基于实测数据的风电场稳态等值建模研究  

Research on wind farm steady-state equivalent modeling based on measured data

作　　者：李牡丹[1] 王印松[2] 刘霜 

机构地区：[1]华北电力大学科技学院,河北保定071003 [2]华北电力大学控制与计算机工程学院,河北保定071003

出　　处：《可再生能源》2018年第3期438-445,共8页Renewable Energy Resources

基　　金：华北电力大学中央高校基本科研业务费专项资金资助项目(9161715008)

摘　　要：针对大型风电场仿真模型复杂,计算量大的特点,提出一种利用实测数据建立大型风电场稳态等值模型的新方法。通过粒子滤波算法对原始风速数据进行预处理,消除噪声干扰对真实数据的影响;考虑到风电场内风机的风况差异问题,采用K-均值聚类算法提取反映风电机组风况差异的特征风速,以简化建模过程;选择特征风速和实测风电功率为输入、输出信号,应用BP神经网络拟合风电场稳态等值模型。该模型考虑了风电场地形地貌、机组分布因素,利用实测数据对模型进行泛化能力分析和精度验证,仿真结果表明,该方法具有一定的准确性与合理性。In order to resolve complexity and large calculation of large-scale wind farm simulation model, a new method of wind farm steady-state equivalent modeling with measured data is proposed in this paper. Utilizing the particle filter algorithm,various interferences in measured wind speed can be eliminated;in consideration of wind regime differences between wind turbines ,the K-means clustering algorithm is introduced to extract the featured wind speed reflecting wind regime differences so as to simplify the modeling process;the featured wind speed and measured power data are collected respectively as the input and output signals,BP neural network is employed to establish the wind farm steady-state equivalent model. The proposed model involves topography and turbines distribution of the actual wind farm,measured data are adopted to analyze its generalization ability and verify its accuracy,simulation results indicates that the modeling method is precise and rational.

关 键 词：风电场 实测数据 BP神经网络 稳态 等值建模 

分 类 号：TK8[动力工程及工程热物理—流体机械及工程] P457.5[天文地球—大气科学及气象学]