Adaptive baseline model for autonomous marine equipment and systems

Area of Science:

• Marine engineering
• Autonomous systems
• Data science

Background:

• The Internet of Things (IoT) and Fourth Industrial Revolution are driving automation in marine equipment.
• Current methods for evaluating marine equipment status struggle with dynamic conditions and static baselines.
• Accurate performance baselines are crucial for autonomous shipping systems.

Purpose of the Study:

• To develop a data-driven method for establishing dynamic performance baselines for marine equipment.
• To address the limitations of static baselines in dynamic operating environments.
• To improve the accuracy and adaptability of status evaluation for autonomous marine systems.

Main Methods:

• Proposed a reference-site (R-S) model to establish an initial baseline, addressing inadequate initial parameters.
• Introduced a dynamic kernel (D-K) model to increase reference sites and update points, reducing data calculation for dynamic baseline updates.
• Determined sliding window capacity using the Kolmogorov-Smirnov method for model implementation.

Main Results:

• The proposed data-driven method successfully established a dynamic performance baseline for marine equipment.
• The method demonstrated improved accuracy and adaptive performance compared to traditional approaches.
• Application to a marine diesel engine's exhaust temperature showed the model's effectiveness.

Conclusions:

• The developed data-driven baseline model effectively adapts to dynamic working conditions in marine environments.
• This approach enhances the reliability of status evaluation for automated and autonomous marine equipment.
• The R-S and D-K models offer a robust solution for creating adaptive performance baselines in maritime applications.