Longitudinal moment Markov chain model of wind power and its application on ultra-short-term prediction

• Published in: 2015 5th International Conference on Electric Utility Deregulation and Restructuring and Power Technologies (DRPT) pp. 1874 - 1878
• Main Authors: Jingwen Sun, Zhihao Yun, Jun Liang, Xiaojuan Yang, Libin Yang, Xueli Wang
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.11.2015
• Subjects: Data models; longitudinal moment; Markov chain model; Markov processes; Power systems; Predictive models; Time series analysis; Wind farms; wind power; Wind power generation; wind power prediction
• DOI: 10.1109/DRPT.2015.7432553

In this paper, a longitudinal moment Markov chain model of wind power time series based on the longitudinal time concept is proposed. This model emphasizes the transition characteristics related to different moments by providing a set of transition probabilities matrices. This matrices set, describing the inherent transition information of moments, gives the necessary probabilistic conditions for optimization decision of power systems containing wind farm. Besides of rapid calculation as conventional Markov chain model has, the proposed model makes the transition information more detailed and accurate. To illustrate the effect of improvement, a wind power prediction (WPP) method on ultra-short-term horizon using the longitudinal moment Markov chain model is put forward. The case study based on actual wind power data under multiple time scales shows that the proposed method achieves a higher prediction precision.