We will let the cylinder be centered at the origin and the axis aligned in the y direction. We will drill into the origin from the positive z-direction.
Note first that at a point (x, y) in xy-plane, the cylinder runs from a z-height of -√(r² - x²) to √(r² - x²), so the height of material removed is 2√(r² - x²). The drill removes all the material directly above or below the disk D = {(x, y): x² + y² ≤ r²}. So the volume of material removed is:
[D]∬2√(r² - x²) dA
Solving this integral, let us find the limits of integration. They are -r≤y≤r, -√(r² - y²)≤x≤√(r² - y²). Thus we have:
[-r, r]∫[-√(r² - y²), √(r² - y²)]∫2√(r² - x²) dx dy
To make some computations regarding sign easier, we will use the fact that [-√(r² - y²), √(r² - y²)]∫2√(r² - x²) dx = 2 [0, √(r² - y²)]∫2√(r² - x²) dx, since 2√(r² - x²) is an even function. Similarly, 2 [0, √(r² - y²)]∫2√(r² - x²) dx is an even function, so by symmetry, [-r, r]∫2 [0, √(r² - y²)]∫2√(r² - x²) dx dy = 2 [0, r]∫2 [0, √(r² - y²)]∫2√(r² - x²) dx dy. Thus we have:
2 [0, r]∫2 [0, √(r² - y²)]∫2√(r² - x²) dx dy
Collecting the constant factors
8 [0, r]∫[0, √(r² - y²)]∫√(r² - x²) dx dy
Extracting a factor of r:
8r [0, r]∫[0, √(r² - y²)]∫√(1 - (x/r)²) dx dy
So now make the substitution θ = arcsin (x/r), so sin θ = x/r, x = r sin θ, and dx = r cos θ dθ. Now to find the new limits of integration, note that when x=0, θ=0, and when x=√(r² - y²), we have sin θ = x/r = √(1 - (y/r)²), so sin² θ = 1 - (y/r)², thus (y/r)² = 1 - sin² θ = cos² θ, thus cos θ = ±y/r. But since the range of arcsin is [-π/2, π/2], cos θ is positive, and since y∈[0, r] is positive, we in fact have cos θ = y/r. So then θ = ±arccos (y/r). However, since x is positive, arcsin (x/r) is positive, and as arccos (y/r) is positive, we in fact have θ = arccos (y/r). So we have:
8r [0, r]∫[0, arccos (y/r)]∫√(1 - sin² θ) r cos θ dθ dy
Simplifying this a bit:
8r² [0, r]∫[0, arccos (y/r)]∫√(cos² θ) cos θ dθ dy
8r² [0, r]∫[0, arccos (y/r)]∫cos² θ dθ dy (since cos θ is positive)
Now, using the formula cos² θ = (cos (2θ) + 1)/2:
4r² [0, r]∫[0, arccos (y/r)]∫cos (2θ) + 1 dθ dy
Integrating:
4r² [0, r]∫1/2 sin (2θ) + θ |[0, arccos (y/r)] dy
4r² [0, r]∫1/2 sin (2 arccos (y/r)) + arccos (y/r) dy
Simplifying a bit:
4r² [0, r]∫sin (arccos (y/r)) cos (arccos (y/r)) + arccos (y/r) dy
4r² [0, r]∫sin (arccos (y/r)) y/r + arccos (y/r) dy
4r² [0, r]∫√(1 - cos² (arccos (y/r))) y/r + arccos (y/r) dy (since arccos (y/r) is positive, so is sin (arccos (y/r)), thus the sign is correct)
4r² [0, r]∫√(1 - (y/r)²) y/r + arccos (y/r) dy
Breaking this into two integrals:
4r² [0, r]∫√(1 - (y/r)²) y/r dy + 4r² [0, r]∫arccos (y/r) dy
Simplifying the first:
2 [0, r]∫√(r² - y²) 2y dy + 4r² [0, r]∫arccos (y/r) dy
Making the substitution u=r² - y², we have that du = -2y dy, and that when y=0, u=² and when y=r, u=0. So we have:
-2 [r², 0]∫√u du + 4r² [0, r]∫arccos (y/r) dy
Integrating:
-4/3 u^(3/2)|[r², 0] + 4r² [0, r]∫arccos (y/r) dy
4/3 r³ + 4r² [0, r]∫arccos (y/r) dy
Now, turning our attention to the second integral, we proceed by parts. Let u=arccos (y/r) and dv = dy, du = -1/√(1 - (y/r)²) * 1/r dy and v=y. So we have:
4/3 r³ + 4r² (y arccos (y/r)|[0, r] + [0, r]∫y/√(1 - (y/r)²) 1/r dy)
Simplifying:
4/3 r³ + 4r² (r arccos (1) + [0, r]∫y/√(r² - y) dy)
4/3 r³ + 4r² [0, r]∫y/√(r² - y²) dy (since arccos (1) = 0)
Making the substitution u = r² - y², we have (again) that du = -2y dy, y=0 ⇒ u=r², and y=r ⇒ u=0, so we have:
4/3 r³ - 2r² [r², 0]∫1/√u du
Integrating:
4/3 r³ - 4r²√u|[r², 0]
4/3 r³ + 4r³
16/3 r³
And we are done.
Edit: corrected arithmetic error near the end. Searching for other possible errors -- give me a minute.
Edit 2: Found a second error, and corrected it.
Edit 3: Okay, this derivation is now correct. Regarding the answers by mr_zone_v and kendizzle -- the shape of the removed material is neither a cylinder nor a sphere. Rather, it is the intersection of two cylinders, which is known as a Steinmetz solid. Thus any volume calculations based on the assumption that it is a sphere or cylinder will be quite wrong.
Edit 4: Wow steve -- that's brilliant! Wish I'd thought of that -- it would have saved me a lot of time.
2007-12-07 08:36:54
·
answer #2
·
answered by Pascal 7
·
3⤊
0⤋