New analyses - collinearity
Added the following analyses
- Collinearity, defined as
(\mathbf{\hat{n} \cdot \hat{b}})^2, where also the tangent vector has been normalised. - An alternative definition of twist, removing the assumption that
\mathbf{u}is orthogonal to\mathbf{\hat{T}}(See Kamien's review). This is stored intwist_alt. This definition uses the normal to the ring,\mathbf{\hat{n}}, and divides the integrand by1-(\mathbf{\hat{n}\cdot \hat{b}})^2The result appears to be more linear, but not too much. I need to write down the derivation.
Edited by Luca Tubiana