1

Not too sure about this.

A surface is described as $y=\phi(x,t)$ Find a unit vector orthogonal to the surface.

I was thinking of a new function $f(x,y,t) = y - \phi(x,t) = 0$ and taking the gradient over the magnitude of the gradient.

Any help appreciated, thanks.

Mike Miller
  • 2,453

1 Answers1

2

Let $F(x,y,z)=0$ be the equation of a surface ,say $S$, such that it is continuously differentiable. According to the definition at any point $P(x_0,y_0,z_0)\in S$, the vector $$\nabla F|_P$$ is the normal vector you are searching for.

Mikasa
  • 67,374
  • I think what confuses me is the surface is defined in terms of time. I'm not sure if that is 2D or 3D. – Mike Miller Sep 29 '13 at 17:20
  • @BritMiller: I meant that your assumption of taking $f$ is right. – Mikasa Sep 29 '13 at 17:23
  • I think I understand. So: $\frac{(-\frac{\partial \phi}{\partial x}, 1)}{\sqrt{(\frac{\partial \phi}{\partial x})^2+1}}$? – Mike Miller Sep 29 '13 at 17:41
  • @BritMiller: It would be $$\frac{(-\frac{\partial \phi}{\partial x}, 1,-\frac{\partial \phi}{\partial t})}{\sqrt{(\frac{\partial \phi}{\partial x})^2+1+(\frac{\partial \phi}{\partial t})^2}}$$ – Mikasa Sep 29 '13 at 17:55
  • @ Babak S. That's what I initially thought, but the answer is given as the one I stated - I just checked. I actually took the partial w.r.t. z rather than t, but I was guessing. – Mike Miller Sep 29 '13 at 18:15
  • As in taking the function to be $f(x,y,z,t)$ I should say. – Mike Miller Sep 29 '13 at 18:33
  • Nice work, Babak! – amWhy Sep 30 '13 at 00:04