torque2026-128

This code formulates numerical experiments to mitigate forecast errors for power ramp limitation in wind farms and documents the results presented in the article "Forecast error mitigation for the ramp-constrained operation of wind farms" presented at the TORQUE 2026 conference.

The numerical experiments are used to assess different method for reducing the impact of forecast error for a wind farm operating with limits on the magnitude of power ramps. The first method consist of inflating the ramp events in the power forecast, resulting in a pessimistic forecast, to encourage a conservative behaviour of the power plant. The second method consist of using a storage system to charge or discharge power during ramp events. The wind farm operation is modelled using an online dispatch optimization problem.

The code is made of four core scripts:

compute_power_forecast_ramp.py: a script to generate a power forecast data set with or without the pessimistic methodology.analysis_ramp_case.py: a script to run an online dispatch optimization simulating the operation of a wind farm with or without a storage system, and with or without the pessimistic methodology.example_plot.py: a script showing an example of how to post-process the results of the numerical experiments.aux_forecast.py: this script provides auxiliary function for the power forecast generation.

Finally, the code includes two files (snakefile and config.yaml) to execute the scripts using the workflow manager Snakemake. Information about dependencies and how to run the code are provided in the README.