Surface area - Double integrals

10 replies · 3K views · last reply
etf
Messages
179
Reaction score
2
Hi!
Here is my task:
Calculate surface area of sphere $$x^{2}+y^{2}+z^{2}=16$$ between $$z=2$$ and $$z=-2\sqrt{3}$$.
Here are 3D graphs of our surfaces:

3D plot of a sphere with a blue horizontal plane at z = 2 cutting off a cap labeled P1, and a green plane at z = -2√3 below; red annotations give x² + y² + z² = 16 and the plane equations.


3D plot of a sphere cut by two planes: a dark blue plane above a horizontal dark green plane, with a cap of the sphere between them, labelled P3 (upper band), P2 (cap below the green plane) and axes x, y, z.


Surface area of interest is P3. It would be P-(P1+P2), where P is surface area of whole sphere. Is it correct?
Here is how I calculated P1:

Handwritten calculus notes on graph paper: sphere equation x^2+y^2+z^2=16 solved for upper and lower hemispheres, a circle of radius sqrt(12) in the xy-plane, partial derivatives of z, and surface integral setups for P1 with polar substitution leading to 8pi(2-sqrt3).


Similarly for P2:

Handwritten calculus working for the surface area of a sphere: z = -sqrt(16 - (x^2+y^2)), a substitution for z = -2 sqrt(3), partial derivatives dz/dx and dz/dy, and a double integral setup for P2 over p and phi with result 2 pi h sqrt(16)/9.


We calculate P using formula 4pi*r*r.
Is it correct?
 
Can't really decipher your answer. It looks like 2πscribble.
 
P1=8*pi*ln2, P2=2*pi*ln(16/9), P=4*pi*16
P=4*pi*16-(8*pi*ln2+2*pi*ln(16/9))=201.0619-21.0358=180.0261
 
etf said:
P1=8*pi*ln2, P2=2*pi*ln(16/9), P=4*pi*16
P=4*pi*16-(8*pi*ln2+2*pi*ln(16/9))=201.0619-21.0358=180.0261
You can check your result using the formula in this article:

http://en.wikipedia.org/wiki/Spherical_segment

It seems there is a disagreement somewhere in your calculations between the formula and your result.
 
etf, have you given any consideration to working the problem in spherical coordinates? You can calculate it directly in about two steps.
 
Doing multiple integrals and not being familiar with integral transforms is not a good thing. Polar, spherical and cylindrical co-ordinates are important to be able to integrate properly, and without them, you won't get very far.

An interesting way to do this problem is to describe each face of the surface and then perform a surface integral over each one:

$$\iint_S \space dS = \iint_{S_1} \space dS_1 + \iint_{S_2} \space dS_2 + \iint_{S_3} \space dS_3$$

So you have a flat top surface ##S_1## described by the plane ##z = 2## and bounded by the region ##D_1## given by ##x^2 + y^2 = 12##. You also have a flat bottom surface ##S_2## describe by the plane ##z = -2 \sqrt{3}## and bounded by the region ##D_2##. Finally, you have the spherical middle section bounded by the planes.
 
Zondrina said:
Doing multiple integrals and not being familiar with integral transforms is not a good thing. Polar, spherical and cylindrical co-ordinates are important to be able to integrate properly, and without them, you won't get very far.

An interesting way to do this problem is to describe each face of the surface and then perform a surface integral over each one:

$$\iint_S \space dS = \iint_{S_1} \space dS_1 + \iint_{S_2} \space dS_2 + \iint_{S_3} \space dS_3$$

So you have a flat top surface ##S_1## described by the plane ##z = 2## and bounded by the region ##D_1## given by ##x^2 + y^2 = 12##. You also have a flat bottom surface ##S_2## describe by the plane ##z = -2 \sqrt{3}## and bounded by the region ##D_2##. Finally, you have the spherical middle section bounded by the planes.
I think the OP was supposed to calculate only the area of the spherical surface lying between the planes z = -2√3 and z = 2. I don't think the areas of S1 or S2 were ever intended to be included.
 
SteamKing said:
I think the OP was supposed to calculate only the area of the spherical surface lying between the planes z = -2√3 and z = 2. I don't think the areas of S1 or S2 were ever intended to be included.

Sorry, I saw the problem a little differently when I first read it this morning. The sphere with its ends cut off seems to be more appropriate, so the ##dS_3## integral would be the only thing required. I believe the conventional method of evaluation then proceeds as usual:

$$\iint_S \space dS = \iint_D \sqrt{z_x^2 + z_y^2 + 1} \space dA$$

The two planes and sphere give two circles in the x-y plane from which the limits can be deduced.
 
Zondrina said:
Sorry, I saw the problem a little differently when I first read it this morning. The sphere with its ends cut off seems to be more appropriate, so the ##dS_3## integral would be the only thing required. I believe the conventional method of evaluation then proceeds as usual:

$$\iint_S \space dS = \iint_D \sqrt{z_x^2 + z_y^2 + 1} \space dA$$

The two planes and sphere give two circles in the x-y plane from which the limits can be deduced.

That is certainly not the "natural" method to use. The surface is not symmetric vertically so you would have to work the top and bottom portions separately. And rectangular coordinates also are not the natural choice.