Suspension Component Forces

[apriliaGT]

Hey every one. First off, yes I have searched, but I haven't really found the answer that I am looking for yet.

I am part of the suspension design team for this years car at GT and we are having a really hard time deciding whether or not our estimates for the forces through our suspension components are correct. So, I thought I might run our method by every one here for a sanity check.

We decided to create a MAT Lab script to estimate all of our forces so we wouldn't have to do all of the tedious calculations by hand and the answers would all be much more exact. So, the way the script runs is with five inputs: the resultant force from the tire in the ground plane, the normal force at the tire, the location of the tire contact patch center, the wheel radius and the suspension geometry (front or rear) from an excel file.

When the script is run it rotates the tire ground plane force about 360 degrees at the contact patch and calculates the moment that both of the tire forces create about the center of the tire contact patches for each 1 degree increment.

Then the script takes the suspension geometry and calculates the unit vector for each suspension component (front upper a-arm, rear upper a-arm, ..., tie rod and pull rod), and uses matrices to solve for the forces in the components assuming that the moment created by the suspension components about the same center point must be the same moment as from the tire forces.

Does that make sense? The only reason I am unsure of the script is that the compressive forces in our upper rear a-arms (pull rod mounts to them) are higher than the lower in most cases, which seems very odd to me.

If need be I can send the script to some one and they can run it to compare the force estimates arrived at by some other manner.

Thanks ahead of time for any help!

[exFSAE]

Have you tried doing a simple front-view statics problem of this, turning each control arm into a 2D element? (Just as a sanity check)

[Ockham]

Does that make sense? The only reason I am unsure of the script is that the compressive forces in our upper rear a-arms (pull rod mounts to them) are higher than the lower in most cases, which seems very odd to me.
That does seem odd, unless your contol arms cross. http://fsae.com/groupee_common/emoticons/icon_biggrin.gif

You might want to modify the script to output values at each stage of calculation. I usually code things that way until I have the whole script working, so that I can see where the flow goes screwy.

[Tim.Wright]

If you have a pull rod then its maybe not incorrect.

Try put only the lateral force and no normal force (it seems you can do this in your script). Then if you have compression in the top you know you have a problem.

Just put a "-1*" on the compression force and everything will be fine ;-)

Tim

[John_Burford]

aprillaGT

Pull Rod suspensions usually create large compressive loads on the upper arm. The smaller the angle between the plane of the upper arm and the pull rod, the larger the compressive load will be for a given tire normal load.

Timo

...Kinematics makes me happy
Compliance makes me sad...

True True

[apriliaGT]

So, I tried reducing the normal load to 0 and that did result in much larger forces (comparatively) in the lower arm and ball joint than the upper, which definitely makes me feel a lot better. Thanks for that idea.

It makes sense that the component with the pull rod attached to it would see the most force, I guess I just didn't really appreciate how much of a contribution it would make.

Even after that sanity check, I would really appreciate it if some one would take a few minutes and see if this script spits out the same (or at least similar) load estimates that other teams have come up with via another method.

Any one up for a potential time saver at the cost of free ninety nine?

Thanks again!

[MegaDeath]

Calculations are good, but real data is even better! Strain gauges are your friend.

[apriliaGT]

I completely agree with you on that one. Unfortunately, our suspension geometry on this years car is significantly different from last years car. So, even with the help of strain gauges, I will not be able to get any truly relevant data until it might be too late to change the design of the components significantly.

I had already hoped to get some strain gauges (along with a slew of other sensors) in the car when it is built so we can back up all of our physical and kinematic calculations. Gotta love having good electrical engineers on the team.

[Nick Renold]

aprillaGT, I have made a spreadsheet in excel that takes in x,y,z suspension points and calculates the forces using the force and moment balance equations and using the Solver functionality in excel. If you want to arrange a swap I'll be happy to compare the forces from both so we can do a double check.

[mmalinowski]

Hello everyone,

I am in the same situation as apriliaGT. I have written my own MATLAB code that allows you to input the coordinates of all of your suspension components, 3 tire forces, and 3 tire moments. It proceeds to calculate a resultant tire force in the lateral and longitudinal direction (to match with the chassis coordinate system) based off of a given steer and slip angle. The calculator later uses simple 3-D equilibrium calculations (r x F approach) to solve for the axial forces in the direction of each respective suspension linkage. My code does not go as far as any 360 degree rotations of the tire.

My program successfully computes 6 unknown forces, however under 70% of the loading conditions, the force seen by the upper a-arm linkages is higher than the ones seen by the lower linkages. I am wondering where I went wrong...

apriliaGT, I was wondering if you were able to figure out what was wrong with your code?

Maybe someone else was able to figure out what was wrong here?

One thing which I thought could be an issue with my program, was the fact that I went directly from the tire forces to the forces within the axis of each of the components. I attempted to modify my code to solve for the reactions at the sphericals of the upright (3 unknows at each attachment point - Fx, Fy, Fz), but I got stuck with too many unknowns when solving for those reactions. Since the directionality of the resultant of each of those components is not known, I had to take the scalar approach to 3-D equilibrium, meaning that I was summing the forces in three directions, as well as taking the moments of each individual force component present about each axis. I had 6 equations, but 9 unknowns - the x,y,and z component of each of the upright spherical reactions.

I appreciate all help and suggestions, and promise to return the helping hand if I ever am able to.

Cheers.

[Buckingham]

You need to add more equations to be able to solve. Think about the following:

Assume your upper A-Arm is parallel to earth ground plane with an ideal frictionless, inertia-less SLA pushrod suspension. The upper ball joint cannot support a 'Z' component. (if you disconnect the upper arm from the upright and try to apply a 'z' force, it will just translate).

[eilonoz]

Hello all
1-like you said. you can't calculate the upright reactions without to consider the direction vectors of the suspension links.
the way of taking sus' links forces with their lines of action, and multiply it by their vectors to the tires' contact patch is correct.
2- about the distribution of the forces: consider the affect of pure normal load at the contact patch on suspension with a push-rod. if the load increases, so does the tensile of the lower arm, which wants to 'take-of' from the chassis. the upper arm almost doesnt get any reaction.
now lets add lateral load at the contact patch, towards the chassis. the tension at the lower arm reduces' while the upper is starting to stretch.
i dont have numbers but the trend should look like this

bye