0
$\begingroup$

I am aware of the following question Evaluating the surface integral in an FEM (Finite Elements Method) procedure. But they use the volumetric coordinates while I want to use the cartesian isoparametric coordinates.

I am using the cartesian coordinate system ($xyz$) and the corresponding isoparametric coordinate system $({\xi \eta \zeta})$. Then the elemental force vector is given by

$\textbf{f}^e = \int_{\partial{V_{xyz}}} \textbf{N}_i \textbf{t}_n d({\partial{V_{xyz}}})$

$= \int_{\partial{V_{\xi \eta \zeta}}} \textbf{N}_i \textbf{t}_n \Vert \textbf{J} \Vert d({\partial{V_{\xi \eta \zeta}}})$

$ = \sum_{i=1}^{\#Gauss.points} \textbf{N}_i \textbf{t}_n \Vert \textbf{J} \Vert W_i $

Assuming a generalized case where the cartesian element may even be inclined i.e., none of the coordinate valued need to be '0'. My confusion here is that since the integral is over a surface of the tetrahedron, how do I evaluate the Jacobian of the surface? I thought the best way in such a case is to use the 3D shape functions directly. Now the 3D shape function vectors $N$ contain all coordinates ($\xi \eta \zeta$) or I cannot say which will be '0', while the iso-parametric triangle for surface is a 2D triangle with only coordinates ($\xi \eta$). So I would ideally need surface integration points in an ($\xi \eta \zeta$) system? How do I go about this? I see that all FEM notes trivialize this. What am I misunderstanding then?

$\endgroup$
2
  • $\begingroup$ How do you define your "cartesian isoparametric coordinate" system? What are the shape functions for a tetrahedral element in this system? And how do you perform the volume integration? I do not see how such a system is particularly useful. $\endgroup$ Jan 9 at 18:34
  • $\begingroup$ by cartesian isoparametric system i only refer to the traditional isoparametric system used in finite element (with lagrange shape functions). I just wanted to differentiate it from the volumetric coordinates used 'only' in triangles and tetrahedrons. it is useful because i could then generalize it for hexahedral elements unlike the volumetric coordinates. $\endgroup$ Jan 9 at 21:55

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.