| The stable operation of distribution system is faced with many challenges because of the inherent randomness,intermittency and volatility of wind power.Especially,with the continuous increase of wind power permeability,higher requirements for the economy and safety of the operation process are put forward by the system.Power prediction and fluctuation suppression are effective ways to deal with the uncertainty of wind power output.Through the short-term power prediction,power sectors can adjust the dispatching timely which reduces the reserve capacity and operation cost.Wind power dispatchability can be improved by fluctuation suppression,which ensures power quality and system stability.Energy storage system in wind farm is an effective solution to compensate power prediction errors and suppress potential power fluctuation.Therefore,short-term prediction of wind power and capacity allocation of energy storage system are the focus of this research.Large power and capacity of lithium iron phosphate battery and small power fast response of super capacitor are composed of hybrid energy storage system(HESS)in this paper.Working principle and output characteristics of wind power system and energy storage system are analyzed.The duality of source and load of energy storage is used to construct the dynamic response model.And the topology and control principle of combined windstorage system are determined by comprehensively considering actual operation characteristics of wind farm.Aiming at the short-term prediction of wind power,Markov chain model and support vector machine model are established based on the historical annual power data of wind farm.The two models are used to predict the 24 h rolling power under typical scenarios,and the prediction error distribution is analyzed statistically.The comparison results verified that the Markov chain model is equipped with higher prediction accuracy and robustness.The control target domain of HESS is developed to suppress wind power fluctuation and compensate prediction error.The control strategy of combined wind-storage system based on wavelet packet and fuzzy theory is proposed.Firstly,the prediction wind power in the period is decomposed adaptively by using wavelet packet decomposition theory,and the low-frequency components that meet the fluctuation criteria are obtained as the target gridconnected power.Secondly,according to comprehensive compensation target of energy storage,the unbalanced power between actual output power of the wind farm and target gridconnected power is decomposed by wavelet packet theory,high and low frequency signals are divided,the power and capacity of HESS are reasonably allocated.Finally,fuzzy control is used for the state of charge(SOC)of HESS to synergistic optimize the power commands.By judging the range of SOC,the charge and discharge power of HESS is modified according to different principles to avoid excessive output.The example is combined with annual measured power data of a 100-megawatt wind farm in Inner Mongolia Autonomous Region,China,and the typical daily wind power output scenes in four seasons are extracted for simulation.Results show that the proposed strategy can reduce the short-term prediction error of wind power through energy storage to improve wind power dispatchability,so that the wind power output can meet the grid-connection goal as far as possible.In terms of enhancing the operation stability of wind power,the rapid response of energy storage can effectively suppress power fluctuations.The proposed strategy has good effects on maintaining the SOC level of HESS and reducing the capacity of energy storage configuration.In this paper,the two measures are combined to provide a certain reference basis for enhancing the controllability of grid-connected wind power. |
PDF Full Text Request