2
$\begingroup$

Most of the calculations I know and can find can solve for the radius if the cord length is known or the angle is known along with the height of the arc segment. However I cannot find, nor figure out how to calculate the radius if only the arc segment and the height of the segment is known.

I have looked up several different methods on this and other site and the following are similar but not exactly what I am looking for.

In CAD I can draw a arc like what is shown at (http://mathworld.wolfram.com/CircularSegment.html). If I fix the mid point at (0,0), the set the cord horizontal, then add a dimension the segment height (i.e. h=2") and then dimension the arc length (i.e. s=10"), the geometry is fully constrained. So I know there is a way to solve this.

If I reversed the inputs and had the arc length (s) known and the radius (R) known I can calculate the segment height but I am having trouble reversing the equation.

$\endgroup$
  • $\begingroup$ What is the height of the arcsegment? The altitude of the triangle with vertex at the centre of the circle and basis the chord? $\endgroup$ – Bernard Apr 12 '16 at 20:34
1
$\begingroup$

As Aretino answered, you need some numerical methods to solve the equation $$L\big(1-\cos(\theta)\big)=2H\theta$$ (remember that there is no analytical solution to $x=\cos(x)$).

However, for $-\frac \pi 2 \leq \theta\leq\frac \pi 2$, you could use as a quite good approximation (see here) $$\cos(\theta) \simeq\frac{\pi ^2-4\theta^2}{\pi ^2+\theta^2}$$ which makes the equation $$2 H \theta^3-5 L \theta^2+2 \pi ^2 H \theta=0$$ which reduces to a quadratic $$2 H \theta^2-5 L \theta+2 \pi ^2 H =0$$ the solution of which providing a reasonable starting point for the iterative process.

For illustration purposes, let me consider $L=10$, $H=2$. The retained solution of the quadratic is $$\theta=\frac{1}{4} \left(25-\sqrt{625-16 \pi ^2}\right)\approx 0.846955$$ while the "exact" solution is $\approx 0.849952$.

Edit

Setting $\lambda=\frac{2H}L$, the initial equation write $$1-\cos(\theta)=\lambda\theta$$ Assuming $-\frac \pi 2 \leq \theta\leq\frac \pi 2$, the approximation leads to $$\theta_0=\frac{5-\sqrt{25-4 \pi ^2 \lambda ^2}}{2 \lambda }$$ Now, peform one iteration of Newton method using the exact equation; this will lead to $$\theta_1=\frac{\theta_0 \lambda +\cos (\theta_0 )-1}{\sin (\theta_0 )-\lambda }+\theta_0$$ For illustration, using $\lambda=\frac 25$ as before, this leads to $$\theta_0\approx 0.846955$$ $$\theta_1\approx 0.849961$$ while the exact solution woudld be $\approx 0.849952$. Continuing the iterations for many more figures, the itrates would be $$\left( \begin{array}{cc} n & \theta_n \\ 0 & 0.846954969749869 \\ 1 & 0.849960525069669 \\ 2 & 0.849952018812858 \\ 3 & 0.849952018744878 \end{array} \right)$$ which is the solution for fifteen significant figures.

$\endgroup$
0
$\begingroup$

If $2\theta$ is the angle related to the given arc, having length $l$, radius $r$ and height $h$, then you have $l=2\theta r$ and $h=r(1-\cos\theta)$. By eliminating $r$ you are left with $l(1-\cos\theta)=2h\theta$, which in general can be solved for $\theta$ only numerically.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.