Digital twin for virtual sensing of ferry quays via a Gaussian Process Latent Force Model Journal: Data/Math Publication type: ARTICLE TYPE
Luigi Sibille, Torodd Skjerve Nord, Alice Cicirello · 2025
Read Digital twin for virtual sensing of ferry quays via a Gaussian Process Latent Force Model Journal: Data/Math Publication type: ARTICLE TYPE by Luigi Sibille, Torodd Skjerve Nord, Alice Cicirello (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-14925.txt.
Introduction
Ferry quays take a beating from weather and from ferries docking, and damage can grow unseen. Luigi Sibille, Torodd Skjerve Nord, and Alice Cicirello built a digital twin of a Norwegian ferry quay in 2025. A digital twin mirrors a real structure, but sensors cannot cover every critical spot, so the twin must estimate motion at places with no sensor, a task called virtual sensing.
Their model pairs a reduced physics model of the quay with a data-driven force model that infers the unknown impact forces. It gave accurate motion estimates at most locations during a real ferry impact, though spots in the impact zone were harder. The authors also study where extra sensors would help most.
Catalog record
| CREATOR | Luigi Sibille, Torodd Skjerve Nord, Alice Cicirello |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 95 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-14925
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Digital twin for virtual sensing of ferry quays via a Gaussian Process Latent Force Model Journal: Data/Math Publication type: ARTICLE TYPE Luigi Sibille, Torodd Skjerve Nord, Alice Cicirello Abstract Ferry quays experience rapid deterioration due to their exposure to harsh maritime environments and operational ferry impacts. Vibration-based structural health monitoring offers a valuable approach to assessing structural integrity and enhancing the understanding of the structural implications of these impacts. However, practical limitations often restrict sensor placement at critical locations. Consequently, virtual sensing techniques become essential for establishing a Digital Twin and estimating the structural response. This study investigates the application of the Gaussian Process Latent Force Model (GPLFM) for virtual sensing on the Magerholm ferry quay, combining in-operation experimental data collected during a ferry impact with a detailed physics-based model. The proposed Physics-Encoded Machine Learning model integrates a reduced-order structural model with a data-driven GPLFM representing the unknown impact forces via their modal contributions. Significant challenges are addressed for the development of the Digital Twin of the ferry quay, including unknown impact characteristics (location, direction, intensity), time-varying boundary conditions, and sparse sensor configurations. Results show that the GPLFM provides accurate acceleration response estimates at most locations, even under simplifying modeling assumptions such as linear time-invariant behavior during the impact phase. Lower accuracy was observed at locations in the impact zone. A numerical study was conducted to explore an optimal real-world sensor placement strategy using a Backward Sequential Sensor Placement approach. Sensitivity analyses were conducted to examine the influence of sensor types, sampling frequencies, and incorrectly assumed damping ratios. The results suggest that the GP latent forces can help accommodate modeling and measurement uncertainties, maintaining acceptable estimation accuracy across scenarios. keywords Digital Twin, Physics-Encoded Machine Learning, Structural Health Monitoring, Virtual Sensing, Gaussian Process Latent Force Model, Ferry Quay, Impact 2025 {Frontmatter} \authormark Luigi Sibille et al . \authormark Luigi Sibille et al. Impact Statement Ferry quays operate under severe environmental and operational conditions that often lead to deterioration before their intended design life. Accurate measurement of structural response is essential to improve understanding of their dynamic behavior and inform maintenance strategies. However, physical constraints, such as water exposure, vehicle traffic, and ferry impacts, prevent sensor placement at many critical locations. This study addresses these limitations by developing a Digital Twin framework for virtual sensing, integrating physics-based modeling with data-driven methods. The work highlights the modeling challenges specific to ferry quay infrastructure and provides a practical approach to estimate structural responses where instrumentation is not feasible. The findings support the development of more effective structural health monitoring practices for engineers, researchers, and infrastructure operators. 1 Introduction Ferry quays play an essential role in the transportation networks of coastal and island regions worldwide. This infrastructure facilitates the transport of people, vehicles, and goods where the construction of traditional structures, such as bridges and tunnels, is not feasible. Particularly in Norway, these quays are critical, supporting over 130 routes and the transportation of approximately 40 million people annually. This usage is expected to increase with the growth in population, vehicle ownership, and tourism. Operational loads, particularly the impacts exerted by ferries during docking, are acknowledged as significant contributors to the rapid degradation of these quays ( Siedziako et al. 2023 ) . These impacts may lead to brittle material failures, the initiation of fatigue cracks, buckling of elements, and the failure of joints, all of which compromise the reliability and integrity of the infrastructure ( Siedziako et al. 2023 ) . The structural responses during such events are not fully understood, despite their crucial importance for applications such as fatigue analysis, system identification, damage detection, and predictive maintenance, which are essential for extending the service life of these structures. Environmental factors, including saltwater corrosion, strong waves, and wind, exacerbate these effects, accelerating the degradation process and potentially reducing the lifespan of these structures ( Xia et al. 2019 ; Wall & Wadsö 2013 ) . A notable incident exemplifying these vulnerabilities occurred during the storm Ingunn in 2024, which led to a failure of the linkspan at the Magerholm ferry quay shortly after the ferry undocked. The specific causes of this failure are yet to be determined, highlighting the importance of investigating how operational and environmental factors impact ferry quay safety. Ferry quays exhibit significant time-varying behavior driven by operational variability, such as interactions with ferries, vehicles, and environmental factors. The impact’s intensities and characteristics for over 1500 impacts were recently assessed by means of temporal moments ( Sibille et al. 2025 ) . Although this approach captured the spatial and temporal characteristics of impact events, it circumvented the challenge of modeling the system’s dynamic behavior. Environmental conditions, particularly tidal levels, have been demonstrated to influence the modal properties, highlighting the time-variant nature of these structures ( Sibille et al. 2025a ) . Nonetheless, these studies do not address the challenges of modeling the complex, time-dependent system under real impact events, nor do they estimate structural responses at unmeasured or inaccessible locations. Digital Twins (DTs) of dynamical systems have gained particular attention in recent years for understanding and predicting the behavior of complex engineering systems and critical infrastructure in operation ( Ye et al. 2019 ) . A DT is a time-evolving digital model based on a two-way interaction between the physical system (usually monitored) and the digital counterpart, that can account for uncertainty, different systems interactions, and can be used for policy and decision making ( Tao et al. 2022 ) . Consequently, the development of a DT based on integrating domain and expert knowledge with physics-based models, and with real-world data (often referred to as observations) obtained through Structural Health Monitoring (SHM), has emerged as a critical tool for guiding high-consequence decision making on complex engineering systems and critical infrastructure ( Haywood-Alexander et al. 2024 ) . Ferry quays exhibit complex and time-variant behavior, making the development of a DT particularly beneficial, as it provides engineers and decision-makers with an up-to-date understanding of the structure’s actual condition and dynamic response. Unlike traditional Finite Element (FE) models, which do not account for temporal variability, model uncertainty, or incomplete knowledge of real boundary conditions, DTs provide a model that evolves in parallel with the physical system, incorporating real-world data from sensors and monitoring systems to continuously reflect the structure’s current state. To effectively build such a responsive and adaptive DT, particularly in environments with limited sensor coverage and complex loading conditions like ferry quays, Physics-Enhanced Machine Learning (PEML) strategies offer a powerful solution ( Cicirello 2024 ) . PEML strategies (also referred to as hybrid physics-data models, grey-box modeling, or scientific machine learning ( Cross et al. 2024 ) ) enable the development of robust and interpretable models based on the integration of physics-based models, knowledge, and real-world data (even if limited, sparse, and noisy). Traditional SHM systems, through sensor networks, provide the measurement data that informs and continuously updates the DT ( Farrar & Worden 2010 ) . However, the placement of sensors in areas subjected to impacts or traffic often presents practical challenges. Operational constraints and the risk of sensor damage in such environments frequently preclude their installation in critical locations. This scenario emphasizes the advantages of developing a DT for virtual