Determination of dynamic wind farm equivalents using heuristic optimization

As a result of the increasing share of wind power the dynamic behavior of power systems will change considerably. To carry out stability studies in the future wind turbine and wind farm dynamic models will be indispensable. Generic models seem to provide the required simplicity and accuracy. But the parameters cannot be derived directly from the mathematical models of the generator and converter system, numerical identification methods are needed. In this paper the authors introduce a new heuristic optimization method called Mean Variance Mapping Optimization (MVMO) which provides excellent performance in terms of the accuracy of the generic model parameters and convergence behavior. The fitness evaluation is performed using time domain simulation in each iteration step. The procedure and the level of accuracy that can be reached are demonstrated using an 18 machine, 90 MW test wind farm consisting of DFIG based wind turbines.