PDA

View Full Version : Chassis FEA (long)??



David Money
10-23-2002, 11:34 AM
I am stuck in a debate with a team member as to how to FEA our frame. We are convinced that we need to physically test our finished frame after construction to verify the FEA, but the question is how? I understand that some teams just bolt one end of the frame to the table and twist the other end and measure the force and deflection. My fellow team member is bringing up the point that the frame is not actually subjected to that type of loading in the driving scenario, but instead, the loading on the frame comes in through the suspension mounting points. Therefore, it is stated that the FEA model needs to be loaded through the suspension mounts to represent what is actually hapening in the competition. Not too much of a problem in FEA, but how do you do this for the physical frame testing? You can't really make a jig to bolt the frame to through the suspension mounts as that would constrain it unlike it is in reality. Any ideas??? I personally think that too much is being read into it. What does ebverybody else think? How has everybody else done it in the past? Thanks!!

Mark Ortiz
10-23-2002, 05:03 PM
It is definitely true that simply testing a frame for torsional rigidity in the shop does not replicate the combinations of loads that it sees on-track. However, it is a simple test that can be FEA-modeled, and used to at least validate the model for that relatively simple loading situation.

To make the test a little more reflective of reality, it is best to attach the beams or other fixtures at the points where the rockers attach to the frame, or where the coilovers attach if they act directly, rather than attaching fixtures at frontmost and rearmost bulkheads.

Despite the crude nature of such tests, they do provide a simple, universally applicable way of comparing one frame to another, even when the tests are performed in different locations by different teams.

It is important not to get obsessed with good performance in such tests, and ignore the rigidities of the link or control arm load points. It is quite possible to devise tests for these points, but the racing industry has no commonly agreed standard for such tests. This means you can't use the results for comparison against other cars, unless you can test the other cars yourself.

steve
10-24-2002, 03:14 AM
Make sure that the fixtures that you use to hold the chassis in place are very stiff, so they don't deflect and mess up your results.

When we did testing like this, the back end of the chassis was bolted doen to a big surface plate, and a bar was put through the chassis in the area of the front suspension, with weights hung at a measured length along the bar to give a certain torque.

A few years ago there was a bunch of SAE papers published on the rigidity of Winston Cup Chassis, we found these papers real useful for our FEA modelling and testing

David Lemire
11-14-2002, 07:33 PM
I am wondering how can we find good load to input into our FEA of the frame... we were trying to get these suspension loads from adams but it always bug and we cannot get any significant loads result out of that... any simple way to calculate them... since only trying to twist the frame makes no sence...

Thanks

David Lemire
Team Leader
McGill Racing Team

David Lemire
11-15-2002, 07:17 PM
The problem we had is that I didn't figure out fast enought that the engine had a separate load... so I included it in our frame weight... which means that I had two times the weight of the engine... now it's all fine I guess...

And thanks for the seat belts

David Lemire
Team Leader
McGill Racing Team

awhittle
11-15-2002, 07:18 PM
Lock the left rear rocker-arm mount in the x,y,z

Lock the right rear rocker-arm mount in the y,z

Lock the left front rocker-arm mount in the z

Apply a load to the right front rocker-arm mount in the z direction

If you get around 2500 to 3000 ft-lb/deg the car will work fine. At 1/2 this the car will be hard to tune. Any more will be overkill. Testing on a table will only verify errors. Most likely it is too late to fix it at that point.

see http://www.ncs-stl.com/Images/Sample.jpg

Hope this helps

Andy Whittle

David Lemire
11-15-2002, 07:43 PM
I have difficulty finding the load... and how can I get result comparable to yours if we don<t have the same initial load...

Also... how this test is representative of a real racing cornering for example... I see it has an approximation of the trosionnal stifness, but it doesn't exactly our frame will sustain the susp. loads...

Thanks to answer again... /infopop/emoticons/icon_confused.gif

David Lemire
Team Leader
McGill Racing Team

awhittle
11-15-2002, 07:56 PM
This test is not a great representation of racing. What it is is my exeriance after building and racing A-Mod autocross cars for 25 years. The stiff ones can be driven. The shocks springs and swaybars behave just like the books claim they should.

Now I am trying to figure out a f125 shifter kart. Wow is this thing messing with my A-Mod head.

Andy