Enhancing Structural Health Monitoring of Ferry Quays through Virtual Sensing and Force Identification
December 2024
In this Smart Infrastructure Blog, Luigi Sibille, PhD Candidate from the Department of Ocean Operations and Civil Engineering at the Norwegian University of Science and Technology (NTNU) and former CSIC research visitor, argues that enhancing structural health monitoring of ferry quays through virtual sensing and force identification is crucial for understanding the forces at play, which in turn is vital for designing more durable structures and developing effective maintenance strategies.
By combining the power of physics and data-driven modelling, the GPLFM represents a transformative step towards smarter, safer, and more sustainable infrastructure. (…) Understanding the forces at play is crucial for designing more durable structures and developing effective maintenance strategies. Luigi Sibille, Department of Ocean Operations and Civil Engineering at the Norwegian University of Science and Technology (NTNU)
Ferry quays are lifelines for coastal communities, providing crucial access to goods, healthcare, and transportation. Despite their importance, little research has been devoted to understanding how these structures respond to the frequent impacts of docking ferries. These impacts are a significant source of wear and tear, accelerating structural degradation over time. While the study “Finite Element Model Updating of a Ferry Dock Bridge” investigated the force identification of ferry quays, we now aim to build on this work using a more advanced methodology. The approach involves employing a physics-enhanced machine learning technique to more accurately assess the forces involved and monitor the structural responses. This innovative method promises to improve the maintenance and longevity of these crucial infrastructures.
The Role of Vibrational Structural Health Monitoring
Structural Health Monitoring (SHM) functions similarly to routine health check-ups for infrastructure. By analysing vibrations and other structural responses, engineers can identify potential issues and monitor the ongoing condition of structures. In recent decades, vibrational SHM has become crucial for providing real-time, detailed insights into the health of various infrastructures, including bridges and buildings. For ferry quays, which undergo repeated docking impacts, SHM is indispensable for comprehending these forces and mitigating damage before it escalates to critical levels.
What are Virtual Sensing and Force Identification?
Rather than relying solely on physical sensors, virtual sensing uses a combination of mathematical models and limited physical data to estimate responses in unmeasured locations, as discussed in “Virtual Sensing of Structural Vibrations Using Dynamic Substructuring”.
Complementing this, force identification pinpoints the specific forces acting on the structure, providing crucial insights into the dynamics of stress and impact. This dual approach is particularly advantageous when it is impractical or costly to install sensors across an entire structure. By applying these techniques, engineers can gain a more comprehensive understanding of a structure’s behaviour while keeping costs and disruptions to a minimum.
The Gaussian Process Latent Force Model (GPLFM)
The research centres on a novel machine learning technique called the Gaussian Process Latent Force Model (GPLFM). This approach bridges the gap between traditional physics-based models and modern data-driven methods, offering improved accuracy and reliability. Simply put, GPLFM combines the known principles of physics with a machine learning framework to predict both the forces acting on a structure and its responses. Here’s how it works:
- Data Collection: Researchers gathered data from six sensors on a working ferry quay. These included four accelerometers to measure vibrations and two linear variable differential transformers (LVDTs) to track movement.
- Finite Element Model (FEM): A detailed computational model of the ferry quay was created, simulating its structural behaviour under various forces.
- Advanced Processing: Sophisticated algorithms, including a Kalman filter and a Rauch-Tung-Striebel smoother, processed the sensor data. These tools work within an augmented state-space model, enabling simultaneous estimation of the ferry impact force and the structural response.
Validation and Accuracy
Validation is a key step in ensuring the reliability of any new method. In this study, the researchers tested their approach in two critical ways:
- Cross-Sensor Validation: They estimated the acceleration response at a location not included in the main model and compared it with data recorded by a separate accelerometer placed there. The close match confirmed the method’s high accuracy.
- Force Validation: Using simulations from the FEM model, the team compared the predicted forces with expected structural behaviour. The strong correlation between simulated and observed results further validated the reliability of the GPLFM.
Applications Beyond Ferry Quays
While this study focuses on ferry quays, the versatility of the GPLFM makes it applicable to a wide range of structural health monitoring challenges. For instance, the method has previously been used to estimate strain in offshore wind turbines and measure friction forces in experimental setups. Its ability to deliver precise results with minimal sensor data makes it an ideal solution for various engineering contexts, from energy infrastructure to building maintenance.
Why This Matters for Ferry Quays
Ferry quays face unique challenges, enduring frequent and often intense docking impacts alongside environmental factors like waves and tides. Understanding the forces at play is crucial for designing more durable structures and developing effective maintenance strategies. The GPLFM provides actionable insights into the stresses these structures endure, enabling engineers to predict and prevent potential failures. This not only enhances the safety and reliability of ferry operations but also reduces maintenance costs over time.
Looking to the Future
As technology evolves, the integration of machine learning, virtual sensing, and force identification into structural health monitoring is set to become standard practice. For ferry quays and similar infrastructure, this means a future where damage can be anticipated and addressed proactively. By combining the power of physics and data-driven modelling, the GPLFM represents a transformative step towards smarter, safer, and more sustainable infrastructure.
In conclusion, this study highlights how innovative approaches like the GPLFM are reshaping the field of structural health monitoring. By providing precise and reliable insights into the forces and responses affecting ferry quays, it paves the way for better-informed engineering decisions and longer-lasting infrastructure that coastal communities can depend on.
The text in this work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Images, including our videos, are Copyright ©University of Cambridge and licensors/contributors as identified. All rights reserved. We make our image and video content available in a number of ways that permit your use and sharing of our content under their respective Terms. For any intended reuse please contact: csic-comms@eng.cam.ac.uk