An Evidence Theory Based Advisory Layer for All-Electric Ship Power Systems

This paper proposes an advisory layer for all-electric ship power systems. The layer considers the limited operating time aspect of generator sets that the sacrifice of instant benefits in order to achieve long-term goals of long service time and system reliability could be beneficial. In this work, the objective is to ensure a satisfying availability for all generators during a mission. To this end, forecasting models are constructed to predict failure rates and load demands as these measures are closely related to system availability. The predictions are then fed into the energy management layer to reconfigure the optimal resource allocation once the quality of service is met. Evidence theory is used to deal with uncertainty in our work. The integration of the layer into existing designs of ship power systems is discussed as well. A simulation is carried out to demonstrate the proposed work.