Improvements:
Reworked namelist parsing so that repeating each namelist is no longer needed. See change 130.
Modified FOMOCO routines to reduce MPI overhead. NFOMO will no longer be reset by dynamic cases to output every step. See change 133.
CFL ramping reworked to accommodate grid sequencing.
Added a ramping factor for rotor disk simulations (variable IRAMP in &RDISK_TRIM for MODEL=0).
Added tri_to_vtk utility to visualize all tri files (including triq_dp) by converting to VTK format.
Bug fixes:
Fixed 2D grid adaption bug.
Fixed missing q.save file when using the SAVE_SPLIT option.
Fixed potential segfault in INTOUT_WRITS.
Fixed bug in trip line ITTYP=103 and enabled the simultaneous specification of ITTYP 102 and 103.
Fixed indexing error for multi-equation turbulence models when using GLS.
Disabled temperature dependent gamma due to incorrect treatment.
Removed features:
DEBUG_SPLIT preprocessor option.
CONSERVE_MEM.
Vortex generator source term BC (IBTYP=601).
Non-point matched fold-over BCs (IBTYP=54, 55, 56).
Steger-Warming flux-split LHS option (ILHS=1).
Yee symmetric TVD and AUSM+ schemes (IRHS=2, 3).
Smoothing options (ISPEC=-1,1,3).
Baldwin-Barth turbulence model (NQT=100).
Simple wall distance calculator (WALLDIST=-1,1).
SCE central scheme (IUPC=0).
Improvements:
Improved force and moment calculation with non-zero turbulent eddy viscosity at boundaries, to accommodate surface roughness models.
Enabled Knopp wall roughness model for Wilcox k-omega turbulence model.
Allowed force and moment calculation on data surface grids.
Added DEBUG=6 output to q.trans file for AFT and LM transition models.
Added skin friction coefficient as variable 6 in DEBUG=1 q.turb file.
Created split-grid restart capability (when using the same number of MPI groups) by using SAVE_SPLIT=.T. in namelist GLOBAL. This avoids residual spikes on restart due to auto-split grid boundaries.
Changed from Spalart Allmaras SA-neg-noft2 to SA-neg (i.e., turned on the ft2 term). This affects NQT=102,104,302, and should improve initial convergence with little effect on final result.
Adde …