MHB How Do You Solve the Integral of ln|a+b*sin(x)| from 0 to 2π?

  • Thread starter Thread starter Saitama
  • Start date Start date
  • Tags Tags
    Integral
Click For Summary
The integral $$\int_0^{2\pi} \ln|a+b\sin x|\,dx$$ is evaluated by first recognizing the need to split the integral at points of discontinuity, specifically at $x=\pi + \arcsin(a/b)$ and $x=2\pi - \arcsin(a/b)$. The absolute value changes the expression within the integral, leading to different cases for integration. After applying substitutions and differentiating under the integral sign, it is concluded that the integral does not depend on $a$, and evaluating as $a$ approaches zero simplifies the expression. Ultimately, the result is found to be $$I = 2\pi \ln \frac{b}{2}$$, providing a complete solution to the problem.
Saitama
Messages
4,244
Reaction score
93
I am trying to evaluate the following integral:
$$\int_0^{2\pi} \ln|a+b\sin x|\,dx$$
where $0<a<b$.

Source: homework - Find the value of the integral $\int_0^{2\pi}\ln|a+b\sin x|dx$ where $0\lt a\lt b$ - Mathematics Stack Exchange

Attempt:
Consider
$$I(a)=\int_0^{2\pi} \ln|a+b\sin x|\,dx$$
It can be shown that:
$$I(a)=2\int_0^{\pi/2} \ln|a^2-b^2\sin^2x|\,dx$$
Differentiate both sides wrt $a$ to get:
$$\frac{dI}{da}=4a\int_0^{\pi/2}\frac{dx}{a^2-b^2\sin^2x}$$
Next, I rewrite the denominator as:
$$a^2-b^2\sin^2x=\frac{a^2+(a^2-b^2)\tan^2x}{\sec^2x}$$
Using the substitution $\tan x=t$, I obtain the following definite integral:
$$\frac{dI}{da}=4a\int_0^{\infty} \frac{dt}{a^2+(a^2-b^2)t^2}$$
$$\Rightarrow \frac{dI}{da}=\frac{2\pi}{\sqrt{a^2-b^2}}$$
This is where I am stuck. I don't find the result correct because as per the question, $b>a$ due to which $\sqrt{a^2-b^2}$ doesn't make sense. :confused:

Any help is appreciated. Thanks!

EDIT: I thought about it some more and I think I see the error. I found an error at this step:
$$I(a)=2\int_0^{\pi/2} \ln|a^2-b^2\sin^2x|\,dx$$
I cannot simply differentiate the above because of the modulus symbol. $|a^2-b^2\sin^2x|=a^2-b^2\sin^2x$ only if $sin(x)<=a/b$, hence,
$$I(a)=2\left(\int_0^{a/b}\ln(a^2-b^2\sin^2x)\,dx+\int_{a/b}^{\pi/2} \ln(b^2\sin^2x-a^2)\,dx\right)$$
If I differentiate the above, I get:
$$\frac{dI}{da}=2\left(\int_0^{a/b}\frac{2a}{a^2-b^2\sin^2x}\,dx+\int_{a/b}^{\pi/2} \frac{-2a}{b^2\sin^2x-a^2}\,dx\right)$$
which is the same integral as before. :confused:
 
Last edited:
Physics news on Phys.org
We are given

[math] I= \int_0^{2 \pi} \ln \left \vert a+b \sin x \right \vert dx [/math].

The reason your evaluation of the integral was incorrect is because there are two discontinuities on [math] [0, 2 \pi] [/math], at [math] x= \pi + \arcsin(a/b) [/math] and [math] 2 \pi - \arcsin(a/b) [/math]. Therefore, we must split the integral at these points. Also, the absolute value of [math] a+b \sin x [/math] on [math] [0, \pi + \arcsin(a/b)] [/math] and [math] [2 \pi - \arcsin(a/b), 2 \pi] [/math] is just [math] a+b \sin x [/math], while on [math] [\pi + \arcsin(a/b), 2 \pi - \arcsin(a/b)] [/math], it is [math] -a-b \sin x [/math].
Putting these together and setting [math] \pi + \arcsin(a/b)=c [/math] and [math] 2 \pi - \arcsin(a/b)=d [/math], we have

[math] I= \int_0^{c} \ln (a+b \sin x) dx + \int_c^{d} \ln (-a-b \sin x) dx +\int_d^{2 \pi} \ln (a+b \sin x) dx [/math].

Considering I to be a function of a and differentiating under the integral sign, we have

[math] \frac{dI}{da} = \int_0^{c} \frac{1}{a+b \sin x} dx +\int_c^{d} \frac{-1}{-a- b \sin x} dx +\int_d^{2 \pi} \frac{1}{a+b \sin x} dx [/math].

Using the substitution [math] z=\tan \frac{x}{2} [/math] and completing the square, the integrand becomes

[math] \frac{2}{a} \frac{1}{(z+\frac{b}{a})^2+(1-\frac{b^2}{a^2})} [/math].

Integrating (without bounds for the moment), we have

[math] \frac{2}{\sqrt{a^2-b^2}} \arctan \frac{a \tan \frac{x}{2} +b}{\sqrt{a^2-b^2}} +C[/math].

But since [math] a < b [/math], this quantity is imaginary! All is not lost, however. We have

[math] \operatorname{arctanh} z = \frac{1}{i} \arctan iz [/math].

Using this relation, the integral becomes

[math] \frac{2}{\sqrt{b^2-a^2}} \operatorname{arctanh} \frac{a \tan \frac{x}{2} +b}{\sqrt{b^2-a^2}} +C[/math],

which is real.
Evaluating between the three sets of bounds, we have

[math] \frac{dI}{da} = \frac{2}{\sqrt{b^2-a^2}} \left ( \operatorname{arctanh} \frac{b}{\sqrt{b^2-a^2}} -\operatorname{arctanh} \frac{b}{\sqrt{b^2-a^2}} \right ) =0 [/math].

We would appear to be stuck. However, all this implies is that I is not a function of a, as we had previously assumed. Therefore, letting a approach 0 in the original integral gives

[math] I= \int_0^{2 \pi} \ln \left \vert b \sin x \right \vert dx [/math].

Splitting up the argument of the ln, and using the fact that the integral of sine on [math] [0, 2 \pi] [/math] is four times the integral on [math] [0, \frac{\pi}{2}] [/math], we have

[math] I= 2 \pi \ln b + 4 \int_0^{\frac{\pi}{2}} \ln \sin x dx [/math].

Taking [math] \sin x = 2 \sin \frac{x}{2} \cos \frac{x}{2} [/math] and splitting up the logarithmic terms, the second term becomes

[math] 2 \pi \ln 2 + 4 \int_0^{\frac{\pi}{2}} \ln \sin \frac{x}{2} dx + 4 \int_0^{\frac{\pi}{2}} \ln \cos \frac{x}{2} dx [/math].

Substituting [math] u= \frac{x}{2} [/math], the integrals become

[math] 8 \int_0^{\frac{\pi}{4}} \ln \sin u \ du + 8 \int_0^{\frac{\pi}{4}} \ln \cos u \ du [/math].

Substituting [math] u= \frac{\pi}{2} –t [/math] into the cosine integral, it becomes

[math] \int_{\frac{\pi}{4}}^{\frac{\pi}{2}} \ln \sin t \ dt [/math].

Therefore,

[math] 4 \int_0^{\frac{\pi}{2}} \ln \sin x dx= 2\pi \ln 2 + 8 \int_0^{\frac{\pi}{2}} \ln \sin x dx [/math].

Cancelling, and remembering the fact that the integral of sine over a period is equal to four times its integral over a quarter-period, we have

[math] 4 \int_0^{\frac{\pi}{2}} \ln \sin x dx=\int_0^{2 \pi} \ln \sin x dx = -2 \pi \ln 2 [/math].

Substituting into the equation for I and combining logarithmic terms, we obtain (finally)

[math] I= 2 \pi \ln \frac{b}{2} [/math].
 
Last edited:
Thread 'Problem with calculating projections of curl using rotation of contour'
Hello! I tried to calculate projections of curl using rotation of coordinate system but I encountered with following problem. Given: ##rot_xA=\frac{\partial A_z}{\partial y}-\frac{\partial A_y}{\partial z}=0## ##rot_yA=\frac{\partial A_x}{\partial z}-\frac{\partial A_z}{\partial x}=1## ##rot_zA=\frac{\partial A_y}{\partial x}-\frac{\partial A_x}{\partial y}=0## I rotated ##yz##-plane of this coordinate system by an angle ##45## degrees about ##x##-axis and used rotation matrix to...

Similar threads

  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 19 ·
Replies
19
Views
4K
Replies
2
Views
3K
Replies
4
Views
4K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 6 ·
Replies
6
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 29 ·
Replies
29
Views
4K
  • · Replies 6 ·
Replies
6
Views
3K