Chassis Torsional Rigidity

[ACXY]

Hi,

I know this topic has been discussed numerous times but as I observe, numbers could be falling as teams attempt to build chassis that weigh in the region of 20-25kg.

Although I am aware of how proper triangulation can improve rigidity, it appears strange how some schools have rigidity of over 1500 lbft/deg despite their chassis having only the bare essentials.

In addition, there has also been talk of 'target' values of 750, 1600 etc. Anybody care to verify this? In addition, I would also appreciate if you can share how much rigidity you are aiming for (or have achieved) this year.

Cheers.

[ACXY]

Hi,

I know this topic has been discussed numerous times but as I observe, numbers could be falling as teams attempt to build chassis that weigh in the region of 20-25kg.

Although I am aware of how proper triangulation can improve rigidity, it appears strange how some schools have rigidity of over 1500 lbft/deg despite their chassis having only the bare essentials.

In addition, there has also been talk of 'target' values of 750, 1600 etc. Anybody care to verify this? In addition, I would also appreciate if you can share how much rigidity you are aiming for (or have achieved) this year.

Cheers.

[Timmay!]

Our chassis weighs 46lbs, its all 1" round tubing. According to the FEA results, if I can do math, we saw 1867ft/lbs per deg.

Granted, our shocks have changed location kinda, which I'm hoping will only help.

[ACXY]

Hi,

Thanks for sharing. Could I ask if your FEA results were based on a 'infinitely' stiff engine? How did you model your engine?

I'm doing chassis FEA on Grape for the 1st time this year and have no idea how 'stiff' the engine is. By varying the stiffness of the engine, I could yield a wide range of values for the chassis stiffness.

I understand that I should back up my FEA with physical testing but I realised that if I do not select an appropriate stiffness from the beginning, I might end up with physical values that are very different from FEA.

[Jersey Tom]

You could model the engine and throw it in your FEA...

For the record I think our chassis was around 50 lbs and 1700ftlbs/deg

[ACXY]

That's sweet. Were those values for your 2004/05 chassis?

Do you have a stiffness value for the engine you use in FEA? What program are you using and how did you model your engine?

Pardon me but I have no idea as of how to properly model an engine in a program like Grape except to inter-link all mounting points to each other. Even then I am not sure of what geometry and material to use for these links.

Thanks once again for any advice provided. Cheers.

[Jersey Tom]

Yea that was our '05 car. '04 was real similar.

We use Solidworks and Cosmosworks.. model the engine itself as the appropriate block of aluminum, and it figures out how stiff it should be. Of course its a hack, since you dont have all the internals and whatever modelled, but hey its somethin. I'd have to imagine an engine is a pretty damn rigid structure though.

Not familiar with Grape.. never heard of it.

What you can do is just test the rigidity experimentally when you build it. From there you can back out a hack of the engine stiffness, and you have that as a starting point for future years.

[ACXY]

We use SolidWorks for CADing too. However, when it comes to chassis FEA, CosmoWorks seem to fail us. Strangely, certain faces of the tubings just don't get meshed. I presume you're using shell mesh?

Anyway, Grape is a free software that does beam element analysis. It's great for doing iterations without having go through the painful process of re-meshing everytime. If you interested, it can be downloaded at
http://www.grapesoftware.mb.ca/features.htm

As for testing rigidity experimentally, would you recommend testing the engine separately? If so, how would you restrain the engine? Fix the rear 4 pick-up points and load the front?

[Jersey Tom]

Yup, you'd want to shell mesh the chassis. Dunno what to tell ya bout it not meshing right. Make sure you have an up to date service pack and you're running an appropriate mesh size. Beam element FEA should go much quicker though.

What I'd think would work.. and you should do this anyway to validate your FEA..

Do FEA of the chassis with no engine in it. Test the torsional rigidity of the chassis itself. See if the two results match up. That's the first check.

Then, you can put in the engine, test the rigidity of the real setup. Go back into FEA and rig up a mock engine that's just basically a box of beams connected that has stiffness properties such that when you run the FEA, it matches up with the real results. Then you have a starting point for the next year's chassis design.

Kinda ghetto, kind of a hassle, but hey it might work. I dunno. I'm tired. I'm goin to bed.

[ACXY]

Hey...Thanks for all your advice Tom. Appreciated it.

Anybody else care to share about the torsional values you aim for or achieve? How about clarifying what's the target value? Thanks!

[Timmay!]

Our FEA was without an engine attached in the FEA. We had a student use Nastran for our analysis. Last year it was only off 5% from physical testing.

[F1engineer]

Imagine a model of a car with a lumped sprung mass at the cofg. 4 lumped unsprung masses. 4 Springs to represent the tyre vertical stiffnesses. 4 springs to represent the suspension. And 2 torsional springs representing the chassis torsional stiffness,
from sprung mass to front suspension and sprung mass to rear suspension.

Taking a typical car you have say 60% mass on the rear. You require say 55% of the lateral load transfer on the front.

The basic requirement for torsion stiffness is that the chassis is stiff enough to achieve the desired load transfer distribution.

The chassis is a spring in series with the suspension and tyre springs. You can't change the tyre stifnesses. You are limint on what you can do to the suspension stiffness and still achieve a decent bump response.

If you do a spreadsheet calculation of springs in series and parallel, you should be able to plot a sensitivity of lltd to chassis torsional stiffnesses.

You will note that the wheel to wheel stiffness is not much use to you, as it is important whether the stiffness is in the front or rear of the chassis.

[Lukin]

As done by Andrew Deakin et al 2000-01-3554