import warnings
from typing import Optional
import numpy as np
from pybamm import IDAKLUJax, SolverError
from pybop import BaseModel, BaseProblem, Dataset
from pybop.parameters.parameter import Inputs, Parameters
[docs]
class FittingProblem(BaseProblem):
"""
Problem class for fitting (parameter estimation) problems.
Extends `BaseProblem` with specifics for fitting a model to a dataset.
Parameters
----------
model : object
The model to fit.
parameters : pybop.Parameter or pybop.Parameters
An object or list of the parameters for the problem.
dataset : Dataset
Dataset object containing the data to fit the model to.
check_model : bool, optional
Flag to indicate if the model should be checked (default: True).
signal : list[str], optional
A list of variables to fit (default: ["Voltage [V]"]).
domain : str, optional
The name of the domain (default: "Time [s]").
additional_variables : list[str], optional
Additional variables to observe and store in the solution (default additions are: ["Time [s]"]).
initial_state : dict, optional
A valid initial state, e.g. the initial open-circuit voltage (default: None).
Additional Attributes
---------------------
dataset : dictionary
The dictionary from a Dataset object containing the signal keys and values to fit the model to.
domain_data : np.ndarray
The domain points in the dataset.
n_domain_data : int
The number of domain points.
target : np.ndarray
The target values of the signals.
"""
def __init__(
self,
model: BaseModel,
parameters: Parameters,
dataset: Dataset,
check_model: bool = True,
signal: Optional[list[str]] = None,
domain: Optional[str] = None,
additional_variables: Optional[list[str]] = None,
initial_state: Optional[dict] = None,
):
super().__init__(
parameters=parameters,
model=model,
check_model=check_model,
signal=signal,
domain=domain,
additional_variables=additional_variables,
initial_state=initial_state,
)
[docs]
self._dataset = dataset.data
[docs]
self._n_parameters = len(self.parameters)
# Check that the dataset contains necessary variables
dataset.check(domain=self.domain, signal=self.signal)
# Unpack domain and target data
[docs]
self._domain_data = self._dataset[self.domain]
[docs]
self.n_data = len(self._domain_data)
self.set_target(dataset)
if self._model is not None:
# Build the model from scratch
if self._model.built_model is not None:
self._model.clear()
self._model.build(
dataset=self._dataset,
parameters=self.parameters,
check_model=self.check_model,
initial_state=self.initial_state,
)
[docs]
def set_initial_state(self, initial_state: Optional[dict] = None):
"""
Set the initial state to be applied to evaluations of the problem.
Parameters
----------
initial_state : dict, optional
A valid initial state (default: None).
"""
if initial_state is not None and "Initial SoC" in initial_state.keys():
warnings.warn(
"It is usually better to define an initial open-circuit voltage as the "
"initial_state for a FittingProblem because this value can typically be "
"obtained from the data, unlike the intrinsic initial state of charge. "
"In the case where the fitting parameters do not change the OCV-SOC "
"relationship, the initial state of charge may be passed to the model "
'using, for example, `model.set_initial_state({"Initial SoC": 1.0})` '
"before constructing the FittingProblem.",
UserWarning,
stacklevel=1,
)
self.initial_state = initial_state
[docs]
def evaluate(self, inputs: Inputs) -> dict[str, np.ndarray[np.float64]]:
"""
Evaluate the model with the given parameters and return the signal.
Parameters
----------
inputs : Inputs
Parameters for evaluation of the model.
Returns
-------
y : np.ndarray
The simulated model output y(t) for self.eis == False, and y(ω) for self.eis == True for the given inputs.
"""
inputs = self.parameters.verify(inputs)
if self.eis:
return self._evaluate(self._model.simulateEIS, inputs)
return self._evaluate(self._model.simulate, inputs)
[docs]
def _evaluate(
self, func, inputs, calculate_grad=False
) -> dict[str, np.ndarray[np.float64]]:
"""
Perform simulation using the specified method and handle exceptions.
Parameters
----------
func : callable
The method to be used for simulation.
inputs : Inputs
Parameters for evaluation of the model.
Returns
-------
dict[str, np.ndarray[np.float64]]
The simulated model output.
"""
try:
if isinstance(self.model.solver, IDAKLUJax):
sol = self._model.solver.get_vars(self.signal)(
self.domain_data, inputs
) # TODO: Add initial_state capabilities
else:
sol = func(
inputs,
self._domain_data,
initial_state=self.initial_state,
)
except (SolverError, ZeroDivisionError, RuntimeError, ValueError) as e:
if isinstance(e, ValueError) and str(e) not in self.exception:
raise # Raise the error if it doesn't match the expected list
error_out = {s: self.failure_output for s in self.signal}
return (error_out, self.failure_output) if calculate_grad else error_out
if self.eis:
return sol
if isinstance(self.model.solver, IDAKLUJax):
return {signal: sol[:, i] for i, signal in enumerate(self.signal)}
if calculate_grad:
signals = self.signal + self.additional_variables
return (
{s: sol[s].data for s in signals},
{s: sol[s].sensitivities for s in signals},
)
return {s: sol[s].data for s in (self.signal + self.additional_variables)}
[docs]
def evaluateS1(self, inputs: Inputs):
"""
Evaluate the model with the given parameters and return the signal and its derivatives.
Parameters
----------
inputs : Inputs
Parameters for evaluation of the model.
Returns
-------
tuple[dict, np.ndarray]
A tuple containing the simulation result y(t) as a dictionary and the sensitivities
dy/dx(t) evaluated with given inputs.
"""
inputs = self.parameters.verify(inputs)
self.parameters.update(values=list(inputs.values()))
y, sens = self._evaluate(self._model.simulateS1, inputs, calculate_grad=True)
if not any([np.isfinite(y[s]).any() for s in self.signal]):
return y, sens
# Extract the sensitivities for all signals at once
param_keys = self.parameters.keys()
dy = np.stack(
[
np.column_stack(
[sens[signal][key].toarray()[:, 0] for key in param_keys]
)
for signal in self.signal
],
axis=1,
)
return y, dy