Establishment of a prediction model for friction stir welding structure defects using finite element method
The friction stir welding structure defect prediction model is established through the finite element method, which solves the problems of collapse and crack defects during welding, realizes the improvement of welding quality and structural optimization, and reduces development costs and waste emissions.
Patent Information
- Application Number
- CN202110076926.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-01-20
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2041-01-20
AI Technical Summary
During friction stir welding, defects such as collapse and cracking are easily generated in the cavity area, and the prior art is difficult to effectively predict and avoid these defects, which affects the welding quality and structural rigidity.
The finite element method is used to establish a defect prediction model for friction stir welding structure, and the welding process is simulated through 3D digital model and finite element model, data parameters are derived, and collapse defect prediction is carried out in combination with regression equations and Delphi method, and the welding process is optimized to avoid defects.
Effectively predict and reduce welding defects, reduce product weight, reduce development costs and waste emissions, and improve welding quality and structural rigidity.
Smart Images

Figure CN112733406B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of welding, and in particular relates to establishing a friction stir welding structure defect prediction model by utilizing a finite element method. Background Art
[0002] The cruising range of new energy vehicles is one of the important goals of various automobile manufacturers at present, and the unsprung mass of the vehicle is the primary goal. Through the mature development of friction stir welding technology, the process application has gradually shifted from the aviation manufacturing industry to the livelihood industry. Friction stir welding technology was first used in the chassis structure of the automobile, turning the original one-piece casting into a two-piece or even multi-piece design; one-piece castings often need to be supplemented for casting shrinkage due to their large structure, resulting in a slow cooling rate and the inability to improve the physical properties of the material. The combination of two-piece (multi-piece) and friction stir welding technology can reduce the volume of individual castings and reduce casting defects. Then, the advantages of multi-piece bonding are used to create hollow areas, which can not only reduce weight but also reduce noise.
[0003] However, various defects such as collapse and cracking often occur in the cavity area during the welding process. The finite element method is used to conduct preliminary evaluation in the design stage, analyze the rigidity and strength of the structure itself, and predict the defects in the friction stir welding process. For this reason, we propose to use the finite element method to establish a friction stir welding structure defect prediction model. Summary of the invention
[0004] The purpose of the present invention is to provide a friction stir welding structure defect prediction model established by using the finite element method to solve the problems raised in the above background technology.
[0005] To achieve the above object, the present invention provides the following technical solution: using the finite element method to establish a friction stir welding structure defect prediction model, including the following steps:
[0006] S1. Draw 3D digital model and establish finite element model: Use UG software to construct the 3D digital model of the friction stir welding structure, establish the finite element model through the 3D digital model, and import various physical performance data;
[0007] S2. Exporting various data parameters of friction stir welding through finite element model: by simulating the contact between the welding head and the area to be welded on the model, applying a constant force to the welding head, observing the change of welding structure, and outputting various data parameters of friction welding;
[0008] S3, inputting the output parameters into the collapse defect prediction model for data analysis: inputting the data parameters in S2 into the collapse defect prediction model, analyzing the data information, and then obtaining a prediction of the collapse and defects that may occur during the friction welding process;
[0009] S4. Friction welding the structure: avoid the collapse and defect prediction results in S3, and then perform friction welding on the structural welding points.
[0010] Preferably, the steps of building the UG software in S1 are as follows:
[0011] S101: First, import the CAD drawing into the UG software and freeze it as a base map;
[0012] S102: Import other CAD plane drawings in turn, set the capture point to a vertical point through 2D capture, and stack the imported layer outlines together, then move the upper layer along the Y axis as a reference, and then import other structural images and freeze your outlines;
[0013] S103: Generate an entity by importing multiple layers of images and then connecting the images to form a 3D model image.
[0014] Preferably, the finite element model in S1 is constructed as follows:
[0015] S111: Discretization of a structure is to divide a structure or a continuum into many units. Therefore, when starting the analysis, the structure must be modeled with appropriate finite elements and the number, type, size and arrangement of the units must be determined;
[0016] S112: Take one of the elements from the region or structure for study, and select an appropriate interpolation mode or displacement mode to approximately describe the displacement field of the element. Since the displacement solution of a complex structure cannot be accurately known in advance under any given load, the difference mode is usually taken as a polynomial form. From the point of view of calculation, the polynomial is simple and meets certain convergence requirements. After the element displacement function is approximated by a polynomial, the problem is transformed into how to calculate the node displacement. After the node displacement is determined, the displacement field is also determined;
[0017] S113: Derivation of element stiffness matrix and load vector. Based on the assumed displacement pattern, the element stiffness matrix and load vector can be derived using equilibrium conditions or appropriate variational principles.
[0018] S114: The total set of equilibrium equations is obtained from the collective unit equations. A continuum or structure is composed of many finite element units. Therefore, when performing finite element analysis on the entire continuum or structure, it is necessary to combine them.
[0019] Preferably, the physical property data in S1 include the thickness of the welding part, the diameter of the welding part, the hardness of the welding part, the elasticity of the welding part and the heat transfer parameters of the welding part.
[0020] Preferably, the process of simulating friction welding in S2 is as follows:
[0021] S21: A constant pressure is applied to the friction welding part through the pressure structure, so that the welding structure can fit tightly during the friction process:
[0022] S22: The welding head is rotated and pressed on the pre-welded multi-piece wheel rim structure through a power mechanism. In the first stage, a constant displacement is pressed down to fully insert the welding needle. In the second stage, welding is performed with a constant pressure and a constant speed. After welding is completed, the welding needle is slowly pulled up to make the multi-piece wheel rim structure fit tightly.
[0023] S23: measuring simulation parameters. After welding is completed, the defects and collapse of the welding parts are detected through testing, and data parameters are output.
[0024] Preferably, when the pressure structure applies pressure for the first time, the first stage is defined by a displacement of 3mm-10mm, and the pressure applied by the pressure structure in the friction stir welding walking state is 1kN-5kN.
[0025] Preferably, the measurement simulation parameters include the number of collapses and defects in the welding part, and the diameters of the collapses and defects at various locations, and the thickness, elastic deformation, plastic deformation and size of the welding node of the welding part are detected.
[0026] Preferably, the collapse defect prediction model in S3 adopts a regression equation and a Delphi method to calculate and predict the result of the collapse defect prediction model.
[0027] Preferably, the Delphi method is a method of asking each member for their predicted value, and then asking for their predicted value again after mutual reference, so that the predicted values tend to be consistent after several rounds of consultation. First, the predicted value of each group of data is asked. After several rounds, the distribution of the predicted value and the prediction result are statistically consulted, and the representative median is used as the final predicted value; and the judgment is made based on the number of data before and after the final predicted value is predicted, and the data information is transmitted to the regression equation for fitting.
[0028] Preferably, the regression equation is fitted using the least squares method:
[0029] By using the least squares method to find its equation, we can calculate the straight line y=bx+a.
[0030] Generally, there is often more than one factor that affects y. Assuming there are x1, x2, ..., xk, k factors, the following linear relationship can usually be considered:
[0031] y=β0+β1x1+β2x2+…+βk x k +ε
[0032] Make n independent observations of y and x1, x2, ..., xk simultaneously to obtain n groups of observations (xt1, xt2, ..., xtk), t = 1, 2, ..., n (n> k+1), which satisfy the relationship:
[0033] y=β0+β1xt1+β2xt2+…+β k x tk +ε t
[0034] Among them, ε1...ε n They are all independent random variables with the same distribution as ε.
[0035] In order to express the above formula in matrix, let:
[0036] So we have Y = Xβ + ε,
[0037] Use the least squares method to find the solution for β
[0038]
[0039] in,
[0040] (X T X) -1 X T ,
[0041] is called the pseudo-inverse of X.
[0042] Compared with the prior art, the present invention has the following beneficial effects:
[0043] The present invention uses the finite element method for analysis, establishes a defect prediction model, and simultaneously establishes various parameter databases. Through preliminary experiments and analysis data verification, the subsequent design and development time is reduced, and the product development cost is reduced. The prediction model, parameter database and data verification are used to achieve structural optimization, reduce product weight, and thus achieve the purpose of reducing waste emissions. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] Figure 1 It is a schematic diagram of the step structure of the present invention. DETAILED DESCRIPTION
[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0046] See also Figure 1 The present invention provides a technical solution: using the finite element method to establish a friction stir welding structure defect prediction model, including the following steps:
[0047] S1. Draw 3D digital model and establish finite element model: Use UG software to construct the 3D digital model of the friction stir welding structure, establish the finite element model through the 3D digital model, and import various physical performance data;
[0048] S2. Exporting various data parameters of friction stir welding through finite element model: by simulating the contact between the welding head and the area to be welded on the model, applying a constant force to the welding head, observing the change of welding structure, and outputting various data parameters of friction welding;
[0049] S3, inputting the output parameters into the collapse defect prediction model for data analysis: inputting the data parameters in S2 into the collapse defect prediction model, analyzing the data information, and then obtaining a prediction of the collapse and defects that may occur during the friction welding process;
[0050] S4. Friction welding the structure: avoid the collapse and defect prediction results in S3, and then perform friction welding on the structural welding points.
[0051] In order to realize the establishment of 3D digital model and improve the speed of establishing 3D digital model, UG software is used for rapid construction. In this embodiment, preferably, the UG software construction steps in S1 are as follows:
[0052] S101: First, import the CAD drawing into the UG software and freeze it as a base map;
[0053] S102: Import other CAD plane drawings in turn, set the capture point to a vertical point through 2D capture, and stack the imported layer outlines together, then move the upper layer along the Y axis as a reference, and then import other structural images and freeze your outlines;
[0054] S103: Generate an entity by importing multiple layers of images and then connecting the images to form a 3D model image.
[0055] In order to establish the finite element model, in this embodiment, preferably, the finite element model in S1 is constructed as follows:
[0056] S111: Discretization of a structure is to divide a structure or a continuum into many units. Therefore, when starting the analysis, the structure must be modeled with appropriate finite elements and the number, type, size and arrangement of the units must be determined;
[0057] S112: Take one of the elements from the region or structure for study, and select an appropriate interpolation mode or displacement mode to approximately describe the displacement field of the element. Since the displacement solution of a complex structure cannot be accurately known in advance under any given load, the difference mode is usually taken as a polynomial form. From the point of view of calculation, the polynomial is simple and meets certain convergence requirements. After the element displacement function is approximated by a polynomial, the problem is transformed into how to calculate the node displacement. After the node displacement is determined, the displacement field is also determined;
[0058] S113: Derivation of element stiffness matrix and load vector. Based on the assumed displacement pattern, the element stiffness matrix and load vector can be derived using equilibrium conditions or appropriate variational principles.
[0059] S114: The total set of equilibrium equations is obtained from the collective unit equations. A continuum or structure is composed of many finite element units. Therefore, when performing finite element analysis on the entire continuum or structure, it is necessary to combine them.
[0060] In order to input the parameters required for the prediction of the model and make the prediction of the model more accurate, in this embodiment, preferably, the physical property data in S1 include the thickness of the welding part, the diameter of the welding part, the hardness of the welding part, the elasticity of the welding part and the heat transfer parameters of the welding part.
[0061] In order to simulate the friction stir welding process of the welded structure and effectively simulate the generation of defects and collapse, in this embodiment, preferably, the process of simulating friction welding in S2 is as follows:
[0062] S21: A constant pressure is applied to the friction welding part through the pressure structure, so that the welding structure can fit tightly during the friction process:
[0063] S22: The welding head is rotated and pressed on the pre-welded multi-piece wheel rim structure through a power mechanism. In the first stage, a constant displacement is pressed down to fully insert the welding needle. In the second stage, welding is performed with a constant pressure and a constant speed. After welding is completed, the welding needle is slowly pulled up to make the multi-piece wheel rim structure fit tightly.
[0064] S23: measuring simulation parameters. After welding is completed, the defects and collapse of the welding parts are detected through testing, and data parameters are output.
[0065] In order to enable the welding structure to achieve a fitting connection and to enable the welding parts to be tightly connected during hot melting so that the welding parts can be adhered together, in this embodiment, preferably, when the pressure structure is first applied, the first stage is defined by a displacement of 6mm, and the pressure applied by the pressure structure in the friction stir welding walking state is 4kN.
[0066] In order to predict the welding conditions, various data information generated during the simulated welding process are statistically calculated. In this embodiment, preferably, the measured simulation parameters include the number of collapses and defects in the welding parts, and the diameters of the collapses and defects at various locations, and the thickness, elastic deformation, plastic deformation and size of the welding nodes of the welding parts are detected.
[0067] In order to better predict the occurrence of welding conditions, the Delphi method is used to distribute the welding parameters and predict the predicted values of each group of parameters, and then the predicted values are fitted to a linear curve through linear regression. In this embodiment, preferably, the collapse defect prediction model in S3 uses linear regression and Delphi method to combine and calculate the results of the collapse defect prediction model.
[0068] In order to achieve the distribution and arrangement of the data appearing in the welding situation, and predict the actual data situation of each data information that appears, and determine the distribution of the predicted values, in this embodiment, preferably, the Delphi method is a method of consulting each member for a predicted value, and consulting each member's predicted value again after mutual reference, and making the predicted values tend to be consistent after several rounds of consultation. First, the predicted value of each group of data is consulted. After several rounds, the distribution of the predicted values and the results of the statistical consultation of the prediction results are calculated, and the representative median is used as the final prediction value; and the judgment is made based on the number of data before and after the final prediction value is predicted, and the data information is transmitted to the regression equation for fitting.
[0069] In order to enable the predicted data to more clearly display the problematic curves that may occur in the actual welding process, in this embodiment, preferably, the regression equation is fitted using the least squares method:
[0070] By using the least squares method to find its equation, we can calculate the straight line y=bx+a.
[0071] Generally, there is often more than one factor that affects y. Assuming there are x1, x2, ..., xk, k factors, the following linear relationship can usually be considered:
[0072] y=β0+β1x1+β2x2x…+β k x k +ε
[0073] Make n independent observations of y and x1, x2, ..., xk simultaneously to obtain n groups of observations (xt1, xt2, ..., xtk), t = 1, 2, ..., n (n> k+1), which satisfy the relationship:
[0074] y=β0+β1x t1 +β2xt2+…+β k x tk +ε t
[0075] Among them, ε1...ε n They are all independent random variables with the same distribution as ε.
[0076] In order to express the above formula in matrix, let:
[0077] So we have Y = Xβ + ε,
[0078] Use the least squares method to find the solution for β
[0079]
[0080] in,
[0081] (X T X) -1 X T ,
[0082] is called the pseudo-inverse of X.
[0083] The working principle and use process of the present invention:
[0084] The first step is to draw a 3D digital model and establish a finite element model: Use UG software to build a 3D digital model of the friction stir welding structure, and use the 3D digital model to establish a finite element model, and import various physical performance data;
[0085] The second step is to derive various data parameters of friction stir welding through the finite element model: by simulating the contact between the welding head and the area to be welded on the model, and applying a constant force to the welding head, observing the changes in the welding structure, and outputting various data parameters of friction welding;
[0086] The third step is to input the output parameters into the collapse defect prediction model for data analysis: the data parameters in S2 are input into the collapse defect prediction model, and the data information is analyzed to obtain the prediction of the collapse and defects that may occur during the friction welding process;
[0087] The fourth step is to perform friction welding on the structure: avoid it by predicting the collapse amount and defects in S3, and then perform friction welding on the structural welding points.
[0088] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. The finite element method is used to establish a friction stir welding structural defect prediction model, which is characterized by: The following steps are included: S1. Draw 3D digital model and establish finite element model: Use UG software to construct the 3D digital model of the friction stir welding structure, establish the finite element model through the 3D digital model, and import various physical performance data; S2. Exporting various data parameters of friction stir welding through finite element model: by simulating the contact between the welding head and the area to be welded on the model, applying a constant force to the welding head, observing the change of welding structure, and outputting various data parameters of friction welding; S3, inputting the output parameters into the collapse defect prediction model for data analysis: inputting the data parameters in S2 into the collapse defect prediction model, analyzing the data information, and then obtaining a prediction of the collapse and defects that may occur during the friction welding process; S4, friction welding is then performed on the structure: the collapse amount and defect prediction results in S3 are used to avoid the problem, and then friction welding is performed on the structural welding part; The steps for building the UG software in S1 are as follows: S101: First, import the CAD drawing into the UG software and freeze it as a base map; S102: input other CAD plane drawings in turn, set the capture point as a vertical point through two-dimensional capture, and stack the imported layer contours together, then move the upper layer along the Y axis as a reference, and then import other structural images, simulate the contour lines, and then freeze; S103: Generate an entity by importing multiple layers of images and then connecting the images to form a 3D model image; The construction of the finite element model in S1 is as follows: S111: Discretization of a structure is to divide a structure or a continuum into many units. Therefore, when starting the analysis, the structure must be modeled with appropriate finite elements and the number, type, size and arrangement of the units must be determined; S112: Take one of the units from the region or structure for study, and select an appropriate interpolation mode or displacement mode to approximately describe the displacement field of the unit. Since the displacement solution of a complex structure cannot be accurately known in advance under any given load, the difference mode is taken as a polynomial form. From the point of view of calculation, the polynomial is simple and meets certain convergence requirements. After the unit displacement function is approximated by a polynomial, the problem is transformed into how to calculate the node displacement. After the node displacement is determined, the displacement field is also determined; S113: Derivation of element stiffness matrix and load vector. Based on the assumed displacement pattern, the element stiffness matrix and load vector can be derived using equilibrium conditions or appropriate variational principles. S114: The total equilibrium equations are obtained from the collective unit equations. The continuum or structure is composed of many finite element units. Therefore, when performing finite element analysis on the entire continuum or structure, it is necessary to combine them; The process of simulating friction welding in S2 is as follows: S21: A constant pressure is applied to the friction welding part through the pressure structure, so that the welding structure can fit tightly during the friction process: S22: The welding head is rotated and pressed on the pre-welded multi-piece wheel rim structure through a power mechanism. In the first stage, a constant displacement is pressed down to fully insert the welding needle. In the second stage, welding is performed with a constant pressure and a constant speed. After welding is completed, the welding needle is slowly pulled up to make the multi-piece wheel rim structure fit tightly. S23: measuring simulation parameters. After welding is completed, the defects and collapse of the welding parts are detected through detection, and data parameters are output; The measurement simulation parameters include the number of collapses and defects in the welding part, and the diameters of the collapses and defects at various locations, and the thickness, elastic deformation, plastic deformation and size of the welding node of the welding part are detected; The collapse defect prediction model in S3 uses linear regression and Delphi method to calculate and predict the collapse defect prediction model result; The Delphi method is to consult each member's prediction value, and then consult each member's prediction value again after mutual reference. After several rounds of consultation, the prediction value is made consistent. First, the prediction value of each group of data is consulted. After several rounds, the distribution of the prediction value and the prediction result are statistically consulted, and the representative median is used as the final prediction value. The judgment is made based on the number of data before and after the final predicted value, and the data information is transmitted to the regression equation for fitting.
2. The method for establishing a friction stir welding structural defect prediction model using the finite element method according to claim 1, characterized in that: The physical property data in S1 include the thickness of the welding part, the diameter of the welding part, the hardness of the welding part, the elasticity of the welding part and the heat transfer parameters of the welding part.
3. The method for establishing a friction stir welding structural defect prediction model using the finite element method according to claim 1, characterized in that: When the pressure structure first applies pressure, the first stage is defined by a displacement of 3mm-10mm, and the pressure applied by the pressure structure in the friction stir welding walking state is 1kN-5kN.
4. The method for establishing a friction stir welding structural defect prediction model using the finite element method according to claim 1, characterized in that: The measurement simulation parameters include the number of collapses and defects at the welding part, and the diameters of the collapses and defects at various locations, and the thickness, elastic deformation, plastic deformation and size of the welding node at the welding part are detected.
5. The method for establishing a friction stir welding structural defect prediction model using the finite element method according to claim 1, characterized in that: The regression equation was fitted using the least squares method: By using the least squares method to find its equation, we can calculate the straight line y=bx+a. There is often more than one factor that affects y. Assuming there are k factors x1, x2, ..., xk, consider the following linear relationship: y=β0+β1x1+β2x2+…β k x k +e Make n independent observations of y and x1, x2, .., xk at the same time to obtain n groups of observations (xt1, xt2, ..., xtk), t = 1, 2, ..., n (n>k+1), which satisfy the relationship: y=β0+β1xt1+β2xt2+…+β k xt k +e t Among them, ε1...ε n They are all independent random variables with the same distribution as ε. In order to express the above formula in matrix, let: So we have Y = Xβ + ε, Use the least squares method to find the solution for β in, (X T X) -1 X T , is called the pseudo-inverse of X.
