Stochastic Modelling of Wind Speed Power Production Correlations



Based on measurements we investigate the velocity-power characteristic of a 2 MW wind turbine. We apply a stochastic analysis where we describe the evolution of the power output with a Langevin equation, with special respect to short-time fluctuations in wind speed.

Standard procedures, such as the IEC 61400-12 standard, are limited due to the fact that only mean values over several minutes of wind speed and power output are considered. According to this, short-time dynamics of wind and power fluctuations are usually not taken into account. We introduce an improved method which enables us to extract these dynamics of the power characteristic from the measured data.

In particular, we get the response dynamics of the power L (u (t)) via the estimation of Kramers-Moyal coefficients, describing its evolution in timeequation image(t) with a Langevin equation where we separate the power output into a relaxation and a noise part. A fixed-point analysis provides the required power characteristic. (© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)