Finite element analysis works by simulating how materials and systems work at a low level. Depending on the nature of your project, you might wonder just how granular an FEA services provider can get with their simulations. Let's look at what influences the granularity of projects in the FEA consulting world.

Computational Power

In the strictest sense, computational power is the only definite limit on analysis granularity. As evidenced by simulations going well into the subatomic realm, you can analyze systems and materials far beyond anything most customers need. The limit, of course, is trying to analyze anything at that level requires processing power at the supercomputing level. Notably, this sort of work tends to be the domain of governments and research institutions.

Most people will never need anything approaching that level of granularity, though. Even if an FEA services firm has the computational power to get within several orders of magnitude of atomic analysis, customers will likely want to invest the computation cycles elsewhere. You might want to test 100 years of weathering on the materials, rather than see how they perform at an extremely granular level.

Understanding of Materials and Systems

FEA service professionals frequently use established knowledge of materials and systems. Most customers don't have jobs that require reassessing how steel beams of standard sizes are going to perform. Data from current and structures will tell us how certain materials and systems will perform.

For example, finite element models of structures typically incorporate data about steel performance from sources like bridges in high-wind regions. There's normally no need to develop a new model of hurricane-force wind performance because the Gulf Coast exists. You can build a model based on the data from actual events, and many software already incorporate these models out of the box.

Iteration of Simulations

At this point, you might wonder how you should decide the level of granularity. Fortunately, you don't have to commit to a level right away. You can run an analysis and test its accuracy and consistency. More importantly, you can continue to iterate through simulations and increase granularity until you reach a desired level of accuracy.

Simulations also typically allow stopping. For example, you might pay for a set amount of computation. The analysis firm can store the state of the simulation at the end of the current run. If you're not satisfied with the accuracy level, you can usually resume where the last run left off.