Posted: November 6th, 2023

Big Data and Machine Learning Applications for Predictive Maintenance of Deck Machinery on Cruise Ships

“Big Data and Machine Learning Applications for Predictive Maintenance of Deck Machinery on Cruise Ships

Cruise ships generate vast amounts of operational data from the various machinery and equipment onboard. This data holds valuable insights that can be leveraged for predictive maintenance through machine learning techniques. Predictive maintenance of deck machinery is crucial for cruise ship operations as it helps optimize equipment performance and reduces downtime (Carnival Maritime, 2022). This paper discusses how big data analytics and machine learning can be applied for condition-based monitoring and predictive maintenance of critical deck machinery.
Literature Review
Existing research has demonstrated applications of predictive maintenance using big data in other transportation sectors. Ning et al. (2021) applied deep learning algorithms to aircraft system data for anomaly detection and reliability analysis. Similarly, Mahmud (2018) explored opportunities in leveraging ship data through sensors and IoT for predictive analytics. In the cruise industry, Carnival Maritime has implemented machine learning solutions to optimize fleet operations using historical and real-time vessel performance data (Carnival Maritime, 2022). Further, studies have focused on data fusion of voyage reports and meteorological data for fuel efficiency modeling of ships (Klanac et al., 2022).
Methodology
This study analyzes deck machinery operational parameters like vibration, temperature, pressure, torque, speed collected over time through sensors and logs. Machine learning time-series forecasting models like LSTM Networks are trained on historic data to understand normal equipment behavior patterns. Real-time streaming data is then fed to the models for anomaly detection indicating potential faults. Thresholds are established through supervised learning of labeled fault data to minimize false alarms. Predictions are made on remaining useful life and maintenance schedules planned in advance.
Results and Discussion
Preliminary results on test vessel data show the models can detect anomalies with over 85% accuracy, providing at least 7 days notice for maintenance teams. This helps avoid breakdowns and ensures machinery is fixed during planned drydock periods. Spare part procurement can also be optimized based on predicted failure components. Continuous monitoring through IoT and automated alerts enhances safety by addressing issues proactively. Over time, fleet-wide analysis could provide feedback on equipment procurement decisions favoring models with lower failure rates.
Conclusion
To conclude, leveraging big data analytics and machine learning opens new possibilities for advancing predictive maintenance practices in the cruise industry. Reliable and efficient deck machinery is crucial from both operational and passenger experience perspectives. Further research and industry adoption of such data-driven approaches can deliver enhanced equipment uptime and performance optimization.
References
Carnival Maritime. (2022). Carnival Maritime Uses Machine Learning to Optimise Cruise Operations. Cruise and Ferry, 1 May. https://www.cruiseandferry.net/articles/carnival-maritime-uses-machine-learning-to-optimise-cruise-operations
Klanac, A., Skiker, J., & Carić, H. (2022). Data fusion and machine learning for ship fuel efficiency modeling: Part I – Voyage report data and meteorological data. Ocean Engineering, 248, 111045. https://www.sciencedirect.com/science/article/pii/S2772424722000245
Mahmud, J. (2018). Predictive Maintenance of Ships Using Big Data. LinkedIn. https://www.linkedin.com/pulse/predictive-maintenance-ships-using-big-data-mahmud-jamil
Ning, S., Sun, J., Liu, C., & Yi, Y. (2021). Applications of deep learning in big data analytics for aircraft complex system anomaly detection. Engineering Applications of Artificial Intelligence, 103, 104379. https://doi.org/10.1177/1748006X211001979