Calculating loads in wishbones etc.

[murpia]

A while ago, I came across a technique for calculating the loads in the various parts of a suspension, such as wishbone legs, pushrod, trackrod etc. However I've lost the reference and plenty of Google searching hasn't turned anything up yet.

Can anyone help me with a reference to the technique?

From what I remember, one took the x,y,z points of all the links, plus the contact patch of the tyre. Then the vectors of all the links were put into a matrix, which was inverted. From the inverted matrix data one could then apply x,y,z components of a force, or a moment about x,y or z at the contact patch and the resultant load in each link would be calculated. 'Dodgy' geometry resulted in a singular matrix as the system would not be properly constrained.

The same method worked for 5-link as well as wishbone, a wishbone being a special case where the outboard points of some of the 5 links are coincident.

Thanks, Ian

 

[GregLocock]

Never seen it. Should work as described, with the exception that "'Dodgy' geometry resulted in a singular matrix as the system would not be properly constrained." is not reliable.





Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[murpia]

Thanks Greg,

I remembered one more thing, the technique wasn't advertised specifically for suspension design, but was a general technique for finding the resultant force on a body subject to an arbitrary set of forces.

To use it for suspension the body represents the upright / wheel assembly and one applies 6 forces in the direction of the suspension links, and a 7th as the contact patch force.

Of course, if anyone has another technique to find the link forces in a suspension, I'd be interested in hearing about that too.

Thanks, Ian

 

[GregLocock]

Well, the usual manual method I'd use is to apply a point unit load at the CP in each of the 6 directions (X Y Z RX RY RZ) and then calculate the arm forces, then use superposition to find the arm load for a given known load case. That's just a bunch of free body diagrams.

But in practice I just apply a whole series of load cases to an ADAMS or equivalent model and let it sort everything out for me - this can handle dynamic loads and non linear geometrical effects, not just pseudo static forces, and the output can be used to drive fatigue life studies.

Another approach might be to do a multibody FEA, if you have lots of balljoints that should work OK.

The big unknown is not the force distribution inside the suspension, it is the loads that need to be applied. Not easy info to get unless you have a decent budget.


Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[murpia]

For the load cases, the assumptions we're making are as follows:

Peak vertical contact patch force = mass + weight transfer + aero load + bump G

Then a tyre mu is assumed, which generates a lateral contact patch force. This lateral force is applied at 10 angles in the xy plane which correspond to the worst case loadings of the 5 suspension links (not including push / pullrod) in tension & compression. For each of the 10 force angles the resultant accelerations and subsequent weight transfer is iterated a few times to refine the results.

Obviously tyre mu and bump G are critical assumptions to get right, but we have some data. Of course we are using a safety factor in the link loads also, then assessing each link for failure in buckling and tension to try to optimise the design.

Regards, Ian

 

[GregLocock]

I think you need to add kerb impact as a special case, if you are circuit racing and expect to be fast.

Rule of thumb is that 3-2-1 (z x y) is OK for light circuit work in non aero cars, 5-4-3 is more like it for durability, and aero cars are different again.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[murpia]

I assume you mean acceleration in G by those numbers?

If so we expect about 6-3-3 from the car based on preliminary data, but using the analysis technique I described above we will calculate link loads based on vertical forces and mu, which we expect will exceed the ~4.5 G combined cornering / braking case.

We want the analysis to cover the worst case 'spin' situation.

Regards, Ian