"""The Kardar–Parisi–Zhang (KPZ) equation describing the evolution of an interface.
.. codeauthor:: David Zwicker <david.zwicker@ds.mpg.de>
"""
from __future__ import annotations
from typing import Callable
import numba as nb
import numpy as np
from ..fields import ScalarField
from ..grids.boundaries import set_default_bc
from ..grids.boundaries.axes import BoundariesData
from ..tools.docstrings import fill_in_docstring
from ..tools.numba import jit
from .base import PDEBase, expr_prod
[docs]
class KuramotoSivashinskyPDE(PDEBase):
r"""The Kuramoto-Sivashinsky equation.
The mathematical definition is
.. math::
\partial_t u = -\nu \nabla^4 u - \nabla^2 u -
\frac{1}{2} \left(\nabla h\right)^2 + \eta(\boldsymbol r, t)
where :math:`u` is the height of the interface in Monge parameterization. The
dynamics are governed by the parameters :math:`\nu` , while :math:`\eta` is Gaussian
white noise, whose strength is controlled by the `noise` argument.
"""
explicit_time_dependence = False
default_bc = "auto_periodic_neumann"
"""Default boundary condition used when no specific conditions are chosen."""
@fill_in_docstring
def __init__(
self,
nu: float = 1,
*,
bc: BoundariesData | None = None,
bc_lap: BoundariesData | None = None,
noise: float = 0,
rng: np.random.Generator | None = None,
):
r"""
Args:
nu (float):
Parameter :math:`\nu` for the strength of the fourth-order term
bc:
The boundary conditions applied to the field.
{ARG_BOUNDARIES}
bc_lap:
The boundary conditions applied to the second derivative of the scalar
field :math:`c`. If `None`, the same boundary condition as `bc` is
chosen. Otherwise, this supports the same options as `bc`.
noise (float):
Variance of the (additive) noise term
rng (:class:`~numpy.random.Generator`):
Random number generator (default: :func:`~numpy.random.default_rng()`)
used for stochastic simulations. Note that this random number generator
is only used for numpy function, while compiled numba code uses the
random number generator of numba. Moreover, in simulations using
multiprocessing, setting the same generator in all processes might yield
unintended correlations in the simulation results.
"""
super().__init__(noise=noise, rng=rng)
self.nu = nu
self.bc = set_default_bc(bc, self.default_bc)
self.bc_lap = self.bc if bc_lap is None else bc_lap
@property
def expression(self) -> str:
"""str: the expression of the right hand side of this PDE"""
expr = f"c + {expr_prod(self.nu, '∇²c')}"
return f"-∇²({expr}) - 0.5 * |∇c|²"
[docs]
def evolution_rate( # type: ignore
self,
state: ScalarField,
t: float = 0,
) -> ScalarField:
"""Evaluate the right hand side of the PDE.
Args:
state (:class:`~pde.fields.ScalarField`):
The scalar field describing the concentration distribution
t (float):
The current time point
Returns:
:class:`~pde.fields.ScalarField`:
Scalar field describing the evolution rate of the PDE
"""
if not isinstance(state, ScalarField):
raise ValueError("`state` must be ScalarField")
state_lap = state.laplace(bc=self.bc, args={"t": t})
result = (
-self.nu * state_lap.laplace(bc=self.bc_lap, args={"t": t})
- state_lap
- 0.5 * state.gradient_squared(bc=self.bc, args={"t": t})
)
result.label = "evolution rate"
return result # type: ignore
def _make_pde_rhs_numba( # type: ignore
self, state: ScalarField
) -> Callable[[np.ndarray, float], np.ndarray]:
"""Create a compiled function evaluating the right hand side of the PDE.
Args:
state (:class:`~pde.fields.ScalarField`):
An example for the state defining the grid and data types
Returns:
A function with signature `(state_data, t)`, which can be called
with an instance of :class:`~numpy.ndarray` of the state data and
the time to obtained an instance of :class:`~numpy.ndarray` giving
the evolution rate.
"""
arr_type = nb.typeof(state.data)
signature = arr_type(arr_type, nb.double)
nu_value = self.nu
laplace = state.grid.make_operator("laplace", bc=self.bc)
laplace2 = state.grid.make_operator("laplace", bc=self.bc_lap)
gradient_sq = state.grid.make_operator("gradient_squared", bc=self.bc)
@jit(signature)
def pde_rhs(state_data: np.ndarray, t: float):
"""Compiled helper function evaluating right hand side."""
result = -laplace(state_data, args={"t": t})
result += nu_value * laplace2(result, args={"t": t})
result -= 0.5 * gradient_sq(state_data, args={"t": t})
return result
return pde_rhs # type: ignore