Cost Component

CostModelComponent

class topfarm.cost_models.cost_model_wrappers.CostModelComponent(input_keys, n_wt, cost_function, cost_gradient_function=None, output_keys=['Cost'], output_unit='', additional_input=[], additional_output=[], max_eval=None, objective=True, maximize=False, output_vals=[0.0], input_units=[], step={}, use_constraint_violation=True, **kwargs)

Wrapper for pure-Python cost functions

__init__(input_keys, n_wt, cost_function[, ...])	Initialize wrapper for pure-Python cost function
compute(inputs, outputs)	Compute cost model
compute_partials(inputs, J)	Compute sub-jacobian parts.

__init__(input_keys, n_wt, cost_function, cost_gradient_function=None, output_keys=['Cost'], output_unit='', additional_input=[], additional_output=[], max_eval=None, objective=True, maximize=False, output_vals=[0.0], input_units=[], step={}, use_constraint_violation=True, **kwargs)

Initialize wrapper for pure-Python cost function

Parameters

• input_keys (list of str) – Inputs to the cost function

• n_wt (int) – Number of wind turbines

• cost_function (function handle) – Function to evaluate cost

• cost_gradient_function (function handle, optional) – Function to evaluate gradient of the cost function

• output_key (list) – Name of output key

• output_unit (str) – Units of output of cost function

• additional_input (list of str) –

  Other (non-design-variable) inputs required by the cost function

  Gradients will not be computed for these inputs

• additional_output (list of str or list of tuples) –

  Other outputs to request

  if list of str: [‘add_out_name’,…]

  if list of tuples [(‘add_out_name’, val),…], where val is a template value of the output The cost function must return: cost, {‘add_out1_name’: add_out1, …}

• max_eval (int) – Maximum number of function evaluations

• objective (boolean) – Must be True for standalone CostModelComponents and the final component in a TopFarmGroup

• maximize (boolean) –

  If True: objective is maximised during optimization

  If False: Objective is minimized during optimization

• output_vals (float or array_like) – Format of output

• input_units (list of str) – Units of the respective input_keys

• step (dict of {str : float}) – Finite difference step size for each input key, e.g. {input_key : step_size, input_key2 : step_size2}

AEPCostModelComponent

class topfarm.cost_models.cost_model_wrappers.AEPCostModelComponent(input_keys, n_wt, cost_function, cost_gradient_function=None, output_unit='', additional_input=[], additional_output=[], max_eval=None, output_key='AEP', **kwargs)

Wrapper for pure-Python cost functions

__init__(input_keys, n_wt, cost_function[, ...])

Initialize Sub class of the CostModelComponent for AEP maximization

__init__(input_keys, n_wt, cost_function, cost_gradient_function=None, output_unit='', additional_input=[], additional_output=[], max_eval=None, output_key='AEP', **kwargs)

Initialize Sub class of the CostModelComponent for AEP maximization

Parameters

• input_keys (list of str) – Inputs to the cost function

• n_wt (int) – Number of wind turbines

• cost_function (function handle) – Function to evaluate cost

• cost_gradient_function (function handle, optional) – Function to evaluate gradient of the cost function

• output_key (list) – Name of output key, default is AEP

• output_unit (str) – Units of output of cost function

• additional_input (list of str) –

  Other (non-design-variable) inputs required by the cost function

  Gradients will not be computed for these inputs

• additional_output (list of str or list of tuples) –

  Other outputs to request

  if list of str: [‘add_out_name’,…]

  if list of tuples [(‘add_out_name’, val),…], where val is a template value of the output The cost function must return: cost, {‘add_out1_name’: add_out1, …}

• max_eval (int) – Maximum number of function evaluations

AEPMaxLoadCostModelComponent

class topfarm.cost_models.cost_model_wrappers.AEPMaxLoadCostModelComponent(input_keys, n_wt, aep_load_function, max_loads, aep_load_gradient=None, output_keys=['AEP', 'loads'], step={}, maximize=True, **kwargs)

Wrapper for pure-Python cost functions

__init__(input_keys, n_wt, ...[, ...])

Initialize Sub class of the CostModelComponent for AEP maximization and load constraints

__init__(input_keys, n_wt, aep_load_function, max_loads, aep_load_gradient=None, output_keys=['AEP', 'loads'], step={}, maximize=True, **kwargs)

Initialize Sub class of the CostModelComponent for AEP maximization and load constraints

Parameters

• input_keys (list of str) – Inputs to the cost function

• n_wt (int) – Number of wind turbines

• aep_load_function (function handle) – Function to evaluate cost

• aep_load_gradient (function handle, optional) – Function to evaluate gradient of the cost function

• output_keys (list) – Name of output key, default is AEP and loads

• step (dict of {str : float}) – Finite difference step size for each input key, e.g. {input_key : step_size, input_key2 : step_size2}

• maximize (boolean) –

  If True: objective is maximised during optimization

  If False: Objective is minimized during optimization

PyWakeAEPCostModelComponent

class topfarm.cost_models.py_wake_wrapper.PyWakeAEPCostModelComponent(windFarmModel, n_wt, wd=None, ws=None, max_eval=None, grad_method=<function autograd>, n_cpu=1, **kwargs)

TOPFARM wrapper for PyWake AEP calculator

__init__(windFarmModel, n_wt[, wd, ws, ...])	Initialize wrapper for PyWake AEP calculator
get_aep4smart_start([ws, wd, type])	Compute AEP with a smart start approach

__init__(windFarmModel, n_wt, wd=None, ws=None, max_eval=None, grad_method=<function autograd>, n_cpu=1, **kwargs)

Initialize wrapper for PyWake AEP calculator

Parameters

• windFarmModel (DeficitModel) – Wake deficit model used

• n_wt (int) – Number of wind turbines

• wd (array_like) – Wind directions to study

• ws (array_like) – Wind speeds to study

• max_eval (int) – Maximum number of function evaluations

• grad_method (function handle) – Selected method to calculate gradients, default is autograd