数学物理方法 15 贝塞尔函数02

$\color{blue}{\S 15.3 其它柱函数}$

$\color{blue}{15.3.1 三类柱函数}$

$1.第一类柱函数-Bessel函数$
$(1)定义 J_{\pm v}(x) = \sum \limits_{k=0}^{\infty}\dfrac{(-1)^k}{k! \Gamma(\pm v + k + 1)} (\dfrac{x}{2})^{2k \pm v} \quad (1) -第一类柱函数$
$2.第二类柱函数-Neuman函数$
$(1)定义 N_v(x) = \dfrac{\cos v\pi J_v(x) - J_{-v}(x)}{\sin v\pi} \quad (2) -第二类柱函数$
$(2)无论v = n与否,J_v(x)与N_v(x)均为v阶的Bessel方程的线性无关的解.$
$通解:y_C(x) = A_vJ_v(x) + B_vN_v(x)$
$\because 当v \neq n：\dfrac{N_v(x)}{J_v(x)} = \dfrac{\cos v\pi J_v(x) - J_{-v}(x)}{\sin v \pi J_v(x)}$
$= \dfrac{1}{\sin v \pi}[\cos v \pi - \dfrac{J_{-v}(x)}{J_v(x)}] \neq 常数$
$当v = n:$
$① N_n(x) = \dfrac{1}{\pi}[\dfrac{\partial J_v(x)}{\partial v} - (-1)^n \dfrac{\partial J_{-v}(x)}{\partial v}]_{v=n} \quad (3)$
$②N_n(x) = \dfrac{2}{\pi}J_n(x) \ln \dfrac{x}{2} - \dfrac{1}{\pi} \sum \limits_{k=0}^{n-1}\dfrac{(n-k-1)!}{k!}(\dfrac{x}{2})^{2k-n}$
$-\dfrac{1}{\pi}\sum \limits_{k=0}^{\infty}\dfrac{(-1)^k}{k!(n+k)!}[\psi(k+1) + \psi(n+k+1)](\dfrac{x}{2})^{2k+n}$
$\psi(1) = -v = -0.577216, \psi(k+1) = -v + 1 + \dfrac{1}{2} + \cdots + \dfrac{1}{k}$
$③N_n(x)是n阶Bessel的解;因为由下可知(3)式满足贝塞尔方程.$
$x^2J_{\pm v}^{\prime \prime}(x) + xJ_{\pm v}^{\prime}(x) + (x^2 - v^2) J_{\pm v}(x) = 0 \quad (1)$
$x^2 \dfrac{d^2}{dx^2}\dfrac{\partial J_v}{\partial v} + x\dfrac{d}{dx} \dfrac{\partial J_v}{\partial v} + (x^2 - v^2) \dfrac{\partial J_v}{\partial v} - 2v J_v(x) = 0 \quad (2)$
$x^2 \dfrac{d^2}{dx^2}\dfrac{\partial J_{-v}}{\partial v} + x\dfrac{d}{dx}\dfrac{\partial J_{-v}}{\partial v} + (x^2 - v^2)\dfrac{\partial J_{-v}}{\partial v} - 2v J_{-v}(x) = 0 \quad (3)$
$[(2) - (-1)^n \cdot (3)]\dfrac{1}{\pi}:(v \to n)$
$x^2N_n^{\prime \prime}(x) + xN_n^{\prime}(x) + (x^2 - n^2) N_n(x) = 0 \quad (4)$

$①N_n(x) = \dfrac{1}{\pi}[\dfrac{\partial J_v(x)}{\partial v} - (-1)^n \dfrac{\partial J_{-v}(x)}{\partial v}]_{v=n} \quad (3)$

$④N_n(x)与J_n(x)线性无关;$
$because x \to 0: J_0(x) = \sum \limits_{k=0}^{\infty}\dfrac{(-1)^k}{(k!)^2}(\dfrac{x}{2})^{2k} \to 1,$
$J_n(x) = \sum \limits_{k=0}^{\infty} \dfrac{(-1)^k}{k!(n+k)!}(\dfrac{x}{2})^{2k+n} \to 0, n \geq 1$
$x \to 0:N_0(x) \approx \dfrac{2}{\pi}J_0(x) \ln \dfrac{x}{2} \approx \dfrac{2}{\pi} \ln \dfrac{x}{2} \to \infty,$
$N_n(x) \approx -\dfrac{(n-1)!}{\pi}(\dfrac{x}{2})^{-n} \to -\infty$

$3.第三类柱函数-Hankel函数$
$(1)定义\left \lbrace \begin{array}{l}H_v^{(1)}(x) = J_v(x) +iN_v(x) \\ H_v^{(2)}(x) = J_v(x) - iN_v(x) \end{array} \right. (4) -第三类柱函数$
$(2)无论v = n与否,H_v^{(1)}(x)与H_v^{(2)}(x)均为v阶的Bessel方程的线性无关的解.$

$4.三类柱函数的关系$
$(1)H_v^{(1)}(x),H_v^{(2)}(x),J_v(x), N_v(x)互相之间均线性无关.$
$(2)三类柱函数的关系为$
$\left \lbrace \begin{array}{l}H_v^{(1)}(x) = J_v(x) + iN_v(x) \\ H_v^{(2)}(x) = J_v(x) - iN_v(x) \\ J_v(x) = \dfrac{H_v^{(1)}(x) +H_v^{(2)}(x)}{2} \\ N_v(x) = \dfrac{H_v^{(1)}(x) - H_v^{(2)}(x)}{2i} \end{array} \right.$

$\color{blue}{15.3.2 球Bessel函数}$

$1.球Bessel方程及其解$
$\Delta u + \lambda u = 0 \stackrel{令u = R(r)\Theta(\theta)\Phi(\varphi)}{----\to}$
$\left \lbrace \begin{array}{l}\Phi^{\prime \prime} + m^2 \Phi = 0 \quad \Phi_m(\varphi) = A_m \cos m \varphi + B_m \sin m \varphi \\ \dfrac{1}{\sin \theta}\dfrac{d}{d \theta}(\sin \theta \dfrac{d \Theta}{d \theta}) + [l(l+1) - \dfrac{m^2}{\sin^2 \theta}]\Theta = 0 \quad \Theta(\theta) = p_l^m(\cos \theta) \\ r^2 R^{\prime \prime} + 2r R^{\prime} + [k^2r^2 - l(l+1)]R = 0 \quad (5) \end{array} \right.$
$x^2y^{\prime \prime} + 2xy^{\prime} + [x^2 -l(l+1)]y = 0 \quad (6)$
$x^2v^{\prime \prime} + xv^{\prime} + [x^2 - (l + \dfrac{1}{2})^2]v = 0 \quad (7)$
$(5)(6)(7)称为球Bessel方程$

$(7) \to v(x) = J_{l+\frac{1}{2}}(x), N_{l+\frac{1}{2}}(x), H_{l+\frac{1}{2}}^{(1)}(x), H_{l+\frac{1}{2}}^{(2)}(x),$
$(6) \to y(x) = \dfrac{1}{\sqrt{x}}J_{l+\frac{1}{2}}(x), \dfrac{1}{\sqrt{x}}N_{l+\frac{1}{2}}(x), \dfrac{1}{\sqrt{x}}H_{l+\frac{1}{2}}^{(1)}(x), \dfrac{1}{\sqrt{x}}H_{l+\frac{1}{2}}^{(2)}(x)$

$2.球Bessel函数$
$(1)定义\left \lbrace \begin{array}{l}j_l(x) = \sqrt{\dfrac{\pi}{2x}}J_{l+\frac{1}{2}}(x) - l阶球Bessel函数 \\ n_l(x) = \sqrt{\dfrac{\pi}{2x}}N_{l+\frac{1}{2}}(x) - l阶Neuman函数 \\ \left. \begin{array}{l}h_l^{(1)}(x) = \sqrt{\dfrac{\pi}{2x}}H_{l+\frac{1}{2}}^{(1)}(x) \\ h_l^{(2)}(x) = \sqrt{\dfrac{\pi}{2x}}H_{l+\frac{1}{2}}^{(2)}(x) \end{array} \right. -l阶球Hankel函数 \end{array} \right.$
$(2)j_l(x),n_l(x),h_l^{(1)}(x), h_l^{(2)}(x),均为球Bessel方程(6)的线性无关的解.$
$n_l(x) \stackrel{x = 0}{--\to}\infty$
$(6) \to y_c(x) = A_lJ_l(x) + B_ln_l(x)$

$3.球Bessel方程的本征值问题$
$\left \lbrace \begin{array}{l}r^2R^{\prime \prime} + 2r R^{\prime} + [k^2r^2 - l(l+1)]R = 0 \quad (5) \\ R(a) = 0, \quad x =kr, y(x) = R(r) \end{array} \right.$
$\left \lbrace \begin{array}{l}x^2y^{\prime \prime} + 2xy^{\prime} + [x^2 -l(l+1)]y = 0 \quad (6) \\ y(ka) = 0 \end{array} \right.$
$(k_m^{l+\frac{1}{2}})^2 = (\dfrac{x_m^{l+\frac{1}{2}}}{a})^2, R(r) = j_l(\dfrac{x_m^{l+\frac{1}{2})}}{a}r), m = 1, 2, \cdots$
$其中, \int_0^a j_l(\dfrac{x_m^{l+\frac{1}{2}}}{a}r)j_l(\dfrac{x_n^{(l+\frac{1}{2})}}{a}r)r^2 dr = \Vert j_n^l \Vert^2 \delta_{mn}$

$\Vert j_n^l \Vert^2 = \dfrac{\pi}{2k_n^{(l+\frac{1}{2})}} \Vert J_n^{(l+\frac{1}{2})} \Vert^2 = \dfrac{\pi}{2k_n^{(l+\frac{1}{2})}} \int_0^a J_{l+\frac{1}{2}}^2(k_n^{l+\frac{1}{2}} r) r dr$

$\color{blue}{15.3.3 虚宗量的Bessel函数}$

$1.虚宗量的Bessel方程及其解$
$\left. \begin{array}{r}\Delta u + \lambda u = 0 \\ \Delta u = 0 \end{array} \right \rbrace \stackrel{令u = R(\rho)\Phi(\varphi)Z(z)}{----\to}$
$\left \lbrace \begin{array}{l}\Phi^{\prime \prime} + n^2 \Phi = 0 \\ Z^{\prime \prime} + \mu Z = 0 \\ \rho^2 R^{\prime \prime} + \rho R^{\prime} + (k^2\rho^2 - n^2) R = 0 \quad (8) \to n阶Bessel方程 \end{array} \right.$
$x = k\rho, R(\rho) = y(x), n \to v,设k^2 = \lambda - \mu \geq 0$
$x^2 y^{\prime \prime}(x) + xy^{\prime}(x) + (x^2 - v^2)y(x) = 0 \quad (9) - n阶Bessel方程$
$若\left \lbrace \begin{array}{l}\lambda - \mu \\ -\mu \end{array} \right. < 0, 记\left. \begin{array}{r}\lambda - \mu \\ - \mu \end{array} \right \rbrace = -k^2$
$(9) \to x^2y^{\prime \prime}(x) + xy^{\prime}(x) - (x^2 + v^2) y(x) = 0 \quad (10) -虚宗量Bessel方程$
$令z = ix$
$z^2 y^{\prime \prime}(z) + zy^{\prime}(z) + (z^2 - v^2)y(z) = 0 \quad (11)$
$y = J_{\pm v}(ix), N_v(ix), H_v^{(1)}(ix), H_v^{(2)}(ix), - 虚宗量Bessel方程的解$

$2.虚宗量的柱函数$
$(1)定义\left \lbrace \begin{array}{l}I_v(x) = i^{-v}J_v(ix) \\ I_{-v}(x) = i^{v}J_{-v}(ix) \end{array} \right. -第一类虚宗量的柱函数(虚宗量Bessel函数)$
$①当v \neq n: y_C(x) = A_vI_v(x) + B_vI_{-v}(x)$
$②当v = n: I_{-n}(x) = I_n(x) \quad I_{-v}(x) \stackrel{x = 0}{--\to} \infty$

$(2)定义:K_v(x) = \dfrac{\pi}{2}\dfrac{I_{-v}(x) + I_v(x)}{\sin v \pi} - 第二类虚宗量的柱函数(Macdona函数)$
$(3)无论v = n与否,y_C(x) = c_vI_v(x) + d_vK_v(x) \quad K_v(x) \stackrel{x=0}{--\to}\infty$
$①当v \neq n: I_v(x)与K_v(x)线性无关;$
$②当v=n: K_n(x) = \dfrac{(-1)^n}{2}[\dfrac{\partial I_{-v}(x)}{\partial v} - \dfrac{\partial I_v(x)}{\partial v}]_{v=n}$

$\color{blue}{15.3.4 小结}$

$第二类柱函数:N_v(x) = \sum \limits_{k=0}^{\infty} \dfrac{\cos v \pi J_v(x) - J_{-v}(x)}{\sin v \pi J_v(x)}$
$第三类柱函数:\left \lbrace \begin{array}{l}H_v^{(1)}(x) = J_v(x) + i N_v(x) \\ H_v^{(2)}(x) = J_v(x) - iN_v(x) \end{array} \right.$

$球Bessel函数\left \lbrace \begin{array}{l}j_l(x) = \sqrt{\dfrac{\pi}{2x}}J_{l+\frac{1}{2}}(x), n_l(x) = \sqrt{\dfrac{\pi}{2x}}N_{l+\frac{1}{2}}(x), \\ h_l^{(1)}(x) = \sqrt{\dfrac{\pi}{2x}}H_{l+\frac{1}{2}}^{(1)}(x), h_l^{(2)}(x) = \sqrt{\dfrac{\pi}{2x}}H_{l+\frac{1}{2}}^{(2)}(x) \end{array} \right.$

$第一类虚宗量的柱函数\left \lbrace \begin{array}{l}I_v(x) = i^{-v}J_v(ix) \\ I_{-v}(x) = i^{v}J_{-v}(ix) \end{array} \right.$
$第二类虚宗量的柱函数: K_v(x) = \dfrac{\pi}{2} \dfrac{I_{-v}(x) + I_v(x)}{\sin v \pi}$

$x^2y^{\prime \prime}(x) + xy^{\prime}(x) + (x^2 - v^2)y(x) = 0 \quad (9)$
$\to y(x) = J_{\pm v}(x), N_v(x), H_{\pm v}(x)$
$x^2y^{\prime \prime} + 2xy^{\prime} + x^2 -l(l+1)]y = 0 \quad (6)$
$\to y(x) = j_l(x), n_l(x), h_l^{(1)}(x), h_l^{(2)}(x)$
$x^2y^{\prime \prime}(x) + xy^{\prime}(x) - (x^2 + v^2)y(x) = 0 \quad (10)$
$\to y(x) = I_v(x), K_v(x)$

$\color{blue}{15.3.5 例题}$

$1.一半径为a高为h的均匀圆柱体,其上、下底面保持温度为零度,而侧面温度为u_0, \\ 试求柱内的稳定温度分布.$
$\left \lbrace \begin{array}{l}\Delta u = 0, 0 \leq \rho \leq a \quad (1) \\ u(a, z) = u_0 \quad (2) \\ u(\rho, 0) = 0 \quad (3) \\ u(\rho, h) = 0 \quad (4) \end{array} \right.$
$解:1) 令u(\rho, z) = R(\rho)Z(z)$
$(1) \to \left \lbrace \begin{array}{l} Z^{\prime \prime} + \mu Z = 0 \quad (5) \\ \rho^2 R^{\prime \prime} + \rho R^{\prime} + (-\mu \rho^2 - 0)R = 0 \quad (6) \end{array} \right.$
$(3) \to Z(0) = 0 \quad (7) \quad (4) \to Z(h) = 0 \quad (8)$

$2)解本征值问题(5)(7)(8):$
$\left \lbrace \begin{array}{l}Z^{\prime \prime} + \mu Z = 0 \quad (5) \to \mu = \dfrac{m^2 \pi^2}{h^2}, m = 1, 2, \cdots \\ Z(0) = 0 \quad (7) \\ Z(h) = 0 \quad (8) \quad Z_m(z) = c_m\sin \dfrac{m \pi}{h} z \end{array} \right.$
$3)解关于\rho的方程(6):$
$\because -\mu = - \dfrac{m^2 \pi^2}{h^2} < 0, 故记 -\mu = -k^2, x = k\rho, y(x) = R(\rho)$
$(6) \to x^2y^{\prime \prime}(x) + xy^{\prime}(x) - (x^2 + 0)y(x) = 0$
$\to R_m(\rho) = y_m(x) = a_mI_0(k_m^0 \rho), k_m^0 = \dfrac{m \pi}{h}, m = 1, 2, \cdots$
$4)叠加，定系数:$
$u(\rho, z) = \sum \limits_{m=1}^{\infty} A_mI_0(k_m^0 \rho) \sin \dfrac{m \pi}{h} z$
$u_0 = \sum \limits_{m=1}^{\infty} A_mI_0(\dfrac{m \pi}{h}a) \sin \dfrac{m \pi}{h} z$
$A_mI_0(\dfrac{m \pi}{h}a) = \dfrac{2}{h}\int_0^{\pi} u_0 \sin \dfrac{n \pi}{h} z dz = \dfrac{2u_0}{m \pi}[1 - (-1)^m]$
$A_{2n} = 0, A_{2n+1} = \dfrac{4 u_0}{(2n+1) \pi I_0 (\dfrac{2n+1}{h} \pi a)}$
$u(\rho, z) = \dfrac{4 u_0}{\pi} \sum \limits_0^{\infty} \dfrac{\sin \dfrac{2n+1}{h} \pi \cdot I_0(\dfrac{2n+1}{h} \pi \rho)}{(2n+1)I_0(\dfrac{2n+1}{h}\pi a)}$

$2.半径为a的均匀导热介质球, 原来的温度为u_0, 将它放在冰水中, 使球面温度保持为零度,\\ 求球内温度的变化.$
$\left \lbrace \begin{array}{l}\dfrac{\partial u}{\partial t} - D \Delta u = 0 \quad (1) \\ u|_{r=0} \to 有限 \quad (2) \\ u|_{r=a} = 0 \quad (3) \\ u|_{t=0} = u_0 \quad (4) \end{array} \right.$
$解:1) 令u(r, t) = R(r)T(t) \quad (5)$
$\left \lbrace \begin{array}{l}T^{\prime}(t) +\lambda DT(t) = 0 \quad (6) \\ r^2 R^{\prime \prime}(r) + 2rR^{\prime}(r) + \lambda r^2 R(r) = 0 \quad (7) \\ R(0) \to 有限,R(a) = 0 \quad (8) \end{array} \right.$

$2)解本征值问题(7)(8):$
$\lambda = k^2 = (\dfrac{x_m^{0+\frac{1}{2}}}{a})^2, R(r) = j_0(\dfrac{x_m^{\frac{1}{2}}}{a}r), m = 1, 2, \cdots$
$\lambda = (\dfrac{m \pi}{a})^2, R_m(r) = \dfrac{a}{m \pi r} \sin \dfrac{m \pi r}{a}, m = 1, 2, \cdots$

$3)解关于t的方程(6):$
$T_m(t) = c_me^{-(\frac{m \pi}{a})^2 Dt}$

$4)叠加，定系数:$
$u(r, t) = \sum \limits_{m=1}^{\infty} c_m \dfrac{a}{m \pi r} \sin \dfrac{m \pi r}{a}e^{-(\frac{m \pi}{a})^2 Dt}$
$u(r, 0) = \sum \limits_{m=1}^{\infty} c_m \dfrac{a}{m \pi r} \sin \dfrac{m \pi r}{a} = \sum \limits{m=1}^{\infty} c_m j_0(\dfrac{m \pi r}{a}) = u_0$
$c_m \dfrac{1}{\Vert j_m^{(0)} \Vert^2} \int_0^a u_0 j_0(\dfrac{x_m^2}{a}r)r^2 dr = \dfrac{u_0 a}{\Vert j_m^{(0)} \Vert^2 m \pi } \int_0^{\pi} r \sin \dfrac{m \pi r}{a} dr = 2u_0 (-1)^{m-1}$

$u(r, t) = \dfrac{2au_0}{\pi r} \sum \limits_{m=1}^{\infty} \dfrac{(-1)^{m-1}}{m} \sin \dfrac{m \pi r}{a} e^{-(\frac{m \pi}{a})^2 Dt}$