Changes in PFLOTRAN 6.0¶
The following has changed in PFLOTRAN 6.0:
PETSc release updated to 3.21
Fixed bug when using relative paths to files
Added buffer erosion and copper corrosion model waste form process model
Fixed bug in mineral tst rate expression affinity factor derivative
Added tightly coupled well model for sco2 and hydrate
Added analytical waste form mechanism
Added error message in ufd decay when a material is not found for the kd values
Added exponential freezing curve in hydrate mode
Added temperature dependence to Lambda reaction sandbox
Added the ability to output cell IDs (local, ghosted, petsc, natural) for debugging purposes
Added the option to read well model constraints from FLOW_CONDITIONS
Added an error message when SUBSURFACE_TRANSPORT is specified in the SIMULATION block but a CHEMISTRY block does not exist.
Deprecated -stochastic command line argument in favor of -multi_realization
Added heterogeneous Waxman-Smits clay conductivity for ERT process model
Added sco2 as a component in HYDRATE mode
Fixed mass balance file headers for GENERAL mode
Added CI scripting to catch trailing spaces in source files
Added heterogeneous Archie parameters to material property
Added error messaging to ensure that ERT electrodes are mapped to a cell
Improved DBASE error messaging when a parameter is not found in the database
Added screen output of number of grid cells, min/max grid cells per process, and number of active and inactive cells
Added wedge and pyramid shape function for geomechanics
Added pressure-regulated source term in source/sink sandbox
Added SCO2 (supercritical CO2) process model
Eliminated the printing of min/max/mean in gold files for regression tests with one cell
Added the ability to print REGIONs to HDF5 files for debugging conceptual models
Added ‘atm’ as a pressure unit
Refactored geomechanics tet shape function
Added error messaging when a CHARACTERISTIC_CURVE is used without a flow mode
Refactored ERT to use a mixture of two solution concentrations when calculating bulk conductivity (e.g., differing ionic strengths)
Implemented stress/strain linkage between geomechanics and ERT
Expanded EOS IF97 density/enthalpy to “region 3”
Added the ability to run transport with solely immobile species
Added the parameter process model
Refactored the source/sink sandbox to use REGIONs instead of coordinates and cell IDs
Added heterogeneous surface electrical conductivity for ERT
Added anisotropic tortuosity
Added error messaging when tracer is factored into ERT electrical conductivity but no tracer transport is simulated
Refactored kinetic reaction inhibition and added SMOOTHSTEP as an option