The Effect of Markov Chain State Size for Synthetic Wind Speed Generation

Abstract

In this study hourly wind speed time series data of Eskisehir region, Turkey have been used for stochastic generation of wind speed data using the transition matrix approach of the Markov chain process. Previous work on Synthetic data generation did not focus on the effects of different choices of wind states. In this work, it was observed that increasing the number of states has a significant benefit in terms of generated data quality. Two different Markov models are constructed. In first model 13 wind states are used whereas in the second model 26 wind states are used to form the transition probability matrices of the model. The algorithm to generate the wind speed time series from the transition probability matrices is presented. The generated data from both models are compared with observed ones. The comparison of the observed wind speed and the generated ones shows that statistical characteristics are satisfactorily preserved. Furthermore increasing the dimension of state space of the Markov model gives more accurate results.