One of the foreign concepts to many people new to FEA is the idea of "Convergence". Convergence is an analytical method that many analysts use to determine the quality of their FEA results. Since FEA uses small elements to solve complex problems, a larger number of smaller elements can sometimes yield more accurate results. But how small is small enough when it comes to element sizes? This is where convergence comes into play. Convergence in Inventor Simulation is actually a series of settings that can be used to automatically make mesh elements smaller, and help determine if results are accurate.
The basic concept is that the mesh will automatically be made of smaller elements and solved until the results of the refined mesh fall within a percentage of the previous mesh. In other words: The smaller mesh is no longer significantly changing the results, and it making it smaller would yield diminishing returns.
In the image below, convergence settings were used to determine the validity of the FEA results. Before using convergence, the stress in the model was calculated to be 5.85 ksi. After turning on convergence (which refines the mesh and makes it smaller), the stress was calculated to be 6.205ksi with only .448% defiation from the previous iteration. This shows us that we can trust our FEA setup and gives us more confidence in our solution.
Contributed by Ben of the Tata Technologies CAD Geeks
