VENTOS – Prediction and generation of Wind scenarios
Safe and economic integration of wind farms to the electricity system requires reliable predictions of the availability of wind resources with time horizons varying between one hour and one week ahead. To achieve this goal, Cepel, within the scope of the Department for Energy Optimization and the Environment (DEA), has been developing the VENTOS Program, a short-term wind power forecasting tool based on quantile regression model.
A quantile regression model estimates the quantile τ (for example, τ = 0.5 for the median) of the response variable as a function of the explanatory variable. In the VENTOS Program the response variable is the wind power production and the explanatory variable is the wind speed represented by B-Splines basis. The probability density function contains all the information of a random variable and the quantiles are a simple way of describing it. Then, based on a quantile regression model, the VENTOS Program can provide hourly forecasts of the wind production probability density functions from hourly wind speed forecast.
The VENTOS Program provides probabilistic forecasts of the wind power production, which is a much better approach to handle with the wind intermittency and uncertainty in wind speed forecasts as compared to point estimations. Nevertheless, the VENTOS Program can also provide point forecasts (medians) of the wind power production for single scenario unit commitment optimization that can be carried out by DESSEM model.
Contact the responsible area via email: