Session: 08-05-01 AI‑Enabled Marine Sensing, Monitoring & Digital Twins

Submission Number: 181089

Data-Driven Sensor Placement for Marine Structures Using Unsupervised Learning Methods 

This paper presents a data-driven approach for optimal sensor placement in marine structures using machine learning methods. The strategic placement of sensors is important in structural health monitoring, as it ensures that the limited measurement data can adequately represent the global structural behavior with sufficient accuracy. Optimal sensor placement in marine structures has commonly been addressed through optimization-based methods, heuristic criteria, or data-driven algorithms. While optimization algorithms can be computationally demanding and heuristic criteria typically provide near-optimal solutions, machine learning algorithms have shown promising results in capturing complex relationships among measurement points. This study aims to apply manifold learning methods to sensor placement by leveraging their ability to capture critical features in high dimensional datasets while reducing computational time necessities. Both linear and nonlinear techniques like Principal Component Analysis (PCA) and Multidimensional Scaling (MDS) are employed to select representative sensor locations. The methods are applied to plate and truss models, which provide suitable representations of common structural components in ships and offshore platforms. A comparison of the results obtained from linear and nonlinear methods enables the identification of effective techniques for the strategic placement of sensors. This application will provide a computationally efficient solution to enhance sensor placement and structural health monitoring in marine applications.

Presenting Author: Burcin Senyasa Istanbul Technical University

Presenting Author Biography: Burcin Senyasa is a PhD candidate and research assistant in the Department of Shipbuilding and Ocean Engineering at Istanbul Technical University. Her research focuses on data-driven optimal sensor placement methods for marine structures, combining structural analysis and machine learning techniques. She received her BSc and MSc degrees from the same department at Istanbul Technical University.

Authors:

Burcin Senyasa Istanbul Technical University
Serdar Aytekin Koroglu Istanbul Technical University