Upper Control Arm constraints 1

[suspguy]

Hello

I am creating a Catia V5 model of a upper control arm , and also doing static analysis of the upper control arm using Catia v5 analysis(elfini solver)
I was having problems trying to contraint the upper control arm to show a realistic loading

its a double wishbone type of suspension

Control arm is A-Arm type
Initially i was constrainig all translation movements at the ball joint and rigidly clamping the bushings, but i believe I am over constraining it because the solution seem to be not right

Any thoughts ??

Thanks for your help in advance....

 

[GregLocock]

You need to allow axial motion of one of the bushes, and release the bj in that direction. You also need to release the bj in y. I'm assuming your force is essentially in y

Cheers

Greg Locock

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

 

[suspguy]

Thanks Greg

The force is acting in y(Cross-car) and x(Fore-aft) direction.
Greg can you also recommend a source like a book or something that helps me determine how to actually constraint a model for FEA.

One more question that I wanted to add was if I did some basic hand calculations and used simple beam theory to calculate the bending moment in arms of the control arm and if I neglect shear should I use the same constraints ??

Can i assume the two bushings to be similar to ball joints(pivot), because I tried that and the results seem to be a little off actually a lot off because in that case bushings ends would have no moment which is actually not true.

Thanjs for your help

 

[GregLocock]

"Greg can you also recommend a source like a book or something that helps me determine how to actually constraint a model for FEA."

Some books on mechanics include a discussion of degrees of freedom, but, to be honest, no I can't remember ever reading about how to constrain FE models. How odd. In my experience a significant proportion of models are over-constrained.

If you have x loads the constraints I gave will give slightly wrong answers, as both inner pivots will react some of the x load. The proper way round that is to model the bushes as well. There's no easy solution to that one. Well, you could figure out the stiffness of each bush and the section of the arm from the bj to that bush, and then assign that proportion of the x load to each bush, and then constrain the bj in x and rz. hmm.

The torsional and coning rates of the bushes will generate moments, I would not have though they were likely to be significant in context.







Cheers

Greg Locock

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

 

[Drej]

> ...recommend a source like a book or something that helps me determine how to actually constraint a model for FEA.

No such thing. What would be the title of this book? And how big would it be? I'll guess and say, er, about a million pages (non-conservative) just to cover the majority of scenarios.



------------
See faq569-1083 for details on how to make best use of Eng-Tips.com

 

[GregLocock]

Well OK, but there must be a proper way of deciding what to use. We spend (literally) hours every year figuring this stuff out. It would be nice to have a recipe book.

Cheers

Greg Locock

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

 

[johnhors]

"I can't remember ever reading about how to constrain FE models. How odd."

Greg - not really odd at, since all you would need to know is covered by basic mechanics - free body diagrams showing all forces acting on a structure - do the applied loads balance ? Before any FE analysis is ever started, this should all be clearly understood.

And, yes most FE models are over constrained because this first step was not carried out.

See

http://www.etm-project.com/Roshaz_Art_1a.htm

 

[suspguy]

I am not a FEA analyst but most FEA guys I have talked to in my company about constraining suspension compoennets, there answer have been " we dont use constraints we used inertia relief analysis"

Well the link that john pointed is pretty good and goes over the basics,
I tried the same methodology for the A-arm type Upper Control Arm, using basic mechanics

1) Applied load at ball joint in fore-aft and cross-car direction (But this is the only load data that I have but I believe to use the 3-2-1 methodolgy we need the forces acting at all three points)

2) For free body diagram purposes I am constaring ball joint location to be contsrained in all three translations(moment =0 at bj) should i consider the front and rear bushing to be rigidly fixed , or pivot points or any thing else which can be used to calculate reaction forces

thanks for your help

 

[GregLocock]

Of course they use inertial relief.

That's the least work for them.

Johnhors, yes that's a good link. However, it does not get around the fundamental problem that you have to know all the loads... and in the case we are discussing the way that the x load is shared by two co linear bushes is not possible to calculate by hand , unless you know how stiff everything is, in which case you already know the answer.



Cheers

Greg Locock

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

 

[johnhors]

Greg - I think you can make an assumption about how the x load is shared between the co-linear bushes (say 0/100 percent or 100/0 for the worst case or a more realistic 40/60 or 60/40), but this shouldn't prevent you from a hand calculation to obtain all the attachment loads. Surely a car suspension mechanism has to be statically determinate ? (i.e. you know all the loads purely from the geometry).

As for the statement that suspguy had quoted to him "we dont use constraints we used inertia relief analysis" - I only hope that inertia relief was used to mop up small residual out of balance forces once all the loads had been applied, otherwise it doesn't bear even thinking about.

 

[suspguy]

Well john is right that inertia relief is used to just balance the residual forces
I was trying to do the same thing I wish i could show you guys my free body diagram
Let me try and explain what I am doing

Inputs

A- Arm pivot points
Fore-Aft and lateral loads at bj
Rigidly fixed bushings

Calculations

Fore aft load at bushings calculated using a aprox factor for bushing stiffness ( or 60/40 and 40/60 factor)

Lateral reaction forces at bushing calculated using a factor based on bushing offset from ball joint and bushing spacing

Moments at bushing calculated using applied and reaction forces
But if we dont assume that how the fore-aft forces are distributed between the bushings the problem is tatically indeterminate

Let me know if my approach is right

Thanks guys for your help

 

[GregLocock]

"Surely a car suspension mechanism has to be statically determinate ?"

No, not at all. None of the suspensions I work on are statically determinate, except the leaf spring, which has its own problems (and is not really determinate either)! That's actually a large part of my job - taking the time history of the loading at the wheel/hub interface (in 6dof) and applying it to my ADAMS model to generate the load at each bush or ball joint. This is then used by the FEA guys for fatigue estimates.

I agree as a first pass you could split the load 50/50 for comparison with a hand calculation, but in general it would be unwise to do so.

Just to add to the confusion, the stiffest rubber bush is often stiffer than the body to which it is attached, and stiffer than the arm to which it is attached.

So suspguy, one approach is to find the axial rate of the uca to body bushes. If they are less than say 2000 N/mm then split the x reaction up between them in proportion to the stiffness of the bushes. Much above that and you have to either identify the compliance of the body, and each leg of the uca in bending, or do an analysis at 3 levels of X force in each bush as johnhors suggests.

Congratulations on trying to do the right thing.

Cheers

Greg Locock

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

 

[suspguy]

I was trying to do hand calculations to get to an approximate solution for the control arm, since I have assumed rigid supports at the bushings using simple statics I am unable to find the bending moment at the bushing ends as it is statically indeterminate

Is there any approximation I can make for the BM at the bushings ends ??

And also would there be any moment in the arm at (where the 2 arms meet) and attach to the Ball joint, I believe there wont be any external moment(as there is a ball joint) but internal moment should exist

Thanks for your help

 

[GregLocock]

It's the internal moment in the arm at the outer ball joint that causes the indeterminacy. There are standard ways of solving this by hand. But they aren't pretty, you'll need to look it up.





Cheers

Greg Locock

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

 

[suspguy]

I know, they are like matrix solutions
it sucks to do that buy hand
explain that to my boss lol

thanks greg