In #283 we revised the 1D-2D line exchange levels. Specifically, the exchange level (also called dpumax) is adjusted if the 1D-2D line crosses an obstacle (also called "levee").
After presenting this to HHNK, we realized that this is not (yet) entirely correct.
This is currently only done for 1D-2D lines that come from a potential_breach/exchange_line (so-called 'displaced' 1D-2D lines). It should be done for all 1D-2D lines. This is backwards incompatible but according to @Jonasvsl and @leendertvanwolfswinkel that is fine.
When intersecting with obstacles, there is the question "what is the 1D-2D line geometry". The 2D end of this line is an arbitrary location in the 2D cell. This should be changed. Currently the 1D-2D line geometry is (for potential_breaches) the actual line inserted by the user, (for exchange lines) the shortest line from the 1D node to the exchange line, (for non-displaced 1D2D) a line of length 0, starting and ending at the 1D node. This is incorrect. It should be:
for potential_breaches, the actual line inserted by the user (so no change there)
for exchange_lines, the line from the 1D node to the 2D cell center
for others the line from the 1D node to the 2D cell center
In #283 we revised the 1D-2D line exchange levels. Specifically, the exchange level (also called
dpumax) is adjusted if the 1D-2D line crosses an obstacle (also called "levee").After presenting this to HHNK, we realized that this is not (yet) entirely correct.