Modeling Doubly Fed Induction Generator Based Wind Farms by Aggregating the Mechanical Parts Using Clustering by Power Coefficient

Abstract

Due to the increase in wind power production and the increasing development of wind farms its impact on power systems cannot be neglected and modeling of wind farms has been converted to an important topic in power engineering. When studying the effects of a wind farm on the power system, Instead of modeling the wind farm in detail, the performance of the wind farm at the point of the common coupling can be considered as an equivalent model includes an equivalent generator and an equivalent turbine. This approach in addition to simplify the system analysis would considerably reduce the simulations burden. In order to improve the accuracy of this equivalent model the mechanical parts of the wind turbines can be replaced with more than one equivalent turbine using clustering algorithms. Traditionally these algorithms consider the wind speed as the feature for clustering. In this paper it is proposed to classify the wind turbines by power coefficient of each turbine instead of the wind speed. The simulation results obtained using this approach is verified against the conventional approach and also the detailed model. Simulations are done in MATLAB. The results indicate that the accuracy of the equivalent model obtained using proposed approach, is considerably improved.

Keywords

• wind farm
• doubly fed induction generator (dfig)
• equivalent model
• k-means algorithm