Geometry#
Struphy models are implemented in curvilinear coordinates and can be run on a variaty of mapped domains. Besides analytical mappings, there are discrete spline mappings available (IGA approach).
The (physical) domain \(\Omega \subset \mathbb R^3\) is an open subset of \(\mathbb R^3\), defined by a diffeomorphism
mapping points \(\boldsymbol{\eta} \in (0, 1)^3 = \hat\Omega\) of the (logical) unit cube to physical points \(\mathbf x \in \Omega\). The corresponding Jacobain matrix \(DF:\hat\Omega \to \mathbb R^{3\times 3}\), its volume element \(\sqrt g: \hat\Omega \to \mathbb R\) and the metric tensor \(G:\hat\Omega \to \mathbb R^{3\times 3}\) are defined by
Only right-handed mappings (\(\textnormal{det}(DF) > 0\)) are admitted.
Contents:
- Available domains
- Base classes
- Evaluation kernels
- Mapping kernels
spline_3d()spline_3d_df()spline_2d_straight()spline_2d_straight_df()spline_2d_torus()spline_2d_torus_df()cuboid()cuboid_df()orthogonal()orthogonal_df()colella()colella_df()hollow_cyl()hollow_cyl_df()powered_ellipse()powered_ellipse_df()hollow_torus()hollow_torus_df()shafranov_shift()shafranov_shift_df()shafranov_sqrt()shafranov_sqrt_df()shafranov_dshaped()shafranov_dshaped_df()
- Transform kernels
- Utilities