Run a model in JutulDarcy

One model compiles for more than one simulator. ps.compile(m) writes a deck for OPM Flow, the default; ps.compile(m, "jutuldarcy") writes the same model for JutulDarcy (an open-source simulator in Julia, measured on 0.4.0). The deck is read by PetroScript’s names either way, so the two runs compare by name.

import petroscript as ps

m = ps.load("opm-tests/spe1/SPE1CASE2.DATA")
jutul = ps.compile(m, "jutuldarcy")
assert (jutul.target, list(jutul.files)) == ("jutuldarcy", ["SPE1CASE2.DATA"])
jutul.write("spe1-jutul")

With a JutulDarcy runner named by PETROSCRIPT_JUTULDARCY (install), run = jutul.run(folder="spe1-jutul/run") runs it, and run.field("oil_production_rate"), run.wells("bottom_hole_pressure") read it as a Flow run is read, one row per report. Cell arrays, the grid and the log are not read from a JutulDarcy run yet. A first run starts Julia and checks the deck first (its pre-flight), so it takes longer than Flow’s; jutul.run(folder=..., preflight=False) skips the check.

What JutulDarcy refuses

The compile checks the model against what JutulDarcy honours exactly, and refuses the rest with ps.CompileError: [target.unsupported] for a keyword it cannot honour (a dissolution-rate limit, DRSDT; a well’s efficiency; a production network; an equation-of-state fluid; every keyword never measured on JutulDarcy), and [target.item_dropped] for an item of a keyword it would read past (a THP limit or lift table on a control, a group control, a stated D-factor). A group or region summary quantity only warns: JutulDarcy does not report it. The compile for OPM Flow is never affected.

SPE1CASE1 states DRSDT (gas may not re-dissolve), so it is refused; SPE1CASE2 states none and compiles. To run SPE1CASE1 anyway, remove the statement, knowing it changes the physics:

case1 = ps.load("opm-tests/spe1/SPE1CASE1.DATA")
try:
    ps.compile(case1, "jutuldarcy")
except ps.CompileError as refused:
    print(str(refused).splitlines()[0])   # [target.unsupported] jutuldarcy cannot honour DRSDT ...
print(case1.find("schedule/*/dissolution_limits"))      # where the model holds it
without = case1.set("schedule/2015-01-01/dissolution_limits", None)
ps.compile(without, "jutuldarcy")                       # compiles: gas may now re-dissolve

examples/two_targets.py compiles SPE1CASE2 for both simulators, runs both and prints the largest difference in oil rate, cumulative oil and BHP. The How PetroScript works says how the results are read.