A method, system and electronic device for predicting the mechanical behavior of a hot melt self-piercing joint

Abstract

The application discloses a hot-melt self-tapping joint mechanical behavior prediction method and system and electronic equipment, and relates to the technical field of mechanical behavior prediction.The application adopts a method combining numerical simulation and test to determine the material parameters of the plate and the screw.A numerical model of the hot-melt self-tapping joint is established according to the joint cross-section morphology obtained through a forming test.The reliability of the numerical model is verified based on the test results of the joint mechanical behavior.The verified numerical model is used to simulate the mechanical behavior of the joint samples of different plate thickness combinations under different loading conditions, so as to extract the training and test data sets of an artificial neural network prediction model.Then, the plate thickness and loading angle of the joint not participating in the training and test are input into the trained mechanical behavior prediction model to obtain the mechanical behavior prediction results of the joint, thereby solving the problems of high test cost, long simulation calculation time and large space storage in the existing hot-melt self-tapping joint mechanical behavior prediction process.

Description

Technical Field

[0001] This invention relates to the field of mechanical behavior prediction technology, and in particular to a method, system and electronic device for predicting the mechanical behavior of a thermoplastic self-tapping screw joint. Background Technology

[0002] Flow Drill Screws (FDS) are a type of single-sided riveting method suitable for connecting multi-layered dissimilar plates. These fasteners possess self-piercing and compression effects, capable of drilling through 2-4 layers of thin plates to achieve a connection. Due to their advantages in connecting dissimilar materials, they are increasingly widely used by major OEMs. Joints are the weakest areas in assembly structures. Understanding the mechanical properties and failure behavior of joints under different loading conditions is crucial for structural integrity, energy dissipation, and design decisions. Experimental testing plays an important role in studying the deformation modes and failure mechanisms of FDS joints, but this method is costly and time-consuming. The rapid development of computer technology has made it possible to simulate joint deformation and failure behavior under actual load conditions using finite element analysis (FEM). However, steel-aluminum hybrid body structures have joints with various plate thicknesses and material grades, and these joints exhibit different failure modes under different loading conditions. Studying the mechanical behavior of these joints one by one through FEM simulation requires even more computation time and enormous storage space. Summary of the Invention

[0003] To address the aforementioned problems in the existing technology, this invention provides a method, system, and electronic device for predicting the mechanical behavior of thermofusion self-tapping screw joints.

[0004] To achieve the above objectives, the present invention provides the following solution:

[0005] A method for predicting the mechanical behavior of a thermoplastic self-tapping screw joint includes:

[0006] The material parameters of the upper plate, lower plate and screws were determined by combining data simulation and experimentation, and a numerical model of the thermoplastic self-tapping joint sample was constructed.

[0007] The reliability of the numerical model is verified based on the experimental results of the joint's mechanical behavior.

[0008] Based on the verified numerical model, the mechanical behavior of hot-melt self-tapping screw joint samples with different plate thickness combinations under different loading conditions was simulated to construct training and test sample datasets.

[0009] A neural network model is trained using training and testing sample datasets to obtain a trained neural network model, and the trained neural network model is used as a mechanical behavior prediction model.

[0010] The plate thickness and loading angle of the thermofused self-tapping screw joint that were not used in the training and testing were input into the mechanical behavior prediction model to obtain the mechanical behavior prediction results of the thermofused self-tapping screw joint.

[0011] Optionally, a numerical model of the thermofusion self-tapping screw joint specimen is constructed, specifically including:

[0012] Based on the cross-sectional morphology of the thermofused self-tapping screw joint specimens with different plate thickness combinations, the cross-sectional profile of the thermofused self-tapping screw joint was drawn to scale, and then a numerical model of the thermofused self-tapping screw joint specimen was established by rotation and stretching.

[0013] Optionally, the reliability of the numerical model is verified based on the experimental results of the joint's mechanical behavior, specifically including:

[0014] Determine the microhardness distribution of the cross-section of the thermoplastic self-tapping screw joint sample;

[0015] The extent of each sub-region of the heat-affected zone is determined based on the microhardness distribution.

[0016] Tensile tests were conducted on equivalent tensile specimens of each sub-region of the heat-affected zone to obtain the failure strain and instability strain of the four sub-regions under tensile stress.

[0017] The failure strain / instability strain-stress triaxiality curve of the parent material is scaled based on the failure strain and the instability strain.

[0018] When the numerical simulation of the joint matches the load-displacement curves and failure modes of the test, the reliability of the numerical model is verified.

[0019] Optionally, the neural network includes a classification artificial neural network model and a regression artificial neural network model.

[0020] Optionally, the failure modes include: lower plate thread peeling failure mode, lower plate tearing failure mode, lower plate thread peeling and lower plate tearing combined failure mode, and screw shear fracture failure mode.

[0021] Optionally, for the prediction of the load-displacement curve, the load data with a displacement interval of a set value on the curve is extracted as a dataset for training and testing of thermofusion self-tapping screw joint samples.

[0022] Optionally, the setting value is 0.2 mm.

[0023] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0024] The mechanical behavior prediction method for thermofusion self-tapping screw joints provided by this invention uses a combination of numerical simulation and experimentation to determine the material parameters of the upper plate, lower plate, and screws. Based on the cross-sectional morphology of the thermofusion self-tapping screw joint, a corresponding numerical model of the joint is established. The reliability of the numerical model is verified based on the experimental results of the joint's mechanical behavior. Based on the verified numerical model, the mechanical behavior of thermofusion self-tapping screw joint samples with different plate thickness combinations under different loading conditions is simulated to construct a sample database, and sample data is extracted to obtain a dataset. A neural network model is trained using the obtained dataset to obtain a mechanical behavior prediction model. The plate thickness and loading angle of the thermofusion self-tapping screw joint, which were not involved in the training and testing, are input into the mechanical behavior prediction model to accurately obtain the predicted mechanical behavior results of the thermofusion self-tapping screw joint. This solves the problems of high experimental costs, long simulation calculation time, and large storage space in the existing thermofusion self-tapping screw joint mechanical behavior prediction process.

[0025] Furthermore, the present invention also provides a system for predicting the mechanical behavior of a thermofusion self-tapping screw joint, the system being used to apply the above-mentioned method for predicting the mechanical behavior of a thermofusion self-tapping screw joint; the system includes:

[0026] The numerical model building module is used to determine the material parameters of the upper plate, lower plate and screws based on a combination of numerical simulation and experiment, and to build a numerical model of the thermoplastic self-tapping joint sample.

[0027] The numerical model verification module is used to verify the reliability of the numerical model based on the test results of the joint mechanical behavior.

[0028] The sample database construction module is used to simulate and construct a training and testing dataset for hot-melt self-tapping screw joint samples with different plate thickness combinations under different loading conditions based on the verified numerical model.

[0029] The prediction model building module is used to train a neural network model using training and testing datasets to obtain a trained neural network model, and to use the trained neural network model as a mechanical behavior prediction model.

[0030] The prediction model applicability module is used to input the plate thickness and loading angle of the thermofused self-tapping screw joint that were not involved in the training and testing into the mechanical behavior prediction model to obtain the mechanical behavior prediction results of the thermofused self-tapping screw joint.

[0031] An electronic device, comprising:

[0032] Memory, used to store computer programs;

[0033] A processor, connected to the memory, is used to retrieve and execute the computer program to implement the above-described method for predicting the mechanical behavior of thermoplastic self-tapping screw joints.

[0034] Since the technical effects achieved by the two structures provided above in this invention are the same as those achieved by the mechanical behavior prediction method for thermofusion self-tapping screw joints provided in this invention, they will not be described again here. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0036] Figure 1 A flowchart illustrating the implementation of the method for predicting the mechanical behavior of a thermoplastic self-tapping screw joint provided by this invention;

[0037] Figure 2 A roadmap for the implementation of the mechanical behavior prediction method for thermofusion self-tapping screw joints provided by the present invention;

[0038] Figure 3 A diagram showing the correspondence between the specimen geometry and stress triaxiality provided by this invention;

[0039] Figure 4 The mechanical property parameter diagram of 6082-T6 aluminum alloy and joint heat-affected zone provided by the present invention; wherein, Figure 4 (a) is the actual stress-plastic strain curve. Figure 4 (b) is the fracture strain / instability strain-stress triaxiality curve. Figure 4 (c) is the normalization factor-cell size relationship diagram. Figure 4 (d) is the engineering stress-strain curve diagram of the four sub-regions of the heat-affected zone; Figure 4 (e) is a scaling factor diagram of fracture strain and instability strain in the four sub-regions of the heat-affected zone;

[0040] Figure 5 A schematic diagram of the geometric dimensions of the FDS connector provided by the present invention;

[0041] Figure 6 A physical image of the FDS connector provided by this invention;

[0042] Figure 7 A schematic diagram of the cross-sectional morphology of the FDS connector provided by the present invention;

[0043] Figure 8A schematic diagram of a three-dimensional finite element model of the cross-shaped FDS joint provided by the present invention; wherein, Figure 8 (a) is a schematic diagram of the internal region of the cross-shaped FDS joint in the three-dimensional finite element model. Figure 8 (b) is a schematic diagram of the finite element model of the cross-shaped FDS joint. Figure 8 (c) is a schematic diagram of the external region of the cross-shaped FDS joint in the three-dimensional finite element model;

[0044] Figure 9 A schematic diagram comparing the simulation and experimental load-displacement curves of the four failure modes provided by this invention;

[0045] Figure 10 A schematic diagram comparing the simulation and experimental failure characteristics of the four failure modes provided by this invention;

[0046] Figure 11 A schematic diagram of the structure of the artificial neural network model provided by this invention;

[0047] Figure 12 A schematic diagram illustrating the correlation between simulated and predicted peak load and failure displacement of the joint provided by this invention; wherein, Figure 12 (a) is a schematic diagram showing the correlation between the simulated and predicted peak loads on the training dataset. Figure 12 (b) is a schematic diagram showing the correlation between the simulated and predicted peak loads on the test dataset. Figure 12 (c) is a schematic diagram illustrating the correlation between simulated and predicted failure displacements on the training dataset. Figure 12 (d) is a schematic diagram illustrating the correlation between the simulated and predicted failure displacements in the test dataset;

[0048] Figure 13 The present invention provides a schematic diagram comparing the load correlation between simulation and prediction, as well as the load-displacement curve; wherein, Figure 13 (a) in the figure is a schematic diagram of the load correlation between the simulation and prediction of the training dataset. Figure 13 (b) in the diagram illustrates the correlation between the simulated and predicted loads on the test dataset. Figure 13 (c) in the figure is a schematic diagram comparing the simulated and predicted load-displacement curves when the failure displacement is at its maximum and minimum.

[0049] Figure 14 A schematic diagram illustrating the correlation between simulated and predicted loads on the joint load-displacement curve; where, Figure 14 (a) is a schematic diagram illustrating the correlation between simulated and predicted loads on the training dataset. Figure 14 (b) is a schematic diagram of the load correlation between the simulation and prediction of the test dataset;

[0050] Figure 15A schematic diagram comparing the simulated and predicted joint load-displacement curves of the test dataset under the four failure modes provided by this invention; wherein, Figure 15 (a) is a schematic diagram comparing the simulated and predicted joint load-displacement curves of the test dataset under the TSB failure mode. Figure 15 (b) is a schematic diagram comparing the simulated and predicted joint load-displacement curves of the test dataset under the TOB failure mode. Figure 15 (c) is a schematic diagram comparing the simulated and predicted joint load-displacement curves of the test dataset under SSF failure mode. Figure 15 (d) is a schematic diagram comparing the joint load-displacement curves of the simulated and predicted test datasets under the TSB+TOB failure mode.

[0051] Figure 16 A schematic diagram comparing the simulation and prediction results of the joint under the TSB failure mode provided by this invention; wherein, Figure 16 (a) is a schematic diagram illustrating the correlation between simulated and predicted loads in the test dataset under the TSB failure mode. Figure 16 (b) is a schematic diagram comparing the load-displacement curves of the test dataset simulation and prediction under the TSB failure mode;

[0052] Figure 17 A schematic diagram comparing the simulation and prediction results of the joint under the TOB failure mode provided by this invention; wherein, Figure 17 (a) is a schematic diagram illustrating the load correlation between the simulated and predicted loads in the test dataset under the TOB failure mode. Figure 17 (b) is a schematic diagram comparing the load-displacement curves of the test dataset simulation and prediction under the TOB failure mode;

[0053] Figure 18 A schematic diagram comparing the simulation and prediction results of the joint under the TSB+TOB failure mode provided by this invention; wherein, Figure 18 (a) is a schematic diagram illustrating the load correlation between simulation and prediction in the test dataset under the TSB+TOB failure mode. Figure 18 (b) is a schematic diagram comparing the load-displacement curves of the test dataset simulation and prediction under the TSB+TOB failure mode.

[0054] Figure 19 A schematic diagram comparing the simulation and prediction results of the joint under the SSF failure mode provided by this invention; wherein, Figure 19 (a) is a schematic diagram illustrating the correlation between simulated and predicted loads in the test dataset under the SSF failure mode. Figure 19 (b) is a schematic diagram comparing the load-displacement curves of the test dataset under the SSF failure mode;

[0055] Figure 20This is a schematic diagram comparing the simulated and predicted joint load-displacement curves of the dataset provided by this invention, which was not used in training and testing. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0057] The purpose of this invention is to provide a rapid and accurate method for predicting the mechanical behavior of hot-melt self-tapping screw joints with different plate thicknesses under different loading conditions, which solves the problems of high test costs, long simulation calculation time and large space storage in the existing process of obtaining the mechanical behavior of hot-melt self-tapping screw joints.

[0058] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0059] like Figure 1 As shown, the method for predicting the mechanical behavior of a thermoplastic self-tapping screw joint provided by the present invention includes:

[0060] Step 100: The material parameters of the upper plate, lower plate and screws were determined based on a combination of numerical simulation and experiment, and a numerical model of the thermoplastic self-tapping screw joint sample was constructed.

[0061] This step primarily involves building a three-dimensional numerical model (i.e., a numerical model) based on the cross-sectional morphology of the joint. The specific building process is as follows:

[0062] Based on the cross-sectional morphology of the thermofused self-tapping screw joint specimens with different plate thickness combinations, the cross-sectional profile of the thermofused self-tapping screw joint was drawn to scale, and then a numerical model of the thermofused self-tapping screw joint specimen was established by rotation and stretching.

[0063] The material softening effect of the heat-affected zone (HAZ) was also considered in the constructed three-dimensional numerical model of the joint. The extent of each sub-region of the HAZ was determined by the microhardness distribution of the joint cross-section. Due to the limited extent of the HAZ, an equivalent substitution test of friction stir welding was used to determine the constitutive relationship and failure parameters of each sub-region of the HAZ.

[0064] Failure modes of joints with different plate thicknesses and loading angles include thread stripping failure mode (TSB), tearing failure mode (TOB), combined thread stripping and tearing failure mode (TSB+TOB), and screw shear fracture failure mode (SSF).

[0065] Step 101: Verify the reliability of the numerical model based on the test results of the joint's mechanical behavior. This step can be implemented as follows:

[0066] Determine the microhardness distribution across the cross-section of the thermofused self-tapping screw joint sample.

[0067] The extent of each sub-region of the heat-affected zone was determined based on the microhardness distribution.

[0068] Tensile tests were conducted on equivalent tensile specimens of each sub-region of the heat-affected zone to obtain the relationship between failure strain and fracture strain of the four sub-regions under tensile stress.

[0069] The failure strain / fracture strain-stress triaxiality curve of the parent material is scaled based on the failure strain relationship and fracture strain.

[0070] When the numerical simulation of the joint matches the load-displacement curves and failure modes of the test, the reliability of the numerical model is verified.

[0071] Step 102: Based on the verified numerical model, simulate the mechanical behavior of hot-melt self-tapping screw joint samples with different plate thickness combinations under different loading conditions to construct a sample database.

[0072] Step 103: Extract sample data from the sample database to obtain training and testing datasets. For example, for predicting the load-displacement curve, extract load data with a set displacement interval on the curve as the training and testing dataset for the thermoplastic self-tapping screw joint specimen. The set value used is 0.2 mm.

[0073] Step 104: Train the neural network model using the obtained dataset to obtain the trained neural network model, and use the trained neural network model as a mechanical behavior prediction model.

[0074] Before training the neural network model, it is necessary to build the model first. In this invention, the neural network models built are a classification artificial neural network (ANN) model and a regression artificial neural network (ANN) model. The classification ANN model is used to classify the failure modes of joints with different plate thicknesses under different loading conditions. The regression ANN model is used to predict the peak load, failure displacement, load-displacement curve, etc. of the joint.

[0075] Furthermore, when using a neural network model to predict the load-displacement curve of a joint, various data processing methods can be employed to improve the prediction accuracy of the load-displacement curve.

[0076] Step 105: Input the plate thickness and loading angle of the thermofused self-tapping screw joint that were not involved in the training and testing into the mechanical behavior prediction model to obtain the mechanical behavior prediction results of the thermofused self-tapping screw joint.

[0077] Furthermore, to improve the accuracy of mechanical behavior prediction, before proceeding to step 100, it is necessary to calibrate the material parameters of the upper plate, lower plate, and screws in the thermoplastic self-tapping screw joint sample. The specific calibration process is as follows:

[0078] 1) For sheet metal, process five different types of specimens. For screws, process only two types of specimens.

[0079] 2) Mechanical property tests on sheet metal and screws, wherein the displacement and strain of sheet metal were obtained using the DIC digital image correlation method.

[0080] 3) Construct simulation models of different types of plates and screws.

[0081] 4) Calibrate the hardening curves and failure parameters of the sheet metal and screws. During the calibration of the failure parameters, the failure models used for the sheet metal and screws were GISSMO and shear-corrected GTN models, respectively.

[0082] Furthermore, in order to perform mechanical predictions for joints with different plate thicknesses and loading conditions, before step 102, it is necessary to prepare FDS joint specimens through forming tests and obtain the mechanical behavior of the joints through mechanical property tests. The specific test process is as follows:

[0083] 1) Processing test pieces for FDS joint forming test.

[0084] 2) Forming tests were conducted on joints with different plate thicknesses. In this test, the specimen structure was a cross-shaped FDS joint.

[0085] 3) Obtain the cross-sectional morphology of joints with different plate thicknesses.

[0086] 4) Randomly select several joints with different plate thicknesses and loading conditions for mechanical performance testing. The test results are used to verify the reliability of the three-dimensional numerical model of the joint.

[0087] Compared with existing technologies, the mechanical behavior prediction method for thermofusion self-tapping screw joints provided by this invention can quickly and accurately predict the failure modes, peak loads, failure displacements, and load-displacement curves of joints with different plate thicknesses under different loading conditions, providing a reference for vehicle body structure design decisions.

[0088] Furthermore, taking the 1.8mm thick DP590 high-strength steel and 3.0mm thick 6082-T6 aluminum alloy combined FDS connector as the main research object, according to... Figure 2 The implementation route shown illustrates the specific implementation process of the above-mentioned method for predicting the mechanical behavior of thermofusion self-tapping screw joints provided by the present invention. The specific implementation process includes:

[0089] Step 1: Calibrate the material parameters of DP590 high-strength steel, 6082-T6 aluminum alloy and hardened steel screws.

[0090] Step 1.1: For DP590 high-strength steel and 6082-T6 aluminum alloy sheets, the required specimens are uniaxial tensile, notched R5, notched R20, shear 0°, and shear 45°. The specimen dimensions are as follows: Figure 3 As shown. For hardened steel screws, due to screw size limitations, only uniaxial tensile and round bar shear specimens are machined. Figure 3 The dimensions of the intermediate sample are shown for illustrative purposes only and are not intended to limit the specific scope of the invention.

[0091] Step 1.2: Mechanical property tests were conducted on different types of specimens, and the load-displacement curves of the specimens were obtained.

[0092] Step 1.3: Based on the experimental boundary conditions, construct a three-dimensional finite element model for all specimens. The average size of the mesh in the finite element model is 0.3 mm.

[0093] Step 1.4: Obtain the constitutive relation and failure parameters of the material using the inverse parametric method for simulation of the joint's three-dimensional numerical model. The constitutive relation and failure parameters of 6082-T6 aluminum alloy are as follows: Figure 4 As shown.

[0094] Step 2: Forming test and mechanical property test of the cross FDS joint.

[0095] Step 2.1: The sheet metal used for hot-melt self-tapping screw connection is machined into a size of 100mm × 40mm along the rolling or extrusion direction. Several sheet metal thicknesses commonly used in vehicle body structures are selected as research objects. For DP590 high-strength steel, the sheet metal thicknesses selected are 1.2mm, 1.5mm, 1.8mm, 2.0mm, and 2.2mm; for 6082-T6 aluminum alloy, the sheet metal thicknesses are 2.0mm, 2.5mm, 3.0mm, 3.5mm, and 4.0mm, for a total of 25 combinations.

[0096] Step 2.2: For ease of study, the process parameters are the same for all plate thickness combinations, namely, drilling speed of 7412 / min, feed speed of 9.6mm / s, and tightening torque of 11.5N·m.

[0097] Step 2.3: Cut the connected joint along the screw axis to obtain the cross-sectional morphology of the joint with different plate thicknesses. The geometric dimensions, physical object, and cross-sectional morphology of the joint are as follows: Figures 5-7As shown.

[0098] Step 2.4: In addition to the plate thickness, this example also studies the mechanical behavior of these joints under different loading conditions, with loading angles of 0°, 15°, 30°, 45°, 60°, 75° and 90°.

[0099] Step 2.5: Randomly select several joints with different plate thicknesses under different loading angles to conduct mechanical property tests. The test results are used to verify the reliability of the joint numerical simulation results.

[0100] Step 3: Simulation of joint mechanical behavior.

[0101] Step 3.1: Based on the cross-sectional morphology of the joints with different plate thicknesses, depict the outline of the joint cross-section to scale, and then establish the three-dimensional geometric model of the joint through rotation and extrusion. When meshing the three-dimensional geometric model of the joint, the mesh size near the thread is 0.05mm, and the mesh size away from the thread is 0.3mm. The three-dimensional numerical model of the cross FDS joint is as follows: Figure 8 As shown.

[0102] Step 3.2: During the joint forming process, due to the high-speed friction between the screw and the sheet metal, the material near the thread undergoes significant thermal softening. Based on this, this example determines the range of each sub-region of the heat-affected zone (HAZ) according to the microhardness distribution of the joint cross-section. The dimensions of each sub-region of the HAZ are small, making it difficult to obtain the mechanical properties of each sub-region through conventional tensile testing. Considering the similarity between the forming principles of friction stir welding and hot-melt self-tapping screw processes, friction stir welding was used to prepare equivalent tensile specimens suitable for mechanical property testing. The stress-strain relationship curves of the four sub-regions of the HAZ are shown below. Figure 4 As shown in (d), the failure strain / instability strain-stress triaxiality curve of the base material is scaled based on the failure strain and fracture strain relationship of the equivalent specimens in the four sub-regions of the heat-affected zone under tensile stress. The scaling factors of the failure strain and fracture strain of the four sub-regions of the heat-affected zone are as follows. Figure 4 As shown in (e).

[0103] Step 3.3, Verification of the joint numerical model. Under the four failure modes, the load-displacement curves and failure characteristics of the joint simulation and experiment are compared as follows: Figure 9 and Figure 10 As shown, the simulation and experimental results are in good agreement, which verifies the reliability of the joint numerical model.

[0104] Step 3.4: Based on the validated joint model, numerical simulations were performed on the mechanical behavior of joints with different plate thicknesses under different loading conditions, totaling 175 combinations. The results of the numerical simulations were used to generate a training and testing database for the artificial neural network.

[0105] Step 4: Joint mechanical behavior prediction based on artificial neural networks

[0106] Step 4.1: Based on the above description, this invention establishes two types of ANN models: a classification ANN model and a regression ANN model. Both ANN models consist of one input layer, five hidden layers, and one output layer. The variables in the input layer are the upper plate thickness, the lower plate thickness, and the loading angle. The loading angle is reduced by a factor of 30 so that its value is on the same order of magnitude as the plate thickness. The output layer variables of the classification ANN model are the failure modes, while the output layer variables of the regression ANN model are the peak load, the failure displacement, and the load-displacement curve. For the prediction of the load-displacement curve, load data with a displacement interval of 0.2 mm on the load-displacement curve are extracted from the sample database as the dataset. 90% of the sample data is used to train the ANN model, and the remaining 10% is used to test the ANN model. The structure of the ANN model is as follows: Figure 11 As shown.

[0107] Step 4.2: Within the plate thickness and loading angle range of the embodiment, the joint exhibits failure modes including thread stripping (TSB), torn plate (TOB), a combination of thread stripping and torn plate (TSB+TOB), and screw shear fracture (SSF). Numerical simulation results show that the number of cases for the four failure modes—TSB, TOB, TSB+TOB, and SSF—are 65, 15, 20, and 75, respectively. As shown in Table 1, the classification ANN model can accurately classify all cases for the four failure modes.

[0108] Table 1. Sample data of failure modes classified using the classification ANN model.

[0109]

[0110] Step 4.3: The performance of the regression ANN model is assessed using relative error (RE) and coefficient of determination (R²). 2 The correlation between the peak load and failure displacement predicted by the regression ANN model and the simulation results is evaluated using root mean square (RMSE) and the results. Figure 12 As shown in Table 2, the relatively high R... 2 The relatively low RE and RMSE values ​​indicate that the regression ANN model established in this embodiment can accurately predict the peak load and failure displacement of joints with different plate thicknesses under different loading conditions.

[0111] Table 2 Performance Indicators of Joint Peak Load and Failure Displacement

[0112]

[0113] Step 4.4: In this embodiment, a regression ANN model was also used to predict the load-displacement curve of the joint. Three data processing methods were employed to improve the prediction accuracy of the ANN model for the load-displacement curve prediction. The first method involved retaining the original data and adding a 0kN load to the load-displacement curves of joints with failure displacements less than 25.84mm, ensuring all curves had the same length. The second method involved using the average failure displacement of 175 samples (15.0mm) as the curve length and removing the load data after the failure displacement from the dataset. Based on the relationship between failure displacement and curve length, the displacement of the original joint load-displacement curve was amplified and reduced. After training and testing, an inverse operation was performed to obtain the true displacement of the curve. The third method involved training and testing the joint load-displacement curves for TSB, TOB, SSF, and TSB+TOB failure modes based on the second method. The data processed using the above three methods was used to predict the load-displacement curves, and the predicted load-displacement curve results are shown below. Figures 13-19 As shown in Table 3, when the third data processing method is used, the load-displacement curves predicted by the ANN model have good consistency with the numerical simulation results. Table 4 shows the combinations of plate thickness and loading angle for the joints corresponding to the number of test dataset instances. In Table 4, numbers 1-18 correspond to... Figures 15-19 The labels in the text.

[0114] Table 3 Comparison of ANN model prediction results using three data processing methods

[0115]

[0116] Table 4 shows the combination of plate thickness and loading angle for the number of test dataset instances.

[0117]

[0118]

[0119] Step 4.5: To verify the feasibility of the trained ANN model on datasets not included in the training dataset, four joints with different combinations of plate thickness and loading angle were randomly selected. These four joints were not included in the 175-sample dataset. At this point, the joint failure mode, peak load, failure displacement, and load-displacement curve shown in the ANN prediction results were identical to those of the numerical model, as follows: Figure 20 As shown, this further verifies the reliability of the ANN model established in this embodiment.

[0120] Based on the above description, this invention first calibrates the constitutive relations and failure parameters of the plate and screw materials using a parametric inverse method. On this basis, joint forming experiments were conducted on different combinations of plate thicknesses to obtain the cross-sectional morphology of the joints. Four joints with randomly selected plate thickness combinations were subjected to mechanical property tests to verify the reliability of the three-dimensional numerical model simulation results. A three-dimensional numerical model was established based on the cross-sectional morphology of the joints with different plate combinations. The reliability of the model was verified by comparing simulation and experimental results. Using the verified three-dimensional numerical model of the joints, a sample database of joints with different plate thickness combinations under different loading conditions was generated. Based on the sample database, classification and regression ANN models were established to classify failure modes and predict peak load, failure displacement, and load-displacement curves, respectively. Three data processing methods were proposed to improve the prediction accuracy of the ANN model when predicting the load-displacement curve of the joint. Finally, the trained ANN model was successfully applied to a dataset that was not used for training.

[0121] Furthermore, this invention also provides a system for predicting the mechanical behavior of thermofusion self-tapping screw joints. The system is used to apply the aforementioned method for predicting the mechanical behavior of thermofusion self-tapping screw joints. The system includes: a numerical model construction module, a numerical model verification module, a sample database construction module, a prediction model construction module, and a prediction model applicability module.

[0122] The numerical model construction module is used to determine the material parameters of the upper plate, lower plate and screws based on a combination of data simulation and experimentation, and to construct a numerical model of the thermoplastic self-tapping screw joint sample.

[0123] The numerical model verification module is used to verify the reliability of the numerical model based on the test results of the joint mechanical behavior.

[0124] The sample database construction module is used to simulate and construct training and test sample datasets based on the validated numerical model, simulating the mechanical behavior of hot-melt self-tapping screw joint samples with different plate thickness combinations under different loading conditions.

[0125] The prediction model building module is used to train a neural network model using training and testing sample datasets, obtain a trained neural network model, and use the trained neural network model as a mechanical behavior prediction model.

[0126] The prediction model applicability module is used to input the plate thickness and loading angle of the thermofused self-tapping screw joint, which were not involved in the training and testing, into the mechanical behavior prediction model to obtain the mechanical behavior prediction results of the thermofused self-tapping screw joint.

[0127] An electronic device, comprising:

[0128] Memory is used to store computer programs.

[0129] A processor, connected to a memory, is used to retrieve and execute a computer program to implement the above-described method for predicting the mechanical behavior of thermoplastic self-tapping screw joints.

[0130] Furthermore, when the computer program in the aforementioned memory is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks.

[0131] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0132] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for predicting the mechanical behavior of a thermoplastic self-tapping screw joint, characterized in that, include: Based on numerical simulation and experiments, the material parameters of the upper plate, lower plate and screws are determined, and a numerical model of the hot melt self-tapping joint sample is constructed. The reliability of the numerical model is verified based on the experimental results of the joint's mechanical behavior. Based on the verified numerical model, the mechanical behavior of hot-melt self-tapping screw joint samples with different plate thickness combinations under different loading conditions was simulated to construct training and test sample datasets. A neural network model is trained using training and testing sample datasets to obtain a trained neural network model, and the trained neural network model is used as a mechanical behavior prediction model. The plate thickness and loading angle of the thermofused self-tapping screw joint that were not used in the training and testing were input into the mechanical behavior prediction model to obtain the mechanical behavior prediction results of the thermofused self-tapping screw joint. The numerical model for constructing the thermofusion self-tapping screw joint sample includes: Based on the cross-sectional morphology of hot-melt self-tapping screw joint samples with different plate thicknesses, the cross-sectional profile of the hot-melt self-tapping screw joint was drawn on a proportional scale, and then a numerical model of the hot-melt self-tapping screw joint sample was established by rotation and stretching. The reliability of the numerical model is verified based on the experimental results of the joint's mechanical behavior, including: Determine the microhardness distribution of the cross-section of the thermoplastic self-tapping screw joint sample; The extent of each sub-region of the heat-affected zone is determined based on the microhardness distribution. Tensile tests were conducted on equivalent tensile specimens of each sub-region of the heat-affected zone to obtain the failure strain and instability strain of the four sub-regions under tensile stress. The failure strain / instability strain-stress triaxiality curve of the parent material is scaled based on the failure strain and the instability strain. When the numerical simulation of the joint matches the load-displacement curves and failure modes of the test, the reliability of the numerical model is verified.

2. The method for predicting the mechanical behavior of a thermoplastic self-tapping screw joint according to claim 1, characterized in that, The neural network includes a classification artificial neural network model and a regression artificial neural network model.

3. The method for predicting the mechanical behavior of a thermoplastic self-tapping screw joint according to claim 1, characterized in that, The failure modes include: lower plate thread peeling failure mode, lower plate tearing failure mode, lower plate thread peeling and lower plate tearing combined failure mode, and screw shear fracture failure mode.

4. The method for predicting the mechanical behavior of a thermoplastic self-tapping screw joint according to claim 1, characterized in that, For the prediction of load-displacement curves, load data with displacement intervals set on the curves are extracted as datasets for training and testing of thermoplastic self-tapping screw joint specimens.

5. The method for predicting the mechanical behavior of a thermoplastic self-tapping screw joint according to claim 4, characterized in that, The set value is 0.2mm.

6. A system for predicting the mechanical behavior of a thermoplastic self-tapping screw joint, characterized in that, The system is used to apply the mechanical behavior prediction method for thermofusion self-tapping screw joints as described in any one of claims 1-5; the system comprises: The numerical model building module is used to determine the material parameters of the upper plate, lower plate and screws based on a combination of data simulation and experimentation, and to build a numerical model of the thermoplastic self-tapping joint specimen. The numerical model verification module is used to verify the reliability of the numerical model based on the test results of the joint mechanical behavior. The sample database construction module is used to simulate and construct training and test sample datasets based on the verified numerical model to simulate the mechanical behavior of hot-melt self-tapping screw joint samples with different plate thickness combinations under different loading conditions. The prediction model building module is used to train a neural network model using training and test sample datasets, obtain a trained neural network model, and use the trained neural network model as a mechanical behavior prediction model. The prediction model applicability module is used to input the plate thickness and loading angle of the thermofused self-tapping screw joint that were not involved in the training and testing into the mechanical behavior prediction model to obtain the mechanical behavior prediction results of the thermofused self-tapping screw joint.

7. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, connected to the memory, is configured to retrieve and execute the computer program to implement the method for predicting the mechanical behavior of a thermoplastic self-tapping screw joint as described in any one of claims 1-5.

Images