Reliability Enhancement via Integration of Extreme Weather Forecast in Power System Operation

According to US Department of Energy, extreme weather is the leading cause of power outages in the United States. Although all the system operators have access to weather forecast data, no existing automated tool can take advantage of this data to guide preventive operation. This paper presents a framework to close this gap. The framework first forecasts weather with accuracy and resolution required for grid operation. The weather forecast data is, then, passed to a structural analysis module, where the failure likelihood of power system elements is estimated. Finally, the failure probabilities are integrated within a day-ahead stochastic unit commitment model to guide preventive operation. A scenario reduction technique is developed and employed alongside an enhanced formulation to achieve computational tractability. The simulation results on a 2000-bus Texas system suggest that the developed model is effective in reducing power outages, with minor additional dispatch cost. The results also suggest that the developed framework is tractable for large-scale systems.