[GeoMechanicsApplication] Fixed the bug related to shape-function-gradients for thermal line el…

[mnabideltares]

📝 Description
There is a bug detected in the calculation of shape function gradients for line thermal element. The shape fuction gradient is calculated based on the local coordinates. Later it has to be multiplied by the inverse of the Jacobian in order to be projected to the global coordinates. However, a such multiplication is missing.

🆕 Changelog

Added a statement to divide the shape-function-gradients to the determinant of Jacobian. In 1D, this is equvalent to multiplication with inverse Jacobian

[rfaasse]

Thanks for splitting this out of the other PR, a very easy to review PR like this! I was just wondering if we can illustrate this bug already without point source flux, but I don't have any other comments, so if that's unnecessarily hard to do without the other PR, this is good to go from my point of view

[WPK4FEM]

Dear Mohammed,
After spending some effort on rewriting the if else structure for the computation of the dN_dX container, the result of that is not better than what you have shown here.
After consulting with Anne and Richard we would appreciate if you ad a comment before line 74 e.g.
// ShapeFunctionsIntegrationsPointsGradients does not allow for the line element in 2D/3D configuration and will produce errors. To circumvent this, the dN_dX_container is separately computed with correct dimensions for the line element.
Regards, Wijtze Pieter

[WPK4FEM]

Thank you.