Research on Energy Management Strategy for Marine Methanol-Electric Hybrid Propulsion System Based on DP-ANFIS Algorithm

To address the challenges of high fuel consumption and emissions in a traditional diesel-powered inland law enforcement vessel, a retrofit design incorporating a methanol-electric hybrid propulsion system was proposed. Corresponding energy management strategies were developed using both rule-based (RL) and dynamic programming (DP) algorithms. To further enhance the energy management performance, a collaborative optimization algorithm integrating DP with an adaptive neural fuzzy inference system (ANFIS) is introduced. This hybrid approach leverages the global optimization capability of DP and the real-time adaptive learning characteristics of ANFIS to achieve efficient power distribution, thereby improving fuel economy and reducing emissions. A simulation model of the ship’s methanol-electric hybrid propulsion system was first developed based on methanol engine bench test data and integrated with models of other powertrain components. The DP algorithm was employed to solve for the optimal control sequence offline. Subsequently, ANFIS was used to learn the DP-derived optimization results online, generating real-time energy allocation rules adapted to actual operating conditions. This method significantly reduces computational burden and enhances system responsiveness. Simulation results demonstrate that, compared to the RL algorithm, the proposed DP-ANFIS algorithm reduces total energy consumption by 78.53% while increasing the battery state of charge (SOC) by 3.24%. In terms of fuel economy, methanol consumption is reduced by 64.95%, and the brake-specific fuel consumption (BSFC) is reduced by 81.26%. Regarding pollutant emissions, CO, HC, and NOx are reduced by 82.91%, 83.4%, and 15.2%, respectively. CO₂ emissions, a key indicator of carbon footprint, are reduced by 81.12%. These performance improvements are comparable to those achieved by a DP-only energy management strategy. Finally, hardware-in-the-loop tests further confirm the practical feasibility of the DP-ANFIS algorithm for real-world engineering applications. This method offers a novel technical pathway for intelligent energy management in marine hybrid propulsion systems, combining significant theoretical value with strong potential for engineering implementation.