Integral Trig Substitution Question

TheRedDevil18
Messages
406
Reaction score
2
I just have a few questions. When using a trig substitution does it have to be under a radical ?

eg, suppose I wanted to integrate (x2)/(x2-9), I used a trig substitution of x = 3sec(t) and got the wrong answer and so apparently I had to use partial fractions
 
Physics news on Phys.org
TheRedDevil18 said:
I just have a few questions. When using a trig substitution does it have to be under a radical ?

eg, suppose I wanted to integrate (x2)/(x2-9), I used a trig substitution of x = 3sec(t) and got the wrong answer and so apparently I had to use partial fractions

In the following integral,

8281587fb9879173e9f84c4aa8e7d423.png


Try using the following substitution:

1e209acbb0838b8008b7b34fb7c8ebeb.png


What do we have?
 
TheRedDevil18 said:
When using a trig substitution does it have to be under a radical ?
No.

TheRedDevil18 said:
eg, suppose I wanted to integrate (x2)/(x2-9), I used a trig substitution of x = 3sec(t) and got the wrong answer and so apparently I had to use partial fractions
Then I'm pretty sure you made a mistake. If you use partial fractions, you need to rewrite the integrand as a proper rational expression, which you can do by polynomial long division. (There's another way, as well.)
You can also do this by trig substitution. Without seeing your work, I can't say why you got a wrong answer.
 
Ok, I'm going to skip some of the substituting parts to this,

3*integral sec3t/tan(t) dt
= (1+tan2t)/(tan(t)) * sec(t) dt

Integrating that, I get

ln(sin(t)) + sec(t)

Is it correct so far ?, I know the last thing I must do is sub the x values, but I just want to make sure if this part is correct
 
TheRedDevil18 said:
Ok, I'm going to skip some of the substituting parts to this,

3*integral sec3t/tan(t) dt
This part looks OK. I'll check the following work in a bit.
 
TheRedDevil18 said:
Ok, I'm going to skip some of the substituting parts to this,

3*integral sec3t/tan(t) dt
= (1+tan2t)/(tan(t)) * sec(t) dt
You lost the factor of 3, and have omitted the integration sign.
BTW, it's easy enough to do in LaTeX - # # \int f(x)dx # # (omit the spaces between the pound signs).
TheRedDevil18 said:
Integrating that, I get

ln(sin(t)) + sec(t)
Nope. The factor of 3 is still missing on both terms. One of them should be 3 sec(t), but the other is wrong.
I suspect that you did this:
$$\int \frac{dt}{sin(t)} = ln(sin(t))$$
... which is incorrect. This is a somewhat tricky integral. Rather than deriving it here, I would advise just looking it up in a table of integrals.
TheRedDevil18 said:
Is it correct so far ?, I know the last thing I must do is sub the x values, but I just want to make sure if this part is correct
 
My integration was wrong. Turns out I had to use integration by parts for that trig integral with the sec^3 and tan and it was very messy. I did the same problem using partial fractions and it was much easier to integrate

Thanks for the help, guys :)
 
You don't need integration by parts. Starting from what you had in post #4
In these integrals, sec(t) = x/3.
$$\int \frac{3sec^3(t)}{tan(t)}dt$$
$$=3\int \frac{dt}{cos^2(t) sin(t)}$$
$$= 3\int sec^2(t)csc(t)dt$$
$$=3\int \frac{sec(t)(tan^2(t) + 1)dt}{tan(t)}$$
$$=3\int sec(t)tan(t)dt + 3 \int \frac{sec(t)dt}{tan(t)}$$
The first integral is very easy, and the second integral simplifies to ##3\int csc(t)dt## which isn't too bad if you look it up in a table.

After undoing the substitution, I get the same result as I got using partial fractions, after a bit of manipulation.
 

Similar threads

Back
Top