A study of battery operational optimization with data-driven clustering

Abstract

Environmental problems have led to continuing efforts to reduce fossil fuel consumption around the world. As a result, interest in battery-based hybrid systems is increasing in the shipbuilding and offshore industries. In particular, battery applications are more efficient for offshore vessels with frequent load variations and high peak power consumption. Propulsion systems are gen-erally classified as direct or electric propulsion. For some vessels, both direct and electric propulsion are used. The electrical power system of a vessel consists of one or multiple grids depending on the status (open/closed) of the bus tie. Owing to the complexity of propulsion and electrical power systems, designing the operation method and specifications of the battery onboard the vessel remains a challenge. Therefore, this paper categorizes and analyzes the data according to the condition of the bus tie. Principal component analysis clustering is applied to define the ship operation mode. The entire profile of a hybrid vessel with the hybrid propulsion sys-tem from a data point of view is analyzed, and an optimized battery operation method is proposed.