THE NEED FOR LARGE AND HIGH-QUALITY DATA VOLUMES FOR AI TRAINING: A REVIEW OF MARINE SCADA SYSTEMS AS PRIMARY DATA SOURCES FOR PREDICTIVE MAINTENANCE
DOI:
https://doi.org/10.32782/mathematical-modelling/2025-8-1-19Keywords:
artificial intelligence, machine learning, marine SCADA systems, predictive maintenance, data analysis, maritime sector, shipboard electrical equipmentAbstract
This article explores the significance and necessity of large, high-quality datasets for the effective application of artificial intelligence (AI) in the maritime sector, particularly for predicting the technical condition of shipboard electrical equipment. The emphasis is placed on analyzing Supervisory Control and Data Acquisition (SCADA) systems to review it as the main data sources for training AI algorithms. The use of AI is increasing its criticality for enhancing operational efficiency, environmental sustainability, and safety in maritime shipping. The paper addresses the primary challenges associated with collection and processing of large data volumes, including the problem of data imbalance, where instances of equipment failure occur much less frequently than normal operational data. Special attention is given to the integration of machine learning algorithms with existing SCADA systems, the challenges of data standardization and quality. Examples of success- ful implementations of predictive maintenance leveraging AI algorithms are demonstrating the potential to significantly reduce operational costs and the risks of unexpected equipment failures. There are financial aspects highlighting substantial economic benefits from implementing advanced predictive maintenance systems. Additionally, specific cases of AI algorithm applications for predicting failures and detecting anomalies in shipboard equipment are analyzed, significantly enhancing the reliability and safety of vessels. Particular attention is given to cybersecurity issues essential for protecting data used in the training and operational phases of AI algorithms. Furthermore, promising directions for future research are outlined, including the development of new algorithms, adaptation of existing data analysis methods to maritime-specific conditions, and a systematic approach to ensuring cybersecurity in shipboard information systems. The conclusions emphasize the necessity of addressing these key tasks for effective and sustainable advancement in contemporary maritime shipping.
References
Pereira M.T., Rocha N., Silva F.G., Moreira M.Â.L., Altinkaya Y.O., & Pereira M.J. Process Optimization in Sea Ports: Integrating Sustainability and Efficiency Through a Novel Mathematical Model. Journal of Marine Science and Engineering. 2025. № 13 (1). P. 119.
Using Optimization in the container shipping industry. Sia Partners Insights. URL: https://www.sia-partners.com/en/insights/publications/using-optimization-container-shipping-industry (date of access: 29.03.2025).
CMA CGM Group announces $20 billion investment in the Americas for maritime transportation and logistics. CMA CGM Group. URL: https://www.cmacgm-group.com/en/news-media/cma-cgm-group-announces-20-billion-investment-americas-maritime-transportation-logistics (date of access: 29.03.2025).
Durlik I., Miller T., Kostecka E., & Tuński T. Artificial Intelligence in Maritime Transportation: A Comprehensive Review of Safety and Risk Management Applications. Applied Sciences. 2024. № 14 (18). P. 8420.
The importance of high-quality training data for building machine learning and deep learning models. ADATA.PRO. URL: https://adata.pro/blog/the-importance-of-high-quality-training- data-for-building-machine-learning-and-deep-learning-models/ (date of access: 29.03.2025).
The true cost of ship downtime: data-driven insights for shipowners. Ship Universe. URL: https://www.shipuniverse.com/the-true-cost-of-ship-downtime-data-driven-insights-for-shipowners/ (date of access: 29.03.2025).
Basurko O., Uriondo Z. Condition-Based Maintenance for medium speed diesel engines used in vessels in operation. Applied Thermal Engineering. 2015. № 80. P. 404–412.
Macdonald F. The Use of AI in Condition-Based Maintenance and Autonomous Navigation. Thetius. URL: https://thetius.com/the-use-of-ai-in-condition-based-maintenance-and- autonomous-navigation/ (date of access: 29.03.2025).
AI-based condition monitoring system for the ship engines of a Shipbuilding company. Infosys. URL: from https://www.infosys.com/services/engineering-services/service-offerings/internet- of-things/case-studies/shipbuilding-company.html (date of access: 29.03.2025).
Taha M. Integrating Artificial Intelligence with SCADA Systems: Enhancing Operational Efficiency, Predictive Maintenance, and Environmental Sustainability. ResearchGate Technical Report. 2024. URL: https://www.researchgate.net/publication/382537416_Title_Integrating_Artificial_Intelligence_with_SCADA_Systems_Enhancing_Operational_Efficiency_Predictive_Maintenance_and_Environmental_Sustainability.
Suryadarma E.H.E., & Ai T.J. Predictive Maintenance in SCADA-Based Industries: A literature review. International Journal of Industrial Engineering and Engineering Management. 2020. № 2 (1). P. 57–70.
Anish W. What is marine electricity and how it is generated? Marine Insight. URL: https://www. marineinsight.com/marine-electrical/what-is-marine-electricity/ (date of access: 29.03.2025).
Knutsen K.E., Manno G., Vartdal B. Beyond condition monitoring in the maritime industry. Strategic research and innovation. 2014. URL: https://www.researchgate.net/publication/263583976_Beyond_condition_monitoring_in_the_maritime_industry.
Predictive maintenance and the smart factory. Deloitte. URL: https://www2.deloitte.com/content/dam/Deloitte/us/Documents/process-and-operations/us-cons-predictive-maintenance.pdf (date of access: 29.03.2025).
Ajaykumar P. Ship hull inspection services market to reach new heights by 2032. Vocal Media. URL: https://vocal.media/journal/ship-hull-inspection-services-market-to-reach-new-heights-by-2032 (date of access: 29.03.2025).
