[HRE17-P03] Predictability of Wind Power Ramp Events based on Dynamical and Statistical Ensemble Prediction
Keywords:Wind power, ramp events, prediction
Abrupt change in wind power generation, known as ramp events, due to fluctuating wind speeds present challenges to the stability of the electric power supply. In Japan, the ramp events are generally induced by extratropical cyclones along their track. Since prediction for the behavior of the cyclones entails uncertainty caused by nonlinear dynamics, probabilistic prediction is more effective. For this study, we developed a regional ensemble prediction method using the Weather Research and Forecasting model (WRF). To obtain dynamically consistent perturbations with a synoptic weather pattern, both initial and lateral boundary perturbations were determined by differences between the control and an ensemble member of the Japan Meteorological Agency (JMA)'s operational one-week ensemble forecast. This method provides 11 ensemble members with a horizontal resolution of 15 km for 75 hours at 30 minutes interval outputs by downscaling JMA's operational global forecast along with the perturbations. Wind power is projected using power curve which is estimated by relationship between area averaged wind speed and wind power generation. In addition to the dynamical ensemble, we expanded the ensemble member using Monte Carlo method to take account of the impact on uncertainty in the power curve. The expanded ensemble has 100 members. Forecast for the ramp events using the expanded ensemble exhibits high statistical consistency and reliability of ability to capture the ramp events compared with the dynamical ensemble prediction.