MUR Blog - Chassis Stiffness FEA of an FSAE Race Car
Producing a reliable MUR race car is our top goal this year. With the design freeze deadline looming close, validating each design to ensure we achieve this goal is essential. One of the most powerful validation tools we have access to is Finite Element Analysis (FEA).We’ll go through how we use FEA to validate the stiffness of our chassis.
Our 2017 combustion chassis team is sticking to a full monocoque structure as we did in 2016. Since the team is familiar with working with a monocoque, our 2017 chassis team has been running a considerable amount of design optimization to achieve high reliability and performance. As such, FEA has been utilized since the start of our design phase. When running FEA, we’ve found that one of the first and most critical jobs to execute is torsional stiffness analysis.
Why is torsional rigidity important for a race car?
Compared to a regular car, traction and cornering balance require greater consideration for a race car. To achieve maximum speed during cornering, it is essential to ensure the car’s tires generate maximum lateral force under a certain given vertical load. During cornering, this load gets transferred from the inner tires to the outer tires. A rigid chassis will provide a highly efficient load transfer between the front and rear suspension; thus, allowing the race car to corner as efficiently as possible according to the suspension design.
The Chassis Twist
So, how does MUR evaluate torsional stiffness in FEA?
If you have any experience in running FEA on a chassis, you should be familiar with the picture above which depicts how the twist can occur in a chassis. When viewing this using FEA, we replicate this twist using boundary and load conditions when running a static structure analysis. Basically, the rear A-arm ends are fixed while two equal loads are applied at the end of the front A-arms in opposite directions. The directional deformation is then recorded to calculate the deformed angle. The chassis’ torsional stiffness is then generated using the applied moment divided by this angle. Once these steps are completed, we’ve found the following tips to be useful for us when carrying out FEA:
Engineering data
This is where the basic work of your FEA lies. Here, an incorrect setting will give you a wrong result and/or produce errors. A good way of managing your engineering data is by first defining your material data once and then exporting it, after which it can be imported to your other projects. Besides that, care must be taken when working with customized materials – ensure all parameters are checked. Our chassis team experienced a pivot error generated at the beginning of our FEA, as one material was set to be highly rigid by mistake, which cost us a huge amount of time debugging the error.
Geometry process
When running FEA, we realized that majority of our time was taken up to prepare the geometry. The main idea is to ensure that the CAD is neat and simple with only the essential characteristics present; a balance is required between representing real-world components and what is actually necessary. Both Space Claim and Design Modeler are tools we find useful. Space Claim helps in simplifying and examining geometries – it will replace Design Modeler in the near future. However, we found that Design Modeler is still a powerful tool that can process geometry at a rapid pace. When running torsional stiffness analysis, we simplified our monocoque as a shell element using ANSYS Composite PrepPost, while the roll hoops and bracing were extracted as beam elements.
Mesh and Connections
The importance of mesh cannot be emphasized enough. The quality of a mesh and the number of nodes used will determine the accuracy of the FEA. MUR uses ANSYS to run all our FEA. ANSYS’ help document is useful to read through to understand the different types of available meshes in ANSYS well. We’ve learnt that applying different mesh types to large face areas and beams will help in optimizing FEA results. Connections are important when running torsional stiffness analysis on an FSAE car because multiple physical parts are required to ensure your analysis relates to a real-world situation as closely as possible. Therefore, the connections used between the roll hoops and monocoque, A-arms and monocoque, and between other mounting tabs are required.
2017 Chassis FEA
Upon utilising these tips ourselves and after putting a lot of effort into familiarising ourselves with ANSYS, we were finally presented with quite a high torsional stiffness after several iterations of FEA. However, what did this actually mean for us? We soon realized that we had not struck a good balance between torsional stiffness and mass. Our target was not to achieve the highest possible torsional stiffness, but a high specific torsional stiffness that was strong enough to bear our suspension loads, while still remaining mass-efficient in our design.
Following this, our strategy was altered; instead of applying loads on the front A-arms, remote deformation was set. After this, the reaction force to calculate corresponding torsional stiffness was recorded. We implemented 26 design iterations to record the specific torsional stiffness change. Our results showed that maximum torsional stiffness would be achieved through using more carbon layers while specific torsional stiffness can be maximised when a combination of the number of carbon fiber layers are used on different surfaces.
Should we stop upon achieving such a positive FEA result though?
No – further verification will always need to be done. At the end of the day, the quality of the results we can get out of ANSYS is entirely reliant on the quality of the inputs we use for it. Confirming results as acceptable is a tough task. The first and easiest thing that can be done to do this is to use our engineering knowledge to question ourselves: Where will maximum deformation occur? How does stress transfer between bodies? Why is the deformation or stress not symmetrical? Answering these questions will help identify simple errors. Following this, certain ANSYS tools may be used to check for other discrepancies. For example, the Contact Tool for connect region quality analysis, force and deformation convergence graphs, solution graph, and so on. Besides that, we’ve found that it is also always useful to compare results with those of other similar projects, or in the case of FSAE, with previous years’ results.
2016 Chassis FEA 2016 Chassis
No FEA result will ever exactly reflect that of a real-world situation – it will only ever do so much. There is also a point of diminishing returns, where a limit is hit in terms of FEA skill, time and the level of accuracy needed. As such, physical testing is imperative once the chassis is fully manufactured. I like to describe FEA as an ocean. People who are skilled well in the art of FEA will find huge joy in running these simulations. Meanwhile, FEA can be a dangerous game to play for those poorly skilled in it. Thus, learning the fundamentals of FEA will provide you with a ship to travel far into this ocean of joy.
Want to hear more from us? Sign up here and never miss a post! No spam, promise.
About the Author:
Shengjie Yu
Senior Chassis Engineer, 2017