Is the modulus of elasticity linear?  Depending on the type of fibers, it probably isn't.  The extreme fiber can be said to fail when it meets maximum strain.  The stress distribution through the thickness of the specimen may not be linear.

Sorry I can't be more helpful - I have a feeling you know this stuff already.  Any progress since you posted this?

Steven Fahey, CET

Thanks for the thought SparWeb. I hadn't thought about the modulus being nonlinear. It should be ok in tension (this is standard modulus carbon, like Tenax HTS). I'm not sure about compression. I'll see if I can find a stress/strain curve. Let's see: if the modulus dropped with increasing strain, then compression stresses would...drop? You'd get a neutral axis shift towards the tension material? I'll have to think about it. That's definitely another possibility for at least part of the effect. The ratio of bending strength to compression strength often seems to be at least 1.3, and can be more.

Someone else mentioned that as the specimen bends you begin to get non-linear curved beam effects, so that the neutral axis shifts towards the compression face and the strain distribution begins to go nonlinear through the thickness. However, curved beam theory is pretty clear that the maximum compression stress gets *bigger* in this case, so it would be a weakening effect (unless you're tension critical). However, for sensible curvatures at failure it's less than 1%.

We've specified more than one thickness for bend testing. Hopefully, when this takes place sometime next year, we'll be able to see whether or not there's a size effect. If there is, then we'll maybe have to do a bit more testing. I'll try to remember to post something here then.

I got wondering about how the neutral axis of bending would be affected, too.  I didn't say anything for fear of foot-in-mouth disease.  The modulus of elasticity in all directions is dependent on the direction of the plies in the lay-up, properties of each layer, and so on.  There are ways to analyze the laminate you have, but the tests you propose may be the only way to know.

Steven Fahey, CET