0

I am trying to determine the volume of the blue shaded area in the first figure. This volume is essentially the volume between two tilted ellipses (one up and one down). Since I am not well versed with double integrals, I will include comments on what my thinking process is so that responders can correct me and I can think about this the correct way.

The volume is given by:

$$V = \int^{2\pi}_0\int^r_0h(\phi)rdrd\phi$$

where $h$ is the height, $r$ is the distance from the left focus and $\phi$ is the angle measured from the focus (see Fig. 2). In Fig. 2 "P" is the left focus and "O" is the center of the ellipse. Now, I guess the way I am seeing this is that there is a slice in the $r$ direction with a height of $h$ (z-axis). The volume is then determined by sweeping this slice along the $\phi$ direction. It turns out that the equation for this slice is given by,

$$h(\phi)r = \frac{R^2\sin^3\beta}{(1-\cos\beta\cos\phi)^2}$$

where $\beta$ is an angle that is constant. $R$ is the semi-minor axis of the ellipse and $\frac{R}{\sin\beta}$ is the semi-major axis of the ellipse as shown in the second figure. From Wikipedia, the polar form relative to the focus is:

$$r(\phi)=\frac{a(1-e^2)}{1\pm e\cos\phi}$$

where $a$ is the semi-major axis and

$$e = \sqrt{1-\frac{b^2}{a^2}} = \sqrt{1-\frac{R^2}{\frac{R^2}{\sin^2\beta}}} = \sqrt{1-\sin^2\beta} = \cos\beta $$

where $b$ is the semi-minor axis.

So, continuing on, this is what I think the double integral becomes

$$V = \int^{2\pi}_0\int^{\frac{a(1-e^2)}{1\pm e\cos\phi}}_0\frac{R^2\sin^3\beta}{(1-\cos\beta\cos\phi)^2}drd\phi$$

or

$$V = 2{R^2\sin^3\beta}\int^{\pi}_0\int^{\frac{a(1-e^2)}{1\pm e\cos\phi}}_0\frac{1}{(1-\cos\beta\cos\phi)^2}drd\phi$$

Since the volume is symmetrical about the $\phi$ direction, the outer integral upper bound is replaced by $\pi$ and the entire integral is multiplied by 2.

The upper bound for the inside integral can be simplified to

$$\frac{a(1-e^2)}{1\pm e\cos\phi} = \frac{\frac{R}{\sin\beta}(1-\cos^2\beta)}{1\pm \cos\beta\cos\phi} = \frac{R\sin\beta}{1\pm \cos\beta\cos\phi} $$

Since the upper bound for the integral on $r$ is a function of $\phi$, it seems to me that the inside integral has to be the one over $r$ and not $\phi$ (please confirm). So after evaluating the inside integral, I get:

$$V = 2{R^2\sin^3\beta}\int^{\pi}_0\frac{1}{(1-\cos\beta\cos\phi)^2}{\frac{R\sin\beta}{1\pm \cos\beta\cos\phi}}d\phi$$

Now before this integral even gets evaluated, I'm starting to get bogged down. From Fig. 2, I think the portion of the ellipse where the second half of the denominator is equal to $1-\cos\beta\cos\phi$ is to the right of the focus "P" and to the left of the focus "P", the right half of the denominator is equal to $1+\cos\beta\cos\phi$.

So should I split the integral up into two portions as in,

$$V = 2{R^3\sin^4\beta}\int_{0}^{\frac{\pi}{2}}\frac{1}{(1-\cos\beta\cos\phi)^2}{\frac{1}{1 - \cos\beta\cos\phi}}d\phi + 2{R^3\sin^4\beta}\int_{\frac{\pi}{2}}^{\pi} \frac{1}{(1-\cos\beta\cos\phi)^2}{\frac{1}{1 + \cos\beta\cos\phi}}d\phi$$

Any help including coming up with the final solution is welcome.

EDIT: Drawing of the volume in x and z plane only has been added for clarification.

enter image description here

enter image description here

enter image description here

enter image description here

rdemo
  • 173
  • Not obvious by the iso-view, I conclude from your ansatz the ellipses are parallel with the xy-plane. By "tilted" you mean the orthogonal projection in z-direction results in two twistied ellipses. Thus another ansatz could be, to integrate the area $A=f(z)$ from the bottom to the top, $V = \int_0^h A(z)dz$. Problem here and also in your approach -- do you know the envelope curve? This could be more or less an hourglass, me think. Or the skin of a soap bubble. – m-stgt Feb 11 '24 at 04:54
  • I have added a drawing of the volume in the x and z plane only to clarify. The black lines at the top and bottom of the solid are the ellipses, which are detailed in Fig. 2. These ellipses bound the top and bottom of the volume. – rdemo Feb 11 '24 at 05:07
  • so the ellipses are inclined, the orthogonal projected they result the same image in the xy-plane. Right? Both ellipses do have the same shape? (That implies angle of inclination of the top is the -angle of the bottom.) – m-stgt Feb 11 '24 at 05:18
  • Yes, both ellipses are exactly the same except that they are inclined differently but at the same absolute angle relative to the x axis. I've added a drawing in the x-y plane. In this view the ellipses are exactly one on top of the other. One other thing, if $B$ is equal to 90 degrees, then the solid is a cylinder with a circular cross section. The integral does evaluate to the volume of a cylinder when $B$ is equal to 90 degrees. – rdemo Feb 11 '24 at 06:00
  • "... then the solid is a cylinder with a circular cross section" -- LOL! Nice, picture you would cut your object in the xy-plane, turn one part in the z-direction by $\pi$ and fit together the two ellipses, you will have a cylinder. No need to torture integrals :) – m-stgt Feb 11 '24 at 06:16
  • Sorry, I just realized that the new drawing wasn't added until I just added it now. Yes, but B = 90 degrees is not really of interest. – rdemo Feb 11 '24 at 07:40
  • OK, I misconceived $\beta$. But you confirmed "both ellipses are exactly the same", so there is the option to fit them together. In one position you'll get a column with parallel top and bottom when fitting the two parts together. Final hint: get the semi axis $a$ and $b$ of the projection in xy-plane, the area of it is $A=\pi a b$ and the rest remains your task :) – m-stgt Feb 11 '24 at 08:13

0 Answers0