Release Notes OVERFLOW 2.3

2.3d 19 February 2021

New Features:
o Added gfortran10 and gfortran10_omp compiler options to Make.sys. Updated CMake compile to handle GFortran 10 as well. (Specific compiler flags are needed for GFortran 10.)

Bugs fixed:
o Fixed usage of Fortran MOD intrinsic to avoid 0 for the 2nd argument.
o Fixed missing OpenMP PRIVATE variables for NQT=103,104.
o Fixed implicit jacobian for K,L-directions in BC#108.


2.3c 29 January 2021

New features:
o 20% speedup in scalar pentadiagonal solver (ILHS=2) (from Chip Jackson).
o Updated Coder transition model from AFT2017b to AFT2019b. (See AIAA-2019-0039 and AIAA-2020-0773 for documentation.) Updated test cases flat_plate_transition and s809.
o Added QCR2013-V and QCR2020 options (IQCR=3 and 4, resp.).
o Added BC#107, viscous wall with specified temperature (read from file). BCFILE is a PLOT3D function file matching the size of the BC subset. Field 1 is Twall (deg. R).
o Added BC#108, viscous wall with specified bleed/suction (read from file). BCFILE is a PLOT3D function file matching the size of the BC subset. Field 1 is MFR, non-dimensionalized by rho_inf*Vref, positive into the domain. Optional field 2 is Twall (deg. R). BCPAR1 is TSWITCH: TSWITCH<0: always set Twall; TSWITCH=0: set Twall where MFR>=0; TSWITCH>0: never set Twall (adiabatic wall). Default is TSWITCH=0 if Twall is supplied.

Rotorcraft interface:
o Fixed rotor PSI calculation when restarting with a new rotation speed.

Rotor disk interface:
o Added Jasim Ahmad’s rdisk_trim_post utility to generate history files of rotor disk control angles and forces and moments from an rdisk_trim file, for plotting.
o The ONEREV file is written whenever a filename is supplied in the rotor namelist file.
o README.rotor_disk explains grid requirements for the rotor disk, including uniform Psi distribution, consistent radial point distribution on all L grid lines, and no grid coning.
o README.rotor_disk describes rotor vs. propeller force and moment definitions.
o Rotor namelist input variable CHORDR is better described as blade chord/R_tip at specified radius fractions.
o Fixed precision problem in rotor disk trim logic.
o Fixed writing ONEREV file when the rotor disk has been split several times in the L-direction.

Bug fixes:
o Fixed a bug in SA turbulence model K-direction convection for 2nd-order or higher-order terms.
o Fixed the copy-to/copy-from boundary condition linearization.
o Fixed uninitialized variable for NQT=103,104.
o Fixed problem with consecutive Scenario.xml motion commands possibly resetting body velocities to zero.
o Fixed problem with near-body grid refinement: hole boundary points were getting interpolation stencils from coarser versions of the same grid, when they should be getting them from the grid that cut the hole (for example).
o Fixed capability to do grid adaption with q.avg output, with data surface grids.


2.3b 12 June 2020

Rotorcraft interface:
o Fixed OMEGA scaling in Helios/OVERFLOW interface (from Rohit Jain, Army CCDC).

Bug fixes:
o Fixed rotor disk model to work for TIPMACH=0.
o Fixed rotor disk model ONEREV files to enforce periodicity.
o Fixed Psi 0 and 360 values for clockwise rotors.
o Fixed rotor disk model C81 table interpolation to enforce no extrapolation off the ends of the table (per Wayne Johnson).

o Fixed some type conversions for ILHS=16,17.
o Fixed dependencies for last item in Make.incl files.
o Cleaned up hinge line option in FOMOCO moment calculation, including pressure, viscous and momentum components as well as total moment. Moments about the hinge line are returned in the x-component, while y- and z-components are set to zero.
o Improved choice of body center for components with rotation, reducing motion integration error in some cases.


2.3a 27 April 2020

New features:
o Added turbulence region type (ITTYP) 104 capability to read a surface function file defining laminar/turbulent (0/1) regions in a grid. This operates in the same way as ITTYP 102 by multiplying the turbulence model source term in the specified subset volume. The user specifies the turbulence region (ITTYP,ITDIR,JTLS,JTLE,KTLS,KTLE,LTLS,LTLE,TLFILE). The file is read into one grid plane in the ITDIR direction and propagated to the rest of the planes. TLFILE is a PLOT3D function file: it is Fortran unformatted; the first record contains the array dimensions in J, K, and L, and the number of field variables (assumed to be 1). The dimension in J should be equal to the range JTLS:JTLE, etc., except in the ITDIR direction, which should have a dimension of 1. The second record contains the transition function, which should be in the range [0,1].
o Created a CMake based build system to mirror the current make system.

Rotorcraft interface:
o Added a capability to couple the rotor disk model with the CAMRAD II comprehensive rotorcraft code. OVERFLOW writes a ONEREV file describing disk loading per blade section, and reads a motion file which defines collective, cyclic, coning, and flapping angles, and tip Mach number. This capability allows rotor trimming (either collective/cyclic control angles or RPM), but does not account for aeroelastic deformation. This motion file conforms to Jasim Ahmad (NASA Ames) definition from Feb. 2020. It is different from the motion file used by the standard rotor interface with rotating blades (EROTOR namelist).

Bugs fixes:
o Fixed a problem with grid adaption and periodic grids, even if you are not refining the periodic grid.
o Fixed axisymmetric BC Jacobians for ILHS=26,27.


2.3 10 October 2019

Major rewrite. This includes new flux linearizations for IRHS=4-6 when using ILHS=26 or 27. Implicit boundary conditions have been included for many, but not all, of the BCs. In addition, similar improvements have been made for the SA model through the use of NQT=103 or 104. CFL ramping is available through new inputs to the &TIMACU namelist. See input parameters RAMP_CFL, CFLMAX_LIMIT, and CFLMIN_LIMIT.


2.2p 8 October 2019

New features:
o Modified GMP constraint implementation to enable CEN=”x,y,z” specification of where the constraint is applied. For example, an Aero6dof Constraint line in the Scenario.xml file,
<Constraint Start=”0″ Duration=”100″ Frame=”Parent” Center=”0,0,0″
Rotate=”1,0,1″ Translate=”1,1,1″ />
specifies that the point (0,0,0) in the Parent coordinate system must have
no translation in x, y, or z, and no rotation about the x- or z-axes.
Note that constraints currently can only restrict motion relative to the
inertial frame (i.e., absolute motion), rather than relative to the parent
o Updated Makelist to automate the compilation test process:
Automatically removes all #include *.F90 statements to avoid duplicate
routines. Rearranges the order of omisoft_global files to ensure
splitm_triq is above splitm. Also suggest using gnu compilers for
checking. (From Chip Jackson.)
o Modified actuator disk BC (BC#44) to hold entropy constant through the disk (from Robert Childs).
o Changed IQCR input parameter so IQCR=1 gives QCR2000, IQCR=2 gives QCR2013.

Rotorcraft interface:
o Fixed a rotor interface onerev file problem which resulted in bad data
being saved at PSI=0 and 360 deg when the PSI angle was very slightly less than 360 deg. This change also removes any restrictions on NPSI (number of psi stations in the onerev file) compared to the number of blades, as well as any implied limits on blade offsets from psi=0, at least in terms of filling the ONEREV data.
o Added a utility program tools/unsupported/onerev2p3d to convert a onerev.txt file (from the OVERFLOW rotor interface) to PLOT3D files onerev.x and onerev.q. These can be plotted using the OVERGRID solution viewer.
o August 2019 Helios updates from Rohit Jain (AFDD).

Bug fixes:
o Fixed Chimera boundary point exchange for DEBUG options.
o Fixed a problem with DEBUG=3 and ITIME=4.
o Fixed IDES=4 option check in VISCHK.
o Fixed problem with interpolation of near-body Q when NBREFINE<0 (not actually doing any near-body refinement).
o Fixed a bug when checking if a TRIQ file subdirectory exists.
o Fixed deallocation of allocatable arrays in GRIDCUT, part of cutlib.F90
(used for generating TRIQ files).
o Fixed a bug in QCR2013 terms, which made it behave like QCR2000.
o Corrected bug in general convection BCs, specifically scaling of averaged
contravariant velocity vs. time metric (from Joe Derlaga).
o Fixed order of arguments in DERHS (DEBUG=3).