Application of ANSYS Mechanical in Welding Simulation
1. Introduction
Welding, as an indispensable process in modern manufacturing, has always held an important position in the field of materials processing. Welding is a complex process involving various disciplines such as arc physics, heat transfer, metallurgy, and mechanics. The heat transfer process, the melting and solidification of metals, phase transformations during cooling, welding stress, and deformation are key concerns for manufacturing departments and designers. The welding stress and deformation generated during welding not only affect the manufacturing process of the welded structure but also its performance. These defects are mainly caused by unreasonable thermal processes during welding. Due to the concentrated, high-energy instantaneous heat input, considerable residual stress and deformation will be generated during and after welding, affecting the machining accuracy and dimensional stability of the structure. Therefore, the quantitative analysis and prediction of stress fields at welding temperatures are of great significance.
Traditional welding temperature field and stress testing relies on the designer's experience or semi-empirical formulas based on statistical foundations. However, such methods have obvious limitations and cannot make forward-looking predictions for new processes, which leads to a sharp increase in experimental costs. Therefore, numerical simulation for welding has shown great advantages.
ANSYS, a world-renowned general-purpose structural analysis software, provides complete analysis functions and comprehensive material constitutive relations, offering technical support for welding simulation. This paper uses ANSYS as a platform to explain the basic theories and simulation processes for welding temperature field simulation, thermal deformation, and stress simulation, providing a valuable reference for enterprise designers.
2. Theoretical Basis of Welding Numerical Simulation
The temperature field, stress, and deformation in welding problems ultimately boil down to solving a system of differential equations. There are generally two main categories of methods for solving these equations: analytical methods and numerical methods. Since analytical methods can only yield solutions when numerous simplifying assumptions are made and the problem is relatively simple, numerical methods are typically used for simulating welding problems. Commonly used numerical methods in welding analysis include: the finite difference method, the finite element method, numerical integration, and the Monte Carlo method.
Finite difference method: The finite difference method solves problems by transforming differential equations into difference equations. For problems involving regular geometric properties and homogeneous material properties, programming is simple and convergence is good. However, this method is often limited to regular difference meshes (square, rectangular, triangular, etc.), and it only considers the effect of nodes, not the contribution of elements between nodes. It is often used to study problems such as welding heat conduction and hydrogen diffusion.
Finite Element Method (FEM): The FEM transforms a continuum into a discretized model composed of a finite number of elements, and solves for the numerical solution of the discretized model using displacement functions. This method is highly flexible and has a wide range of applications, thus it is widely used in fields such as welding heat conduction, welding thermo-elastic-plastic stress, deformation, and fracture analysis of welded structures.
Numerical integration method: This method uses Simpson's rule and other methods to solve problems where it is difficult to find the antiderivative. This method avoids solving complex antiderivative problems and can achieve high accuracy with fewer points.
Monte Carlo method: This method is based on stochastic simulation technology and performs numerical simulation of stochastic processes without changing the original process.
Welding simulation is usually based on the above theories to simulate problems such as welding heat conduction and thermo-elastic-plastic stress. However, the appropriate selection of heat source function and calculation of post-weld stress require designers to choose suitable mathematical models.
2.1 Commonly Used Heat Source Models in Welding Numerical Simulation
The welding thermal process is one of the main factors affecting welding quality and productivity; therefore, accurate simulation of the welding thermal process is a prerequisite for accurate welding stress-deformation analysis. Early studies on the analysis of the welding thermal process involved extensive theoretical research, proposing various heat source distribution models:
Centralized heat source: Rosenthai-Rykalin formula
This method, as a typical analytical method, assumes that the heat source is concentrated at one point. This approach is only applicable when the study area is far from the heat source. At the same time, this method cannot describe the distribution pattern of the heat source and has a significant impact on the fusion zone and the heat-affected zone.
Planar heat sources: Gaussian heat source, double elliptical heat source
Gaussian heat source
The Gaussian heat source distribution assumption assumes a symmetrical distribution of welding heat sources. This method works well at low welding speeds, but at higher welding speeds, the heat source distribution becomes symmetrical, leading to larger errors. This method is suitable for situations where the arc stiffness is weak and the arc's impact on the molten pool is relatively small.
While the Gaussian distribution provides a heat source distribution, it does not account for the effect of welding torch movement on this distribution. In reality, due to the different rates of heating and cooling in the weld, the heating area in front of the arc is smaller than the heating area behind it.
Double elliptical heat source
Volumetric heat sources: semi-ellipsoidal heat source, double-ellipsoidal heat source
Semi-ellipsoidal heat source
For gas metal arc welding (GMAW) or high-energy current welding, the heat flux density of the welding heat source acts not only on the workpiece surface but also along the workpiece thickness. In this case, the welding heat source should be considered as a volumetric heat source. To account for the distribution of arc heat flux along the workpiece thickness, an ellipsoidal model can be used to describe it.
In reality, due to the movement of the electric arc along the welding direction, the heat flow of the arc is asymmetrically distributed. Due to the influence of welding speed, the heating zone in front of the arc is smaller than that behind it; the heating zone is not a single hemispherical symmetrical about the arc's centerline, but rather a double hemispherical sphere, and the shapes of the hemispherical spheres in front of and behind the arc are also different.
Double ellipsoidal heat source
2.2 Commonly Used Methods for Simulating Welding Deformation
The dynamic stress-strain process generated by welding, and the subsequent residual stress and residual deformation, are important factors leading to weld cracks and a decrease in joint strength and performance. Therefore, the following theories have been developed for the calculation of weld deformation and residual stress:
Analytical Method: One-Dimensional Residual Plastic Deformation Analytical Method
This method, based on welding deformation theory, determines the relationship between the longitudinal plastic deformation of welded joint shrinkage and welding process parameters and conditions. It requires extensive experience and is particularly suitable for beam structures with regular, uniform cross-sections.
Inherent strain method: Inherent strain can be regarded as the source of residual stress.
The inherent strain during welding includes plastic strain, temperature strain, and phase transformation strain. After one welding thermal cycle, the temperature strain of the welded component is zero, and the inherent strain is the sum of the residual plastic strain and phase transformation strain. During welding, inherent strain exists in and around the weld seam. Therefore, understanding the distribution of inherent strain allows for the prediction of residual stress and structural deformation using only one elastic finite element method calculation. However, this method also focuses on the deformation of the structure after welding and is an approximate method, not considering the entire welding heat transfer process.
Thermo-elastic-plastic finite element method: Recording the welding heat transfer process and describing the stress and deformation of the dynamic process.
The thermo-elastic-plastic finite element method first analyzes the welding thermal process to obtain the transient temperature field of the welded structure, and then uses this result to calculate welding stress and deformation. Because this calculation is nonlinear, it is computationally intensive and is generally used to study the mechanical behavior of welded joints, rather than for the overall study of large and complex structures.
3 Welding Simulation Cases
3.1 Welding Simulation Based on ANSYS Workbench Platform
For the following components, laser welding was used. Using ANSYS Workbench as the platform, the temperature field and stress field changes of the model were simulated.
ANSYS Workbench, as a unified multi-field coupling analysis platform, supports data collaboration. Therefore, the coupling project for this welding analysis is created in Workbench, as shown in the figure below.
In this example, we are only illustrating the welding simulation process, so the material is assumed to be linearly elastic structural steel. The material parameters are entered in EngineerData as follows:
ANSYS Workbench uses ANSYS Meshing as a basis for meshing the model. The two welded parts and the weld seam in this model are meshed in a hexahedral manner. In addition, the software provides a large number of size functions, local control and other functions to perform high-quality meshing for geometric models with different features.
Performing transient thermal analysis of the welding process using the Workbench platform requires the Moving_Heat_Flux plugin, developed based on ANSYS Workbench. This plugin, embedded in the Workbench interface, provides a moving heat source distribution method based on the planar Gaussian heat source method. Users can specify parameters such as welding torch movement speed, welding current, power, and welding time. In addition, other boundary conditions required for transient thermal analysis, such as Convection, need to be input for heat transfer process analysis. The welding-related parameters input in this case are shown below:
For large-scale simulation problems like this, it is recommended to use HPC (High-Performance Computing) to fully utilize computer hardware capabilities and significantly improve solution efficiency. The final transient thermal analysis results for welding under these parameters are as follows:
Based on transient thermal analysis, post-weld stress analysis can be performed. Using the coupled analysis process of ANSYS Workbench established above, the thermal analysis temperature field is transferred to the structural field for stress analysis via the importload method.
Simultaneously, constraints were applied to the component based on the actual working conditions, and stress analysis was performed. The stress cloud diagram at a certain moment is shown below:
3.2 Welding Simulation Based on ANSYS Classic Interface
As mentioned earlier, there are many limitations when using Workbench as a platform for welding simulation, such as the inability to select other types of heat source models. Therefore, users can perform welding simulation based on ANSYS Classic. When performing welding simulation based on ANSYS Classic, it can be done in a command stream manner, with welding parameters read in as parameters, which is very convenient for optimizing welding analysis.
In this example, the welding temperature field simulation uses a welding plate with dimensions of 200mm x 200mm x 6mm and a specimen material of Q235A, with material parameters shown in the table below. To ensure penetration, both steel plates are beveled at 45°. Arc welding is used, with the following parameters: welding current 180A, arc voltage 20V, welding speed 4.8mm/s, welding heat input 0.75kJ/mm, welding efficiency η=0.825, and the heat transfer coefficient between the structure and air is 15W/(m²*℃).
Create a geometric model of this component in ANSYS Classic using Solid70. The completed model is shown in the figure below:
Create a complete material parameter table using the MP command, as shown in the following figure:
The model is locally meshed using commands such as `esize`, generating a hexahedral mesh with high quality. The finite element model is as follows:
This example also uses a Gaussian heat source for simulation, and the relevant welding process is expressed as parameters to provide a basis for later optimization. A typical command flow is as follows:
A fixed constraint is applied to the bottom of the model, and iterative calculations are performed according to the solution parameters set in APDL. The iteration curve is shown in the figure:
After calculation, the temperature field distribution cloud map of the welded part can be obtained, as shown in the figure below at a certain moment:
4. Summary
As described above, ANSYS software can be used to easily simulate the temperature and stress fields during the welding process. Currently, Workbench only supports welding simulation as a plugin and can only consider the heat source distribution of a planar Gaussian heat source. If other heat source distribution methods are required, APDL programming based on the classic version of ANSYS is necessary. In addition, users can also use the birth and death element method for welding simulation. It should be noted that the birth and death element method simulates the weld filling process by controlling the birth and death of the control unit. This method can simulate more complex heat input conditions. Since the heat source distribution and the birth and death element are two different calculation methods, they cannot be used together.
ANSYS software provides strong technical support for welding simulation through its complete material constitutive relations and solution capabilities. Therefore, designers can use it to perform welding simulations and provide reference for setting welding process parameters such as current and voltage, thereby optimizing the welding process.
