferably, the first layer of the sequential model is a fully connected layer, comprising 128 neurons, the input dimension is 2, and ReLU is used as the activation function;

[0017] The second layer of the sequential model is a fully connected layer, which contains 64 neurons and uses ReLU as the activation function. The first and second layers introduce the Dropout layer.

[0018] The third layer of the sequential model is a fully connected layer, which contains 64 neurons and uses ReLU as the activation function;

[0019] The fourth layer of the sequential model is a fully connected layer, which contains 32 neurons and uses ReLU as the activation function;

[0020] The output layer of the sequential model is a fully connected layer, which contains 1 neuron and uses a linear activation function to output real values.

[0021] Preferably, the crack growth prediction data of the new complex integral structure is stored in an Excel file.

[0022] On the other hand, the present application provides a device for constructing a fatigue crack growth parameter library of a complex integral structure, comprising:

[0023] A crack extension data acquisition module is used to construct a complex overall structure finite element model, perform fatigue crack extension simulation analysis on the complex overall structure finite element model, and obtain crack extension data, wherein the crack extension data includes crack length a, load cycle number n, and stress intensity factor K;

[0024] The proxy model building module is used to simplify the stress intensity factor calculation formula and build a proxy model In the formula, ΔK is the proxy stress intensity factor, a is the crack length, and β is the proxy coefficient;

[0025] A neural network model building module is used to build a neural network model, and train the neural network model based on the crack extension data to obtain a trained neural network model;

[0026] A data generation module is used to obtain crack propagation data of a new complex overall structure, and predict the crack propagation data of the new complex overall structure under different working conditions through a trained neural network model to obtain crack propagation prediction data of multiple new complex overall structures, wherein the crack propagation prediction data includes crack length a, number of load cycles n, stress intensity factor K and its corresponding proxy coefficient β in a proxy model, and the crack propagation data of the new complex overall structure is combined with the crack propagation prediction data to construct a parameter library of the entire crack propagation process of the new overall structure.

[0027] In a third aspect, the present application provides an electronic device, including:

[0028] one or more processors;

[0029] Memory;

[0030] One or more applications, the one or more applications are stored in the memory and are configured to be executed by the one or more processors, and the one or more applications are configured to implement the fatigue crack growth parameter library construction method for a complex integral structure as described in any one of the above.

[0031] In the final aspect, the present application provides a computer storage medium, which stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by a processor to implement the method for constructing a fatigue crack extension parameter library for a complex integral structure as described in any of the above items.

[0032] The complex integral structure fatigue crack propagation parameter library construction method of the present application solves the deficiencies in the existing technology in the prediction of fatigue crack propagation parameters. It expands the data set by using numerical simulation, and then uses proxy models and neural network technology to calculate and predict stress intensity factors. Based on this, a parameter library of the entire fatigue crack propagation process is constructed to achieve fast and accurate crack propagation evaluation, improve the accuracy and efficiency of fatigue crack propagation prediction, and provide strong technical support for structural health monitoring and safety assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solution provided by the present application, the following is a brief introduction to the accompanying drawings. Obviously, the accompanying drawings described below are only some embodiments of the present application.

[0034] Figure 1 This is a schematic diagram of the process of constructing a fatigue crack propagation parameter library for a complex integral structure of the present application.

[0035] Figure 2 It is a schematic diagram of crack length and its prediction results in one embodiment of the present application.

[0036] Figure 3 This is a schematic diagram of a new overall structural crack extension parameter library in one embodiment of the present application.

[0037] Figure 4 This is a crack length and load cycle number curve based on a parameter library in an embodiment of the present application.

[0038] Figure 5 This is a stress intensity factor and load cycle number curve based on a parameter library in an embodiment of the present application.

[0039] Figure 6 This is a curve of proxy coefficient and load cycle number based on a parameter library in one embodiment of the present application.

[0040] Figure 7 Schematic diagram of the device for constructing the fatigue crack propagation parameter library of the complex integral structure of the present application.

[0041] Figure 8 A schematic diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solutions and advantages of the implementation of this application clearer, the technical solutions in the embodiments of this application will be described in more detail below in conjunction with the drawings in the embodiments of this application.

[0043] like Figure 1 As shown, the present application provides a method for constructing a fatigue crack growth parameter library of a complex integral structure, comprising:

[0044] S10, constructing a finite element model of the complex overall structure, and performing fatigue crack propagation simulation analysis on the finite element model of the complex overall structure, thereby obtaining crack propagation data, including crack length a, load cycle number n, and stress intensity factor K.

[0045] In some embodiments of the present application, finite element software such as Abaqus or Ansys may be used to model a finite element model of a complex overall structure.

[0046] In the present application, the Zencrack crack extension simulation software can be used to perform fatigue crack extension simulation analysis on the finite element model of the complex overall structure. After the finite element model of the complex overall structure is constructed in finite element software such as Abaqus, it can be imported into the Zencrack crack extension analysis software. By introducing the initial crack and setting the fatigue analysis type, material parameters and stress ratio, simulation analysis is carried out to obtain crack extension data.

[0047] In this application, due to the presence of multiple structures such as hole edges, R zones, and long beams in the crack extension area, there are problems of cross-region, increased number of cracks, and local singularity in the simulation analysis process. During the simulation calculation process, the finite element model is adjusted and a slit of predetermined width is introduced to simulate the extension of the crack to match the crack extension position of the previous step, thereby completing the crack extension calculation on a structure with continuous structure and complex form. Among them, the width of the slit can be compared with the stress intensity factor K value before and after the simulation analysis, and regression analysis is used to ensure that the error meets the requirements, so as to ensure that the data meets the continuity and scientific requirements before and after the simulation calculation of cross-region and increased number of cracks.

[0048] In a preferred embodiment of the present application, crack propagation test data can be obtained by performing fatigue crack propagation tests on complex overall structures of different materials under different working conditions. By comparing the crack propagation test data with the crack propagation data obtained by simulation, the effectiveness and scientificity of the simulation structure can be verified.

[0049] In the cross-region fatigue crack propagation simulation of complex overall structures, multiple sets of crack propagation data can be obtained. Through regression analysis, the data that meet the accuracy and continuity requirements are spliced, and the data are cleaned and standardized, and invalid data are eliminated, so as to obtain a large amount of high-precision simulated crack propagation data for subsequent training.

[0050] S20, the calculation formula of stress intensity factor K is simplified and a proxy model is established, namely:

[0051]

[0052] Where ΔK is the proxy stress intensity factor;

[0053] a is the crack length;

[0054] β is the proxy coefficient.

[0055] S30, constructing a neural network model, and training the neural network model with the crack extension data in step S10 to obtain a trained neural network model.

[0056] In this application, the neural network model adopts a sequential model, which is composed of multiple layers stacked in sequence, and data is transmitted from front to back between these layers. The input of the neural network model is crack extension data including crack length a, load cycle number n and stress intensity factor K, and the output of the neural network model is the proxy coefficient β in the proxy model and the predicted crack length a, load cycle number n and stress intensity factor K.

[0057] The first layer is a fully connected layer (Dense Layer), which contains 128 neurons and the input dimension is 2, that is, the input layer has 2 features. Here, it is assumed that the input data has two features, and ReLU (Rectified Linear Unit) is used as the activation function, which can introduce nonlinearity and enable the model to learn more complex patterns.

[0058] The second layer is a fully connected layer, containing 64 neurons, using ReLU as the activation function; the first two layers introduce the Dropout layer and use regularization technology to prevent the model from overfitting. During each training, 20% (0.2) of the neurons are randomly turned off, the depth of the hidden layer is increased, and the Dropout technology is used to improve the model's ability to capture complex patterns of data and reduce the risk of overfitting. It is suitable for more complex data prediction tasks. The final output layer is used to generate the final prediction value;

[0059] The third layer is a fully connected layer containing 64 neurons and using ReLU as the activation function.

[0060] The fourth layer is still another fully connected layer, containing 32 neurons and using ReLU as the activation function.

[0061] Finally, the output layer selects a fully connected layer, which contains 1 neuron and uses a linear activation function. Because this is a regression task, the linear activation function can directly output real values.

[0062] In the compilation of the neural network model, Adam is used to adaptively adjust the learning rate, so that the neural network model can converge faster during training. At the same time, it is less sensitive to the initial learning rate and is suitable for most tasks. MAPE (mean absolute error) and R are also introduced. 2 (coefficient of determination) to control the error.

[0063] In the neural network model training process of the present application, the simulated crack extension data is divided into a training set and a validation set, the neural network model is trained using the training set, and the validation set is used for testing and evaluation. By making the neural network model traverse the entire training set n times, before each weight update, the neural network model will process m samples, display the progress bar and the loss information of each epoch during the training process, and use the metrics (Metrics) to customize the callback function to calculate and record the calculation accuracy and loss value during the training process, so as to facilitate the analysis and adjustment of the structure and parameters of the neural network model after training, so as to obtain the best prediction effect of the stress intensity factor K.

[0064] S40, obtaining crack propagation data of the new complex overall structure, predicting the crack propagation data of the new complex overall structure under different working conditions through the trained neural network model, and obtaining a large amount of crack propagation prediction data of the new complex overall structure, including crack length a, number of load cycles n, stress intensity factor K and its corresponding proxy coefficient β in the proxy model, merging the crack propagation data of the new complex overall structure with the crack propagation prediction data to construct a parameter library of the entire crack propagation process of the new overall structure.

[0065] like Figure 2 The left figure shows the embodiment of the present application. The original crack extension data and its predicted data of the input layer characteristics are as follows: Figure 2 The figure on the right shows the original crack extension data and its predicted data of the crack length a in this embodiment of the present application. It can be seen from the curve in the figure that the prediction result is good.

[0066] In some embodiments of the present application, the crack extension prediction data of the new complex integral structure may be stored in an Excel file.

[0067] like Figure 3 Shown is a new parameter library of the entire process of overall structural crack propagation constructed in this embodiment of the present application.

[0068] like Figures 4 to 6 The following are respectively based on the embodiment of this application Figure 3 The parameter library shown obtains the curves of crack length a and load cycle number n, the curves of stress intensity factor K and load cycle number n, and the curves of proxy coefficient β and load cycle number n.

[0069] The complex integral structure fatigue crack propagation parameter library construction method of the present application solves the deficiencies in the existing technology in the prediction of fatigue crack propagation parameters. It expands the data set by using numerical simulation, and then uses proxy models and neural network technology to calculate and predict stress intensity factors. Based on this, a parameter library of the entire fatigue crack propagation process is constructed to achieve fast and accurate crack propagation evaluation, improve the accuracy and efficiency of fatigue crack propagation prediction, and provide strong technical support for structural health monitoring and safety assessment.

[0070] like Figure 7 As shown, based on the above content, the present application also provides a device for constructing a fatigue crack growth parameter library of a complex integral structure, and the device 100 includes:

[0071] A crack extension data acquisition module 101 is used to construct a complex overall structure finite element model, perform fatigue crack extension simulation analysis on the complex overall structure finite element model, and obtain crack extension data, wherein the crack extension data includes crack length a, load cycle number n, and stress intensity factor K;

[0072] The proxy model building module 102 is used to simplify the stress intensity factor calculation formula and build a proxy model In the formula, ΔK is the proxy stress intensity factor, a is the crack length, and β is the proxy coefficient;

[0073] A neural network model building module 103 is used to build a neural network model, and train the neural network model based on the crack extension data to obtain a trained neural network model;

[0074] The data generation module 104 is used to obtain crack propagation data of a new complex overall structure, and predict the crack propagation data of the new complex overall structure under different working conditions through the trained neural network model to obtain crack propagation prediction data of multiple new complex overall structures, wherein the crack propagation prediction data includes crack length a, number of load cycles n, stress intensity factor K and its corresponding proxy coefficient β in the proxy model, and the crack propagation data of the new complex overall structure is combined with the crack propagation prediction data to construct a parameter library of the entire crack propagation process of the new overall structure.

[0075] The specific processing process of the apparatus 100 for constructing a fatigue crack growth parameter library for a complex integral structure of the present application may refer to the contents or steps in the method for constructing a fatigue crack growth parameter library for a complex integral structure, which will not be described in detail here.

[0076] like Figure 8 As shown, an embodiment of the present application also provides an electronic device 200, which includes a processing device 201, which can perform various appropriate actions and processes according to a program stored in a ROM (read-only memory) 202 or a program loaded from a storage device 208 to a RAM (random access memory) 203. The processing device 201 may include, for example, a general-purpose microprocessor (such as a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (for example, an application-specific integrated circuit (ASIC)), etc. The processing device 201 may also include an onboard memory for caching purposes. The processing device 201 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0077] In RAM 203, various programs and data required for the operation of electronic device 200 are stored. Processing device 201, ROM 202 and RAM 203 are connected to each other via bus 204. Processing device 201 performs various operations of the method flow according to the embodiment of the present application by executing the program in ROM 202 and / or RAM 203. It should be noted that the program can also be stored in one or more memories other than ROM 202 and RAM 203. Processing device 201 can also perform various operations of the method or flow according to the embodiment of the application by executing the program stored in the one or more memories.

[0078] According to an embodiment of the present application, the electronic device 200 may further include an I / O interface 205 (input / output interface), which is also connected to the bus 204. The electronic device 200 may further include one or more of the following components connected to the I / O interface 205: an input device 206 including a keyboard, a mouse, etc.; an output device 207 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage device 208 including a hard disk, etc.; and a communication device 209 including a network interface card such as a LAN card, a modem, etc. The communication device 209 performs communication processing via a network such as the Internet. A drive is also connected to the I / O interface 205 as needed. Removable media, such as magnetic disks, optical disks, magneto-optical disks, semiconductor memories, etc., are installed on the drive as needed so that the computer program read therefrom is installed into the storage device 208 as needed.

[0079] The present application also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, various operations of the above methods or processes according to the embodiments of the present application are implemented.

[0080] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A method for constructing a fatigue crack growth parameter library for a complex integral structure, characterized in that: include: Constructing a finite element model of a complex overall structure, performing fatigue crack propagation simulation analysis on the finite element model of the complex overall structure, and obtaining crack propagation data, wherein the crack propagation data includes crack length a, load cycle number n, and stress intensity factor K; Simplify the stress intensity factor calculation formula and build a proxy model In the formula, ΔK is the proxy stress intensity factor, a is the crack length, and β is the proxy coefficient; Constructing a neural network model, and training the neural network model based on the crack extension data to obtain a trained neural network model; The crack propagation data of the new complex overall structure is obtained, and the crack propagation data of the new complex overall structure under different working conditions is predicted by the trained neural network model to obtain a plurality of new crack propagation prediction data of the complex overall structure, wherein the crack propagation prediction data includes crack length a, number of load cycles n, stress intensity factor K and its corresponding proxy coefficient β in the proxy model, and the crack propagation data of the new complex overall structure and the crack propagation prediction data are combined to construct a parameter library of the whole process of crack propagation of the new overall structure.

2. The method for constructing a fatigue crack growth parameter library of a complex integral structure according to claim 1, characterized in that: The finite element model of the complex overall structure is constructed based on finite element software, and the finite element software includes Abaqus and Ansys.

3. The method for constructing a fatigue crack growth parameter library of a complex integral structure according to claim 2, characterized in that: The fatigue crack propagation simulation analysis is performed on the complex integral structure finite element model by crack propagation simulation software, wherein the crack propagation simulation software includes Zencrack.

4. The method for constructing a fatigue crack growth parameter library of a complex integral structure according to claim 3, characterized in that: During the fatigue crack propagation simulation analysis of the complex integral structure finite element model, a slit with a predetermined width is introduced to simulate the extension of the crack to match the crack propagation position.

5. The method for constructing a fatigue crack growth parameter library of a complex integral structure according to claim 4, characterized in that: The width of the slit is determined by comparing the values ​​of the stress intensity factors before and after the simulation analysis, and using a regression analysis method to ensure that the error meets the requirements.

6. The method for constructing a fatigue crack growth parameter library of a complex integral structure according to any one of claims 1 to 5, characterized in that: The neural network model adopts a sequential model, which is composed of multiple layers stacked in sequence, and data is transmitted from front to back between the multiple layers.

7. The method for constructing a fatigue crack growth parameter library of a complex integral structure according to claim 6, characterized in that: The first layer of the sequential model is a fully connected layer, which contains 128 neurons, the input dimension is 2, and ReLU is used as the activation function; The second layer of the sequential model is a fully connected layer, which contains 64 neurons and uses ReLU as the activation function. The first and second layers introduce the Dropout layer. The third layer of the sequential model is a fully connected layer, which contains 64 neurons and uses ReLU as the activation function; The fourth layer of the sequential model is a fully connected layer, which contains 32 neurons and uses ReLU as the activation function; The output layer of the sequential model is a fully connected layer, which contains 1 neuron and uses a linear activation function to output real values.

8. The method for constructing a fatigue crack growth parameter library of a complex integral structure according to claim 1, characterized in that: The crack extension prediction data of the new complex integral structure is stored in an Excel file.

9. A device for constructing a fatigue crack growth parameter library of a complex integral structure, characterized in that: include: A crack extension data acquisition module is used to construct a complex overall structure finite element model, perform fatigue crack extension simulation analysis on the complex overall structure finite element model, and obtain crack extension data, wherein the crack extension data includes crack length a, load cycle number n, and stress intensity factor K; The proxy model building module is used to simplify the stress intensity factor calculation formula and build a proxy model In the formula, ΔK is the proxy stress intensity factor, a is the crack length, and β is the proxy coefficient; A neural network model building module is used to build a neural network model, and train the neural network model based on the crack extension data to obtain a trained neural network model; A data generation module is used to obtain crack propagation data of a new complex overall structure, and predict the crack propagation data of the new complex overall structure under different working conditions through a trained neural network model to obtain crack propagation prediction data of multiple new complex overall structures, wherein the crack propagation prediction data includes crack length a, number of load cycles n, stress intensity factor K and its corresponding proxy coefficient β in a proxy model, and the crack propagation data of the new complex overall structure and the crack propagation prediction data are combined to construct a parameter library of the entire crack propagation process of the new overall structure.

10. An electronic device, characterized in that: include: one or more processors; Memory; One or more applications, wherein the one or more applications are stored in the memory and are configured to be executed by the one or more processors, and the one or more applications are configured to implement the method for constructing a fatigue crack growth parameter library for a complex integral structure as described in any one of claims 1 to 8.

11. A computer storage medium, characterized in that: The computer storage medium stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set are loaded and executed by the processor to implement the method for constructing a fatigue crack growth parameter library of a complex integral structure as described in any one of claims 1 to 8.

Images