Dynamic multi-turbine multi-state model of wind farm based on historical wind data

In order to enhance the accuracy of the dynamic equivalence of wind farm (WF) under different wind conditions (WCs), this paper proposed a Dynamic Multi-Turbine Multi-State (DMTMS) Model of WF based on the historical wind data. The proposed model could represent the dynamic characteristics of WF under different WCs with high accuracy. Support vector clustering (SVC), whose cluster partition is completed by the genetic algorithm (GA), is adopted so as to handle the varietion of wind energy with the pre-fault active power of wind turbines (WT) as input parameters. Equivalence model of cable is established with the principle of maintaining the terminal voltage of wind turbines unchanged. The model is demonstrated on a WF consisting of 133 WTs connected to the grid with a transmission line. Dynamic characteristics of DMTMS are compared against the detail WF model under different WCs. Results demonstrated that the DMTMS model can adapt to different wind conditions.