A fatigue life prediction method and related device

Abstract

The present application provides a fatigue life prediction method and related devices, which relate to the fields of composite material mechanics and experimental mechanics technology. In the present application, fatigue test parameters are determined, and fatigue test operations are performed on target composite materials using the fatigue test parameters to obtain time-frequency scaling diagrams of the target composite materials at different fatigue times. According to the fatigue test parameters and the time-frequency scaling diagrams at different fatigue times, fatigue life prediction operations are performed to obtain the fatigue life of the target composite materials. That is, the fatigue life test is realized by the present invention, and in the present invention, the time-frequency scaling diagrams of the target composite materials at different fatigue times are used. The data can represent the usage data of the target composite materials in actual scenarios. Therefore, fatigue life prediction using the data can improve the accuracy of life prediction.

Description

Technical Field

[0001] The present application relates to the technical field of composite material mechanics and experimental mechanics, and more specifically, to a fatigue life prediction method and related devices. Background Art

[0002] Composite materials are new materials that are made by optimizing the combination of material components with different properties using advanced material preparation technology. Fatigue life prediction of composite materials is an important research field that aims to evaluate the durability of materials under cyclic stress.

[0003] Fatigue life prediction of composite materials can better optimize the structure of composite materials. Therefore, how to achieve fatigue life prediction of composite materials is a technical problem that technical personnel in this field need to solve urgently. Summary of the invention

[0004] In view of this, the present application provides a fatigue life prediction method and related devices to solve the problem of the urgent need to predict the fatigue life of composite materials.

[0005] In order to solve the above technical problems, this application adopts the following technical solutions:

[0006] A fatigue life prediction method, comprising:

[0007] Determine fatigue test parameters;

[0008] Using the fatigue test parameters, a fatigue test operation is performed on the target composite material to obtain a time-frequency scaling diagram of the target composite material under different fatigue times;

[0009] A fatigue life prediction operation is performed according to the fatigue test parameters and the time-frequency scaling diagrams under different fatigue times to obtain the fatigue life of the target composite material.

[0010] Optionally, using the fatigue test parameters, a fatigue test operation is performed on the target composite material to obtain a time-frequency scaling diagram of the target composite material under different fatigue times, including:

[0011] Determine fatigue times;

[0012] For each fatigue number, using the fatigue test parameters, controlling the fatigue test device to perform a preset fatigue test action on the target composite material, the number of executions of the preset fatigue test action being the fatigue number;

[0013] After the preset fatigue test action is executed, a time-frequency scaling diagram of the target composite material is determined.

[0014] Optionally, determining a time-frequency scaling diagram of the target composite material comprises:

[0015] Acquiring image data of the target composite material;

[0016] A deformation analysis operation is performed on the image data to obtain a time-frequency scale diagram.

[0017] Optionally, performing a fatigue life prediction operation according to the fatigue test parameters and the time-frequency scaling diagrams under different fatigue times to obtain the fatigue life of the target composite material includes:

[0018] Convert the time-frequency scale images under different fatigue times into pixel matrices;

[0019] Performing a normalization operation on the pixel matrix to obtain a target image vector;

[0020] The fatigue life prediction model is called to process the fatigue test parameters and the target image vector to obtain the fatigue life of the target composite material.

[0021] Optionally, the training process of the fatigue life prediction model includes:

[0022] Acquire training samples; the training samples include input parameters and fatigue life reference values; the input parameters include test parameter samples and image vector samples;

[0023] The fatigue life prediction model is trained using the training samples until the training stop condition is met; in the fatigue life prediction model, the number of hidden layer nodes is a preset number; the transfer function is a Sigmoid function; the output function of the output layer is a linear function; and the training function is a trainlm function.

[0024] Optionally, obtaining training samples includes:

[0025] Performing a preset fatigue test operation on the actual composite material by a fatigue test device to obtain a first training sample;

[0026] Performing finite element simulation operation on the simulated composite material to obtain a second training sample;

[0027] The first training sample and the second training sample are combined to obtain a training sample.

[0028] Optionally, performing a finite element simulation operation on the simulated composite material to obtain a second training sample includes:

[0029] Using the test parameter samples, finite element simulation is performed on the simulated composite material. During the simulation process, data enhancement operation is performed by adding noise with a preset signal-to-noise ratio to obtain a scale map sample;

[0030] converting the scale map samples into image vector samples;

[0031] The test parameter sample and the image vector sample are combined to obtain a second training sample.

[0032] A fatigue life prediction device, comprising:

[0033] A parameter determination module, used for determining fatigue test parameters;

[0034] A fatigue test module, used to perform a fatigue test operation on a target composite material using the fatigue test parameters, and obtain a time-frequency scaling diagram of the target composite material under different fatigue times;

[0035] The life prediction module is used to perform fatigue life prediction operations according to the fatigue test parameters and the time-frequency scaling diagrams under different fatigue times to obtain the fatigue life of the target composite material.

[0036] An electronic device comprises at least one processor and a memory connected to the processor, wherein:

[0037] The memory is used to store computer programs;

[0038] The processor is used to execute the computer program so that the electronic device can implement the above fatigue life prediction method.

[0039] A fatigue life prediction system includes the above electronic device and further includes:

[0040] Fatigue testing equipment and cameras;

[0041] The fatigue testing device is used to perform fatigue testing operations based on the control of the electronic equipment;

[0042] The camera is used to collect image data of the target composite material and send the image data to the electronic device.

[0043] The present application provides a fatigue life prediction method and related devices. In the present application, fatigue test parameters are determined, and fatigue test operations are performed on target composite materials using the fatigue test parameters to obtain time-frequency scale diagrams of the target composite materials at different fatigue times. According to the fatigue test parameters and the time-frequency scale diagrams at different fatigue times, fatigue life prediction operations are performed to obtain the fatigue life of the target composite materials. That is, the fatigue life test is realized by the present invention, and in the present invention, the time-frequency scale diagrams of the target composite materials at different fatigue times are used, and the data can represent the usage data of the target composite materials in actual scenarios. Therefore, fatigue life prediction using the data can improve the accuracy of life prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are merely embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0045] Figure 1 A flowchart of a fatigue life prediction method provided by an embodiment of the present invention;

[0046] Figure 2 A schematic diagram of the structure of a fatigue life prediction system provided by an embodiment of the present invention;

[0047] Figure 3 A flow chart of a method for determining a time-frequency scaling diagram provided by an embodiment of the present invention;

[0048] Figure 4 A schematic diagram of a composite material provided by an embodiment of the present invention;

[0049] Figure 5 A schematic diagram of a time-frequency scaling diagram provided by an embodiment of the present invention;

[0050] Figure 6 A flow chart of a method for life prediction provided by an embodiment of the present invention;

[0051] Figure 7 A schematic diagram of a model structure provided by an embodiment of the present invention;

[0052] Figure 8 A schematic diagram of a scenario of a fatigue life prediction method provided by an embodiment of the present invention;

[0053] Fig. 9 A schematic diagram of the structure of a fatigue life prediction device provided by an embodiment of the present invention;

[0054] Fig.10 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention;

[0055] Fig.11 A schematic diagram of the structure of a fatigue life prediction system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0057] Composite materials have different mechanical properties from metal materials, and their fatigue damage mechanisms are also different from those of metal materials. Macroscopically, fatigue damage of composite materials can be measured by the degree of strength degradation. In recent years, a lot of research has been done on using residual strength to characterize fatigue damage of composite materials, and a variety of residual strength models have been established to predict fatigue life. However, the premise of the application of existing models is to use residual strength test data to confirm the relevant parameters in the model. However, the cost of residual strength tests is high, and there is often no historical data available in actual engineering, which seriously restricts the application of existing models.

[0058] To this end, in the present invention, in order to simplify the processing scheme for fatigue life prediction of composite materials, after obtaining the time-frequency scaling diagram using a fatigue testing device, the correlation between fatigue test parameters, the time-frequency scaling diagram and fatigue life is used to train a machine learning or neural network model. The machine learning or neural network method can establish a nonlinear mapping relationship between input signals and output results, which can improve the accuracy of fatigue life prediction of composite materials, and the neural network has a strong ability to identify unique system states in complex environments.

[0059] In the present invention, only fatigue test parameters and time-frequency scaling diagrams need to be collected. Compared with the parameters used in the residual strength model, the amount of parameter collection is reduced, thereby reducing costs.

[0060] Based on the above content, an embodiment of the present invention provides a fatigue life prediction method, and the execution subject can be an electronic device such as a processor, a controller, and a server.

[0061] Reference Figure 1 , a fatigue life prediction method may include:

[0062] S11. Determine fatigue test parameters.

[0063] In the embodiment of the present invention, since it is necessary to use a fatigue testing device to obtain a time-frequency scaling diagram, it is necessary to determine fatigue testing parameters used when using the fatigue testing device.

[0064] The structural diagram of the fatigue testing device can be shown as follows: Figure 2 shown.

[0065] Figure 2The sample in the embodiment of the present invention is the target composite material. The fatigue test device is used to apply pressure to the target composite material. Before applying pressure to the composite material, black and white dots need to be applied to the composite material, which is generally used to determine the time-frequency scale diagram later.

[0066] The digital image acquisition high-speed camera performs image acquisition operation on the target composite material, and then the digital image correlation technology (Digital Image Correlation, DIC) is called to obtain the time-frequency scale map.

[0067] The core principle of digital image correlation technology is to use image comparison to measure the deformation and displacement of an object. This technology requires applying random black and white dots on the surface of the object, which form a specific texture to facilitate subsequent image processing. The working process of DIC is usually divided into several steps.

[0068] First, before the experiment begins, the researchers add texture to the surface of the object to be tested (i.e., the target composite material in the embodiment of the present invention). Then, a high-resolution camera (i.e., Figure 2 The digital image acquisition high-speed camera in the image acquisition system captures the initial state of the object to form a reference image. When external force is applied, the surface of the object will deform, and the deformed image will be captured again.

[0069] The DIC system then uses an algorithm to identify the displacement of each point by comparing the two images. Specifically, DIC compares a small area (sub-image) in the initial image with the deformed image to calculate the displacement and strain of each point. This process is not only efficient, but also provides full-field strain distribution, avoiding the limitations of traditional measurement methods.

[0070] exist Figure 2 Based on the structure of the fatigue testing device shown, fatigue testing parameters may include: loading frequency, load amplitude, stress ratio, etc.

[0071] The loading frequency refers to the number of times the load is applied to the specimen per second, usually expressed in Hertz (Hz).

[0072] Load Amplitude represents the peak value of a material between positive loads. It is the difference between the highest and lowest load values ​​in a fatigue cycle.

[0073] The stress ratio refers to the ratio of the minimum load to the maximum load during cyclic loading of the specimen (or the ratio of the minimum stress to the maximum stress of the specimen).

[0074] In practical applications, technicians can set the values ​​of loading frequency, load amplitude, and stress ratio according to the requirements and the bearing capacity of the composite material. For example, the loading frequency can be set to 5 / S, the load amplitude to 100N, and the stress ratio to 2.

[0075] S12. Perform a fatigue test operation on the target composite material using the fatigue test parameters to obtain a time-frequency scaling diagram of the target composite material under different fatigue times.

[0076] Specifically, the fatigue test device supports three test modes when performing fatigue tests, namely three-point bending test, tensile test and compression test. Different test modes use different test equipment, such as Figure 2 As shown in the figure, the three-point bending test uses a three-point bending device. In addition, if a tensile test is required, the three-point bending device can be replaced by a tensile device.

[0077] After the fatigue test parameters are determined, fatigue test operations can be performed on the target composite material according to the fatigue test parameters.

[0078] Specifically, the present invention takes the three-point bending test as an example to explain the fatigue test process. If the loading frequency is 5 / S, the load amplitude is set to 100N, and the stress ratio is set to 2, the target composite material is placed Figure 2 At the sample, the fatigue test device is started, pressed down five times in one second, each pressing force is 100N, and the force on the target composite material when lifted is 100N / 2=50N.

[0079] After multiple presses, you can use Figure 2 The digital image acquisition high-speed camera performs image acquisition operation on the target composite material to obtain its corresponding time-frequency scale diagram. During image acquisition, an irradiation light source can be used to provide sufficient brightness.

[0080] S13, performing a fatigue life prediction operation according to the fatigue test parameters and the time-frequency scaling diagrams under different fatigue times to obtain the fatigue life of the target composite material.

[0081] Specifically, after obtaining the fatigue test parameters and the time-frequency scaling diagram under different fatigue times, a machine learning model or a neural network model can be used to predict fatigue life to obtain the fatigue life of the target composite material. The fatigue life of the target composite material refers to the predicted number of uses, such as 10,000 times.

[0082] In the embodiment of the present invention, the BP (back propagation) neural network is suitable for a series of applications such as nonlinear classification prediction, pattern recognition and function approximation. Therefore, the embodiment of the present invention uses the BP neural network model to predict fatigue life. In addition, a convolutional neural network can also be used.

[0083] The basic idea of ​​BP algorithm is that the learning process consists of two processes: forward propagation of signals and back propagation of errors. The input is input from the input layer, processed by the hidden layer, and then transmitted to the output layer. If the actual output of the output layer does not match the expected output, the back propagation phase of the error is entered.

[0084] When the BP neural network model is used in the present invention, the fatigue test parameters and the time-frequency scale diagram under different fatigue times are taken as input, and the output obtained is the fatigue life.

[0085] In this embodiment, fatigue test parameters are determined, and fatigue test operations are performed on the target composite material using the fatigue test parameters to obtain a time-frequency scale diagram of the target composite material at different fatigue times. A fatigue life prediction operation is performed based on the fatigue test parameters and the time-frequency scale diagram at different fatigue times to obtain the fatigue life of the target composite material. That is, the fatigue life test is achieved through the present invention, and in the present invention, the time-frequency scale diagram of the target composite material at different fatigue times is used, and the data can represent the usage data of the target composite material in the actual scene. Therefore, fatigue life prediction using this data can improve the accuracy of life prediction.

[0086] The above embodiment mentions using the fatigue test parameters to perform fatigue test operations on the target composite material to obtain a time-frequency scaling diagram of the target composite material under different fatigue times. The specific implementation thereof is now introduced.

[0087] Reference Figure 3 , which may include:

[0088] S21. Determine the number of fatigue times.

[0089] In this embodiment, a total number of fatigue tests, such as 1500 times, can be set manually according to the test requirements and the bearing capacity of the target composite material. In order to detect the deformation of the target composite material during the stress process, its deformation can be detected after a certain number of fatigue tests.

[0090] For example, the deformation condition can be detected when the fatigue times are 0, 500, 1000 and 1500 times, and the fatigue times in the embodiment of the present invention are 0, 500, 1000 and 1500 times.

[0091] S22. For each fatigue number, using the fatigue test parameters, control the fatigue test device to perform a preset fatigue test action on the target composite material, where the number of executions of the preset fatigue test action is the fatigue number.

[0092] Specifically, the operation process of the fatigue testing device is as follows:

[0093] If the loading frequency is 5 / S, the load amplitude is set to 100N, the stress ratio is set to 2, and the target composite material is placed Figure 2 At the sample, the fatigue test device is started, pressed down five times in one second, each pressing force is 100N, and the force on the target composite material when lifted is 100N / 2=50N.

[0094] The fatigue test rig performs the same process each time, performing a total of 1,500 compressions on the target composite material for the three-point bending test.

[0095] After 1500 presses, the target composite material may have deformed. The deformation diagram can be found in Figure 4 . Figure 4 In the example, the size of the composite material can be 60 mm×15 mm×9 mm. Before pressing, it is the composite material that has not been subjected to fatigue testing, and after pressing, it is the composite material that has been subjected to fatigue testing.

[0096] S23, after the preset fatigue test action is executed, determining a time-frequency scaling diagram of the target composite material.

[0097] After the preset fatigue test action is completed, the image data of the target composite material is collected by using the high-speed camera, and the digital image correlation technology is used to perform deformation analysis operation on the image data to obtain a time-frequency scale diagram.

[0098] It should be noted that when the image data of the target composite material is collected using the above-mentioned high-speed camera, the image data of the entire composite material can be collected to ensure the integrity of the data.

[0099] However, in actual applications, the deformation of the composite material mainly occurs between the pressure head and the support of the three-point bending device, and the deformation of the parts outside the support is relatively small. Therefore, in the embodiment of the present invention, the image data between the two supports can be mainly obtained. In addition, the composite material in the embodiment of the present invention is symmetrically placed on the support, and the deformation between the two supports and the pressure head is similar, and the image data between one support and one pressure head can also be collected.

[0100] In the above embodiment, the fatigue times are set to 0, 500, 1000 and 1500 times. At each fatigue time, the image data of the target composite material is collected, and the image data is subjected to deformation analysis using digital image correlation technology to obtain a time-frequency scale diagram. The time-frequency scale diagram can be referred to Figure 5 shown. Figure 5 Different colors represent different strains exy, which are determined by using the Lagrange method in digital image correlation technology.

[0101] It should be noted that if the fatigue times are set to 0, 500, 1000 and 1500 times, the time-frequency scale diagram can be determined once before the fatigue test, that is, when the fatigue times is 0, and then the button is pressed 500 times to determine the time-frequency scale diagram when the fatigue times is 500 times, and then the button is pressed 500 times to determine the time-frequency scale diagram when the fatigue times is 1000 times, and then the button is pressed 500 times to determine the time-frequency scale diagram when the fatigue times is 1500 times.

[0102] After obtaining the fatigue test parameters and the time-frequency scaling diagram under different fatigue times, the fatigue life prediction operation can be performed according to the fatigue test parameters and the time-frequency scaling diagram under different fatigue times to obtain the fatigue life of the target composite material. Figure 6 , which may include:

[0103] S31. Convert the time-frequency scale diagram under different fatigue times into a pixel matrix.

[0104] Specifically, in order to ensure that the model can better process the image, the image can be resized to obtain an image that meets the model input requirements.

[0105] Then, in order to reduce the amount of data processing of the model, the image can be grayed out to simplify the matrix and increase the calculation speed.

[0106] The time-frequency scale image is converted into a pixel matrix through grayscale processing, and the grayscale value of each pixel can be extracted from the pixel matrix.

[0107] It should be noted that, since there are multiple time-frequency scale images, grayscale processing can be performed on each image to obtain a corresponding pixel matrix.

[0108] S32. Perform a normalization operation on the pixel matrix to obtain a target image vector.

[0109] Specifically, normalization is a way to simplify calculations, that is, to transform a dimensional expression into a dimensionless expression and become a scalar. Normalizing the pixel matrix can avoid the influence of the numerical value in the pixel matrix on subsequent processing. During normalization, the pixel values ​​in the pixel matrix are normalized to the range of 0-1, and then the normalized values ​​are flattened into a one-dimensional vector to obtain the target image vector.

[0110] It should be noted that since there are multiple time-frequency scale diagrams, each time-frequency scale diagram corresponds to a target image vector. Multiple target image vectors can be concatenated into one vector and input into the model, or multiple target image vectors can be input into the model separately.

[0111] S33, calling a fatigue life prediction model, processing the fatigue test parameters and the target image vector, and obtaining the fatigue life of the target composite material.

[0112] Specifically, the fatigue life prediction model can be as follows Figure 7 As shown, the input is the loading frequency, load amplitude, stress ratio, and the target image vector obtained after image processing of the full-field deformation map (i.e., the time-frequency scale map mentioned above).

[0113] The fatigue life prediction model is a three-layer BP network. In the three-layer BP network, the input layer has four layers, namely x1, x2, x3 and x4. 1 、x 2 、x 3 and x 4 They correspond to loading frequency, load amplitude, stress ratio and target image vector respectively. The number of hidden layer nodes is the preset number, such as 8, H1-H8, and the output parameter y is 1. The Sigmoid function is used as the transfer function, the output function of the output layer is the linear function purelin, and the training function is the trainlm function to improve the processing ability and accuracy of the model.

[0114] In addition, the following network training parameters are set: the maximum number of training times is 50000, the learning rate is set to 0.0001, and the minimum error of the training target is set to 0.0000001. The network is trained using the training samples as the input of the network and the corresponding composite material fatigue life as the output of the network.

[0115] In the embodiment of the present invention, the fatigue life prediction model is obtained by training with training samples. Specifically, the training process of the fatigue life prediction model includes:

[0116] 1) Obtain training samples.

[0117] The training samples include input parameters and fatigue life reference values; the input parameters include test parameter samples and image vector samples. The test parameter samples refer to the above fatigue test parameters and also include the three data of loading frequency, load amplitude and stress ratio. The image vector samples refer to the above target image vector.

[0118] Fatigue life reference values ​​are obtained based on tests or simulations.

[0119] The training samples of the present invention are determined in two ways, respectively as follows:

[0120] 1.1. Perform a preset fatigue test operation on an actual composite material through a fatigue test device to obtain a first training sample.

[0121] Specifically, some real composite materials can be selected as test samples, and fatigue testing operations can be performed using a fatigue testing device in accordance with the above-mentioned testing method to obtain time-frequency scaling diagrams corresponding to different fatigue testing parameters. The time-frequency scaling diagrams can be converted into target image vectors, and the fatigue testing parameters and the target image vectors can be combined to obtain the first training sample.

[0122] In actual scenarios, the training of neural network models requires a certain number of samples with certain variability as input. For the signals captured by fatigue testing equipment and high-speed cameras in laboratory environments, they themselves have certain variability due to the influence of environmental factors, and only need to repeat the full-field deformation measurement under fatigue load a certain number of times.

[0123] 1.2. Perform finite element simulation on the simulated composite material to obtain a second training sample.

[0124] Specifically, since the test time of fatigue testing is relatively long in an experimental environment, in order to reduce the time and improve the efficiency, a simulation method can be used to obtain training samples.

[0125] Specifically, the actual composite material is simulated to obtain the simulated composite material, and then the test parameter sample is set, and the simulated composite material is subjected to finite element simulation using the test parameter sample. The simulation process is not affected by the external environment. In order to ensure that the simulation result is closer to the actual experimental environment, a data enhancement operation can be performed by adding noise with a preset signal-to-noise ratio to obtain a scale map sample. The scale map sample can refer to the above-mentioned time-frequency scale map. After obtaining the scale map sample, the scale map sample is converted into an image vector sample, and the test parameter sample and the image vector sample are combined to obtain a second training sample.

[0126] In this embodiment, during the simulation process, noise is introduced to simulate the actual experimental environment, thereby creating a data set with variability for training.

[0127] 1.3. Combine the first training sample and the second training sample to obtain a training sample.

[0128] Specifically, the first training sample and the second training sample are directly aggregated to obtain a training sample.

[0129] It should be noted that, in practical applications, training samples may be obtained by using both the test and simulation methods at the same time, or by using only one of the test and simulation methods, depending on the actual configuration.

[0130] 2) Using the training samples, the fatigue life prediction model is trained until a training stop condition is met.

[0131] During training, the above training samples are used as the input of the network, and the corresponding fatigue life is used as the output to establish the data set required for neural network training to train the network. The trained network can be used to predict the fatigue life of composite materials.

[0132] During training, the maximum number of training times is 50,000, the learning rate is set to 0.0001, and the minimum error of the training target is set to 0.0000001. Training can be stopped when the number of training times reaches 50,000 or the loss function changes very little.

[0133] The trained network can be used to predict the fatigue life of composite materials. Specifically, the fatigue test parameters and the target image vector are input into the trained neural network, and the output of the network is the fatigue life prediction of the composite material.

[0134] In this embodiment, the digital image correlation method is used to measure the full-field deformation of the composite material under fatigue load, and the neural network is used to learn the full-field deformation evolution of the composite material during fatigue. The nonlinear mapping relationship between the input signal and the output result can be established when the characteristics of the composite material plate are unknown, which is fast and accurate. The end-to-end architecture of machine learning can achieve better performance than the classical algorithm.

[0135] Reference Figure 8 , the whole implementation process of fatigue life prediction is given. Specifically, the fatigue test platform based on digital image correlation (refer to Figure 2 ), through fatigue testing and digital image correlation method calculation, the full-field deformation information of composite material fatigue is obtained, such as Figure 5The time-frequency scaling diagram (also called full-field deformation diagram) under different fatigue times is shown. The stress ratio, frequency and other parameters and the vector corresponding to the time-frequency scaling diagram are used as the input data of the neural network, and the corresponding composite material fatigue life is used as the output of the network to establish the data set required for neural network training to train the network. The trained network can be used to predict the fatigue life of composite materials.

[0136] In actual scenarios, the data set required for neural network training consists of vectors, stress ratios, frequencies and other parameters corresponding to the time-frequency scale diagrams under different fatigue times. By testing the fatigue properties of multiple groups of composite materials, a total of 100 samples are obtained for training the neural network. This can be obtained through two methods: experiments and simulations.

[0137] Then build a neural network model or machine learning model. For example, in the model, set the number of hidden layer nodes to 8, use the Sigmoid function as the transfer function, and use the output function of the output layer as a three-layer BP network with the linear function purelin. Select the trainlm function as the training function, and set the following network training parameters: the maximum number of training times is 50,000, the learning rate is set to 0.0001, and the training target minimum error is set to 0.0000001.

[0138] The network is trained by taking the dataset as the input and the corresponding composite material fatigue life as the output.

[0139] Use the trained neural network to predict the fatigue life of composite materials: After the neural network training is completed, the vector, stress ratio, frequency and other parameters of the full-field deformation diagram under different fatigue times are extracted and input into the trained neural network. At this time, the output of the network is the prediction of the fatigue life of the composite material.

[0140] In this embodiment, a composite material fatigue life prediction method based on machine learning and digital image correlation is provided. The method has simple implementation steps and can quickly and accurately predict the composite material fatigue life. It has broad application prospects.

[0141] Based on the above-mentioned embodiment of the fatigue life prediction method, in another implementation of the present invention, a fatigue life prediction device is provided, referring to Fig. 9 , which may include:

[0142] A parameter determination module 101 is used to determine fatigue test parameters;

[0143] A fatigue testing module 102 is used to perform a fatigue testing operation on a target composite material using the fatigue testing parameters to obtain a time-frequency scaling diagram of the target composite material under different fatigue times;

[0144] The life prediction module 103 is used to perform fatigue life prediction operations according to the fatigue test parameters and the time-frequency scaling diagrams under different fatigue times to obtain the fatigue life of the target composite material.

[0145] In one implementation, the fatigue testing module 102 includes:

[0146] The times determination submodule is used to determine the times of fatigue;

[0147] A fatigue test submodule, for controlling the fatigue test device to perform a preset fatigue test action on the target composite material using the fatigue test parameters for each fatigue number, wherein the execution number of the preset fatigue test action is the fatigue number;

[0148] The graph determination submodule is used to determine the time-frequency scale graph of the target composite material after the preset fatigue test action is executed.

[0149] In one implementation, the graph determination submodule includes:

[0150] A data acquisition unit, used to acquire image data of the target composite material;

[0151] The deformation analysis unit is used to perform a deformation analysis operation on the image data to obtain a time-frequency scale diagram.

[0152] In one implementation, the life prediction module 103 includes:

[0153] A conversion submodule, used to convert the time-frequency scale map under different fatigue times into a pixel matrix;

[0154] A normalization submodule, used for performing a normalization operation on the pixel matrix to obtain a target image vector;

[0155] The life prediction submodule is used to call the fatigue life prediction model, process the fatigue test parameters and the target image vector, and obtain the fatigue life of the target composite material.

[0156] In one implementation, a model generation module is also included, and the model generation module includes:

[0157] A sample acquisition submodule is used to acquire training samples; the training samples include input parameters and fatigue life reference values; the input parameters include test parameter samples and image vector samples;

[0158] A training submodule is used to train the fatigue life prediction model using the training samples until the training stop condition is met; in the fatigue life prediction model, the number of hidden layer nodes is a preset number; the transfer function is a Sigmoid function; the output function of the output layer is a linear function; and the training function is a trainlm function.

[0159] In one implementation, the sample acquisition submodule includes:

[0160] A first acquisition unit, configured to perform a preset fatigue test operation on an actual composite material through a fatigue test device to obtain a first training sample;

[0161] A second acquisition unit is used to perform a finite element simulation operation on the simulated composite material to obtain a second training sample;

[0162] A combining unit is used to combine the first training sample and the second training sample to obtain a training sample.

[0163] In one implementation, the second obtaining unit includes:

[0164] A simulation subunit is used to perform finite element simulation on the simulated composite material using the test parameter sample, and during the simulation process, a data enhancement operation is performed by adding noise with a preset signal-to-noise ratio to obtain a scale map sample;

[0165] a conversion subunit, configured to convert the scale map samples into image vector samples;

[0166] The combining subunit is used to combine the test parameter sample and the image vector sample to obtain a second training sample.

[0167] In this embodiment, fatigue test parameters are determined, and fatigue test operations are performed on the target composite material using the fatigue test parameters to obtain a time-frequency scale diagram of the target composite material at different fatigue times. A fatigue life prediction operation is performed based on the fatigue test parameters and the time-frequency scale diagram at different fatigue times to obtain the fatigue life of the target composite material. That is, the fatigue life test is achieved through the present invention, and in the present invention, the time-frequency scale diagram of the target composite material at different fatigue times is used, and the data can represent the usage data of the target composite material in the actual scene. Therefore, fatigue life prediction using this data can improve the accuracy of life prediction.

[0168] It should be noted that, for the working process of each module, sub-module, unit and sub-unit in the embodiment of the present invention, please refer to the corresponding description in the above embodiment, which will not be repeated here.

[0169] Another embodiment of the present invention provides an electronic device, comprising at least one processor and a memory connected to the processor, wherein:

[0170] The memory is used to store computer programs;

[0171] The processor is used to execute the computer program so that the electronic device can implement the above fatigue life prediction method.

[0172] refer to Fig.10 As shown, it shows a schematic diagram of the structure of an electronic device suitable for implementing the embodiment of the present application. The electronic device in the embodiment of the present application may include but is not limited to fixed terminals such as mobile phones, laptops, PDAs (personal digital assistants), PADs (tablet computers), desktop computers, etc. Fig.10 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0173] like Fig.10 As shown, the electronic device may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 to a random access memory (RAM) 603. When the electronic device is powered on, various programs and data required for the operation of the electronic device are also stored in the RAM 603. The processing device 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0174] Typically, the following devices may be connected to the I / O interface 605: an input device 606 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 608 including, for example, a memory card, a hard disk, etc.; and a communication device 609. The communication device 609 may allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Although Fig.10 An electronic device having various devices is shown, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0175] Reference Fig.11 , an embodiment of the present invention provides a fatigue life prediction system, including an electronic device 202 for executing a fatigue life prediction method, and further comprising:

[0176] Fatigue testing device 201 and camera 203; camera 203 can be Figure 2 Digital image acquisition in high-speed camera.

[0177] The fatigue testing device is used to perform fatigue testing operations based on the control of the electronic equipment;

[0178] The camera is used to collect image data of the target composite material and send the image data to the electronic device.

[0179] Among them, the electronic device communicates with the fatigue testing device and the camera respectively, and the electronic device is used to control the fatigue testing device to perform fatigue testing, such as sending fatigue test parameters and fatigue times to the fatigue testing device, the fatigue testing device performs fatigue testing operations based on these parameters, and provides real-time feedback on the test progress. When the test times are reached, such as 0 times, 500 times, 1000 times, and 1500 times, instructions can be sent to the electronic device, and the electronic device controls the camera to collect image data, performs deformation analysis on the image data, and obtains a time-frequency scale diagram.

[0180] For the specific working process of the electronic equipment, fatigue test device and camera, please refer to the corresponding instructions above.

[0181] An embodiment of the present application also provides a computer program product including computer-readable instructions. When the computer-readable instructions are executed on an electronic device, the electronic device implements any fatigue life prediction method provided in the embodiment of the present application.

[0182] A computer-readable storage medium is also provided in an embodiment of the present application. The storage medium carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement any fatigue life prediction method provided in the embodiment of the present application.

[0183] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A fatigue life prediction method, characterized in that: include: Determine fatigue test parameters; Using the fatigue test parameters, a fatigue test operation is performed on the target composite material to obtain a time-frequency scaling diagram of the target composite material under different fatigue times; The time-frequency scale diagram is obtained by performing deformation analysis on the image data of the target composite material; According to the fatigue test parameters and the time-frequency scale diagram under different fatigue times, a fatigue life prediction operation is performed to obtain the fatigue life of the target composite material, which includes: converting the time-frequency scale diagram under different fatigue times into a pixel matrix, normalizing the pixel matrix to obtain a target image vector, calling a fatigue life prediction model, processing the fatigue test parameters and the target image vector to obtain the fatigue life of the target composite material; the fatigue life of the target composite material refers to the predicted number of uses; the fatigue life prediction model is a three-layer BP network, in which the inputs are x1, x2, x3 and x4, respectively, and x1, x2, x3 and x4 correspond to the loading frequency, load amplitude, stress ratio and target image vector respectively; in the fatigue test operation, the total number of fatigue times is the number that matches the bearing capacity of the target composite material; The fatigue life prediction model is obtained by training using training samples; the training samples include: A first training sample is obtained by performing a preset fatigue test operation on an actual composite material through a fatigue test device, and a second training sample is obtained by performing a finite element simulation operation on a simulated composite material; during the finite element simulation process, data enhancement operation is performed by adding noise with a preset signal-to-noise ratio.

2. The fatigue life prediction method according to claim 1, characterized in that: Using the fatigue test parameters, a fatigue test operation is performed on the target composite material to obtain a time-frequency scaling diagram of the target composite material under different fatigue times, including: Determine fatigue times; For each fatigue number, using the fatigue test parameters, controlling the fatigue test device to perform a preset fatigue test action on the target composite material, the number of executions of the preset fatigue test action being the fatigue number; After the preset fatigue test action is executed, a time-frequency scaling diagram of the target composite material is determined.

3. The fatigue life prediction method according to claim 2, characterized in that: Determining a time-frequency scaling diagram of the target composite material comprises: Acquiring image data of the target composite material; A deformation analysis operation is performed on the image data to obtain a time-frequency scale diagram.

4. The fatigue life prediction method according to claim 1, characterized in that: The training process of the fatigue life prediction model includes: Acquire training samples; the training samples include input parameters and fatigue life reference values; the input parameters include test parameter samples and image vector samples; The fatigue life prediction model is trained using the training samples until the training stop condition is met; in the fatigue life prediction model, the number of hidden layer nodes is a preset number; the transfer function is a Sigmoid function; the output function of the output layer is a linear function; and the training function is a trainlm function.

5. The fatigue life prediction method according to claim 4, characterized in that: Get training samples, including: Performing a preset fatigue test operation on the actual composite material by a fatigue test device to obtain a first training sample; Performing finite element simulation operation on the simulated composite material to obtain a second training sample; The first training sample and the second training sample are combined to obtain a training sample.

6. The fatigue life prediction method according to claim 5, characterized in that: Perform finite element simulation on the simulated composite material to obtain a second training sample, including: Using the test parameter samples, finite element simulation is performed on the simulated composite material. During the simulation process, data enhancement operation is performed by adding noise with a preset signal-to-noise ratio to obtain a scale map sample; converting the scale map samples into image vector samples; The test parameter sample and the image vector sample are combined to obtain a second training sample.

7. A fatigue life prediction device, characterized in that: include: A parameter determination module, used for determining fatigue test parameters; A fatigue test module, used to perform a fatigue test operation on a target composite material using the fatigue test parameters, and obtain a time-frequency scaling diagram of the target composite material under different fatigue times; The time-frequency scale diagram is obtained by performing deformation analysis on the image data of the target composite material; A life prediction module is used to perform fatigue life prediction operations according to the fatigue test parameters and the time-frequency scaling diagrams under different fatigue times to obtain the fatigue life of the target composite material, which includes: converting the time-frequency scaling diagrams under different fatigue times into a pixel matrix, normalizing the pixel matrix to obtain a target image vector, calling a fatigue life prediction model, processing the fatigue test parameters and the target image vector to obtain the fatigue life of the target composite material; the fatigue life of the target composite material refers to the predicted number of uses; the fatigue life prediction model is a three-layer BP network, in which the inputs are x1, x2, x3 and x4, respectively, and x1, x2, x3 and x4 correspond to the loading frequency, load amplitude, stress ratio and target image vector respectively; in the fatigue test operation, the total number of fatigue times is the number that matches the bearing capacity of the target composite material; The fatigue life prediction model is obtained by training using training samples; the training samples include: A first training sample is obtained by performing a preset fatigue test operation on an actual composite material through a fatigue test device, and a second training sample is obtained by performing a finite element simulation operation on a simulated composite material; during the finite element simulation process, data enhancement operation is performed by adding noise with a preset signal-to-noise ratio.

8. An electronic device, characterized in that: The method comprises at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs; The processor is used to execute the computer program so that the electronic device can implement the fatigue life prediction method according to any one of claims 1 to 6.

9. A fatigue life prediction system, characterized in that: The electronic device according to claim 8 further comprises: Fatigue testing equipment and cameras; The fatigue testing device is used to perform fatigue testing operations based on the control of the electronic equipment; The camera is used to collect image data of the target composite material and send the image data to the electronic device.

Images