Moment of inertia

[skfegan]

Currently the code is using the diagonal elements of the moment of inertia. I think perhaps it should be using the eigenvalues of the moment of inertia tensor.

[ioanaapapa]

I think we still want to use the diagonal elements for the torque averaging but the eigenvalues are used to sort the corresponding eigenvectors so that the direction of the axis is always consistent. I changed the moment of inertia calculation in #228 , using the same rotational coordinates as for the torques to calculate the eigenvalues, and the moment of inertia axes are now consistent with the torque axes, and results are a much closer match to previous code results. I think that might be fixing this?