Using Extreme Wind-Speed Probabilistic Forecasts to Optimize Unit Scheduling Decision

With the increasing proportion of renewable energy in power systems, the use of probabilistic forecasts to provide uncertainty information for future decision analyses is inevitable. In addition, the resilience of power systems to extreme events is also a matter of concern. This paper proposes a decision-making model for unit scheduling, using ensemble numerical weather predictions. In this analysis, if wind speeds exceed the cut-out value of a wind turbine, then it is defined as a predicted event of concern. The economic value to decision-makers of using probabilistic wind-speed forecasts will be analyzed. Based on the analysis results, the optimal probability threshold is determined and is used to decide whether, in the unit scheduling, wind turbines should be shut down in advance. This study uses four typhoon events as test cases to verify the forecast performance of the proposed numerical weather prediction model and determine different optimal probability thresholds for shutting down wind turbines (decision probability thresholds) in different months and seasons, because if the same decision probability threshold is used as throughout the year, it not only fails to achieve maximum economic value but also endangers the safety of power system operation.