Study On The Short Term Wind Speed Forecasting In Wind Farm

Developing and making use of new energy is important strategy in 21 centuries in China. Wind energy has been increasingly embraced and wind generation is one of the most prospective new energy due to its exhaustless. But wind power has the disadvantages of intermittence and randomicity, which will bring challenge to the safety and stabilization of power grid and then restrict the scale of wind power development. Short term wind speed forecasting and wind power forecasting in wind farm is an effective approach for the above problem. The wind farms in China are mostly centralized and large scaled ones, while the power grids construction is weak. Short term wind speed forecasting and wind power forecasting in wind farm is more needed in China. The wind power forecasting is originated from the wind speed forecasting mainly. This paper studied short term wind speed forecasting methods. The main works are as follows:Firstly, the statistics method is used to analyse the time series characteristic of wind speed and the wind speed forecasting method and its application. Wind speed is a non-stationary time series. Secondly, the wavelet analysis with "digital microscope" reputation is used to analyse the have noted wind speed datas. The wavelet decomposition and reconstruction is used in the wind speed datas in Hong Kong and Hexi corridor, the wind speed time series with tendency are decomposed into a low frequency component and several high frequency components. The high frequency signals and the low frequency are forecasted with least square support vector machines. The forecasting result of the original time series is the superposition of the respective forecasting. The short term wind speed forecasting method using the wavelet analysis and LS-SVM is proposed in this thesis. The results of test examples in MATLAB show the wavelet is effective tool for the non-stationary wind speed time series forecasting, the low frequency component and several high frequency components are decomposed well, and the forecasting accuracy is effectively improved using LS-SVM. The simulation results of the wind speed datas in Hong Kong and Hexi corridor shown that the above scheme is feasible.