University Calculus · Calculus III大学微积分 · 微积分 III

Unit C7: Vector Fields, Line Integrals, and Green's Theorem单元 C7:向量场、线积分与格林定理

From the work done by a force along a curve to the loop integrals of Green's Theorem, this unit builds the line integral and the conservative-field toolkit at the heart of multivariable calculus.从力沿曲线所做的功,到格林定理(Green's theorem)的环路积分,本单元搭建起多元微积分核心的线积分(line integral)与保守场(conservative field)工具箱。

Calculus III微积分 III Multivariable多元 Vector Calculus向量微积分 MIT 18.02 / GT 2551 / Princeton MAT 201
Read me first.请先读我。 This unit assumes you are comfortable with partial derivatives, gradients, and double integrals. We start with vector fields, build scalar and vector line integrals, prove the Fundamental Theorem for Line Integrals, characterize conservative fields by path independence, and finish with Green's Theorem and its curl and divergence forms. Work the examples by hand before checking the collapsibles, and verify each quiz answer against the worked solution.本单元假设你已能熟练运用偏导数、梯度与二重积分。我们从向量场(vector field)出发,建立标量与向量的线积分,证明线积分基本定理(fundamental theorem of line integrals),用路径无关(path independence)刻画保守场,最后讲格林定理及其旋度(curl)形式与散度形式。先动手做每道例题,再展开折叠的解答;每道测验答案都要对照详解核验。

Vector Fields向量场

A scalar function assigns a number to each point of space. A vector field assigns a vector. Velocity of a fluid, the gravitational pull near a planet, and the electric force around a charge are all naturally described this way: at every point there is both a magnitude and a direction.标量函数给空间中每个点指定一个数;向量场(vector field)则指定一个向量。流体的速度、行星附近的引力、电荷周围的电场力,都天然地用这种方式描述:每一点既有大小,又有方向。

Key idea.核心思想。 A vector field on a region of the plane or space is a vector-valued function that attaches a vector to each point. In two dimensions $\mathbf{F}(x,y) = P(x,y)\,\mathbf{i} + Q(x,y)\,\mathbf{j}$; in three dimensions $\mathbf{F}(x,y,z) = P\,\mathbf{i} + Q\,\mathbf{j} + R\,\mathbf{k}$. The component functions $P, Q, R$ are ordinary scalar functions.平面或空间某区域上的向量场,是给每个点附上一个向量的向量值函数。二维时 $\mathbf{F}(x,y) = P(x,y)\,\mathbf{i} + Q(x,y)\,\mathbf{j}$;三维时 $\mathbf{F}(x,y,z) = P\,\mathbf{i} + Q\,\mathbf{j} + R\,\mathbf{k}$。分量函数 $P, Q, R$ 都是普通的标量函数。
Vector field in the plane and in space平面与空间中的向量场
$$ \mathbf{F}(x,y) = \langle P(x,y),\, Q(x,y)\rangle, \qquad \mathbf{F}(x,y,z) = \langle P(x,y,z),\, Q(x,y,z),\, R(x,y,z)\rangle. $$

The most important family of examples comes from differentiation. If $f$ is a scalar function, its gradient $\nabla f$ is a vector field that points in the direction of steepest increase of $f$. A field that arises this way is called a gradient field, and the function $f$ is its potential.最重要的一类例子来自求导。若 $f$ 是标量函数,其梯度 $\nabla f$ 就是一个向量场,指向 $f$ 上升最快的方向。这样产生的场称为梯度场(gradient field),函数 $f$ 则是它的势函数(potential function)。

Gradient field梯度场
$$ \nabla f = \left\langle \frac{\partial f}{\partial x},\ \frac{\partial f}{\partial y},\ \frac{\partial f}{\partial z} \right\rangle. $$

Remark.注记。 Sketching a field by hand means evaluating $\mathbf{F}$ at a grid of points and drawing the resulting arrows. Long arrows mark fast flow or strong force; the arrows are tangent to the trajectories that a particle carried by the field would follow.手绘向量场,就是在一组网格点上算出 $\mathbf{F}$ 的值并画出对应的箭头。箭头越长,表示流动越快或力越强;这些箭头与被场带动的质点所走轨迹相切。

Two families of fields recur so often that they deserve names. A central or radial field has the form $\mathbf{F} = g(r)\,\langle x, y\rangle$, where $r = \sqrt{x^2+y^2}$; every arrow points along the line through the origin. A rotational field has the form $\mathbf{F} = h(r)\,\langle -y, x\rangle$; every arrow is perpendicular to that line. Gravitational and electrostatic fields are central with $g(r) = c/r^3$, so their magnitude $|\mathbf{F}| = c/r^2$ follows the inverse-square law. The velocity field of a rotating disk is rotational with $h(r)$ constant.有两类场反复出现,值得专门命名。中心场(径向场)形如 $\mathbf{F} = g(r)\,\langle x, y\rangle$,其中 $r = \sqrt{x^2+y^2}$,每个箭头都沿着过原点的直线方向。旋转场形如 $\mathbf{F} = h(r)\,\langle -y, x\rangle$,每个箭头都垂直于那条直线。引力场与静电场是 $g(r) = c/r^3$ 的中心场,故其大小 $|\mathbf{F}| = c/r^2$ 服从平方反比律。旋转圆盘的速度场则是 $h(r)$ 取常数的旋转场。

Common error.常见错误。 Students often confuse a vector field $\mathbf{F}(x,y) = \langle P, Q\rangle$ with the scalar function $f(x,y)$ whose graph is a surface. A vector field has no graph as a surface: it is a rule that plants an arrow at each point of the plane. In particular $\nabla f$ is a vector field, but $f$ itself is a scalar function. Writing $\nabla f$ as a single number, or drawing $\mathbf{F}$ as a height, are both category mistakes that cause sign and dimension errors downstream.学生常把向量场 $\mathbf{F}(x,y) = \langle P, Q\rangle$ 和图像为曲面的标量函数 $f(x,y)$ 混为一谈。向量场没有曲面图像:它是在平面每一点插上一支箭头的规则。特别地,$\nabla f$ 是向量场,而 $f$ 本身是标量函数。把 $\nabla f$ 写成一个数,或把 $\mathbf{F}$ 当成高度来画,都是类型错误,会在后续引发符号与维数错误。
Worked Example 1.1: sketching a radial field例题 1.1:勾画一个径向场

Describe $\mathbf{F}(x,y) = \langle x, y\rangle$.描述 $\mathbf{F}(x,y) = \langle x, y\rangle$。

At the point $(x,y)$ the vector is $\langle x,y\rangle$, which is the position vector itself. So every arrow points directly away from the origin, and its length is $\sqrt{x^2+y^2}$, the distance from the origin.在点 $(x,y)$ 处的向量是 $\langle x,y\rangle$,正是位置向量本身。于是每个箭头都直接背离原点,其长度为 $\sqrt{x^2+y^2}$,即到原点的距离。

$$ |\mathbf{F}(x,y)| = \sqrt{x^2+y^2}. $$

The field is radial and grows linearly with distance: short arrows near the origin, long arrows far away. This is the gradient field of $f(x,y) = \tfrac12(x^2+y^2)$, since $\nabla f = \langle x, y\rangle$.该场是径向的,且随距离线性增长:原点附近箭头短,远处箭头长。它是 $f(x,y) = \tfrac12(x^2+y^2)$ 的梯度场,因为 $\nabla f = \langle x, y\rangle$。

Worked Example 1.2: a rotational field例题 1.2:一个旋转场

Describe $\mathbf{F}(x,y) = \langle -y, x\rangle$.描述 $\mathbf{F}(x,y) = \langle -y, x\rangle$。

The vector $\langle -y, x\rangle$ is perpendicular to the position vector $\langle x, y\rangle$, since their dot product is $(-y)(x) + (x)(y) = 0$. Its length is again $\sqrt{x^2+y^2}$.向量 $\langle -y, x\rangle$ 垂直于位置向量 $\langle x, y\rangle$,因为二者的标量积为 $(-y)(x) + (x)(y) = 0$。其长度同样是 $\sqrt{x^2+y^2}$。

So the arrows circle the origin counterclockwise with speed proportional to the radius. This is the model velocity field of a rigid rotation. As we will see in Section 6, this field is not a gradient field.于是箭头绕原点逆时针环行,速度与半径成正比。这是刚性转动的标准速度场模型。正如第 6 节将看到的,这个场不是梯度场。

Worked Example 1.3: a gravitational field and its potential例题 1.3:引力场及其势函数

Show that the inverse-square field $\mathbf{F} = -\dfrac{\langle x, y, z\rangle}{(x^2+y^2+z^2)^{3/2}}$ is the gradient of $f = \dfrac{1}{\sqrt{x^2+y^2+z^2}}$ away from the origin.证明:在原点之外,平方反比场 $\mathbf{F} = -\dfrac{\langle x, y, z\rangle}{(x^2+y^2+z^2)^{3/2}}$ 是 $f = \dfrac{1}{\sqrt{x^2+y^2+z^2}}$ 的梯度。

Write $r = \sqrt{x^2+y^2+z^2}$, so $f = r^{-1}$. Differentiate using $\dfrac{\partial r}{\partial x} = \dfrac{x}{r}$:记 $r = \sqrt{x^2+y^2+z^2}$,则 $f = r^{-1}$。利用 $\dfrac{\partial r}{\partial x} = \dfrac{x}{r}$ 求导:

$$ \frac{\partial f}{\partial x} = -r^{-2}\cdot\frac{\partial r}{\partial x} = -r^{-2}\cdot\frac{x}{r} = -\frac{x}{r^3}. $$

By symmetry the $y$ and $z$ partials are $-y/r^3$ and $-z/r^3$, so由对称性,关于 $y$ 和 $z$ 的偏导分别为 $-y/r^3$ 与 $-z/r^3$,于是

$$ \nabla f = -\frac{\langle x, y, z\rangle}{r^3} = \mathbf{F}. $$

This is exactly the Newtonian gravitational and Coulomb electrostatic field. Because it is a gradient field, the work it does between two points depends only on the radial distances of those points, a fact we exploit in Sections 4 and 5.这正是牛顿引力场与库仑静电场。因为它是梯度场,它在两点之间所做的功只依赖于这两点的径向距离,这一事实我们将在第 4、5 节加以利用。

Which vector field is the gradient of $f(x,y) = x^2 y$?下列哪个向量场是 $f(x,y) = x^2 y$ 的梯度?
1.1
$\langle 2xy,\, x^2 y\rangle$
$\langle 2xy,\, x^2\rangle$
$\langle x^2,\, 2xy\rangle$
$\langle 2x,\, x^2\rangle$
Correct. $\partial f/\partial x = 2xy$ and $\partial f/\partial y = x^2$, so $\nabla f = \langle 2xy,\, x^2\rangle$.正确。$\partial f/\partial x = 2xy$,$\partial f/\partial y = x^2$,故 $\nabla f = \langle 2xy,\, x^2\rangle$。
Differentiate $f = x^2 y$ with respect to $x$ (treating $y$ constant) and then with respect to $y$. The result is $\langle 2xy,\, x^2\rangle$.把 $f = x^2 y$ 先对 $x$ 求导(把 $y$ 当常数),再对 $y$ 求导,结果是 $\langle 2xy,\, x^2\rangle$。

Line Integrals of Scalar Functions标量函数的线积分

An ordinary integral $\int_a^b f\,dx$ adds up values of $f$ along a straight segment of the $x$-axis. A line integral generalizes this by letting the path of integration be any curve $C$ in the plane or in space. We add up values of a scalar function along that curve, weighting by arc length.普通积分 $\int_a^b f\,dx$ 沿 $x$ 轴上一段直线累加 $f$ 的值。线积分(line integral)把它推广开来,允许积分路径是平面或空间中的任意曲线 $C$。我们沿该曲线累加一个标量函数的值,并以弧长为权。

Key idea.核心思想。 Parametrize the curve by $\mathbf{r}(t) = \langle x(t), y(t)\rangle$ for $a \le t \le b$. The element of arc length is $ds = |\mathbf{r}'(t)|\,dt$. The line integral of a scalar function $f$ with respect to arc length sums $f\,ds$ over the curve.把曲线参数化为 $\mathbf{r}(t) = \langle x(t), y(t)\rangle$,$a \le t \le b$。弧长元为 $ds = |\mathbf{r}'(t)|\,dt$。标量函数 $f$ 关于弧长的线积分,就是沿曲线累加 $f\,ds$。
Line integral with respect to arc length关于弧长的线积分
$$ \int_C f\,ds = \int_a^b f\big(x(t), y(t)\big)\,\big|\mathbf{r}'(t)\big|\,dt, \qquad |\mathbf{r}'(t)| = \sqrt{x'(t)^2 + y'(t)^2}. $$

If $f \equiv 1$, the integral returns the length of $C$. If $f$ is the linear mass density of a wire bent along $C$, the integral returns the total mass. The value does not depend on the chosen parametrization or on the direction of travel, because $ds \ge 0$ is intrinsic to the curve.若 $f \equiv 1$,积分给出 $C$ 的长度。若 $f$ 是沿 $C$ 弯曲的金属丝的线密度,积分给出总质量。其值既不依赖所选的参数化,也不依赖行进方向,因为 $ds \ge 0$ 是曲线本身固有的量。

Mass and arc length as special cases质量与弧长作为特例
$$ \text{length}(C) = \int_C 1\,ds, \qquad \text{mass} = \int_C \rho(x,y)\,ds. $$
Worked Example 2.1: integrating along a parabola例题 2.1:沿抛物线积分

Evaluate $\displaystyle\int_C 2x\,ds$ where $C$ is the parabola $y = x^2$ from $(0,0)$ to $(1,1)$.求 $\displaystyle\int_C 2x\,ds$,其中 $C$ 是抛物线 $y = x^2$ 从 $(0,0)$ 到 $(1,1)$ 的一段。

Parametrize by $x = t$, $y = t^2$, $0 \le t \le 1$. Then $\mathbf{r}'(t) = \langle 1, 2t\rangle$ and $|\mathbf{r}'(t)| = \sqrt{1 + 4t^2}$.参数化为 $x = t$,$y = t^2$,$0 \le t \le 1$。则 $\mathbf{r}'(t) = \langle 1, 2t\rangle$,且 $|\mathbf{r}'(t)| = \sqrt{1 + 4t^2}$。

$$ \int_C 2x\,ds = \int_0^1 2t\,\sqrt{1+4t^2}\,dt. $$

Let $u = 1 + 4t^2$, so $du = 8t\,dt$ and $2t\,dt = \tfrac14\,du$. As $t$ runs $0 \to 1$, $u$ runs $1 \to 5$.令 $u = 1 + 4t^2$,则 $du = 8t\,dt$,$2t\,dt = \tfrac14\,du$。当 $t$ 从 $0$ 到 $1$ 时,$u$ 从 $1$ 到 $5$。

$$ \int_0^1 2t\sqrt{1+4t^2}\,dt = \frac14\int_1^5 \sqrt{u}\,du = \frac14\cdot\frac{2}{3}\,u^{3/2}\Big|_1^5 = \frac{1}{6}\left(5^{3/2} - 1\right). $$
Worked Example 2.2: mass of a wire on a helix例题 2.2:螺旋线上金属丝的质量

A wire follows the helix $\mathbf{r}(t) = \langle \cos t,\ \sin t,\ t\rangle$, $0 \le t \le 2\pi$, with linear density $\rho(x,y,z) = z$. Find its mass.一根金属丝沿螺旋线 $\mathbf{r}(t) = \langle \cos t,\ \sin t,\ t\rangle$ 分布,$0 \le t \le 2\pi$,线密度为 $\rho(x,y,z) = z$。求其质量。

The velocity is $\mathbf{r}'(t) = \langle -\sin t,\ \cos t,\ 1\rangle$, so the speed is constant:速度为 $\mathbf{r}'(t) = \langle -\sin t,\ \cos t,\ 1\rangle$,故速率是常数:

$$ |\mathbf{r}'(t)| = \sqrt{\sin^2 t + \cos^2 t + 1} = \sqrt{2}. $$

On the curve $z = t$, so $\rho = t$ and在曲线上 $z = t$,故 $\rho = t$,于是

$$ \text{mass} = \int_C z\,ds = \int_0^{2\pi} t\,\sqrt{2}\,dt = \sqrt{2}\cdot\frac{t^2}{2}\Big|_0^{2\pi} = \sqrt{2}\,\frac{(2\pi)^2}{2} = 2\sqrt{2}\,\pi^2. $$

The constant speed of a helix is what makes its arc-length integrals so clean: $ds = \sqrt{2}\,dt$ throughout.螺旋线的速率恒定,正是它的弧长积分如此简洁的原因:处处 $ds = \sqrt{2}\,dt$。

Worked Example 2.3: a piecewise path例题 2.3:一条分段路径

Evaluate $\displaystyle\int_C x\,ds$ where $C$ goes from $(0,0)$ to $(1,0)$ along the $x$-axis, then from $(1,0)$ to $(1,1)$ straight up.求 $\displaystyle\int_C x\,ds$,其中 $C$ 先沿 $x$ 轴从 $(0,0)$ 到 $(1,0)$,再竖直向上从 $(1,0)$ 到 $(1,1)$。

A line integral over a piecewise curve is the sum of the integrals over each smooth piece. On $C_1$ use $\mathbf{r}(t) = \langle t, 0\rangle$, $0 \le t \le 1$, so $ds = dt$ and the integrand is $x = t$:分段曲线上的线积分是各光滑段上积分之和。在 $C_1$ 上取 $\mathbf{r}(t) = \langle t, 0\rangle$,$0 \le t \le 1$,故 $ds = dt$,被积量为 $x = t$:

$$ \int_{C_1} x\,ds = \int_0^1 t\,dt = \tfrac12. $$

On $C_2$ use $\mathbf{r}(t) = \langle 1, t\rangle$, $0 \le t \le 1$, so again $ds = dt$ but now $x = 1$ throughout:在 $C_2$ 上取 $\mathbf{r}(t) = \langle 1, t\rangle$,$0 \le t \le 1$,故仍有 $ds = dt$,但此时处处 $x = 1$:

$$ \int_{C_2} x\,ds = \int_0^1 1\,dt = 1. $$

Adding the pieces gives $\int_C x\,ds = \tfrac12 + 1 = \tfrac32$.两段相加得 $\int_C x\,ds = \tfrac12 + 1 = \tfrac32$。

Common error.常见错误。 A frequent slip is replacing $ds = |\mathbf{r}'(t)|\,dt$ with simply $dt$, or with $dx$. The factor $|\mathbf{r}'(t)|$ is the speed and it is rarely $1$. Only when the curve is parametrized by arc length, so that $|\mathbf{r}'(t)| = 1$, does $ds = dt$ hold. Dropping the speed factor turns a true arc-length integral into a coordinate integral and gives the wrong length, mass, or average value. Always compute $|\mathbf{r}'(t)| = \sqrt{x'(t)^2 + y'(t)^2}$ before integrating.一个常见的失误是把 $ds = |\mathbf{r}'(t)|\,dt$ 直接换成 $dt$ 或 $dx$。因子 $|\mathbf{r}'(t)|$ 是速率,很少等于 $1$。只有当曲线按弧长参数化、即 $|\mathbf{r}'(t)| = 1$ 时,$ds = dt$ 才成立。丢掉速率因子会把真正的弧长积分变成坐标积分,给出错误的长度、质量或平均值。积分前务必先算出 $|\mathbf{r}'(t)| = \sqrt{x'(t)^2 + y'(t)^2}$。
Going deeper: why the scalar line integral is parametrization independent深入探究:为什么标量线积分与参数化无关

Claim: $\int_C f\,ds$ does not depend on how $C$ is parametrized. Suppose $\mathbf{r}(t)$, $a \le t \le b$, and a second parametrization $\mathbf{R}(\tau) = \mathbf{r}(\phi(\tau))$, $\alpha \le \tau \le \beta$, where $\phi$ is a smooth increasing change of variable with $\phi(\alpha) = a$ and $\phi(\beta) = b$.断言:$\int_C f\,ds$ 不依赖 $C$ 的参数化方式。设有 $\mathbf{r}(t)$,$a \le t \le b$,以及第二种参数化 $\mathbf{R}(\tau) = \mathbf{r}(\phi(\tau))$,$\alpha \le \tau \le \beta$,其中 $\phi$ 是光滑的递增换元,满足 $\phi(\alpha) = a$,$\phi(\beta) = b$。

By the chain rule $\mathbf{R}'(\tau) = \mathbf{r}'(\phi(\tau))\,\phi'(\tau)$, and since $\phi' > 0$ we have $|\mathbf{R}'(\tau)| = |\mathbf{r}'(\phi(\tau))|\,\phi'(\tau)$. Substitute into the second integral:由链式法则 $\mathbf{R}'(\tau) = \mathbf{r}'(\phi(\tau))\,\phi'(\tau)$,又因 $\phi' > 0$,故 $|\mathbf{R}'(\tau)| = |\mathbf{r}'(\phi(\tau))|\,\phi'(\tau)$。代入第二个积分:

$$ \int_\alpha^\beta f\big(\mathbf{R}(\tau)\big)\,|\mathbf{R}'(\tau)|\,d\tau = \int_\alpha^\beta f\big(\mathbf{r}(\phi(\tau))\big)\,|\mathbf{r}'(\phi(\tau))|\,\phi'(\tau)\,d\tau. $$

Now set $t = \phi(\tau)$, so $dt = \phi'(\tau)\,d\tau$, and the limits become $a$ to $b$:再令 $t = \phi(\tau)$,则 $dt = \phi'(\tau)\,d\tau$,积分限变为 $a$ 到 $b$:

$$ = \int_a^b f\big(\mathbf{r}(t)\big)\,|\mathbf{r}'(t)|\,dt. $$

The two integrals are equal, so the value belongs to the curve, not the parametrization. The same substitution with a decreasing $\phi$ (a reversal) introduces $|\phi'| = -\phi'$ and swaps the limits, and the two sign changes cancel, which is why reversing direction also leaves $\int_C f\,ds$ unchanged.两个积分相等,可见其值属于曲线本身,而非参数化。若 $\phi$ 递减(即反向),同样的换元会带来 $|\phi'| = -\phi'$ 并交换积分限,这两处符号变化相互抵消,这正是反转方向也不改变 $\int_C f\,ds$ 的原因。

For the line integral $\int_C f\,ds$, the arc-length element along $\mathbf{r}(t)$ equals:对线积分 $\int_C f\,ds$,沿 $\mathbf{r}(t)$ 的弧长元等于:
2.1
$\mathbf{r}'(t)\,dt$
$\mathbf{r}(t)\,dt$
$|\mathbf{r}'(t)|\,dt$
$|\mathbf{r}(t)|\,dt$
Correct. $ds = |\mathbf{r}'(t)|\,dt$ is the speed times $dt$, the scalar length of the displacement.正确。$ds = |\mathbf{r}'(t)|\,dt$ 是速率乘以 $dt$,即位移的标量长度。
Arc length accumulates the magnitude of the velocity: $ds = |\mathbf{r}'(t)|\,dt$, a scalar.弧长累加的是速度的大小:$ds = |\mathbf{r}'(t)|\,dt$,是一个标量。

Line Integrals of Vector Fields向量场的线积分

The central object of this unit is the line integral of a vector field, which measures work. If a force $\mathbf{F}$ pushes a particle along a curve $C$, only the component of $\mathbf{F}$ along the direction of motion does work. We integrate that tangential component over the path.本单元的核心对象是向量场的线积分,它度量功。若力 $\mathbf{F}$ 推动质点沿曲线 $C$ 运动,只有 $\mathbf{F}$ 沿运动方向的分量做功。我们把这个切向分量沿路径积分。

Key idea.核心思想。 The work done by a vector field $\mathbf{F}$ along an oriented curve $C$ is the integral of the tangential component $\mathbf{F}\cdot\mathbf{T}$ with respect to arc length, which collapses to a clean dot product against $d\mathbf{r}$.向量场 $\mathbf{F}$ 沿定向曲线 $C$ 所做的功,是切向分量 $\mathbf{F}\cdot\mathbf{T}$ 关于弧长的积分,它可化简为对 $d\mathbf{r}$ 的一个简洁标量积。
Work integral of a vector field向量场的功积分
$$ \int_C \mathbf{F}\cdot d\mathbf{r} = \int_C \mathbf{F}\cdot\mathbf{T}\,ds = \int_a^b \mathbf{F}\big(\mathbf{r}(t)\big)\cdot\mathbf{r}'(t)\,dt. $$

In components, with $\mathbf{F} = \langle P, Q\rangle$ and $d\mathbf{r} = \langle dx, dy\rangle$, the same integral is written in differential form:写成分量,取 $\mathbf{F} = \langle P, Q\rangle$ 与 $d\mathbf{r} = \langle dx, dy\rangle$,同一积分可写成微分形式:

Differential form微分形式
$$ \int_C \mathbf{F}\cdot d\mathbf{r} = \int_C P\,dx + Q\,dy. $$

Orientation matters.定向至关重要。 Reversing the direction of travel along $C$ flips the sign of the velocity $\mathbf{r}'(t)$, so $\int_{-C} \mathbf{F}\cdot d\mathbf{r} = -\int_C \mathbf{F}\cdot d\mathbf{r}$. This is unlike the scalar arc-length integral of Section 2, which is unaffected by direction.沿 $C$ 反向行进会翻转速度 $\mathbf{r}'(t)$ 的符号,故 $\int_{-C} \mathbf{F}\cdot d\mathbf{r} = -\int_C \mathbf{F}\cdot d\mathbf{r}$。这与第 2 节的标量弧长积分不同,后者不受方向影响。

Worked Example 3.1: work along a line segment例题 3.1:沿线段做功

Compute $\displaystyle\int_C \mathbf{F}\cdot d\mathbf{r}$ for $\mathbf{F} = \langle y, x\rangle$ along the segment from $(0,0)$ to $(2,3)$.对 $\mathbf{F} = \langle y, x\rangle$,沿从 $(0,0)$ 到 $(2,3)$ 的线段计算 $\displaystyle\int_C \mathbf{F}\cdot d\mathbf{r}$。

Parametrize $\mathbf{r}(t) = \langle 2t, 3t\rangle$, $0 \le t \le 1$, so $\mathbf{r}'(t) = \langle 2, 3\rangle$. On the path, $\mathbf{F} = \langle 3t, 2t\rangle$.参数化 $\mathbf{r}(t) = \langle 2t, 3t\rangle$,$0 \le t \le 1$,则 $\mathbf{r}'(t) = \langle 2, 3\rangle$。在路径上 $\mathbf{F} = \langle 3t, 2t\rangle$。

$$ \mathbf{F}\cdot\mathbf{r}'(t) = (3t)(2) + (2t)(3) = 12t. $$ $$ \int_C \mathbf{F}\cdot d\mathbf{r} = \int_0^1 12t\,dt = 6. $$
Worked Example 3.2: orientation flips the sign例题 3.2:定向翻转符号

Compute $\displaystyle\int_C \mathbf{F}\cdot d\mathbf{r}$ for $\mathbf{F} = \langle -y, x\rangle$ around the upper half of the unit circle from $(1,0)$ to $(-1,0)$.对 $\mathbf{F} = \langle -y, x\rangle$,沿单位圆上半部分从 $(1,0)$ 到 $(-1,0)$ 计算 $\displaystyle\int_C \mathbf{F}\cdot d\mathbf{r}$。

Use $\mathbf{r}(t) = \langle \cos t, \sin t\rangle$, $0 \le t \le \pi$, so $\mathbf{r}'(t) = \langle -\sin t, \cos t\rangle$. On the circle $\mathbf{F} = \langle -\sin t, \cos t\rangle$.取 $\mathbf{r}(t) = \langle \cos t, \sin t\rangle$,$0 \le t \le \pi$,则 $\mathbf{r}'(t) = \langle -\sin t, \cos t\rangle$。在圆上 $\mathbf{F} = \langle -\sin t, \cos t\rangle$。

$$ \mathbf{F}\cdot\mathbf{r}'(t) = \sin^2 t + \cos^2 t = 1, \qquad \int_C \mathbf{F}\cdot d\mathbf{r} = \int_0^\pi 1\,dt = \pi. $$

Traversing the same arc backward, from $(-1,0)$ to $(1,0)$, gives $-\pi$. The magnitude is fixed by the path; the sign is fixed by the orientation.把同一段弧反向走,从 $(-1,0)$ 到 $(1,0)$,得 $-\pi$。大小由路径决定,符号由定向(orientation)决定。

Worked Example 3.3: the integral can depend on the path例题 3.3:积分可能依赖路径

Compute $\displaystyle\int_C y\,dx$ for $\mathbf{F} = \langle y, 0\rangle$ along two different routes from $(0,0)$ to $(1,1)$.对 $\mathbf{F} = \langle y, 0\rangle$,沿从 $(0,0)$ 到 $(1,1)$ 的两条不同路线计算 $\displaystyle\int_C y\,dx$。

Route 1, the straight segment.路线 1,直线段。 Use $\mathbf{r}(t) = \langle t, t\rangle$, $0 \le t \le 1$, so $dx = dt$ and $y = t$:取 $\mathbf{r}(t) = \langle t, t\rangle$,$0 \le t \le 1$,故 $dx = dt$,$y = t$:

$$ \int_{C_1} y\,dx = \int_0^1 t\,dt = \tfrac12. $$

Route 2, along the parabola路线 2,沿抛物线 $y = x^2$. Use $\mathbf{r}(t) = \langle t, t^2\rangle$, so $dx = dt$ and $y = t^2$:$y = x^2$。取 $\mathbf{r}(t) = \langle t, t^2\rangle$,故 $dx = dt$,$y = t^2$:

$$ \int_{C_2} y\,dx = \int_0^1 t^2\,dt = \tfrac13. $$

The two answers differ, $\tfrac12 \ne \tfrac13$, so the work done by $\mathbf{F} = \langle y, 0\rangle$ depends on the path, not just the endpoints. This field is therefore not conservative, a point Section 5 makes precise: indeed $P_y = 1 \ne 0 = Q_x$.两个答案不同,$\tfrac12 \ne \tfrac13$,故 $\mathbf{F} = \langle y, 0\rangle$ 所做的功依赖路径,而不只是端点。因此该场不是保守场,这一点第 5 节会精确说明:事实上 $P_y = 1 \ne 0 = Q_x$。

Common error.常见错误。 Do not evaluate $\mathbf{F}$ at the parameter, $\mathbf{F}(t)$, and forget to substitute the curve into the field. You must plug $\mathbf{r}(t)$ into $\mathbf{F}$ to get $\mathbf{F}(\mathbf{r}(t))$ before dotting with $\mathbf{r}'(t)$. A second frequent mistake is to dot the field with the position $\mathbf{r}(t)$ instead of the velocity $\mathbf{r}'(t)$. The work integrand is $\mathbf{F}(\mathbf{r}(t)) \cdot \mathbf{r}'(t)$, the field along the curve dotted with the velocity.不要直接把 $\mathbf{F}$ 写成参数的函数 $\mathbf{F}(t)$,而忘了把曲线代入场。必须先把 $\mathbf{r}(t)$ 代入 $\mathbf{F}$ 得到 $\mathbf{F}(\mathbf{r}(t))$,再与 $\mathbf{r}'(t)$ 做标量积。第二个常见错误是把场与位置 $\mathbf{r}(t)$ 而非速度 $\mathbf{r}'(t)$ 做标量积。功的被积量是 $\mathbf{F}(\mathbf{r}(t)) \cdot \mathbf{r}'(t)$,即曲线上的场与速度做标量积。
If $\int_C \mathbf{F}\cdot d\mathbf{r} = 7$, what is $\int_{-C} \mathbf{F}\cdot d\mathbf{r}$, where $-C$ is $C$ traversed in the opposite direction?若 $\int_C \mathbf{F}\cdot d\mathbf{r} = 7$,则 $\int_{-C} \mathbf{F}\cdot d\mathbf{r}$ 等于多少?其中 $-C$ 是 $C$ 的反向。
3.1
$7$
$0$
$14$
$-7$
Correct. Reversing orientation negates $\mathbf{r}'(t)$, so the work integral changes sign: $-7$.正确。反转定向使 $\mathbf{r}'(t)$ 变号,故功积分变号:$-7$。
The vector line integral is orientation-dependent: reversing the path negates the integral, giving $-7$.向量线积分依赖定向:反转路径使积分变号,得 $-7$。
In differential form, $\int_C \mathbf{F}\cdot d\mathbf{r}$ for $\mathbf{F} = \langle P, Q\rangle$ equals:写成微分形式,对 $\mathbf{F} = \langle P, Q\rangle$,$\int_C \mathbf{F}\cdot d\mathbf{r}$ 等于:
3.2
$\int_C P\,dx + Q\,dy$
$\int_C P\,dy + Q\,dx$
$\int_C P\,dx - Q\,dy$
$\int_C Q\,dx + P\,dy$
Correct. $\mathbf{F}\cdot d\mathbf{r} = \langle P,Q\rangle\cdot\langle dx, dy\rangle = P\,dx + Q\,dy$.正确。$\mathbf{F}\cdot d\mathbf{r} = \langle P,Q\rangle\cdot\langle dx, dy\rangle = P\,dx + Q\,dy$。
The dot product pairs $P$ with $dx$ and $Q$ with $dy$: $\int_C P\,dx + Q\,dy$.标量积把 $P$ 与 $dx$ 配对、$Q$ 与 $dy$ 配对:$\int_C P\,dx + Q\,dy$。

The Fundamental Theorem for Line Integrals线积分基本定理

The Fundamental Theorem of Calculus says $\int_a^b f'(x)\,dx = f(b) - f(a)$: integrating a derivative recovers the net change of the original function. There is an exact analogue for line integrals of gradient fields.微积分基本定理说 $\int_a^b f'(x)\,dx = f(b) - f(a)$:对导数积分,恢复出原函数的净变化。对梯度场的线积分,有一个完全对应的结论。

Key idea.核心思想。 If a vector field is a gradient, $\mathbf{F} = \nabla f$, then its line integral depends only on the endpoints of the path, not on the route taken. The accumulated work is the net change in the potential $f$.若向量场是梯度,$\mathbf{F} = \nabla f$,则它的线积分只依赖路径的端点,与所走的路线无关。累积的功就是势函数 $f$ 的净变化。
Fundamental Theorem for Line Integrals线积分基本定理
$$ \int_C \nabla f\cdot d\mathbf{r} = f\big(\mathbf{r}(b)\big) - f\big(\mathbf{r}(a)\big), $$ $$ \text{where } C \text{ runs from } \mathbf{r}(a) = A \text{ to } \mathbf{r}(b) = B. $$

Two immediate consequences. First, for a gradient field the integral is the same over any two curves sharing the same endpoints (path independence). Second, around any closed loop the integral is zero, since the start and end points coincide.两个直接推论。其一,对梯度场,端点相同的任意两条曲线上积分相等,这就是路径无关(path independence)。其二,绕任意闭合回路积分为零,因为起点与终点重合。

Going deeper: proof of the theorem深入探究:定理的证明

Let $\mathbf{F} = \nabla f$ and parametrize $C$ by $\mathbf{r}(t)$, $a \le t \le b$. Substitute into the work integral:设 $\mathbf{F} = \nabla f$,把 $C$ 参数化为 $\mathbf{r}(t)$,$a \le t \le b$,代入功积分:

$$ \int_C \nabla f\cdot d\mathbf{r} = \int_a^b \nabla f\big(\mathbf{r}(t)\big)\cdot \mathbf{r}'(t)\,dt. $$

By the multivariable chain rule, the integrand is exactly the $t$-derivative of the composition $g(t) = f(\mathbf{r}(t))$:由多元链式法则,被积量恰是复合函数 $g(t) = f(\mathbf{r}(t))$ 对 $t$ 的导数:

$$ \frac{d}{dt}\,f\big(\mathbf{r}(t)\big) = \nabla f\big(\mathbf{r}(t)\big)\cdot\mathbf{r}'(t) = g'(t). $$

Now apply the single-variable Fundamental Theorem of Calculus to $g$:再对 $g$ 应用一元微积分基本定理:

$$ \int_a^b g'(t)\,dt = g(b) - g(a) = f\big(\mathbf{r}(b)\big) - f\big(\mathbf{r}(a)\big). $$

This is the claimed result. The route from $A$ to $B$ never enters the answer.这就是所断言的结论。从 $A$ 到 $B$ 的具体路线从未进入答案。

Worked Example 4.1: evaluating with a potential例题 4.1:用势函数求值

Compute $\displaystyle\int_C \mathbf{F}\cdot d\mathbf{r}$ for $\mathbf{F} = \langle 2xy, x^2\rangle$ along any path from $(0,0)$ to $(3,1)$.对 $\mathbf{F} = \langle 2xy, x^2\rangle$,沿从 $(0,0)$ 到 $(3,1)$ 的任意路径计算 $\displaystyle\int_C \mathbf{F}\cdot d\mathbf{r}$。

First recognize a potential: we need $f$ with $f_x = 2xy$ and $f_y = x^2$. Integrating $f_x$ in $x$ gives $f = x^2 y + h(y)$; then $f_y = x^2 + h'(y) = x^2$ forces $h'(y) = 0$, so $f = x^2 y$.先找势函数:需要 $f$ 满足 $f_x = 2xy$ 与 $f_y = x^2$。把 $f_x$ 对 $x$ 积分得 $f = x^2 y + h(y)$;再由 $f_y = x^2 + h'(y) = x^2$ 推出 $h'(y) = 0$,故 $f = x^2 y$。

$$ \int_C \mathbf{F}\cdot d\mathbf{r} = f(3,1) - f(0,0) = (9)(1) - 0 = 9. $$
Worked Example 4.2: path independence checked two ways例题 4.2:两种方法验证路径无关

For $\mathbf{F} = \langle 2xy, x^2\rangle$ with potential $f = x^2 y$, verify the answer of Worked Example 4.1 by integrating directly along the straight segment from $(0,0)$ to $(3,1)$.对带势函数 $f = x^2 y$ 的 $\mathbf{F} = \langle 2xy, x^2\rangle$,沿从 $(0,0)$ 到 $(3,1)$ 的直线段直接积分,以验证例题 4.1 的答案。

Parametrize $\mathbf{r}(t) = \langle 3t, t\rangle$, $0 \le t \le 1$, so $\mathbf{r}'(t) = \langle 3, 1\rangle$. On the path $\mathbf{F} = \langle 2(3t)(t),\ (3t)^2\rangle = \langle 6t^2,\ 9t^2\rangle$.参数化 $\mathbf{r}(t) = \langle 3t, t\rangle$,$0 \le t \le 1$,则 $\mathbf{r}'(t) = \langle 3, 1\rangle$。在路径上 $\mathbf{F} = \langle 2(3t)(t),\ (3t)^2\rangle = \langle 6t^2,\ 9t^2\rangle$。

$$ \mathbf{F}\cdot\mathbf{r}'(t) = (6t^2)(3) + (9t^2)(1) = 27t^2. $$ $$ \int_C \mathbf{F}\cdot d\mathbf{r} = \int_0^1 27t^2\,dt = 27\cdot\frac{t^3}{3}\Big|_0^1 = 9. $$

The direct computation agrees with the potential method. The Fundamental Theorem saves the parametrization work entirely once a potential is in hand, which is why recognizing a gradient field is so valuable.直接计算与势函数方法一致。一旦手握势函数,基本定理就完全省去了参数化的功夫,这正是识别梯度场如此宝贵的原因。

Worked Example 4.3: a three-dimensional potential例题 4.3:一个三维势函数

Compute $\displaystyle\int_C \mathbf{F}\cdot d\mathbf{r}$ for $\mathbf{F} = \langle yz,\ xz,\ xy\rangle$ from $(1,1,1)$ to $(2,3,4)$.对 $\mathbf{F} = \langle yz,\ xz,\ xy\rangle$,从 $(1,1,1)$ 到 $(2,3,4)$ 计算 $\displaystyle\int_C \mathbf{F}\cdot d\mathbf{r}$。

Look for $f$ with $f_x = yz$. Integrating in $x$ gives $f = xyz + h(y,z)$. Then $f_y = xz + h_y = xz$ forces $h_y = 0$, and $f_z = xy + h_z = xy$ forces $h_z = 0$, so $h$ is constant and $f = xyz$.寻找满足 $f_x = yz$ 的 $f$。对 $x$ 积分得 $f = xyz + h(y,z)$。再由 $f_y = xz + h_y = xz$ 推出 $h_y = 0$,由 $f_z = xy + h_z = xy$ 推出 $h_z = 0$,故 $h$ 为常数,$f = xyz$。

$$ \int_C \mathbf{F}\cdot d\mathbf{r} = f(2,3,4) - f(1,1,1) = (2)(3)(4) - (1)(1)(1) = 24 - 1 = 23. $$

No path was specified, and none is needed: the field is a gradient, so the answer is fixed by the endpoints.题目未给路径,也无需给定:该场是梯度场,故答案由端点确定。

Common error.常见错误。 The Fundamental Theorem applies only when $\mathbf{F}$ is actually a gradient field. Plugging endpoints into a potential for a non-conservative field gives a wrong answer, because no potential exists. Always confirm $\mathbf{F} = \nabla f$ first, either by exhibiting $f$ or by passing the component test of Section 5. A related slip is to compute $f(A) - f(B)$ with the endpoints reversed; the theorem is $f(\text{end}) - f(\text{start})$, matching the orientation of $C$.基本定理只在 $\mathbf{F}$ 确实是梯度场时才适用。对非保守场把端点代入某个"势函数"会得到错误答案,因为根本不存在势函数。务必先确认 $\mathbf{F} = \nabla f$,办法是给出 $f$,或通过第 5 节的分量判别法。一个相关失误是把端点写反、算成 $f(A) - f(B)$;定理是 $f(\text{终点}) - f(\text{起点})$,与 $C$ 的定向一致。
If $\mathbf{F} = \nabla f$, then $\oint_C \mathbf{F}\cdot d\mathbf{r}$ around any closed curve $C$ equals:若 $\mathbf{F} = \nabla f$,则绕任意闭曲线 $C$ 的 $\oint_C \mathbf{F}\cdot d\mathbf{r}$ 等于:
4.1
the length of $C$$C$ 的长度
$0$
$f$ at the start point起点处的 $f$ 值
always positive恒为正
Correct. The start and end points coincide on a closed curve, so $f(B) - f(A) = 0$.正确。闭曲线上起点与终点重合,故 $f(B) - f(A) = 0$。
On a closed loop $A = B$, so the net change in potential is $f(B) - f(A) = 0$.闭合回路上 $A = B$,故势函数的净变化为 $f(B) - f(A) = 0$。

Conservative Fields and Path Independence保守场与路径无关

A field whose line integral depends only on endpoints is called conservative. Section 4 showed that every gradient field is conservative. The converse holds on reasonable domains, so on such domains the three statements below are equivalent.线积分只依赖端点的场称为保守场(conservative field)。第 4 节已表明每个梯度场都是保守场。在合理的定义域上其逆命题也成立,因此在这类定义域上,下面三条陈述等价。

Key idea.核心思想。 On an open connected region, these are equivalent: (1) $\mathbf{F} = \nabla f$ for some potential $f$; (2) $\int_C \mathbf{F}\cdot d\mathbf{r}$ is path independent; (3) $\oint_C \mathbf{F}\cdot d\mathbf{r} = 0$ around every closed loop. A field satisfying any of these is conservative.在开连通区域上,以下三条等价:(1) 存在某势函数 $f$ 使 $\mathbf{F} = \nabla f$;(2) $\int_C \mathbf{F}\cdot d\mathbf{r}$ 路径无关;(3) 绕每条闭合回路 $\oint_C \mathbf{F}\cdot d\mathbf{r} = 0$。满足其中任一条的场即为保守场。

To test a planar field $\mathbf{F} = \langle P, Q\rangle$ quickly, use the mixed-partials condition. If $\mathbf{F}$ is conservative then $P = f_x$ and $Q = f_y$, so $P_y = f_{xy} = f_{yx} = Q_x$ by Clairaut's theorem.要快速检验平面场 $\mathbf{F} = \langle P, Q\rangle$,可用混合偏导条件。若 $\mathbf{F}$ 保守,则 $P = f_x$,$Q = f_y$,由克莱罗定理 $P_y = f_{xy} = f_{yx} = Q_x$。

Component test for a conservative planar field平面保守场的分量判别法
$$ \frac{\partial P}{\partial y} = \frac{\partial Q}{\partial x}. $$

Remark.注记。 On a simply connected domain (one with no holes), the test is also sufficient: $P_y = Q_x$ guarantees a potential exists. On a domain with a hole, the equality is necessary but not sufficient; the classic counterexample is the vortex field of Going Deeper below.在单连通定义域(即没有"洞"的区域)上,该判别法也是充分的:$P_y = Q_x$ 保证势函数存在。在有洞的定义域上,该等式必要但不充分;经典反例是下文"深入探究"中的涡旋场。

Worked Example 5.1: testing and finding a potential例题 5.1:判别并求势函数

Is $\mathbf{F} = \langle 3 + 2xy,\ x^2 - 3y^2\rangle$ conservative? If so, find $f$.$\mathbf{F} = \langle 3 + 2xy,\ x^2 - 3y^2\rangle$ 是保守场吗?若是,求 $f$。

Test: $P = 3 + 2xy$ gives $P_y = 2x$; $Q = x^2 - 3y^2$ gives $Q_x = 2x$. They agree on all of the plane, which is simply connected, so $\mathbf{F}$ is conservative.判别:$P = 3 + 2xy$ 给出 $P_y = 2x$;$Q = x^2 - 3y^2$ 给出 $Q_x = 2x$。二者在整个平面上相等,而平面单连通,故 $\mathbf{F}$ 保守。

Integrate $P$ in $x$: $f = 3x + x^2 y + h(y)$. Then $f_y = x^2 + h'(y)$ must equal $Q = x^2 - 3y^2$, so $h'(y) = -3y^2$ and $h(y) = -y^3$.把 $P$ 对 $x$ 积分:$f = 3x + x^2 y + h(y)$。再令 $f_y = x^2 + h'(y)$ 等于 $Q = x^2 - 3y^2$,得 $h'(y) = -3y^2$,$h(y) = -y^3$。

$$ f(x,y) = 3x + x^2 y - y^3 + \text{constant}. $$
Worked Example 5.2: a field that fails the test例题 5.2:未通过判别的场

Is $\mathbf{F} = \langle x^2 y,\ xy^2\rangle$ conservative on the plane?$\mathbf{F} = \langle x^2 y,\ xy^2\rangle$ 在平面上是保守场吗?

Compute the mixed partials. With $P = x^2 y$ we get $P_y = x^2$. With $Q = xy^2$ we get $Q_x = y^2$.计算混合偏导。由 $P = x^2 y$ 得 $P_y = x^2$;由 $Q = xy^2$ 得 $Q_x = y^2$。

$$ P_y = x^2, \qquad Q_x = y^2. $$

These are not equal as functions (they agree only on the lines $y = \pm x$), so $P_y \ne Q_x$ and $\mathbf{F}$ is not conservative. No potential exists, and the work it does will in general depend on the path. There is nothing to integrate to find $f$, so we stop as soon as the test fails.作为函数它们并不相等(仅在直线 $y = \pm x$ 上相等),故 $P_y \ne Q_x$,$\mathbf{F}$ 不是保守场。势函数不存在,它做的功一般依赖路径。既无 $f$ 可求,判别一失败便就此停手。

Worked Example 5.3: using a potential to evaluate a loop-free integral例题 5.3:用势函数计算非闭合积分

Evaluate $\displaystyle\int_C \mathbf{F}\cdot d\mathbf{r}$ for $\mathbf{F} = \langle e^y,\ x e^y\rangle$ from $(0,0)$ to $(2,\ln 3)$.对 $\mathbf{F} = \langle e^y,\ x e^y\rangle$,从 $(0,0)$ 到 $(2,\ln 3)$ 计算 $\displaystyle\int_C \mathbf{F}\cdot d\mathbf{r}$。

Test first: $P = e^y$ gives $P_y = e^y$; $Q = x e^y$ gives $Q_x = e^y$. They agree on the whole plane, which is simply connected, so $\mathbf{F}$ is conservative.先判别:$P = e^y$ 给出 $P_y = e^y$;$Q = x e^y$ 给出 $Q_x = e^y$。二者在整个平面上相等,而平面单连通,故 $\mathbf{F}$ 保守。

Find $f$: integrate $P = e^y$ in $x$ to get $f = x e^y + h(y)$. Then $f_y = x e^y + h'(y)$ must equal $Q = x e^y$, so $h'(y) = 0$ and $f = x e^y$.求 $f$:把 $P = e^y$ 对 $x$ 积分得 $f = x e^y + h(y)$。再令 $f_y = x e^y + h'(y)$ 等于 $Q = x e^y$,得 $h'(y) = 0$,$f = x e^y$。

$$ \int_C \mathbf{F}\cdot d\mathbf{r} = f(2, \ln 3) - f(0,0) = 2e^{\ln 3} - 0 = 2\cdot 3 = 6. $$
Common error.常见错误。 The component test $P_y = Q_x$ is only conclusive on a simply connected domain. On a region with a hole, passing the test does not prove the field is conservative, as the vortex field below shows. A second pitfall is to test the wrong pair: the planar condition compares $\partial P/\partial y$ with $\partial Q/\partial x$, not $P_x$ with $Q_y$. Mixing up which partial goes with which component is the single most common reason a correct conservative field is wrongly rejected.分量判别 $P_y = Q_x$ 只在单连通定义域上才下得了结论。在有洞的区域上,通过判别并不能证明场保守,下文涡旋场即为明证。第二个陷阱是检验了错误的配对:平面条件比较的是 $\partial P/\partial y$ 与 $\partial Q/\partial x$,而非 $P_x$ 与 $Q_y$。弄混哪个偏导配哪个分量,是把正确的保守场误判出局的头号原因。
Going deeper: why simple connectedness is needed深入探究:为什么需要单连通性

Consider the vortex field on the plane minus the origin:考虑挖去原点的平面上的涡旋场:

$$ \mathbf{F} = \left\langle \frac{-y}{x^2+y^2},\ \frac{x}{x^2+y^2}\right\rangle. $$

A direct computation gives $P_y = Q_x = \dfrac{y^2 - x^2}{(x^2+y^2)^2}$, so the component test is satisfied everywhere the field is defined.直接计算得 $P_y = Q_x = \dfrac{y^2 - x^2}{(x^2+y^2)^2}$,故分量判别在场有定义之处处处满足。

Yet integrate around the unit circle $\mathbf{r}(t) = \langle\cos t, \sin t\rangle$, $0 \le t \le 2\pi$. On the circle $\mathbf{F} = \langle -\sin t, \cos t\rangle$ and $\mathbf{r}'(t) = \langle -\sin t, \cos t\rangle$, so $\mathbf{F}\cdot\mathbf{r}' = 1$ and然而沿单位圆 $\mathbf{r}(t) = \langle\cos t, \sin t\rangle$,$0 \le t \le 2\pi$ 积分。在圆上 $\mathbf{F} = \langle -\sin t, \cos t\rangle$,$\mathbf{r}'(t) = \langle -\sin t, \cos t\rangle$,故 $\mathbf{F}\cdot\mathbf{r}' = 1$,于是

$$ \oint_C \mathbf{F}\cdot d\mathbf{r} = \int_0^{2\pi} 1\,dt = 2\pi \ne 0. $$

A nonzero loop integral means $\mathbf{F}$ is not conservative, even though $P_y = Q_x$. The obstruction is the hole at the origin: the domain is not simply connected, so the test alone cannot certify a potential.回路积分非零意味着 $\mathbf{F}$ 不是保守场,尽管 $P_y = Q_x$。障碍就在原点处的洞:定义域不单连通,单凭判别无法确证势函数存在。

On a simply connected region, $\mathbf{F} = \langle P, Q\rangle$ is conservative if and only if:在单连通区域上,$\mathbf{F} = \langle P, Q\rangle$ 是保守场,当且仅当:
5.1
$P_x = Q_y$
$P = Q$
$P_y = Q_x$
$P_x = Q_x$
Correct. The mixed-partial condition $P_y = Q_x$ characterizes conservative fields on a simply connected domain.正确。混合偏导条件 $P_y = Q_x$ 刻画了单连通定义域上的保守场。
From $P = f_x$ and $Q = f_y$, equality of mixed second partials gives $P_y = Q_x$.由 $P = f_x$、$Q = f_y$,二阶混合偏导相等给出 $P_y = Q_x$。

Green's Theorem格林定理

Green's Theorem connects a line integral around a closed curve in the plane to a double integral over the region it encloses. It is the planar prototype of the great integral theorems of vector calculus, and it turns many loop integrals into routine area integrals.格林定理(Green's theorem)把平面上绕闭曲线的线积分,与它所围区域上的二重积分联系起来。它是向量微积分诸多重要积分定理的平面原型,能把许多回路积分变成常规的面积分。

Key idea.核心思想。 Let $C$ be a piecewise-smooth, simple closed curve oriented counterclockwise, bounding a region $D$. If $P$ and $Q$ have continuous partials on $D$, then the counterclockwise circulation equals the double integral of $Q_x - P_y$ over $D$.设 $C$ 是逆时针定向的分段光滑简单闭曲线,围出区域 $D$。若 $P$、$Q$ 在 $D$ 上有连续偏导,则逆时针环量(circulation)等于 $Q_x - P_y$ 在 $D$ 上的二重积分。
Green's Theorem (circulation form)格林定理(环量形式)
$$ \oint_C P\,dx + Q\,dy = \iint_D \left(\frac{\partial Q}{\partial x} - \frac{\partial P}{\partial y}\right)\,dA. $$

The orientation convention is essential: $C$ is traversed counterclockwise, so the region $D$ stays on the left. A clockwise loop flips the sign. Notice that if $\mathbf{F}$ is conservative then $Q_x - P_y = 0$ throughout $D$, and Green's Theorem reproduces the closed-loop result $\oint_C \mathbf{F}\cdot d\mathbf{r} = 0$ from Section 4.定向约定至关重要:$C$ 逆时针环行,使区域 $D$ 始终位于左侧。顺时针回路则翻转符号。注意,若 $\mathbf{F}$ 保守,则在整个 $D$ 上 $Q_x - P_y = 0$,格林定理便重现了第 4 节的闭合回路结论 $\oint_C \mathbf{F}\cdot d\mathbf{r} = 0$。

A useful corollary computes area as a line integral. Choosing $P, Q$ so that $Q_x - P_y = 1$, for instance $P = -\tfrac{y}{2}$ and $Q = \tfrac{x}{2}$, gives:一个有用的推论把面积写成线积分。选取 $P, Q$ 使 $Q_x - P_y = 1$,例如 $P = -\tfrac{y}{2}$、$Q = \tfrac{x}{2}$,便得:

Area by Green's Theorem用格林定理求面积
$$ \text{Area}(D) = \frac12 \oint_C x\,dy - y\,dx. $$
Worked Example 6.1: a loop integral via the double integral例题 6.1:用二重积分算回路积分

Evaluate $\displaystyle\oint_C (y^2)\,dx + (3xy)\,dy$ where $C$ is the boundary of the unit disk $x^2 + y^2 \le 1$, oriented counterclockwise.求 $\displaystyle\oint_C (y^2)\,dx + (3xy)\,dy$,其中 $C$ 是单位圆盘 $x^2 + y^2 \le 1$ 的边界,逆时针定向。

Here $P = y^2$ and $Q = 3xy$, so $Q_x = 3y$ and $P_y = 2y$, giving $Q_x - P_y = y$.这里 $P = y^2$,$Q = 3xy$,故 $Q_x = 3y$,$P_y = 2y$,得 $Q_x - P_y = y$。

$$ \oint_C P\,dx + Q\,dy = \iint_D y\,dA. $$

The region $D$ is symmetric about the $x$-axis and $y$ is an odd function of $y$, so the integral vanishes:区域 $D$ 关于 $x$ 轴对称,而 $y$ 是 $y$ 的奇函数,故积分为零:

$$ \iint_D y\,dA = 0. $$
Worked Example 6.2: area of an ellipse例题 6.2:椭圆的面积

Find the area enclosed by the ellipse $\mathbf{r}(t) = \langle a\cos t,\ b\sin t\rangle$, $0 \le t \le 2\pi$.求椭圆 $\mathbf{r}(t) = \langle a\cos t,\ b\sin t\rangle$($0 \le t \le 2\pi$)所围的面积。

Use the area formula. With $x = a\cos t$, $y = b\sin t$, we get $dx = -a\sin t\,dt$ and $dy = b\cos t\,dt$, so使用面积公式。取 $x = a\cos t$,$y = b\sin t$,得 $dx = -a\sin t\,dt$,$dy = b\cos t\,dt$,于是

$$ x\,dy - y\,dx = (a\cos t)(b\cos t)\,dt - (b\sin t)(-a\sin t)\,dt = ab(\cos^2 t + \sin^2 t)\,dt = ab\,dt. $$ $$ \text{Area} = \frac12\int_0^{2\pi} ab\,dt = \frac12\,ab\,(2\pi) = \pi ab. $$
Worked Example 6.3: a hard loop integral made easy例题 6.3:化难为易的回路积分

Evaluate $\displaystyle\oint_C (x^4 - y)\,dx + (x + y^4)\,dy$, where $C$ is the boundary of the square $0 \le x \le 1$, $0 \le y \le 1$, counterclockwise.求 $\displaystyle\oint_C (x^4 - y)\,dx + (x + y^4)\,dy$,其中 $C$ 是正方形 $0 \le x \le 1$、$0 \le y \le 1$ 的边界,逆时针。

Direct parametrization would mean four separate segment integrals. Green's Theorem replaces all of them with one double integral. Here $P = x^4 - y$ and $Q = x + y^4$, so直接参数化意味着四段分别积分。格林定理把它们全部换成一个二重积分。这里 $P = x^4 - y$,$Q = x + y^4$,故

$$ Q_x - P_y = (1) - (-1) = 2. $$

The integrand is the constant $2$, so the loop integral is just $2$ times the area of the square:被积量是常数 $2$,故回路积分就是正方形面积的 $2$ 倍:

$$ \oint_C P\,dx + Q\,dy = \iint_D 2\,dA = 2\cdot\text{Area}(D) = 2\cdot 1 = 2. $$

The high powers $x^4$ and $y^4$ never matter, because they sit in the components whose relevant partials cancel. Recognizing this in advance is the whole art of applying Green's Theorem.高次项 $x^4$ 与 $y^4$ 始终无关紧要,因为它们恰好落在那些相关偏导会抵消的分量里。事先看出这一点,正是运用格林定理的全部技巧所在。

Worked Example 6.4: a non-constant integrand in polar coordinates例题 6.4:极坐标下的非常数被积量

Evaluate $\displaystyle\oint_C (-y^3)\,dx + (x^3)\,dy$, where $C$ is the circle $x^2 + y^2 = 4$, counterclockwise.求 $\displaystyle\oint_C (-y^3)\,dx + (x^3)\,dy$,其中 $C$ 是圆 $x^2 + y^2 = 4$,逆时针。

With $P = -y^3$ and $Q = x^3$ we get $Q_x - P_y = 3x^2 - (-3y^2) = 3(x^2+y^2)$. Green's Theorem turns the loop into取 $P = -y^3$,$Q = x^3$,得 $Q_x - P_y = 3x^2 - (-3y^2) = 3(x^2+y^2)$。格林定理把回路化为

$$ \oint_C P\,dx + Q\,dy = \iint_D 3(x^2+y^2)\,dA. $$

The disk of radius $2$ calls for polar coordinates, $x^2+y^2 = r^2$ and $dA = r\,dr\,d\theta$:半径为 $2$ 的圆盘宜用极坐标,$x^2+y^2 = r^2$,$dA = r\,dr\,d\theta$:

$$ \iint_D 3r^2\,dA = \int_0^{2\pi}\!\!\int_0^2 3r^2\cdot r\,dr\,d\theta = \int_0^{2\pi}\!\!\int_0^2 3r^3\,dr\,d\theta = \int_0^{2\pi} \frac{3r^4}{4}\Big|_0^2 d\theta = \int_0^{2\pi} 12\,d\theta = 24\pi. $$
Common error.常见错误。 Green's Theorem applies only to a closed curve and only with the correct orientation. If $C$ is not closed, the theorem does not apply directly; you may need to close it with an auxiliary segment and subtract that segment's contribution. If $C$ is traversed clockwise, the boundary orientation is reversed and the result picks up a minus sign: $\oint_{\text{cw}} = -\iint_D (Q_x - P_y)\,dA$. A third trap is using the theorem when $P$ or $Q$ fails to have continuous partials inside $D$, as with the vortex field, whose singularity at the origin voids the hypotheses.格林定理只适用于闭曲线,且必须定向正确。若 $C$ 不闭合,定理不能直接用;你可能需要用一条辅助线段把它补成闭合曲线,再减去该线段的贡献。若 $C$ 顺时针环行,边界定向被反转,结果带一个负号:$\oint_{\text{cw}} = -\iint_D (Q_x - P_y)\,dA$。第三个陷阱是在 $P$ 或 $Q$ 于 $D$ 内部偏导不连续时硬用定理,如涡旋场,其原点处的奇异性使定理的前提失效。
Green's Theorem rewrites $\oint_C P\,dx + Q\,dy$ (counterclockwise) as:格林定理把 $\oint_C P\,dx + Q\,dy$(逆时针)改写为:
6.1
$\iint_D (Q_x - P_y)\,dA$
$\iint_D (P_x - Q_y)\,dA$
$\iint_D (P_y - Q_x)\,dA$
$\iint_D (P_x + Q_y)\,dA$
Correct. The circulation form integrates $Q_x - P_y$ over the enclosed region $D$.正确。环量形式在所围区域 $D$ 上对 $Q_x - P_y$ 积分。
The integrand is $\partial Q/\partial x - \partial P/\partial y$, that is $Q_x - P_y$.被积量是 $\partial Q/\partial x - \partial P/\partial y$,即 $Q_x - P_y$。

Going Deeper深入探究

Green's Theorem is a special case of a deeper pattern. Two reformulations point toward the rest of vector calculus, and a brief look at the underlying machinery explains why the theorem is true.格林定理是一个更深层模式的特例。两种改写指向向量微积分的其余部分,而对底层机制的简要一瞥,便能说明定理为何成立。

Key idea.核心思想。 Green's Theorem has two faces. The circulation form measures swirl through the scalar curl $Q_x - P_y$. The flux form measures outward flow through the boundary and equals the integral of the divergence $P_x + Q_y$. In three dimensions these become Stokes' Theorem and the Divergence Theorem.格林定理有两副面孔。环量(circulation)形式通过标量旋度(curl)$Q_x - P_y$ 度量旋转。通量(flux)形式度量穿过边界的向外流动,等于散度 $P_x + Q_y$ 的积分。在三维中,它们分别变为斯托克斯定理与散度定理。
Flux (divergence) form of Green's Theorem格林定理的通量(散度)形式
$$ \oint_C \mathbf{F}\cdot\mathbf{n}\,ds = \iint_D \left(\frac{\partial P}{\partial x} + \frac{\partial Q}{\partial y}\right)\,dA, $$ $$ \text{where } \mathbf{n} \text{ is the outward unit normal to } C. $$

The scalar quantity $Q_x - P_y$ is the two-dimensional curl of $\mathbf{F}$, and $P_x + Q_y$ is its divergence. Curl measures local rotation; divergence measures local source strength. The rotational field $\langle -y, x\rangle$ from Section 1 has curl $2$ and divergence $0$: pure swirl, no source. The radial field $\langle x, y\rangle$ has curl $0$ and divergence $2$: pure outflow, no swirl.标量 $Q_x - P_y$ 是 $\mathbf{F}$ 的二维旋度(curl),而 $P_x + Q_y$ 是它的散度。旋度度量局部转动,散度度量局部源的强度。第 1 节的旋转场 $\langle -y, x\rangle$ 旋度为 $2$、散度为 $0$:纯旋转,无源。径向场 $\langle x, y\rangle$ 旋度为 $0$、散度为 $2$:纯外流,无旋转。

Going deeper: sketch of why Green's Theorem holds深入探究:格林定理成立的思路

Prove the special case $\oint_C P\,dx = -\iint_D P_y\,dA$ for a region $D$ that is vertically simple: $a \le x \le b$, $g_1(x) \le y \le g_2(x)$.对竖直方向简单的区域 $D$($a \le x \le b$,$g_1(x) \le y \le g_2(x)$),证明特例 $\oint_C P\,dx = -\iint_D P_y\,dA$。

Evaluate the right side by integrating $P_y$ in $y$ first, using the single-variable Fundamental Theorem:先对 $y$ 积分 $P_y$,用一元基本定理算出右端:

$$ \iint_D P_y\,dA = \int_a^b \big[\,P(x, g_2(x)) - P(x, g_1(x))\,\big]\,dx. $$

Now trace the boundary $C$ counterclockwise. The lower curve $y = g_1(x)$ runs left to right and contributes $\int_a^b P(x, g_1)\,dx$; the upper curve $y = g_2(x)$ runs right to left and contributes $-\int_a^b P(x, g_2)\,dx$. The vertical sides have $dx = 0$ and contribute nothing.再逆时针描出边界 $C$。下方曲线 $y = g_1(x)$ 从左到右,贡献 $\int_a^b P(x, g_1)\,dx$;上方曲线 $y = g_2(x)$ 从右到左,贡献 $-\int_a^b P(x, g_2)\,dx$。竖直两侧 $dx = 0$,无贡献。

$$ \oint_C P\,dx = \int_a^b P(x, g_1)\,dx - \int_a^b P(x, g_2)\,dx = -\iint_D P_y\,dA. $$

An identical argument across horizontally simple regions gives $\oint_C Q\,dy = \iint_D Q_x\,dA$. Adding the two identities yields Green's Theorem; general regions are handled by cutting them into simple pieces, whose interior boundaries cancel in pairs.对水平方向简单的区域作同样论证得 $\oint_C Q\,dy = \iint_D Q_x\,dA$。两式相加即得格林定理;一般区域则切成若干简单片处理,其内部边界成对抵消。

Worked Example 7.1: curl and divergence of a field例题 7.1:场的旋度与散度

Compute the scalar curl and the divergence of $\mathbf{F} = \langle x^2 - y,\ x + y^2\rangle$.计算 $\mathbf{F} = \langle x^2 - y,\ x + y^2\rangle$ 的标量旋度与散度。

Here $P = x^2 - y$ and $Q = x + y^2$.这里 $P = x^2 - y$,$Q = x + y^2$。

$$ \text{curl} = Q_x - P_y = (1) - (-1) = 2, \qquad \text{div} = P_x + Q_y = (2x) + (2y) = 2x + 2y. $$

The curl is constant, so the field has uniform swirl. The divergence varies with position: it is positive in the first quadrant (a source) and negative in the third (a sink).旋度是常数,故场具有均匀的旋转。散度随位置变化:在第一象限为正(源),在第三象限为负(汇)。

Worked Example 7.2: outward flux across a circle例题 7.2:穿过圆周的向外通量

Find the outward flux of $\mathbf{F} = \langle x, y\rangle$ across the circle $x^2 + y^2 = 9$ using the flux form of Green's Theorem.用格林定理的通量形式,求 $\mathbf{F} = \langle x, y\rangle$ 穿过圆 $x^2 + y^2 = 9$ 的向外通量。

The divergence is $P_x + Q_y = 1 + 1 = 2$, a constant. The flux form converts the boundary integral into a double integral of the divergence over the enclosed disk $D$:散度为 $P_x + Q_y = 1 + 1 = 2$,是常数。通量形式把边界积分化为散度在所围圆盘 $D$ 上的二重积分:

$$ \oint_C \mathbf{F}\cdot\mathbf{n}\,ds = \iint_D (P_x + Q_y)\,dA = \iint_D 2\,dA = 2\cdot\text{Area}(D). $$

The disk has radius $3$, so its area is $9\pi$ and the flux is $18\pi$. This matches intuition: the radial field $\langle x, y\rangle$ flows uniformly outward, so its net outflow grows with the enclosed area.圆盘半径为 $3$,故面积为 $9\pi$,通量为 $18\pi$。这与直觉相符:径向场 $\langle x, y\rangle$ 均匀向外流动,故其净外流随所围面积增大。

Going deeper: deriving the area corollary and handling a hole深入探究:推导面积推论并处理"洞"

The area formula.面积公式。 Apply Green's Theorem with $P = -\tfrac{y}{2}$ and $Q = \tfrac{x}{2}$. Then对 $P = -\tfrac{y}{2}$、$Q = \tfrac{x}{2}$ 应用格林定理。则

$$ Q_x - P_y = \frac12 - \left(-\frac12\right) = 1, $$

so the double integral collapses to the area itself:于是二重积分坍缩为面积本身:

$$ \frac12\oint_C x\,dy - y\,dx = \iint_D 1\,dA = \text{Area}(D). $$

Any choice with $Q_x - P_y = 1$ works; the symmetric choice above is simply the most memorable.任何使 $Q_x - P_y = 1$ 的选取都行;上面这种对称选取只是最便于记忆而已。

A region with a hole.带洞的区域。 Green's Theorem extends to a region $D$ between two curves, an outer boundary $C_{\text{out}}$ counterclockwise and an inner boundary $C_{\text{in}}$ clockwise, by cutting $D$ along a bridge and observing the bridge is traversed twice in opposite directions, so it cancels. The result is格林定理可推广到夹在两条曲线之间的区域 $D$:外边界 $C_{\text{out}}$ 逆时针、内边界 $C_{\text{in}}$ 顺时针。沿一条"桥"把 $D$ 切开,注意这座桥被反向走了两次而相互抵消。结果是

$$ \oint_{C_{\text{out}}} P\,dx + Q\,dy \;-\; \oint_{C_{\text{in}}} P\,dx + Q\,dy = \iint_D (Q_x - P_y)\,dA, $$

with both boundaries oriented counterclockwise. For the vortex field of Section 5, $Q_x - P_y = 0$ on the annulus that avoids the origin, so the two loop integrals must be equal. That is why every loop encircling the origin once gives the same value $2\pi$: the integral cannot be deformed across the puncture, but it is unchanged by any deformation that does not cross it. This is the seed of the homotopy invariance that runs through the rest of vector calculus.其中两条边界都按逆时针定向。对第 5 节的涡旋场,在避开原点的圆环上 $Q_x - P_y = 0$,故两个回路积分必相等。这正是每条绕原点一圈的回路都给出同一值 $2\pi$ 的原因:积分无法越过那个穿孔变形,但任何不跨越它的变形都不改变积分值。这便是贯穿向量微积分其余部分的同伦不变性的雏形。

Common error.常见错误。 Curl and divergence are different quantities and must not be swapped. The scalar curl is $Q_x - P_y$ and measures rotation; the divergence is $P_x + Q_y$ and measures spreading. Mixing the signs or the pairings, for instance writing divergence as $Q_x - P_y$, inverts the physical meaning. Memory aid: divergence adds the matched partials $P_x + Q_y$ with a plus, while curl crosses the components and subtracts, $Q_x - P_y$.旋度与散度是不同的量,不可互换。标量旋度是 $Q_x - P_y$,度量转动;散度是 $P_x + Q_y$,度量扩散。弄混符号或配对,例如把散度写成 $Q_x - P_y$,会颠倒物理含义。记忆口诀:散度把对应偏导相加 $P_x + Q_y$(取加号),旋度则交叉分量相减 $Q_x - P_y$。
The divergence of $\mathbf{F} = \langle P, Q\rangle$ in the plane is:平面中 $\mathbf{F} = \langle P, Q\rangle$ 的散度是:
7.1
$Q_x - P_y$
$P_y - Q_x$
$P_x - Q_y$
$P_x + Q_y$
Correct. Divergence sums the rate of spreading in each coordinate: $P_x + Q_y$.正确。散度把各坐标方向上的扩散速率相加:$P_x + Q_y$。
Divergence is $\partial P/\partial x + \partial Q/\partial y$. The expression $Q_x - P_y$ is the scalar curl instead.散度是 $\partial P/\partial x + \partial Q/\partial y$。而 $Q_x - P_y$ 是标量旋度。

Flashcards单词卡

0 / 12 flipped0 / 12 已翻转
Vector field $\mathbf{F}$ in the plane平面中的向量场 $\mathbf{F}$
$\mathbf{F}(x,y) = \langle P(x,y),\, Q(x,y)\rangle$: a vector attached to each point.$\mathbf{F}(x,y) = \langle P(x,y),\, Q(x,y)\rangle$:给每个点附上一个向量。
Gradient field $\nabla f$梯度场 $\nabla f$
$\nabla f = \langle f_x, f_y, f_z\rangle$. Points toward steepest increase of the potential $f$.$\nabla f = \langle f_x, f_y, f_z\rangle$。指向势函数 $f$ 上升最快的方向。
Scalar line integral $\int_C f\,ds$标量线积分 $\int_C f\,ds$
$\int_a^b f(\mathbf{r}(t))\,|\mathbf{r}'(t)|\,dt$. Direction independent; $f\equiv1$ gives arc length.$\int_a^b f(\mathbf{r}(t))\,|\mathbf{r}'(t)|\,dt$。与方向无关;$f\equiv1$ 给出弧长。
Work integral $\int_C \mathbf{F}\cdot d\mathbf{r}$功积分 $\int_C \mathbf{F}\cdot d\mathbf{r}$
$\int_a^b \mathbf{F}(\mathbf{r}(t))\cdot\mathbf{r}'(t)\,dt = \int_C P\,dx + Q\,dy$. Reverses sign under reversed orientation.$\int_a^b \mathbf{F}(\mathbf{r}(t))\cdot\mathbf{r}'(t)\,dt = \int_C P\,dx + Q\,dy$。反转定向时变号。
Fundamental Theorem for Line Integrals线积分基本定理
$\int_C \nabla f\cdot d\mathbf{r} = f(B) - f(A)$. Depends only on the endpoints.$\int_C \nabla f\cdot d\mathbf{r} = f(B) - f(A)$。只依赖端点。
Closed-loop integral of a gradient field梯度场的闭合回路积分
$\oint_C \nabla f\cdot d\mathbf{r} = 0$, since the start and end points coincide.$\oint_C \nabla f\cdot d\mathbf{r} = 0$,因为起点与终点重合。
Conservative field, three equivalent tests保守场的三个等价判据
(1) $\mathbf{F}=\nabla f$; (2) path independence; (3) zero around every closed loop.(1) $\mathbf{F}=\nabla f$;(2) 路径无关;(3) 绕每条闭合回路积分为零。
Component test (simply connected domain)分量判别(单连通定义域)
$\mathbf{F}=\langle P,Q\rangle$ is conservative iff $P_y = Q_x$.$\mathbf{F}=\langle P,Q\rangle$ 保守当且仅当 $P_y = Q_x$。
Green's Theorem (circulation form)格林定理(环量形式)
$\oint_C P\,dx + Q\,dy = \iint_D (Q_x - P_y)\,dA$, $C$ counterclockwise around $D$.$\oint_C P\,dx + Q\,dy = \iint_D (Q_x - P_y)\,dA$,$C$ 逆时针绕 $D$。
Area by Green's Theorem用格林定理求面积
$\text{Area}(D) = \tfrac12\oint_C x\,dy - y\,dx$.
Scalar curl and divergence in the plane平面中的标量旋度与散度
curl $= Q_x - P_y$ (swirl); div $= P_x + Q_y$ (source strength).旋度 $= Q_x - P_y$(旋转);散度 $= P_x + Q_y$(源的强度)。
Vortex field caution涡旋场的警示
$\langle -y,x\rangle/(x^2+y^2)$ satisfies $P_y=Q_x$ but loops to $2\pi$: the domain has a hole, so it is not conservative.$\langle -y,x\rangle/(x^2+y^2)$ 满足 $P_y=Q_x$,但回路积分为 $2\pi$:定义域有洞,故它不保守。

Unit Quiz单元测验

The gradient field of $f(x,y) = x e^{y}$ is:$f(x,y) = x e^{y}$ 的梯度场是:
Q.1
$\langle e^y,\, e^y\rangle$
$\langle x e^y,\, e^y\rangle$
$\langle e^y,\, x e^y\rangle$
$\langle x e^y,\, x e^y\rangle$
Correct. $f_x = e^y$ and $f_y = x e^y$, so $\nabla f = \langle e^y,\, x e^y\rangle$.正确。$f_x = e^y$,$f_y = x e^y$,故 $\nabla f = \langle e^y,\, x e^y\rangle$。
Differentiate $x e^y$ in $x$ to get $e^y$, then in $y$ to get $x e^y$.把 $x e^y$ 对 $x$ 求导得 $e^y$,再对 $y$ 求导得 $x e^y$。
For the scalar line integral $\int_C 1\,ds$, the value is:标量线积分 $\int_C 1\,ds$ 的值是:
Q.2
the area under $C$$C$ 下方的面积
the length of $C$$C$ 的长度
always zero恒为零
the work along $C$沿 $C$ 所做的功
Correct. Integrating the constant $1$ with respect to arc length returns the length of the curve.正确。把常数 $1$ 关于弧长积分,得到曲线的长度。
With $f \equiv 1$, $\int_C 1\,ds = \int_a^b |\mathbf{r}'(t)|\,dt$, which is the arc length of $C$.当 $f \equiv 1$ 时,$\int_C 1\,ds = \int_a^b |\mathbf{r}'(t)|\,dt$,即 $C$ 的弧长。
$\mathbf{F} = \langle 2xy,\, x^2\rangle$ is conservative with potential $f = x^2 y$. Then $\int_C \mathbf{F}\cdot d\mathbf{r}$ from $(1,1)$ to $(2,3)$ equals:$\mathbf{F} = \langle 2xy,\, x^2\rangle$ 保守,势函数为 $f = x^2 y$。则从 $(1,1)$ 到 $(2,3)$ 的 $\int_C \mathbf{F}\cdot d\mathbf{r}$ 等于:
Q.3
$11$
$12$
$1$
$13$
Correct. $f(2,3) - f(1,1) = (4)(3) - (1)(1) = 12 - 1 = 11$.正确。$f(2,3) - f(1,1) = (4)(3) - (1)(1) = 12 - 1 = 11$。
Use the Fundamental Theorem: $f(2,3) - f(1,1) = 12 - 1 = 11$.用基本定理:$f(2,3) - f(1,1) = 12 - 1 = 11$。
Which field is not conservative on the whole plane?下列哪个场在整个平面上不是保守场?
Q.4
$\langle 2x,\, 2y\rangle$
$\langle y,\, x\rangle$
$\langle e^x,\, 1\rangle$
$\langle -y,\, x\rangle$
Correct. For $\langle -y, x\rangle$, $P_y = -1 \ne 1 = Q_x$, so it fails the component test.正确。对 $\langle -y, x\rangle$,$P_y = -1 \ne 1 = Q_x$,故未通过分量判别。
Apply $P_y = Q_x$. Only $\langle -y, x\rangle$ fails: $P_y = -1$ but $Q_x = 1$.用 $P_y = Q_x$ 检验。只有 $\langle -y, x\rangle$ 不满足:$P_y = -1$ 而 $Q_x = 1$。
By Green's Theorem, $\oint_C -y\,dx + x\,dy$ (counterclockwise around region $D$) equals:由格林定理,$\oint_C -y\,dx + x\,dy$(逆时针绕区域 $D$)等于:
Q.5
$0$
$\text{Area}(D)$
$2\,\text{Area}(D)$
$\tfrac12\,\text{Area}(D)$
Correct. Here $Q_x - P_y = 1 - (-1) = 2$, so the integral is $\iint_D 2\,dA = 2\,\text{Area}(D)$.正确。这里 $Q_x - P_y = 1 - (-1) = 2$,故积分为 $\iint_D 2\,dA = 2\,\text{Area}(D)$。
With $P = -y$, $Q = x$: $Q_x - P_y = 1 + 1 = 2$, so the loop integral is $2\,\text{Area}(D)$.取 $P = -y$,$Q = x$:$Q_x - P_y = 1 + 1 = 2$,故回路积分为 $2\,\text{Area}(D)$。
The scalar curl $Q_x - P_y$ of $\mathbf{F} = \langle 3y,\, -x\rangle$ is:$\mathbf{F} = \langle 3y,\, -x\rangle$ 的标量旋度 $Q_x - P_y$ 是:
Q.6
$2$
$-4$
$4$
$0$
Correct. $Q_x = -1$ and $P_y = 3$, so $Q_x - P_y = -1 - 3 = -4$.正确。$Q_x = -1$,$P_y = 3$,故 $Q_x - P_y = -1 - 3 = -4$。
With $P = 3y$, $Q = -x$: $Q_x = -1$, $P_y = 3$, so the curl is $-4$.取 $P = 3y$,$Q = -x$:$Q_x = -1$,$P_y = 3$,故旋度为 $-4$。

Readiness Checklist掌握自查清单

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