Event Informed Capacity Planning for Port Electrification with Multiscale Hourly Demand Scenarios

Authors:
Preprint
Publisher website
Google Scholar
Abstract
Electrifying ports transfers energy use from fossil fuel to electricity, but annual energy projections alone do not establish the grid connection and substation capacities required during coincident vessel service events. This study develops a capacity planning framework that converts pathways for throughput, vessel mix, equipment electrification, onshore power supply (OPS), terminal migration and flexibility into ensembles of annual hourly demand trajectories. A multiscale graph state-space model with conformal prediction (MGSS-CP) is used as the mapping from operational events to load rather than recursively forecasting one hour ahead for ten years. Annual maxima provide capacity quantiles, capacity gaps and reinforcement deferral. On the chronological 2025 hold-out, MGSS-CP attains a one-hour mean absolute error of 1.06 MW and an R2 of 0.957, while its 90% interval coverage remains 0.899 to 0.901 across the evaluated horizons. Under the fully electrified pathway, annual demand reaches 625.7 GWh and the central scenario peak reaches 122.6 MW by 2035. Scaling annual energy at the 2025 load factor gives 102.0 MW. It therefore understates the peak by 20.6 MW and the additional capacity above 2025 by about 40%. Alternative event placement raises the peak to about 129 MW, whereas an assumed 18% peak shaving case lowers it to about 100.5 MW and indicates 22.1 MW of gross reinforcement deferral. The results show why port capacity planning needs a validated hourly representation of operational events, explicit long-term pathways and uncertainty statements appropriate to each horizon.

This publication has 12 references indexed in Scilit: