RedlineMan

Geo;

I think the only load that is useful to model for is a straight t-bone, or in the case of a header tube, a simple roll. You can drive your self nuts trying to model for all kinds of loads. You have to pick something fairly easy to model (at our level), and also generic to the situation. Anything else makes it far too cumbersome to get real useful data, I'd think.

As for door bars; Flat horizontal tubes creating a rectangle to the front downtube and main hoop leg would only have the strength of the material itself in any direction you cared to force it, a rectangle being a weak geometric form. Putting "stringers" between these two tubes would then render it a beam, but then the positioning of these stringers is important, angling being best for a load trying to mash or diamond this rectangle. These stringers would add a lot to the strength, but that means a lot of mass too.

Triangles are the strongest geometric form. An X means two triangles. Quite strong in the plain of the X. Add the rolled/formed gussets and you get torsional strength as well; many forms of strength in one assembly. Light too.

Hard to describe...

NOTE: Modeling should be done with the ubiquitous standard 1.75x.095 mild steel first. Add 1.75x.120 if you like, for reference. Following that,
1.5D tube in both thicknesses would be of interest as well, for reference.

Geo

Modeling perpendicular impacts are overly simplistic. If all impacts occured this way, the model would be the paragidm. But now many impacts actually happen this way? Yes, modeling various scenarios would make a person crazy, but the only way to make an informed decision is with more than just simple data.

I agree with you about the two horizontal tubes. Add a third and add verticle supports between them. Lots of strength. Lots of mass. They key IMHO is to maximize strength and rigidity with minimal mass.

keith

D'oh! - how do I always come in late on these threads? One of you guys needs to call me when we're discussing cagework.

Nothing worse than reading through a thread, having comments and/or rebuttals, and seeing them brought up and/or answered before the thread's end!

Geo

Calling Dr. Keith, calling Dr. Keith. We need a second opinion.....

I always like your input. Of course, I know we have pretty similar views on cages.

keith

'cept I'm a Dapper Dan man!

(actually, I'm a reinforced, kinked X man)

;^)

bruinbro

Back from a weekend in NorCal. The beam model and the shell model seem to agree, so I will forge ahead and cobble up the Pyramid, Pyramid+Sill+Stringer, NASCAR and gusseted configurations and load them in a variety of ways.

BTW, a 20 g load on a 3000 lb car generates stresses like 650 ksi using linear stress analysis. I used something more moderate (4 g on a 3000 lb car) loading to compare the beam model with the shell model so I can still do linear analysis, which should be OK since we are looking at comparisons, not absolute performance.

Bro

Cory M

That will work. The 20g load was just a shot in the dark and was based on some LMP carbon tub testing requirements I read about a while back. Those cars are a lot lighter and faster than anything we'd build a cage for.

Thanks for doing the FEA. This has been a good discussion and I'm looking forward to seeing the results.

bruinbro

Here is the first results of the Basic X design. Shown are the deflection results, as stress results probably don't mean much for macro design comparisons. Please let me know if there is anything else or different that you would like to see before I march on to the other designs.

Basic geometry



Basic boundary conditions



Broadside crash results



Forward angled crash results



Head-on crash results



Let the nit-picking begin!

Bro

Geo

I don't mind showing my ignorance here. Can you help me with the scale on the right and how to relate it.

bruinbro

Best advise I give the newbies in my profession (spacecraft engineering) is to ask questions when you don't understand something.

The scale on the right denotes the magnitude of the deflection of the beam elements in inches. It is the vector sum of the X,Y and Z deflections. In the broadside and angled loadings most of the deflection is in the negative Z direction, while in the head-on crash the deflection is in the X and Y directions.

After looking ar my post I see it is not clear which way X,Y and Z are. X is sort of left to right, Y is up and Z is sort of out of the page.

Hope this helps,

Bro

Geo

Interesting stuff. What is the assumption regarding the anchors? In a properly built cage the tops of the X would be supported by additional tubes creating additional load paths.

I'm kind of shocked at the deflection. It indicated the tubing would have to stretch considerably to create that much deflection. Pretty sobering if nothing else.

I'm really interested in seeing the other models.

bruinbro

In the broadside and angled crashes the 4 posts are rigidly clamped ( all 6 degrees of freedom are fixed i.e. X, Y and Z translations along with X, Y and Z moments). In the head-on crash the rear vertical post ends are clamped while the front vertical post ends are allowed to move in all manners except for translating in Z.

I wouldn't read too much into the deflection numbers as this is a linear (non-time variant) analysis and reality would load the bars in a somewhat different manner. However, the differences between the deflection numbers for the various designs should be indicative of the relative performance of each design.

Bro

Geo

Thanks for the education. This is the most interesting stuff I've read/seen about cages in quite some time.

If you're up for it (after this round) perhaps you may want to consider modeling a more thorough cage. The reason for this is as I noted before, with a properly designed and constructed cage, each attachment point or node forms multiple load paths that I suspect can effect the results.

Thanks for doing all this.

bruinbro

Sure, if you only want design comparison numbers. As I said earlier (and was pointed out by complexx) FEA modeling of realsitic crash loading is extremely complicated and not only would be a non-linear analysis but would have to factor in a lot of other variables such as body stiffness, cage/body interface characteristics, weld joint chararacteristics and mass property changes over time. Not for the faint of heart nor for my humble capabilities.

I'm waiting for others to chime in with their desirements before I continue on.

Bro

bruinbro

As expected, removing the sill tube really didn't do anything in the broadside and angled crashes, but look at the head-on.



Of course, this assumed no vertical support from the car body nor interaction with the rest of the cage.

Bro

ETA: Just realized it's kinda hard to read the exponent sign on the scale. All of the numbers so far have negative or positive zero exponents. For example, the broadside crash result maximum is 8.17D-01 which is 0.817". The 2.11D+00 would be 2.11"

Bro