Griffiths problem 2.7 Messy Integral

7 replies · 6K views · last reply
Waxterzz
Messages
82
Reaction score
0
Hi all,
Handwritten-style typeset derivation of a potential integral: substitution u = cos θ, the integral of (z − Ru) over (R² + z² − 2Rzu)^(3/2) from −1 to 1, and a closed-form result with a bracketed term involving |z − R| and |z + R|.


"Integral can be done by partial fractions - or look it up" So second line, that's what I want to do.

How to deal with this? What substitution can I use? Never encountered partial fractions with non-integer exponents.

Someone give me a tip?

Thanks in advance
 
Waxterzz said:
Hi all,
Three lines of a mathematical derivation: a surface integral over theta from 0 to pi is rewritten with substitution u = cos theta into an integral from -1 to 1, then evaluated to a closed-form expression with terms involving (z - R)/|z - R| and (-z - R)/|z + R|.


"Integral can be done by partial fractions - or look it up" So second line, that's what I want to do.

How to deal with this? What substitution can I use? Never encountered partial fractions with non-integer exponents.

Someone give me a tip?

Thanks in advance
I do think the author must have meant integration by parts. The term without a ## u ## in the numerator that has a ## (A+Bu)^{3/2} ## in the denominator is a straightforward integration and the ## u \, du/(A+Bu)^{3/2} ## can be readily integrated by parts.
 
  • Like
Likes   Reactions: Waxterzz and zaidalyafey
You can directly use integration by parts.
 
  • Like
Likes   Reactions: Waxterzz
Charles Link said:
I do think the author must have meant integration by parts.

zaidalyafey said:
You can directly use integration by parts.

I get something extremely messy. Here is the first part, the second part with integration by parts,I'll do it again tomorrow.

So the first term

integral (z du / (R² + z² - 2 Rz u) ^3/2 )

d(R² + z² - 2 Rz u) = -2 R z du

The integral becomes

-1/2R * integral ( d(R² + z² - 2 Rz u) / (R² + z² - 2 Rz u) ^3/2) )

Then I get1/R * (1/ (R² + z² -2Rzu) ^ 1/2)

The second part I'm going to do again tomorrow, and I'm going to try to put it in latex (have to learn latex quick) because posting like this is unreadable. But the first part is correct right?
 
You can differentiate z - Ru and integrate ( du / (R² + z² - 2 Rz u)^3/2)
 
  • Like
Likes   Reactions: Waxterzz
From first to second line is simply the substitution given. The integration is from second line to third.
 
  • Like
Likes   Reactions: Waxterzz
zaidalyafey said:
You can differentiate z - Ru and integrate ( du / (R² + z² - 2 Rz u)^3/2)

mathman said:
From first to second line is simply the substitution given. The integration is from second line to third.
Ok, so partial integration. integral (fdg) = fg - integral (gdf)

f = z-Ru
dg = du / (R² +z² -2zRu) ^ 3/2

For g I get
1 / ( (Rz) * (R² + z² -2*R*z*u)^1/2)

fg = (z-Ru) / ( (Rz) * (R² + z² -2*R*z*u)^1/2)

now the integral part - integral ( g d f)

df = -R du

So

The integral to solve

Rdu / ((Rz) * (R² + z² -2*R*z*u)^1/2))

Becomes

1/z * (2/-2*R*z) * (R² + z² -2*R*z*u)^1/2

Combining two terms, and rearranging

I get solutionDEAR GOD

hhahaha

(Okay I really need to learn latex now haha)
Charles Link said:
I do think the author must have meant integration by parts. The term without a ## u ## in the numerator that has a ## (A+Bu)^{3/2} ## in the denominator is a straightforward integration and the ## u \, du/(A+Bu)^{3/2} ## can be readily integrated by parts.

Thank u also.

How to give everyone upvotes in this thread?
 
Waxterzz said:
How to give everyone upvotes in this thread?

You can hit the like button in the bottom right corner of each post.
 
  • Like
Likes   Reactions: Waxterzz