In this article, we’ll talk about three different types of elements in finite element modeling: Solid, Shell, and Solid-Shell, which is recently introduced in ANSYS. We’ll be looking at the similarities and differences between these three types.
1. Geometrical Characteristics.
Solid elements are 3D and typically have 4, 8, 10, or 20 nodes per element. They occupy volume within the model space. Solid shell elements are also 3D, but they only have 8 nodes per element and occupy surfaces within the model space. Shell elements, on the other hand, are 2D and may consist of 2, 3, 4, 6, or 8 nodes.
2. Geometry Limitations.
3D elements can be used for any geometry, while Shell and Solid-Shell elements are only suitable for structures whose two dimensions are much larger than the third dimension – like thin-walled plates. A general rule of thumb is to consider a body thin-walled if the smallest dimension (wall thickness) is at least 20 times less than the largest dimension (length).
3. Degrees of freedom (DOF)
DOF refer to the unknowns for which a mathematical solution will be developed for that node in FEM. Solid and Solid-Shell elements calculate results for the X, Y, and Z direction translations at each node within the element. However, Shell elements provide even more information as they also give rotation information in addition to the translation at each node, resulting in 6 DOF for each node.
4. Geometry Preprocessing.
All geometries are 3D in the real world, but we can take a 3D CAD model and generate a 3D mesh with solid elements after basic clean up. Solid shell elements might require specific preprocessing to comply with the requirements of the meshing algorithm, as they are essentially a sweep mesh. The meshing algorithm generates a mesh on the face at one end and literally sweeps it along the length/width of the component, meaning that the two faces (source and target face) must be identical.
Shell elements being 2D, require that the geometry also be 2D. 3D bodies will have to be converted into planes. Usually, a thin walled plate would be represented by a single plane at the mid surface of the cross section.
5.Meshing Freedom.
One downside of Solid Shell elements is that they require individual meshing operations for each body, unlike Solid or Shell elements, where you can define the mesh for multiple bodies with a single operation. This can be a big problem if the model has many components, as it would be time-consuming and prone to errors. It’s because each meshing operation requires unique source and target faces due to sweep mesh requirement.
6.Meshing and Solution Time
This is pretty straightforward. Shell elements are the clear winner when it comes to run time. Depending on the model size and hardware capabilities, utilizing shell elements can reduce run time many folds without affecting accuracy. Solid Shell elements take less time than solid elements but more than shell elements.
7.Through Thickness Results
If accurate stress and strain profiles through the wall thickness are required, shell elements are not suitable. This is because the thin wall assumption assumes that stress and strain do not vary significantly within the thickness.
8.Shear Locking
Shear locking is a problem we might face when using linear 3D elements. Linear elements cannot capture bending, which means the straight edges of the element remain straight when the element deforms. This makes a body meshed with linear 3D elements artificially stiff during bending, inducing fictitious shear stresses. To overcome this problem, we can use quadratic 3D elements or more elements through the wall thickness. However, these solutions are computationally expensive.
Shell elements, on the other hand, account for bending at each node, providing accurate stiffness representation. Solid Shell elements are specially designed to avoid shear locking. According to ANSYS documentation for SOLSH190, this element technology uses a suite of special kinematic formulations to avoid locking when the shell thickness becomes extremely small. However, SOLSH190 fails to pass the patch test if the element is distorted in the thickness direction due to its shell-like behavior. It is fully compatible with 3D constitutive relations and provides more accurate predictions for thick shells compared to classical shell elements that are based on plane stress assumptions.
CONCLUSION
When it comes to accuracy, 3D quadratic elements are the best, but they are also the most expensive. If you have the hardware and time to spare, go for it! Shell and solid shell elements are cheaper but rely on approximations for thin-walled structures.
If hardware limitations force you to consider shell or solid shell elements, be prepared for preprocessing and geometry preparation. If your model has both solid and shell elements, you’ll need to consider the behavior at the interface between the two.
Solid shell elements offer a sweet spot between accuracy and ease of use. They’re more accurate than shell elements but don’t require the interface considerations of models with both solid and shell elements. However, they do require a sweep mesh and individual meshing commands for each body.
To decide which type of element to use, consider your personal skill and experience, the level of result accuracy required, and the amount of time available. Whether you’re working on a personal, academic, or professional project, time is finite, and you need to choose the right element type for your needs.
I hope this article has been helpful to you in your endeavors, and good luck!
暂无评论内容