Instability-Aware Digital Twin Control for Grid-Constrained Low-Carbon Smart Ports

Abstract

Modern smart ports combine terminal electrification, shore power, renewable generation, storage, and digital operational control under uncertain vessel traffic and time-varying grid limits. This study develops a simulation-based digital-twin control prototype that synchronizes vessel flow, berth occupancy, equipment activity, shore-power demand, renewable supply, battery state, grid import, and modelled emissions. A policy-weighted Instability Propagation Index (IPI) classifies lower-stress, critical-transition, and instability-dominated operating regimes and is embedded in an NSGA-II rolling-horizon controller for berth allocation, crane scheduling, vessel sequencing, shore-power timing, flexible loads, renewable curtailment, and battery dispatch. The confirmatory analysis uses a 720-h coupled congestion-and-grid-limit benchmark with 30 paired common-random-number replications, service-preservation controls, terminal-horizon accounting, optimizer-seed verification, and distribution-robust statistics. Relative to operation-focused, energy-focused, and decoupled policies, the proposed controller reduces operational energy by 8.9–21.2%, modelled grid-plus-queue emissions by 8.5–21.0%, peak grid import by 5.7–15.1%, and mean replication-level maximum IPI by 24.8–42.3%, while achieving the lowest composite vessel delay and highest berth-utilization efficiency. A 23 MW renewable portfolio with an 8 MW/16 MWh battery reduces net grid import from 14,510 to 8,370 MWh, peak import from 54.9 to 46.2 MW, modelled emissions from 8,040 to 4,847 tCO₂e, and mean maximum IPI from 0.82 to 0.68. The results demonstrate coordinated logistics–energy control within an externally grounded synthetic benchmark whose principal digital-twin, equipment, and shore-power operating scales are supported by published real-terminal evidence. Deployment at a specific port would nevertheless require local calibration of arrival, service, grid, emission, and control parameters.