- cross-posted to:
- [email protected]
- cross-posted to:
- [email protected]
Crossposted from https://fedia.io/m/[email protected]/t/4559012
Maybe engineering was a mistake.
Crossposted from https://fedia.io/m/[email protected]/t/4559012
Maybe engineering was a mistake.
The elastic modulus of concrete is calculated as 57,000 times the square root of f’c, where f’c is the 28-day compressive strength of a concrete cylinder. (The force required to break the cylinder divided by the cross-sectional area.)
This formula, the 57,000sqrt(f’c), is a vitally important formulas. The elastic modulus is used to predict deflections of concrete elements and other calculations.
This equation also cannot be derived from first principles; it’s purely empirical. Engineers just made and busted a whole lot of concrete cylinders and beams and fit a curve to it.
This method has been used for basically every concrete building in the United States. Sleep well tonight.
Why would people not sleep well knowing that? Derivation from first principles is only as good as our representation of those first principles; no model perfectly represents its system and complex models especially so. This is why we do experiments in the first place, to confirm that we haven’t bungled the math. That concrete buildings are designed based on the result of exhaustive testing rather than an abstractly idealized vision of the world should be far more reassuring than the alternative.
Because we don’t know why it’s true so we don’t know when it might be false. Doubly so for something like concrete which is a mixture whose proportions and ingredients can vary. Obviously most people in well-designed buildings are under little risk of having concrete pillars collapse on them, but generally it’s better for safety if we know how stuff works.
Why? An experimentally derived model is generally not the same as one derived from first principles true, but that has no impact on it’s applicability. For example, we know experimentally from within the set of controlled parameters when it is false and when the experimental model collapses, and as with a FP-derived model we can make make accurate extrapolations about other situations with known conditions from there.
(I’m sure I have better things to be doing with my time besides reading matsci papers from 1936, but this turns out to be a fascinating topic. Particularly, the use of the Elastic Moduli was known at the time to be a way of simplifying the actual behavior of concrete, which is (for a given value of “not”) not an elastic material. However it looks like as far back as [old enough they’re using Fortran] we have been able to derive an accurate model of concrete behaviors (and have been aware of how one could be created), but it took decades for computing power to progress to the point that we could test it, as it relies on the simultaneous physical interaction of ~30 different parameters across the time domain in what I’m blearily going to describe as an ‘ur-FEA’…)
(seriously it turns out aggregate modeling is a fascinating topic, can’t believe I’ve never found this one before)
This is why factor of safety is a thing. Bigger factor of safety = greater confidence it won’t break. Builders cheapening out on material costs worries me more.