Data-Driven Propulsion Load Profile Prediction for All-Electric Ships

To achieve high-efficiency power and energy man-agement in a modern hybrid shipboard power system, an accurate shipload profile has become the pre-condition for opti-mization of the power and energy control. With the introduction of the all-electric ship concept, propulsion load is determined by the electrical power demand instead of the propeller mechanical torque. In this study, a data-driven load profile prediction model of the ship propulsion power is developed with machine learning methods. A typical ferry boat operation data is utilized as the case study to verify the prediction accuracy. The prediction model is approached via three different regression models and the predicted results are studied and compared. The marine power load forecasting can improve power system efficiency and reduce the energy cost through a better understanding of the power system demand pattern.