Perplexing integration - sqrt(1+sinx)

  • Thread starter Thread starter peripatein
  • Start date Start date
  • Tags Tags
    Integration
Join the discussion
Registration is free. Ask a follow-up in this thread, or start your own.
17 replies · 18K views
peripatein
Messages
868
Reaction score
0
Hello,

I am trying to integrate sqrt(1+sinx) dx but I keep getting 2sqrt(1-sinx) (having initially substituted u=sinx), which is wrong! Could anyone please tell me what I am doing wrong?
 
Physics news on Phys.org
We can't really know what you did wrong unless you tell us what you did! So, can you show us your work?
 
As I wrote, I substituted u=sinx and got that the integral should be equal to 2sqrt(1-sinx). Which is wrong, according to Wolfram and a simple differentiation test. I am simply not sure where I faltered. May you please advise?
 
I would have thought so, except that I was instructed to use that substitution. :s
It should still be correct, granted the algebra is. Yet the answer is incorrect!
 
peripatein, you need to show us your work. We cannot read your mind.
 
u=sinx, du=cosxdx
Hence, int=[sqrt(1+u)/sqrt(1-u^2)] du=[1/sqrt(1-u)] du.
Hence, int=2sqrt(1-sinx)
 
What do you mean? Dividing by sqrt(1+u)?
 
peripatein said:
What do you mean? Dividing by sqrt(1+u)?
Do the integration:

[itex]\displaystyle \int \frac{1}{\sqrt{1-u}}\,du\ .[/itex]
 
But isn't that 2sqrt(1-u)?
 
peripatein said:
But isn't that 2sqrt(1-u)?
sign error.
 
Wrong link. In any case, that is NOT the answer. Question is, why?
 
They're the same answer *if* [itex]\cos x=\sqrt{1-\sin^2 x}[/itex]. That's not always the case, but you implicitly assumed that this is an identity in your u-substitution.
 
peripatein said:
Wrong link. In any case, that is NOT the answer. Question is, why?
What is the correct answer?

You have to be careful about sign behavior of the functions you're working with.

[itex]\displaystyle \cos(x)=\sqrt{1-\sin^2(x)}\ \[/itex] only when cos(x) ≥ 0.

Otherwise, [itex]\displaystyle \ \cos(x)=-\sqrt{1-\sin^2(x)\ .}[/itex]
 
Okay, thanks!
I am also trying to evaluate the improper integral 1/[(x-1)^1/3] between 0 and 4. May I do so thus?:

{ lim b->1- int 1/[(x-1)^1/3] dx between 0 and b } + { lim b->1+ int 1/[(x-1)^1/3] dx between b and 4 }?

Is there a another, better way to evaluate it?
 
peripatein said:
Okay, thanks!
I am also trying to evaluate the improper integral 1/[(x-1)^1/3] between 0 and 4. May I do so thus?:

{ lim b->1- int 1/[(x-1)^1/3] dx between 0 and b } + { lim b->1+ int 1/[(x-1)^1/3] dx between b and 4 }?

Is there a another, better way to evaluate it?
It's best to start a new thread for this.

It's a completely different problem.