Matern process-based simulation of wind speed time series

This work proposes a practical methodological framework for simulating long-term discrete wind speed time series, for applications to renewable energy systems. The framework is based on the Matern stochastic process, that is used to identify and extract the short and long-term periodic components present in the mean and variance of a wind speed time series, and to characterize the residual random component. The procedure of construction of synthetic wind speed time series is developed, rooted in a fast numerical simulation algorithm for the Matern process. We empirically validate the usefulness of the proposed framework by numerical experiments based on three real historical wind speed datasets. The results show consistent statistical similarities between the historical and simulated wind speed time series of data.

Automated summary

This paper proposes a practical methodological framework for simulating long-term wind speed time series for renewable energy systems. The framework based on the Matern stochastic process is used to identify and extract the periodic components of wind speed, and characterize the residual random component. The results show consistent statistical similarities between historical and simulated wind speed time series.