Predictive Maintenance Study, EUROGATE Container Terminal Limassol, Cyprus

Limassol is a medium-sized multi-purpose port on the south coast of Cyprus. As part of a European Union funded research project (aerOS, 2022 to 2025), researchers and the terminal operator, EUROGATE Container Terminal Limassol, developed an Internet of Things (IoT) and machine-learning framework to predict faults in CHE, focusing on four electric straddle carriers in the monitored architecture; the inverter dataset analyzed in the paper draws on telemetry from 15 straddle carriers, with six inverter incidents. Multiple sensors were installed on hydraulic systems, inverters and other critical components to capture multi-modal time-series data such as vibration, inclination, temperature, speed, torque and hydraulic pressure. The data was sent from IoT gateways on each machine using the MQTT protocol, pre-processed, and used to train several predictive models such as artificial neural networks, decision trees, random forests, XGBoost and Gaussian Naive Bayes to detect inverter over-temperature faults; a separate statistical model checked the health of the hydraulic system. (Aslam et al., 2025)

The best-performing model, based on artificial neural networks, achieved around 98.7% accuracy and a 98.0% F1-score in predicting inverter over-temperature faults linked to fan failures, clogged filters and similar issues. (Aslam et al., 2025)

The project was funded under a European research grant (Grant Agreement 101069732), so specific capital costs to the terminal are not published, . The authors note that unplanned maintenance of port equipment can lead to operational disruptions, including unexpected delays and long waiting times. The authors report that the hydraulic anomaly model reached 83.3% precision and that further testing is needed, since the dataset contained few anomalies. (Aslam et al., 2025)

The authors note that model choice must fit the computational limits of the edge infrastructure, which favors lightweight models. (Aslam et al., 2025)

Overall, the Limassol case shows that combining IoT sensing with machine-learning analytics can deliver high-accuracy fault prediction for port equipment, although the study did not measure operational, cost or environmental benefits. (Aslam et al., 2025)