System
Once the domain and model are defined, the next step is to define the system. This is done by creating a System object.
System
A class representing a system of equations.
| Parameters: |
|
|---|
Source code in splitfxm/system.py
class System:
"""
A class representing a system of equations.
Parameters
----------
model : Model
The model for which to solve the system of equations.
scheme : Schemes
The discretization scheme to use for the system
scheme_opts : dict
A dictionary of options for the scheme.
"""
def __init__(self, model, scheme, scheme_opts):
"""
Initialize a System object.
"""
self._model = model
self._scheme = scheme
self._scheme_opts = scheme_opts
def residuals(self, d: Domain):
"""
Calculate the residuals for the system of equations.
Parameters
----------
d : Domain
The domain for which to calculate the residuals.
Returns
-------
rhs_list : list of ndarray
The list of residual arrays for each cell in the domain.
"""
cells = d.cells()
# Interior indices
ilo = d.ilo()
ihi = d.ihi()
# Get list of boundary points in each direction
nb_left = d.nb(btype.LEFT)
nb_right = d.nb(btype.RIGHT)
# Fetch equation for model
eq = self._model.equation()
rhs_list = []
for i in range(ilo, ihi + 1):
# Define the neighborhood and band around the current cell
cell_sub = [cells[i + offset]
for offset in range(-nb_left, nb_right + 1)]
# Send two-sided stencil
# Let model decide computation
rhs = eq.residuals(cell_sub, self._scheme, self._scheme_opts)
rhs_list.append(rhs)
# Stack the residuals to form a 2D array
rhs_array = np.vstack(rhs_list)
return rhs_array
__init__(model, scheme, scheme_opts)
Initialize a System object.
Source code in splitfxm/system.py
def __init__(self, model, scheme, scheme_opts):
"""
Initialize a System object.
"""
self._model = model
self._scheme = scheme
self._scheme_opts = scheme_opts
residuals(d)
Calculate the residuals for the system of equations.
| Parameters: |
|
|---|
| Returns: |
|
|---|
Source code in splitfxm/system.py
def residuals(self, d: Domain):
"""
Calculate the residuals for the system of equations.
Parameters
----------
d : Domain
The domain for which to calculate the residuals.
Returns
-------
rhs_list : list of ndarray
The list of residual arrays for each cell in the domain.
"""
cells = d.cells()
# Interior indices
ilo = d.ilo()
ihi = d.ihi()
# Get list of boundary points in each direction
nb_left = d.nb(btype.LEFT)
nb_right = d.nb(btype.RIGHT)
# Fetch equation for model
eq = self._model.equation()
rhs_list = []
for i in range(ilo, ihi + 1):
# Define the neighborhood and band around the current cell
cell_sub = [cells[i + offset]
for offset in range(-nb_left, nb_right + 1)]
# Send two-sided stencil
# Let model decide computation
rhs = eq.residuals(cell_sub, self._scheme, self._scheme_opts)
rhs_list.append(rhs)
# Stack the residuals to form a 2D array
rhs_array = np.vstack(rhs_list)
return rhs_array