Research On Large Scale Wind Farms Operation Characteristics And Optimization Control

Since the new century,the rapid growth of the global economy has brought the rapid growth of energy demand.As fossil energy exhaustion,development and utilization of new energy has become the theme of today's energy revolution.And the wind power is the world's most rapid development new energy,it has achieved ten consecutive years of rapid growth,which is the installed capacity of about 20% a year.However,due to the characteristics of the wind power,its volatility,randomness and uncontrollable problem is serious,grid problem and wind power consumptive problem,have gradually become stumbling blocks restricting the wind power development.Due to the limited conditions in china,whether wind resource conditions or system load capacity has a large gap with Europe and the United States.“Wind Curtailment” pressure is particularly obvious.In this context,large scale wind farms need to improve their control ability,and makes them become "grid friendly" power sources.Around this goal,this paper carried out research in the following aspects: wind power data preprocessing,the smoothing effect in large scale wind farms,large scale wind farms cluster modelling,short-term wind power forcast and optimal control of large-scale wind farms.The details are as follows:In the process of data mining,data preprocessing is important.This paper based on Markov chain theory,establishing a two-way supplementary model,to complement the missing points in wind data.Simulation results show that the model has high accuracy and can meet the requirements of data preprocessing.In this paper,through data analysis,the smoothing effect on the spatial scales has been quantitatively analyzed.Two properties have been found: one is that within the wind farms,varies with the extent of space greatens,wind power fluctuations become smaller,and the other is that,the combination between distantly regions,could better stabilize the fluctuations.Based on the virtual power plant theory and fuzzy-C clustering method,a cluster model of large scale wind farms has been established.In this model,the wind farms have been turned into multiple virtual wind generators.Inside a virtual wind generator,scheduling and controlling is based on the "principle of coherency".As characteristics and laws are the same in wind power singnal of the same frequency part,we choose the wavelet packet transform as the signal analysis method.The short-term wind power forcast model is based on the optimal wavelet packet transform and least squares support vector machine(LS-SVM)hybrid method.Actual data validation show that,the optimal wavelet packet transform improve the prediction precision of the model.Based on the single wind power generator model,set up the full adjustable strategy of the virtual wind generator.Within wind,strategy chooses generator torgue for small range adjustment,and pitche control and generator torgue together for large range adjustment.So as to realize the whole virtual wind generator full adjustable strategy.Finally,based on the short-term wind power forcast,the full adjustable strategy of the virtual wind generator and particle swarm optimization algorithm,optimal control strategy of large scale wind farms has been stated.It provides a new way to make wind farms into "grid friendly" power sources.