Ubuntu Manpage: flow - Eclipse compatible parallel reservoir simulator

Provided by: libopm-simulators-bin_2025.10+ds-4_amd64

NAME

       flow - Eclipse compatible parallel reservoir simulator

SYNOPSIS

       flow [OPTIONS] [ECL_DECK_FILENAME]

DESCRIPTION

Flow, an advanced reservoir simulator for ECL-decks provided by the Open Porous Media project.

Recognized options:
-h,--help
Print this help message and exit

--help-all
Print all parameters, including obsolete, hidden and deprecated ones.

--action-parsing-strictness=STRING
Set strictness of parsing process for ActionX and PyAction. Available options are normal (do not
apply keywords that have not been tested for ActionX or PyAction) and low (try to apply all
keywords, beware: the simulation outcome might be incorrect). Default: "normal"

--add-corners=BOOLEAN
Whether to add overlap cells to the partition that do not share an intersection with other cells
in the partition, but just a corner. . Default: "false"

--allow-distributed-wells=BOOLEAN
Allow the perforations of a well to be distributed to interior of multiple processes. Default:
false

--allow-splitting-inactive-wells=BOOLEAN
Allow inactive (never non-shut) wells to be split across multiple domains. Default: true

--alternative-well-rate-init=BOOLEAN
Use alternative well rate initialization procedure. Default: true

--check-group-constraints-inner-well-iterations=BOOLEAN
Allow checking of group constraints during inner well iterations. Default: true

--check-satfunc-consistency=BOOLEAN
Whether or not to check saturation function consistency requirements. Default: false

--continue-on-convergence-error=BOOLEAN
Continue with a non-converged solution instead of giving up if we encounter a time step size
smaller than the minimum time step size. Default: false

--convergence-monitoring=BOOLEAN
Enable convergence monitoring. Default: false

--convergence-monitoring-cut-off=INTEGER
Cut off limit for convergence monitoring. Default: 6

--convergence-monitoring-decay-factor=SCALAR
Decay factor for convergence monitoring. Default: 0.75

--cpr-reuse-interval=INTEGER
Reuse preconditioner interval. Used when CprReuseSetup is set to 4, then the preconditioner will
be fully recreated instead of reused every N linear solve, where N is this parameter. Default: 30

--cpr-reuse-setup=INTEGER
Reuse preconditioner setup. Valid options are 0: recreate the preconditioner for every linear
solve, 1: recreate once every timestep, 2: recreate if last linear solve took more than 10
iterations, 3: never recreate, 4: recreated every CprReuseInterval. Default: 4

--dbhp-max-rel=SCALAR
Maximum relative change of the bottom-hole pressure in a single iteration. Default: 1

--dp-max-rel=SCALAR
Maximum relative change of pressure in a single iteration. Default: 0.3

--ds-max=SCALAR
Maximum absolute change of any saturation in a single iteration. Default: 0.2

--dwell-fraction-max=SCALAR
Maximum absolute change of a well's volume fraction in a single iteration. Default: 0.2

--ecl-deck-file-name=STRING
The name of the file which contains the ECL deck to be simulated. Default: ""

--ecl-output-double-precision=BOOLEAN
Tell the output writer to use double precision. Useful for 'perfect' restarts. Default: false

--ecl-output-interval=INTEGER
The number of report steps that ought to be skipped between two writes of ECL results. Default: -1

--edge-weights-method=INTEGER
Choose edge-weighing strategy: 0=uniform, 1=trans, 2=log(trans). Default: 1

--enable-adaptive-time-stepping=BOOLEAN
Use adaptive time stepping between report steps. Default: true

--enable-async-ecl-output=BOOLEAN
Write the ECL-formated results in a non-blocking way (i.e., using a separate thread). Default:
true

--enable-async-vtk-output=BOOLEAN
Dispatch a separate thread to write the VTK output. Default: true

--enable-drift-compensation=BOOLEAN
Enable partial compensation of systematic mass losses via the source term of the next time step.
Default: false

--enable-dry-run=STRING
Specify if the simulation ought to be actually run, or just pretended to be. Default: "auto"

--enable-ecl-output=BOOLEAN
Write binary output which is compatible with the commercial Eclipse simulator. Default: true

--enable-esmry=BOOLEAN
Write ESMRY file for fast loading of summary data. Default: false

--enable-logging-fallout-warning=BOOLEAN
Developer option to see whether logging was on non-root processors. In that case it will be
appended to the *.DBG or *.PRT files. Default: false

--enable-opm-rst-file=BOOLEAN
Include OPM-specific keywords in the ECL restart file to enable restart of OPM simulators from
these files. Default: false

--enable-storage-cache=BOOLEAN
Store previous storage terms and avoid re-calculating them. Default: true

--enable-terminal-output=BOOLEAN
Print high-level information about the simulation's progress to the terminal. Default: true

--enable-tuning=BOOLEAN
Honor some aspects of the TUNING keyword. Default: false

--enable-vtk-output=BOOLEAN
Global switch for turning on writing VTK files. Default: false

--enable-well-operability-check=BOOLEAN
Enable the well operability checking. Default: true

--enable-well-operability-check-iter=BOOLEAN
Enable the well operability checking during iterations. Default: false

--enable-write-all-solutions=BOOLEAN
Write all solutions to disk instead of only the ones for the report steps. Default: false

--force-disable-fluid-in-place-output=BOOLEAN
Do not print fluid-in-place values after each report step even if requested by the deck. Default:
false

--force-disable-resv-fluid-in-place-output=BOOLEAN
Do not print reservoir volumes values after each report step even if requested by the deck.
Default: false

--full-time-step-initially=BOOLEAN
Always attempt to finish a report step using a single substep. Default: false

--gpu-device-id=INTEGER
Choose device ID for cusparseSolver or openclSolver, use 'nvidia-smi' or 'clinfo' to determine
valid IDs. Default: 0

--ignore-keywords=STRING
List of Eclipse keywords which should be ignored. As a ':' separated string. Default: ""

--ilu-fillin-level=INTEGER
The fill-in level of the linear solver's ILU preconditioner. Default: 0

--ilu-redblack=BOOLEAN
Use red-black partitioning for the ILU preconditioner. Default: false

--ilu-relaxation=SCALAR
The relaxation factor of the linear solver's ILU preconditioner. Default: 0.9

--ilu-reorder-spheres=BOOLEAN
Whether to reorder the entries of the matrix in the red-black ILU preconditioner in spheres
starting at an edge. If false the original ordering is preserved in each color. Otherwise why try
to ensure D4 ordering (in a 2D structured grid, the diagonal elements are consecutive). Default:
false

--imbalance-tol=SCALAR
Tolerable imbalance of the loadbalancing (default: 1.1). Default: 1.1

--initial-time-step-in-days=SCALAR
The size of the initial time step in days. Default: 1

--inj-mult-damp-mult=SCALAR
Injection multiplier dampening factor (dampening multiplied by this each time oscillation is
detected). Default: 0.9

--inj-mult-min-damp-factor=SCALAR
Minimum injection multiplier dampening factor (maximum dampening level). Default: 0.05

--inj-mult-osc-threshold=SCALAR
Injection multiplier oscillation threshold (used for multiplier dampening). Default: 0.1

--input-skip-mode=STRING
Set compatibility mode for the SKIP100/SKIP300 keywords. Options are 100 (skip SKIP100..ENDSKIP,
keep SKIP300..ENDSKIP) [default], 300 (skip SKIP300..ENDSKIP, keep SKIP100..ENDSKIP) and all (skip
both SKIP100..ENDSKIP and SKIP300..ENDSKIP) . Default: "100"

--linear-solver=STRING
Configuration of solver. Valid options are: cprw (default), ilu0, dilu, cpr (an alias for cprw),
cpr_quasiimpes, cpr_trueimpes, cpr_trueimpesanalytic, amg or hybrid (experimental). Alternatively,
you can request a configuration to be read from a JSON file by giving the filename here, ending
with '.json.'. Default: "cprw"

--linear-solver-ignore-convergence-failure=BOOLEAN
Continue with the simulation like nothing happened after the linear solver did not converge.
Default: false

--linear-solver-max-iter=INTEGER
The maximum number of iterations of the linear solver. Default: 200

--linear-solver-print-json-definition=BOOLEAN
Write the JSON definition of the linear solver setup to the DBG file. Default: true

--linear-solver-reduction=SCALAR
The minimum reduction of the residual which the linear solver must achieve. Default: 0.01

--linear-solver-restart=INTEGER
The number of iterations after which GMRES is restarted. Default: 40

--linear-solver-verbosity=INTEGER
The verbosity level of the linear solver (0: off, 2: all). Default: 0

--load-step=INTEGER
Load serialized state from .OPMRST file. Either a specific report step, or 0 to load last stored
report step. Default: -1

--local-domains-ordering-measure=STRING
Subdomain ordering measure. Allowed values are 'maxpressure', 'averagepressure' and 'residual'.
Default: "maxpressure"

--local-domains-well-neighbor-levels=INTEGER
Number of neighbor levels of cells around wells to include in the same domain during NLDD
partitioning. Default: 1

--local-domains-partitioning-imbalance=SCALAR
Subdomain partitioning imbalance tolerance. 1.03 is 3 percent imbalance. Default: 1.03

--local-domains-partitioning-method=STRING
Subdomain partitioning method. Allowed values are 'zoltan', 'simple', and the name of a partition
file ending with '.partition'. Default: "zoltan"

--local-solve-approach=STRING
Choose local solve approach. Valid choices are jacobi and gauss-seidel. Default: "gauss-seidel"

--local-tolerance-scaling-cnv=SCALAR
Set lower than 1.0 to use stricter convergence tolerance for local solves. Default: 0.1

--local-tolerance-scaling-mb=SCALAR
Set lower than 1.0 to use stricter convergence tolerance for local solves. Default: 1

--local-well-solve-control-switching=BOOLEAN
Allow control switching during local well solutions. Default: true

--matrix-add-well-contributions=BOOLEAN
Explicitly specify the influences of wells between cells in the Jacobian and preconditioner
matrices. Default: false

--max-inner-iter-ms-wells=INTEGER
Maximum number of inner iterations for multi-segment wells. Default: 100

--max-inner-iter-wells=INTEGER
Maximum number of inner iterations for standard wells. Default: 50

--max-local-solve-iterations=INTEGER
Max iterations for local solves with NLDD nonlinear solver. Default: 20

--max-newton-iterations-with-inner-well-iterations=INTEGER
Maximum newton iterations with inner well iterations. Default: 8

--max-pressure-change-ms-wells=SCALAR
Maximum relative pressure change for a single iteration of the multi-segment well model. Default:
1e+06

--max-residual-allowed=SCALAR
Absolute maximum tolerated for residuals without cutting the time step size. Default: 1e+07

--max-single-precision-days=SCALAR
Maximum time step size where single precision floating point arithmetic can be used solving for
the linear systems of equations. Default: 20

--max-temperature-change=SCALAR
Maximum absolute change of temperature in a single iteration. Default: 5

--max-welleq-iter=INTEGER
Maximum number of iterations to determine solution the well equations. Default: 30

--maximum-number-of-group-switches=INTEGER
Maximum number of times a group can switch to the same control. Default: 3

--maximum-number-of-well-switches=INTEGER
Maximum number of times a well can switch to the same control. Default: 3

--maximum-water-saturation=SCALAR
Maximum water saturation. Default: 1

--metis-params=STRING
Configuration of Metis partitioner. You can request a configuration to be read from a JSON file by
giving the filename here, ending with extension '.json.' See
http://glaros.dtc.umn.edu/gkhome/fetch/sw/metis/manual.pdf for available METIS options. Default:
"default"

--milu-variant=STRING
Specify which variant of the modified-ILU preconditioner ought to be used. Possible variants are:
ilu (default, plain ILU), milu_1 (lump diagonal with dropped row entries), milu_2 (lump diagonal
with the sum of the absolute values of the dropped row entries), milu_3 (if diagonal is positive
add sum of dropped row entries, otherwise subtract them), milu_4 (if diagonal is positive add sum
of dropped row entries, otherwise do nothing. Default: "ilu"

--min-strict-cnv-iter=INTEGER
Minimum number of Newton iterations before relaxed tolerances can be used for the CNV convergence
criterion. Default: -1

--min-strict-mb-iter=INTEGER
Minimum number of Newton iterations before relaxed tolerances can be used for the MB convergence
criterion. Default -1 means that the relaxed tolerance is used when maximum number of Newton
iterations are reached. Default: -1

--min-time-step-based-on-newton-iterations=SCALAR
The minimum time step size (in days for field and metric unit and hours for lab unit) can be
reduced to based on newton iteration counts. Default: 0

--min-time-step-before-shutting-problematic-wells-in-days=SCALAR
The minimum time step size in days for which problematic wells are not shut. Default: 0.01

--network-max-outer-iterations=INTEGER
Maximum outer number of iterations in the network solver before giving up. Default: 10

--network-max-pressure-update-in-bars=SCALAR
Maximum pressure update in the inner network pressure update iterations. Default: 5

--network-max-strict-outer-iterations=INTEGER
Maximum outer iterations in network solver before relaxing tolerance. Default: 10

--network-max-sub-iterations=INTEGER
Maximum number of sub-iterations to update network pressures (within a single well/group control
update). Default: 20

--network-pressure-damping-factor=SCALAR
Damping factor in the inner network pressure update iterations. Default: 0.1

--newton-max-iterations=INTEGER
The maximum number of Newton iterations per time step. Default: 20

--newton-max-relax=SCALAR
The maximum relaxation factor of a Newton iteration. Default: 0.5

--newton-min-iterations=INTEGER
The minimum number of Newton iterations per time step. Default: 2

--newton-relaxation-type=STRING
The type of relaxation used by Newton method. Default: "dampen"

--nldd-local-linear-solver=STRING
Configuration of NLDD local linear solver. Valid options are: ilu0 (default), dilu, cpr_quasiimpes
and amg. Alternatively, you can request a configuration to be read from a JSON file by giving the
filename here, ending with '.json.'. Default: "ilu0"

--nldd-local-linear-solver-max-iter=INTEGER
The maximum number of iterations of the NLDD local linear solver. Default: 200

--nldd-local-linear-solver-reduction=SCALAR
The minimum reduction of the residual which the NLDD local linear solver must achieve. Default:
0.01

--nldd-num-initial-newton-iter=INTEGER
Number of initial global Newton iterations when running the NLDD nonlinear solver. Default: 1

--nonlinear-solver=STRING
Choose nonlinear solver. Valid choices are newton or nldd. Default: "newton"

--num-local-domains=INTEGER
Number of local domains for NLDD nonlinear solver. Default: 0

--num-overlap=INTEGER
Numbers of layers overlap in parallel partition. Default: 1

--num-satfunc-consistency-sample-points=INTEGER
Maximum number of reported failures for each individual saturation function consistency check.
Default: 5

--nupcol-group-rate-tolerance=SCALAR
Tolerance for acceptable changes in VREP/RAIN group rates. Default: 0.001

--opencl-ilu-parallel=BOOLEAN
Parallelize ILU decomposition and application on GPU. Default: true

--opencl-platform-id=INTEGER
Choose platform ID for openclSolver, use 'clinfo' to determine valid platform IDs. Default: 0

--output-dir=STRING
The directory to which result files are written. Default: ""

--output-extra-convergence-info=STRING
Provide additional convergence output files for diagnostic purposes. "none" gives no extra output
and overrides all other options, "steps" generates an INFOSTEP file, "iterations" generates an
INFOITER file. Combine options with commas, e.g., "steps,iterations" for multiple outputs.
Default: "none"

--output-interval=INTEGER
Specify the number of report steps between two consecutive writes of restart data. Default: 1

--output-mode=STRING
Specify which messages are going to be printed. Valid values are: none, log, all (default).
Default: "all"

--owner-cells-first=BOOLEAN
Order cells owned by rank before ghost/overlap cells. Default: true

--parameter-file=STRING
An .ini file which contains a set of run-time parameters. Default: ""

--parsing-strictness=STRING
Set strictness of parsing process. Available options are normal (stop for critical errors), high
(stop for all errors) and low (as normal, except do not stop due to unsupported keywords even if
marked critical. Default: "normal"

--partition-method=INTEGER
Choose partitioning strategy: 0=simple, 1=Zoltan, 2=METIS, 3=Zoltan with all cells of well
represented by one vertex. Default: 3

--pre-solve-network=BOOLEAN
Whether to pre solve and iterate the network model at start-up. Default: true

--pressure-max=SCALAR
Maximum absolute pressure. Default: 1e+99

--pressure-min=SCALAR
Minimum absolute pressure. Default: -1e+99

--pressure-scale=SCALAR
Scaling of pressure primary variable. Default: 1

--pri-var-oscilation-threshold=SCALAR
The threshold value for the primary variable switching conditions after its meaning has switched
to hinder oscillations. Default: 1e-05

--print-parameters=INTEGER
Print the values of the run-time parameters at the start of the simulation. Default: 2

--project-saturations=BOOLEAN
Option for doing saturation projection. Default: false

--regularization-factor-wells=SCALAR
Regularization factor for wells. Default: 100

--relaxed-linear-solver-reduction=SCALAR
The minimum reduction of the residual which the linear solver need to achieve for the solution to
be accepted. Default: 0.01

--relaxed-max-pv-fraction=SCALAR
The fraction of the pore volume of the reservoir where the volumetric error (CNV) may be violated
during strict Newton iterations. Default: 0.03

--relaxed-pressure-tol-msw=SCALAR
Relaxed tolerance for the MSW pressure solution. Default: 10000

--relaxed-well-flow-tol=SCALAR
Relaxed tolerance for the well flow residual. Default: 0.001

--save-step=STRING
Save serialized state to .OPMRST file. Either a specific report step, "all" to save all report
steps or ":x" to save every x'th step.Use negative values of "x" to keep only the last written
step, or "last" to save every step, keeping only the last. Default: ""

--scale-linear-system=BOOLEAN
Scale linear system according to equation scale and primary variable types. Default: false

--sched-restart=BOOLEAN
When restarting: should we try to initialize wells and groups from historical SCHEDULE section.
Default: false

--serial-partitioning=BOOLEAN
Perform partitioning for parallel runs on a single process. Default: false

--shut-unsolvable-wells=BOOLEAN
Shut unsolvable wells. Default: true

--solve-welleq-initially=BOOLEAN
Fully solve the well equations before each iteration of the reservoir model. Default: true

--solver-continue-on-convergence-failure=BOOLEAN
Continue instead of stop when minimum solver time step is reached. Default: false

--solver-growth-factor=SCALAR
The factor time steps are elongated after a successful substep. Default: 2

--solver-max-growth=SCALAR
The maximum factor time steps are elongated after a report step. Default: 3

--solver-max-restarts=INTEGER
The maximum number of breakdowns before a substep is given up and the simulator is terminated.
Default: 10

--solver-max-time-step-in-days=SCALAR
The maximum size of a time step in days. Default: 365

--solver-min-time-step=SCALAR
The minimum size of a time step in days for field and metric and hours for lab. If a step cannot
converge without getting cut below this step size the simulator will stop. Default: 1e-12

--solver-restart-factor=SCALAR
The factor time steps are elongated after restarts. Default: 0.33

--solver-verbosity=INTEGER
Specify the "chattiness" of the non-linear solver itself. Default: 1

--strict-inner-iter-wells=INTEGER
Number of inner well iterations with strict tolerance. Default: 40

--strict-outer-iter-wells=INTEGER
Number of newton iterations for which wells are checked with strict tolerance. Default: 6

--temperature-max=SCALAR
Maximum absolute temperature. Default: 1e+09

--temperature-min=SCALAR
Minimum absolute temperature. Default: 0

--threads-per-process=INTEGER
The maximum number of threads to be instantiated per process ('-1' means 'automatic'). Default: 2

--time-step-after-event-in-days=SCALAR
Time step size of the first time step after an event occurs during the simulation in days.
Default: -1

--time-step-control=STRING
The algorithm used to determine time-step sizes. Valid options are: 'pid' (default),
'pid+iteration', 'pid+newtoniteration', 'iterationcount', 'newtoniterationcount' and 'hardcoded'.
Default: "pid+newtoniteration"

--time-step-control-decay-damping-factor=SCALAR
The decay rate of the time step decrease when the target iterations is exceeded. Default: 1

--time-step-control-decay-rate=SCALAR
The decay rate of the time step size of the number of target iterations is exceeded. Default: 0.75

--time-step-control-file-name=STRING
The name of the file which contains the hardcoded time steps sizes. Default: "timesteps"

--time-step-control-growth-damping-factor=SCALAR
The growth rate of the time step increase when the target iterations is undercut. Default: 3.2

--time-step-control-growth-rate=SCALAR
The growth rate of the time step size of the number of target iterations is undercut. Default:
1.25

--time-step-control-safety-factor=SCALAR
Value to be multiplied with the time step control tolerance to ensure that the target relative
change is lower than the tolerance. Default: 0.8

--time-step-control-target-iterations=INTEGER
The number of linear iterations which the time step control scheme should aim for (if applicable).
Default: 30

--time-step-control-target-newton-iterations=INTEGER
The number of Newton iterations which the time step control scheme should aim for (if applicable).
Default: 8

--time-step-control-tolerance=SCALAR
The tolerance used by the time step size control algorithm. Default: 0.1

--time-step-verbosity=INTEGER
Specify the "chattiness" during the time integration. Default: 1

--tolerance-cnv=SCALAR
Local convergence tolerance (Maximum of local saturation errors). Default: 0.01

--tolerance-cnv-energy=SCALAR
Local energy convergence tolerance (Maximum of local energy errors). Default: 0.01

--tolerance-cnv-energy-relaxed=SCALAR
Relaxed local energy convergence tolerance that applies for iterations after the iterations with
the strict tolerance. Default: 1

--tolerance-cnv-relaxed=SCALAR
Relaxed local convergence tolerance that applies for iterations after the iterations with the
strict tolerance. Default: 1

--tolerance-energy-balance=SCALAR
Tolerated energy balance error relative to (scaled) total energy present. Default: 1e-07

--tolerance-energy-balance-relaxed=SCALAR
Relaxed tolerated energy balance error that applies for iterations after the iterations with the
strict tolerance. Default: 1e-06

--tolerance-mb=SCALAR
Tolerated mass balance error relative to total mass present. Default: 1e-07

--tolerance-mb-relaxed=SCALAR
Relaxed tolerated mass balance error that applies for iterations after the iterations with the
strict tolerance. Default: 1e-06

--tolerance-pressure-ms-wells=SCALAR
Tolerance for the pressure equations for multi-segment wells. Default: 1000

--tolerance-well-control=SCALAR
Tolerance for the well control equations. Default: 1e-07

--tolerance-wells=SCALAR
Well convergence tolerance. Default: 0.0001

--update-equations-scaling=BOOLEAN
Update scaling factors for mass balance equations during the run. Default: false

--use-gmres=BOOLEAN
Use GMRES as the linear solver. Default: false

--use-implicit-ipr=BOOLEAN
Compute implict IPR for stability checks and stable solution search. Default: true

--use-multisegment-well=BOOLEAN
Use the well model for multi-segment wells instead of the one for single-segment wells. Default:
true

--use-update-stabilization=BOOLEAN
Try to detect and correct oscillations or stagnation during the Newton method. Default: true

--water-only-threshold=SCALAR
Cells with water saturation above or equal is considered one-phase water only. Default: 1

flow October 2025 FLOW(1)