There is an inconsistency between the Peregrine and Peregrine HDF5 scan path extraction, which causes issues with the AdditiveFOAM simulations. The two behaviors are as follows:
- In Peregrine database scan path extraction, the initial point for the part's scan path is a spot melt with
tParam = 0.0. This sets the initial location of the laser and the initial time of the simulation is t = 0 s, which is assumed by the additivefoam/solidification_region_reduced app.
- In PeregrineHDF5 database scan path extract, the initial point for the part's scan path is a spot melt with
tParam = X s with X being the elapsed time in the layer before the part begins melting. This helps with part melting order, but means that the AdditiveFOAM(MynaApp).update_region_start_and_end_times() function gives the wrong behavior by assuming the initial elapsed time is 0.0 s.