Radius of Curvature


By Arun Iyer on Tuesday, December 25, 2001 - 06:01 pm:

could anyone prove that the radius of curvature of a curve is given by
r=[1+(dy/dx)2 ]3/2 /[d2 y/dx2 ]

love arun


By Kerwin Hui on Tuesday, December 25, 2001 - 08:47 pm:
Arun,

There are two approaches that I know of. You may like to fill in the details of the proofs. One way is to use the relation

R=1/k = ds/dy

where s is the arc length, y is the angle between the positive x-axis and the tangent. Using the chain rule and the fact that tany = dy/dx, ds2 = dx2 +dy2, we get the required result.

The other way is a work directly from the definition. The circle of curvature is the circle which has y, dy/dx and d2 y/dx2 all agreeing with our curve at the given point. Now from the general equation of circle and these conditions, we can derive the radius of the circle, i.e. the radius of curvature.

Kerwin


By Arun Iyer on Wednesday, December 26, 2001 - 06:17 pm:

Kerwin,
the first way you are talking about is relatively new to me....

is k in the above equation related to serret-frenet formulae???
the second way...
can you explain it a bit please??

love arun
By Kerwin Hui on Wednesday, December 26, 2001 - 06:50 pm:
Arun,

k is called the curvature, defined as the rate at which direction of the tangent changes along the curve. To see why R=ds/dy, consider a circle of radius R, we have s=Ry providing one measures s from (0,-R).

The second way is very messy. WLOG, let us assume our point of interest is the origin, and our curve has y ' =p, y ' ' =q at the origin. We have the general equation of circle

(X-a)2 +(Y-b)2 = R2

Differentiating, we get

dY/dX=(a-X)/(Y-b)=-a/b at (0,0)

d2 Y/dX2 = -[(X-a)2 +(Y-b)2 ]/(Y-b)3 = (a2 +b2 )/b3 at (0,0).

Now force the conditions:

(1) a2 +b2 = R2

(2) -a/b=p

(3) (a2 +b2 )/b3 = q

and check all these conditions are satisfied if R=(1+p2 )3/2/q

Kerwin


By Arun Iyer on Wednesday, December 26, 2001 - 07:09 pm:

gee thanks i got it now...

love arun