基于PCA和CA-ST方法的风电场等值建模研究

EQUIVALENT MODELING OF WIND FARM BASED ON PCA AND CA-ST METHODS

在风电场等值建模的研究中发现,当聚类指标选取太少时,机组信息不全面,导致等值模型和详细模型之间存在较大误差;当指标选取太多时,增加聚类工作量,且变量间的相关性和数据上的冗余性若不处理,也将影响结果的准确性。基于此,该文选取包含风电机组机械、电气特性变量的动态数据作为聚类指标,使用主成分分析(principal component analysis,PCA)消除其强相关性和冗余性;提出相关性分析和显著性检验(correlation analysissignificance testing,CA-ST)聚类方法,建立风电场等值模型。仿真结果表明,基于PCA和CA-ST方法得到的等值模型在表征风电场的输出特性上具有较高精度,且在改变风速情况下对风速扰动和短路故障工况都具有较好的适应性。

In the research of equivalent modeling of wind farm,it is found that when the selected clustering indices are inadequate,the information of wind turbines is not comprehensive,leading to large errors between equivalent models and detailed models.When excessive clustering indices are selected,the workload of clustering increases,and if the correlation between variables and the redundancy on the data are not processed,the accuracy of clustering results will also be affected.This paper selects the dynamic data including mechanical characteristics variables and electrical characteristics variables of wind turbines as clustering indices and uses principal component analysis(PCA)to eliminate the strong correlation and redundancy between data indices.The clustering methods based on correlation analysis and significance testing(CA-ST)are proposed to establish the equivalent model of wind farm.The simulation results show that the high accuracy of the equivalent model of wind farm obtained by PCA and CA-ST in characterizing the output characteristics of wind farm.The established equivalent model also has adaptability to wind speed disturbance and short circuit fault when the wind speed changes.