Engineered CAD models are typically not ready to simulate using FEA as-is, and require modification before doing so. Oftentimes, there can be features on CAD models that are unessential to the accurate simulation of objects, or can actually create false information.
Such features might include intricate details, minute fillets, or tiny bolt holes that, while critical for manufacturing or aesthetics, do not significantly influence the overall structural behavior during simulation.
Including these superfluous features can unnecessarily complicate the meshing process, extend computational times, and may introduce errors or localized stress concentrations that aren't representative of real-world behavior.
It's crucial to differentiate between design details that are relevant for analysis and those that can be safely omitted.
Simplifying a CAD model by removing these non-essential features can lead to faster, more efficient, and more accurate simulations. This streamlined model, often referred to as a "defeatured" model, allows engineers to focus on critical design aspects and obtain results that are more reflective of real-world performance, without being obscured by unnecessary complexities.
FEA works by breaking down an object into a vast number (thousands to millions) of finite elements, such as little cubes/squares or tetrahedrons/triangles. This technique is known as Discretization, or "Meshing", and converts solid body CAD objects into tessellations of triangles (mesh bodies).
The greater number of individual finite elements there are in an FEA study, the more accurate the simulation will be, but it will also require more processing power and time from a computer to solve the scenario
Conversely, the fewer number of individual finite elements there are in an FEA study, the less accurate the simulation will be, but it will also require less processing power and time, so can be done more rapidly
Determining the "best" number of finite elements is finding the Goldilocks "just right" amount, and depends largely on the geometry of the model and the scenario
Low-stakes scenarios - simple, small objects, with minimal requirements of safety and performance - can utilize fewer finite elements, as learning
High-stakes scenarios - large, complex objects/assemblies, with great requirements of safety and performance - typically lean towards maximizing the number of finite elements, as accuracy of simulation is more important than cost or time
One common approach is to simulate using FEA multiple times, starting with the fewest number of individual finite elements, and gradually increasing the number of finite elements with each additional simulation
This method allows engineers to be efficient, gleaning key pieces of information about their designs rapidly, and progressively refining their designs and their understanding of how their design reacts to the scenario(s) it will perform under
Constraints, often referred to as boundary conditions, restrict the movement of nodes in the FEA model in one or more directions. They simulate how the component is held, supported, or attached in the real world.
Without properly defined constraints, the FEA model would be free to move in space, and the analysis wouldn't be representative of the actual physical scenario. Additionally, over-constraining or under-constraining a model can lead to incorrect results.
Loads represent the external forces or effects applied to the model. These can come in many forms, depending on the specific analysis being carried out.
Examples include mechanical forces (tension, compression, torsion), pressures, thermal loads (temperature changes), gravitational loads, and more.
Loads are essential to simulate the actual working conditions of the component or structure. Without them, the analysis wouldn't generate any meaningful stress, strain, or deformation results. Incorrectly specified loads can lead to misinterpretation of a component's behavior under real-world conditions.
Material properties must be assigned to each component to ensure accurate simulation.
This means not only identifying the right material, but ensuring the associated characteristics — such as elasticity, thermal conductivity, and density — are correctly inputted.
Overlooking or generalizing these properties can lead to significant discrepancies between the simulation and real-world performance, potentially compromising the validity of the study and the safety or effectiveness of the end product.
Precise and accurate assignment of constraints, loads, and materials ensure that the FEA model replicates the real-world conditions as closely as possible. Methods to ensure proper assignment of their properties include:
Thorough Understanding: The engineer should have a comprehensive understanding of the component's operational environment and its interaction with surrounding structures.
Empirical Data: Wherever possible, use real-world measurements and data to define loads. For instance, using sensors or load cells to measure forces in a physical system.
Sensitivity Analysis: Perform analyses with varying conditions to understand how sensitive the results are to changes in loads or constraints. This can help identify if certain conditions are overly dominant or if they've been incorrectly specified.
Review & Double-Check: It's always beneficial to have peer reviews. Another set of eyes might catch errors or inconsistencies that were initially missed.
Physical Testing & Validation: Compare FEA results with experimental results from physical testing. If the two align closely, it adds confidence to the accuracy of the constraints and loads specified in the simulation.
Iterative Approach: Start with simpler models and boundary conditions, and progressively refine and iterate based on the insights gained from each analysis.
When simulating structural loads to your Generative Study, it is important to understand the correct interactions between the part your are generating and any touching (contacting) parts. Correct contact assignment is critical for proper simulation of adequately-capable parts.
Contact types include:
Bonded: A bonded contact type is where two bodies are perfectly bonded together at their point of meeting with no relative movement allowed throughout the analysis.
Offset Bonded: An offset bonded contact type will bond together bodies that are not touching each other.
Separation: A separation contact type where contact bodies cannot penetrate each other but can partially or fully separate from one another. You can also add a coefficient of static friction.
Sliding: A sliding type contact is where contact bodies cannot penetrate or separate from each other. However, the surfaces can freely slide together in a face tangential direction relative to one another.
Rough: A rough contact type is where bodies cannot penetrate each other but they can partially or fully separate from one another. However, while in contact, they cannot slide relative to each other.
Free/No Contact: A free contact type prevents interaction between the selected entities.