Basis vectors derivation from the position vector in general curvilinear coordinates

264 Views Asked by At

In general coordinates, the basis vectors are defined as the derivatives of teh posituon vector: $$\vec e_i = \frac { {\partial \vec R}} {\partial x^i}$$

Nonetheless, the position vector is given by:

$$\vec R= \sum_i x^i \vec e_i$$

and when we do the derivative we have:

$$\frac{\partial \vec R}{\partial x^n} = \sum_i(\frac{\partial x^i}{\partial x^n} \vec e_i + x^i \frac{\partial \vec e_i}{\partial x^n}) = \vec e_n + \sum_ix^i\frac{\partial \vec e_i}{\partial x^n} $$

Since the term $\sum_ix^i\frac{\partial \vec e_i}{\partial x^n}$ is not null in general, then the question is: where am I messing up?

1

There are 1 best solutions below

3
On

You should be considering $\vec R=\sum_i R^i\vec e_i$ (instead of $R=\sum_i x^i\vec e_i$) where each $R^i$ are functions of the euclidean coordinates $x^j$. It is necessary that the Jacobian matrix $J\vec R=\left[\frac{\partial R^i}{\partial x^j}\right]$ has determinant different from zero in the domain of definition of $R$.

A new basis for the space in these coordinates are $$\vec \partial_k=\sum_s\frac{\partial R^s}{\partial x^k}\vec e_s,$$ which give you, may be, a non-orthogonal neither non-orthonormal frame of the space.