Realistic Frame Load Cases in FEA

[Buhweet]

Hi,

I'm attempting to load the frame in a manner similar to real life in my FEA software. Torsion testing is not too much of a drama but I don't think it simulates reality very well.

The biggest drama seems to be how to effectively restrain the chassis. I have found all the forces at the tyres under the worst case, but applying only these to the chassis isn't quite right, it needs to be restrained somehow.

In reality the cornering/braking loads are applied through the CG and reacted through the tyres. Seems simple at first but how do you apply a force to the CG when it isn't actually a real point on the car? Or apply the forces at the tyres and restrain the CG?

I've thought about breaking down the inertia loads and distributing them throughout the nodes in the chassis, but this will probably take a very long time.

I'm using ANSYS and apparently it can do dynamic loading but it seems like you need to be a real brainiac to get your head around it.

Buhweet

[Angry Joe]

Ideally, the car should be modelled with full suspension, only the shocks should be replaced with rigid links. I'm pretty sure there is a way to model a pivot for the bellcranks, but I don't know how to do it. Constrain the car at 3 of the 4 uprights and apply a load at the 4th. I'm sure there are more accurate ways to simulate loads on the chassis, but you don't want to waste too much time on it - that was one of my mistakes http://fsae.com/groupee_common/emoticons/icon_biggrin.gif

[Buhweet]

Hi Angry Joe,

I tried your suggestion to constrain 3 uprights and load the fourth, the problem I found here was if you apply the previously calculated reactions to the un-restrained upright the sum of the forces is less than the force applied the CG. In other words the chassis is not loaded enough.
If I increase the force applied to the upright to account for this, the forces in the wishbones are too high (the chassis is overloaded). I know it's safer to use the higher loads, but I'm worried the frame'll end up being too heavy.

Buhweet

[Angry Joe]

Unless you really know what you are doing in FEA, I would not use it to determine weather a part will fail, or even gather deflection data. In my opinion its best use is comparing relative designs - does one design deflect more than the other? Are there tubes that aren't taking much load, or tubes that are taking too much? I'm sure with enough time you could get accurate data on deflection and failure, but that time is probably better spent elsewhere (at least in my opinion - that's all I am capable of doing with it).

[Denny Trimble]

Joe, what you're describing is trend analysis, and it's good for comparing different shapes when you're not confident that you're getting accurate numbers. We did this in the initial design phase of our monocoque, because we hadn't done tests on layups and corner sections yet.

But especially with steel structures, if you get your constraints and loads right, you can have a very good correlation between results and reality. Basically it comes down to loads and constraints.

I guess for buhweet's original question, if you were to take the sprung mass and CG height at 10 stations along the length of the car, and apply inertial loads to the nodes at those stations, you could get an accurate load case for cornering. We usually do a point load from the driver's CG and the engine's CG, and worry more about torsional stiffness because bending and lateral are usually higher, and to meet our stiffness targets, stress isn't very high.

[Buhweet]

Denny,

That was my next step, to distribute the inertia loads throughout the nodes on the car. It seemes like it will be pretty time consuming so I was trying to avoid it. But I'm about to bite the bullet and try it right now. I'll let you guys know how it goes.

Buhweet

[Buhweet]

I had a go at trying to artificially load the model dynamically, but it got too hard (and time consuming). What I come up with was loading the wheels with the calculated worst case (cornering and braking) and restraining key points around the CG, eg cockpit and engine mts.

The thing was, as most people told me, the actual stresses in the frame are quite low for normal (or even extreme) driving conditions once the chassis is designed for torsional rigidity.

I suppose the method used to mount all the components to the chassis is where it gets tricky. To effectively transfer loads from the component to the frame with minimal bending.

Buhweet