Contents I. Introduction
This paper presents a new method for estimating the upper and the lower bounds of wind speed. The reduction of CO2 is of main concern after the Kyoto Protocol in 1997 [1], [2]. As a result, power utilities are quite interested in renewable energy such as wind power generation, solar energy generation, geothermal energy generation, etc. European countries have developed wind power positively and occupy 70% of the wind power generation in the world. Similarly, Japan also accepts wind power generation to protect the environment. Although wind power generation seems attractive, it is not easy to predict output due to the uncertainty of the wind speed. In practice, it is reported the power generation output is affected by the wind speed remarkably. In other words, the power players are interested in the prediction of the wind speed in consideration of the uncertainty. So far, a lot of methods have been developed to predict short-term wind speed [3]–[9]. However, they are not so sophisticated techniques since the prediction of wind speed is much more difficult as a time series prediction problem. In this paper, an efficient method is proposed to predict the upper and the lower bound of wind speed as well as the average. The proposed method is based on the kernel machine technique that has elegant theoretical background. This paper makes use of Gaussian Process (GP) with Bayesian estimation to construct the prediction model. The Proposed method is tested for real wind speed data.
