Optimal Energy Management of Hybrid Power System with Two-Scale Dynamic Programming

AbstractHybrid power system (HPS) is the power systemconsists of renewable energy sources and traditional energysources used together to increase system efficiency and reduceoperation cost. Energy management is one of the main issues inoperating the HPS, which needs to be optimized with respect to thecurrent and future change in generation, demand, and marketprice, particularly for HPS with strong renewable penetration.Optimal energy management strategies such as dynamicprogramming (DP) may become significantly suboptimal understrong uncertainty in prediction of renewable generation andutility price. In order to reduce the impact of such uncertainties, atwo-scale dynamic programming scheme is proposed in this studyto optimize the operational benefit based on multi-scale prediction.The proposed idea is illustrated with a simple HPS which consistsof wind turbine and battery storage with grid connection. Thesystem is expected to satisfy certain load demand while minimizingthe cost via peak-load shaving. First, a macro-scale dynamicprogramming (MASDP) is performed for the long term period,based on long term ahead prediction of hourly electricity price andwind energy (speed). The battery state-of-charge (SOC) is thusobtained as the macro-scale reference trajectory. The micro-scaledynamic programming (MISDP) is then applied with a short terminterval, based on short term-hour ahead auto-regressive movingaverage (ARMA) prediction of hourly electricity price and windenergy. The nodal SOC values from the MASDP result are used asthe terminal condition for the MISDP. The simulation resultsshow that the proposed method can significantly decrease theoperation cost, as compared with the single scale DP method.6 Halama