Load Prediction of Ship Load Profiles for Hybrid Marine Energy Management Systems Using Neural Networks

Shipping is the backbone of the global industry, as most goods are being shipped on seaways. This work presents an implementation of neural networks for the predictions of load demand profiles within a hybrid ship propulsion system powered by a fuel cell and a battery. Implementing these algorithms in an energy management system (EMS) ensures the safer operation of ship propulsion components within their operational limits. The methodology is benchmarked against measured ship load profiles. The developed methodology achieves a high prediction accuracy of up to 97% using a Long Short-Term Memory (LSTM) algorithm, which enables higher efficiencies of the propulsion system.