In a significant stride towards enhancing maritime vessel maintenance, researchers have developed an intelligent system aimed at bolstering the predictive maintenance capabilities of the Colombian Navy. This innovative project, dubbed the “Colombian Integrated Platform Supervision and Control System” (SISCP-C), is designed to extend the lifespan of naval vessels, increase their operational availability, and optimize life cycle costs. The lead author of this research, Carlos E. Pardo B. from the Department of Electrical and Electronics Engineering at Universidad del Norte in Barranquilla, Colombia, explains that the system provides crucial information to crew members, enabling them to make informed decisions for more intelligent and efficient maintenance strategies.
The challenge at hand is the limited availability of normal operating data, a common hurdle in maritime environments. To overcome this, the researchers employed synthetic data generation techniques, including tabular variational autoencoders, conditional tabular generative adversarial networks, and Gaussian copulas. Among these, tabular variational autoencoders proved to be the most effective. These autoencoders were used to generate synthetic datasets under normal conditions for the Wärtsilä 6L26 diesel engine. These datasets were then used to train several unsupervised anomaly detection models, including one-class support vector machines, classical autoencoders, and long short-term memory-based autoencoders. The long short-term memory (LSTM) autoencoders emerged as the top performers in terms of detection metrics.
The research highlights that dedicated multivariate LSTM autoencoders were subsequently trained for each engine subsystem. By calculating the mean absolute error of the reconstructions, a subsystem-specific health index was computed, which is used to estimate the remaining useful life of the components. This approach not only enhances the predictive maintenance capabilities but also ensures the sustainability of the vessels over time, aligning with the Colombian Navy’s strategic direction established in the Naval Development Plan 2042.
The commercial impacts of this research are substantial. For maritime sectors, the ability to predict the remaining useful life of critical components and detect anomalies early can lead to significant cost savings. By optimizing maintenance schedules and reducing downtime, shipping companies and naval operations can enhance their operational efficiency and extend the lifespan of their vessels. This technology also opens up opportunities for maritime tech startups and established players to develop advanced predictive maintenance solutions tailored to the unique challenges of the maritime industry.
As Pardo B. notes, “The system provides useful information to the crew, enabling informed decision-making for intelligent and efficient maintenance strategies.” This insight underscores the practical benefits of the research, which was recently published in the journal ‘Systems’ (translated from Spanish).
In summary, this research represents a significant advancement in the field of predictive maintenance for maritime vessels. By leveraging synthetic data generation and advanced anomaly detection techniques, the Colombian Navy and other maritime sectors can look forward to more efficient and cost-effective maintenance strategies, ultimately enhancing the sustainability and operational availability of their fleets.