# I Contravariant components and classic values into code

Tags:
1. Jan 5, 2017

### eliot13

Hello,

I made a small application that simulates the transport a vector on a spherical sphere. The goal is to check the following general equation :

$$\text{D}v^{i}= \text{d}v^{i} +v^{k}\Gamma_{jk}^{i}\text{d}y^{j} = 0$$

with $$v^{i}$$ the contravariant components of vector.

Into my case, I use the variables $$\theta$$ and $$\phi$$.

Following the definition above, I get the 2 following relations :

$$\text{D}v^{\theta} = \text{d}v^{\theta} - v^{\varphi}\sin\theta\,\cos\theta\,\text{d}\varphi=0$$

$$\text{D}v^{\phi} = \text{d}v^{\varphi} + \cot\theta\,(v^{\theta}\text{d}\varphi + v^{\varphi}\text{d}\theta)=0$$

Here's a capture of initial situation on this link : http://imgur.com/a/b6fSP

You can see transported vector which is cyan color and in yellow the geodesic. Basis vectors (which define the tangential plane shadowed) are in pink color (vertical = e_theta and horizontal = e_phi).

My problem is that in code, I use classical values for the transported vector, I mean like I would use basis vectors normalized. But actually, for checking these above equations, I need to use contravariant components knowing that norm of e_theta and e_phi of local basis is not equal to 1 ( norm(e_theta) = R and norm(e_phi)=R sin(theta).

To circumvent this issue, I have applied these factors to the classical components that I use in my code :

$$\vec{V} = \alpha \vec{e_{\theta}}_{Normalized}+\beta \vec{e_{\phi}}_{Normalized}$$

so I have the norm of my vector V equal to :

$$||(\vec{V})|| = \sqrt{\alpha^2+\beta^2}$$

Knowing the norm of e_theta and e_phi, I think I have to take :

$$\vec{V} = \alpha/R \vec{e_\theta}+\beta/(R \sin(\theta)) \vec{e_\phi}$$

Then in my code, I replace "alpha" and "beta" (which are used initially and which enable to reproduce the transport) by :

$$v^{\theta} = \alpha/R$$

and

$$v^{\phi} = \beta/(R \sin(\theta))$$

All of this in order to check the 2 relations at the top.

Unfortunately, I can't get a value equal to zero for these 2 top equations, even with the factors 1/R and 1/(R si(theta)). On the other hand, I have to make notice that transport is well reproduced at the animation, the only issue is that I have not the vanishing of these 2 equations.

Do you think this conversion between components used into the code (alpha and beta) and contravariant components (v^{\theta} and v^{\phi}) is good and justified ?

Right now, I don't know how to fix this problem.

Thanks

Last edited: Jan 5, 2017