Surface integral over a cylinder problem

multiple integralmultivariable-calculus

I have a generic cylinder $x^2+y^2=a^2, 0≤z≤h$ labeled as $G$. It has an outward-pointing unit normal vector $\vec{n}$.

As seen in the picture below:

cylinder

Now I am asked to solve this integral on $G$:

$$\iint\limits_{G}\vec{n}⋅\vec{n}ds$$

Note: I know that the most convenient way to solve that problem, is just to do the calculation of the dot product which equals to 1. From there I can easily proceed to the calculation of the surface area of $G$ which equals to $2\pi ah$.

And here is my problem. I did not pay attention that $\vec{n}⋅\vec{n}=1$ in our case. I went straight forward and noticed that $\vec{n}=∇f(x,y,z)=(2x,2y,0)$. Then I used the known equation of surface integral:

$$\iint\limits_{S}f(x,y,z)ds=\iint\limits_{D}f(x,y,g(x,y))\sqrt{(z_x)^2+(z_y)^2+1} dA$$

And tried to solve the following integral:

$$\iint\limits_{S}(2x,2y,0)⋅(2x,2y,0)ds=4\iint\limits_{S}x^2+y^2ds=4\iint\limits_{D}x^2+y^2dA=4\int_{θ=0}^{2\pi}\int_{r=0}^{a}r^3drdθ$$

Here I got a wrong answer. The problem is that there is no reference to value of $z$ which limits the height of that cylinder. I think that I have missed something here, because my way of solving seems logical to me in general, but I am sure that there is a right way to deal with the height constraint here that I can't understand well.

Best Answer

The small problem is that $\vec n$ needs to be normalized. But your bigger problem is that you are calculating the integral on the wrong surface. When you integrate $r$ from $0$ to $a$, and $\theta$ from $0$ to $2\pi$ (not $4\pi$), you are calculating the integral on the bottom cap of the cylinder, not on the side. So solving the first issue, $$\vec n=\frac{1}{2\sqrt{x^2+y^2}}(2x,2y,0)$$ Then the integrand will be $1$. For the second issue, the first integral is along the circumference, $dl$ from $0$ to $2\pi a$, and $dz$ from $0$ to $h$