Unit C8: Surface Integrals, Stokes’, and the Divergence Theorem第 C8 单元:曲面积分(surface integral)、斯托克斯定理(Stokes' theorem)与散度定理(divergence theorem)
Curl and divergence, parametric surfaces, scalar and flux surface integrals, and the two capstone theorems of vector calculus.旋度(curl)与散度(divergence)、参数曲面、标量曲面积分与通量(flux)积分,以及向量微积分的两大压轴定理。
curl)与散度(divergence),再学习如何在曲面上积分,既包括标量函数(用于面积与质量),也包括向量场(用于通量)。最终成果是把格林定理(Green's theorem)推广到空间的斯托克斯定理(Stokes' theorem),以及把闭曲面通量化为三重积分的散度定理(divergence theorem)。最后一节展示四个边界定理如何统一为一个模式,让你一眼就能选对工具。
Curl and Divergence旋度(curl)与散度(divergence)
vector field)$\mathbf{F} = \langle P, Q, R \rangle$ 上有两种方式。旋度 $\nabla \times \mathbf{F}$ 是一个向量,刻画场的局部旋转;散度 $\nabla \cdot \mathbf{F}$ 是一个标量,刻画单位体积的局部流出量。这两个算子正是斯托克斯定理与散度定理的微分原料。Two identities are used constantly. First, the curl of a gradient is zero: $\nabla \times (\nabla f) = \mathbf{0}$, because mixed partials commute. Hence a field with nonzero curl cannot be conservative. Second, the divergence of a curl is zero: $\nabla \cdot (\nabla \times \mathbf{F}) = 0$. Both follow from equality of mixed second partials.有两个恒等式会不断用到。第一,梯度的旋度为零:$\nabla \times (\nabla f) = \mathbf{0}$,因为混合偏导可交换。因此旋度非零的场不可能是保守场(conservative field)。第二,旋度的散度为零:$\nabla \cdot (\nabla \times \mathbf{F}) = 0$。两者都源于二阶混合偏导相等。
It is worth keeping the geometric meaning in view while you compute. The divergence at a point is the rate at which an infinitesimal amount of fluid is being created (positive divergence) or destroyed (negative divergence) there, measured per unit volume; a divergence-free field is incompressible. The curl points along the axis about which the field swirls fastest, and its length is twice the local angular velocity of a tiny paddle wheel placed in the flow. A constant field such as $\mathbf{F} = \langle 1, 0, 0 \rangle$ has zero divergence and zero curl: nothing is created and nothing rotates. A rigid rotation $\mathbf{F} = \langle -y, x, 0 \rangle$ has zero divergence but constant curl $\langle 0, 0, 2 \rangle$, which is exactly twice the angular velocity of the rotation.计算时不妨始终把几何意义放在心里。某点的散度是无穷小流体在该处被产生(正散度)或被消灭(负散度)的速率,按单位体积度量;散度为零的场是不可压缩的。旋度指向场旋转最快的轴,其长度等于放入流中的微小桨轮局部角速度的两倍。常向量场如 $\mathbf{F} = \langle 1, 0, 0 \rangle$ 的散度与旋度均为零:既无产生,也无旋转。刚体旋转 $\mathbf{F} = \langle -y, x, 0 \rangle$ 散度为零,但有常旋度 $\langle 0, 0, 2 \rangle$,恰好是旋转角速度的两倍。
scalar),所以 $\nabla \times (\nabla \cdot \mathbf{F})$ 毫无意义,因为旋度需要向量输入。反之 $\nabla \times \mathbf{F}$ 是含三个分量的向量。第二个常见失误是旋度中间分量的符号错误:行列式展开给出的是 $P_z - R_x$,而非 $R_x - P_z$。每次都写出符号行列式,而不要凭记忆默写三个分量。Worked Example 1.1: curl and divergence of a field例题 1.1:场的旋度与散度
Let $\mathbf{F} = \langle xz,\; xyz,\; -y^2 \rangle$. Compute the curl and the divergence.
For the curl, use the determinant formula with $P = xz$, $Q = xyz$, $R = -y^2$:
$$ \nabla \times \mathbf{F} = \left\langle R_y - Q_z,\; P_z - R_x,\; Q_x - P_y \right\rangle = \left\langle -2y - xy,\; x - 0,\; yz - 0 \right\rangle. $$So $\nabla \times \mathbf{F} = \langle -2y - xy,\; x,\; yz \rangle$. For the divergence,
$$ \nabla \cdot \mathbf{F} = \frac{\partial}{\partial x}(xz) + \frac{\partial}{\partial y}(xyz) + \frac{\partial}{\partial z}(-y^2) = z + xz + 0 = z(1 + x). $$设 $\mathbf{F} = \langle xz,\; xyz,\; -y^2 \rangle$。求旋度与散度。
求旋度,用行列式公式,取 $P = xz$,$Q = xyz$,$R = -y^2$:
$$ \nabla \times \mathbf{F} = \left\langle R_y - Q_z,\; P_z - R_x,\; Q_x - P_y \right\rangle = \left\langle -2y - xy,\; x - 0,\; yz - 0 \right\rangle. $$故 $\nabla \times \mathbf{F} = \langle -2y - xy,\; x,\; yz \rangle$。求散度:
$$ \nabla \cdot \mathbf{F} = \frac{\partial}{\partial x}(xz) + \frac{\partial}{\partial y}(xyz) + \frac{\partial}{\partial z}(-y^2) = z + xz + 0 = z(1 + x). $$Worked Example 1.2: testing a field for a potential例题 1.2:检验场是否有势函数
Decide whether $\mathbf{F} = \langle 2xy + z^2,\; x^2,\; 2xz \rangle$ can be conservative, and if so find a potential.
Because the curl of a gradient is zero, a necessary condition for $\mathbf{F} = \nabla f$ is $\nabla \times \mathbf{F} = \mathbf{0}$. Compute with $P = 2xy + z^2$, $Q = x^2$, $R = 2xz$:
$$ \nabla \times \mathbf{F} = \langle R_y - Q_z,\; P_z - R_x,\; Q_x - P_y \rangle = \langle 0 - 0,\; 2z - 2z,\; 2x - 2x \rangle = \mathbf{0}. $$The curl vanishes, and the domain is all of space, which is simply connected, so $\mathbf{F}$ is conservative. To recover $f$, integrate $P$ in $x$: $f = x^2 y + xz^2 + h(y, z)$. Then $f_y = x^2 + h_y$ must equal $Q = x^2$, so $h_y = 0$ and $h = k(z)$. Finally $f_z = 2xz + k'(z)$ must equal $R = 2xz$, so $k'(z) = 0$. A potential is $f(x,y,z) = x^2 y + x z^2$.
判断 $\mathbf{F} = \langle 2xy + z^2,\; x^2,\; 2xz \rangle$ 是否可能为保守场,若是则求出一个势函数(potential function)。
由于梯度的旋度为零,$\mathbf{F} = \nabla f$ 的必要条件是 $\nabla \times \mathbf{F} = \mathbf{0}$。取 $P = 2xy + z^2$,$Q = x^2$,$R = 2xz$ 计算:
$$ \nabla \times \mathbf{F} = \langle R_y - Q_z,\; P_z - R_x,\; Q_x - P_y \rangle = \langle 0 - 0,\; 2z - 2z,\; 2x - 2x \rangle = \mathbf{0}. $$旋度为零,且定义域为整个空间,是单连通的,故 $\mathbf{F}$ 是保守场。为求 $f$,将 $P$ 对 $x$ 积分:$f = x^2 y + xz^2 + h(y, z)$。则 $f_y = x^2 + h_y$ 须等于 $Q = x^2$,故 $h_y = 0$,$h = k(z)$。最后 $f_z = 2xz + k'(z)$ 须等于 $R = 2xz$,故 $k'(z) = 0$。一个势函数为 $f(x,y,z) = x^2 y + x z^2$。
Worked Example 1.3: a field that swirls but does not spread例题 1.3:只旋转不扩散的场
For $\mathbf{F} = \langle -y,\; x,\; 0 \rangle$, compute both operators and interpret them.
Divergence: $\nabla \cdot \mathbf{F} = \partial_x(-y) + \partial_y(x) + \partial_z(0) = 0$. The field neither creates nor destroys fluid; the flow is incompressible. Curl: with $P = -y$, $Q = x$, $R = 0$,
$$ \nabla \times \mathbf{F} = \langle 0 - 0,\; 0 - 0,\; 1 - (-1) \rangle = \langle 0, 0, 2 \rangle. $$The curl is a constant vector pointing along the positive $z$-axis with length $2$. A paddle wheel dropped anywhere in this field spins about the $z$-axis with angular velocity $1$, since the curl is twice the angular velocity. This is the prototype rotational field, and its nonzero curl confirms at a glance that it cannot be a gradient.
对 $\mathbf{F} = \langle -y,\; x,\; 0 \rangle$,计算两个算子并加以解释。
散度:$\nabla \cdot \mathbf{F} = \partial_x(-y) + \partial_y(x) + \partial_z(0) = 0$。该场既不产生也不消灭流体,流动是不可压缩的。旋度:取 $P = -y$,$Q = x$,$R = 0$,
$$ \nabla \times \mathbf{F} = \langle 0 - 0,\; 0 - 0,\; 1 - (-1) \rangle = \langle 0, 0, 2 \rangle. $$旋度是沿 $z$ 轴正向、长度为 $2$ 的常向量。在此场中任意位置放入桨轮,它都绕 $z$ 轴以角速度 $1$ 旋转,因为旋度是角速度的两倍。这是旋转场的原型,其非零旋度一眼便确认它不可能是梯度。
Going deeper: why the curl of a gradient vanishes深入探究:为何梯度的旋度恒为零
Let $f$ have continuous second partials. The third component of $\nabla \times (\nabla f)$ is
$$ \frac{\partial}{\partial x}\!\left(\frac{\partial f}{\partial y}\right) - \frac{\partial}{\partial y}\!\left(\frac{\partial f}{\partial x}\right) = f_{yx} - f_{xy}. $$By Clairaut's Theorem the mixed partials are equal, so this component is zero. The same cancellation occurs in the other two components, giving $\nabla \times (\nabla f) = \mathbf{0}$. This is the differential form of the statement that conservative fields are irrotational.
The companion identity $\nabla \cdot (\nabla \times \mathbf{F}) = 0$ is proved the same way. Write $\nabla \times \mathbf{F} = \langle R_y - Q_z,\; P_z - R_x,\; Q_x - P_y \rangle$ and take its divergence:
$$ \nabla \cdot (\nabla \times \mathbf{F}) = (R_y - Q_z)_x + (P_z - R_x)_y + (Q_x - P_y)_z. $$Expanding gives six second-order terms, $R_{yx} - Q_{zx} + P_{zy} - R_{xy} + Q_{xz} - P_{yz}$. Provided $\mathbf{F}$ has continuous second partials, Clairaut pairs them: $R_{yx} = R_{xy}$, $Q_{zx} = Q_{xz}$, and $P_{zy} = P_{yz}$. Every term cancels against its mate, so the sum is identically zero. Geometrically this says a curl field is always source-free, which is why the curl flux of a field through a closed surface is automatically zero, a fact we will see again in the Divergence Theorem.
设 $f$ 有连续二阶偏导。$\nabla \times (\nabla f)$ 的第三个分量为
$$ \frac{\partial}{\partial x}\!\left(\frac{\partial f}{\partial y}\right) - \frac{\partial}{\partial y}\!\left(\frac{\partial f}{\partial x}\right) = f_{yx} - f_{xy}. $$由克莱罗定理(Clairaut's Theorem),混合偏导相等,故此分量为零。其余两个分量发生同样的抵消,于是 $\nabla \times (\nabla f) = \mathbf{0}$。这是"保守场无旋"这一论断的微分形式。
配套恒等式 $\nabla \cdot (\nabla \times \mathbf{F}) = 0$ 以同样方式证明。写出 $\nabla \times \mathbf{F} = \langle R_y - Q_z,\; P_z - R_x,\; Q_x - P_y \rangle$ 并取其散度:
$$ \nabla \cdot (\nabla \times \mathbf{F}) = (R_y - Q_z)_x + (P_z - R_x)_y + (Q_x - P_y)_z. $$展开得六个二阶项 $R_{yx} - Q_{zx} + P_{zy} - R_{xy} + Q_{xz} - P_{yz}$。只要 $\mathbf{F}$ 有连续二阶偏导,克莱罗定理将它们两两配对:$R_{yx} = R_{xy}$,$Q_{zx} = Q_{xz}$,$P_{zy} = P_{yz}$。每一项都与其配对项抵消,故和恒为零。几何上这说明旋度场总是无源的,这正是场穿过闭曲面的旋度通量自动为零的原因,我们在散度定理中会再次见到这一事实。
Parametric Surfaces and Areas参数曲面(parametrized surface)与面积
normal vector)。其长度记录参数化如何拉伸面积,这正是曲面积分所需。For a graph $z = g(x,y)$ over a region $D$, take $\mathbf{r}(x,y) = \langle x, y, g(x,y) \rangle$. Then $\mathbf{r}_x \times \mathbf{r}_y = \langle -g_x, -g_y, 1 \rangle$, so the area element specializes to the familiar form below.对区域 $D$ 上的图像 $z = g(x,y)$,取 $\mathbf{r}(x,y) = \langle x, y, g(x,y) \rangle$。则 $\mathbf{r}_x \times \mathbf{r}_y = \langle -g_x, -g_y, 1 \rangle$,于是面积元素退化为下面熟悉的形式。
Two parametrizations recur so often they are worth memorizing. The sphere of radius $a$ uses spherical angles, $\mathbf{r}(\phi, \theta) = \langle a \sin\phi \cos\theta,\; a\sin\phi\sin\theta,\; a\cos\phi \rangle$ with $0 \le \phi \le \pi$ and $0 \le \theta \le 2\pi$, and there $\left|\mathbf{r}_\phi \times \mathbf{r}_\theta\right| = a^2 \sin\phi$, so $dS = a^2 \sin\phi\, d\phi\, d\theta$. The cylinder of radius $a$ uses $\mathbf{r}(\theta, z) = \langle a\cos\theta,\; a\sin\theta,\; z\rangle$, for which $\left|\mathbf{r}_\theta \times \mathbf{r}_z\right| = a$, giving $dS = a\, d\theta\, dz$. Recognizing the right parameters at the start turns most surface integrals into routine double integrals.有两种参数化反复出现,值得记牢。半径为 $a$ 的球面用球面角,$\mathbf{r}(\phi, \theta) = \langle a \sin\phi \cos\theta,\; a\sin\phi\sin\theta,\; a\cos\phi \rangle$,其中 $0 \le \phi \le \pi$,$0 \le \theta \le 2\pi$,此时 $\left|\mathbf{r}_\phi \times \mathbf{r}_\theta\right| = a^2 \sin\phi$,故 $dS = a^2 \sin\phi\, d\phi\, d\theta$。半径为 $a$ 的柱面用 $\mathbf{r}(\theta, z) = \langle a\cos\theta,\; a\sin\theta,\; z\rangle$,此时 $\left|\mathbf{r}_\theta \times \mathbf{r}_z\right| = a$,给出 $dS = a\, d\theta\, dz$。一开始就认出正确的参数,能把大多数曲面积分化为常规的二重积分。
Worked Example 2.1: area of a portion of a paraboloid例题 2.1:抛物面一部分的面积
Find the area of the part of the paraboloid $z = x^2 + y^2$ lying below the plane $z = 4$.
With $g(x,y) = x^2 + y^2$ we have $g_x = 2x$, $g_y = 2y$, so
$$ A = \iint_D \sqrt{1 + 4x^2 + 4y^2}\; dA. $$The region $D$ is the disk $x^2 + y^2 \le 4$. In polar coordinates the integrand becomes $\sqrt{1 + 4r^2}$ and $dA = r\, dr\, d\theta$:
$$ A = \int_0^{2\pi}\!\!\int_0^2 \sqrt{1 + 4r^2}\; r\, dr\, d\theta = 2\pi \cdot \frac{1}{12}\Big[(1 + 4r^2)^{3/2}\Big]_0^2 = \frac{\pi}{6}\left(17^{3/2} - 1\right). $$求抛物面 $z = x^2 + y^2$ 位于平面 $z = 4$ 下方部分的面积。
取 $g(x,y) = x^2 + y^2$,则 $g_x = 2x$,$g_y = 2y$,故
$$ A = \iint_D \sqrt{1 + 4x^2 + 4y^2}\; dA. $$区域 $D$ 是圆盘 $x^2 + y^2 \le 4$。在极坐标下被积函数变为 $\sqrt{1 + 4r^2}$,$dA = r\, dr\, d\theta$:
$$ A = \int_0^{2\pi}\!\!\int_0^2 \sqrt{1 + 4r^2}\; r\, dr\, d\theta = 2\pi \cdot \frac{1}{12}\Big[(1 + 4r^2)^{3/2}\Big]_0^2 = \frac{\pi}{6}\left(17^{3/2} - 1\right). $$Worked Example 2.2: surface area of a sphere from its parametrization例题 2.2:由参数化求球面的面积
Recover the classical area $4\pi a^2$ of a sphere of radius $a$ from the parametrization $\mathbf{r}(\phi,\theta) = \langle a\sin\phi\cos\theta,\; a\sin\phi\sin\theta,\; a\cos\phi\rangle$.
The two tangent vectors are
$$ \mathbf{r}_\phi = \langle a\cos\phi\cos\theta,\; a\cos\phi\sin\theta,\; -a\sin\phi\rangle, \qquad \mathbf{r}_\theta = \langle -a\sin\phi\sin\theta,\; a\sin\phi\cos\theta,\; 0\rangle. $$Their cross product has length $\left|\mathbf{r}_\phi \times \mathbf{r}_\theta\right| = a^2 \sin\phi$ (a standard computation; the $z$-component is $a^2\sin\phi\cos\phi$ and the horizontal part contributes $a^2\sin^2\phi$, summing under the root to $a^2\sin\phi$ since $\sin\phi \ge 0$ on $[0,\pi]$). Therefore
$$ A = \int_0^{2\pi}\!\!\int_0^{\pi} a^2 \sin\phi\, d\phi\, d\theta = a^2 (2\pi)\big[-\cos\phi\big]_0^{\pi} = a^2 (2\pi)(2) = 4\pi a^2. $$由参数化 $\mathbf{r}(\phi,\theta) = \langle a\sin\phi\cos\theta,\; a\sin\phi\sin\theta,\; a\cos\phi\rangle$ 重新得出半径为 $a$ 的球面的经典面积 $4\pi a^2$。
两个切向量为
$$ \mathbf{r}_\phi = \langle a\cos\phi\cos\theta,\; a\cos\phi\sin\theta,\; -a\sin\phi\rangle, \qquad \mathbf{r}_\theta = \langle -a\sin\phi\sin\theta,\; a\sin\phi\cos\theta,\; 0\rangle. $$其叉积长度为 $\left|\mathbf{r}_\phi \times \mathbf{r}_\theta\right| = a^2 \sin\phi$(标准计算;$z$ 分量为 $a^2\sin\phi\cos\phi$,水平部分贡献 $a^2\sin^2\phi$,在根号下相加得 $a^2\sin\phi$,因为在 $[0,\pi]$ 上 $\sin\phi \ge 0$)。因此
$$ A = \int_0^{2\pi}\!\!\int_0^{\pi} a^2 \sin\phi\, d\phi\, d\theta = a^2 (2\pi)\big[-\cos\phi\big]_0^{\pi} = a^2 (2\pi)(2) = 4\pi a^2. $$Worked Example 2.3: area of a cylinder cut by a plane例题 2.3:被平面截切的柱面面积
Find the area of the part of the cylinder $x^2 + y^2 = 1$ between $z = 0$ and the slanted plane $z = 2 + x$.
Parametrize by $\mathbf{r}(\theta, z) = \langle \cos\theta, \sin\theta, z\rangle$, so $dS = d\theta\, dz$. On the cylinder $x = \cos\theta$, so the surface runs from $z = 0$ up to $z = 2 + \cos\theta$, which is positive for every $\theta$. Hence
$$ A = \int_0^{2\pi}\!\!\int_0^{2 + \cos\theta} dz\, d\theta = \int_0^{2\pi} (2 + \cos\theta)\, d\theta = 2(2\pi) + 0 = 4\pi. $$The $\cos\theta$ term integrates to zero over a full revolution, so the slanted top contributes the same area as a flat cut at the average height $z = 2$.
求柱面 $x^2 + y^2 = 1$ 介于 $z = 0$ 与斜平面 $z = 2 + x$ 之间部分的面积。
用 $\mathbf{r}(\theta, z) = \langle \cos\theta, \sin\theta, z\rangle$ 参数化,故 $dS = d\theta\, dz$。在柱面上 $x = \cos\theta$,于是曲面从 $z = 0$ 延伸到 $z = 2 + \cos\theta$,后者对每个 $\theta$ 都为正。因此
$$ A = \int_0^{2\pi}\!\!\int_0^{2 + \cos\theta} dz\, d\theta = \int_0^{2\pi} (2 + \cos\theta)\, d\theta = 2(2\pi) + 0 = 4\pi. $$$\cos\theta$ 项在一整圈上积分为零,故斜顶贡献的面积与在平均高度 $z = 2$ 处的平切相同。
Going deeper: deriving the graph area element from the cross product深入探究:由叉积导出图像面积元素
For the graph parametrization $\mathbf{r}(x,y) = \langle x, y, g(x,y)\rangle$, the tangent vectors are $\mathbf{r}_x = \langle 1, 0, g_x\rangle$ and $\mathbf{r}_y = \langle 0, 1, g_y\rangle$. Their cross product is
$$ \mathbf{r}_x \times \mathbf{r}_y = \begin{vmatrix} \mathbf{i} & \mathbf{j} & \mathbf{k} \\ 1 & 0 & g_x \\ 0 & 1 & g_y \end{vmatrix} = \langle -g_x,\; -g_y,\; 1\rangle. $$Its length is $\sqrt{g_x^2 + g_y^2 + 1}$, which is exactly the area-scaling factor. The geometric content is that a unit square in the $xy$-plane is tilted into a parallelogram on the surface whose area exceeds the flat square by precisely the secant of the tilt angle. Since the surface normal makes angle $\gamma$ with the vertical and $\cos\gamma = 1/\sqrt{1 + g_x^2 + g_y^2}$, the area element is $dA / \cos\gamma = \sqrt{1 + g_x^2 + g_y^2}\, dA$, confirming the formula from pure geometry.
对图像参数化 $\mathbf{r}(x,y) = \langle x, y, g(x,y)\rangle$,切向量为 $\mathbf{r}_x = \langle 1, 0, g_x\rangle$ 与 $\mathbf{r}_y = \langle 0, 1, g_y\rangle$。其叉积为
$$ \mathbf{r}_x \times \mathbf{r}_y = \begin{vmatrix} \mathbf{i} & \mathbf{j} & \mathbf{k} \\ 1 & 0 & g_x \\ 0 & 1 & g_y \end{vmatrix} = \langle -g_x,\; -g_y,\; 1\rangle. $$其长度为 $\sqrt{g_x^2 + g_y^2 + 1}$,恰是面积缩放因子。几何含义是:$xy$ 平面上的单位正方形被倾斜成曲面上的平行四边形,其面积比平正方形大出恰好倾角的正割。由于曲面法向量与竖直方向成角 $\gamma$ 且 $\cos\gamma = 1/\sqrt{1 + g_x^2 + g_y^2}$,故面积元素为 $dA / \cos\gamma = \sqrt{1 + g_x^2 + g_y^2}\, dA$,从纯几何角度证实了该公式。
Surface Integrals of Scalar Functions标量函数的曲面积分(surface integral)
For a graph $z = g(x,y)$ this reads $\iint_S f\, dS = \iint_D f(x, y, g(x,y))\, \sqrt{1 + g_x^2 + g_y^2}\, dA$. The orientation of $S$ is irrelevant here, because $dS$ is a positive scalar element of area. Beyond mass, scalar surface integrals deliver the centroid of a shell, its moment of inertia, and the average value of any quantity over the surface, in every case by inserting the appropriate weight $f$ and dividing by the area when an average is wanted.对图像 $z = g(x,y)$,公式写作 $\iint_S f\, dS = \iint_D f(x, y, g(x,y))\, \sqrt{1 + g_x^2 + g_y^2}\, dA$。此处 $S$ 的定向(orientation)无关紧要,因为 $dS$ 是正的标量面积元素。除质量外,标量曲面积分还能给出薄壳的形心、转动惯量,以及任意量在曲面上的平均值,每次都是代入合适的权函数 $f$,需要平均值时再除以面积。
Worked Example 3.1: mass of a conical shell例题 3.1:锥形薄壳的质量
A thin shell is the cone $z = \sqrt{x^2 + y^2}$ for $0 \le z \le 1$, with surface density $\rho(x,y,z) = z$. Find its mass $m = \iint_S z\, dS$.
With $g = \sqrt{x^2 + y^2}$ we get $g_x = x/\sqrt{x^2+y^2}$ and $g_y = y/\sqrt{x^2+y^2}$, so $g_x^2 + g_y^2 = 1$ and $\sqrt{1 + g_x^2 + g_y^2} = \sqrt{2}$. On the cone $z = \sqrt{x^2 + y^2} = r$ in polar coordinates, and $D$ is the disk $r \le 1$:
$$ m = \iint_D \sqrt{x^2 + y^2}\, \sqrt{2}\; dA = \sqrt{2}\int_0^{2\pi}\!\!\int_0^1 r \cdot r\, dr\, d\theta = \sqrt{2}\,(2\pi)\,\frac{1}{3} = \frac{2\sqrt{2}\,\pi}{3}. $$一个薄壳是圆锥 $z = \sqrt{x^2 + y^2}$($0 \le z \le 1$),面密度为 $\rho(x,y,z) = z$。求其质量 $m = \iint_S z\, dS$。
取 $g = \sqrt{x^2 + y^2}$,得 $g_x = x/\sqrt{x^2+y^2}$,$g_y = y/\sqrt{x^2+y^2}$,故 $g_x^2 + g_y^2 = 1$,$\sqrt{1 + g_x^2 + g_y^2} = \sqrt{2}$。在锥面上,极坐标下 $z = \sqrt{x^2 + y^2} = r$,且 $D$ 是圆盘 $r \le 1$:
$$ m = \iint_D \sqrt{x^2 + y^2}\, \sqrt{2}\; dA = \sqrt{2}\int_0^{2\pi}\!\!\int_0^1 r \cdot r\, dr\, d\theta = \sqrt{2}\,(2\pi)\,\frac{1}{3} = \frac{2\sqrt{2}\,\pi}{3}. $$Worked Example 3.2: integrating over a sphere例题 3.2:在球面上积分
Evaluate $\iint_S z^2\, dS$ where $S$ is the unit sphere $x^2 + y^2 + z^2 = 1$.
Use the spherical parametrization, on which $z = \cos\phi$ and $dS = \sin\phi\, d\phi\, d\theta$ since $a = 1$. Then
$$ \iint_S z^2\, dS = \int_0^{2\pi}\!\!\int_0^{\pi} \cos^2\phi \, \sin\phi\, d\phi\, d\theta = 2\pi \left[-\frac{\cos^3\phi}{3}\right]_0^{\pi} = 2\pi \cdot \frac{2}{3} = \frac{4\pi}{3}. $$As a sanity check, by symmetry $\iint_S x^2\, dS = \iint_S y^2\, dS = \iint_S z^2\, dS$, and their sum is $\iint_S (x^2 + y^2 + z^2)\, dS = \iint_S 1\, dS = 4\pi$. Dividing by three gives $4\pi/3$, matching the direct computation.
求 $\iint_S z^2\, dS$,其中 $S$ 是单位球面 $x^2 + y^2 + z^2 = 1$。
用球面参数化,其上 $z = \cos\phi$,且由于 $a = 1$,$dS = \sin\phi\, d\phi\, d\theta$。则
$$ \iint_S z^2\, dS = \int_0^{2\pi}\!\!\int_0^{\pi} \cos^2\phi \, \sin\phi\, d\phi\, d\theta = 2\pi \left[-\frac{\cos^3\phi}{3}\right]_0^{\pi} = 2\pi \cdot \frac{2}{3} = \frac{4\pi}{3}. $$作为合理性验算,由对称性 $\iint_S x^2\, dS = \iint_S y^2\, dS = \iint_S z^2\, dS$,三者之和为 $\iint_S (x^2 + y^2 + z^2)\, dS = \iint_S 1\, dS = 4\pi$。除以三得 $4\pi/3$,与直接计算一致。
Worked Example 3.3: average height of a hemispherical dome例题 3.3:半球穹顶的平均高度
Find the average value of the height $z$ over the upper unit hemisphere $S$.
The average is $\bar z = \dfrac{1}{A(S)} \iint_S z\, dS$. The hemisphere has area $A(S) = 2\pi$. With the spherical parametrization restricted to $0 \le \phi \le \pi/2$, $z = \cos\phi$ and $dS = \sin\phi\, d\phi\, d\theta$:
$$ \iint_S z\, dS = \int_0^{2\pi}\!\!\int_0^{\pi/2} \cos\phi \sin\phi\, d\phi\, d\theta = 2\pi \left[\frac{\sin^2\phi}{2}\right]_0^{\pi/2} = 2\pi \cdot \frac{1}{2} = \pi. $$Therefore $\bar z = \pi / (2\pi) = \tfrac{1}{2}$. The average height of a unit dome is exactly half its radius, a clean result that students can quote as a check on harder problems.
求高度 $z$ 在上半单位半球面 $S$ 上的平均值。
平均值为 $\bar z = \dfrac{1}{A(S)} \iint_S z\, dS$。半球面面积 $A(S) = 2\pi$。把球面参数化限制在 $0 \le \phi \le \pi/2$,有 $z = \cos\phi$,$dS = \sin\phi\, d\phi\, d\theta$:
$$ \iint_S z\, dS = \int_0^{2\pi}\!\!\int_0^{\pi/2} \cos\phi \sin\phi\, d\phi\, d\theta = 2\pi \left[\frac{\sin^2\phi}{2}\right]_0^{\pi/2} = 2\pi \cdot \frac{1}{2} = \pi. $$因此 $\bar z = \pi / (2\pi) = \tfrac{1}{2}$。单位穹顶的平均高度恰为其半径的一半,是一个干净的结论,学生可在更难的题目中引用作为验算。
Going deeper: why the scalar integral is independent of the parametrization深入探究:为何标量积分与参数化无关
Suppose $\tilde{\mathbf{r}}(s,t) = \mathbf{r}(u(s,t), v(s,t))$ is a smooth reparametrization with nonzero Jacobian $J = \partial(u,v)/\partial(s,t)$. By the chain rule the tangent vectors transform so that $\tilde{\mathbf{r}}_s \times \tilde{\mathbf{r}}_t = J\,(\mathbf{r}_u \times \mathbf{r}_v)$, hence $\left|\tilde{\mathbf{r}}_s \times \tilde{\mathbf{r}}_t\right| = |J|\,\left|\mathbf{r}_u \times \mathbf{r}_v\right|$. Substituting into the integral and applying the change-of-variables theorem for double integrals, the factor $|J|$ is exactly absorbed by $ds\, dt = dA / |J|$ relative to $du\, dv$. The two integrals are equal:
$$ \iint_{\tilde D} f\,\left|\tilde{\mathbf{r}}_s \times \tilde{\mathbf{r}}_t\right|\, ds\, dt = \iint_D f\,\left|\mathbf{r}_u \times \mathbf{r}_v\right|\, du\, dv. $$So $\iint_S f\, dS$ is a genuine property of the surface and the function, not of how we happened to chart it. The absolute value is essential here, and it is precisely why scalar surface integrals are blind to orientation, in contrast to the flux integrals of the next section.
设 $\tilde{\mathbf{r}}(s,t) = \mathbf{r}(u(s,t), v(s,t))$ 是一个雅可比行列式 $J = \partial(u,v)/\partial(s,t)$ 非零的光滑重参数化。由链式法则,切向量按 $\tilde{\mathbf{r}}_s \times \tilde{\mathbf{r}}_t = J\,(\mathbf{r}_u \times \mathbf{r}_v)$ 变换,故 $\left|\tilde{\mathbf{r}}_s \times \tilde{\mathbf{r}}_t\right| = |J|\,\left|\mathbf{r}_u \times \mathbf{r}_v\right|$。代入积分并对二重积分应用换元定理,因子 $|J|$ 恰好被相对于 $du\, dv$ 的 $ds\, dt = dA / |J|$ 吸收。两个积分相等:
$$ \iint_{\tilde D} f\,\left|\tilde{\mathbf{r}}_s \times \tilde{\mathbf{r}}_t\right|\, ds\, dt = \iint_D f\,\left|\mathbf{r}_u \times \mathbf{r}_v\right|\, du\, dv. $$所以 $\iint_S f\, dS$ 是曲面与函数本身的真实属性,与我们恰好如何为它建图无关。这里绝对值至关重要,它正是标量曲面积分对定向不敏感的原因,与下一节的通量积分形成对照。
Flux Integrals通量(flux)积分
oriented surface)$S$ 的通量,度量场沿所选单位法向量(normal vector)$\mathbf{n}$ 方向穿过 $S$ 的净速率。定向至关重要:反转 $\mathbf{n}$ 会使通量变号。通量就是法向分量 $\mathbf{F} \cdot \mathbf{n}$ 的曲面积分。The vector $\mathbf{r}_u \times \mathbf{r}_v$ already carries both the direction of the normal and the area scaling, so no separate normalization is needed. For a graph oriented upward, $d\mathbf{S} = \langle -g_x, -g_y, 1 \rangle\, dA$. For a closed surface the convention is the outward normal.向量 $\mathbf{r}_u \times \mathbf{r}_v$ 已经同时携带了法向方向与面积缩放,无需另行单位化。对向上定向的图像,$d\mathbf{S} = \langle -g_x, -g_y, 1 \rangle\, dA$。对闭曲面(closed surface),约定取外法向(outward normal)。
Worked Example 4.1: flux through a hemisphere例题 4.1:穿过半球面的通量
Find the upward flux of $\mathbf{F} = \langle 0, 0, z \rangle$ through the upper unit hemisphere $z = \sqrt{1 - x^2 - y^2}$.
Here $g = \sqrt{1 - x^2 - y^2}$, $g_x = -x/g$, $g_y = -y/g$, so the upward normal element is $d\mathbf{S} = \langle x/g,\; y/g,\; 1 \rangle\, dA$. On the surface $\mathbf{F} = \langle 0, 0, g \rangle$, hence $\mathbf{F} \cdot d\mathbf{S} = g\, dA$ over the unit disk:
$$ \iint_S \mathbf{F}\cdot d\mathbf{S} = \iint_{r \le 1} \sqrt{1 - r^2}\; r\, dr\, d\theta = 2\pi \int_0^1 r\sqrt{1 - r^2}\, dr = 2\pi \cdot \frac{1}{3} = \frac{2\pi}{3}. $$求 $\mathbf{F} = \langle 0, 0, z \rangle$ 穿过上半单位半球面 $z = \sqrt{1 - x^2 - y^2}$ 的向上通量。
此处 $g = \sqrt{1 - x^2 - y^2}$,$g_x = -x/g$,$g_y = -y/g$,故向上的法向元素为 $d\mathbf{S} = \langle x/g,\; y/g,\; 1 \rangle\, dA$。在曲面上 $\mathbf{F} = \langle 0, 0, g \rangle$,因此在单位圆盘上 $\mathbf{F} \cdot d\mathbf{S} = g\, dA$:
$$ \iint_S \mathbf{F}\cdot d\mathbf{S} = \iint_{r \le 1} \sqrt{1 - r^2}\; r\, dr\, d\theta = 2\pi \int_0^1 r\sqrt{1 - r^2}\, dr = 2\pi \cdot \frac{1}{3} = \frac{2\pi}{3}. $$Worked Example 4.2: flux of a radial field through a plane patch例题 4.2:径向场穿过平面块的通量
Find the upward flux of $\mathbf{F} = \langle x, y, z\rangle$ through the part of the plane $z = 1 - x - y$ lying over the triangle $D$ with vertices $(0,0)$, $(1,0)$, $(0,1)$.
With $g = 1 - x - y$ we have $g_x = g_y = -1$, so the upward element is $d\mathbf{S} = \langle 1, 1, 1\rangle\, dA$. On the surface $z = 1 - x - y$, so
$$ \mathbf{F}\cdot d\mathbf{S} = \big(x + y + (1 - x - y)\big)\, dA = 1\, dA. $$The integrand collapses to the constant $1$, so the flux is just the area of the triangle $D$, which is $\tfrac{1}{2}$. Therefore $\iint_S \mathbf{F}\cdot d\mathbf{S} = \tfrac{1}{2}$. Recognizing such collapses saves a great deal of integration.
求 $\mathbf{F} = \langle x, y, z\rangle$ 穿过平面 $z = 1 - x - y$ 位于以 $(0,0)$、$(1,0)$、$(0,1)$ 为顶点的三角形 $D$ 上方那部分的向上通量。
取 $g = 1 - x - y$,则 $g_x = g_y = -1$,故向上的元素为 $d\mathbf{S} = \langle 1, 1, 1\rangle\, dA$。在曲面上 $z = 1 - x - y$,于是
$$ \mathbf{F}\cdot d\mathbf{S} = \big(x + y + (1 - x - y)\big)\, dA = 1\, dA. $$被积函数坍缩为常数 $1$,故通量就是三角形 $D$ 的面积 $\tfrac{1}{2}$。因此 $\iint_S \mathbf{F}\cdot d\mathbf{S} = \tfrac{1}{2}$。认出这类坍缩能省下大量积分。
Worked Example 4.3: flux through a closed cylinder, surface by surface例题 4.3:逐面计算穿过闭柱面的通量
Find the outward flux of $\mathbf{F} = \langle x, y, 0\rangle$ through the closed cylinder $x^2 + y^2 = 1$, $0 \le z \le 2$, including its top and bottom disks.
The closed surface has three pieces. On the top ($z = 2$) and bottom ($z = 0$) the outward normals are $\pm\langle 0, 0, 1\rangle$, but $\mathbf{F}$ has no $z$-component, so $\mathbf{F}\cdot\mathbf{n} = 0$ and both caps contribute zero flux. On the curved side, parametrize by $\mathbf{r}(\theta,z) = \langle\cos\theta,\sin\theta,z\rangle$; the outward normal element is $\mathbf{r}_\theta \times \mathbf{r}_z = \langle\cos\theta,\sin\theta,0\rangle\, d\theta\, dz$, and on the surface $\mathbf{F} = \langle\cos\theta,\sin\theta,0\rangle$, so $\mathbf{F}\cdot d\mathbf{S} = (\cos^2\theta + \sin^2\theta)\, d\theta\, dz = d\theta\, dz$. Hence
$$ \iint_S \mathbf{F}\cdot d\mathbf{S} = \int_0^2\!\!\int_0^{2\pi} d\theta\, dz = 2\pi \cdot 2 = 4\pi. $$As a check, $\nabla\cdot\mathbf{F} = 2$ and the solid has volume $2\pi$, so the Divergence Theorem predicts $2 \cdot 2\pi = 4\pi$, in agreement.
求 $\mathbf{F} = \langle x, y, 0\rangle$ 穿过闭柱面 $x^2 + y^2 = 1$($0 \le z \le 2$,含上下两个圆盘)的外向通量。
闭曲面有三块。在顶面($z = 2$)和底面($z = 0$)上,外法向为 $\pm\langle 0, 0, 1\rangle$,但 $\mathbf{F}$ 无 $z$ 分量,故 $\mathbf{F}\cdot\mathbf{n} = 0$,两个端盖贡献的通量均为零。在侧曲面上,用 $\mathbf{r}(\theta,z) = \langle\cos\theta,\sin\theta,z\rangle$ 参数化;外法向元素为 $\mathbf{r}_\theta \times \mathbf{r}_z = \langle\cos\theta,\sin\theta,0\rangle\, d\theta\, dz$,在曲面上 $\mathbf{F} = \langle\cos\theta,\sin\theta,0\rangle$,故 $\mathbf{F}\cdot d\mathbf{S} = (\cos^2\theta + \sin^2\theta)\, d\theta\, dz = d\theta\, dz$。因此
$$ \iint_S \mathbf{F}\cdot d\mathbf{S} = \int_0^2\!\!\int_0^{2\pi} d\theta\, dz = 2\pi \cdot 2 = 4\pi. $$作为验算,$\nabla\cdot\mathbf{F} = 2$,立体体积为 $2\pi$,故散度定理预测 $2 \cdot 2\pi = 4\pi$,一致。
Going deeper: flux is independent of the parametrization but not the orientation深入探究:通量与参数化无关,但与定向有关
If two parametrizations $\mathbf{r}(u,v)$ and $\tilde{\mathbf{r}}(s,t)$ trace the same surface, the change of variables theorem shows that $\iint_D \mathbf{F} \cdot (\mathbf{r}_u \times \mathbf{r}_v)\, dA$ has the same magnitude under both, because the Jacobian of the reparametrization rescales the cross product exactly as it rescales $dA$. What can change is the sign: if the new parametrization induces the opposite normal direction, the Jacobian is negative and the flux changes sign. This is why every flux statement must fix an orientation of $S$.
若两个参数化 $\mathbf{r}(u,v)$ 与 $\tilde{\mathbf{r}}(s,t)$ 描出同一曲面,换元定理表明 $\iint_D \mathbf{F} \cdot (\mathbf{r}_u \times \mathbf{r}_v)\, dA$ 在两者下大小相同,因为重参数化的雅可比行列式对叉积的缩放恰与它对 $dA$ 的缩放相同。可能变的是符号:若新参数化诱导出相反的法向方向,雅可比行列式为负,通量便变号。这正是每个通量陈述都必须为 $S$ 固定一个定向的原因。
Stokes' Theorem斯托克斯定理(Stokes' theorem)
Green's theorem)在三维中的推广。边界(boundary)$\partial S$ 的定向与 $S$ 的法向必须满足右手定则:拇指沿 $\mathbf{n}$ 指向时,四指卷曲方向即 $\partial S$ 的正方向。A powerful consequence is that the surface flux of a curl depends only on the boundary curve, not on the surface filling it. So if two surfaces share the same boundary, the curl flux through them is equal. This often lets you replace a complicated surface by a flat disk.一个有力的推论是:旋度的曲面通量只依赖于边界曲线,而与填充它的曲面无关。于是若两个曲面共享同一边界,穿过它们的旋度通量相等。这常让你用一个平圆盘替换掉复杂的曲面。
Worked Example 5.1: verifying Stokes on a disk例题 5.1:在圆盘上验证斯托克斯定理
Let $\mathbf{F} = \langle -y, x, z \rangle$ and let $S$ be the disk $x^2 + y^2 \le 1$ in the plane $z = 0$, oriented upward, with boundary the unit circle $C$ traversed counterclockwise.
Circulation side: parametrize $C$ by $\mathbf{r}(t) = \langle \cos t, \sin t, 0 \rangle$, so $\mathbf{r}'(t) = \langle -\sin t, \cos t, 0 \rangle$ and $\mathbf{F} = \langle -\sin t, \cos t, 0 \rangle$. Then
$$ \oint_C \mathbf{F}\cdot d\mathbf{r} = \int_0^{2\pi} (\sin^2 t + \cos^2 t)\, dt = 2\pi. $$Flux side: $\nabla \times \mathbf{F} = \langle 0, 0, 2 \rangle$, and the upward normal element on the flat disk is $d\mathbf{S} = \langle 0, 0, 1\rangle\, dA$, so
$$ \iint_S (\nabla \times \mathbf{F})\cdot d\mathbf{S} = \iint_{r \le 1} 2\, dA = 2 \cdot \pi = 2\pi. $$The two sides agree, as Stokes' Theorem requires.
设 $\mathbf{F} = \langle -y, x, z \rangle$,$S$ 为平面 $z = 0$ 内的圆盘 $x^2 + y^2 \le 1$,向上定向,其边界为逆时针绕行的单位圆 $C$。
环量一侧:用 $\mathbf{r}(t) = \langle \cos t, \sin t, 0 \rangle$ 参数化 $C$,则 $\mathbf{r}'(t) = \langle -\sin t, \cos t, 0 \rangle$,$\mathbf{F} = \langle -\sin t, \cos t, 0 \rangle$。于是
$$ \oint_C \mathbf{F}\cdot d\mathbf{r} = \int_0^{2\pi} (\sin^2 t + \cos^2 t)\, dt = 2\pi. $$通量一侧:$\nabla \times \mathbf{F} = \langle 0, 0, 2 \rangle$,平圆盘上向上的法向元素为 $d\mathbf{S} = \langle 0, 0, 1\rangle\, dA$,故
$$ \iint_S (\nabla \times \mathbf{F})\cdot d\mathbf{S} = \iint_{r \le 1} 2\, dA = 2 \cdot \pi = 2\pi. $$两侧相等,正如斯托克斯定理所要求。
Worked Example 5.2: replacing a hard surface by an easy one例题 5.2:用简单曲面替换困难曲面
Compute $\iint_S (\nabla\times\mathbf{F})\cdot d\mathbf{S}$ for $\mathbf{F} = \langle -y, x, xyz\rangle$, where $S$ is the upper hemisphere $x^2 + y^2 + z^2 = 1$, $z \ge 0$, oriented upward.
Computing the curl flux directly over the curved hemisphere is unpleasant. But $S$ has the same boundary as the flat disk $D: x^2 + y^2 \le 1$ in the plane $z = 0$, namely the unit circle $C$ traversed counterclockwise. By Stokes, both curl fluxes equal $\oint_C \mathbf{F}\cdot d\mathbf{r}$, so we may integrate over the disk instead. On $z = 0$ the field reduces to $\langle -y, x, 0\rangle$, and with $\mathbf{r}(t) = \langle\cos t, \sin t, 0\rangle$,
$$ \oint_C \mathbf{F}\cdot d\mathbf{r} = \int_0^{2\pi} \big(\sin^2 t + \cos^2 t\big)\, dt = 2\pi. $$So the hemisphere curl flux is $2\pi$, obtained without ever parametrizing the hemisphere. The $xyz$ term in $\mathbf{F}$ vanishes on the boundary and therefore never enters the answer.
计算 $\iint_S (\nabla\times\mathbf{F})\cdot d\mathbf{S}$,其中 $\mathbf{F} = \langle -y, x, xyz\rangle$,$S$ 为上半球面 $x^2 + y^2 + z^2 = 1$($z \ge 0$),向上定向。
在弯曲的半球面上直接计算旋度通量很麻烦。但 $S$ 与平面 $z = 0$ 内的平圆盘 $D: x^2 + y^2 \le 1$ 共享同一边界,即逆时针绕行的单位圆 $C$。由斯托克斯定理,两个旋度通量都等于 $\oint_C \mathbf{F}\cdot d\mathbf{r}$,故可改在圆盘上积分。在 $z = 0$ 上场化为 $\langle -y, x, 0\rangle$,取 $\mathbf{r}(t) = \langle\cos t, \sin t, 0\rangle$,
$$ \oint_C \mathbf{F}\cdot d\mathbf{r} = \int_0^{2\pi} \big(\sin^2 t + \cos^2 t\big)\, dt = 2\pi. $$所以半球面旋度通量为 $2\pi$,全程无需参数化半球面。$\mathbf{F}$ 中的 $xyz$ 项在边界上为零,故根本不进入答案。
Worked Example 5.3: Stokes on a triangle in a slanted plane例题 5.3:斜平面内三角形上的斯托克斯定理
Let $\mathbf{F} = \langle z, x, y\rangle$ and let $C$ be the triangle with vertices $(1,0,0)$, $(0,1,0)$, $(0,0,1)$, oriented counterclockwise as seen from the first octant. Find $\oint_C \mathbf{F}\cdot d\mathbf{r}$.
Rather than parametrize three line segments, use Stokes with $S$ the flat triangle in the plane $x + y + z = 1$. The curl is $\nabla\times\mathbf{F} = \langle 1 - 0,\; 1 - 0,\; 1 - 0\rangle = \langle 1, 1, 1\rangle$. The plane has upward unit normal $\mathbf{n} = \tfrac{1}{\sqrt 3}\langle 1, 1, 1\rangle$, so $(\nabla\times\mathbf{F})\cdot\mathbf{n} = 3/\sqrt 3 = \sqrt 3$, a constant. Therefore
$$ \oint_C \mathbf{F}\cdot d\mathbf{r} = \iint_S (\nabla\times\mathbf{F})\cdot d\mathbf{S} = \sqrt 3 \cdot A(S). $$The triangle has side length $\sqrt 2$ and is equilateral, so its area is $A(S) = \tfrac{\sqrt 3}{4}(\sqrt 2)^2 = \tfrac{\sqrt 3}{2}$. Hence the circulation is $\sqrt 3 \cdot \tfrac{\sqrt 3}{2} = \tfrac{3}{2}$.
设 $\mathbf{F} = \langle z, x, y\rangle$,$C$ 为以 $(1,0,0)$、$(0,1,0)$、$(0,0,1)$ 为顶点的三角形,从第一卦限看为逆时针定向。求 $\oint_C \mathbf{F}\cdot d\mathbf{r}$。
与其参数化三条线段,不如用斯托克斯定理,取 $S$ 为平面 $x + y + z = 1$ 内的平三角形。旋度为 $\nabla\times\mathbf{F} = \langle 1 - 0,\; 1 - 0,\; 1 - 0\rangle = \langle 1, 1, 1\rangle$。该平面的向上单位法向为 $\mathbf{n} = \tfrac{1}{\sqrt 3}\langle 1, 1, 1\rangle$,故 $(\nabla\times\mathbf{F})\cdot\mathbf{n} = 3/\sqrt 3 = \sqrt 3$,是常数。因此
$$ \oint_C \mathbf{F}\cdot d\mathbf{r} = \iint_S (\nabla\times\mathbf{F})\cdot d\mathbf{S} = \sqrt 3 \cdot A(S). $$该三角形边长为 $\sqrt 2$ 且为等边三角形,故面积为 $A(S) = \tfrac{\sqrt 3}{4}(\sqrt 2)^2 = \tfrac{\sqrt 3}{2}$。于是环量为 $\sqrt 3 \cdot \tfrac{\sqrt 3}{2} = \tfrac{3}{2}$。
Going deeper: why Stokes' Theorem holds, via Green's Theorem深入探究:借助格林定理说明斯托克斯定理为何成立
The proof reduces the spatial statement to the planar Green's Theorem. Take a surface that is a graph $z = g(x,y)$ over a region $D$, oriented upward, with boundary curve $C$ projecting to the boundary $\partial D$. Parametrize the circulation by pushing the planar boundary up onto the surface. Writing $\mathbf{F} = \langle P, Q, R\rangle$ and substituting $z = g(x,y)$, the line integral $\oint_C \mathbf{F}\cdot d\mathbf{r}$ becomes a planar line integral over $\partial D$ with integrand involving $P$, $Q$, and the chain-rule terms $R\, g_x$ and $R\, g_y$.
Applying Green's Theorem to that planar integral converts it to a double integral over $D$ of a combination of partial derivatives. Collecting terms, that combination is exactly $(\nabla\times\mathbf{F})\cdot\langle -g_x, -g_y, 1\rangle$, which is the integrand of the curl flux through the graph. Hence $\oint_C \mathbf{F}\cdot d\mathbf{r} = \iint_S (\nabla\times\mathbf{F})\cdot d\mathbf{S}$ for a graph. A general orientable surface is cut into graph-like patches whose internal boundaries cancel in pairs, leaving only the outer boundary, which extends the result to the surfaces met in practice.
证明把空间陈述归约为平面的格林定理。取一个区域 $D$ 上的图像曲面 $z = g(x,y)$,向上定向,其边界曲线 $C$ 投影到边界 $\partial D$。把平面边界提升到曲面上来参数化环量。写 $\mathbf{F} = \langle P, Q, R\rangle$ 并代入 $z = g(x,y)$,线积分(line integral)$\oint_C \mathbf{F}\cdot d\mathbf{r}$ 变为 $\partial D$ 上的平面线积分,被积量含 $P$、$Q$ 以及链式法则项 $R\, g_x$ 与 $R\, g_y$。
对该平面积分应用格林定理,将其化为 $D$ 上若干偏导数组合的二重积分。整理各项,该组合恰为 $(\nabla\times\mathbf{F})\cdot\langle -g_x, -g_y, 1\rangle$,正是穿过图像的旋度通量的被积量。于是对图像有 $\oint_C \mathbf{F}\cdot d\mathbf{r} = \iint_S (\nabla\times\mathbf{F})\cdot d\mathbf{S}$。一般可定向曲面被切成若干图像状的小块,其内部边界两两抵消,只剩外边界,这就把结论推广到实际遇到的曲面。
The Divergence Theorem散度定理(divergence theorem)
triple integral),后者往往简单得多。曲面必须是闭的、向外定向的,并且必须界出立体区域 $E$。Intuitively, the divergence is the local source strength of the field. Summing these sources over the solid gives the net amount of the field leaving through the bounding surface, the total flux. Interior flux cancels between adjacent cells, leaving only the boundary contribution.直观上,散度是场的局部源强度。把这些源在立体上求和,便得到场穿过界面流出的净量,即总通量。相邻小格之间内部通量相互抵消,只剩边界的贡献。
Worked Example 6.1: flux out of a sphere by divergence例题 6.1:用散度求球面的外向通量
Find the outward flux of $\mathbf{F} = \langle x, y, z \rangle$ through the sphere $x^2 + y^2 + z^2 = a^2$.
Computing flux directly is tedious, but the divergence is simple: $\nabla \cdot \mathbf{F} = 1 + 1 + 1 = 3$. Let $E$ be the solid ball of radius $a$. Then
$$ \iint_{\partial E} \mathbf{F}\cdot d\mathbf{S} = \iiint_E 3\, dV = 3 \cdot \operatorname{Vol}(E) = 3 \cdot \frac{4}{3}\pi a^3 = 4\pi a^3. $$求 $\mathbf{F} = \langle x, y, z \rangle$ 穿过球面 $x^2 + y^2 + z^2 = a^2$ 的外向通量。
直接算通量很繁琐,但散度很简单:$\nabla \cdot \mathbf{F} = 1 + 1 + 1 = 3$。设 $E$ 为半径 $a$ 的实心球。则
$$ \iint_{\partial E} \mathbf{F}\cdot d\mathbf{S} = \iiint_E 3\, dV = 3 \cdot \operatorname{Vol}(E) = 3 \cdot \frac{4}{3}\pi a^3 = 4\pi a^3. $$Worked Example 6.2: a non-constant divergence over a box例题 6.2:立方体上的非常数散度
Find the outward flux of $\mathbf{F} = \langle x^2,\; y^2,\; z^2\rangle$ through the surface of the cube $0 \le x, y, z \le 1$.
Six face integrals would work, but the divergence makes it one triple integral. Here $\nabla\cdot\mathbf{F} = 2x + 2y + 2z$, so
$$ \iint_{\partial E} \mathbf{F}\cdot d\mathbf{S} = \iiint_E (2x + 2y + 2z)\, dV = \int_0^1\!\!\int_0^1\!\!\int_0^1 (2x + 2y + 2z)\, dx\, dy\, dz. $$By symmetry each of the three terms integrates to the same value. For the first, $\int_0^1 2x\, dx = 1$ and the other two integrations over the unit cube contribute factors of $1$, giving $1$; the three terms sum to $1 + 1 + 1 = 3$. The total outward flux is $3$.
求 $\mathbf{F} = \langle x^2,\; y^2,\; z^2\rangle$ 穿过立方体 $0 \le x, y, z \le 1$ 表面的外向通量。
分六个面积分也行,但用散度只需一个三重积分。此处 $\nabla\cdot\mathbf{F} = 2x + 2y + 2z$,故
$$ \iint_{\partial E} \mathbf{F}\cdot d\mathbf{S} = \iiint_E (2x + 2y + 2z)\, dV = \int_0^1\!\!\int_0^1\!\!\int_0^1 (2x + 2y + 2z)\, dx\, dy\, dz. $$由对称性,三项各自积分得到相同的值。对第一项,$\int_0^1 2x\, dx = 1$,另两次在单位立方体上的积分各贡献因子 $1$,得 $1$;三项之和为 $1 + 1 + 1 = 3$。总外向通量为 $3$。
Worked Example 6.3: closing an open surface to apply the theorem例题 6.3:封闭开曲面以套用定理
Find the upward flux of $\mathbf{F} = \langle 0, 0, z\rangle$ through the open paraboloid $S: z = x^2 + y^2$, $0 \le z \le 1$, oriented with an upward component.
$S$ is open, so the Divergence Theorem does not apply directly. Close the solid bowl $E$ by adding the top disk $T: z = 1$, $x^2 + y^2 \le 1$, oriented upward. On the closed surface $\partial E = S_{\text{down}} \cup T$, where the paraboloid must be oriented outward (downward, away from the solid) to match the outward convention. With $\nabla\cdot\mathbf{F} = 1$,
$$ \iint_{\partial E} \mathbf{F}\cdot d\mathbf{S} = \iiint_E 1\, dV = \operatorname{Vol}(E) = \int_0^{2\pi}\!\!\int_0^1\!\!\int_{r^2}^1 r\, dz\, dr\, d\theta = 2\pi\int_0^1 (r - r^3)\, dr = 2\pi\cdot\tfrac{1}{4} = \tfrac{\pi}{2}. $$The flux through the top disk is $\iint_T \langle 0,0,1\rangle\cdot\langle 0,0,1\rangle\, dA = \pi$. The closed-surface flux equals the outward paraboloid flux plus the top flux: $\tfrac{\pi}{2} = \Phi_{\text{out}} + \pi$, so the outward (downward) paraboloid flux is $-\tfrac{\pi}{2}$. The problem wants the upward orientation, which is the negative, giving $+\tfrac{\pi}{2}$.
求 $\mathbf{F} = \langle 0, 0, z\rangle$ 穿过开抛物面 $S: z = x^2 + y^2$($0 \le z \le 1$,带向上分量定向)的向上通量。
$S$ 是开的,故散度定理不能直接套用。给碗状立体 $E$ 添上顶部圆盘 $T: z = 1$,$x^2 + y^2 \le 1$,向上定向,将其封闭。在闭曲面 $\partial E = S_{\text{down}} \cup T$ 上,抛物面必须向外定向(向下,背离立体)以符合外向约定。由 $\nabla\cdot\mathbf{F} = 1$,
$$ \iint_{\partial E} \mathbf{F}\cdot d\mathbf{S} = \iiint_E 1\, dV = \operatorname{Vol}(E) = \int_0^{2\pi}\!\!\int_0^1\!\!\int_{r^2}^1 r\, dz\, dr\, d\theta = 2\pi\int_0^1 (r - r^3)\, dr = 2\pi\cdot\tfrac{1}{4} = \tfrac{\pi}{2}. $$穿过顶部圆盘的通量为 $\iint_T \langle 0,0,1\rangle\cdot\langle 0,0,1\rangle\, dA = \pi$。闭曲面通量等于抛物面外向通量加顶部通量:$\tfrac{\pi}{2} = \Phi_{\text{out}} + \pi$,故抛物面的外向(向下)通量为 $-\tfrac{\pi}{2}$。题目要的是向上定向,即其相反数,给出 $+\tfrac{\pi}{2}$。
Going deeper: the divergence as flux density深入探究:作为通量密度的散度
Apply the Divergence Theorem to a small ball $B_\varepsilon$ of radius $\varepsilon$ centered at a point $\mathbf{p}$. Since $\nabla \cdot \mathbf{F}$ is continuous, it is nearly constant on $B_\varepsilon$, so
$$ \iint_{\partial B_\varepsilon} \mathbf{F}\cdot d\mathbf{S} = \iiint_{B_\varepsilon} (\nabla\cdot\mathbf{F})\, dV \approx (\nabla\cdot\mathbf{F})(\mathbf{p}) \cdot \operatorname{Vol}(B_\varepsilon). $$Dividing by the volume and letting $\varepsilon \to 0$ gives
$$ (\nabla\cdot\mathbf{F})(\mathbf{p}) = \lim_{\varepsilon \to 0} \frac{1}{\operatorname{Vol}(B_\varepsilon)} \iint_{\partial B_\varepsilon} \mathbf{F}\cdot d\mathbf{S}. $$This is the precise sense in which the divergence is the outward flux per unit volume at a point, independent of any coordinate system.
对以点 $\mathbf{p}$ 为心、半径 $\varepsilon$ 的小球 $B_\varepsilon$ 套用散度定理。由于 $\nabla \cdot \mathbf{F}$ 连续,它在 $B_\varepsilon$ 上近似为常数,故
$$ \iint_{\partial B_\varepsilon} \mathbf{F}\cdot d\mathbf{S} = \iiint_{B_\varepsilon} (\nabla\cdot\mathbf{F})\, dV \approx (\nabla\cdot\mathbf{F})(\mathbf{p}) \cdot \operatorname{Vol}(B_\varepsilon). $$除以体积并令 $\varepsilon \to 0$,得
$$ (\nabla\cdot\mathbf{F})(\mathbf{p}) = \lim_{\varepsilon \to 0} \frac{1}{\operatorname{Vol}(B_\varepsilon)} \iint_{\partial B_\varepsilon} \mathbf{F}\cdot d\mathbf{S}. $$这正是散度作为某点单位体积外向通量的精确含义,与任何坐标系无关。
Summary of the Vector Calculus Theorems向量微积分定理小结
Reading these as a dimensional ladder: the Fundamental Theorem connects a curve to its two endpoints, Green's and Stokes' connect a surface to its bounding curve, and the Divergence Theorem connects a solid to its bounding surface. In each case integrating the derivative inside reproduces the boundary data outside.把它们读作一架维度阶梯:基本定理把曲线与其两个端点相连,格林定理和斯托克斯定理把曲面与其界曲线相连,散度定理把立体与其界曲面相连。每种情形下,对内部的导数积分都重现了外部的边界数据。
Worked Example 7.1: choosing the right theorem例题 7.1:选对定理
You must evaluate the outward flux of $\mathbf{F} = \langle x^3, y^3, z^3 \rangle$ through the closed unit sphere. Which theorem applies, and what is the value?
The surface is closed and bounds a solid, so the Divergence Theorem is the right tool. Compute $\nabla\cdot\mathbf{F} = 3x^2 + 3y^2 + 3z^2 = 3\rho^2$ in spherical coordinates. Then
$$ \iiint_E 3\rho^2\, dV = \int_0^{2\pi}\!\!\int_0^\pi\!\!\int_0^1 3\rho^2 \cdot \rho^2 \sin\phi\, d\rho\, d\phi\, d\theta = 3 \cdot \frac{1}{5} \cdot 2 \cdot 2\pi = \frac{12\pi}{5}. $$If instead the problem asked for circulation around a single closed curve, Stokes' Theorem would be the tool; for an endpoint difference, the Fundamental Theorem.
需要求 $\mathbf{F} = \langle x^3, y^3, z^3 \rangle$ 穿过闭单位球面的外向通量。该用哪个定理,其值是多少?
曲面是闭的且界出一个立体,故散度定理是正确工具。在球坐标下算 $\nabla\cdot\mathbf{F} = 3x^2 + 3y^2 + 3z^2 = 3\rho^2$。则
$$ \iiint_E 3\rho^2\, dV = \int_0^{2\pi}\!\!\int_0^\pi\!\!\int_0^1 3\rho^2 \cdot \rho^2 \sin\phi\, d\rho\, d\phi\, d\theta = 3 \cdot \frac{1}{5} \cdot 2 \cdot 2\pi = \frac{12\pi}{5}. $$若题目问的是绕单条闭曲线的环量,则用斯托克斯定理;若问端点之差,则用基本定理。
Worked Example 7.2: the inverse-square field and a hidden singularity例题 7.2:平方反比场与隐藏的奇点
Consider $\mathbf{F} = \dfrac{\langle x, y, z\rangle}{(x^2 + y^2 + z^2)^{3/2}}$, the inverse-square field. Show that $\nabla\cdot\mathbf{F} = 0$ away from the origin, yet the outward flux through any sphere centered at the origin is $4\pi$, not $0$.
A direct computation gives $\nabla\cdot\mathbf{F} = 0$ everywhere except the origin, where $\mathbf{F}$ is undefined. On the sphere of radius $a$, the outward unit normal is $\mathbf{n} = \langle x, y, z\rangle / a$, so on the surface $\mathbf{F}\cdot\mathbf{n} = \dfrac{x^2+y^2+z^2}{a \cdot a^3} = \dfrac{a^2}{a^4} = \dfrac{1}{a^2}$, a constant. Then
$$ \iint_S \mathbf{F}\cdot d\mathbf{S} = \frac{1}{a^2}\cdot A(S) = \frac{1}{a^2}\cdot 4\pi a^2 = 4\pi. $$The flux is independent of radius and nonzero, even though the divergence vanishes throughout the shell. The resolution is that the Divergence Theorem fails here because the origin, where $\mathbf{F}$ blows up, lies inside. This is the mathematics behind Gauss' law in electrostatics: the flux measures the enclosed charge, which the smooth divergence away from it cannot see.
考虑平方反比场 $\mathbf{F} = \dfrac{\langle x, y, z\rangle}{(x^2 + y^2 + z^2)^{3/2}}$。证明在原点之外 $\nabla\cdot\mathbf{F} = 0$,然而穿过任何以原点为心的球面的外向通量是 $4\pi$,而非 $0$。
直接计算给出除原点外处处 $\nabla\cdot\mathbf{F} = 0$,而在原点处 $\mathbf{F}$ 无定义。在半径 $a$ 的球面上,外向单位法向为 $\mathbf{n} = \langle x, y, z\rangle / a$,故在曲面上 $\mathbf{F}\cdot\mathbf{n} = \dfrac{x^2+y^2+z^2}{a \cdot a^3} = \dfrac{a^2}{a^4} = \dfrac{1}{a^2}$,是常数。则
$$ \iint_S \mathbf{F}\cdot d\mathbf{S} = \frac{1}{a^2}\cdot A(S) = \frac{1}{a^2}\cdot 4\pi a^2 = 4\pi. $$尽管散度在整个壳层中为零,通量却与半径无关且非零。原因在于:散度定理在此失效,因为 $\mathbf{F}$ 在其处爆破的原点落在内部。这正是静电学中高斯定律背后的数学:通量度量的是所围电荷,而远离电荷处的光滑散度看不到它。
Worked Example 7.3: combining two theorems in one problem例题 7.3:一题中综合两个定理
Let $\mathbf{F} = \langle y, -x, z^3\rangle$ and let $S$ be the closed boundary of the solid cylinder $x^2 + y^2 \le 1$, $0 \le z \le 2$. Find the outward flux through $S$, and separately the circulation of $\mathbf{F}$ around the top rim $C$ (the circle $x^2 + y^2 = 1$, $z = 2$, counterclockwise from above).
For the flux, the surface is closed, so use the Divergence Theorem. $\nabla\cdot\mathbf{F} = 0 + 0 + 3z^2 = 3z^2$, hence
$$ \iint_S \mathbf{F}\cdot d\mathbf{S} = \iiint_E 3z^2\, dV = \int_0^{2\pi}\!\!\int_0^1\!\!\int_0^2 3z^2\, r\, dz\, dr\, d\theta = (2\pi)\left(\tfrac{1}{2}\right)\big[z^3\big]_0^2 = (2\pi)\left(\tfrac{1}{2}\right)(8) = 8\pi. $$For the circulation, use Stokes on the flat top disk $D$. The curl is $\nabla\times\mathbf{F} = \langle 0, 0, -2\rangle$, and the upward normal on $D$ is $\langle 0, 0, 1\rangle$, so $(\nabla\times\mathbf{F})\cdot\mathbf{n} = -2$. Thus $\oint_C \mathbf{F}\cdot d\mathbf{r} = \iint_D (-2)\, dA = -2\pi$. The two theorems answer two different questions about the same field on the same solid.
设 $\mathbf{F} = \langle y, -x, z^3\rangle$,$S$ 为实心柱 $x^2 + y^2 \le 1$($0 \le z \le 2$)的闭边界。求穿过 $S$ 的外向通量,并另外求 $\mathbf{F}$ 绕顶边 $C$(圆 $x^2 + y^2 = 1$,$z = 2$,从上方看逆时针)的环量。
对通量,曲面是闭的,故用散度定理。$\nabla\cdot\mathbf{F} = 0 + 0 + 3z^2 = 3z^2$,因此
$$ \iint_S \mathbf{F}\cdot d\mathbf{S} = \iiint_E 3z^2\, dV = \int_0^{2\pi}\!\!\int_0^1\!\!\int_0^2 3z^2\, r\, dz\, dr\, d\theta = (2\pi)\left(\tfrac{1}{2}\right)\big[z^3\big]_0^2 = (2\pi)\left(\tfrac{1}{2}\right)(8) = 8\pi. $$对环量,在平顶圆盘 $D$ 上用斯托克斯定理。旋度为 $\nabla\times\mathbf{F} = \langle 0, 0, -2\rangle$,$D$ 上向上的法向为 $\langle 0, 0, 1\rangle$,故 $(\nabla\times\mathbf{F})\cdot\mathbf{n} = -2$。于是 $\oint_C \mathbf{F}\cdot d\mathbf{r} = \iint_D (-2)\, dA = -2\pi$。两个定理回答的是关于同一场、同一立体的两个不同问题。
Going deeper: the four theorems as one generalized Stokes statement深入探究:四个定理统一为一个广义斯托克斯陈述
The repeating pattern is not a coincidence. In the language of differential forms, every one of these results is a special case of a single identity, the generalized Stokes' Theorem,
$$ \int_{\Omega} d\omega = \int_{\partial \Omega} \omega, $$where $\omega$ is a differential form, $d$ is the exterior derivative, $\Omega$ is an oriented region, and $\partial\Omega$ is its boundary. Choosing $\omega$ to be a $0$-form (a function) recovers the Fundamental Theorem for line integrals; a $1$-form on a planar region recovers Green's Theorem; a $1$-form on a surface recovers the classical Stokes' Theorem; and a $2$-form on a solid recovers the Divergence Theorem. The gradient, curl, and divergence are the three faces of the single operator $d$ acting in dimension three. This is why all four theorems share the form "integral of a derivative over a region equals integral over the boundary," and learning to see them as one statement is what separates a memorized list from genuine understanding.
这个反复出现的模式并非巧合。用微分形式(differential form)的语言,这些结果中的每一个都是同一个恒等式(即广义斯托克斯定理)的特例,
其中 $\omega$ 是一个微分形式,$d$ 是外微分,$\Omega$ 是定向区域,$\partial\Omega$ 是其边界。取 $\omega$ 为 $0$ 形式(一个函数)便得线积分基本定理;平面区域上的 $1$ 形式便得格林定理;曲面上的 $1$ 形式便得经典斯托克斯定理;立体上的 $2$ 形式便得散度定理。梯度、旋度与散度,是单一算子 $d$ 在三维中作用的三副面孔。这正是四个定理都呈"导数在区域上的积分等于在边界上的积分"形式的原因,而学会把它们看作一个陈述,正是把一张死记的清单与真正的理解区分开来的关键。
Flashcards闪卡
Unit Quiz单元测验
Readiness Checklist学习自测清单
Tap each item you can do without notes. 点击你无需参考资料即可完成的项目。0 / 8 mastered0 / 8 已掌握
- Compute the curl and divergence of a vector field from the determinant and sum formulas.用行列式公式与求和公式计算向量场的旋度与散度。
- Parametrize a surface and find the area element $\left|\mathbf{r}_u \times \mathbf{r}_v\right|\, dA$.参数化曲面并求出面积元素 $\left|\mathbf{r}_u \times \mathbf{r}_v\right|\, dA$。
- Evaluate a scalar surface integral for area or mass, including the graph special case.计算用于面积或质量的标量曲面积分,含图像特例。
- Set up and evaluate a flux integral, fixing the orientation of the surface correctly.建立并计算通量积分,正确确定曲面的定向。
- Apply Stokes' Theorem to convert a circulation into a curl flux and verify both sides.运用斯托克斯定理把环量化为旋度通量,并验证两侧相等。
- Apply the Divergence Theorem to convert a closed surface flux into a triple integral.运用散度定理把闭曲面通量化为三重积分。
- Use the identities $\nabla\times(\nabla f) = \mathbf{0}$ and $\nabla\cdot(\nabla\times\mathbf{F}) = 0$.运用恒等式 $\nabla\times(\nabla f) = \mathbf{0}$ 与 $\nabla\cdot(\nabla\times\mathbf{F}) = 0$。
- Choose the correct theorem (Fundamental, Green, Stokes, Divergence) from the geometry of the problem.根据题目的几何选出正确的定理(基本定理、格林、斯托克斯、散度)。