Solving Bessel's Equation: Homework

  • Thread starter Thread starter TheFerruccio
  • Start date Start date
Click For Summary

Homework Help Overview

The discussion revolves around solving a differential equation of the form x^2y'' + 5xy' + (x^2 - 12)y = 0, specifically seeking a general solution in terms of Bessel functions. The original poster attempts to apply a substitution y = u/x^2, but expresses uncertainty regarding the substitution process and its implications for finding roots of the indicial equation.

Discussion Character

  • Exploratory, Assumption checking, Problem interpretation

Approaches and Questions Raised

  • Participants discuss the need to correctly calculate derivatives using the quotient rule, as u is a function of x. There are attempts to clarify the nature of the substitution and its role in transforming the original equation into a Bessel differential equation. Some participants question the independence of solutions when ν is an integer and the implications for expressing the general solution.

Discussion Status

The conversation reflects a mix of attempts to clarify the substitution process and the correct application of Bessel functions. Some participants provide guidance on the need for a second independent solution and the proper handling of the Bessel functions, while others express confusion over previous posts and corrections made during the discussion.

Contextual Notes

There are indications of confusion regarding the nature of the substitution and the treatment of derivatives, as well as the specific requirements for expressing solutions in terms of Bessel functions. The discussion also highlights the challenge of ensuring linear independence of solutions when ν is an integer.

TheFerruccio
Messages
216
Reaction score
0

Homework Statement



Find a general solution in terms of Bessel functions. (Use the indicated transformations and show the details)

Homework Equations



[tex]x^2y''+5xy'+(x^2-12)y = 0[/tex]
[tex]y = \frac{u}{x^2}[/tex]

The Attempt at a Solution



I know that the answer needs to be in the form of Bessel functions of the first and second kind, depending on what my roots are.

What is throwing me off is the substitution. The book never explained the nature of substitution for this type of problem. I just literally substituted in that variable like they told me, but I have an inkling that this is extremely wrong, and is going in the completely wrong direction:

[tex]x^2(6ux^{-4})+5x(-2ux^{-3})+(x^2-12)ux^{-2} = 0[/tex]
[tex]-16ux^{-2} + u = 0[/tex]

After this, I reached a dead end. I do not know how to relate this to finding my roots of r from the indicial equation. I have a sense that I need to use the series method first, then somewhere in the middle of the problem, do the substitution.
 
Physics news on Phys.org
You didn't calculate the derivatives correctly. Remember that u is a function of x, so when you calculate y'' and y', you have to use the quotient rule.
 
vela said:
You didn't calculate the derivatives correctly. Remember that u is a function of x, so when you calculate y'' and y', you have to use the quotient rule.

Ohhh, I thought [tex]u[/tex] was just a constant.
 
EDIT: I made two mistakes here. First: I forgot to convert the substitution back. Second: I forgot that the values of [tex]\nu[/tex] differ by an integer, which necessitates using the Bessel function of the second kind for the other solution.


Here's an attempt to solve it. I believe that, since [tex]u[/tex] is a function of [tex]x[/tex], then I am to use the product rule to find all the associated derivatives, then plug into the original equation in order to come up with Bessel's differential equation.

[tex]y = \frac{u}{x^2}[/tex]
[tex]-2ux^{-3}+u'x^{-2}[/tex]
[tex]6ux^{-4}-2u'x^{-3}+u''x^{-2}-2u'x^{-3}[/tex]


Plugging in...

[tex]x^{2}(6x^{-4}u-2x^{-3}u'+x^{-2}u''-2x^{-3}u')+5x(-2x^{-3}u+x^{-2}u')+x^2(x^{-2}u)-12(x^{-2}u) = 0[/tex]

This leaves me with...

[tex]u''+\frac{u'}{x}+(\frac{-16}{x^2}+1) = 0[/tex]

multiplying by [tex]x^2[/tex]...

[tex]x^2u''+xu' + (x^2-16)u = 0[/tex]

so

[tex]\nu=\pm4[/tex]

Using the Bessel function of the first kind of order [tex]\nu[/tex] I get the first solution...

[tex]u_1(x) = y_1(x)x^2 = c_1x^4\sum_{m=0}^{\infty}\frac{(-1)^mx^{2m}}{2^{2m+4}m!\Gamma(m+5)}[/tex]

The second solution uses the Bessel function of the second kind.

[tex]u_2(x) = \frac{1}{\sin{4\pi}}[J_{4}(x)\cos{4\pi}-J_{-4}(x)][/tex]
 
Last edited:
Two problems: First, the solutions to the differential equation for u(x), not y(x), are the Bessel functions. Second, when ν is an integer, Jν and J aren't linearly independent, so you still need to find a second independent solution for u(x).
 
Disregard the previous post I made. I have the problem of solving my confusion while typing the post, which results in countless edits and, consequently, likely confusion for the helpers here who are contributing their time. Sorry.
 
TheFerruccio said:
The second solution uses the Bessel function of the second kind.

[tex]u_2(x) = \frac{1}{\sin{4\pi}}[J_{4}(x)\cos{4\pi}-J_{-4}(x)][/tex]
This doesn't work either because sin 4π=0 and J4=J-4. You need to write it as a limit. In any case, you're probably just expected to express u(x) as a linear combination of J4(x) and Y4(x). You don't need to write out what the functions actually equal.
 
vela said:
This doesn't work either because sin 4π=0 and J4=J-4. You need to write it as a limit. In any case, you're probably just expected to express u(x) as a linear combination of J4(x) and Y4(x). You don't need to write out what the functions actually equal.

Again, right (read the first part of my post, I stuck an edit there yesterday, I realized both mistakes and later said to disregard that post, lol).

Anyway, here's the solution I got, for the sake of closure (not the mathematical closure):

[tex]y(x) = c_1\frac{J_4(x)}{x^2}+c_2\frac{Y_4(x)}{x^2}[/tex] (edited for typo)

where
[tex]J_4(x)=x^4\sum_{m=0}^{\infty}\frac{(-1)^mx^{2m}}{2^{2m+4}m!(M+4)!}[/tex]

and

[tex]Y_4(x) = \frac{2}{\pi}x^4J_4(x)(\ln(\frac{x}{2})+ \gamma)+\frac{x^4}{\pi}\sum_{m=0}^{\infty}\frac{(-1)^{m-1}(\sum_{j=1}^{\infty}\frac{1}{j}+\sum_{j=1}^{m+4}\frac{1}{j})}{2^{2m+4}m!(m+4)!}x^{2m}-\frac{x^{-4}}{\pi}\sum_{m=0}^{3}\frac{(3-m)!}{2^{2m-4}m!}x^{2m}[/tex]

where

[tex]\gamma = \lim_{n \to 0}(\sum_{k=1}^{n}\frac{1}{k} - \ln(n))[/tex] AKA the Euler–Mascheroni constant, which is [tex]\approx .577[/tex]
Thanks for the help!
 
Last edited:

Similar threads

  • · Replies 12 ·
Replies
12
Views
2K
  • · Replies 8 ·
Replies
8
Views
2K
  • · Replies 11 ·
Replies
11
Views
2K
  • · Replies 5 ·
Replies
5
Views
3K
Replies
8
Views
2K
  • · Replies 4 ·
Replies
4
Views
3K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 15 ·
Replies
15
Views
2K