Modelling Wind Farm with Doubly Fed Induction Generators Based on Characteristic Fusion

In this paper, a novel modelling approach based on characteristic fusion is proposed and used to build a static equivalent model of wind farm. Firstly, the modelling framework based on characteristic fusion is given. Secondly, the basic characteristics of wind farm including characteristic of wind turbine generator (WTG), wind speed spatial distribution and characteristic of wind farm are analyzed according to the framework. Then detailed modelling process is provided utilizing SVR as a fusion tool. This approach combines the advantages of both mechanism and non-mechanism methods with both satisfactory fitting ability and generalization ability. It only requires the maximum and minimum value of wind speed among the wind farm, rather than accurate wake model as mechanism method nor massive measurement data as non-mechanism method. Numerical simulation indicates the effectiveness and robustness of the proposed method. When available data is reduced or includes bad measurement, the proposed method can still keep favorable performance.