Research On Wind Farm Modeling Based On Direct Drive Permanent Magnet Synchronous Wind Turbine

Wind power generation,as one of the most mature,large-scale development conditions and commercial development prospects of renewable energy power generation,has achieved large-scale development and utilization in the global scale.Due to the randomness and fluctuation of wind power generation,its large-scale grid connection will have a huge and not negligible impact on the voltage stability,transient stability,and power quality of the power grid.It has become one of the important topics to study the influence of large-scale wind power on the operating characteristics of the power system after grid connection,and the primary issue of research is wind farm modeling.The equivalent modeling method of wind farm composed of direct drive permanent magnet synchronous wind turbine is studied and a method of wind farm equivalent modeling based on principal component analysis is proposed.Equivalent modeling of wind farms mainly includes three aspects:grouping of wind turbines in a wind farm,calculation of equivalent wind turbine parameters,and equivalent calculation of a collector system.The grouping of wind turbines in a wind farm is generally based on the principle that the wind turbines have the same or similar operating points.Although all state variables that characterize the operating point of a wind turbine can be used as an index,it can be accurately classified,but for large-scale wind farms,its calculation is too large to be applied.In view of the certain correlation between various state variables,all state variables,β、T_m、T_e、ω_m、ω_r、u_d、u_q、i_d、i_qthat reflect the operating points of the wind power unit,are obtained by analyzing the operating characteristics of the wind power unit.Then,three leading variables,β、T_e、ω_rare extracted by using the principal component analysis method as cluster indexes.Since the dominant variable can represent more than 90%of all state variables,the clustering calculation with the dominant variable as the clustering index can accurately reflect the operating point of the wind turbine and greatly reduce the calculation amount.After the clustering is completed,the capacity-weighted method is used to calculate the parameters of the same group of equivalent wind turbines,and the equivalent wind farm can be set to several wind turbines.Finally,the common electrical connection methods between wind turbines in a wind farm are introduced,and the equivalent impedance of the collector system is calculated using the single impedance equivalent method.Since the wind speed is not the only factor that determines the operating state of the wind turbine,taking wind speed as clustering index will lead to large error.For comparative analysis,the wind farm is firstly clustered with the dominant variable and wind speed as the grouping index.Then,according to different clustering results,two equivalent models and detailed models of wind farm are built on Matlab/Simulink simulation platform.The analysis and comparison of the dynamic characteristics of the three models in wind speed fluctuation and grid fault verify the correctness and high accuracy of the proposed wind farm equivalent modeling method based on principal component analysis.