Continuous-Time ARMA Models for Data-Based Wind Speed Models

The focus of this paper is on the dynamic analysis of power systems including wind generation. A continuous-time implementation of the well-known Autoregressive Moving Average processes is proposed. This model is based on measured data and built through a four-step procedure that retains both the probability distribution and the autocorrelation of the actual wind speed. The effect of the data sampling rate on the stochastic properties of the wind speed is evaluated through the comparison of real-world hourly and minutely data. Finally, the proposed model is included in a benchmark system and its transient behavior is discussed.