Study On The Cooperative Prediction Of Large-scale Of Wind Power

Nowadays，countries all around world gradually pay more attention to theenergy crisis and environmental problems，and new energy generation comes intomore people’s vision. As the most commercially valuable large-scale development ofclean energy, wind power in China has been vigorously developed. Compared toconventional energy sources, random changes in wind speed and wind directionmake wind power have the characteristics of fluctuant, intermittent anduncontrollable, and have a adversely affected the power system stability andreliability.When large-scale wind farms connected to the grid operation, significantlywind power fluctuations will bring adversely impact to power balance and frequencyregulation. Improve the accuracy of wind power prediction is an important means toachieve the target of operational safety and economy of power systems includinglarge-scale wind power.This article is based on the actual measured wind power data of wind farms inJilin Province, firstly it introduced the reason of wind power data loss, analyzed theadvantages and disadvantages of the four different data completing strategies, andproposed the adaptive neuro-fuzzy inference system to complete and optimize thelosing data of one wind turbine or large scales; Then for fluctuation characteristicsof wind power, using the wavelet theory to decompose and predict differentfrequency component using different predictive methods, and getting the model ofcooperative prediction, aimed at the non-real-time decomposition predictive valuedeteriorate with the increasing of prediction steps in the model, introduced thereal-time wavelet decomposition theory; Furthermore it analyzed prediction error. Itintroduced the concept and the cause of prediction error, and the current evaluationsystem of prediction error description. Then it combined with practical examples,indicating the guidance and significance which wind power prediction error bring to wind power forecasting work; Finally, set up the database management system oflarge-scale wind farm, it can collect, store the real-time wind power data, and givedata support to wind power prediction; meanwhile we can extract and modifyaccording to the user’s personal needs, it can be used in wind farm management,currently the platform has been applied to a wind farm in western of Jilin Province.