You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/benchmarks/linear elasticity/plate-with-hole.md
+22-13Lines changed: 22 additions & 13 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -96,31 +96,40 @@ $$
96
96
97
97
In order to solve the weak formulation, the finite-element method (FEM) can be used. This method discretizes the domain $\Omega$ into so called finite elements that can for example be triangles or quadrilaterals in 2D. On these elements, ansatz functions are defined such that they are continous on the boundaries between elements. These functions form a basis for the solution space for an approximate solution $\boldsymbol{u}_h$ of the problem.
98
98
99
-
## Comparison of approximate solution with analytical solution
99
+
## Solution metrics
100
100
101
-
The approxiamte solution and the analytical solution can be compared with the $L_2$ norm which is defined as
101
+
The approximate solution is compared with the analytical solution using the $L_2$ norm which is defined as
102
102
103
103
$$
104
-
\Vert \boldsymbol{u}\Vert_{L_2} = \sqrt{\int_\Omega \Vert\boldsymbol{u}(\boldsymbol{x})\Vert_2^2 \mathrm d \boldsymbol x}
104
+
\Vert \boldsymbol{f}\Vert_{L_2} = \sqrt{\int_\Omega \left|\boldsymbol{f}(\boldsymbol{x})\right|^2 \mathrm d \boldsymbol x}
105
105
$$
106
106
107
-
and the error in the $L_2$ norm
107
+
This norm is computed for the error between the finite element solution and the analytical solution of displacements i.e., $\Vert \boldsymbol{u}-\boldsymbol{u}_h\Vert_{L_2}$.
108
+
109
+
The maximum displacement error is computed at the nodes of the finite element mesh with respect to the analytical solution.
With these metrices, we can perform a convergence analysis for different approximations $\boldsymbol{u}_h$ which differ in the element size $h$. Plotting the error over the used element-size in a log-log plot lets us determine the convergence order of the approximation.
0 commit comments