Method and system for predicting rotary friction welding process parameters

Abstract

The invention belongs to the field of rotary friction welding, and discloses a method and a system for predicting rotary friction welding process parameters, and the method comprises the following steps: acquiring peak temperature and peak equivalent plastic strain of multiple groups of all grid points by utilizing a three-dimensional thermal coupling model about a first rotary friction welding head established by Abaqus software; processing is carried out through a BP neural network prediction model meeting the prediction precision, and the ultimate tensile strength corresponding to all grid points in multiple groups is obtained; screening the ultimate tensile strength to obtain the lowest ultimate tensile strength, and taking the lowest ultimate tensile strength as the ultimate tensile strength value of the welded joint under the group of parameters; and the maximum value of the ultimate tensile strength in the first rotary friction welding head is further screened out, and a set of rotary friction welding parameters corresponding to the maximum value serve as final rotary friction welding parameters. Aiming at the problem of welding size diversification, an optimal welding process parameter combination is quickly screened according to mechanical properties, and a guiding effect is provided for actual production of metal material transmission shaft rotary friction welding, so that the welding production efficiency is greatly improved.

Description

Technical Field

[0001] The invention belongs to the field of rotary friction welding, and in particular relates to a method and a system for predicting rotary friction welding process parameters. Background Art

[0002] As a large structural component, the automobile drive shaft requires a connection process for assembly. Welding is the main connection process for metal drive shafts. Fusion welding is a traditional welding method for drive shafts, and its joints have a high sensitivity to thermal cracking. In addition to thermal cracks, welding defects such as pores, lack of fusion, and incomplete penetration also often occur. Therefore, friction welding, as a hot-pressed solid-phase welding method, can be effectively applied to the welding of drive shafts. Under the action of external force, it uses the heat generated by the mutual friction between the surfaces of the weldments to make the contact surface of the weldments reach a thermoplastic state, and then applies a forging force to quickly forge to complete the solid-phase welding.

[0003] Process parameters are the main influencing parameters for friction welding to obtain excellent joints, but a large number of process experimental studies are often required to determine the optimal process parameter combination for a specific size, and the research cycle is cumbersome and lengthy. In addition, establishing a finite element model to simulate the changes in temperature field and plastic flow field during welding can help analyze the deformation behavior of the joint, but it has not been closely integrated with process experiments to provide more effective help for improving the production efficiency of rotary friction welding of metal materials. The drive shaft varies depending on the type of car, resulting in a large number of repeated process tests and numerical simulations under different models to conduct in-depth research and analysis on the welding of drive shafts of specific sizes, which greatly increases the experimental research cycle, and thus seriously reduces the actual production efficiency of the industrialization of metal drive shaft welding. Summary of the invention

[0004] In view of the above-mentioned problems, a method and system for predicting rotary friction welding process parameters are provided. The present invention aims at the problem of diversified welding sizes, quickly screens the optimal welding process parameter combination according to mechanical properties, and provides guidance for the actual production of rotary friction welding of metal material transmission shafts, in order to greatly improve welding production efficiency.

[0005] To achieve the above object, the present invention adopts the following technical solution: A method for predicting rotary friction welding process parameters includes the following steps: A three-dimensional thermomechanical coupling model of the first rotary friction welding joint is established using Abaqus software to obtain the peak temperature and peak equivalent plastic strain of all grid points of the first rotary friction welding joint under multiple sets of rotary friction welding parameters, wherein a set of rotary friction welding parameters corresponds to a set of peak temperature and peak equivalent plastic strain of grid points, and the first rotary friction welding joint is a rotary friction welding joint with process parameters to be predicted; After processing the peak temperature and peak equivalent plastic strain of the grid points of the first rotary friction welding joint through the BP neural network prediction model with satisfactory accuracy, the grid points are grouped according to different rotary friction welding parameters to obtain the ultimate tensile strength corresponding to all grid points of the first rotary friction welding joint under all groups of rotary friction welding parameters, wherein one group of rotary friction welding parameters corresponds to the ultimate tensile strength corresponding to all grid points of a group of first rotary friction welding joints; Screening the ultimate tensile strengths corresponding to all grid points of the first rotary friction welding joint under each set of rotary friction welding parameters, obtaining the lowest ultimate tensile strength among the ultimate tensile strengths corresponding to all grid points in the first rotary friction welding joint, and using it as the ultimate tensile strength of the first rotary friction welding joint under the set of parameters; Under all groups of rotational friction welding parameters, the maximum value of the ultimate tensile strength in the first rotational friction welding joint is screened out, and a group of rotational friction welding parameters corresponding to the maximum value is used as the final rotational friction welding parameters of the first rotational friction welding joint.

[0006] Preferably, the establishment and training process of the BP neural network prediction model includes: A three-dimensional thermal-mechanical coupling model of the second rotary friction welding joint is established based on Abaqus software, and the three-dimensional thermal-mechanical coupling model of the second rotary friction welding joint is corrected by using the welding temperature measured at the joint when the second rotary friction welding joint is welded under preset rotary friction welding parameters, so as to obtain a final three-dimensional thermal-mechanical coupling model of the second rotary friction welding joint, wherein the second rotary friction welding joint is the rotary friction welding joint used to establish the three-dimensional thermal-mechanical coupling model; Using the three-dimensional thermal-mechanical coupling model of the second rotary friction welding joint finally established by the Abaqus software, the peak temperature and the peak equivalent plastic strain of all grid points of the second rotary friction welding joint under the preset rotary friction welding parameters are obtained; A BP neural network prediction model is established with the peak temperature and peak equivalent plastic strain of all grid points of the second rotary friction welding joint as input and the ultimate tensile strength of all grid points of the second rotary friction welding joint as output; the ultimate tensile strength of all grid points of the second rotary friction welding joint is calculated by the microhardness values ​​of all grid points of the second rotary friction welding joint, and the microhardness values ​​of all grid points of the second rotary friction welding joint are experimentally measured at all grid points of the second rotary friction welding joint; The BP neural network prediction model is trained using the peak temperature and peak equivalent plastic strain of all grid points of the second rotary friction welding joint and the ultimate tensile strength of all grid points of the second rotary friction welding joint to obtain a BP neural network prediction model that meets the prediction accuracy.

[0007] Preferably, the second rotary friction welding head and the first rotary friction welding head are rotating bodies made of the same material but with different dimensions; Alternatively, the second rotary friction welding head and the first rotary friction welding head are rotating bodies made of the same material and with the same size specifications. Preferably, the process of establishing the three-dimensional thermal-mechanical coupling model of the first rotary friction welding joint established by the Abaqus software includes: Select metal materials of preset material and specification, weld the metal materials by using rotary friction welding process, and measure the temperature at different positions during the welding process; Combined with the constitutive equation of the selected metal material, based on Abaqus software, the coupled CEL method and the strain-compensated Arrhenius material model are used to establish a three-dimensional thermal-mechanical coupling model. During the process of establishing the three-dimensional thermal-mechanical coupling model, meshing is performed on each area along the axial cross section with the welding joint surface as the center. The three-dimensional thermal-mechanical coupling model is used to simulate the temperature field and plastic flow field of each area of ​​the joint during the welding process. Combined with the measured temperatures at different positions during the welding process, the established three-dimensional thermal-mechanical coupling model is corrected to obtain the three-dimensional thermal-mechanical coupling model established by the Abaqus software.

[0008] Preferably, the first rotary friction welding joint is a round bar or a round tube.

[0009] Preferably, the parameters of the rotary friction welding include rotation speed, friction pressure and upset pressure.

[0010] Preferably, when the peak temperature and the peak equivalent plastic strain of the grid points of the first rotary friction welding joint are processed by the BP neural network prediction model that has satisfied the prediction accuracy: The peak temperature and peak equivalent plastic strain of all grid points of the first rotary friction welding joint are first normalized, and then the normalized results are input into the BP neural network prediction model that has met the prediction accuracy for processing. After that, the BP neural network prediction model that has met the prediction accuracy outputs the ultimate tensile strength corresponding to all grid points.

[0011] Preferably, the BP neural network prediction model adopts a multi-layer feedforward neural network, and the BP neural network prediction model includes an input layer, a hidden layer and an output layer. Neurons in each hidden layer and output layer are given specific thresholds and activation functions, and weight connections are used between layers; The output of the hidden layer is calculated using the following formula: ))

[0012] in: is the regularization correction, is the residual connection correction, is the weight matrix, X is the input vector; is the bias term; is the tanh activation function, defined as ; When training the BP neural network prediction model, the adjusted determination coefficient Adjusted R 2 Determine the prediction accuracy of the BP neural network prediction model and adjust the determination coefficient Adjusted R 2 The calculation formula is as follows:

[0013]

[0014] in: is the true value of the i-th sample, is the average of the true values, is the predicted value of the i-th sample, is the number of samples, is the number of independent variables.

[0015] The present invention also provides a system for predicting rotary friction welding process parameters, comprising: An input data acquisition unit is used to obtain the peak temperature and peak equivalent plastic strain of all grid points of the first rotary friction welding joint under a plurality of preset sets of rotary friction welding parameters using a three-dimensional thermomechanical coupling model of the first rotary friction welding joint established by Abaqus software, wherein a set of rotary friction welding parameters corresponds to a set of peak temperature and peak equivalent plastic strain of grid points, and the first rotary friction welding joint is a rotary friction welding joint of which process parameters are to be predicted; A calculation unit: used to process the peak temperature and peak equivalent plastic strain of the grid points of the first rotary friction welding joint through a BP neural network prediction model that has satisfied the prediction accuracy, and then group them according to different rotary friction welding parameters to obtain the ultimate tensile strength corresponding to all grid points of the first rotary friction welding joint under all groups of rotary friction welding parameters, wherein one group of rotary friction welding parameters corresponds to a group of ultimate tensile strength corresponding to all grid points of the first rotary friction welding joint; The first screening unit is used to screen the ultimate tensile strengths corresponding to all grid points of the first rotary friction welding joint under each set of rotary friction welding parameters, obtain the lowest ultimate tensile strength among the ultimate tensile strengths corresponding to all grid points in the first rotary friction welding joint, and use it as the ultimate tensile strength of the first rotary friction welding joint under the set of parameters; The second screening unit is used to screen out the maximum value of the ultimate tensile strength in the first rotary friction welding joint under all groups of rotary friction welding parameters, and use a group of rotary friction welding parameters corresponding to the maximum value as the final rotary friction welding parameters of the first rotary friction welding joint.

[0016] Preferably, the system for predicting rotary friction welding process parameters of the present invention further includes: A first modeling unit: used to establish a three-dimensional thermal-mechanical coupling model; the three-dimensional thermal-mechanical coupling model includes a three-dimensional thermal-mechanical coupling model of a first rotary friction welding joint established by Abaqus software and a three-dimensional thermal-mechanical coupling model of a second rotary friction welding joint established by Abaqus software, wherein the second rotary friction welding joint is a rotary friction welding joint used to establish the three-dimensional thermal-mechanical coupling model, and the second rotary friction welding joint and the first rotary friction welding joint are rotating bodies with the same material and different dimensions, or the second rotary friction welding joint and the first rotary friction welding joint are rotating bodies with the same material and the same dimensions; The second modeling unit: used to establish a BP neural network prediction model.

[0017] The present invention has the following beneficial effects: In the method for predicting the process parameters of rotary friction welding of the present invention, the three-dimensional thermomechanical coupling model of the first rotary friction welding joint established by the Abaqus software can obtain the peak temperature and peak equivalent plastic strain of all grid points under multiple groups of rotary friction welding parameters, wherein one group of rotary friction welding parameters corresponds to the peak temperature and peak equivalent plastic strain of one group of grid points, and the first rotary friction welding joint is a rotary friction welding joint whose process parameters are to be predicted; then, through the BP neural network prediction model that has satisfied the prediction accuracy, the ultimate tensile strength corresponding to all grid points of the first rotary friction welding joint under each group of rotary friction welding parameters can be obtained, wherein one group of rotary friction welding parameters corresponds to the ultimate tensile strength corresponding to all grid points of a group of the first rotary friction welding joint; then, screening is performed to obtain the lowest ultimate tensile strength among the ultimate tensile strengths corresponding to all grid points in the first rotary friction welding joint, and then screening is performed again, and a group of rotary friction welding parameters corresponding to the maximum value of the lowest ultimate tensile strength is used as the final rotary friction welding parameters of the first rotary friction welding joint, thereby realizing the prediction of the rotary friction welding process parameters of the rotary friction welding joint of the preset material and preset size specifications. By using the three-dimensional thermomechanical coupling model of the first rotary friction welding joint established based on the Abaqus software of the present invention, multiple simulations can be performed to obtain multiple inputs of the BP neural network prediction model that have met the prediction accuracy, so that for the rotary friction welding joints of the same material and the same size, only one actual rotary friction welding experiment is required, which greatly reduces the number and cost of actual experiments; at the same time, for the rotary friction welding joints of the same material and different sizes, the same BP neural network prediction model that has met the prediction accuracy can be used to predict the ultimate tensile strength corresponding to all grid points of the rotary friction welding joint, so when it is necessary to predict the rotary friction welding joints of the same material and different sizes, it is only necessary to re-establish the three-dimensional thermomechanical coupling model of the rotary friction welding joints of different sizes based on the Abaqus software, and there is no need to train the BP neural network prediction model, so the training cost is greatly reduced. In summary, the present invention aims at the problem of diversified welding sizes, quickly screens the optimal welding process parameter combination according to mechanical properties, and provides guidance for the actual production of rotary friction welding of metal material transmission shafts, so as to greatly improve the welding production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0019] Figure 1A three-dimensional thermal-mechanical coupling model diagram established in an embodiment of the present invention; Figure 2 for Figure 1 The axial temperature cross-section diagram of the three-dimensional thermal-mechanical coupling model is shown; Figure 3 for Figure 1 The axial equivalent plastic strain cross-section diagram of the three-dimensional thermal-mechanical coupling model shown; Figure 4 This is a graph showing the tensile strength test results of the joint of the sample test set in the embodiment of the present invention. Figure 5 A flow chart for predicting rotary friction welding process parameters in an embodiment of the present invention; Figure 6 This is a flow chart of the BP neural network calculation used in the embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to explain the present invention more intuitively and demonstrate the advantages of the present invention, the following embodiments are used to specifically illustrate the implementation mode and related implementation details of the present invention.

[0021] In the technical solution of the present invention, a rotary friction welding experiment is carried out on a metal material transmission shaft of a given size, and the temperature change of each point in different axial regions of the joint during the welding process is measured. Combined with the temperature change curve, a precise corresponding three-dimensional thermal-mechanical coupling model of rotary friction welding is established based on the Abaqus software CEL method to simulate the change distribution of the temperature field and plastic flow field of the joint during the welding process. The peak temperature T and peak equivalent plastic strain ε of each point in different regions of the joint are extracted from the three-dimensional thermal-mechanical coupling simulation. p The hardness distribution of each point in different areas of the experimental welded joint was measured and converted into the ultimate tensile strength. The peak temperature T and peak equivalent plastic strain ε of each point were used to calculate the ultimate tensile strength. p As the input parameter, the tensile strength of the joint at the corresponding point in each area of ​​the joint is used as the output parameter. Through sufficient data for learning and training, a BP neural network model is established. When using the model, the peak temperature T and peak equivalent plastic strain ε at different points of the joint are used. p As input value, predict the change of ultimate tensile strength distribution at each point.

[0022] The present invention is particularly suitable for predicting the mechanical properties of rotary friction welding joints of the same material in any size. Specifically, another material of the same size is selected, a corresponding three-dimensional thermomechanical coupling model is established according to the above method, corresponding characteristic data are proposed, a sample database is obtained, and the established BP neural network model is used to predict the ultimate tensile strength of each point under different conditions, and the minimum value is determined as the ultimate tensile strength of the joint, so as to achieve rapid screening of the optimal process parameter combination for high-performance joints.

[0023] For details, see Figure 5 and Figure 6The method for predicting the process parameters of rotary friction welding of the present invention comprises the following steps: Step 1: Select a metal material rotating body of fixed size, adopt a rotary friction welding process to weld, and measure the temperature of different positions of the rotary friction welding joint during the welding process.

[0024] Step 2. Combined with the material constitutive equation, based on Abaqus software, the coupled CEL (coupled Euler-Lagrange) method and the strain-compensated Arrhenius material model are used to establish a three-dimensional thermal-mechanical coupling model of the rotary friction welding joint. This three-dimensional thermal-mechanical coupling model can simulate the temperature field and plastic flow field of each region of the joint during the rotary friction welding process. Combined with the welding temperature measured in the experiment in step 1, the established three-dimensional thermal-mechanical coupling model of the rotary friction welding joint is corrected to ensure the accuracy of the numerical simulation of the three-dimensional thermal-mechanical coupling model.

[0025] Step 3: Extract the peak temperature T and peak equivalent plastic strain ε of each grid point of the rotational friction welding joint along the axial direction from the three-dimensional thermal-mechanical coupling model of the rotational friction welding joint. p .

[0026] Step 4: Based on the position of each grid point in step 3, the microhardness value of each grid point in the corresponding area of ​​the rotational friction welding joint is measured in the experiment of step 1, and converted into the ultimate tensile strength of the grid points in each area. Step 5: The peak temperature (T) and the peak equivalent plastic strain (ε p ) as the input value, and the ultimate tensile strength of each grid point obtained in the experiment in step 1 as the output value, and a BP neural network prediction model is established.

[0027] Step six, train and learn the BP neural network prediction model established in step five, divide the data obtained in step three and step four into training data set, verification data set and test data set, repeat the training and verification process until the BP neural network prediction model reaches high prediction accuracy.

[0028] Step seven, select another metal material rotating body (referred to as the second metal material rotating body) with the same material and different size specifications as the metal material rotating body in step one, repeat steps one and two, and establish a three-dimensional thermal-mechanical coupling model of the rotational friction welding joint corresponding to the second metal material rotating body. The three-dimensional thermal-mechanical coupling model is referred to as the second three-dimensional thermal-mechanical coupling model. Step 8, changing the welding process parameters of the second three-dimensional thermal-mechanical coupling model, repeating step 3, and then using the BP neural network prediction model established in step 5 to predict the ultimate tensile strength of each grid point in the second three-dimensional thermal-mechanical coupling model under different rotary friction welding parameters; Step nine, filter the data obtained in step eight, select the lowest value of the ultimate tensile strength in each grid point under each group of parameters as the ultimate tensile strength value of the welded joint under this group of parameters, and further select the maximum value of the ultimate tensile strength under different rotational friction welding parameters, and use the rotational friction welding parameters corresponding to the maximum value as the final rotational friction welding parameters of the second metal material rotating body selected in step seven, thereby realizing rapid screening of the optimal process parameter combination for high-performance joints.

[0029] The above technical solution of the present invention is applicable to rotating bodies such as round bars and round tubes. The parameters of the rotary friction welding include the rotation speed, friction pressure and upsetting pressure. The experimental temperature measurement method is to drill holes at specific positions of the welding material and embed thermocouples to complete the temperature collection during the welding process.

[0030] The specific calculation formula of the constitutive equation is as follows:

[0031]

[0032] in: is the flow stress; is a specific coefficient; is the thermal deformation activation energy of the material; is the material constant; is the gas constant; is the deformation temperature of the material; Z is the Zener-Hollomon parameter.

[0033] When the established three-dimensional thermal-mechanical coupling model of the rotary friction welding joint is modified, the three-dimensional thermal-mechanical coupling model can accurately simulate the changes in the temperature field and the plastic flow field during the friction welding process when the correlation coefficient R>0.99 after temperature comparison and parameter adjustment.

[0034] In the process of establishing the three-dimensional thermal-mechanical coupling model of the rotary friction welding joint, the meshing is performed by taking the welding joint surface as the center and dividing the meshes in each area along the axial cross section.

[0035] The peak temperature (T) and the peak equivalent plastic strain (ε p ) is the highest temperature and the highest equivalent plastic strain reached at this point from the beginning to the end of welding. It can represent the plastic deformation state of this point under the action of thermal-mechanical coupling during the welding process, and directly determines the microstructure and mechanical properties of this point.

[0036] In the above scheme of the present invention, when measuring the microhardness value of each grid point in the corresponding area of ​​the rotary friction welding joint, the actual rotary friction welding joint is firstly gridded according to step three, and a microhardness tester is used for measurement, including Vickers hardness, Rockwell hardness and Brinell hardness.

[0037] The specific calculation method for converting the microhardness value of each grid point in the corresponding area of ​​the rotary friction welding joint into the ultimate tensile strength of the grid points in each area is as follows:

[0038] in: is the ultimate tensile strength; is the Vickers hardness value, is the correction factor, To compensate for the error.

[0039] The BP neural network prediction model in the present invention adopts a multi-layer feedforward neural network, which includes an input layer, a hidden layer and an output layer. The neurons in each hidden layer and the output layer are given specific thresholds and activation functions, and weights are used to connect the layers. The BP neural network prediction model uses Z-score to normalize the input value. In the BP neural network prediction model, the output of the hidden layer is calculated using the following formula: ))

[0040] in: is regularization correction; Correction for residual connection; is the weight matrix, which determines how the input affects the output of the hidden layer; X is the input vector, which contains the input data passed to the hidden layer; It is a bias term that helps the model fit the data better; is the tanh activation function. It is defined as .

[0041] The BP neural network prediction model of the present invention uses mean square error (MSE) as the error function of the model, and its calculation method is as follows:

[0042] in: is the number of training samples; is the number of outputs; is the predicted value of the jth output of the i-th sample; is the true value of the jth output of the ith sample.

[0043] The BP neural network prediction model of the present invention adopts a gradient descent optimization algorithm to update network weights.

[0044] When training and optimizing the BP neural network prediction model of the present invention, the proportions of the training data set, the validation data set, and the test data set are 70%, 15%, and 15%, respectively. 2 ) determines the prediction accuracy of the prediction model, and its calculation formula is as follows: Adjusted R 2 When the value is >0.95, the prediction model is considered accurate and can achieve the prediction target:

[0045]

[0046] in: is the true value of the i-th sample; is the average of the true values; is the predicted value of the i-th sample; is the number of samples; is the number of independent variables.

[0047] Example: This embodiment takes Al-5.2Zn-1.8Mg-0.8Cu-0.2Si-0.15Er-0.15Zr high-strength aluminum alloy as an example, combines the rotary friction welding process experiment and finite element simulation to provide a sample database for the BP neural network prediction model, and realizes the prediction of the rotary friction welding process parameters. The method for predicting the rotary friction welding process parameters of the rotating body of the material specifically includes the following steps: Step (1), select Al-5.2Zn-1.8Mg-0.8Cu-0.2Si-0.15Er-0.15Zr high-strength aluminum alloy rod as welding material, with a size of φ22mm×45mm (diameter×length), grind its end face and clean it with anhydrous ethanol to remove the surface oxide layer. Use a continuous drive friction welding machine with equipment model ZCC-10TxDJ. The welding parameters are a speed of 1100rpm, a friction pressure of 15kN, and a top forging pressure of 15kN. Punch a hole along the axial direction of the welding rod to embed a thermocouple, and the punching positions are 2mm, 4mm, 6mm, 8mm, and 12mm away from the welding joint surface, and measure the welding temperature at each point.

[0048] Step (2), based on the constitutive equation of Al-5.2Zn-1.8Mg-0.8Cu-0.2Si-0.15Er-0.15Zr high-strength aluminum alloy, the three-dimensional thermal-mechanical coupling model of the rotary friction welding joint is established using Abaqus software, as shown in Figure 1 As shown in the figure, the flash morphology of the model is good and there is no obvious welding defect. The three-dimensional thermal-mechanical coupling model of the rotary friction welding joint is used to simulate the temperature field and plastic flow field of the joint during the welding process. Figure 2 and Figure 3As shown in Figure 2, the joint temperature and equivalent plastic strain are symmetrically distributed. The simulated temperature field is compared with the measured temperature in step (1). Figure 2 The temperature distribution patterns shown are similar, and the simulation error of the three-dimensional thermal-mechanical coupling model of the rotary friction welding joint is only 1.0%.

[0049] Step (3), after welding is completed, the joint surface is taken as the center, and the grids are divided along the axial direction with a row and column spacing of 0.5 mm. A total of 500 points are divided, and the peak temperature (T) and peak equivalent plastic strain (ε) of each point in the welding process are extracted from the three-dimensional thermal-mechanical coupling model. p ) as shown in Table 1: Table 1

[0050] Step (4), using a Vickers microhardness tester to characterize the microhardness of 500 points of the experimental joint corresponding to the three-dimensional thermal-mechanical coupling model, converted into ultimate tensile strength, the correction factor is =0.88, error compensation is X =8, the results of some tensile strength values ​​are shown in Table 2: Table 2

[0051] Step (5) is to compare the peak temperature (T) and the peak equivalent plastic strain (ε p ) as the input value and the ultimate tensile strength of the corresponding point as the output value, 350 groups of corresponding values ​​are selected as the training data set, 75 groups of corresponding values ​​are selected as the verification data set, and 75 groups of corresponding values ​​are selected as the test data set.

[0052] Step (6), establish a BP neural network prediction model, train it with the sample training set of step (5), verify it with the validation data set, test it with the sample test set, repeat the training and verification, and the test set results are as follows: Figure 4 As shown, the errors of each point are within 8%, and finally, the BP neural network prediction model reaches the required prediction accuracy requirement.

[0053] Step (7), select the same high-strength aluminum alloy material with the size of φ12mm×φ22mm×45mm (inner diameter×outer diameter×length), weld it, and the welding parameters are rotation speed 1300rpm, friction pressure 15kN, and upsetting pressure 15kN. Repeat step (1).

[0054] Step (8), based on its constitutive equation, a three-dimensional thermomechanical coupling model is established using Abaqus software, and the simulated temperature field of the three-dimensional thermomechanical coupling model is compared with the actual temperature measured in the experiment. The simulation error of the three-dimensional thermomechanical coupling model is only 0.9%.

[0055] Step (9), changing the welding rotation speed in the three-dimensional thermal-mechanical coupling model, the first set of rotational friction welding parameters are rotation speed 900 rpm, friction pressure 15 kN, and upsetting pressure 15 kN. The second set of rotational friction welding parameters are rotation speed 1700 rpm, friction pressure 15 kN, and upsetting pressure 15 kN. The third set of rotational friction welding parameters are rotation speed 2100 rpm, friction pressure 15 kN, and upsetting pressure 15 kN. The peak temperature (T) and peak equivalent plastic strain (ε) of each point in the welding process under each set of rotational friction welding parameters are extracted. P ).

[0056] Step (10), using the BP neural network prediction model trained in the above step (6), predict the ultimate tensile strength of each point under different rotation speed conditions of the size (φ12mm×φ22mm×45mm (inner diameter×outer diameter×length)), as shown in Table 3: Table 3

[0057] Combined with Table 3, the lowest value of the ultimate tensile strength is selected as the ultimate tensile strength of the joint under the rotation speed condition; under the three different process conditions (i.e., rotary friction welding parameters), the lowest ultimate tensile strengths are 378.13 MPa, 380.77 MPa and 372.32 MPa, respectively. Therefore, the second set of rotary friction welding parameters corresponding to the lowest ultimate tensile strength of 380.77 MPa is used as the optimal parameters for rotary friction welding of the high-strength aluminum alloy bar with the specifications (φ12 mm×φ22 mm×45 mm (inner diameter×outer diameter×length)) selected in step (7).

[0058] The present invention also provides a system for predicting rotary friction welding process parameters, which is used to implement the method for predicting rotary friction welding process parameters of each embodiment of the present invention, including: An input data acquisition unit is used to obtain the peak temperature and peak equivalent plastic strain of all grid points of the first rotary friction welding joint under multiple groups of rotary friction welding parameters using a three-dimensional thermal-mechanical coupling model of the first rotary friction welding joint established by Abaqus software, wherein a group of rotary friction welding parameters corresponds to a group of peak temperature and peak equivalent plastic strain of grid points, and the first rotary friction welding joint is a rotary friction welding joint of which process parameters are to be predicted; A calculation unit: used to process the peak temperature and peak equivalent plastic strain of the grid points of the first rotary friction welding joint through a BP neural network prediction model that has satisfied the prediction accuracy, and then group them according to different rotary friction welding parameters to obtain the ultimate tensile strength corresponding to all grid points of the first rotary friction welding joint under all groups of rotary friction welding parameters, wherein one group of rotary friction welding parameters corresponds to a group of ultimate tensile strength corresponding to all grid points of the first rotary friction welding joint; The first screening unit is used to screen the ultimate tensile strengths corresponding to all grid points of the first rotary friction welding joint under each set of rotary friction welding parameters, obtain the lowest ultimate tensile strength among the ultimate tensile strengths corresponding to all grid points in the first rotary friction welding joint, and use it as the ultimate tensile strength of the first rotary friction welding joint under the set of parameters; The second screening unit is used to screen out the maximum value of the ultimate tensile strength in the first rotary friction welding joint under all groups of rotary friction welding parameters, and use a group of rotary friction welding parameters corresponding to the maximum value as the final rotary friction welding parameters of the first rotary friction welding joint.

[0059] Furthermore, the system for predicting the rotary friction welding process parameters of the present invention also includes: A first modeling unit: used to establish a three-dimensional thermal-mechanical coupling model; the three-dimensional thermal-mechanical coupling model includes a three-dimensional thermal-mechanical coupling model of a first rotary friction welding joint established by Abaqus software and a three-dimensional thermal-mechanical coupling model of a second rotary friction welding joint established by Abaqus software, wherein the second rotary friction welding joint is a rotary friction welding joint used to establish the three-dimensional thermal-mechanical coupling model, and the second rotary friction welding joint and the first rotary friction welding joint are rotating bodies with the same material and different dimensions, or the second rotary friction welding joint and the first rotary friction welding joint are rotating bodies with the same material and the same dimensions; The second modeling unit: used to establish a BP neural network prediction model.

[0060] Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0061] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for predicting rotary friction welding process parameters, characterized in that: The process includes the following: A three-dimensional thermomechanical coupling model of the first rotary friction welding joint established by Abaqus software is used to obtain the peak temperature and peak equivalent plastic strain of all grid points of the first rotary friction welding joint under multiple groups of rotary friction welding parameters, wherein a group of rotary friction welding parameters corresponds to a group of peak temperature and peak equivalent plastic strain of grid points, and the first rotary friction welding joint is a rotary friction welding joint with process parameters to be predicted; After processing the peak temperature and peak equivalent plastic strain of the grid points of the first rotary friction welding joint by using a BP neural network prediction model that has met the prediction accuracy, the grid points are grouped according to different rotary friction welding parameters to obtain the ultimate tensile strength corresponding to all grid points of the first rotary friction welding joint under all groups of rotary friction welding parameters, wherein one group of rotary friction welding parameters corresponds to the ultimate tensile strength corresponding to all grid points of a group of first rotary friction welding joints; Screening the ultimate tensile strengths corresponding to all grid points of the first rotary friction welding joint under each set of rotary friction welding parameters, obtaining the lowest ultimate tensile strength among the ultimate tensile strengths corresponding to all grid points in the first rotary friction welding joint, and using it as the ultimate tensile strength of the first rotary friction welding joint under the set of parameters; Under all groups of rotational friction welding parameters, the maximum value of the ultimate tensile strength in the first rotational friction welding joint is screened out, and a group of rotational friction welding parameters corresponding to the maximum value is used as the final rotational friction welding parameters of the first rotational friction welding joint.

2. A method for predicting rotary friction welding process parameters according to claim 1, characterized in that: The establishment and training process of the BP neural network prediction model includes: A three-dimensional thermal-mechanical coupling model of the second rotary friction welding joint is established based on Abaqus software, and the three-dimensional thermal-mechanical coupling model of the second rotary friction welding joint is corrected by using the welding temperature measured at the joint when the second rotary friction welding joint is welded under preset rotary friction welding parameters, so as to obtain a final three-dimensional thermal-mechanical coupling model of the second rotary friction welding joint, wherein the second rotary friction welding joint is the rotary friction welding joint used to establish the three-dimensional thermal-mechanical coupling model; Using the three-dimensional thermal-mechanical coupling model of the second rotary friction welding joint finally established by the Abaqus software, the peak temperature and the peak equivalent plastic strain of all grid points of the second rotary friction welding joint under the preset rotary friction welding parameters are obtained; A BP neural network prediction model is established with the peak temperature and peak equivalent plastic strain of all grid points of the second rotary friction welding joint as input and the ultimate tensile strength of all grid points of the second rotary friction welding joint as output; the ultimate tensile strength of all grid points of the second rotary friction welding joint is calculated by the microhardness values ​​of all grid points of the second rotary friction welding joint, and the microhardness values ​​of all grid points of the second rotary friction welding joint are experimentally measured at all grid points of the second rotary friction welding joint; The BP neural network prediction model is trained using the peak temperature and peak equivalent plastic strain of all grid points of the second rotary friction welding joint and the ultimate tensile strength of all grid points of the second rotary friction welding joint to obtain a BP neural network prediction model that meets the prediction accuracy.

3. A method for predicting rotary friction welding process parameters according to claim 2, characterized in that: The second rotary friction welding head and the first rotary friction welding head are rotating bodies with the same material and different dimensions; Alternatively, the second rotary friction welding head and the first rotary friction welding head are rotating bodies made of the same material and with the same size specifications.

4. A method for predicting rotary friction welding process parameters according to any one of claims 1 to 3, characterized in that: The process of establishing the three-dimensional thermal-mechanical coupling model of the first rotary friction welding joint established by the Abaqus software includes: Select metal materials of preset material and specification, weld the metal materials by using rotary friction welding process, and measure the temperature at different positions during the welding process; Combined with the constitutive equation of the selected metal material, based on Abaqus software, the coupled CEL method and the strain-compensated Arrhenius material model are used to establish a three-dimensional thermal-mechanical coupling model. During the process of establishing the three-dimensional thermal-mechanical coupling model, meshing is performed on each area along the axial cross section with the welding joint surface as the center. The three-dimensional thermal-mechanical coupling model is used to simulate the temperature field and plastic flow field of each area of ​​the joint during the welding process. The established three-dimensional thermal-mechanical coupling model is corrected in combination with the measured temperatures at different positions during the welding process, thereby obtaining the three-dimensional thermal-mechanical coupling model established by the Abaqus software.

5. The method for predicting rotary friction welding process parameters according to claim 1, characterized in that: The first rotational friction welding joint is a round bar or a round tube.

6. The method for predicting rotary friction welding process parameters according to claim 1, characterized in that: The parameters of rotary friction welding include rotation speed, friction pressure and upsetting pressure.

7. The method for predicting rotary friction welding process parameters according to claim 1, characterized in that: When the peak temperature and peak equivalent plastic strain of the grid points of the first rotary friction welding joint are processed by the BP neural network prediction model that has met the prediction accuracy: The peak temperature and peak equivalent plastic strain of all grid points of the first rotary friction welding joint are first normalized, and then the normalized results are input into the BP neural network prediction model that has met the prediction accuracy for processing. After that, the BP neural network prediction model that has met the prediction accuracy outputs the ultimate tensile strength corresponding to all grid points.

8. The method for predicting rotary friction welding process parameters according to claim 1, characterized in that: The BP neural network prediction model adopts a multi-layer feedforward neural network. The BP neural network prediction model includes an input layer, a hidden layer and an output layer. Neurons in each hidden layer and output layer are given specific thresholds and activation functions, and weight connections are used between layers. The output of the hidden layer is calculated using the following formula: )) in: is the regularization correction, is the residual connection correction, is the weight matrix, X is the input vector; is the bias term; is the tanh activation function, defined as ; When training the BP neural network prediction model, the adjusted determination coefficient Adjusted R 2 Determine the prediction accuracy of the BP neural network prediction model and adjust the determination coefficient Adjusted R 2 The calculation formula is as follows: in: is the true value of the i-th sample, is the average of the true values, is the predicted value of the i-th sample, is the number of samples, is the number of independent variables.

9. A system for predicting rotary friction welding process parameters, characterized in that: include: An input data acquisition unit is used to obtain the peak temperature and peak equivalent plastic strain of all grid points of the first rotary friction welding joint under multiple groups of rotary friction welding parameters using a three-dimensional thermal-mechanical coupling model of the first rotary friction welding joint established by Abaqus software, wherein a group of rotary friction welding parameters corresponds to a group of peak temperature and peak equivalent plastic strain of grid points, and the first rotary friction welding joint is a rotary friction welding joint of which process parameters are to be predicted; A calculation unit: used to process the peak temperature and peak equivalent plastic strain of the grid points of the first rotary friction welding joint through a BP neural network prediction model that has satisfied the prediction accuracy, and then group them according to different rotary friction welding parameters to obtain the ultimate tensile strength corresponding to all grid points of the first rotary friction welding joint under all groups of rotary friction welding parameters, wherein one group of rotary friction welding parameters corresponds to a group of ultimate tensile strength corresponding to all grid points of the first rotary friction welding joint; The first screening unit is used to screen the ultimate tensile strengths corresponding to all grid points of the first rotary friction welding joint under each set of rotary friction welding parameters, obtain the lowest ultimate tensile strength among the ultimate tensile strengths corresponding to all grid points in the first rotary friction welding joint, and use it as the ultimate tensile strength of the first rotary friction welding joint under the set of parameters; The second screening unit is used to screen out the maximum value of the ultimate tensile strength in the first rotary friction welding joint under all groups of rotary friction welding parameters, and use a group of rotary friction welding parameters corresponding to the maximum value as the final rotary friction welding parameters of the first rotary friction welding joint.

10. The system for predicting rotary friction welding process parameters according to claim 9, characterized in that: Also includes: A first modeling unit: used to establish a three-dimensional thermal-mechanical coupling model; the three-dimensional thermal-mechanical coupling model includes a three-dimensional thermal-mechanical coupling model of a first rotary friction welding joint established by Abaqus software and a three-dimensional thermal-mechanical coupling model of a second rotary friction welding joint established by Abaqus software, wherein the second rotary friction welding joint is a rotary friction welding joint used to establish the three-dimensional thermal-mechanical coupling model, and the second rotary friction welding joint and the first rotary friction welding joint are rotating bodies with the same material and different dimensions, or the second rotary friction welding joint and the first rotary friction welding joint are rotating bodies with the same material and the same dimensions; The second modeling unit: used to establish a BP neural network prediction model.

Images