Short-Term Forecasting of Wind Power Plant Generation for System Stability and Provision of Ancillary Services

Renewable wind generation presents many challenges for accurate and timely generation forecasts due to site topology, turbine layout and their potential wake effects and local network constraints, particularly for generators operating in fast changing, low or high wind resource conditions. Internationally, short term generation forecasting for wind power plants has become a critical issue for market operators and transmission networks to adequately manage system reserve margins and allow for the delivery of ancillary services by wind generators. The authors have investigated and developed machine learning techniques in combination with robust site and plant condition heuristics to produce a short-term generation forecast of wind power plant generation for the 5 to 7 minute timeframes to meet new requirements in the Australian National Electricity Market (NEM) for wind power plant sourced unconstrained generation forecasts. Similarly, wind power plants will need reliable short term generation forecasts to be able to participate in contingency and regulation raise markets in the NEM to offset a significant increase in the generator costs for ancillary services over the last three years. The forecasting system will be implemented as a standalone device designed to integrate directly with wind power plant SCADA systems and therefore can be deployed at many new and existing wind generators. Preliminary generation forecasting results are very encouraging and already demonstrate a substantial improvement in the accuracy of short term forecasts in comparison to simple generation state estimation and market operator generated forecasts for actual wind power plant case studies for a wide range of generator sites. Accurate and timely short term forecasts will have a critical role to play in ensuring improved system stability for power systems with a large or increasing proportions of wind power plant generation. The proposed system has the potential to fulfill a missing vital technological component for improved wind generator revenue through improved dispatch and ancillary services income streams and cost reductions and lead to increased power system stability and reliability.