Operational Implementation and Evaluation of a Solar and Wind Integrated Forecast Tool (SWIFT) in the Hawaiian Islands

The Hawaiian Islands are experiencing considerable growth in wind and solar power production. Furthermore, the temporal and spatial variability of the wind and solar resource can be considerable owing to the islands' terrain, marine- and trade wind-influenced climate. Thus, detailed knowledge of the sensitivity of the island grids to variable wind and solar generation is crucial; system operators need better foreknowledge of imminent changes in wind power and PV output and the potential for large-scale ramp events occurring on short time scales. Here we present an overview of the development, validation, institution, and operational performance of SWIFT, (Solar and Wind Integration Forecast Tool), a service for Hawaii's electric utilities that provides short-term forecasts (~15 minutes up to next day) of wind and solar-based generation that is updated several times per day. SWIFT was developed in conjunction with WindNET , a dedicated sensor network comprised of remote monitoring platforms (e.g. radiometer, sodar, lidar) strategically placed around wind projects sites in Hawaii under a project funded through the Department of Energy and supported by the Hawaiian Electric Company, and through an Electric Power Research Institute sponsored project to develop solar forecasting services for Hawaii.
 The primary purpose of SWIFT is to equip system operators with timely and accurate foreknowledge of ‘next-hour' and ‘next-day' wind and solar-based generation so that they can readily accommodate the increasing penetration of this type of generation while economically balancing other resources and maintaining high levels of grid reliability. SWIFT incorporates state-of-the-art physics-based and statistical forecasting techniques with innovative adapations for Hawaiian weather regimes while developing, through feedback from utility operators, and ongoing analysis of system performance, technological improvements that are customized for Hawaii's unique climate and utility operating environment. Work presented here will feature 1) discussion of metrics used to characterize forecast accuracy, 2) an assessment of SWIFT's wind and solar forecast performance under the most challenging forecast scenarios (e.g. rapidly changing sky conditions leading to short-term large-scale ramp events); and 3) a summary of how operator feedback was used to further refine forecast presentation and ease of use.

Other regions of the country, such as California, Texas, the mid-Atlantic and northeastern U.S., are also experiencing rapid growth in the penetration of renewables in their grid operations. Thus, forecast systems such as SWIFT will be needed to enable operators to economically balance generation and load while maintaining high levels of reliability on systems experiencing increasing penetration of renewables on their grid operations.