Why is this proof of the Frent-Serret formulae wrong?

132 Views Asked by At

I'm aware that similar questions have been asked, but none of them appear relevant to my own question.

By definition and by simple geometric observations, $N,B,T$ are all mutually perpendicular unit vectors that satisfy the following, where $\kappa$ is the curvature of some curve and $\tau$ is its torsion.

$$\begin{align*} \\N\cdot B=B\cdot T&=N\cdot T=0 \\ T'&=\kappa N \\ B'&=-\tau N\ \\ \tau&=-N\cdot B' \\ \kappa&=T'\cdot N \end{align*}$$

Since they are all perpendicular to one another, and of unit length, $N\cdot N=B\cdot B=T\cdot T=1$, and $N\cdot B=B\cdot T=T\cdot N=0$, where I use $\cdot$ to denote the dot product. The derivative of zero is just zero, so:

$$\begin{align*}&0=(N\cdot B)'=N'\cdot B+B'\cdot N=N'\cdot B-\tau=0 \\&0=(N\cdot T)'=N'\cdot T+T'\cdot N=N'\cdot T+\kappa=0 \\\therefore\quad\tau&=N'\cdot B\implies\tau B=N'\cdot(B\cdot B)\implies N'=\tau B \\\therefore\quad\kappa&=-N'\cdot T\implies -\kappa T=N'\cdot(T\cdot T)\implies N'=-\kappa T \end{align*}$$

But the formulae would have me believe that instead $N'$ is not equal to either $\tau B$ or $-\kappa T$, but instead is equal to their sum. My workings would suggest that $\tau B-\kappa T=\mathbf{2}N'$, which is apparently wrong! Where did I make a mistake?

1

There are 1 best solutions below

1
On BEST ANSWER

The error starts with the two conclusions with "$\therefore$". For example, it doesn't make sense to write $(N'\cdot B)\cdot B=N'\cdot (B\cdot B)$ because you're confused with scalar multiplication of a number with a vector, and the dot product between two vectors.

What you can do to fix the problem is to write $N'=aT+bN+cB$, and then use dot product to get $a,b,$ and $c$. Namely $$a=N'\cdot T=-\kappa, b=N'\cdot N=0,c=N'\cdot B=\tau,$$ hence $$N'=-\kappa T+\tau B.$$