- Allow for VoronoiFVM v3, PreallocationTools v1
- Update
prampsuch that it always solves for the first bias. - Use one-sided "upwinded" gradient evolution
Up to now, electric field was evaluated by averaged gradients in the inner of the domain and by one-sided gradients at the boundaries. This inconsistency led to some mesh size dependent and thus unphysical concentration jumps in the first and last grid cells.
Now the we always use one-sided evaluation, switching the direction in the center of the domain. This gives a seemingly physically more consistent picture.
The issue is semi-solved and should be investigated further.
- Bump version to 1.0
- Update README
Release as presented at Lake arrowhead workshop
- rename internal files
- rename ICMPBP to SolverCore
- rename ICMBPData to AugmentedPBData
- Add SciMLPublic
- Ensure explicit imports
- Move code from pluto notebook to normal files
- Add lots of docs with help of Claude