A two-level real-time energy management strategy incorporating load forecasting for hydrogen fuel cell-powered ships

Abstract

This paper proposes a two-level real-time energy management strategy (EMS) for hydrogen fuel cell-powered ships, incorporating load forecasting, with the aim of comprehensively enhance both the energy economy and system durability. The proposed EMS is divided into two parts: ship load forecasting (SLF) and two-level real-time management (TRM). SLF employs low-pass filtering (LPF), variational mode decomposition (VMD), and an adaptive-weight lightweight gradient boosting machine (AW-LightGBM) to effectively address the uncertainty and high volatility of ship loads. This enables multi-step load forecasting at the second-level timescale, yielding a future ship loads sequence (FSLS) that is then passed on to the TRM. The TRM uses a two-level model predictive control (MPC) framework: the upper-level MPC optimizes the overall efficiency of two sets of proton exchange membrane fuel cells (PEMFCs), while the lower-level MPC minimizes the energy costs and the performance degradation costs of two sets of PEMFCs and the lithium battery (LB). Finally, the two types of energy supply devices are scheduled according to the control signals generated by the lower-level MPC. Hardware-in-the-loop (HIL) experiments indicate that, compared to other strategies, the proposed EMS significantly enhances both the energy economy and system durability of hydrogen fuel cell-powered ships during operations. Under various scenarios, the proposed EMS can achieve maximum reductions of up to 12.8% in total operating costs, 25.2% in performance degradation costs of the PEMFCs, and 17.8% in performance degradation costs of the LB, respectively.