Unit 4, lesson 28

Triple Integrals in Rectangular Coordinates

How a triple integral adds up density times volume over a solid, how to set up its bounds from a projection and vertical segments, and how it gives volume, mass, centers of mass and averages.

  • Theory part 55
  • Practice part 56
  • Triple Integrals

Picture a block of cheese whose density changes from one corner to the other. To weigh it on paper, cut it into tiny cubes, multiply each cube’s density by its volume, and add. A triple integral is the limit of that bookkeeping, and the hard part in practice is not the adding but describing the solid so the bounds come out right.

Key terms

Triple Riemann sums

Definition Let ff be defined on a closed box B=[a,b]×[c,d]×[r,s]B = [a, b] \times [c, d] \times [r, s]. Divide [a,b][a, b], [c,d][c, d] and [r,s][r, s] into ll, mm and nn subintervals of equal lengths Δx\Delta x, Δy\Delta y and Δz\Delta z. This cuts BB into lmnlmn subboxes BijkB_{ijk}, each of volume ΔV=Δx Δy Δz\Delta V = \Delta x \, \Delta y \, \Delta z. Choose a sample point Pijk∗=(xijk∗,yijk∗,zijk∗)P_{ijk}^* = (x_{ijk}^*, y_{ijk}^*, z_{ijk}^*) in each BijkB_{ijk}. The triple Riemann sum is

∑i=1l∑j=1m∑k=1nf(Pijk∗) ΔV.\sum_{i=1}^{l} \sum_{j=1}^{m} \sum_{k=1}^{n} f(P_{ijk}^*) \, \Delta V.

The triple integral of ff over BB is

∭Bf(x,y,z) dV=lim⁡l,m,n→∞∑i=1l∑j=1m∑k=1nf(Pijk∗) ΔV\begin{aligned} &\iiint_B f(x, y, z) \, dV \\[0.4em] &= \lim_{l, m, n \to \infty} \sum_{i=1}^{l} \sum_{j=1}^{m} \sum_{k=1}^{n} f(P_{ijk}^*) \, \Delta V \end{aligned}

if this limit exists and is the same for every choice of sample points. Then ff is called integrable on BB. Every continuous function on BB is integrable. For a bounded solid EE, enclose EE in a box BB and set F=fF = f on EE and F=0F = 0 on the rest of BB. Then ∭Ef dV\iiint_E f \, dV is defined as ∭BF dV\iiint_B F \, dV when that integral exists. This integral does not depend on the box chosen, and it exists in particular when EE is closed and bounded, ff is continuous on EE, and the boundary of EE consists of finitely many pieces, each the graph of a continuous function over a closed bounded region in one of the coordinate planes. Such a boundary has zero volume, so the jump of FF across it does no harm. All the solids below are of this kind.

In plain words Chop the solid into tiny boxes, multiply the value of ff in each box by the box’s volume, and add. The integral is what these sums settle on as the boxes shrink.

Example Take f(x,y,z)=z2f(x, y, z) = z^2 on B=[0,1]×[0,1]×[0,2]B = [0, 1] \times [0, 1] \times [0, 2] and cut each edge in half, giving 8 subboxes with ΔV=12⋅12⋅1=14\Delta V = \frac{1}{2} \cdot \frac{1}{2} \cdot 1 = \frac{1}{4}. At the midpoints, four boxes have z=12z = \frac{1}{2} and four have z=32z = \frac{3}{2}, so the sum is

4⋅14⋅14+4⋅94⋅14=52.4 \cdot \frac{1}{4} \cdot \frac{1}{4} + 4 \cdot \frac{9}{4} \cdot \frac{1}{4} = \frac{5}{2}.

The exact value, computed in the next section, is 83≈2.667\frac{8}{3} \approx 2.667, so even this coarse sum is close.

Iterated integrals and Fubini’s theorem

Definition Fubini’s theorem for boxes says that if ff is continuous on B=[a,b]×[c,d]×[r,s]B = [a, b] \times [c, d] \times [r, s], then

∭Bf dV=∫rs∫cd∫abf(x,y,z) dx dy dz,\begin{aligned} &\iiint_B f \, dV \\[0.4em] &= \int_r^s \int_c^d \int_a^b f(x, y, z) \, dx \, dy \, dz, \end{aligned}

and the same holds for each of the six orders of integration. The iterated integral is evaluated from the inside out, holding the outer variables fixed while integrating the inner one.

In plain words On a box you can integrate one variable at a time, in any order you like, and the answer does not change.

Example For the Riemann sum example above,

∫02∫01∫01z2 dx dy dz=∫02z2 dz=83.\begin{aligned} \int_0^2 \int_0^1 \int_0^1 z^2 \, dx \, dy \, dz &= \int_0^2 z^2 \, dz \\[0.4em] &= \frac{8}{3}. \end{aligned}

A second one, over B=[0,1]×[−1,2]×[0,3]B = [0, 1] \times [-1, 2] \times [0, 3]. Integrating xx, then yy, then zz gives

∭Bxyz2 dV=∫03∫−12yz22 dy dz=∫033z24 dz=274.\begin{aligned} \iiint_B xyz^2 \, dV &= \int_0^3 \int_{-1}^{2} \frac{y z^2}{2} \, dy \, dz \\[0.4em] &= \int_0^3 \frac{3 z^2}{4} \, dz = \frac{27}{4}. \end{aligned}

Region descriptions

Definition A solid EE is of type 1 if

E={(x,y,z)∣(x,y)∈D,u1(x,y)≤z≤u2(x,y)},\begin{gathered} E = \{(x, y, z) \mid (x, y) \in D, \\ u_1(x, y) \le z \le u_2(x, y)\}, \end{gathered}

where DD is the projection of EE onto the xyxy-plane, DD is a closed bounded region of type I or type II, and u1≤u2u_1 \le u_2 are continuous on DD. If ff is continuous on EE, then

∭Ef dV=∬D[∫u1(x,y)u2(x,y)f dz]dA.\begin{aligned} &\iiint_E f \, dV \\[0.4em] &= \iint_D \left[ \int_{u_1(x, y)}^{u_2(x, y)} f \, dz \right] dA. \end{aligned}

Types 2 and 3 are defined the same way with xx or yy as the inner variable, projecting onto the yzyz-plane or the xzxz-plane.

In plain words Look at the shadow of the solid on a coordinate plane. Above each point of the shadow, a segment runs from a bottom surface to a top surface. The inner integral covers that segment, and a double integral covers the shadow.

Example Let EE be the tetrahedron bounded by the coordinate planes and the plane x+y+z=1x + y + z = 1. Its shadow on the xyxy-plane is the triangle 0≤x≤10 \le x \le 1, 0≤y≤1−x0 \le y \le 1 - x, and above each shadow point zz runs from 00 to 1−x−y1 - x - y. Then

∭Ez dV=∫01∫01−x∫01−x−yz dz dy dx=∫01∫01−x(1−x−y)22 dy dx=∫01(1−x)36 dx=124.\begin{aligned} &\iiint_E z \, dV \\[0.4em] &= \int_0^1 \int_0^{1 - x} \int_0^{1 - x - y} z \, dz \, dy \, dx \\[0.4em] &= \int_0^1 \int_0^{1 - x} \frac{(1 - x - y)^2}{2} \, dy \, dx \\[0.4em] &= \int_0^1 \frac{(1 - x)^3}{6} \, dx = \frac{1}{24}. \end{aligned}

Changing the order of integration

Definition For ff continuous on a solid EE that can be described in more than one of the three types, every valid description gives the same value of ∭Ef dV\iiint_E f \, dV. To change the order, first write EE as a set of inequalities, then read off a new projection and new inner bounds from those same inequalities. The new limits must describe exactly the same set of points.

In plain words Never shuffle the limits around directly. Go back to the solid, sketch or list its inequalities, and rebuild the bounds for the new order from scratch.

Example The integral

∫01∫0x2∫0yf dz dy dx\int_0^1 \int_0^{x^2} \int_0^y f \, dz \, dy \, dx

covers the solid 0≤x≤10 \le x \le 1, 0≤y≤x20 \le y \le x^2, 0≤z≤y0 \le z \le y. For the order dy dz dxdy \, dz \, dx, fix xx. The inequalities 0≤z≤y≤x20 \le z \le y \le x^2 say that zz runs from 00 to x2x^2 and then yy from zz to x2x^2, so the integral equals

∫01∫0x2∫zx2f dy dz dx.\int_0^1 \int_0^{x^2} \int_z^{x^2} f \, dy \, dz \, dx.

For the order dx dy dzdx \, dy \, dz, the same inequalities read 0≤z≤y≤10 \le z \le y \le 1 and y≤x≤1\sqrt{y} \le x \le 1, so the integral equals

∫01∫z1∫y1f dx dy dz.\int_0^1 \int_z^1 \int_{\sqrt{y}}^1 f \, dx \, dy \, dz.

With f=1f = 1 all three give 110\frac{1}{10}, a good check that nothing was lost.

Volume

Definition The volume of a solid EE for which the integral exists is

V(E)=∭E1 dV.V(E) = \iiint_E 1 \, dV.

In plain words Integrating the constant 11 just adds up the volumes of the tiny boxes.

Example The tetrahedron cut from the first octant by 2x+y+z=42x + y + z = 4 has intercepts 22, 44 and 44. Its shadow is 0≤x≤20 \le x \le 2, 0≤y≤4−2x0 \le y \le 4 - 2x, and zz runs from 00 to 4−2x−y4 - 2x - y, so

V=∫02∫04−2x(4−2x−y) dy dx=∫02(4−2x)22 dx=163.\begin{aligned} V &= \int_0^2 \int_0^{4 - 2x} (4 - 2x - y) \, dy \, dx \\[0.4em] &= \int_0^2 \frac{(4 - 2x)^2}{2} \, dx = \frac{16}{3}. \end{aligned}

This agrees with the formula one sixth times the product of the intercepts, 16⋅2⋅4⋅4=163\frac{1}{6} \cdot 2 \cdot 4 \cdot 4 = \frac{16}{3}.

Mass

Definition If a solid EE has continuous density δ(x,y,z)≥0\delta(x, y, z) \ge 0, measured in mass per unit volume, its mass is

m=∭Eδ(x,y,z) dV.m = \iiint_E \delta(x, y, z) \, dV.

In plain words Each tiny box has mass about density times volume. Add them all.

Example Give the tetrahedron bounded by the coordinate planes and x+y+z=1x + y + z = 1 the density δ=y\delta = y, so it gets heavier away from the xzxz-plane. Then

m=∫01∫01−xy(1−x−y) dy dx=∫01(1−x)36 dx=124.\begin{aligned} m &= \int_0^1 \int_0^{1 - x} y(1 - x - y) \, dy \, dx \\[0.4em] &= \int_0^1 \frac{(1 - x)^3}{6} \, dx = \frac{1}{24}. \end{aligned}

Moments and center of mass

Definition For a solid EE with continuous density δ\delta and mass m>0m > 0, the first moments about the three coordinate planes are

Myz=∭Ex δ dV,Mxz=∭Ey δ dV,Mxy=∭Ez δ dV,\begin{gathered} M_{yz} = \iiint_E x \, \delta \, dV, \\[0.4em] M_{xz} = \iiint_E y \, \delta \, dV, \\[0.4em] M_{xy} = \iiint_E z \, \delta \, dV, \end{gathered}

and the center of mass is

(xˉ,yˉ,zˉ)=(Myzm,Mxzm,Mxym).(\bar{x}, \bar{y}, \bar{z}) = \left( \frac{M_{yz}}{m}, \frac{M_{xz}}{m}, \frac{M_{xy}}{m} \right).

The moment of inertia about the zz-axis is Iz=∭E(x2+y2) δ dVI_z = \iiint_E (x^2 + y^2) \, \delta \, dV, and IxI_x, IyI_y are defined the same way with the squared distance to that axis.

In plain words The center of mass is the balance point. Each coordinate is a weighted average of that coordinate over the solid, with the density as the weight.

Example For the tetrahedron with δ=y\delta = y from the mass example, the same bounds give

Myz=1120,Mxz=160,Mxy=1120.\begin{gathered} M_{yz} = \frac{1}{120}, \qquad M_{xz} = \frac{1}{60}, \\[0.4em] M_{xy} = \frac{1}{120}. \end{gathered}

Dividing by m=124m = \frac{1}{24} gives the center of mass (15,25,15)\left( \frac{1}{5}, \frac{2}{5}, \frac{1}{5} \right). With constant density the center would be at (14,14,14)\left( \frac{1}{4}, \frac{1}{4}, \frac{1}{4} \right). The heavier side pulls the balance point toward larger yy and away from the other two directions, as you would expect.

Average value

Definition If ff is integrable on a solid EE with V(E)>0V(E) > 0, the average value of ff on EE is

favg=1V(E)∭Ef(x,y,z) dV.f_{\text{avg}} = \frac{1}{V(E)} \iiint_E f(x, y, z) \, dV.

In plain words Total up ff over the solid and divide by the volume, just as an ordinary average divides a total by a count.

Example On the tetrahedron bounded by the coordinate planes and x+y+z=1x + y + z = 1, which has volume 16\frac{1}{6}, take f=x2+y2+z2f = x^2 + y^2 + z^2, the squared distance from the origin. Each of the three terms integrates to 160\frac{1}{60}, so

favg=6⋅360=310.f_{\text{avg}} = 6 \cdot \frac{3}{60} = \frac{3}{10}.

Check yourself

  1. Evaluate the integral of x+y+zx + y + z over the box [0,1]×[0,2]×[0,3][0, 1] \times [0, 2] \times [0, 3].
Show the answer

The box has volume 66. Integrating term by term gives 3+6+93 + 6 + 9, so

∭B(x+y+z) dV=18.\iiint_B (x + y + z) \, dV = 18.

For example, the xx term is 12⋅2⋅3=3\frac{1}{2} \cdot 2 \cdot 3 = 3.

  1. Rewrite the integral below in the order dx dz dydx \, dz \, dy.
∫01∫01−x∫01−x−yf dz dy dx\int_0^1 \int_0^{1 - x} \int_0^{1 - x - y} f \, dz \, dy \, dx
Show the answer

The solid is x,y,z≥0x, y, z \ge 0 with x+y+z≤1x + y + z \le 1. Its shadow on the yzyz-plane is 0≤y≤10 \le y \le 1, 0≤z≤1−y0 \le z \le 1 - y, and for each such point xx runs from 00 to 1−y−z1 - y - z, so the integral is

∫01∫01−y∫01−y−zf dx dz dy.\int_0^1 \int_0^{1 - y} \int_0^{1 - y - z} f \, dx \, dz \, dy.
  1. Find the center of mass of the cube [0,1]3[0, 1]^3 with density δ=z\delta = z.
Show the answer

The mass is m=∫01z dz=12m = \int_0^1 z \, dz = \frac{1}{2} and Mxy=∫01z2 dz=13M_{xy} = \int_0^1 z^2 \, dz = \frac{1}{3}, so zˉ=23\bar{z} = \frac{2}{3}. The density does not depend on xx or yy and the cube is symmetric about x=12x = \frac{1}{2} and y=12y = \frac{1}{2}, so the center of mass is (12,12,23)\left( \frac{1}{2}, \frac{1}{2}, \frac{2}{3} \right).

How this lesson connects

It also uses Double integrals and iterated bounds and Three-dimensional coordinate geometry and density.