An implementation of synthetic generation of wind data series

Wind power fluctuation is a major concern of large scale wind power grid integration. To test methods proposed for wind power grid integration, a large amount of wind data with time series are necessary and will be helpful to improve the methods. Meanwhile, due to the short operation history of most wind farms as well as limitations of data collections, the data obtained from wind farms could not satisfy the needs of data analysis. Consequently, synthetic generation of wind data series could be one of the effective solutions for this issue. In this paper, a method is presented for generating wind data series using Markov chain. Due to the high order Markov chain, the possibility matrix designed for a wind farm could cost a lot of memory, which is a problem with current computer technologies. Dynamic list will be introduced in this paper to reduce the memory required. Communication errors are un-avoidable on long way signal transmission between the control centre and wind farms. Missing of data always happens in the historical wind data series. Using these data to generate wind data series may result in some mistakes when searching related elements in the probability matrix. An adaptive method will be applied in this paper to solve the problem. The proposed method will be verified using a set of one-year historical data. The results show that the method could generate wind data series in an effective way.