Research On Equivalence Of Wind Farms And Probabilistic Load Flow Calculation Considering Dependence

With the maturing of wind power technology,new wind farms tend to be large-scale and centralized.Inherent randomness,volatility,and dependence of wind power bring huge challenges to the safe operation of power system.Probability load flow(PLF)is a powerful tool to study the influence of wind power integration on the system operation.In view of the PLF problem of power system containing wind farms,this thesis studies the equivalent model of wind farm,wind speed model between wind farms and PLF calculation method considering correlation.This thesis proposes a new equivalent model and PLF algorithm.That provides a new thought for economic dispatch and safety analysis of power system.In order to reduce complexity of simulation model of wind farm while ensuring accuracy,and considering wake effect and different control methods in actual wind farm,a two-step clustering method based on Analytic Hierarchy Process is proposed.Firstly,the input wind speed is calculated under the influence of wake effect.The first step clusters the DFIGs according to pitch angle action situation,and the second step re-clusters the DFIGs by their dynamic characteristics according to disturbed curves of rotor current.Then parameters of equivalent wind turbine generators and collecting power lines are derived respectively to ensure power output characteristics and voltage difference constant.Thus a multi-machine equivalent model of wind farm is obtained.Finally simulation results show that,comparing with traditional equivalent method,this way can reflect accurately both initial operating point and dynamic behavior.There are varying degrees of wind speed correlation between different wind farms.It is significant for economic dispatch and safety analysis of power system to build exact wind speed correlation model.Wind speed marginal distribution is estimated by non-parametric kernel density estimation.Based on single fitting indexes,Copula entropy is introduced as a correlation index.An optimal Copula selection method based on entropy weighted fuzzy comprehensive evaluation is proposed.And then wind power correlation model is built adopting optimized Copula function.The validity of proposed method is tested by two wind farms in North China.In view of the contradiction between calculation precision and speed of the existing PLF calculation method,a three point estimate method based on Nataf transform is introduced to calculate PLF.This method can effectively deal with the correlation between wind farms,and can guarantee high calculation accuracy with reducing computing time.This method and Monte Carlo method is used respectively to calculate PLF of IEEE30 node system containing many wind farms.Finally,efficiency and accuracy of this method are verified through results comparison.