Ellipsoid surface area
|
05-30-2021, 06:04 PM
(This post was last modified: 08-05-2022 04:57 PM by Albert Chan.)
Post: #3
|
|||
|
|||
RE: Ellipsoid surface area
This version use Carlson symmetric elliptic integrals, to avoid integration.
\(F\!\left(\phi, m\right) = \sin(\phi) R_F\!\left(\cos^{2}\!\left(\phi\right), 1 - m \sin^{2}\!\left(\phi\right), 1\right)\) \(F\!\left(\phi, m\right) - E\!\left(\phi, m\right) = \frac{1}{3} m \sin^{3}\!\left(\phi\right) R_D\!\left(\cos^{2}\!\left(\phi\right), 1 - m \sin^{2}\!\left(\phi\right), 1\right)\) We force R arguments to get close, with Duplication theorems. Assume x,y,z > 0 (x or y can be zero, but not both) Let λ = √(xy) + √(yz) + √(xz) > 0: RF(x,y,z) = RF(x+λ, y+λ, z+λ)*2 RD(x,y,z) = RD(x+λ, y+λ, z+λ)*2 + 3/(√(z)*(z+λ)) If arguments close enough, we can stop. RF(x,x,x) = x^(-1/2) RD(x,x,x) = x^(-3/2) From AM-GM inequality, we have (x+y)/2 ≥ √(xy) λ = √(xy) + √(yz) + √(xz) ≤ (x+y)/2 + (y+z)/2 + (x+z)/2 = x+y+z Code assumed λ*(1+1e-12) > x+y+z, we consider x,y,z close enough. Code: #cas >>> mp.pretty = True >>> elliprf(1,2,3), elliprd(1,2,3) # see https://mpmath.org/doc/current/functions/elliptic.html (0.726945935468908, 0.290460281028991) CAS> RFD(1,2,3) → [0.726945935469, 0.290460281029] CAS> ellipsoid_area(1,2,3) → 48.8821463026 Numerical Recipes, Elliptic Integrals: https://core.ac.uk/download/pdf/211378009.pdf HP-41 Module: Elliptic Functions and Orthogonal Polynomials, ellipsoid area example (page 9), using RG |
|||
« Next Oldest | Next Newest »
|
Messages In This Thread |
Ellipsoid surface area - Albert Chan - 05-28-2021, 12:37 AM
RE: Ellipsoid surface area - Albert Chan - 05-28-2021, 08:05 PM
RE: Ellipsoid surface area - Albert Chan - 05-30-2021 06:04 PM
RE: Ellipsoid surface area - Albert Chan - 05-31-2021, 01:38 AM
RE: Ellipsoid surface area - Albert Chan - 08-05-2022, 04:40 PM
RE: Ellipsoid surface area - Albert Chan - 08-05-2022, 04:51 PM
|
User(s) browsing this thread: 1 Guest(s)