Resonate Chassis Frequency

[Ampto]

Dear fellow FSAE peers.

Firstly I'll introduce my self, I'm the chassis/dynamics engineer for this years team at the University of Ballarat. The team as been on and off again over the last few years and we have raised the project from the dead after a 5 year break (I'm still debating weather this was a good idea!). We have finished most of the design phase and we are keeping things pretty simple at this stage.

Now onto that engineering stuff...

How many teams are now using modal frequency simulation of the chassis as a design parameter?

I've approached the design of the chassis by finding the best compromise between mass and stiffness rather then just aimlessly triangulating to meet x amount of stiffness. It just so happens that the frequency of an object relates those two perfectly. I've used both the standard torsional twisting model and the frequency model to help direct the design of the chassis and compare the two methods for accurately, i figure if torsional stiffness is increasing as well as the torsional mode frequency i must be doing something right!

I've seen a few things floating around the forums and well as a few papers but i haven't seen a decent discussion on this.

Cheers Patrick

[BillCobb]

This technique is very effective for the purpose you state. But, you must use the loaded chassis (meaning all suspension pieces attached, powertrain installed and lumped masses [battery, fuel etc. in place]). Then you look at the modal decomposition to see which modes have the highest participation in a constrained maneuver (yeah, like cornering). Then you look at those mode shapes and you will see where to concentrate your structural modifications/reinforcements. THEN you must redo the analysis because your change means a change in the normal modes. Not sure what package you are using, but NASTRAN (for example) is well equiped to do this. BTW, the modal decomp is a well known technology to reduce the analysis time for any FEM because only the participating modes of the model that are important for the test procedure need to be there. The rest are non-players. This can have a huge time and cost saving benefit to your analysis because you are not solving the consitutive equations (by matrix inversion). But, the who enchilada has to be assembled, either from normal modes or parts.

[Big Bird]

Bravo, Patrick. Effective engineering, I say. Nice work.

Cheers,

Geoff