Difference between revisions of "User:Tohline/Appendix/Ramblings/ConcentricEllipsodalT8Coordinates"
Line 967: | Line 967: | ||
<td align="left"> | <td align="left"> | ||
<math>~ | <math>~ | ||
-\frac{ | -\frac{\eta^{-1}}{(1+\eta^2)} \biggl[ - \frac{q^2 y}{x^2} \biggr] | ||
= | = | ||
\frac{x}{(x^2 + q^4y^2)} \, | \frac{x}{(x^2 + q^4y^2)} \, ; | ||
</math> | |||
</td> | |||
</tr> | |||
</table> | |||
and, | |||
<table border="0" cellpadding="5" align="center"> | |||
<tr> | |||
<td align="right"> | |||
<math>~\frac{\partial\lambda_3}{\partial y} </math> | |||
</td> | |||
<td align="center"> | |||
<math>~=</math> | |||
</td> | |||
<td align="left"> | |||
<math>~ | |||
-\frac{\eta^{-1}}{(1+\eta^2)} \biggl[ \frac{q^2 }{x} \biggr] | |||
= | |||
-\frac{x^2}{y(x^2+ q^4y^2)} | |||
</math> | </math> | ||
</td> | </td> |
Revision as of 22:35, 25 January 2021
Concentric Ellipsoidal (T8) Coordinates
| Tiled Menu | Tables of Content | Banner Video | Tohline Home Page | |
Background
Building on our general introduction to Direction Cosines in the context of orthogonal curvilinear coordinate systems, and on our previous development of T3 (concentric oblate-spheroidal) and T5 (concentric elliptic) coordinate systems, here we explore the creation of a concentric ellipsoidal (T8) coordinate system. This is motivated by our desire to construct a fully analytically prescribable model of a nonuniform-density ellipsoidal configuration that is an analog to Riemann S-Type ellipsoids.
Note that, in a separate but closely related discussion, we made attempts to define this coordinate system, numbering the trials up through "T7." In this "T7" effort, we were able to define a set of three, mutually orthogonal unit vectors that should work to define a fully three-dimensional, concentric ellipsoidal coordinate system. But we were unable to figure out what coordinate function, <math>~\lambda_3(x, y, z)</math>, was associated with the third unit vector. In addition, we found the <math>~\lambda_2</math> coordinate to be rather strange in that it was not oriented in a manner that resembled the classic spherical coordinate system. Here we begin by redefining the <math>~\lambda_2</math> coordinate such that its associated <math>~\hat{e}_3</math> unit vector lies parallel to the x-y plane.
Realigning the Second Coordinate
The first coordinate remains the same as before, namely,
<math>~\lambda_1^2</math> |
<math>~=</math> |
<math>~x^2 + q^2 y^2 + p^2 z^2 \, .</math> |
This may be rewritten as,
<math>~1</math> |
<math>~=</math> |
<math>~\biggl( \frac{x}{a}\biggr)^2 + \biggl( \frac{y}{b}\biggr)^2 + \biggl(\frac{z}{c}\biggr)^2 \, ,</math> |
where,
<math>~a = \lambda_1 \, ,</math> |
<math>~b = \frac{\lambda_1}{q} \, ,</math> |
<math>~c = \frac{\lambda_1}{p} \, .</math> |
By specifying the value of <math>~z = z_0 < c</math>, as well as the value of <math>~\lambda_1</math>, we are picking a plane that lies parallel to, but a distance <math>~z_0</math> above, the equatorial plane. The elliptical curve that defines the intersection of the <math>~\lambda_1</math>-constant surface with this plane is defined by the expression,
<math>~\lambda_1^2 - p^2z_0^2</math> |
<math>~=</math> |
<math>~x^2 + q^2 y^2 </math> |
<math>~\Rightarrow~~~1</math> |
<math>~=</math> |
<math>~\biggl( \frac{x}{a_{2D}}\biggr)^2 + \biggl( \frac{y}{b_{2D}}\biggr)^2 \, ,</math> |
where,
<math>~a_{2D} = \biggl(\lambda_1^2 - p^2z_0^2 \biggr)^{1 / 2} \, ,</math> |
<math>~b_{2D} = \frac{1}{q} \biggl(\lambda_1^2 - p^2z_0^2 \biggr)^{1 / 2} \, .</math> |
At each point along this elliptic curve, the line that is tangent to the curve has a slope that can be determined by simply differentiating the equation that describes the curve, that is,
<math>~0</math> |
<math>~=</math> |
<math>~\frac{2x dx}{a_{2D}^2} + \frac{2y dy}{b_{2D}^2}</math> |
<math>~\Rightarrow~~~\frac{dy}{dx}</math> |
<math>~=</math> |
<math>~- \frac{2x}{a_{2D}^2} \cdot \frac{b_{2D}^2}{2y} = - \frac{x}{q^2y} \, .</math> |
<math>~\Rightarrow~~~\Delta y</math> |
<math>~=</math> |
<math>~- \biggl( \frac{x}{q^2y} \biggr)\Delta x \, .</math> |
The unit vector that lies tangent to any point on this elliptical curve will be described by the expression,
<math>~\hat{e}_2</math> |
<math>~=</math> |
<math>~ \hat\imath~ \biggl\{ \frac{\Delta x}{[ (\Delta x)^2 + (\Delta y)^2 ]^{1 / 2}} \biggr\} + \hat\jmath~ \biggl\{ \frac{\Delta y}{[ (\Delta x)^2 + (\Delta y)^2 ]^{1 / 2}} \biggr\} </math> |
|
<math>~=</math> |
<math>~ \hat\imath~ \biggl\{ \frac{1}{[ 1 + x^2/(q^4y^2) ]^{1 / 2}} \biggr\} - \hat\jmath~ \biggl\{ \frac{x/(q^2y)}{[ 1 + x^2/(q^4y^2) ]^{1 / 2}} \biggr\} </math> |
|
<math>~=</math> |
<math>~ \hat\imath~ \biggl\{ \frac{q^2y}{[ x^2 + q^4y^2 ]^{1 / 2}} \biggr\} - \hat\jmath~ \biggl\{ \frac{x}{[ x^2 + q^4y^2 ]^{1 / 2}} \biggr\} \, .</math> |
As we have discovered, the coordinate that gives rise to this unit vector is,
<math>~\lambda_2</math> |
<math>~=</math> |
<math>~\frac{x}{y^{1/q^2}} \, .</math> |
Other properties that result from this definition of <math>~\lambda_2</math> are presented in the following table.
Direction Cosine Components for T8 Coordinates | |||||||||||
<math>~n</math> | <math>~\lambda_n</math> | <math>~h_n</math> | <math>~\frac{\partial \lambda_n}{\partial x}</math> | <math>~\frac{\partial \lambda_n}{\partial y}</math> | <math>~\frac{\partial \lambda_n}{\partial z}</math> | <math>~\gamma_{n1}</math> | <math>~\gamma_{n2}</math> | <math>~\gamma_{n3}</math> | |||
<math>~1</math> | <math>~(x^2 + q^2 y^2 + p^2 z^2)^{1 / 2} </math> | <math>~\lambda_1 \ell_{3D}</math> | <math>~\frac{x}{\lambda_1}</math> | <math>~\frac{q^2 y}{\lambda_1}</math> | <math>~\frac{p^2 z}{\lambda_1}</math> | <math>~(x) \ell_{3D}</math> | <math>~(q^2 y)\ell_{3D}</math> | <math>~(p^2z) \ell_{3D}</math> | |||
<math>~2</math> | <math>~\frac{x}{ y^{1/q^2}}</math> | <math>~\frac{1}{\lambda_2}\biggl[\frac{x q^2 y }{(x^2 + q^4y^2)^{1 / 2}}\biggr] </math> | <math>~\frac{\lambda_2}{x}</math> | <math>~-\frac{\lambda_2}{q^2 y}</math> | <math>~0</math> | <math>~\frac{q^2 y }{(x^2 + q^4y^2)^{1 / 2}} </math> | <math>~- \frac{x }{(x^2 + q^4y^2)^{1 / 2}} </math> | <math>~0</math> | |||
<math>~3</math> | --- | --- | --- | --- | --- | --- | --- | --- | |||
|
The associated unit vector is, then,
<math>~\hat{e}_2</math> |
<math>~=</math> |
<math>~ \hat\imath~\biggl[ \frac{q^2 y }{(x^2 + q^4y^2)^{1 / 2}} \biggr] - \hat\jmath~\biggl[ \frac{x }{(x^2 + q^4y^2)^{1 / 2}} \biggr] \, . </math> |
It is easy to see that <math>~\hat{e}_2 \cdot \hat{e}_2 = 1</math>. We also see that,
<math>~\hat{e}_1 \cdot \hat{e}_2</math> |
<math>~=</math> |
<math>~ x \ell_{3D}\biggl[ \frac{q^2 y }{(x^2 + q^4y^2)^{1 / 2}} \biggr] - q^2y \ell_{3D} \biggl[ \frac{x }{(x^2 + q^4y^2)^{1 / 2}} \biggr] = 0 \, , </math> |
so it is clear that these first two unit vectors are orthogonal to one another.
Search for the Third Coordinate
Cross Product of First Two Unit Vectors
The cross-product of these two unit vectors should give the third, namely,
<math>~\hat{e}_3 = \hat{e}_1 \times \hat{e}_2</math> |
<math>~=</math> |
<math>~ \hat\imath~\biggl[ {e}_{1y}{e}_{2z} - {e}_{1z}{e}_{2y} \biggr] + \hat\jmath~\biggl[ {e}_{1z}{e}_{2x} - {e}_{1x}{e}_{2z} \biggr] + \hat{k}~\biggl[ {e}_{1x}{e}_{2y} - {e}_{1y}{e}_{2x} \biggr] </math> |
|
<math>~=</math> |
<math>~ \hat\imath~\biggl[ \frac{x p^2z\ell_{3D}}{(x^2 + q^4y^2)^{1 / 2}} \biggr] + \hat\jmath~\biggl[ \frac{q^2 y p^2z \ell_{3D}}{(x^2 + q^4y^2)^{1 / 2}} \biggr] - \hat{k}~\biggl[ \frac{x^2\ell_{3D} }{(x^2 + q^4y^2)^{1 / 2}} ~+~ \frac{q^4 y^2\ell_{3D} }{(x^2 + q^4y^2)^{1 / 2}} \biggr] </math> |
|
<math>~=</math> |
<math>~ \frac{\ell_{3D}}{(x^2 + q^4y^2)^{1 / 2}} \biggl\{~ \hat\imath~( x p^2z ) + \hat\jmath~(q^2 y p^2z ) - \hat{k}~(x^2 + q^4y^2) ~\biggr\} \, . </math> |
Inserting these component expressions into the last row of the T8 Direction Cosine table gives …
Direction Cosine Components for T8 Coordinates | |||||||||||
<math>~n</math> | <math>~\lambda_n</math> | <math>~h_n</math> | <math>~\frac{\partial \lambda_n}{\partial x}</math> | <math>~\frac{\partial \lambda_n}{\partial y}</math> | <math>~\frac{\partial \lambda_n}{\partial z}</math> | <math>~\gamma_{n1}</math> | <math>~\gamma_{n2}</math> | <math>~\gamma_{n3}</math> | |||
<math>~1</math> | <math>~(x^2 + q^2 y^2 + p^2 z^2)^{1 / 2} </math> | <math>~\lambda_1 \ell_{3D}</math> | <math>~\frac{x}{\lambda_1}</math> | <math>~\frac{q^2 y}{\lambda_1}</math> | <math>~\frac{p^2 z}{\lambda_1}</math> | <math>~(x) \ell_{3D}</math> | <math>~(q^2 y)\ell_{3D}</math> | <math>~(p^2z) \ell_{3D}</math> | |||
<math>~2</math> | <math>~\frac{x}{ y^{1/q^2}}</math> | <math>~\frac{1}{\lambda_2}\biggl[\frac{x q^2 y }{(x^2 + q^4y^2)^{1 / 2}}\biggr] </math> | <math>~\frac{\lambda_2}{x}</math> | <math>~-\frac{\lambda_2}{q^2 y}</math> | <math>~0</math> | <math>~\frac{q^2 y }{(x^2 + q^4y^2)^{1 / 2}} </math> | <math>~- \frac{x }{(x^2 + q^4y^2)^{1 / 2}} </math> | <math>~0</math> | |||
<math>~3</math> | --- | --- | --- | --- | --- | <math>~\frac{x p^2 z\ell_{3D}}{(x^2 + q^4y^2)^{1 / 2}} </math> | <math>~\frac{q^2 y p^2 z\ell_{3D}}{(x^2 + q^4y^2)^{1 / 2}} </math> | <math>~-\frac{(x^2 + q^4 y^2)\ell_{3D}}{(x^2 + q^4y^2)^{1 / 2}} </math> | |||
|
Associated h3 Scale Factor
After working through various scenarios on my whiteboard today (21 January 2021), I propose that,
<math>~\frac{\partial \lambda_3}{\partial x}</math> |
<math>~=</math> |
<math>~\frac{xp^2z}{(x^2 + q^4y^2)} \, ;</math> |
<math>~\frac{\partial \lambda_3}{\partial y}</math> |
<math>~=</math> |
<math>~\frac{q^2 y p^2z}{(x^2 + q^4y^2)} \, ;</math> |
and |
<math>~\frac{\partial \lambda_3}{\partial z}</math> |
<math>~=</math> |
<math>~-1 \, .</math> |
This means that,
<math>~h_3^{-2}</math> |
<math>~=</math> |
<math>~\sum_{i=1}^3 \biggl( \frac{\partial \lambda_3}{\partial x_i}\biggr)^2</math> |
|
<math>~=</math> |
<math>~ \biggl[ \frac{xp^2z}{(x^2 + q^4y^2)} \biggr]^2 + \biggl[ \frac{q^2 y p^2z}{(x^2 + q^4y^2)} \biggr]^2 + \biggl[ -1 \biggr]^2 </math> |
|
<math>~=</math> |
<math>~ \frac{p^4z^2(x^2 + q^4y^2)}{(x^2 + q^4y^2)^2} + 1 </math> |
|
<math>~=</math> |
<math>~ \frac{(x^2 + q^4y^2 +p^4z^2)}{(x^2 + q^4y^2)} </math> |
|
<math>~=</math> |
<math>~ \frac{1}{\ell_{3D}^2 (x^2 + q^4y^2)} </math> |
<math>~\Rightarrow~~~ h_3</math> |
<math>~=</math> |
<math>~ \ell_{3D} (x^2 + q^4y^2)^{1 / 2} \, . </math> |
This seems to work well because, when combined with the three separate expressions for <math>~\partial \lambda_3/\partial x_i</math>, this single expression for <math>~h_3</math> generates all three components of the third unit vector, that is, all three direction cosines, <math>~\gamma_{3i}</math>. All of the elements of this new "EUREKA moment" result have been entered into the following table.
Direction Cosine Components for T8 Coordinates | |||||||||||
<math>~n</math> | <math>~\lambda_n</math> | <math>~h_n</math> | <math>~\frac{\partial \lambda_n}{\partial x}</math> | <math>~\frac{\partial \lambda_n}{\partial y}</math> | <math>~\frac{\partial \lambda_n}{\partial z}</math> | <math>~\gamma_{n1}</math> | <math>~\gamma_{n2}</math> | <math>~\gamma_{n3}</math> | |||
<math>~1</math> | <math>~(x^2 + q^2 y^2 + p^2 z^2)^{1 / 2} </math> | <math>~\lambda_1 \ell_{3D}</math> | <math>~\frac{x}{\lambda_1}</math> | <math>~\frac{q^2 y}{\lambda_1}</math> | <math>~\frac{p^2 z}{\lambda_1}</math> | <math>~(x) \ell_{3D}</math> | <math>~(q^2 y)\ell_{3D}</math> | <math>~(p^2z) \ell_{3D}</math> | |||
<math>~2</math> | <math>~\frac{x}{ y^{1/q^2}}</math> | <math>~\frac{1}{\lambda_2}\biggl[\frac{x q^2 y }{(x^2 + q^4y^2)^{1 / 2}}\biggr] </math> | <math>~\frac{\lambda_2}{x}</math> | <math>~-\frac{\lambda_2}{q^2 y}</math> | <math>~0</math> | <math>~\frac{q^2 y }{(x^2 + q^4y^2)^{1 / 2}} </math> | <math>~- \frac{x }{(x^2 + q^4y^2)^{1 / 2}} </math> | <math>~0</math> | |||
<math>~3</math> | --- | <math>~\ell_{3D}(x^2 + q^4 y^2)^{1 / 2}</math> | <math>~\frac{xp^2z}{(x^2 + q^4y^2)} </math> | <math>~\frac{q^2 y p^2z}{(x^2 + q^4y^2)}</math> | <math>~-1</math> | <math>~\frac{x p^2 z\ell_{3D}}{(x^2 + q^4y^2)^{1 / 2}} </math> | <math>~\frac{q^2 y p^2 z\ell_{3D}}{(x^2 + q^4y^2)^{1 / 2}} </math> | <math>~-\frac{(x^2 + q^4 y^2)\ell_{3D}}{(x^2 + q^4y^2)^{1 / 2}} </math> | |||
|
What is the Third Coordinate Function, λ3
The remaining $64,000 question is, "What is the actual expression for <math>~\lambda_3(x, y, z)</math> ? "
Notice that the (partial) derivatives of <math>~\lambda_3</math> with respect to <math>~x</math> and <math>~y</math> may be rewritten, respectively, in the form
<math>~\biggl( \frac{q^2 y}{p^2z} \biggr) \frac{\partial \lambda_3}{\partial x}</math> |
<math>~=</math> |
<math>~ \frac{q^2 y}{x(1 + \eta^2)} = \frac{\eta}{(1 + \eta^2)} \, , </math> and, |
<math>~\biggl( \frac{x}{p^2z} \biggr) \frac{\partial \lambda_3}{\partial y}</math> |
<math>~=</math> |
<math>~ \frac{q^2y}{x(1 + \eta^2)} = \frac{\eta}{(1 + \eta^2)} \, , </math> |
where,
<math>~\eta</math> |
<math>~\equiv</math> |
<math>~ \frac{q^2y}{x} ~~~~\Rightarrow~~ \frac{\partial \ln\eta}{\partial \ln x} = -1 \, , ~~~~\frac{\partial \ln\eta}{\partial \ln y} = +1 \, . </math> |
Then, after searching through the CRC Mathematical Handbook's pages of familiar derivative expressions, we appreciate that
<math>~\frac{d}{dx_i} \biggl[\frac{1}{\cosh\gamma}\biggr]</math> |
<math>~=</math> |
<math>~- \biggl[ \frac{\tanh\gamma}{\cosh\gamma} \biggr] \frac{d\gamma}{dx_i} \, .</math> |
Hence, it will be useful to adopt the mapping,
<math>~\eta</math> |
<math>~~~\rightarrow~~~</math> |
<math>~\sinh \gamma \, ,</math> |
because the right-hand side of both partial-derivative expressions becomes,
<math>~\frac{\eta}{(1+\eta^2)}</math> |
<math>~~~\rightarrow~~~</math> |
<math>~\frac{\sinh\gamma}{\cosh^2\gamma} = \frac{\tanh\gamma}{\cosh\gamma} \, .</math> |
Guess A
In particular, this suggests that we set,
<math>~\lambda_3</math> |
<math>~=</math> |
<math>~\frac{A}{\cosh\gamma} \, ,</math> |
where,
<math>~\gamma</math> |
<math>~\equiv</math> |
<math>~\sinh^{-1}\eta = \pm \cosh^{-1}[\eta^2 + 1]^{1 / 2} \, .</math> |
In other words,
<math>~\lambda_3</math> |
<math>~=</math> |
<math>~A[\eta^2 + 1]^{-1 / 2} \, .</math> |
Let's check the first and second partial derivatives.
<math>~\frac{\partial \lambda_3}{\partial x}</math> |
<math>~=</math> |
<math>~ - \frac{A}{2} \biggl[ \frac{2\eta}{(\eta^2 + 1)^{3 / 2}} \biggr] \frac{\partial \eta}{\partial x} </math> |
Guess B
What if,
<math>~\lambda_3</math> |
<math>~=</math> |
<math>~\frac{1}{2}\ln(1+\eta^2) \, .</math> |
Then,
<math>~\frac{d\lambda_3}{d\eta}</math> |
<math>~=</math> |
<math>~ \frac{\eta}{(1+\eta^2)} \, , </math> |
in which case we find,
<math>~\frac{\partial\lambda_3}{\partial x_i} </math> |
<math>~=</math> |
<math>~\frac{d\lambda_3}{d\eta} \cdot \frac{\partial\eta}{\partial x_i}</math> |
which means,
<math>~\frac{\partial\lambda_3}{\partial x} </math> |
<math>~=</math> |
<math>~ \frac{\eta}{(1+\eta^2)} \biggl[ - \frac{q^2 y}{x^2} \biggr] = - \frac{x}{(x^2 + q^4y^2)} \biggl[ \frac{q^4 y^2}{x^2} \biggr] \, . </math> |
Guess C
What if,
<math>~\lambda_3</math> |
<math>~=</math> |
<math>~\frac{1}{2} \ln(1+\eta^{-2}) \, .</math> |
Then,
<math>~\frac{d\lambda_3}{d\eta}</math> |
<math>~=</math> |
<math>~- \frac{\eta^{-3}}{(1+\eta^{-2})} = -\frac{ \eta^{-1}}{(1+\eta^{2})} </math> |
in which case we find,
<math>~\frac{\partial\lambda_3}{\partial x_i} </math> |
<math>~=</math> |
<math>~\frac{d\lambda_3}{d\eta} \cdot \frac{\partial\eta}{\partial x_i}</math> |
which means,
<math>~\frac{\partial\lambda_3}{\partial x} </math> |
<math>~=</math> |
<math>~ -\frac{\eta^{-1}}{(1+\eta^2)} \biggl[ - \frac{q^2 y}{x^2} \biggr] = \frac{x}{(x^2 + q^4y^2)} \, ; </math> |
and,
<math>~\frac{\partial\lambda_3}{\partial y} </math> |
<math>~=</math> |
<math>~ -\frac{\eta^{-1}}{(1+\eta^2)} \biggl[ \frac{q^2 }{x} \biggr] = -\frac{x^2}{y(x^2+ q^4y^2)} </math> |
See Also
© 2014 - 2021 by Joel E. Tohline |