Single Parameter Exponential Time Series Smoothening to Dampen Wind Power Irregularities in Grid-Integrated Doubly Fed Induction Generator

Stochastic behavior of wind leads to power irregularities and subsequent frequency fluctuations in renewable sources based small scale community power system, connected to weak AC network. To take care of such variability, this paper proposes a smoothening method based on one-period ahead prediction of reference grid power with a doubly fed induction generator (DFIG) for a wind energy conversion system (WECS). The variability of output wind power is levelled through a battery energy storage (BES) integrated at the DC link of back to back connected power converters in a DFIG. The grid power prediction is performed using a single parameter exponential smoothening (SPES) algorithm, which provides exponential weights to the time series wind data to dampen the irregularities in wind pattern. For a realistic response, the scheme is implemented for wind data obtained from an actual wind farm site located at Kutch district in Gujarat. Both simulation and test results are displayed to demonstrate the validity of the SPES algorithm in smoothening the wind power irregularities.