Please login first
A Machine Learning–Enabled Energy Management Tool for Flexible Port Operations and DER Integration
* , ,
1  Institute for Energy Engineering, Universitat Politècnica de Valencia, Camino de Vera s/n, 46022 Valencia, Spain
Academic Editor: El Manaa Barhoumi

Abstract:

Port decarbonization requires integrated energy solutions that reduce fossil-fuel dependence while maintaining safety, reliability, and operational continuity. This contribution proposes a machine learning–enabled energy management tool designed to support flexible port operations by forecasting demand on operational timescales and coordinating distributed energy resources (DERs) within a modular, deployment-ready architecture.

The tool combines short-term load forecasting with scenario-based representations of port activity. Instead of relying on fully instrumented real-time submetering, operational states are inferred using available indicators such as time-of-day patterns, workload proxies, and configurable activity levels (e.g., low/medium/high handling intensity). These inputs are used to estimate demand trajectories and flexibility opportunities, enabling the scheduling of on-site generation and storage without disrupting core operations.

The proposed framework is modular and intended for phased deployment, allowing progressive integration as local operational and metering data become available. A demonstration case study considers renewable and flexible resources that are relevant to port environments, including photovoltaic pavements for low-speed operational areas to supply auxiliary loads (lighting, signage, monitoring), wave energy-converting seawalls integrated into coastal protection infrastructure, and vertical-axis wind turbines for turbulent and space-constrained locations. System robustness is enhanced through battery energy storage systems for peak smoothing and backup-oriented flexibility. In addition, regenerative braking in mobile equipment is discussed as an energy-recovery pathway in frequent start–stop duty cycles. A hydrogen-based supply concept for fuel-cell terminal tractors is included as a decarbonization option for internal logistics, aligned with port electrical infrastructure and safety constraints.

Overall, the contribution provides integration criteria and a scalable roadmap toward net-zero port operations, emphasizing practical implementation steps and data requirements for moving from proof-of-concept to pilot deployment.

Keywords: Machine learning; Energy management; Port decarbonization; Distributed energy resources; Flexibility; Renewable integration.

 
 
Top