Composites in non-redundant safety critical role

[CHagen]

My question is regarding reliably transmitting torsion via a shaft in a safety critical application. It seems like composites are often kept out of such scenarios where there is no redundancy and safety is critical of the component. I want to know of some scenarios where they are used like this and how their design is treated.
My intension is to build a steering shaft for an automotive application with no joints between the rack and steering wheel. High torsional stiffness, but with very weak compressive and bending stiffness is required. My idea is to create a tube with a wave form along it's length and a large OD (~2.5") to keep stiffness high and stress low. I considered metals, but hydroforming is the only method I can think to produce this in a metal.

 

[SWComposites]

787 and A350 wings and fuselages, along with other aircraft, use composite materials.

 

[Sparweb]

You won't like these answers but I'll try to make it humorous...

1)
I attempted a similar project using a CF tube a few years ago. The replaced component was steel, and CF promised a weight saving of 50%. Once the part was designed and tested, it was obvious (pretty quickly) that the contact stresses at the mounting points would require a redesign of the mounts, plus a steel sleeve bonded over the CF tube at the mounts to distribute the contact stress. Not only did the mount redesign give back most of the weight savings, but the sleeve made inspection of the CF impossible. You will face similar challenges using CF. None of the common matrix resins used for composites have high enough bearing strength to withstand concentrated forces like bearing or sliding.

2)
A machine like the one you described was invented over 1000 years ago and was nearing perfection 100 years ago. Then the automobile was invented and all development on the wooden cart axle was frozen in time. Since wood is an "old" material, is seems to have no value or use in the modern world. Since engineers who consider their work to be safety-critical design today have very little contact with wood, I don't know how you would bridge the knowledge gap. Even if I could convince you that it could be an appropriate choice, you would probably face an uphill battle unless you can build a prototype "in secret" and show-em-all you're right.

I have a somewhat biased opinion against the biased opinions that many people have against materials that are "out of fashion".

http://www.sparweb.ca