zoukankan      html  css  js  c++  java
  • 拉普拉斯算子的极坐标、柱坐标和球坐标表示

    documentclass{article}

    usepackage{amsmath}

    usepackage{amsthm}

    usepackage{amsfonts}

    usepackage{ctex}

    usepackage{mathrsfs}

    egin{document}

    section{极坐标变换下的Laplace算子}

    对于函数$u=u(x,y)$,其中$(x,y)in D_{xy}subseteq mathbb R^2ackslash{(0,0)}$,构造极坐标变换

    egin{equation}

    x=r cos heta ,

    end{equation}

    egin{equation}

    y=rsin heta,

    end{equation}

    其中$(r, heta)in D_{r heta}subseteq(0,+infty ) imesleft[ 0,2pi ight),$计算得雅可比行列式

    $dfrac{partial (x,y)}{partial (r, heta)}=r>0$,因此(1)式和(2)式表示一个双射$T: (r, heta )mapsto (x,y)$,从而映射(函数)$(r, heta)mapsto u$存在。为了方便,我们还假设$u$是二阶连续可微的,使得$u$对$x$,$y$的混合偏导数与求导顺序无关。

     

    由(1)(2)易得$r^2=x^2+y^2$,两边对变量$x$求导得$rdfrac{partial r}{partial x}=x$,所以$dfrac{partial r}{partial x}=dfrac{x}{r}$,同理可得$dfrac{partial r}{partial y}=dfrac{y}{r}$。由(1)(2)也易得$xsin heta=ycos heta$,两边对变量$x$求导得$sin heta+xdfrac{partial heta}{partial x}cos heta=-ydfrac{partial heta}{partial x}sin heta$,即$dfrac{partial heta}{partial x}=-dfrac{sin heta}{xcos heta+ysin heta}$,为了使得表达式简洁,我们在分子分母都乘以非零的$r$并将(1)(2)分别代入式中的$x$,$y$得$dfrac{partial heta}{partial x}=-dfrac{y}{r^2}$,同理可得$dfrac{partial heta}{partial y}=dfrac{x}{r^2}$.

     

    为了计算$ abla^2 u$,用链式法则先求对变量$x$的一阶偏导数并代入上面的结论和化简得

    [dfrac{partial u}{partial x}=dfrac{partial u}{partial r}dfrac{partial r}{partial x}+dfrac{partial u}{partial heta}dfrac{partial heta}{partial x}=dfrac{partial u}{partial r}dfrac{x}{r}-dfrac{partial u}{partial heta}dfrac{y}{r^2},]

    [dfrac{partial u}{partial y}=dfrac{partial u}{partial r}dfrac{partial r}{partial y}+dfrac{partial u}{partial heta}dfrac{partial heta}{partial y}=dfrac{partial u}{partial r}dfrac{y}{r}+dfrac{partial u}{partial heta}dfrac{x}{r^2},]

    运用求导的乘积法则和链式法则得

    egin{equation}

    dfrac{partial^2 u}{partial x^2}=dfrac{partial}{partial x}left(dfrac{partial u}{partial x} ight)=left(dfrac{partial^2 u}{partial r^2}dfrac{partial r}{partial x}+dfrac{partial^2 u}{partial rpartial heta}dfrac{partial heta}{partial x} ight)dfrac{partial r}{partial x}+dfrac{partial u}{partial r}dfrac{partial^2 r}{partial x^2}+left(dfrac{partial^2 u}{partial rpartial heta}dfrac{partial r}{partial x}+dfrac{partial^2 u}{partial heta^2}dfrac{partial heta}{partial x} ight)dfrac{partial heta}{partial x}+dfrac{partial u}{partial heta}dfrac{partial^2 heta}{partial x^2},

    end{equation}

    将$x$换成$y$得

    egin{equation}

    dfrac{partial^2 u}{partial y^2}=dfrac{partial}{partial y}left(dfrac{partial u}{partial y} ight)=left(dfrac{partial^2 u}{partial r^2}dfrac{partial r}{partial y}+dfrac{partial^2 u}{partial rpartial heta}dfrac{partial heta}{partial y} ight)dfrac{partial r}{partial y}+dfrac{partial u}{partial r}dfrac{partial^2 r}{partial y^2}+left(dfrac{partial^2 u}{partial rpartial heta}dfrac{partial r}{partial y}+dfrac{partial^2 u}{partial heta^2}dfrac{partial heta}{partial y} ight)dfrac{partial heta}{partial y}+dfrac{partial u}{partial heta}dfrac{partial^2 heta}{partial y^2}.

    end{equation}

    注意由于假设二阶连续可微,所以两个混合偏导数用同一个记号表示。下面计算所需的四个二阶偏导数。$dfrac{partial r}{partial x}=dfrac{x}{r}$的两边求对变量$x$的偏导数得[dfrac{partial^2 r}{partial x^2}=dfrac{r^2-rxdfrac{partial r}{partial x}}{r^3}=dfrac{y^2}{r^3},]

    这里利用了一个小技巧,为了能够通过关系式$r^2=x^2+y^2$来化简最后的结果,分子分母同时乘以非零的$r$。把$x$和$y$互换即得[dfrac{partial^2 r}{partial y^2}=dfrac{x^2}{r^3};]

    $dfrac{partial heta}{partial x}=-dfrac{y}{r^2}$的两边求对变量$x$的偏导数得[dfrac{partial^2 heta}{partial x^2}=dfrac{2y}{r^3}dfrac{partial r}{partial x}=frac{2xy}{r^4}],同理可得

    [dfrac{partial^2 heta}{partial y^2}=-frac{2xy}{r^4},]

    将这些结果代入(3)(4)式得

    [dfrac{partial^2 u}{partial x^2}=dfrac{1}{r^4}left[ r^2dfrac{partial^2 u}{partial r^2}x^2+left(rdfrac{partial u}{partial r}+dfrac{partial^2 u}{partial heta^2} ight)y^2+2left(dfrac{partial u}{partial heta}-rdfrac{partial^2 u}{partial rpartial heta} ight)xy ight],]

    [dfrac{partial^2 u}{partial y^2}=dfrac{1}{r^4}left[ left(rdfrac{partial u}{partial r}+dfrac{partial^2 u}{partial heta^2} ight)x^2+r^2dfrac{partial^2 u}{partial r^2}y^2+2left(rdfrac{partial^2 u}{partial rpartial heta}-dfrac{partial u}{partial heta} ight)xy ight],]

    两式相加得

    [dfrac{partial^2 u}{partial x^2}+dfrac{partial^2 u}{partial y^2}=dfrac{1}{r^4}left(r^2dfrac{partial^2 u}{partial r^2}+rdfrac{partial u}{partial r}+dfrac{partial^2 u}{partial heta^2} ight)left(x^2+y^2 ight),]

    注意到$r^2=x^2+y^2$,所以在极坐标变换下二维Laplace算子的表达式为

    [oxed{ abla^2 =dfrac{partial^2 }{partial r^2}+dfrac{1}{r}dfrac{partial }{partial r}+dfrac{1}{r^2}dfrac{partial^2 }{partial heta^2}}]

    section{柱面坐标变换下的Laplace算子}

    函数$u=u(x,y,z)$的柱面坐标变换是指

    [x= ho cosphi,]

    [y= ho sinphi,]

    [z=z,]

    于是由上一节的讨论易得Laplace算子在柱面坐标变换下的表示为

    [oxed{ abla^2 =dfrac{partial^2 }{partial ho^2}+dfrac{1}{ ho}dfrac{partial }{partial ho}+dfrac{1}{ ho^2}dfrac{partial^2 }{partial phi^2}+dfrac{partial^2 }{partial z^2}}]

    section{球面坐标变换下的Laplace算子}

    对于函数$u=u(x,y,z)$,其中$(x,y,z)in Omega_{xyz}subseteq mathbb R^3ackslash{(0,0,0)}$,构造球面坐标变换

    [x=rsin hetacosphi,]

    [y=rsin hetasinphi,]

    [z=rcos heta,]

    其中$(r, heta,phi)in Omega_{r hetaphi}subseteq(0,+infty) imes[0,pi] imes[0,2pi)$.类似于极坐标变换的情形,以上三式给出了一个双射$S:(r, heta,phi)mapsto (x,y,z)$,从而函数$(r, heta,phi)mapsto u$存在。为了方便我们仍然假设$u$是二阶连续可微的。

     

    为了能够利用第一节的结论,我们令$ ho=rsin heta$(这实际上是引入了柱面坐标$( ho,phi,z)$),于是$x= ho cosphi$,$y= hosinphi$,对$dfrac{partial^2 u}{partial x^2}+dfrac{partial^2 u}{partial y^2}$可以利用极坐标变换的结论得

    egin{equation}

    dfrac{partial^2 u}{partial x^2}+dfrac{partial^2 u}{partial y^2}=dfrac{partial^2 u }{partial ho^2}+dfrac{1}{ ho}dfrac{partial u}{partial ho}+dfrac{1}{ ho^2}dfrac{partial^2 u}{partial phi^2}

    end{equation}

    同理,由$ ho=rsin heta$,$z=rcos heta$得

    egin{equation}

    dfrac{partial^2 u}{partial ho^2}+dfrac{partial^2 u}{partial z^2}=dfrac{partial^2 u }{partial r^2}+dfrac{1}{r}dfrac{partial u}{partial r}+dfrac{1}{r^2}dfrac{partial^2 u}{partial heta^2}

    end{equation}

    (5)和(6)相加得

    egin{equation}

    dfrac{partial^2 u}{partial x^2}+dfrac{partial^2 u}{partial y^2}+dfrac{partial^2 u}{partial z^2}=dfrac{1}{ ho}dfrac{partial u}{partial ho}+dfrac{1}{ ho^2}dfrac{partial^2 u}{partial phi^2}+dfrac{partial^2 u }{partial r^2}+dfrac{1}{r}dfrac{partial u}{partial r}+dfrac{1}{r^2}dfrac{partial^2 u}{partial heta^2}

    end{equation}

    我们还需计算$dfrac{partial u}{partial ho}$,注意到$(z, ho)mapsto (r, heta)$是一个极坐标变换,于是由第一节的讨论可知

    [dfrac{partial u}{partial ho}=dfrac{partial u}{partial r}dfrac{ ho}{r}+dfrac{partial u}{partial heta}dfrac{z}{r^2},]

    将$z=rcos heta$和$ ho=rsin heta$代入得

    [dfrac{partial u}{partial ho}=dfrac{partial u}{partial r}sin heta+dfrac{1}{r}dfrac{partial u}{partial heta}cos heta,]

    代入(7)式得到Laplace算子在球面坐标变换下的表示为

    [oxed{ abla^2=dfrac{partial^2}{partial r^2}+dfrac{2}{r}dfrac{partial}{partial r}+dfrac{1}{r^2}left(dfrac{partial^2}{partial heta^2}+cot hetadfrac{partial}{partial heta}+csc^2 heta dfrac{partial^2}{partial phi^2} ight)}]

    end{document}

  • 相关阅读:
    关于jar包的创建及引用相关问题
    关于Android系统一次能创建多大的Bitmap?
    Android开发导入第三方Jar包
    Javascript操作表格
    installanywhere 打包j2ee的方法II(转)
    C#网页自动登录和提交POST信息的多种方法
    JavaScript手册javascript语法javascript函数
    javascript 禁止复制网页
    [Android实例] Handler+ExecutorService(线程池)+MessageQueue模式+缓存模式(转)
    android消息处理系统原理
  • 原文地址:https://www.cnblogs.com/Eufisky/p/15005858.html
Copyright © 2011-2022 走看看