Study On Wind Speed And Short-term Power Prediction

In recent years, environmental pressure in China is increasing, the installed capacity of wind power has rapid growth. The output power of wind farm has randomness characteristic, when the wind power into the grid, the proportion of wind power more than a certain coefficent, will brings the serious influence on the security of the power system. So when large-scale wind power into the grid, should need to make prediction of short-term power.Based on the analysis of wind speed statistical rules, do some research on short term wind power prediction. The main contents are as follows:1, Analysed statistical characteristics of wind speed distribution, and fit the statistical distribution parameters.Turned the NWP data into wind turbine’s wind speed through the spatial transform, calculating the wind farm’s wind wake, and get the prediction of the wind power. The example shows that the model has high forecasting accuracy;2、Make the selection of variables through the Neural network, get every input vector’s influence to wind power. Using the prediction sample as the center, finds the similar samples as the training set. At the same time, get the initial weights of neural network through genetic algorithm, the example shows that the prediction precision can be effectively improved;3、Combining the two kinds of forecasting method, compare the forecast accuracy of different combinations method, and analyzes the causes of error.Last, based on the above research results, developed the wind power prediction system, and give a brief introduction of the system, including the system’s architecture, the interface of the system.This paper get support from the Green Energy Technology Key Laboratory of Guangdong Province, wind power control and grid connected of Guangdong Province, wind power control and grid connected of the National Local United Engineering Laboratory,securing loans of from the State Natural Science Fund Project(61273172), Guangdong Province of strategic emerging industry technology research project(2012A032300013) fund.