Storage Sizing with Peak-Shaving Policy for Wind Farm Based on Cyclic Markov Chain Model

Jiaojiao Dong,Feng Gao,Xiaohong Guan,Qiaozhu Zhai,Jiang Wu
DOI: https://doi.org/10.1109/tste.2016.2637916
IF: 8.31
2017-01-01
IEEE Transactions on Sustainable Energy
Abstract:Due to the intermittent nature of wind power, large-scale wind farm integration creates technical challenges, such as increase of peak/valley net load difference and uncertainty of generation. Energy storage is essential in providing flexibility and ensuring system reliability. Storage sizing problem is widely studied for a given demand curve, and the needed storage capacity to achieve a certain level of peak-shaving performance is not analyzed. In this paper, a probabilistic model of storage sizing with peak-shaving policy optimization under required matching probability is established to minimize net cost considering time-variant energy price. Storage is used not only for reducing energy deficit and keeping generation reliable but also for energy shifting to obtain higher profit. A cyclic nonhomogeneous Markov chain (CNHMC) steady-state analysis method is proposed, serving as a more efficient way to test probability constraint than commonly used time-consuming sequential Monte-Carlo simulation. CNHMC is used in stored power modeling representing diurnal variation of wind power and load. Probability constraint is tested by obtained analytical expression of matching probability. Numerical test shows that reliable power supply is achieved with little profit sacrifice, peak/valley net load difference decreases with little increment on storage capacity, and the proposed solution method is fast and accurate.
What problem does this paper attempt to address?