You are here:

Differential Equations/Non-constant 2nd order differential eqn


Hello Bob,

I am currently working on the following equation. I have tried quite a few tricks to solve the problem but can not find an analytic solution. In the case where L=0, I do find an analytic solution.

Here is the equation:

with R=b*Sqrt[1-X^2], L>0, b>0, A real

With L=0 I have a soution as I said earlier, but also when I take out the term L/R in (1+L/R)^3. I feel that this exact term is making the equation difficult to solve, so I am trying ways to bypass that, but did not come up with anything good. I also tried linearizing the terms by supposing    L << R but it does not do me any good.

Do you have any advice?

Best regards,



Try writing the whole ode out as:-

y'' +ay' + by = 0   (1)

Put y=f(x)u(x),   (2), f and u ti be determined.

then y'=f'u + fu'

so y''=f''u + 2f'u' + fu''

so (1) => f''u + 2f'u' + fu'' + a(f'u + fu') + bfu = 0

so fu'' + (2f' + af)u' + c (=other terms)u = 0

Now choose an f to kill the u' term, so we solve 2f' + af = 0,

the solution of this being f(x) = exp (- \int a(x)dx).

Hence we now just need to solve fu'' + cu = 0,

which we can write as u'' + du = 0, where d = c/f,

and we use the same method, so that u(x) = exp (- \int d(x)dx),

and those both go into (2) to give y.

Try for ALL your digital wants and needs.

Differential Equations

All Answers

Answers by Expert:

Ask Experts


Bob Marlow


Questions on partial differential equations, particularly solution of time-dependent ones, and particularly parabolic. I have some knowledge of ODEs, but am out of practice with them.


My thesis was in this area.

Headingley Signs

International Journal of Computers and Fluids, July 2011. Moving mesh methods for solving parabolic partial differential equations R. Marlow , M.E. Hubbard, P.K. Jimack.

PhD in computer science

©2017 All rights reserved.