Undergrad Gravitational potential of an ellipsoid

Click For Summary
The discussion revolves around the derivation of the potential formula outside a homogeneous ellipsoid of density, as presented in Landau's work. The potential is expressed through an integral involving the ellipsoid's semi-axes and density, with specific conditions for the variable ##\xi##. A multipole expansion is utilized to express the potential at a point, incorporating the mass quadrupole tensor. There is a concern regarding the signs in the original formula, suggesting a possible error that aligns with discrepancies found in a referenced document. The conversation concludes with a verification of the derivation from a reputable source.
ergospherical
Science Advisor
Homework Helper
Education Advisor
Insights Author
Messages
1,100
Reaction score
1,387
There is a formula for the potential ##\varphi## outside of a homogenous ellipsoid of density ##\mu## in Landau\begin{align*}
\varphi = -\pi \mu abck \int_{\xi}^{\infty} \left(1- \dfrac{x^2}{a^2 + s} + \dfrac{y^2}{b^2 + s} + \dfrac{z^2}{c^2+s} \right) \frac{ds}{R_s} \ \ \ (1)
\end{align*}where ##R_s = \sqrt{(a^2+s)(b^2+s)(c^2+s)}## and ##\xi## satisfies ## \dfrac{x^2}{a^2 + \xi} + \dfrac{y^2}{b^2 + \xi} + \dfrac{z^2}{c^2+\xi} = 1##. How is this formula obtained?

So far I can only find an expression in the limit of ##r \gg a,b,c##. The mass quadrupole tensor is ##Q_{\alpha \beta} = \displaystyle{\int_{\mathcal{V}}} \mu(3x_{\alpha} x_{\beta} - r^2 \delta_{\alpha \beta}) dV## therefore the potential at a point ##\mathbf{r}## has a multipole expansion ##\varphi(\mathbf{r}) = - \dfrac{km}{r} + \dfrac{1}{6} Q_{\alpha \beta} \partial^2_{\alpha \beta} \dfrac{1}{r} + \mathrm{etc}##. Due to the axial symmetry, the matrix ##Q## can be brought to diagonal form by aligning the coordinate system with the principal axes and the non-zero components are\begin{align*}
Q_{xx} &= \mu \int_{\mathcal{V}} d^3 x (2x^2 - y^2 - z^2)\\
&= \mu abc \int_0^{2\pi} d\phi \int_0^{\pi} d\theta \int_0^1 dr\, r^4 ([2a^2 \cos^2{\phi} - b^2 \sin^2{\phi}]\sin^3{\theta} - c^2 \cos^2{\theta} \sin{\theta} ) \\
&= \dfrac{\mu abc}{5} \int_0^{2\pi} d\phi \int_0^{\pi} d\theta \, ([2a^2 \cos^2{\phi} - b^2 \sin^2{\phi}]\sin^3{\theta} - c^2 \cos^2{\theta} \sin{\theta} ) \\
&= \dfrac{2\mu abc}{15} \int_0^{2\pi} d\phi \, (4a^2 \cos^2{\phi} - 2b^2 \sin^2{\phi} - c^2 ) \\
&= \dfrac{4\pi \mu abc}{15} (2a^2 - b^2 - c^2 )
\end{align*}where the "spherical-like" coordinate transformations ##x = ar\sin{\theta} \cos{\phi}##, etc. map the ellipsoid into the unit sphere. Similarly ##Q_{yy} = \dfrac{4\pi \mu abc}{15} (2b^2 - a^2 - c^2 )## and ##Q_{zz} = \dfrac{4\pi \mu abc}{15} (2c^2 - a^2 - b^2 )##. Now\begin{align*}
\partial_{\beta} \dfrac{1}{r} &= - \dfrac{1}{r^2} \cdot \dfrac{1}{2r} \partial_{\beta} r^2 = -\dfrac{1}{2r^3} \partial_{\beta} (x_{\gamma} x_{\gamma}) = -\dfrac{x_{\beta}}{r^3} \\ \\
\implies \partial^2_{\alpha \beta} \dfrac{1}{r} &= -\partial_{\alpha} \dfrac{x_{\beta}}{r^3} = -\frac{1}{r^3} \delta_{\alpha \beta} + \frac{3x_{\alpha} x_{\beta}}{r^5} = \frac{3x_{\alpha} x_{\beta} - r^2 \delta_{\alpha \beta}}{r^5}
\end{align*}therefore ##\partial^2_{xx} \dfrac{1}{r} = \dfrac{2x^2 - y^2 -z^2}{r^5}## and etc. therefore \begin{align*}
\varphi(\mathbf{r}) &= -\dfrac{km}{r} + \dfrac{2\pi \mu abc}{45 r^5} \left\{ (2a^2 - b^2 - c^2 )(2x^2 - y^2 -z^2) + \mathrm{y \ \ first} + \mathrm{z \ \ first} \right\} \\
&= -\dfrac{km}{r} + \dfrac{2\pi \mu abc}{45 r^5} \left\{ 6(a^2 x^2 + y^2b^2 + c^2 z^2) -3((b^2+c^2)x^2 + (a^2+c^2)y^2 + (a^2+b^2)z^2) \right\}
\end{align*}I haven't checked yet if this is consistent with ##(1)## as ##r## gets very big, but I'm more interested to know how Landau derived the exact expression?
 
  • Like
Likes LCSphysicist and vanhees71
Physics news on Phys.org
Golly, that was fast! Looks like exactly it, thanks.
 
Is it possible that it should be (minus signs wrong in eq 1 from ergospherical):

## \varphi = -\pi \mu abc \int_{\xi}^{\infty} \left(1- \dfrac{x^2}{a^2 + s} - \dfrac{y^2}{b^2 + s} - \dfrac{z^2}{c^2+s} \right) \frac{ds}{R_s} ##

The same error is present in equation (72) the document from Stanford above, but not in equation (3).
 
Topic about reference frames, center of rotation, postion of origin etc Comoving ref. frame is frame that is attached to moving object, does that mean, in that frame translation and rotation of object is zero, because origin and axes(x,y,z) are fixed to object? Is it same if you place origin of frame at object center of mass or at object tail? What type of comoving frame exist? What is lab frame? If we talk about center of rotation do we always need to specified from what frame we observe?

Similar threads

  • · Replies 2 ·
Replies
2
Views
1K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 1 ·
Replies
1
Views
470
  • · Replies 42 ·
2
Replies
42
Views
5K
  • · Replies 33 ·
2
Replies
33
Views
3K
  • · Replies 9 ·
Replies
9
Views
2K
  • · Replies 6 ·
Replies
6
Views
2K
Replies
5
Views
1K
  • · Replies 3 ·
Replies
3
Views
575
  • · Replies 3 ·
Replies
3
Views
2K