数学物理方法 10 格林函数法

$\color{blue}{第十章 格林函数法}$

$\left \lbrace \begin{array}{l}行波法:无界空间波动问题,有局限性 \\ 分离变量法:各种有界问题,其解为无穷级数 \\ 积分变换法:各种无界问题,其解为无限积分 \end{array} \right.$
$1.格林函数法:其解为含有格林函数的有限积分。$
$由\S10.2:\left \lbrace \begin{array}{l} \Delta u = - h(M) \\ u |_{\sigma} = f(M) \end{array} \right. \to$
$\color{blue}{u(M) = \iiint_{\tau} G(M, M_0) h(M_0) d \tau_0 - \iint_{\sigma} f(M_0) \dfrac{\partial G}{\partial n_0} d \sigma_0}$
$\color{blue}{G(M, M_0) - 狄氏格林函数}$
$2.格林函数:点源函数,点源产生的场和影响若外力f(x, t) 只在 \xi 点,\tau 时起作用$
$\left \lbrace \begin{array}{l}u_{tt} = a^2 u_{xx} + f(x, t), f(x, t) = \left \lbrace \begin{array}{l}0, x \neq \xi, t \neq \tau \\ f(\xi, \tau), x = \xi, t = \tau \end{array} \right. \\ u|_{x=0} = 0, u|_{x=l} = 0 \\ u|_{t=0} = 0, u_t|_{t=0} = 0 \\ u(x, t) -格林函数,即G(x, t| \xi, \tau); f(x, t) -点源 \end{array} \right.$

$3.为何引入格林函数法:$
$5.2 \to \left \lbrace \begin{array}{l}\Delta u = - h(M) \\ u|_{\sigma} = f(M) \end{array} \right. \to$
$u(M) = \iiint_{\tau}G(M, M_0)h(M_0) d\tau_0 - \iint_{\sigma} f(M_0) \dfrac{\partial G}{\partial n_0} d \sigma_0$
$(1)解的形式(有限积分)便于理论分析和研究$
$(2)以统一的形式研究各类定解问题$
$(3)对于线性问题,格林函数一旦求出,就可以算出任意源的场,关键就是求点源$

$\color{blue}{\S 10.1 \delta函数}$

$\color{blue}{10.1.1 \delta函数的引入}$

$1.物理背景$
$(1)金属线:总质量m = 1,集中在x = 0处,$
$则密度: \rho(x) = \lim \limits _{\Delta x \to 0} \dfrac{\Delta m}{\Delta x} = \left \lbrace \begin{array}{l}0, x \neq 0 \\ \infty, x = 0 \end{array} \right.$
$\int_{-\infty}^{\infty} \rho(x) dx = 1$

$(2)带电导线:总电量q = 1,集中在x = 0处$
$则电荷密度:\rho(x) = \lim \limits _{\Delta x \to 0} \dfrac{\Delta q}{\Delta x} = \left \lbrace \begin{array}{l}0, x \neq 0 \\ \infty, x = 0 \end{array} \right.$
$\int_{-\infty}^{\infty} \rho(x) dx = 1$

$2.定义:$
$\color{blue}{\left \lbrace \begin{array}{l}\delta(x) = \left \lbrace \begin{array}{l}0, x \neq 0 \\ \infty, x = 0 \end{array} \right. -\delta函数 \\ \int_{-\infty}^{\infty} \delta(x) dx = 1 \end{array} \right.}$

$\color{blue}{一般：\left \lbrace \begin{array}{l}\delta(x-x_0) = \left \lbrace \begin{array}{l}0, x \neq x_0 \\ \infty, x = x_0 \end{array} \right. -\delta函数 \\ \int_{-\infty}^{\infty} \delta(x-x_0) dx = 1 \end{array} \right.}$

$3.注意:$
$(1)\delta-密度函数和点源函数$
$若在x = x_0点放有m质量,总质量m,则\rho(x) = m \delta(x - x_0)$
$若在x=x_0点放有电量为q的点电荷,总电量为q,则\rho(x) = q \delta(x-x_0)$

$(2)\delta 广义函数$

$\color{blue}{10.1.2 \delta函数的性质}$

$设f(x)在(-\infty, \infty)连续,则$
$1.\int_{-\infty}^{\infty} f(x)\delta(x - x_0) dx = f(x_0) \quad [\int_{-\infty}^{\infty} f(x) \delta(x) dx = f(0)]$
$注意:\delta也能表示连续分布的函数$
$f(t) = \int_{-\infty}^{\infty} f(\tau) \delta(\tau - t) d\tau = \int_a^b f(\tau) \delta (\tau - t) d \tau$

$附:判断函数相等的一种方法:$
$设f(x)与g(x)都是定义在(a, b)区间上的函数,若对于定义在(a, b)区间上的任意连续\\ 函数\varphi(x)都有如下等式成立:\int_a^b f(x)\varphi(x) dx = \int_a^b g(x) \varphi(x) dx, 则必有:\\ f(x) = g(x), 特别：若\int_a^b \varphi(x)g(x) dx = 0, 则必有g(x) = 0$

$2.若定义\dfrac{d}{dx} \delta(x) = \delta^{\prime}(x) - \delta函数的导数,则$
$(1)\int_{-\infty}^{\infty} f(x) \delta^{\prime}(x-x_0) dx = - f^{\prime}(x_0)$
$(2)(x-x_0)\delta^{\prime}(x-x_0) = - \delta(x-x_0)$
$(3)\int_{-\infty}^{\infty}f(x) \delta^{(n)}(x-x_0)dx = (-1)^nf^{(n)}(x_0)$

$3.\delta[\varphi(x)] = \sum \limits _{i=1}^{n}\dfrac{\delta(x-x_i)}{|\varphi^{\prime}(x_i)|},其中\varphi(x_i) = 0$

$\color{blue}{10.1.3 高维\delta函数}$

$1.定义：$
$(1)\left \lbrace \begin{array}{l}\delta(M-M_0) = \left \lbrace \begin{array}{l}0, M \neq M_0, \\ \infty, M = M_0, \end{array} \right. \\ \iiint_{-\infty}^{\infty}\delta(M-M_0)dv = 1, 为三维函数,dv = dxdydz \end{array} \right.$
$其中\delta(M-M_0)=\delta(x-x_0,y-y_0,z-z_0)= \delta(x-x_0)\delta(y-y_0)\delta(z-z_0)$

$(2)\left \lbrace \begin{array}{l}\delta(M-M_0) = \left \lbrace \begin{array}{l}0, M \neq M_0, \\ \infty, M = M_0, \end{array} \right. \\ \iint_{-\infty}^{\infty} \delta(M-M_0)dxdy = 1 \quad 为二维函数 \end{array} \right.$
$其中\delta(M-M_0) = \delta(x-x_0, y-y_0) = \delta(x-x_0)\delta(y-y_0)$

$2.性质:$
$(1)\iiint_{-\infty}^{\infty}f(M)\delta(M-M_0)dxdydz = f(x_0, y_0, z_0) = f(M_0)$
$(2)\iint_{-\infty}^{\infty}f(M)\delta(M-M_0)dxdy = f(x_0,y_0) = f(M_0)$

$\color{blue}{10.1.4 例题}$

$1.\int_1^2 \sin x \delta(x - \dfrac{1}{2})dx = 0$
$2.\int_1^2 \sin x \delta(x) dx = 0$
$3.\int_{-\infty}^{\infty} \int_{-\infty}^{\infty} \sin(x+y) \delta(x+2) \delta(y-1)dxdy = \sin(-1)$
$4.长为1，密度为\rho的弦两端固定,初位移为零,初始时刻在x = x_0点受到一横向\\ 冲量I_0.试写出弦的横震动的定解问题.$
$\left \lbrace \begin{array}{l}u_{tt} = a^2 u_{xx} \\ u|_{x=0} = 0, u|_{x=l} = 0 \\ u|_{t=0} = 0 \\ u_t | _{t=0} = \dfrac{I_0}{\rho}\delta(x-x_0) \end{array} \right.$

$\color{blue}{\S 10.2 泊松方程的狄氏问题}$

$\left \lbrace \begin{array}{l}\Delta u = -h(M), M \in \tau \quad (1) \\ u|_{\sigma=0} = g(M) \quad (2) \end{array} \right.$

$\color{blue}{10.2.1 格林公式}$

$1.为何引入格林公式$
$(1)积分公式的起点是通过直接积分或分部积分将未知函数从微分号下解脱出来$
$(2)我们要求解的三类数值方程中均含有\Delta,格林公式是将未知函数,从微分算法\Delta\\ 下解脱出来的工具.设u(x, y, z), v(x, y, z)在\tau中具有连续的二阶导数,在\bar \tau上具\\ 有连续的一阶导数,则有如下格林公式:$
$2.格林第一公式$
$\int_{\tau} u \Delta v d \tau + \int_{\tau} \nabla u \cdot \nabla v d \tau = \int_{\sigma} u \dfrac{\partial v}{\partial n} d \sigma \quad (3)$
$\int_{\tau} v \Delta u d \tau + \int_{\tau} \nabla u \cdot \nabla v d \tau = \int_{\sigma} v \dfrac{\partial u}{\partial n} d \sigma \quad (4)$

$3.格林第二公式$
$\int_{\tau} u \Delta v d \tau - \int_{\tau} v \Delta u d \tau = \int_{\sigma}(u \dfrac{\partial v}{\partial n} - v \dfrac{\partial u}{\partial n}) d\sigma \quad (5)$
$意义:(1)将u,v,\Delta u, \Delta v的值与u, v, \dfrac{\partial v}{\partial n}, \dfrac{\partial u}{\partial n}的边值联系起来.$
$(2)u,v堆成$
$(3)若已知v, \Delta v = 0,及v|_{\sigma},则由格林公式可能求得\left \lbrace \begin{array}{l}\Delta u = -h(M) \quad(1) \\ u|_{\sigma} = g(M) \quad (2) \end{array} \right.之解.$

$4.球面平均值公式$
$(1)定义:$
$\bar u(r, t) = \dfrac{1}{4 \pi r^2}\iint_{S_r^{M_0}} u(M, t) ds$
$= \dfrac{1}{4 \pi}\iint_{S_r^{M_0}}u(M, t) d \Omega \quad d\Omega = \dfrac{ds}{r^2} = \sin \theta d \theta d \varphi$
$-u(M, t)在以M_0为中心，r为半径的球面S_r^{M_0}上的平均值.$
$(2)显然 u(M_0, t_0) = \lim \limits_{r \to 0} \bar u(r, t_0)$

$\color{blue}{10.2.2 积分公式-格林函数法}$

$1.(三维)狄氏积分公式:M,M_0 \in \tau$
$\left \lbrace \begin{array}{l}\Delta u = -h(M) \quad (1) \\ u|_{\sigma} = f(M) \quad (2) \end{array} \right. \quad \left \lbrace \begin{array}{l}\Delta G = -\delta(M-M_0) \quad (3) \\ G|_{\sigma} = 0 \quad (4) \end{array} \right.$
$则(3) \to G = \dfrac{1}{4 \pi r} (10.3.4), r= \sqrt{(x-x_0)^2 +(y-y_0)^2 + (z-z_0)^2}$
$[(1)\cdot G - (3) \cdot u]在[\tau - \tau_{\varepsilon}]中积分有:$
$\int_{\tau - \tau_{\varepsilon}}[G\Delta u - u \Delta G d \tau = \int_{\tau - \tau_{\varepsilon}} u \delta(M-M_0) d \tau - \int_{\tau - \tau_{\varepsilon}} Gh(M) d \tau$
$对左边用格林第二公式有:$
$\int_{\sigma + \sigma_{\varepsilon}}(G(\dfrac{\partial u}{\partial n} - u \dfrac{\partial G}{\partial n}) d \sigma = \int_{\tau - \tau_{\varepsilon}} u \delta(M-M_0) d \tau - \int_{\tau - \tau_{\varepsilon}} Gh(M) d \tau$
$\int_{\sigma}(G\dfrac{\partial u}{\partial n} - u\dfrac{\partial G}{\partial n}) d \sigma - u(M_0) = -\int_{\tau} G(M, M_0)h(M) d \tau$
$将边界条件(2)(4)代入上式有:$
$u(M_0) = \int_{\tau}G(M, M_0)h(M)d \tau - \int_{\sigma}f(M) \dfrac{\partial G}{\partial n} d \sigma$
$\color{blue}{u(M) = \int_{\tau}G(M, M_0)h(M_0)d \tau_0 - \int_{\sigma}f(M_0)\dfrac{\partial G}{\partial n_0} d \sigma_0}$
$\because G(M, M_0) = G(M_0, M)$

$2.狄氏积分公式的物理意义:$
$第一项:体内源产生的场的和。$
$第二项:边界上源产生的场的和。$

$3.(二维)狄氏积分公式$
$\left \lbrace \begin{array}{l}\Delta u(M) = -h(M) \quad M \in \sigma \\ u|_l = f(M) \end{array} \right.$
$u(M) = \iint_{\sigma}G(M, M_0)h(M_0)d \sigma_0 - \int_l f(M_0) \dfrac{\partial G}{\partial n_0} dl_0$

$\color{blue}{10.2.3 小结}$

$1.\int_{\tau}u\Delta v d\tau - \int_{\tau}v \Delta u d \tau = \int_{\sigma}(u\dfrac{\partial v}{\partial n} - v \dfrac{\partial u}{\partial n}) d \sigma \quad (5)$
$2.\left \lbrace \begin{array}{l}\Delta u = -h(M), M \in \tau \quad (1) \\ u|_{\sigma} = f(M) \quad (2) \end{array} \right.$
$u(M) = \iiint_{\tau} G(M, M_0)h(M_0) d\tau_0 - \iint_{\sigma} f(M_0)\dfrac{\partial G}{\partial n_0} d \sigma_0 \quad (6)$
$\left \lbrace \begin{array}{l}\Delta G = -\delta(M-M_0) \quad (3) \\ G|_{\sigma} = 0 \quad (4) \end{array} \right. G(M, M_0) = G(M_0, M)$

$\color{blue}{\S 10.3 格林函数}$

$\color{blue}{10.3.1 泊松方程的格林函数}$

$1.三维:\Delta G = -\delta(M-M_0)$
$\Delta G = \dfrac{1}{r^2}\dfrac{\partial}{\partial r}(r^2 \dfrac{\partial G}{\partial r} = -\delta(r)$
$(1)若r \neq 0: G = -C_1 \dfrac{1}{r}$
$(2)若r = 0: 考虑\iiint_{\tau_{\varepsilon}}\Delta G dv = -\iiint_{\tau_{\varepsilon}}\delta(r) dv = -1$
$又\iiint_{\tau_{\varepsilon}}\Delta G dv = \iiint_{\tau_{\varepsilon}} \nabla \cdot \nabla G dv = \iint_{\sigma_{\varepsilon}} \nabla G \cdot d \vec \sigma = C_1 4 \pi$
$\to C_1 = -\dfrac{1}{4 \pi},$
$(1),(2) \to \color{blue}{G = \dfrac{1}{4 \pi r} - 泊松方程格林函数}$

$2.二维:\Delta G = - \delta(M-M_0)$
$\Delta G = \dfrac{1}{r}\dfrac{\partial}{\partial r}(r\dfrac{\partial G}{\partial r}) = - \delta(r), r = \sqrt{(x-x_0)^2 + (y-y_0)^2}$
$\iint_{\sigma}\nabla \cdot \nabla u d \sigma = \int_l \nabla u \cdot d \vec l$
$\color{blue}{G(M, M_0) = \dfrac{1}{2 \pi} \ln \dfrac{1}{r}}$

$\color{blue}{10.3.2 狄氏格林函数}$

$1.三维：$
$\left \lbrace \begin{array}{l}\Delta G = - \delta(x-x_0, y-y_0, z-z_0), M \in \tau \\ G|_{\sigma} = 0 \end{array} \right.$
$令G(M, M_0) = F(M, M_0) + g(M, M_0)$
$使\Delta F(M, M_0) = -\delta(M - M_0) \quad M \in \tau, 则$
$\color{blue}{G(M, M_0) = \dfrac{1}{4 \pi r} + g -狄氏格林函数}$
$\color{blue}{\left \lbrace \begin{array}{l}\Delta g = 0, M \in \tau \\ g|_{\sigma} = -\dfrac{1}{4 \pi r}|_{\sigma} \end{array} \right.}$

$2.二维:$
$\left \lbrace \begin{array}{l}\Delta G = -\delta(x-x_0, y-y_0) \\ G|_l = 0 \end{array} \right.$
$\color{blue}{G = \dfrac{1}{2 \pi} \ln \dfrac{1}{r} + g -狄氏格林函数}$
$\color{blue}{\left \lbrace \begin{array}\Delta g = 0, M \in \sigma \\ g|_l = -\dfrac{1}{2 \pi} \ln \dfrac{1}{r} |_l \end{array} \right.}$

$3.狄氏格林函数的物理意义:$
$\left \lbrace \begin{array}{l}\Delta G = - \delta(M - M_0), M \in \tau \\ G | _{\sigma} = 0 \end{array} \right.$
$G(M, M_0) = \dfrac{1}{4 \pi r} + g \quad \left \lbrace \begin{array}{l} \Delta g = 0, M \in \tau \\ g |_{\sigma} = -\dfrac{1}{4 \pi r} | _{\sigma} \end{array} \right.$
$G-M点点位\left \lbrace \begin{array}{l}\varepsilon_0提供:\dfrac{1}{4 \pi \varepsilon_0} \dfrac{\varepsilon_0}{r} = \dfrac{1}{4 \pi r} \\ 感应电荷提供v:\left \lbrace \begin{array}{l}\Delta v = 0, M \in \tau \because v = g \\ v|_{\sigma} = -\dfrac{1}{4 \pi r} | _{\sigma} \end{array} \right. \end{array} \right.$
$求G \to 求M点点位 \to 求感应电荷产生的点位$

$对于三维:即求:\left \lbrace \begin{array}{l}\Delta g = 0, M \in \tau \\ g|_{\sigma} = -\dfrac{1}{4 \pi r} | _{\sigma} \end{array} \right.$
$对于二维:即求:\left \lbrace \begin{array}{l}\Delta g = 0, M \in \sigma \\ g|_l = - \dfrac{1}{2 \pi} \ln \dfrac{1}{r} | _l \end{array} \right.$

$\color{blue}{10.3.3 用电像法求狄氏格林函数}$

$1.问题引入:$
$求解球内狄氏问题:\left \lbrace \begin{array}{l}\Delta u = 0, \rho < a \\ u|_{\rho = a} = f(M) \end{array} \right.$
$解:u(M) = -\iint_{\sigma} f(M_0) \dfrac{\partial G}{\partial n_0} d \sigma_0$
$G(M, M_0) = \dfrac{1}{4 \pi r} + g; \left \lbrace \begin{array}{l}\Delta g = 0, M \in \rho < a \\ g|_{\rho = a} = - \dfrac{1}{4 \pi r} | _{\rho = a} \end{array} \right.$
$求u \to 求G \to 求M点电位 \to 求感应电荷产生的点位g$

$2.用电像法求g:$
$(1)分析:若能在\sigma外的某点M_1放一适当的负q,则$
$\Delta(\dfrac{-q}{4 \pi \varepsilon_0 r_1}) = 0, M \in \rho < a$
$使:-\dfrac{q}{4 \pi \varepsilon_0 r_1} |_{\rho = a} = -\dfrac{1}{4 \pi r} | _{\rho = a}$
$则g = -\dfrac{q}{r \pi \varepsilon_0 r_1}$
$\therefore 求g \to a)确定M_1的位置; b)确定q大小问题$

$(2)求球域的G$
$a) r_1 = ?记|OM_0| = \rho_0, |OM_1| = \rho_1,使\rho_0 \cdot \rho_1 = a^2, 即\dfrac{\rho_0}{a} = \dfrac{a}{\rho_1}$
$则称M_1为M_0关于球面\rho = a的像$
$b)q = ? \dfrac{1}{r}|_{\sigma} = ?$
$\because \Delta OM_0M \sim \Delta OM_1M$
$\therefore \dfrac{\rho_0}{a} = \dfrac{a}{\rho_1} = \dfrac{r}{r_1}, 即\dfrac{1}{r}|_{\rho = a} = \dfrac{a/\rho_0}{r_1} | _{\rho = a}$
$g = \dfrac{-\varepsilon_0 a/ \rho_0}{4 \pi \varepsilon_0 r_1} = \dfrac{-a/\rho_0}{4 \pi r_1}, \quad q = \dfrac{\varepsilon_0 a}{\rho_0}$
$G = \dfrac{1}{4 \pi r} - \dfrac{a/\rho_0}{4 \pi r_1}, -q = -\dfrac{\varepsilon_0 a}{\rho_0}是\varepsilon_0的电像$

$(3)电像法:这种在像点放一虚构的点电荷，来等效代替边界面上的感应电荷所产生的\\ 点位的方法称之为电像法。$

$3.求u(M)$
$\dfrac{\partial G}{\partial n} = \dfrac{1}{4 \pi}[\dfrac{\partial}{\partial \rho}(\dfrac{1}{r}) - \dfrac{a}{\rho_0}\dfrac{\partial}{\partial \rho}(\dfrac{1}{r_1})]$
$r = \sqrt{\rho^2 + \rho_0^2 - 2 \rho \rho_0 \cos \gamma} \quad r_1 = \sqrt{\rho^2 + \rho_1^2 - 2 \rho \rho_1 \cos \gamma}$
$\dfrac{\partial}{\partial \rho}\dfrac{1}{r} = \dfrac{\rho_0^2 - \rho^2 - r^2}{2 \rho r^3} \to \dfrac{\partial}{\partial \rho}(\dfrac{1}{r})_{\rho=a} = \dfrac{\rho_0^2 - a^2 - r^2}{2a r^3}$
$类似有:\dfrac{a}{\rho_0}\dfrac{\partial}{\partial \rho}(\dfrac{1}{r_1})_{\rho=a} = \dfrac{a}{\rho_0} \dfrac{\rho_1^2 - a^2 - r_1^2}{2ar_1^3} = \dfrac{a^2 - r^2 - \rho_0^2}{2ar^3}$
$\dfrac{\partial G}{\partial n}|_{\rho = a} = \dfrac{1}{4 \pi a} \dfrac{\rho_0^2 - a^2}{r^3}$

$u(M) = -\iint_{\sigma} f(M_0) \dfrac{\partial G}{\partial n_0} d \sigma_0$
$= \dfrac{1}{4 \pi a} \int_0^{2 \pi} \int_0^{\pi}f(\theta_0, \varphi_0) \dfrac{a^2 - \rho^2}{(a^2 + \rho^2 - 2a \rho \cos \gamma)^{\frac{3}{2}}}a^2 \sin \theta_0 d \theta_0 d \varphi_0$

$\color{blue}{u(M) = \dfrac{a}{4 \pi}\int_0^{2 \pi} \int_0^{\pi} f(\theta_0, \varphi_0) \dfrac{a^2 - \rho^2}{(a^2 + \rho^2 - 2a \rho \cos \gamma)^{\frac{3}{2}}} \sin \theta_0 d \theta_0 d \varphi_0 -球的泊松积分公式}$
$其中，\cos \gamma = \sin \theta \sin \theta_0 \cos(\varphi - \varphi_0) + \cos \theta \cos \theta_0$

$\color{blue}{10.3.4 注释}$

$1.\cos \gamma = ?$
$设\vec I为\vec{OM}方向单位向量,\vec{I_0}为\vec{OM_0}方向单位向量, 则$
$\vec I = x \vec i + y \vec j + z \vec k = \sin \theta \cos \varphi \vec i + \sin \theta \sin \varphi \vec j + \cos \theta \vec k$
$\vec{I_0} = x_0 \vec i + y_0 \vec j + z_0 \vec k = \sin \theta_0 \cos \varphi_0 \vec i + \sin \theta_0 \sin \varphi_0 \vec j + \cos \theta_0 \vec k$
$\therefore \vec I \cdot \vec I_0 = |\vec I \cdot \vec I_0| \cos \gamma = cos \gamma$
$=\sin \theta \cos \varphi \sin \theta_0 \cos \varphi_0 + \sin \theta \sin \varphi \sin \theta_0 \sin \varphi_0 + \cos \theta \cos \theta_0$

$2.对于球外的狄氏问题:$
$\left \lbrace \begin{array}{l}\Delta u_1 = 0, \rho > a \\ u_1|_{\rho = a} = f(M) \end{array} \right.$
$u_1(M) = -u(M)$
$\color{blue}{u_1(M) = \dfrac{-a}{4 \pi}\int_0^{2 \pi} \int_0^{\pi} f(\theta_0, \varphi_0) \dfrac{a^2 - \rho^2}{(a^2 + \rho^2 - 2a \rho \cos \gamma)^{\frac{3}{2}}} \sin \theta_0 d \theta_0 d \varphi_0}$

$3.对于圆内的狄氏问题:$
$G = \dfrac{1}{2 \pi} \ln \dfrac{1}{r} +g$
$\left \lbrace \begin{array}{l}\Delta u = 0, \rho < a \\ u|_{\rho=a} = f(\varphi) \end{array} \right. \quad \left \lbrace \begin{array}{l}\Delta g = 0, M \in \sigma \\ g|_{\rho = a} = -\dfrac{1}{2 \pi} \ln \dfrac{1}{r} |_{\rho = a} \end{array} \right.$
$\color{blue}{u(M) = \dfrac{1}{2 \pi}\int_0^{2\pi} f(\varphi_0) \dfrac{a^2 - \rho^2}{a^2 + \rho^2 - 2a \rho \cos(\varphi - \varphi_0)} d \varphi_0}$

$4.对于圆外的狄氏问题:$
$\left \lbrace \begin{array}{l}\Delta u_1 = 0, \rho > a \\ u_1 |_{\rho = a} = f(\varphi) \end{array} \right. \quad u_1(M) = -u(M)$
$\color{blue}{u(M) = \dfrac{1}{2 \pi}\int_0^{2\pi}\int_0^{2 \pi} f(\varphi_0)\dfrac{\rho^2 - a^2}{a^2 + \rho^2 - 2a\rho \cos(\varphi - \varphi_0)} d \varphi_0}$

$\color{blue}{10.3.5 小结}$

$\Delta G = -\delta(M - M_0) \to G = \dfrac{1}{4 \pi r}$
$\left \lbrace \begin{array}{l} \Delta G = -\delta(M-M_0), M \in \tau \\ G|_{\sigma} = 0 \end{array} \right. \to G(M, M_0) = \dfrac{1}{4 \pi r} + g$
$\rho < a: G = \dfrac{1}{4 \pi r} - \dfrac{a/\rho_0}{4 \pi r_1}$
$\left \lbrace \begin{array}{l}\Delta g = 0 \\ g|_{\sigma} = -\dfrac{1}{4 \pi r}|_{\sigma} \end{array} \right.$
$\left \lbrace \begin{array}{l}\Delta u = 0, \rho < a \\ u|_{\rho = a} =f(M) \end{array} \right.$
$u(M) = \dfrac{a}{4 \pi}\int_0^{2\pi}\int_0^{\pi}f(\theta_0, \varphi_0)\dfrac{a^2 - \rho^2}{(a^2 + \rho^2 - 2a \rho \cos \gamma)^{\frac{3}{2}}}\sin \theta_0 d \theta_0 d \varphi_0$