Thanks for your insightful comments. It’s also nice to hear your perspective on the 1.6-watt magnitude. I admit that it seems to me that 1.6 watts gets lost in the noise of chain lube, jockey wheel bearings and tire rolling resistance. On the other hand, when you put 1.6 watts in terms of time and weight, it certainly seems more significant.
Thanks for clarifying your use of “heat.” It’s broad–all losses end up as heat, eventually–but I see what you mean. That said, consider out-of-plane bending of the crankarm (as when you stand on the pedal at BDC). When you unload the bottom pedal, the crank returns to its unloaded position without adding to forward propulsion and without literally heating the arm. (We agree that there’s a very small amount of hysteresis; I suspect some of the remaining losses are as low-frequency sound waves). That bending absolutely happens in a “real” pedal stroke, and that’s one component of the total strain energy that isn’t returned to the drivetrain.
But beyond these overly simplistic sub-cases, we’re just speculating. While I have some intuitive guesses about the subject, I don’t know the answer to the energy return question. Finding the answer would require either really well-controlled physical testing or, as you suggest, a more involved FEA model.
In order to properly address the power loss question with FEA, I would need highly detailed force data from a complete pedal revolution. You mentioned Metrigear, and it seems pretty likely that the Metrigear guys have those data. But if the boundary conditions (including force input and direction vs. time) are properly understood, such an analysis should give pretty good results. It would require a full transient solution with many load steps, but it’s doable. Sure, you could use a constant force input, but you’d still need to do a similarly involved analysis; why not go whole hog?
Your postscript hints at another point: carbon fiber cranks will certainly damp more power than aluminum ones, although how much more is unclear. Regardless of the degree of damping, there are situations to which a 700-gram aluminum crank would be much better suited than a 600-gram carbon crank. If I were facing off with Mark Cavendish in Tour de France field sprints–I can dream, can’t I?–I’d definitely want a stiff aluminum crank.
Thanks again for your thoughtful feedback.