Return electric_field_enabled option to ionflow

[AndreySchadel]

Abstract

Greeting developers! I've been working on version 2.6.0 because option "electric_field_enabled" was available. It was quite important to correct reproduce the ion concentration in flames where the concentration of electrons and cations were the same.

Motivation

However, I tried to update to Ubuntu 24.04 and when i tried to install cantera 2.6.0. I was unable to do it due to the errors produced by Cython in Python 3.12. I looked through the source code and found lines of code written below.

Is it possible to toggle back this feature?

Possible Solutions

`size_t Flow1D::getSolvingStage() const
{
throw NotImplementedError("Flow1D::getSolvingStage",
"Not used by '{}' objects.", type());
}

void Flow1D::setSolvingStage(const size_t stage)
{
throw NotImplementedError("Flow1D::setSolvingStage",
"Not used by '{}' objects.", type());
}

void Flow1D::solveElectricField(size_t j)
{
throw NotImplementedError("Flow1D::solveElectricField",
"Not used by '{}' objects.", type());
}

void Flow1D::fixElectricField(size_t j)
{
throw NotImplementedError("Flow1D::fixElectricField",
"Not used by '{}' objects.", type());
}

bool Flow1D::doElectricField(size_t j) const
{
throw NotImplementedError("Flow1D::doElectricField",
"Not used by '{}' objects.", type());
}`

References

https://github.com/Cantera/cantera/tree/3.1/src/oneD/Flow1D.cpp

[speth]

Hi Andrey,

All the capabilities for solving the electric field are still in place, but have been modified a bit in Cantera 3.0 and later, such that the IonFlow class is used automatically whenever the ionized-gas transport model is selected. This was done in Cantera/cantera#1514. See ion_free_flame.py for an example.

[AndreySchadel]

The above image is done by Cantera 3.1.0:
Result_phi_0.8_ion_cantera_3,1,0.csv

from pathlib import Path
import cantera as ct
from numpy import genfromtxt


p = ct.one_atm
tburner = 368.0
reactants = 'CH4:0.06, O2:0.15, Ar:0.79'  # premixed gas composition
width = 0.01 # m 
loglevel = 1  # amount of diagnostic output (0 to 5)
refine_grid = True  # 'True' to enable refinement

gas = ct.Solution(r'/home/andreyubuntu24/Canteradev/Mechanism/MECHANISM - MarioDeRenzo/reducedS26R134_0.yaml')
gas.TPX = tburner, p, reactants
mdot = 0.19  # kg/m^2/s

f = ct.BurnerFlame(gas, width=width)
f.burner.mdot = mdot
zloc, tvalues = genfromtxt('/home/andreyubuntu24/Canteradev/Repository/CH4_flame/temp.txt', delimiter=',', comments='#').T
zloc /= max(zloc)
f.flame.set_fixed_temp_profile(zloc, tvalues)
f.set_refine_criteria(ratio=2.0, slope=0.2, curve=0.4, prune=0.01)
f.set_initial_guess()
f.set_profile('H', [0.0, 0.05, 1.0], [0.0, 0.001, 0.0])
f.set_profile('OH', [0.0, 0.05, 1.0], [0.0, 0.001, 0.0])
f.set_profile('CH3', [0.0, 0.05, 1.0], [0.0, 0.001, 0.0])
f.set_profile('O', [0.0, 0.05, 1.0], [0.0, 0.0001, 0.0])
f.energy_enabled = False
f.show()

f.solve(loglevel, auto=True)
f.solve(loglevel=loglevel, stage=2)

if "native" in ct.hdf_support():
    output = Path() / "Result_phi_0.8_ion.h5"
else:
    output = Path() / "Result_phi_0.8_ion.yaml"
output.unlink(missing_ok=True)

f.save(output, name="ion", description="solution with ionic transport")

f.save('/home/andreyubuntu24/Canteradev/Repository/CH4_flame/Result_phi_0.8_ion_cantera_3,1,0.csv', basis="mole", overwrite=True)         

The above image is done by Cantera 2.6.0 with option "electric_field_enabled":
Result_phi_0.8_ion_cantera_2,6,0.txt

import cantera as ct
import numpy as np

p = ct.one_atm
tburner = 368.0
reactants = 'CH4:0.06, O2:0.15, Ar:0.79'  # premixed gas composition
width = 0.01 # m 
loglevel = 1  # amount of diagnostic output (0 to 5)
refine_grid = True  # 'True' to enable refinement

gas = ct.Solution(r'/home/andreywslubuntu/Canteradev/Mechanism/MECHANISM - MarioDeRenzo/reducedS26R134_0.yaml')
gas.TPX = tburner, p, reactants
mdot = 0.19  # kg/m^2/s

f = ct.BurnerFlame(gas, width=width)
f.burner.mdot = mdot
zloc, tvalues = np.genfromtxt('/home/andreywslubuntu/Canteradev/Repository/CH4_flame/temp.txt', delimiter=',', comments='#').T
zloc /= max(zloc)
f.flame.set_fixed_temp_profile(zloc, tvalues)
f.set_refine_criteria(ratio=3.0, slope=0.3, curve=1)
f.set_initial_guess()
f.set_profile('H', [0.0, 0.05, 1.0], [0.0, 0.001, 0.0])
f.set_profile('OH', [0.0, 0.05, 1.0], [0.0, 0.001, 0.0])
f.set_profile('CH3', [0.0, 0.05, 1.0], [0.0, 0.001, 0.0])
f.set_profile('O', [0.0, 0.05, 1.0], [0.0, 0.0001, 0.0])
f.show_solution()
f.energy_enabled = False

f.transport_model = 'Mix'
f.solve(loglevel, refine_grid)
f.save('/home/andreywslubuntu/Canteradev/Repository/CH4_flame/Result_phi_0.8_cantera_2,6,0.xml', 'Mix', 'solution with Mix transport')
f.write_csv('/home/andreywslubuntu/Canteradev/Repository/CH4_flame/Result_phi_0.8_mix_cantera_2,6,0.txt', quiet=False)

g = ct.IonBurnerFlame(gas, width=width)
g.burner.mdot = mdot
g.flame.set_fixed_temp_profile(zloc, tvalues)
g.set_initial_guess()
g.restore('/home/andreywslubuntu/Canteradev/Repository/CH4_flame/Result_phi_0.8_cantera_2,6,0.xml', 'Mix')
g.transport_model = 'Ion'
g.show_solution()
g.energy_enabled = False
g.electric_field_enabled = True
g.set_refine_criteria(ratio=2.0, slope=0.2, curve=0.4, prune=0.01)
g.solve(loglevel, refine_grid)
g.save('/home/andreywslubuntu/Canteradev/Repository/CH4_flame/Result_phi_0.8_cantera_2,6,0.xml', 'Ion', 'solution with Ion transport')

g.write_csv('/home/andreywslubuntu/Canteradev/Repository/CH4_flame/Result_phi_0.8_ion_cantera_2,6,0.txt', quiet=False)

I tried to reproduce the same behaviour. However failed to because this option is not presented there. I believe stage 2 doesn't activate as it should be. The number of positive and negative charged particles are the same but doesn't reproduce the real behaviour. Although both algorithms use the same mechanism.

reducedS26R134_0.txt
This is the mechanism for instance. Change the format from TXT to Yaml.