DaveW

While browsing on an unrelated topic I came across this little gem. It's an analysis of the bending and torsional stiffness of the 928's chassis. Probably only of interest to the geeks or racers here, as it's quite technical (but with some interesting shaded images).

There's also some directly useful info as it shows exactly where to add layers of carbon fibre to the chassis to significantly stiffen it up.

Whoops, adding the link helps! http://www.vrand.com/pub/JPL_nafem.pdf

Cheers
DaveW

JEC_31

File - Save!

Thanks for the link, Dave.

I have studied Finite Element Analysis applications for plastic injection molding, but haven't got into automotive structural stiffness until now. It's pretty amazing engineering made possible with the magic of computers because we can't have the 10,000 math slaves it would take to make the millions of mathematical calculations necessary.

And how cool do they get? They use the 928 for their example. Awesome!

Any of you racers out there printing this out and calling around for carbon-epoxy quotes?

PorKen

Wow! I wonder how they input the point cloud for the computer model?

(At first glance, I thought it have come from the pic below, but they are German.)


clic pic for site

JEC_31

Actually Porken, judging by the looks of the images shown of the unibody, they got ahold of a "factory original" CAD file and not one created by scanning a "point cloud" from an actual car unibody with a CMM which is quite expensive and time-consuming. I know because I run CMMs with CAD interface working in the auto industry and have had to do such crazy things.

For studying the structual stiffness, I would think that it is very important to study a model containing every litttle detail of that unibody. If you undertake to scan an actual car, there's details that you ARE NOT going to be able to reach (or even define clearly with the scan) and therefore miss out on and possibly skew your analysis. So you really do need a "nominal" CAD file from the OEM.

I used to have a CAD file of the left half of a Stratus unibody, but left it at my previous employer because it was bigger (file size) than my backup CD could hold. We were building a fixture for some interior trim and someone sent us the wrong CAD file. So these things do get distributed throughout the industry, and with older cars they don't really enforce the "proprietary top secret" rules as long as nobody's using them to make bootleg service parts.

Yes - I am a Techno Geek.

928drvr86.5

Funny Element as we call the group here at John Deere Experimental (finite element) has not been perfected enough to eliminate the use of stress coating and strain gaging to find equivalent microstrain. It's funny because the Finite Element guys always say "Put a strain gage in X location, our analysis model shows there to be a high concentration of strain in that area." As it turns out most of the time there is not any appreciable strain in the areas that Funny Element says there should be, and all kinds of strain where it should be modest to moderate. I am sure some day it will be a more reliable option than it is right now.

928drvr86.5

At present time the best way to accurately determine strain and analyze stress is to apply a brittle coating to the test subject, perform stressful but common operations to the part, look for cracks in the coating (denoting strain), and apply strain gages in the areas of visible cracking.

SharkSkin

IIRC the 928 was the last car that Porsche designed with "Pencil and paper". Unless they switched to CAD for 87+ it's not clear how a "Factory Original" cad file would exist... But it's interesting to note that the analysis was performed on a non-sunroof tub, like the one in Ken's pic. BTW, they left out the battery box, so what's the fudge factor on that?

928SS

as a computer guy, I've lost my cherry on that type of analysis a long time ago too - nothing like a real world test, although sometimes we get lucky w/approximations....

drnick

so does all of that mean the picture is usefull? is it possible to apply this information and stiffen the chasis substantialy??

mulik51

It looks like somebody's dissertation! At least it is done in the right format.

Klim

DaveW

drnick, assuming the guy is competent, and there's no reasons stated above to believe he isn't. then it's roughly a 10% increase in stiffness you're looking at by applying those patches.

DaveW

IcemanG17

Okay for the non-engineers of the list, what is the real world advantage in stiffness that the study concluded & why would this be necessary (besides for racers?)
Brian

SharkSkin

I would think it's not worth doing for street cars. For race cars, there looks to be some good info, but otherwise I think it's mostly of use for showing off the modeling technique.

drnick

10% sounds worthwhile especialy as you dont need to weld in a lot of pipes and add weight/loose space. looks like you could get at all the major areas noted fairly easily, except the transmission tunnel.

Thaddeus

It might be a worthwhile thing to do to restore some of the original stiffness to a chassis that's been being bunged around for 20 years... some of those patches intuitively make sense... some of them do not... which is why this is a very cool thing.