Optimal scheduling of a renewable energy-based park power system: A novel hybrid SDDP/MPC approach
Lei Qiang, Yuan Huang, Xiao Xu, Feng Zhu, Yuyan Yang, Junyong Liu, Weihao Hu
Abstract
This paper proposes a hybrid real-time scheduling scheme for a park power system at multi-time scale, including stochastic dual dynamic programming (SDDP) and model predictive control (MPC). This approach realizes hierarchical suppression of real-time power fluctuation of distributed generation and load. Multiple uncertainties, including generation and consumer load, are considered to achieve better economic scheduling. A novel multistage SDDP feedback correction scheduling scheme with an hour as the discrete time unit is constructed, which embeds MPC into SDDP scheduling time precision, focusing on maintaining the power to the distribution network as constant as possible. In addition, the response performance of the 5G macro base station in the real-time tracking control procedure is also explored in this work. The base station absorbs a small range of uncertainties by using the migration characteristics of the communication load. An example of park system scheduling verifies the effectiveness of the proposed approach, and realizes hierarchical suppression of real-time power fluctuation of distributed generation and load caused by prediction errors that vary over time-scales.