load paths for chassis

[carmaxxx]

i have a rough draft of a space frame chassis ready, and i want to simulate acceleration/braking/cornering loads on it using ansys.
what i want to know is, for each of the load cases, where should the forces be applied, how much load should be applied and which nodes should be constrained.
i have a brief idea about it by reading forums etc but i want to be sure.
thanks

[fade]

start at the tire and work your way to the reaction loads on the chassis

[carmaxxx]

yeah.. thats the way i did.. and i came to the following condiditons:

1. acceleration : the tyre reaction forces travel through the a-arms to the chassis through the a arm mounts. in acceleration, the chassis is being pulled along by the a arm mounts. this causes a longitudinal pull at every mounting point in the direction of travel at the rear, and in the opposite direction at the front.

2. braking : the braking forces are transmitted to the chassis from the a arm mounting, all in the direction opposite to motion of the car (rearwards)

3. cornering : for simulation, the front mounts are constrained while all rear mounts are given a suitable for to produce a moment about the rear roll centre. the same is done for the front.

4. for combined accn/braking and cornering, the above steps are combined.

i have calculated the total accelerating force to be approx. 4.5kN, to be divided equally amongst the 8 rear suspension mounting points.

do u feel this is a correct method?
can u give me your idea of the magnitude of forces acting?

[carmaxxx]

and of course i forgot to mention the weight of the car acting at the node where the shocks are mounted.

[Ben C-M]

Assuming that all your loads are going through the a-arms is very wrong. You need to basically do a FBD of your upright and determine how the load is split between the a-arms and the shock (via pushrod or pullrod).

In acceleration and braking, calculate the load transfer at max longitudinal g's to find the force going through your shock, which basically takes all the vertical load (assuming your a-arms aren't too crazy). From there you have to make sure that you consider the a-arm angles when feeding your a-arm loads into the chassis. Any longitudinal force acting at the a-arm mounts is accompanied by a lateral force as well.

For cornering, the load going through the shocks on one side are going to be alot higher then the loads on the others. Do lateral weight transfer to see the loads going through the shock.

I'm not really sure that testing these situations is the best method. The chassis should be way stronger then just meeting these cases. I would just calculate the max loads at each a-arm mount and shock mount, make sure you've got a good safety factor, then analyse to find your stiffness, because that is what you're trying to optimize. Do a search cause there are some good discussions on how to properly constrain your chassis when doing stiffness FEA's.

/Ben

[carmaxxx]

of course the chassis is going to be stronger than that.. but i wanted to do these simulations to optimise triangulation to minimise flex.

im not quite sure how to split the load correctly between the a arms and the shock. if u cud throw some light on that for me...

[Kerry]

You could also include the suspension in the FE model. Make the links rigid if you aren't looking at the stresses in them. You could also do some kind of cantilevered arm down from the hubs (or wherever your wheels mount) so you can apply the forces at the ground plane. Let your solver figure out the loads at the suspension points.

If I remember correctly, solving for the reactions at the mounts on the chassis is a statically indeterminate problem. I know adding suspension into your model is time intensive, but it makes it really easy to change loading conditions and test different hypothetical situations (what if our weight distribution was...).

[Big Bird]

Another methodology is to leave the chassis completely unrestrained and do a normal modes (natural frequency) analysis on the chassis. The major manufacturers do this in part to assess structural efficiency. Natural frequency is proportional to stiffness and inversely proportional to mass, and therefore provides a good benchmark figure to compare stiffness-to-weight ratios of different designs. Also, you get animated results for the various flexing modes, and from this you can see weak points in your overall chassis design. Hope this all makes sense. We used to perform these analyses in Nastran, can't remember the solution number (101 maybe?).

In 2003 we had a first natural frequency (torsional) of the bare chassis of around 80Hz, for what that is worth. If you try this route, remember that the first 6 results you will get are considered "rigid body modes" in the 6 degrees of freedom - so are essentially nonsense. The 7th result is the one of interest.

Hope this helps,

[carmaxxx]

thanks both u guys... helped a lot..
the frequency analysis things seems interesting!

[Bazanaius]

Hey guys, sorry to bring up an old thread but I did some searching and this thread seemed related to what I wanted to ask..

I'm going to follow the method outlined by Kerry above, and in an FSAE paper (havnt got it with me so can't remember title) that I read about optimising FSAE chassis. My question is just checking what people think about my method for calculating the loads to apply at the contact patch.

Obviously I have the weight of the vehicle, including load transfer in longitudinal and lateral directions. This then gives me loads in forward/back and side/side directions on the contact patch, factoring for the friction coefficient of the tyre on the road (how accurate do people take this to be? A recent project at uni used 1.6, am I ok to use this?). It also gives me a vertical load to apply.

Is this then it? Do people account for hitting kerbs etc? What sort of loads would this apply, assuming it acts through the centre of the wheel, perpendicular to the wheel axis? What sort of safety factor would be realistic. I have heard numbers of anything from 1.2 on a thread here, to 2 and above for other structural projects I've done. Obviously the strength of the chassis is the most important factor (esp with respect to safety) and so I certainly don't want to underengineer the thing, but then again I'd like to optimise for weight as much as possible.

Do people just optimise for stiffness then use the reccomended materials in the rules? I'm not sure I'd be happy about this, as I like to 'know' that something will work!
thinking about it now, I guess you could optimise for stiffness with your wireframe FEA model as long as the loads are in the correct proportions, then calculate the required strength of the members for your given worst case scenario load cases (e.g hitting a kerb whilst accelerating out of a corner or something)?

any thoughts would be much appreciated :-)

[murpia]

Have a look here. (http://forums.autosport.com/showthread.php?s=&threadid=98827)

For FSAE assuming no wings you might choose mu = 1.6 for your tyres and bump inputs of 4 or 5g.

As to safety factors, decide amongst your team if you prefer to apply those to the input loads, or the component strengths, or a bit of both...

Regards, Ian

[Bazanaius]

hi murpia,

thanks for your reply. I will try and get a copy of 'Racing and sports Car design' as mentioned in that thread.

And I'll have a chat to the team about safety factors.

If anyone else has any input that would be great as well!

cheers!

b

[Bazanaius]

Lol, just had a look around and the cheapest that I can find that book is £150! it seems it's been out of print a while.
anyone know where it might be sold a bit cheaper?
B

[murpia]

Did you actually read the thread?

I didn't post it to suggest you buy a book but because it contains a detailed discussion of suitable 'g' based load factors, suggestions on making an analysis more realistic and a discussion of differing safety factors for different load cases...

Regards, Ian

[Bazanaius]

I did read the thread, and thought the book might be handy in addition to what was said.
If you've read it and think otherwise than that's great - will save me £150! :-P

So from that thread..
-lateral and longitudinal accelns can be modelled as roughly 2g in each direction (+/-). Given that there is no aero on our FSAE cars, take 1g vertical load.

-it's worth investigating all possible combinations of loading. some are significantly more likely than others, but for the sake of a few more simulations...

- when optimising for weight approach a FOS of 1.0, particularly if you are confident of your material properties. Just watch out for buckling of members, e.g. in wishbones.

When you say 'bump inputs' above - do you mean when hitting a kerb?

[murpia]

Originally posted by Bazanaius:
So from that thread..
-lateral and longitudinal accelns can be modelled as roughly 2g in each direction (+/-). Given that there is no aero on our FSAE cars, take 1g vertical load.
If 2.0g is what your performance simulation suggests, then that's what you should use. The point is to _predict_ the performance of your car.

When you say 'bump inputs' above - do you mean when hitting a kerb?
Well, a kerb is usually bumpy... Bump inputs are any suspension input that result in a vertical acceleration of the sprung mass. They are often specified in 'g' as that can easily be measured by a vertical accelerometer on the sprung mass.

Regards, Ian