Skip to main content

​
Overview

The Remedial Action Optimization (RAO) module provides advanced optimization capabilities for power system security management. It optimizes the use of remedial actions to ensure network security while minimizing costs.

​
Main Classes

​
Rao

Main class for running remedial action optimization. 
from pypowsybl.rao import create_rao

rao = create_rao()

​
Methods

​
run
method
Run a remedial action optimization.Parameters:
  • crac (Crac): CRAC (Contingency list, Remedial Actions and additional Constraints)
  • network (Network): Network to optimize
  • parameters (Parameters, optional): RAO parameters
  • rao_provider (str): RAO provider name (default: “SearchTreeRao”)
  • loop_flow_glsk (Glsk, optional): Loop flow GLSK for loop flow computation
Returns: RaoResultExample:
result = rao.run(
    crac=crac,
    network=network,
    parameters=params,
    loop_flow_glsk=glsk
)
​
run_voltage_monitoring
method
Run voltage monitoring after RAO.Parameters:
  • crac (Crac): CRAC object
  • network (Network): Network
  • rao_result (RaoResult): Result from previous RAO run
  • load_flow_parameters (loadflow.Parameters, optional): Load flow parameters
  • provider_str (str): Provider string
  • monitoring_glsk (Glsk, optional): GLSK for monitoring
Returns: RaoResult
​
run_angle_monitoring
method
Run angle monitoring after RAO.Parameters:
  • crac (Crac): CRAC object
  • network (Network): Network
  • rao_result (RaoResult): Result from previous RAO run
  • load_flow_parameters (loadflow.Parameters, optional): Load flow parameters
  • provider_str (str): Provider string
  • monitoring_glsk (Glsk, optional): GLSK for monitoring
Returns: RaoResult

​
Crac

Represents the CRAC (Contingency list, Remedial Actions and additional Constraints) data. 
from pypowsybl.rao import Crac

# Load from file
crac = Crac.from_file_source(network, 'crac.json')

# Load from buffer
import io
with open('crac.json', 'rb') as f:
    crac = Crac.from_buffer_source(network, io.BytesIO(f.read()))
​
from_file_source
classmethod
Load CRAC from a file.Parameters:
  • network (Network): Network object
  • crac_file (str or PathLike): Path to CRAC file
Returns: Crac
​
from_buffer_source
classmethod
Load CRAC from a buffer.Parameters:
  • network (Network): Network object
  • crac_source (BytesIO): CRAC data buffer
Returns: Crac

​
Glsk

Represents Generation and Load Shift Keys for sensitivity analysis. 
from pypowsybl.rao import Glsk
Used in RAO for loop flow computation and monitoring.

​
Parameters

Comprehensive parameters for RAO execution. 
from pypowsybl.rao import (
    Parameters,
    ObjectiveFunctionParameters,
    RangeActionOptimizationParameters,
    TopoOptimizationParameters
)

params = Parameters(
    objective_function_parameters=ObjectiveFunctionParameters(),
    range_action_optimization_parameters=RangeActionOptimizationParameters(),
    topo_optimization_parameters=TopoOptimizationParameters()
)
​
objective_function_parameters
ObjectiveFunctionParameters
Parameters for the objective function.
​
range_action_optimization_parameters
RangeActionOptimizationParameters
Parameters for range action optimization (PST, HVDC, injections).
​
topo_optimization_parameters
TopoOptimizationParameters
Parameters for topological action optimization.
​
multithreading_parameters
MultithreadingParameters
Parameters for parallel execution.
​
second_preventive_rao_parameters
SecondPreventiveRaoParameters
Parameters for second preventive RAO execution.
​
not_optimized_cnecs_parameters
NotOptimizedCnecsParameters
Parameters for CNECs that should not be optimized.
​
loadflow_and_sensitivity_parameters
LoadFlowAndSensitivityParameters
Load flow and sensitivity computation parameters.
​
provider_parameters
Dict[str, str]
Provider-specific parameters.

​
Loading and Saving Parameters

​
from_file_source
classmethod
Load parameters from a file.Parameters:
  • parameters_file (str or PathLike): Path to parameters file
Returns: ParametersExample:
params = Parameters.from_file_source('rao_params.json')
​
from_buffer_source
classmethod
Load parameters from a buffer.Parameters:
  • parameters_source (BytesIO): Parameters data buffer
Returns: Parameters
​
serialize
method
Save parameters to a file.Parameters:
  • output_file (str): Output file path
Example:
params.serialize('output_params.json')
​
serialize_to_binary_buffer
method
Serialize parameters to a binary buffer.Returns: BytesIO
​
to_json
method
Convert parameters to JSON dictionary.Returns: Dict[str, Any]

​
ObjectiveFunctionParameters

Parameters for the RAO objective function. Attributes:
  • objective_function_type: Type of objective function
  • unit: Unit for optimization (MW, AMPERE)
  • curative_min_obj_improvement: Minimum improvement required for curative actions
  • enforce_curative_security: Whether to enforce security in curative perimeter

​
RangeActionOptimizationParameters

Parameters for optimizing range actions (PST, HVDC, injections). Key Attributes:
  • max_mip_iterations: Maximum iterations for MIP solver
  • pst_ra_min_impact_threshold: Minimum impact threshold for PST actions
  • pst_sensitivity_threshold: Sensitivity threshold for PST
  • pst_model: PST model type
  • hvdc_ra_min_impact_threshold: Minimum impact threshold for HVDC actions
  • injection_ra_min_impact_threshold: Minimum impact threshold for injection actions
  • ra_range_shrinking: Range shrinking strategy
  • solver: Solver to use (e.g., “CBC”, “SCIP”)
  • relative_mip_gap: Relative MIP gap tolerance
  • solver_specific_parameters: Solver-specific parameters

​
TopoOptimizationParameters

Parameters for topological action optimization. Key Attributes:
  • max_preventive_search_tree_depth: Maximum depth for preventive search tree
  • max_curative_search_tree_depth: Maximum depth for curative search tree
  • predefined_combinations: Use predefined action combinations
  • relative_min_impact_threshold: Relative minimum impact threshold
  • absolute_min_impact_threshold: Absolute minimum impact threshold
  • skip_actions_far_from_most_limiting_element: Skip actions far from constraints
  • max_number_of_boundaries_for_skipping_actions: Max boundaries for action skipping

​
MultithreadingParameters

Parameters for parallel execution. Attributes:
  • available_cpus: Number of CPUs available for parallel computation

​
RaoResult

Result object returned by RAO execution. 
# RaoResult is returned by rao.run()
result = rao.run(crac=crac, network=network, parameters=params)

​
Utility Functions

​
create_rao
function
Create a RAO object for running optimization.Returns: RaoExample:
from pypowsybl.rao import create_rao

rao = create_rao()

​
RaoLogFilter

Filter for RAO log messages. 
from pypowsybl.rao import RaoLogFilter
import logging

# Create filter to show only OpenRAO logs
filter = RaoLogFilter()
​
filter
method
Filter log records to show only OpenRAO package logs.Parameters:
  • record (LogRecord): Log record to filter
Returns: bool - True if record should be displayed

​
Complete Example

import pypowsybl as pp
from pypowsybl.rao import (
    create_rao,
    Crac,
    Parameters,
    RangeActionOptimizationParameters
)

# Load network
network = pp.network.load('network.xiidm')

# Load CRAC
crac = Crac.from_file_source(network, 'crac.json')

# Configure parameters
range_params = RangeActionOptimizationParameters()
range_params.max_mip_iterations = 10
range_params.solver = "CBC"

params = Parameters(
    range_action_optimization_parameters=range_params
)

# Create and run RAO
rao = create_rao()
result = rao.run(
    crac=crac,
    network=network,
    parameters=params,
    rao_provider="SearchTreeRao"
)

# Process results
print(f"RAO optimization completed")
print(f"Result: {result}")

​
Parameter File Example

{
  "objective-function": {
    "type": "MAX_MIN_MARGIN_IN_MEGAWATT",
    "curative-min-obj-improvement": 0.0,
    "enforce-curative-security": true
  },
  "range-actions-optimization": {
    "max-mip-iterations": 10,
    "pst-sensitivity-threshold": 0.01,
    "solver": "CBC",
    "relative-mip-gap": 0.001
  },
  "topological-actions-optimization": {
    "max-preventive-search-tree-depth": 5,
    "max-curative-search-tree-depth": 3,
    "relative-minimum-impact-threshold": 0.01
  }
}

​
Notes

  • RAO requires a valid CRAC file defining contingencies, CNECs, and remedial actions
  • The SearchTreeRao provider is the default and most commonly used
  • Parameters can be loaded from JSON files or configured programmatically
  • GLSK files are optional but required for loop flow computation

Build docs developers (and LLMs) love

Get started for freeTalk to us