1

I have a question about the sign of curvature. Here is the problem I met:

Problem: Detemine the intrinsic equations of the catenary $$\mathbf x=(a \cosh (t/a), \, t) \; (a:\text{constant)}$$ i, e, Find the cuvature and torsion of the curve parametrized by the arc length $s. $

Torsion is identically zero, so it does not matter and also I've had no difficulty to find the value of $|\kappa|$, since by computation, $$|\mathbf x'|=\cosh (t/a), \; |\mathbf {x'\times x''}|=1/a(\cosh(t/a))\\[.5em] \Rightarrow |\kappa|=\frac{|\bf {x'\times x''}|}{|\bf {x'}|^3}=\frac{1}{|a|\cosh^2(t/a)}$$ And also $s=\int_0^t|\mathbf{x'}|dt=a\sinh(t/a)$ gives $$|\kappa|=\frac{|a|}{s^2+a^2}$$ But my textbook just says: 'Eliminating $t$ gives $\kappa=\frac{a}{s^2+a^2}$, which is the required result.' without any further explanations. The rest of solution is the same exatly what I wrote above.

I know the signed curvature $\kappa$ is defined by $\kappa=\mathbf t'(s)\cdot \mathbf n(s)$, where the binormal vector $\mathbf n(s)$ is selected to be continuously defined, so that the sign of $\kappa$ is determined by the direction of $\mathbf t'$ and $\mathbf n$. $\cdots (*)$

But I have no idea that why the sign of $\kappa$ is that easily determined on the problem, and the practical way to apply (*) on computation. Is there a general way to determining the sign of $\kappa$? Thanks.

JJLEE
  • 183
  • My computation gives $$\kappa(t)=\frac{\det\bigl((\dot x(t),\ddot x(t)\bigr)}{|\dot x(t)|^3}=-\frac{1}{a\cosh^2(t/a)}.$$Hence the sign of $\kappa$ solely depends on $a$. – Michael Hoppe Apr 26 '21 at 13:00
  • @MichaelHoppe Does that hold for 2-variable curve? – JJLEE Apr 27 '21 at 04:09
  • Yes, it's simply $\frac{\text{orientated normal-component of the acceleration}}{\text{square of $x$’s velocity}}$. Please refer to https://math.stackexchange.com/questions/1272514/shouldnt-these-two-definitions-for-curvature-agree/1272699#1272699 and https://math.stackexchange.com/questions/2899242/curvature-function-fracr-times-rr3/2899454#2899454 – Michael Hoppe Apr 27 '21 at 09:52
  • @MichaelHoppe Then I guess it is not reasonable to give a sign to curvature on 3-dim curve, right? I think it is not easy to define orientation in $\Bbb R^3$ compared with the case in $\Bbb R^2$. Also I think the author choose the sign of $\kappa$ optionally in $\Bbb R^3$. – JJLEE Apr 28 '21 at 03:40

0 Answers0