@allanleal, regarding our conversation in the thread about constraining the H2O(g)-H2O(l) phase boundary in P-T space, I tested a few different constraint and initial condition combos:
- constrain amount of aqueous phase , start with an amount of gaseous phase
- constrain amount of aqueous phase, start with an amount of aqueous phase
- constraint amount of gaseous phase, start with an amount of aqueous phase
- constrain amount of gaseous phase, start with an amount of gaseous phase
The first three succeeded, but the last failed with the following error message:
And here's the code that produced it:
import reaktoro as rkt
db = rkt.ThermoFunDatabase("slop98")
# initialize object to hold phases
phases = rkt.Phases(db)
# add species to aq. phase; add aq. phase to system
aqueous_phase = rkt.AqueousPhase(["H2O@"]) # aq. species
phases.add(aqueous_phase)
# add species to gas phase; add gas phase to system
gaseous_phase = rkt.GaseousPhase(["H2O"]) # gases
phases.add(gaseous_phase)
# create system
system = rkt.ChemicalSystem(phases)
# set specifications for equilibrium solver
specs = rkt.EquilibriumSpecs(system)
specs.temperature()
specs.phaseAmount("GaseousPhase")
# set constraints and calculate equilibrium state
solver = rkt.EquilibriumSolver(specs)
conditions = rkt.EquilibriumConditions(specs)
conditions.temperature(150, "C")
conditions.phaseAmount("GaseousPhase", 1e-6)
conditions.setLowerBoundPressure(1.0, "bar")
conditions.setUpperBoundPressure(100.0, "bar")
# initialize state
state = rkt.ChemicalState(system)
state.set("H2O", 1, "kg")
result = solver.solve(state, conditions)
if result.optima.succeeded is True:
print("The equilibrium calculation succeeded!")
else:
print("The equilibrium calculation failed!")
print(state)
Intuitively, the equilibrium state should not depend on the initial condition in this case, but I guess there's an algorithmic reason for why it did not converge?
@allanleal , we have been obtaining saturation indices for various scales. We were unable to obtain the Saturation Indices for Tri-Calcium Phosphate (TCP). Although we used the following atoms for speciation in the aqeous phase, we could not get the SI for TCP.
solution = AqueousPhase(speciate("H O Na Cl Ca C Si Mg P S K Zn")). Please let us know how could we obtain it ?
Hello! You'll need to list the species you want as shown below:
AqueousPhase("H2O H+ OH- Na+ Cl-") # species names need to be exactly as found in the databaseThere is no capability at the moment to remove/exclude species with a certain oxidation state from the definition of a phase.
I'm trying to compile Reaktoro to better understands how it works, (and see if I can do a few pull requests for missing python binfdings (e.g. ElementComposition iterator)), but I run into a strange issue with the tests in ChemicalProps:
> state.setTemperature(100.0)#, 'celsius')
E TypeError: setTemperature(): incompatible function arguments. The following argument types are supported:
E 1. (self: reaktoro.reaktoro4py.ChemicalState, arg0: autodiff::detail::Real<1ul, double>) -> None
E 2. (self: reaktoro.reaktoro4py.ChemicalState, arg0: autodiff::detail::Real<1ul, double>, arg1: str) -> NoneThe problem is in the first argument. Where are the python overload for real defined ?
I've not found a solution for this, as I'm afraid it is something happening in pybind11 (since I've tried many workarounds to eliminate this issue)
Could you please confirm if different compilers (or different versions) were used?
autodiff was installed with conda, and the main gcc is used
and the main gcc is used
Is this gcc from your system, right?
unless gcc is treated differently than the other packages ?
Let me show you how the cmake config messages appear here if you have the conda env activated
-- CCache: Found ccache installed.
-- CCache: Using ccache to potentially speed up the build operation.
-- The CXX compiler identification is GNU 9.4.0
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /home/allan/miniconda3/envs/reaktoro/bin/x86_64-conda-linux-gnu-c++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- CondaAware: Conda environment detected!
-- CondaAware: Found environment variable CONDA_PREFIX=/home/allan/miniconda3/envs/reaktoro
-- CondaAware: Setting PYTHON_EXECUTABLE=/home/allan/miniconda3/envs/reaktoro/bin/python
-- CondaAware: Set CONDA_AWARE_PREFIX=/home/allan/miniconda3/envs/reaktoro
-- CondaAware: Setting CMAKE_INSTALL_PREFIX=CONDA_AWARE_PREFIX=/home/allan/miniconda3/envs/reaktoro
-- CondaAware: Appended /home/allan/miniconda3/envs/reaktoro to CMAKE_PREFIX_PATH
-- CondaAware: Appended /home/allan/miniconda3/envs/reaktoro/include to include directories
-- CondaAware: Appended /home/allan/miniconda3/envs/reaktoro/lib to link directories
You can see above that the compiler from the conda environment was used, instead of system's compiler.
I guess I must have something out of order the first time
Sorry about that, it also solved the bug #243 ... I guess default parameters are not the same
Let me know if you encounter further issues.
for coupling with my own software it would be easier no to go through conda (but I recognize it makes everything much easier in all cases)
phreeqc4rktincludes some changes in the PHREEQC source code. It will probably continue to exist for a while.
state = ChemicalState(system)
I'm getting an error in following line which used to work before:
state.temperature(25.0, "celsius")
TypeError: temperature(): incompatible function arguments. The following argument types are supported:
1. (self: reaktoro.reaktoro4py.ChemicalState, arg0: autodiff::detail::Real<1ul, double>) -> None
2. (self: reaktoro.reaktoro4py.ChemicalState, arg0: autodiff::detail::Real<1ul, double>, arg1: str) -> None
3. (self: reaktoro.reaktoro4py.ChemicalState) -> autodiff::detail::Real<1ul, double>Is there any solution to fix this?
There was a new package produced this morning,
v2.0rc19, wondering if this is the issue
I'll fix the issue with
v2.0rc19