|
rizer.cantera_ext
C++ Cantera 1-D plasma extension (custom Domain1D models and solvers)
|
The parabolic half of PlasmaChannel1D::eval(): conduction in both pools, Fickian species diffusion, and the two terms that diffusion drives in the energy equations. More...
#include "numeric/PropertyTable.h"#include "cantera/base/ctexceptions.h"#include <cstddef>#include <vector>Go to the source code of this file.
Classes | |
| struct | rizer::RadialDiffusionCoefficients |
| Everything the diffusion block needs that the grid and the state do not carry: the transport coefficients, the wall ghost, and the per-species properties the two energy terms close with. More... | |
| struct | rizer::WallEnergyOutflow |
| Energy leaving through the wall ghost face per unit time, per unit axial length and per radian [W/m] – the convention of V_cell and of HydroStepResult::boundary_total_energy_out. More... | |
Namespaces | |
| namespace | rizer |
Functions | |
| void | rizer::addRadialDiffusion (const PlasmaChannel1D &domain, Cantera::span< const double > x, const RadialDiffusionCoefficients &c, std::size_t j, double lhsTg, double lhsTe, double &dTg, double &dTe, double *dYdt, double *scratch) |
| Add one node's diffusive contributions to the rates the caller holds. | |
| WallEnergyOutflow | rizer::wallEnergyOutflow (const PlasmaChannel1D &domain, Cantera::span< const double > x, const RadialDiffusionCoefficients &c, const double *species_enthalpy) |
| What conduction and species diffusion move through the wall face at the state x. | |
The parabolic half of PlasmaChannel1D::eval(): conduction in both pools, Fickian species diffusion, and the two terms that diffusion drives in the energy equations.
Assembly, not arithmetic. numeric/DiffusiveFlux.h holds the per-face flux and the cell measure; this file turns them into the divergences one node contributes, with the wall ghost and the axis symmetry resolved. The hyperbolic side is split the same way, numeric/EulerFlux.h under HydroStep.
Not a sub-step. HydroStep advances a state over an interval; this runs inside CVODE's residual, once per node of the [jmin, jmax] Jacobian window, so it allocates nothing and owns no state. The caller supplies the scratch it needs.
It adds into the caller's rates rather than returning its own. The contributions are divided by the node's capacitances and accumulated in the order the energy equations are written, and that order is load-bearing: floating-point addition is not associative, so summing the terms separately and adding once would move the last bits of every P4 run.
Definition in file RadialDiffusion.h.