Learning uncertainty of wind speed forecasting using a fuzzy multiplexer of Gaussian processes

The smart power systems of the future will be able to accommodate wind power at a maximum efficiency by utilizing available information. For instance, information pertained to wind speed is essential in forecasting the overall amount of power generated by wind farms. Information is used to offset the inherent stochasticity of wind power and improve wind speed forecasting precision. In this work, an intelligent methodology for quantifying the uncertainty of wind speed pertained to forecasting is introduced. The introduced methodology adopts a set of Gaussian processes to assemble a model of the uncertainty of the forecasted speed. Results are taken on a set of real-world wind speed data.