Is there an easier way to integrate this equation using a triangle?

  • Context: Undergrad 
  • Thread starter Thread starter Philosophaie
  • Start date Start date
  • Tags Tags
    Calculus Integral
Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
9 replies · 2K views
Philosophaie
Messages
456
Reaction score
0
I need to take the integral of an equation. I am trying to do Integration by Parts but I am coming up short.

I need to integrate (x^2 + y^2 + z^2)^(-3/2) dy for a Astronomy problem.
I tried:

u = x^2 + y^2 + z^2
du = 2*y dy

v=u^(-3/2)
dv = -3/2*u^(-5/2)

I can not do anything with the "y" in du.

Can you help?
 
Physics news on Phys.org
Philosophaie said:
I need to take the integral of an equation.
First off, what you have below is NOT an equation -- an equation has two expressions with = in between.
Philosophaie said:
I am trying to do Integration by Parts but I am coming up short.

I need to integrate (x^2 + y^2 + z^2)^(-3/2) dy for a Astronomy problem.
Like this?
$$\int (x^2 + y^2 + z^2)^{-3/2}dy$$

It would help to see the actual problem, to verify that the integral you set up is appropriate for the question. Assuming for the moment that you have the right integral, I don't think integration by parts is the way to go. I would use trig substitution instead.
Philosophaie said:
I tried:

u = x^2 + y^2 + z^2
du = 2*y dy

v=u^(-3/2)
dv = -3/2*u^(-5/2)

I can not do anything with the "y" in du.

Can you help?
 
If x, y, and z are independent variables, as they normally are, this is just [itex]\int (a^2+ y^2)^{3/2} dy[/itex] with [itex]a^2= x^2+ z^2[/itex]. The substitution [itex]y= asin(\theta)[/itex] shoud work.
 
HallsofIvy said:
If x, y, and z are independent variables, as they normally are, this is just [itex]\int (a^2+ y^2)^{3/2} dy[/itex] with [itex]a^2= x^2+ z^2[/itex]. The substitution [itex]y= asin(\theta)[/itex] shoud work.
The exponent is -3/2, per the OP.
 
Would it not be y = a*tan(t)?
1

For
img4.gif
set
img5.gif
. In this case we talk about sine-substitution.
2
For
img6.gif
set
img7.gif
. In this case we talk about tangent-substitution.
3
For
img8.gif
set
img9.gif
. In this case we talk about secant-substitution.
[tex]a = \sqrt{x^{2} + z^{2}}[/tex]

[tex]y = a*\tan{(t)}[/tex]
[tex]dy=a*(/sec{(t)})^2 dt[/tex]
[tex]t = \arctan{(\frac{x^{2} + y^{2} + z^{2}}{a})}[/tex]

[tex]\int \frac{a*(\sec(t))^2}{(a^3 (\tan{(t)}^{3})}dt[/tex]

I get lost from here.
 
Philosophaie said:
Would it not be y = a*tan(t)?
Yes, this would be the right trig substitution.

Philosophaie said:
1
For
img4.gif
set
img5.gif
. In this case we talk about sine-substitution.
2
For
img6.gif
set
img7.gif
. In this case we talk about tangent-substitution.
3
For
img8.gif
set
img9.gif
. In this case we talk about secant-substitution.
[tex]a = \sqrt{x^{2} + z^{2}}[/tex]

[tex]y = a*\tan{(t)}[/tex]
[tex]dy=a*(/sec{(t)})^2 dt[/tex]
[tex]t = \arctan{(\frac{x^{2} + y^{2} + z^{2}}{a})}[/tex]

[tex]\int \frac{a*(\sec(t))^2}{(a^3 (\tan{(t)}^{3})}dt[/tex]

I get lost from here.
The integral should be ##\int \frac{a\sec^2(t)dt}{a^3 \sec^3(t)}##

If you draw a right triangle with t as the acute angle, and the opposite side is y, adjacent side is a, the hypotenuse is ##\sqrt{y^2 + a^2}##. So ##\sec(t) = \frac{\sqrt{y^2 + a^2}}{a}##. Solve this equation for ##\sqrt{y^2 + a^2}## that is in the denominator.
 
[tex]\text{Where do you get "}(a*sec(t))^{3} \text{ in the denominator?}[/tex]
Mark44 said:
The integral should be ∫asec 2 (t)dta 3 sec 3 (t) \int \frac{a\sec^2(t)dt}{a^3 \sec^3(t)}
 
Last edited:
In the original integral, the denominator is ##(y^2 + a^2)^{3/2}##, right? Based on the trig substitution you chose, and on the right triangle I described, ##\sec(t) = \frac{\sqrt{y^2 + a^2}}{a} = \frac 1 a (y^2 + a^2)^{1/2}##, or ##(y^2 + a^2)^{1/2} = a \sec(t)##. Use this to rewrite what you started with in the denominator.
 
[tex]a*sec(t) \text{ does not equal }\sqrt{y^2 + a^2}[/tex]
[tex]a*\tan{(t)} = \sqrt{y^2 + a^2}[/tex]
 
Last edited:
Philosophaie said:
[tex]a*sec(t) \text{ does not equal }\sqrt{y^2 + a^2}[/tex]
[tex]a*\tan{(t)} = \sqrt{y^2 + a^2}[/tex]
No, this is not how things are in my triangle, which I described in an earlier post.

Snapshot.jpg

In my triangle, tan(t) = y/a and ##\sec(t) = \frac{\sqrt{y^2 + a^2}} a##

Have you actually drawn this triangle? I never bothered memorizing the formulas you wrote in post #5 -- I just draw a right triangle and label the sides and hypotenuse according to whether the expression in the radical is a sum or difference. If you don't use the triangle, you're doing things the hard way, IMO.