Changes between Version 1 and Version 2 of LineTransfer

Timestamp:

01/10/17 16:23:56 (8 years ago)

Author:

Jonathan

Comment:

—

LineTransfer

@@ -1 @@
-So - data structures?
+== Data structures ==
+There are two primary data structures used by the linetransfer routines in AstroBEAR
 
-Type ZoneLayout
-
-integer ::  ZoneLayout(3,2,nZones)
-integer :: ZoneMembers(nZones, MaxProcsPerZone)
-
-Type PatchBoundaryDataPerZonePerDirection(nMyZones, nRays)
-  integer :: mGlobal !originating patch boundaries - used for sorting
-  integer :: Rays(2,2) ! given indexing based on direction and level
-  REAL :: Depth (:,:) 
-  LOGICAL :: Integrated ! stores whether each ray is an integrated value - or a local value
-  REAL :: zvals(:,:) ! stores exit value for 'z' data.  used to determine which rays should be accumulated.
-end type
-
-Patch->Zones->RayBoundaries -> Communication to Upwind Zones
-
-Each processor orders Patch Boundaries within a zone and integrates the depth.  At this point every patch boundary should be integrated and is suitable for updating physical zones.  So now - loop over patches and find boundaries and update optical depth to all of the cells.
-
-SUBROUTINE Calculate_Patch_Boundary_Data_For_Local_Zones(Info, direction)
-! Calculates optical depth from grid edge to each cell and creates a Patch_Boundary_Data_Per_Zone_Per_Direction object(s)
+1.  !PatchDef - This stores information about patches needed to do the ray tracing
+1.  !RayGroupDef - This stores information about a particular group of parallel rays
 
 
-END
+The basic algorithm to calculate optical depths is as follows:
 
-SUBROUTINE Integrate_Patch_Boundary_Data_For_Zone(zone, direction)
+==  Get Local Depths ==
+Integrate across each patch and store the local (patch) contribution to the optical depth at each cell center as well as the total optical depth across the patch.
 
-END
+== Collect Patches ==
+For each patch, build a list of shadow-patches and send/receive the optical depth across each shadow patch.
+
+== Add Shadows ==
+For each patch, sort the shadow patches by their earliest 'z-value'
+Then accumulate shadows for each patch
+
+
+Currently the code just calculates the optical depth to each zone.
+
+The Krumholz et al scheme, calculates ionization rates as well as optical depths when it traverses the tree.
+
+1.)  Modify tree traversal scheme to also calculate the ionization rate in addition to the optical depth
+2.)  Add routines to take ionization rate and calculate heating/cooling rates and calculate CFL
+3.)  Add routine to update energy and ionization values
+4.)  Add sub-cycling
+5.)  Will need to store ionization & heating/cooling rates in addition to optical depths at patch boundaries for AMR to work with sub-cycling so we can update ghost zones
+6.)  Implement healpix tree traversal scheme.
+