eval_model

Evaluate chemical kinetic models against experimental ignition-delay data.

members:

pyteck.eval_model.calculate_error_function(ignition_delays_exp, ignition_delays_sim, standard_dev)

Calculate the error and deviation functions for a dataset.

Cases that did not ignite—indicated by a simulated ignition delay of zero or a non-finite value—are excluded from the averages, so that a single non-ignition does not drive the whole dataset’s error to infinity (see issues #1 and #18).

Parameters:

• ignition_delays_exp (numpy.ndarray) – Experimental ignition delays

• ignition_delays_sim (numpy.ndarray) – Simulated ignition delays (zero or non-finite where no ignition occurred)

• standard_dev (float) – Standard deviation of the experimental data

Returns:

• error_func (float) – Mean squared logarithmic error over the igniting cases (nan if none of the cases ignited)

• dev_func (float) – Mean logarithmic deviation over the igniting cases (nan if none of the cases ignited)

pyteck.eval_model.create_simulations(dataset, properties)

Set up individual simulations for each ignition delay value.

Parameters:

• dataset (str) – Name of dataset file

• properties (pyked.chemked.ChemKED) – ChemKED object with full set of experimental properties

Returns:

simulations – List of simulation cases (BaseSimulation subclass instances), one per datapoint

Return type:

list of BaseSimulation

pyteck.eval_model.estimate_std_dev(indep_variable, dep_variable)

Estimate standard deviation of experimental data via a spline fit.

Parameters:

• indep_variable (numpy.ndarray or list of float) – Independent variable (e.g., temperature, pressure)

• dep_variable (numpy.ndarray or list of float) – Dependent variable (e.g., ignition delay)

Returns:

standard_dev – Standard deviation of difference between data and best-fit line

Return type:

float

pyteck.eval_model.evaluate_model(model_name, spec_keys_file, dataset_file, data_path='data', model_path='models', results_path='results', model_variant_file=None, num_threads=None, print_results=False, restart=False, skip_validation=False)

Evaluate the ignition delay error of a model for a given dataset.

Parameters:

• model_name (str) – Chemical kinetic model filename

• spec_keys_file (str) – Name of YAML file identifying important species

• dataset_file (str) – Name of file with list of data files

• data_path (str, optional) – Local path for data files (default: "data")

• model_path (str, optional) – Local path for the model file (default: "models")

• results_path (str, optional) – Local path for creating results files (default: "results")

• model_variant_file (str, optional) – Name of YAML file identifying ranges of conditions for variants of the kinetic model (default: None)

• num_threads (int, optional) – Number of CPU threads to use for running simulations in parallel. The default (None) uses the number of available cores minus one.

• print_results (bool, optional) – If True, print results of the model evaluation to screen (default: False).

• restart (bool, optional) – If True, reuse existing results files and only compute new cases (default: False).

• skip_validation (bool, optional) – If True, skip validation of ChemKED files (default: False).

Returns:

output – Dictionary with all information about model evaluation results

Return type:

dict

pyteck.eval_model.get_changing_variable(cases)

Identify variable changing across multiple cases.

Parameters:

cases (list of pyked.chemked.DataPoint) – List of DataPoint with experimental case data

Returns:

variable – Values of the changing experimental variable

Return type:

list of float

pyteck.eval_model.min_deviation = 0.1

minimum allowable standard deviation for experimental data

Type:

float

pyteck.eval_model.read_dataset_list(dataset_file)

Read the list of dataset files, skipping blank or whitespace-only lines.

Parameters:

dataset_file (str or pathlib.Path) – Name of file listing dataset files, one per line

Returns:

Names of the dataset files, stripped of surrounding whitespace and with blank lines removed

Return type:

list of str

pyteck.eval_model.select_variant_suffix(variant, properties)

Build the model-file suffix for a model variant from a case’s properties.

Some models ship as several files that differ by bath gas and/or nominal pressure. The model_variant mapping records, for each such model, the filename suffix to use for each bath gas and pressure. This selects the suffix appropriate for a given experimental case.

Parameters:

• variant (dict) – Model-variant entry, optionally with "bath gases" and/or "pressures" maps from a bath-gas name / pressure to a filename suffix

• properties (pyked.chemked.DataPoint) – Experimental case properties (composition is a dict keyed by species name, and pressure is a pint quantity)

Returns:

Suffix to append to the model filename (empty if no variant applies)

Return type:

str

pyteck.eval_model.simulation_worker(sim_tuple)

Worker for multiprocessing of simulation cases.

Parameters:

sim_tuple (tuple) – Contains a BaseSimulation instance and the parameters needed to set up and run the case: (sim, model_file, model_spec_key, path, restart)

Returns:

sim – Simulation case with results ready for process_results

Return type:

BaseSimulation