In this blog I would like to guide you through how we use the finite element method to tune our designs.
Finite Element Analysis, more commonly known as FEA, is a software based technique. This tool helps both the designer and the engineer reach a solution that is both aesthetically pleasing and structurally sound.
By addressing the risks at an early stage a more streamline development process is achieved with time to market and reliability benefits. Giving the designers a good view of what is happening in a clean visual way helps find new ways to solve form-giving and functional challenges.
Here at Attention we use FEA analysis in many forms: to determine extreme ultimate loads on more aesthetic plastic elements such as the Vestas skylight we designed, or to explore the effects of hard-core fatigue loads in pump shafts or purely functional parts.
In both of these cases, the process starts with good old fashioned hand calculations. Understanding the load paths and constraints is an important step towards an accurate analysis. Often these hand calculations are backed up with 2D free-body diagrams to communicate to the team what we expect to find as we go deeper. This approach to our engineered features ensures that that the team is on the same page and provides us with a good benchmark/foundation for the coming FEA results.
Now the FEA is prepared within the software environment. The first step is to ‘mesh the part’, which means the 3D part is converted into elements that define the domain of a continuous material structure. The type of mesh elements that are created during the meshing depend on the geometry of the part, as well as the type of analysis that we are carrying out.
Put simply the elements are like sticks and the nodes are like tennis balls. Every element or stick is connected to a node, and every node connects to several elements. If the mesh was fine enough each of these elements could be thought of as an atom and every element the bond between the atoms. What the software will calculate is how the applied loads on the 3D part affect the bonds or elements between these nodes – the output been a graphical view of the stresses and strain in the part. Are the sticks between the tennis balls breaking, close to breaking or perhaps totally redundant….
Next the loads and boundary conditions are added. The loads could be torque loads, direct forces, pressures, or gravity loads… in fact there are many ways to simulate real life situations.
The boundary conditions and constraints are ways of simulating the relation of the part to its environment; they could be cylindrical constraints that allow rotation to simulate bearings, or fixed constraints simulating holding parts clamped in positions. I won’t go any deeper into the other possibilities in this blog, but there are lots of other more complex cases.
Once the loads, constraints, and mesh are in place, we only need to add the material properties and we are ready to run a basic analysis. A good analysis can take from 5 minutes to 24 hours to run depending on the complexity of the mesh, the size and complexity of the part or assembly, and the amount of computing power available.
So now what happens?
Stress on a gearbox shaft.
Now we go into the FEA post-processing stage and look at the stresses and displacements. Do they match what we expected from the hand calculations? If so we can begin to trust the results. If not, it’s back to evaluating the hand calculations and examining the set-up of the analysis, more often than not it’s a decimal place error or some other little gremlin that needs to be found.
If things look good, we can focus on the stress concentrations; are they where we expected? Are there any areas where we go beyond the values allowed for infinite life (a level where if the loads are low enough the part could last forever)? Or are we looking at a part with a finite life – such as a part on a racing car that should only survive one race plus a safety factor of a few laps.
At this stage the work by no means finished; the areas of too much risk are re-designed and adjusted to satisfy safety limits. Areas that are too well engineered and too strong are reworked to bring the cost of the parts down. We run the analysis again and again with the new design iterations until both the designers and engineers are happy with the results.
So many people would now say the FEA work is complete, but this is where the good project team doesn’t stop. The same team who carried out the analysis will now be involved in the life-time testing of the prototypes, either directly or on the sideline.
Depending on the requirements from the product specification and the Design Failure Mode and Effect Analysis (DFMEA) these tests can be quite thorough, involving max loads and life time loads to validate the product as it goes into the market or on field trials.
So to summarize this blog, FEA is a tool in the design and engineering tool kit that allows teams to simulate a product or component in a virtual environment, applying the anticipated loads and evaluating the calculated outcomes. It should always be used together with more traditional calculations and live tests to verify results and give a good insight into performance and life expectation of a product.
If you have any questions on how we can help you improve your products or processes with the aid of FEA please feel free to contact us for an informal chat on the possibilities, we look forward to hearing from you!