A prediction method for the yield load and damage degree of an explosion-welded composite material

Abstract

The present invention relates to a method for predicting the yield load and damage degree of an explosion-welded composite material, belonging to the technical field of composite material damage prediction, and solves the problems of high cost, poor safety, low accuracy and low efficiency in the damage research of explosion-welded composite materials in the prior art. Obtain the damage image of the composite material to be tested, and input the damage image into the first image recognition model to obtain the number and size of holes and cracks in the image; input the damage image into the second image recognition model to obtain the wavelength and amplitude of the wavy composite interface; obtain the cross-sectional area change rate of the composite material to be tested, and input the cross-sectional area change rate, the number of holes and its minimum and maximum values, the number of cracks and its minimum and maximum values, the wavelength, and the amplitude into the first load damage prediction model to obtain the yield load and damage level of the composite material to be tested. Implement a method for predicting the damage degree of explosion-welded composite materials with low cost, high safety, and accuracy and speed.

Description

Technical Field

[0001] The present invention relates to the technical field of composite material damage prediction, and particularly relates to a method for predicting the yield load and damage degree of an explosively welded composite material. Background Art

[0002] Explosion welding technology is a process that uses the energy generated by explosive detonation to achieve solid-state connection of the same or different materials, and is widely used in the preparation of composite materials. In fields such as aerospace, automotive manufacturing, and petrochemical industry, many key structural components use explosively welded composite materials. If there are undetected damages in these materials, catastrophic accidents may occur during service. Studying material damage can evaluate the safety of the structure in advance, timely discover potential problems and take measures for repair or replacement, and avoid major accidents caused by sudden material failure. For example, if the wing structural components of an aircraft use damaged explosively welded composite materials, they may break during flight due to being unable to withstand the huge aerodynamic force, endangering flight safety. Through the study and detection of material damage, the safety of the wing structure can be ensured and the smooth progress of the flight mission can be guaranteed.

[0003] In the prior art, the methods for studying the damage of explosively welded composite materials mainly include: (1) Experimental research methods, such as tensile tests, nanoindentation tests, etc. By actual welding and testing to analyze the damage characteristics, although it can intuitively reflect the mechanical behavior and damage of the composite material, the experiment requires a large amount of materials and equipment, and involves the use of explosives, with high costs. The stress state of the sample during the test is relatively simple, which is different from the complex stress state that the composite material may bear in actual applications, and may not be able to fully reflect its mechanical properties and damage characteristics. There are safety risks in the explosion welding process and strict safety measures are required. (2) Microstructural analysis methods, such as scanning electron microscopy (SEM), transmission electron microscopy (TEM), X-ray diffraction (XRD), etc. This method can analyze the microstructure and damage characteristics of explosively welded composite materials. Although it can accurately observe the interface microstructure, such as grain refinement, adiabatic shear bands, etc., and provide theoretical support for the damage mechanism, it requires high-precision microanalysis equipment, with high costs, and the microanalysis data volume is large, requiring professional analysis techniques. (3) Numerical simulation methods, through techniques such as finite element analysis, smoothed particle hydrodynamics (SPH), etc., to simulate the damage mechanism during the explosion welding process. Although computer simulation reduces the number of experiments and significantly reduces the R & D cost, it can simulate complex phenomena such as stress, strain, and temperature fields during the welding process, revealing details that are difficult to observe in experiments, but the simulation results depend on the accuracy of the model. Complex physical phenomena may be difficult to fully simulate, requiring high-performance computing resources, and having high requirements for the technical level of the operator. The simulation results need to be verified by experiments to ensure their reliability.

[0004] Therefore, there is a need to provide a method for predicting the damage degree of explosive welded composite materials with low cost, high safety, and accuracy and speed. Summary of the Invention

[0005] In view of the above analysis, the embodiments of the present invention aim to provide a method for predicting the yield load and damage degree of explosive welded composite materials to solve the problems of high cost, poor safety, low accuracy, and low efficiency in the damage research of explosive welded composite materials in the prior art.

[0006] The embodiments of the present invention provide a method for predicting the yield load and damage degree of explosive welded composite materials, which is characterized by including:

[0007] Obtain the damage image of the composite material to be tested, input the damage image into the first image recognition model to obtain the number and size of holes and cracks in the composite material to be tested, and count the minimum and maximum values of the holes and cracks;

[0008] Input the damage image into the second image recognition model to obtain the wavelength and amplitude of the wavy composite interface in the composite material to be tested;

[0009] Obtain the cross-sectional area change rate of the composite material to be tested, and input the cross-sectional area change rate, the number of holes and their minimum and maximum values, the number of cracks and their minimum and maximum values, the wavelength, and the amplitude into the first load damage prediction model to obtain the yield load and damage level of the composite material to be tested.

[0010] Based on a further improvement of the above method, the damage image at least includes , , , X-ray scan images; the first image recognition model is a 3D ResNet network model, the second image recognition model is a model composed of a U-NET model and a regression head, and the first load damage prediction model is a machine learning model.

[0011] Based on a further improvement of the above method, dilated convolution is introduced into each residual block in the 3D ResNet network model, and an SE attention module is introduced after each residual block.

[0012] Based on a further improvement of the above method, the first load damage prediction model is a support vector machine SVM model, which is implemented by hard parameter sharing, and the output parameters are the yield load and damage level respectively;

[0013] The SVM model is trained based on a first sample data set; wherein, each sample data item in the first sample data set includes: cross-sectional area change rate, number of holes and its minimum and maximum values, number of cracks and its minimum and maximum values, wavelength, amplitude, yield load, and damage level.

[0014] Based on a further improvement of the above method, a first sample data set is constructed in the following manner:

[0015] Prepare a variety of composite materials based on explosive welding parameters. For each composite material, obtain its corresponding first sample data in the following manner to form a first sample data set:

[0016] Cut 3 in-situ tensile specimens from the composite material by turning on a machine tool; select 1 in-situ tensile specimen to conduct an in-situ tensile test on an in-situ tensile test platform to obtain the load-displacement curve corresponding to the specimen, and then obtain the yield load; select 1 in-situ tensile specimen to obtain multiple X-ray scan images based on the in-situ tensile test platform, and input them into a second image recognition model to obtain the wavelength and amplitude of the wavy composite interface in the composite material to be tested; select 1 in-situ tensile specimen to conduct an in-situ tensile test on the in-situ tensile test platform to obtain the cross-sectional area change rate, number and size distribution of holes, number and size distribution of cracks of the specimen under different tensile loads. Based on the cross-sectional area change rate, number of holes and its minimum and maximum values, number of cracks and its minimum and maximum values, wavelength, and amplitude of the specimen under each tensile load, mark the damage level of the specimen. Take the cross-sectional area change rate, number of holes and its minimum and maximum values, number of cracks and its minimum and maximum values, wavelength, amplitude, yield load, and damage level of the specimen under each tensile load as a first sample data item.

[0017] Based on a further improvement of the above method, preparing a composite material based on explosive welding parameters includes:

[0018] Input the explosive welding parameters into a second prediction model, and judge whether the output wavelength and amplitude meet the predetermined requirements. If they meet, prepare a composite material based on the explosive welding parameters; if not, use a genetic algorithm to optimize the input parameters, and take the optimized parameters as the explosive welding parameters.

[0019] Based on a further improvement of the above method, the second prediction model is trained based on a second sample data set; wherein, each sample data item in the second sample data set includes: control parameters in each explosive welding process and the corresponding wavelength and amplitude, and the control parameters include: collision speed, collision angle, explosive thickness, mass ratio of explosive per unit area to flyer plate mass, distance between flyer plate and substrate, flyer plate thickness;

[0020] The wavelength and amplitude in each sample data are obtained in the following manner: Select the flyer plate material and the substrate material based on service requirements, and composite the flyer plate and the substrate based on the control parameters to form a specific explosively welded composite plate. Prepare an observation sample based on the composite plate, and conduct optical microscopy on the observation sample to obtain the wavelength and wave amplitude of the wavy interface of the explosive composite interface of the composite plate.

[0021] Based on a further improvement of the above method, the composite material to be tested is obtained by explosively welding a TA1 material plate as the flyer plate and a Q235 material plate as the substrate;

[0022] The preparation of the observation sample based on the composite material includes:

[0023] Cut an observation sample segment from the composite material, mechanically polish the laminated surface of the heterogeneous materials, immerse the observation sample segment in a metallographic etchant for a first preset duration, rinse the surface with clean water and then dry it, and then use a cotton ball to dip into a nitric acid alcohol solution with a volume concentration and wipe it on the surface of the specimen. After etching for a second preset duration, clean and dry it to finally obtain an observation sample.

[0024] Based on a further improvement of the above method, the in-situ tensile test is carried out on the in-situ tensile test platform to obtain the number and size distribution of holes and the number and size distribution of cracks in the specimen under different tensile loads, including:

[0025] Set the initial tensile load and the tensile loading rate;

[0026] Place the in-situ tensile specimen into a mechanical fixing device and apply the initial tensile load. Every time the tensile load increases by 1 N, the in-situ tensile test platform rotates the fixing device to take specimen slices of the in-situ tensile specimen at different angles, input the specimen slices into a first image recognition model to obtain the number and size of holes and cracks corresponding to the specimen slice, and count the minimum and maximum values of holes and cracks in the specimen under the current tensile load.

[0027] Based on a further improvement of the above method, the predetermined requirements include: the wavelength should be in the range of 660 - 690 and the amplitude should be in the range of 60 - 90 ;

[0028] The implementation of the optimization of the input parameters using the genetic algorithm includes:

[0029] B1: Initialize the population, set the population size N and the preset number of iterations T. Each individual in the population includes: collision velocity, collision angle, explosive thickness, mass ratio of explosive per unit area to the mass of the flyer plate, distance between the flyer plate and the substrate, and flyer plate thickness;

[0030] The collision velocity, collision angle, distance between the flying plate and the substrate, and thickness of the flying plate are binary coded, and the thickness of the explosive, the mass ratio of the explosive per unit area to the mass of the flying plate is Gray coded;

[0031] B2: Calculate the fitness value of each individual according to the fitness function, where the fitness function is:

[0032] ,

[0033] where, is the minimum value of the wavelength in the predetermined requirements, is the maximum value of the wavelength in the predetermined requirements, the wavelength obtained by inputting the i-th individual into the second prediction model, is the minimum value of the amplitude in the predetermined requirements, the maximum value of the amplitude in the predetermined requirements, the amplitude obtained by inputting the i-th individual into the second prediction model, is a constant, i = 1, 2, 3,.., N;

[0034] B3: Selection operation, select individuals with higher fitness from the current population as parents based on the tournament algorithm for generating the next generation;

[0035] B4: Crossover operation, for the parameter group with binary coding, use the single-point crossover method, and for the parameter group with Gray coding, use the multi-point crossover method;

[0036] B5: Mutation operation, select the individual with the highest fitness from the population for local search. For each selected individual, generate several neighborhood solutions in its vicinity, search for a solution better than the current solution among the neighborhood solutions. If a better solution is found, replace the current solution, and put the individual after local search back into the population;

[0037] B6: Generate a new population, merge the parent and offspring individuals, select a preset number of individuals as the new generation population according to the fitness value, and determine whether the iteration stop condition is reached. If so, use the optimal individual in the current population as the optimized operating parameter.

[0038] Compared with the prior art, the present invention can at least achieve one of the following beneficial effects:

[0039] 1. The present invention provides a method for predicting the yield load and damage degree of an explosion - welded composite material. By combining the damage research of the explosion - welded composite material with machine - learning techniques, potential laws in the material damage data are mined, thereby discovering the relationship between damage patterns and characteristics that are difficult to detect by traditional methods, and comprehensively understanding the damage situation of the composite material. Compared with traditional experimental research methods, microstructure analysis methods, and numerical simulation methods, the proposed solution of the present invention can quickly and accurately obtain damage results and the yield load corresponding to the current material state, with lower costs and higher safety.

[0040] 2. The present invention provides a method for predicting the yield load and damage degree of an explosion - welded composite material. When preparing the composite material, if the current explosion - welding parameters do not meet the requirements of the predetermined wavelength and amplitude, the genetic algorithm is used to optimize the current explosion - welding parameters to obtain the explosion - welding parameters that meet the requirements of the predetermined wavelength and amplitude. When preparing the composite material, the present invention reduces the preparation difficulty. It is no longer necessary for technicians to adjust according to experience. Only relying on the provided genetic algorithm can quickly obtain the explosion - welding parameters adapted to the current predetermined requirements. By combining neighborhood - solution optimization in the mutation operation of the genetic algorithm, high - quality solutions can be quickly found through efficient local search. This can not only improve the search efficiency and solution quality of the algorithm, but also enhance the robustness and global search ability of the algorithm, thus showing better performance when solving complex optimization problems.

[0041] In the present invention, the above - mentioned technical solutions can also be combined with each other to achieve more preferred combination schemes. Other features and advantages of the present invention will be described in the subsequent specification. Moreover, some advantages can be made obvious from the specification or understood by implementing the present invention. The objectives and other advantages of the present invention can be realized and obtained through the content specifically pointed out in the specification and the drawings. Brief Description of the Drawings

[0042] The drawings are only for the purpose of showing specific embodiments and are not considered as a limitation to the present invention. Throughout the drawings, the same reference signs represent the same components;

[0043] Figure 1 It is an example diagram of a method for predicting the yield load and damage degree of an explosion - welded composite material in an embodiment of the present invention;

[0044] Figure 2 It is an example diagram of the damage - level standard in an embodiment of the present invention. Detailed Description of the Embodiments

[0045] The preferred embodiments of the present invention will be specifically described below in conjunction with the accompanying drawings. The accompanying drawings form a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, rather than to limit the scope of the present invention.

[0046] Hard Parameter Sharing is a common strategy in multi-task learning. Its core idea is that multiple tasks share the parameters of the main part of the model, and only independent parameters are reserved for each task in the output layer. This method is widely used in deep learning and has advantages such as reducing model complexity, accelerating the training process, and reducing the risk of overfitting.

[0047] A specific embodiment of the present invention discloses a method for predicting the yield load and damage degree of an explosion-welded composite material, as Figure 1 shown, including:

[0048] S1: Obtain the damage image of the composite material to be tested, and input the damage image into the first image recognition model to obtain the number and size of holes and cracks in the composite material to be tested, and count the minimum and maximum values of the holes and cracks.

[0049] Exemplarily, the composite material to be tested can be a whole steel-titanium composite plate obtained by explosively welding a TA1 material plate as the flyer plate and a Q235 material plate as the base plate.

[0050] The damage image in the present invention refers to an image obtained by transmission scanning with an X-ray imaging device. The present invention does not specifically limit the number of damage images, but at least includes , , , , and the X-ray scanning images of . It can be understood that the more damage images input, the more accurate the prediction result.

[0051] The first image recognition model is a 3D convolutional neural network, which can directly process three-dimensional slice data, make full use of the spatial correlation between slices in multiple directions, better identify the three-dimensional structural features of holes and cracks, and improve the accuracy of prediction. Common 3D convolutional neural networks include 3D ResNet, 3D U-Net Lite, 3D DenseNet, 3D U-Net, VoxNet, etc. The present invention does not specifically limit the model adopted, and those skilled in the art can select a network model according to needs to realize the identification of holes and cracks. Exemplarily, the present invention adopts a 3D ResNet network model, introduces dilated convolution in each residual block, and introduces an SE attention module after each residual block to improve the prediction accuracy and training efficiency of the 3D ResNet network.

[0052] The training process of the first image recognition model is as follows:

[0053] Obtain the first training sample set for training the first image recognition model. The samples in the first training sample set are image slices obtained by transmission scanning of each X-ray imaging device, and the first training sample set needs to ensure that it covers holes and cracks of different sizes and shapes. Manually annotate each slice, and the annotation content includes the position, quantity, and size information of the holes and cracks. Professional image annotation tools such as OpenCV, Scikit-image, LabelMe, CVAT, ITK-SNAP, etc. can be used to mark the boundaries or key points of each hole and crack, and at the same time record the corresponding quantity and size data. And operations such as normalization and denoising can be performed as needed to improve the recognition accuracy.

[0054] Divide the annotated data set into a training set, a validation set, and a test set, and finally obtain the first image recognition model that can be used to recognize the quantity and size of holes and cracks based on the training set, the validation set, and the test set. Generally speaking, the training set is used for model training, the validation set is used to adjust the hyperparameters of the model and monitor the training process to prevent overfitting, and the test set is used to evaluate the final performance of the model. Divide according to a certain ratio, such as 7:2:1 or 8:1:1.

[0055] The present invention does not specifically limit operations such as the network structure design, loss function, optimizer selection, and learning rate adjustment of the first image recognition model, as long as it can achieve the recognition of the quantity and size of holes and cracks.

[0056] Obtain multiple damaged images of the composite material to be tested, input the multiple damaged images into the first image recognition model to obtain the quantity and size of the holes and cracks in the composite material to be tested, and statistically calculate the minimum and maximum values of the holes and cracks based on the current prediction results.

[0057] S2: Input the damaged image into the second image recognition model to obtain the wavelength and amplitude of the wavy composite interface in the composite material to be tested.

[0058] The second image recognition model is a model combined by a U-NET model and a regression head, and is used to recognize the wavelength and amplitude of the wavy composite interface in the damaged image. The input damaged image is the X-ray scan image.

[0059] The training process of the second image recognition model is as follows:

[0060] Obtain the second training sample set for training the second image recognition model. The samples in the second training sample set are each damaged image with a wavy composite interface. Manually annotate the wavelength and amplitude information, and preprocess the images, including operations such as resizing the images, normalizing pixel values, and grayscale conversion. Resizing the images can unify all images to a fixed size for easy model processing; normalizing pixel values can map pixel values to or interval to accelerate the model convergence speed; grayscale conversion can convert color images into grayscale images, reduce the data dimension, and at the same time can also highlight the texture features of the images, which is beneficial to the recognition of the wavy interface.

[0061] Divide the annotated data set into a training set, a validation set, and a test set to implement the training of the second image recognition model, and finally obtain a network that can accurately output the wavelength and amplitude of the wavy composite interface.

[0062] The present invention does not make specific limitations on operations such as the network structure design, loss function, optimizer selection, and learning rate adjustment of the second image recognition model, as long as it can accurately output the wavelength and amplitude of the wavy composite interface.

[0063] S3: Obtain the cross-sectional area change rate of the composite material to be tested, and input the cross-sectional area change rate, the number of holes and its minimum and maximum values, the number of cracks and its minimum and maximum values, the wavelength, and the amplitude into the first load damage prediction model to obtain the yield load and damage level of the composite material to be tested.

[0064] The first load damage prediction model is a machine learning model, such as a neural network model, a decision tree model, a support vector machine model, a convolutional neural network (CNN) model, a recurrent neural network (RNN) model, a generative adversarial network (GAN) model, etc.

[0065] Exemplarily, the present invention adopts a support vector machine SVM model, adaptively modifies its structure in a hard parameter manner, limits the input parameters to the cross-sectional area change rate, the number of holes and its minimum and maximum values, the number of cracks and its minimum and maximum values, the wavelength, and the amplitude, and the output parameters are the yield load and damage level. The SVM model is trained based on the first sample data set; wherein, each data item in the first sample data set includes: the cross-sectional area change rate, the number of holes and its minimum and maximum values, the number of cracks and its minimum and maximum values, the wavelength, the amplitude, the yield load, and the damage level.

[0066] Among them, the damage level is obtained based on the cross-sectional area change rate, the number of holes and their minimum and maximum values, the number of cracks and their minimum and maximum values, the wavelength, and the amplitude annotation. Different damage level criteria can be quantified for different composite materials. Exemplarily, the composite material to be measured selected in the present invention is a whole steel-titanium composite plate composed of a TA1 material plate and a Q235 material plate constructed according to a predetermined requirement. The damage level criteria under specified wavelength requirements and specified amplitude requirements are as Figure 2 shown.

[0067] Among them, the first sample data set is constructed in the following manner:

[0068] Based on explosive welding parameters, a variety of composite materials are prepared. For each composite material, the corresponding first sample data is obtained in the following manner to form a first sample data set:

[0069] At least 3 in-situ tensile specimens are intercepted from the composite material by turning on a machine tool;

[0070] One in-situ tensile specimen is selected to conduct an in-situ tensile test on an in-situ tensile test platform to obtain the load-displacement curve corresponding to the specimen, and then the yield load is obtained;

[0071] One in-situ tensile specimen is selected to obtain multiple X-ray scan images based on the in-situ tensile test platform, and they are input into the second image recognition model to obtain the wavelength and amplitude of the wavy composite interface in the composite material to be measured;

[0072] One in-situ tensile specimen is selected to conduct an in-situ tensile test on an in-situ tensile test platform to obtain the cross-sectional area change rate, the number and size distribution of holes, and the number and size distribution of cracks of the specimen under different tensile loads. Based on the cross-sectional area change rate, the number of holes and their minimum and maximum values, the number of cracks and their minimum and maximum values, as well as the wavelength, amplitude, yield load, and damage level of the specimen under each tensile load, the cross-sectional area change rate, the number of holes and their minimum and maximum values, the number of cracks and their minimum and maximum values, as well as the wavelength, amplitude, yield load, and damage level of the specimen under each tensile load are used as a first sample data.

[0073] Exemplarily, when intercepting in-situ tensile specimens from the composite material, in order to prevent special situations such as accidental damage to the material during the test, several more in-situ tensile specimens can be intercepted for backup to ensure that the first sample data can be generated.

[0074] Among them, the cross-sectional area change rate S refers to the ratio of the initial cross-sectional area A1 of the in-situ tensile specimen intercepted from the composite material to the change in the cross-sectional area A2 under different tensile loads during the in-situ tensile test. Specifically:

[0075] 。

[0076] Exemplarily, the composite material can be a monolithic steel-titanium composite plate obtained by explosively welding a TA1 material plate with a length of 300 mm, a width of 300 mm, and a height of 10 mm as the flying plate and a Q235 material plate with a length of 300 mm, a width of 300 mm, and a height of 30 mm as the substrate. The in-situ tensile specimen intercepted is a specimen segment with a length of 20 mm, a width of 20 mm, and a height of 10 mm, where the thickness of the TA1 and Q235 materials is 5 mm each.

[0077] The in-situ tensile test platform consists of an X-ray source, a thermo-mechanical coupling in-situ loading device, a detector, and a three-dimensional reconstruction system. The main model of the X-ray microscope is XPloreVista2000 4D, and the spatial resolution is ,and the microtube voltage is between The thermo-mechanical coupling in-situ loading device consists of two fixture structures, one end is fixed, and the other end applies a load.

[0078] The in-situ tensile test is carried out based on the in-situ tensile test platform to obtain the number and size distribution of holes and the number and size distribution of cracks in the specimen under different tensile loads, including:

[0079] Set the initial tensile load and tensile loading rate;

[0080] Place the in-situ tensile specimen into the mechanical fixing device and apply the initial tensile load. For every 1 N increase in the tensile load, the in-situ tensile test platform rotates the fixing device to photograph the specimen slices of the in-situ tensile specimen at different angles. Input the specimen slices into the first image recognition model to obtain the number and size of holes and cracks corresponding to the specimen slices, and count the minimum and maximum values of holes and cracks in the specimen under the current tensile load.

[0081] In order to immediately capture the propagation of internal cracks at the material interface, the present invention constantly controls the tensile loading rate at Place the tensile specimen into the mechanical fixing device and apply the tensile load. When the predetermined load threshold is reached, stop loading and rotate the fixing device to photograph the specimen from different angles, thereby obtaining slices at different angles. It can be understood that those skilled in the art can set the tensile loading rate according to experience, and the predetermined load threshold is the yield load obtained from another in-situ tensile specimen.

[0082] Prepare the composite material based on the preset explosive welding parameters, including:

[0083] Input the explosive welding parameters into the second prediction model, and determine whether the output wavelength and amplitude meet the predetermined requirements. If they meet, prepare the composite material based on the explosive welding parameters; if not, use the genetic algorithm to optimize the input parameters and use the optimized parameters as the explosive welding parameters.

[0084] Among them, the predetermined requirements include: the wavelength should be in the range of 660 - 690 and the amplitude should be in the range of 60 - 90. The predetermined requirements can be set by technicians according to actual needs.

[0085] The second prediction model is trained based on the second sample dataset; among them, each sample data in the second sample dataset includes: the control parameters in each explosive welding process and the corresponding wavelength and amplitude, and the control parameters include: collision velocity, collision angle, explosive thickness, mass ratio of explosive per unit area to flyer plate mass, distance between flyer plate and substrate, flyer plate thickness;

[0086] The wavelength and amplitude in each sample data are obtained through the following methods: select the flyer plate material and substrate material based on business requirements, and compound the flyer plate and the substrate based on the control parameters to form a specific explosive welding composite plate. Prepare an observation sample based on the composite plate, and conduct optical microscopy on the observation sample to obtain the wavelength and wave amplitude of the wavy interface of the explosive composite interface of the composite plate.

[0087] The preparation of the observation sample based on the composite material includes:

[0088] Cut an observation sample segment from the composite material, mechanically polish the laminated surface of the heterogeneous materials, immerse the observation sample segment in the metallographic etching agent for a first preset time, rinse the surface with clean water and then dry it, and then use a cotton ball to stick the nitric acid alcohol solution with a volume concentration of to wipe the surface of the sample, etch for a second preset time and then clean and dry to finally obtain the observation sample. Exemplarily, the first preset time is 15s and the second preset time is 10s.

[0089] The optimization of the input parameters by using the genetic algorithm includes:

[0090] B1: Initialize the population, set the population size N and the preset number of iterations T. Each individual in the population includes: collision velocity, collision angle, explosive thickness, mass ratio of explosive per unit area to flyer plate mass, distance between flyer plate and substrate, flyer plate thickness;

[0091] The collision velocity, collision angle, distance between the flying plate and the substrate, and thickness of the flying plate are binary-coded, and the thickness of the explosive, the mass ratio of the explosive per unit area to the mass of the flying plate is Gray-coded;

[0092] B2: Calculate the fitness value of each individual according to the fitness function, where the fitness function is:

[0093] ,

[0094] where, is the minimum value of the wavelength in the predetermined requirement, is the maximum value of the wavelength in the predetermined requirement, the wavelength obtained by inputting the i-th individual into the second prediction model, is the minimum value of the amplitude in the predetermined requirement, the maximum value of the amplitude in the predetermined requirement, the amplitude obtained by inputting the i-th individual into the second prediction model, is a constant, i = 1, 2, 3,.., N;

[0095] B3: Selection operation, select individuals with higher fitness from the current population based on the tournament algorithm as parents for generating the next generation;

[0096] B4: Crossover operation, for the parameter group with binary coding, adopt single-point crossover method, and for the parameter group with Gray coding, adopt multi-point crossover method;

[0097] B5: Mutation operation, select the individual with the highest fitness from the population for local search. For each selected individual, generate several neighborhood solutions near it, search for a solution better than the current solution among the neighborhood solutions. If a better solution is found, replace the current solution and put the individual after local search back into the population;

[0098] Preferably, the neighborhood solutions are generated by the following method:

[0099] ,

[0100] where, is the newly generated neighborhood solution, is the current individual, , represents rounding up, t is the current iteration number, and T is the preset iteration number;

[0101] B6: Generate a new population, merge the parent and offspring individuals, select a preset number of individuals as the new generation population according to the fitness value, and determine whether the iteration stop condition is reached. If so, use the optimal individual in the current population as the optimized operating parameter.

[0102] Understandably, the implementation of initialization operations, crossover operations, mutation operations, etc. in genetic algorithms is common knowledge in the art, and the present invention does not make specific limitations here.

[0103] Compared with the prior art, a method for predicting the yield load and damage degree of an explosively welded composite material provided in this embodiment combines the damage research of explosively welded composite materials with machine learning technology. By mining the potential laws in the material damage data, the relationship between damage patterns and characteristics that are difficult to detect by traditional methods can be discovered, and the damage situation of the composite material can be understood more comprehensively. Compared with traditional experimental research methods, microstructure analysis methods, and numerical simulation methods, the solution proposed by the present invention can quickly and accurately obtain the damage results and the yield load corresponding to the current material state, with lower cost and higher safety. When preparing the composite material, if the current explosive welding parameters do not meet the requirements of the predetermined wavelength and amplitude, a genetic algorithm is used to optimize the current explosive welding parameters to obtain the explosive welding parameters that meet the requirements of the predetermined wavelength and amplitude. When preparing the composite material, the present invention reduces the preparation difficulty and no longer requires technicians to adjust according to experience. Only relying on the provided genetic algorithm can quickly obtain the explosive welding parameters that adapt to the current predetermined requirements. By combining neighborhood solution optimization in the mutation operation of the genetic algorithm, high-quality solutions can be quickly found through efficient local search, which can not only improve the search efficiency and solution quality of the algorithm, but also enhance the robustness and global search ability of the algorithm, thus showing better performance when solving complex optimization problems.

[0104] Those skilled in the art can understand that all or part of the processes for implementing the methods of the above embodiments can be completed by instructing relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. Among them, the computer-readable storage medium is a magnetic disk, an optical disc, a read-only memory, or a random access memory, etc.

[0105] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention.

Claims

1. A method for predicting the yield load and damage degree of an explosively welded composite material, characterized in that Including: Obtain the damage image of the composite material to be tested, input the damage image into the first image recognition model to obtain the quantity and size of holes and cracks in the composite material to be tested, and count the minimum and maximum values of the holes and cracks; Input the damage image into the second image recognition model to obtain the wavelength and amplitude of the wavy composite interface in the composite material to be tested; Obtain the cross-sectional area change rate of the composite material to be tested, and input the cross-sectional area change rate, the quantity of holes and its minimum and maximum values, the quantity of cracks and its minimum and maximum values, the wavelength, and the amplitude into the first load damage prediction model to obtain the yield load and damage level of the composite material to be tested; The damage image at least includes , , , X-ray scan images; the first image recognition model is a 3D ResNet network model, the second image recognition model is a model composed of a U-NET model and a regression head, and the first payload damage prediction model is a machine learning model; Introduce dilated convolution into each residual block in the 3D ResNet network model, and introduce an SE attention module after each residual block.

2. The prediction method for the yield load and damage degree of an explosion-welded composite material according to claim 1, wherein, The first load damage prediction model is a support vector machine (SVM) model, which is implemented by hard parameter sharing, and the output parameters are the yield load and damage level respectively; The SVM model is trained based on the first sample data set; wherein, each data item in the first sample data set includes: cross-sectional area change rate, quantity of holes and its minimum and maximum values, quantity of cracks and its minimum and maximum values, wavelength, amplitude, yield load, and damage level.

3. The prediction method for the yield load and damage degree of an explosion welded composite material according to claim 2, characterized in that, Construct the first sample data set in the following manner: Prepare various composite materials based on explosive welding parameters. For each composite material, obtain the corresponding first sample data in the following manner to form the first sample data set: Cut 3 in-situ tensile specimens from the composite material by turning on a machine tool; Select 1 in-situ tensile specimen to conduct an in-situ tensile test based on the in-situ tensile test platform to obtain the load-displacement curve corresponding to the specimen, and further obtain the yield load; Select 1 in-situ tensile specimen to obtain multiple X-ray scan images based on the in-situ tensile test platform, and input them into the second image recognition model to obtain the wavelength and amplitude of the wavy composite interface in the composite material to be tested; Select 1 in-situ tensile specimen to conduct an in-situ tensile test based on the in-situ tensile test platform to obtain the cross-sectional area change rate, the quantity and size distribution of holes, and the quantity and size distribution of cracks of the specimen under different tensile loads. Mark the damage level of the specimen based on the cross-sectional area change rate, the quantity of holes and its minimum and maximum values, the quantity of cracks and its minimum and maximum values, wavelength, and amplitude of the specimen under each tensile load. Take the cross-sectional area change rate, the quantity of holes and its minimum and maximum values, the quantity of cracks and its minimum and maximum values, wavelength, amplitude, yield load, and damage level of the specimen under each tensile load as a first sample data item.

4. The prediction method for the yield load and damage degree of an explosion-welded composite material according to claim 3, characterized in that Prepare the composite material based on the explosive welding parameters, including: Input the explosive welding parameters into the second prediction model to judge whether the output wavelength and amplitude meet the predetermined requirements. If they meet, prepare the composite material based on the explosive welding parameters; if not, use the genetic algorithm to optimize the explosive welding parameters, and use the optimized parameters as the explosive welding parameters.

5. A method for predicting the yield load and damage degree of an explosively welded composite material according to claim 4, characterized in that The second prediction model is trained based on a second sample data set; wherein, each sample data in the second sample data set includes: control parameters in each explosion welding process and the corresponding wavelength and amplitude, and the control parameters include: collision speed, collision angle, explosive thickness, mass ratio of explosive per unit area to flyer plate mass, distance between flyer plate and substrate, and flyer plate thickness; The wavelength and amplitude in each sample data are obtained through the following method: Select the flyer plate material and substrate material based on service requirements, and compound the flyer plate and the substrate based on the control parameters to form a specific explosion welded composite plate. Prepare an observation sample based on the composite plate, and perform optical microscopy on the observation sample to obtain the wavelength and wave amplitude of the wavy interface of the explosion composite interface of the composite plate.

6. The prediction method for the yield load and damage degree of an explosion welded composite material according to claim 5, characterized in that The composite material to be tested is obtained by compounding a TA1 material plate as the flyer plate and a Q235 material plate as the substrate through explosion welding technology; Preparing the observation sample based on the composite material includes: Cut an observation sample segment from the composite material, mechanically polish the interface of the heterogeneous material layer, immerse the observation sample segment in a metallographic etchant for a first preset time, rinse the surface with clean water and then dry it, and then use a cotton ball to stick a nitric acid alcohol solution with a volume concentration of to wipe the surface of the sample, wash and dry it after etching for a second preset time, and finally obtain an observation sample.

7. A method for predicting the yield load and damage degree of an explosively welded composite material according to claim 6, characterized in that Performing an in-situ tensile test based on the in-situ tensile test platform to obtain the number and size distribution of holes and the number and size distribution of cracks in the specimen under different tensile loads, including: Set the initial tensile load and tensile loading rate; Place the in-situ tensile specimen into the mechanical fixing device and apply the initial tensile load. Every time the tensile load increases by 1 N, the in-situ tensile test platform rotates the fixing device to photograph the specimen slices of the in-situ tensile specimen at different angles, input the specimen slices into the first image recognition model to obtain the number and size of holes and cracks corresponding to the specimen slices, and count the minimum and maximum values of holes and cracks in the specimen under the current tensile load.

8. A prediction method for the yield load and damage degree of an explosion-welded composite material according to claim 7, characterized in that The predetermined requirements include: the wavelength should be within 660 - 690 and the amplitude should be within 60 - 90 ; The implementation of optimizing the input parameters using the genetic algorithm includes: B1: Initialize the population, set the population size N and the preset number of iterations T. Each individual in the population includes: collision speed, collision angle, explosive thickness, mass ratio of explosive per unit area to flyer plate mass, distance between flyer plate and substrate, and flyer plate thickness; Perform binary encoding on the collision speed, collision angle, distance between flyer plate and substrate, and flyer plate thickness, and perform Gray code encoding on the explosive thickness and mass ratio of explosive per unit area to flyer plate mass; B2: Calculate the fitness value of each individual according to the fitness function, where the fitness function is: ， wherein, is the minimum value of the wavelength in the predetermined requirement, is the maximum value of the wavelength in the predetermined requirement, is the wavelength obtained by the i-th individual inputting into the second prediction model, is the minimum value of the amplitude in the predetermined requirement, is the maximum value of the amplitude in the predetermined requirement, is the amplitude obtained by the i-th individual inputting into the second prediction model, is a constant, i = 1, 2, 3,.., N; B3: Selection operation, implement the selection operation based on the tournament algorithm; B4: Crossover operation, for the parameter group with binary encoding, use the single-point crossover method, and for the parameter group with Gray code encoding, use the multi-point crossover method; B5: Mutation operation. Select the individual with the highest fitness from the population for local search. For each selected individual, generate several neighborhood solutions in its vicinity, search for a solution better than the current solution among the neighborhood solutions. If a better solution is found, replace the current solution and put the individual after local search back into the population; B6: Generate a new population, merge the parent and offspring individuals, select a preset number of individuals as the new generation population according to the fitness value, and determine whether the iteration stop condition is reached. If so, use the optimal individual in the current population as the optimized operating parameter.

Images