Physical interpretation of machine learning-based recognition of defects for the risk management of existing bridge heritage

Engineering Failure Analysis, Год журнала: 2023, Номер 149, С. 107237 - 107237

Опубликована: Апрель 5, 2023

The challenge of the research work presented in paper is to combine growing interest monitoring health condition existing bridge heritage through systematic and periodic visual inspections with automated recognition typical defects, which can greatly facilitate assessment defect evolution over time. study focused on identification defects Reinforced Concrete (RC) bridges exploiting different Deep Learning (DL) approaches techniques interpret obtained predictions. Ensuring safety infrastructures typically a technical economic issue. Still, case engineering infrastructure heritage, there are viaducts high historical, cultural, symbolic value. For them, accurate knowledge characterization possible degradation processes become particularly important order define intervention strategies that conservation requirements. With aim develop non-invasive investigation protocols for continuous effective control their evolution, database RC images was collected, most recurrent typologies were classified by domain experts. Some Convolutional Neural Networks (CNNs) algorithms applied dataset automatically recognizing all but specific novel contribution interpretation results form humanly explainable directly implementable new tools inspections. To results, Class Activation Maps (CAMs) employed within available eXplainable Artificial Intelligence (XAI) techniques, allow observe activation zones nearly perfectly highlight type given image. besides suggesting network works better than others if effectively recognized, have been evaluated quasi-quantitative procedure compared qualitative CNNs models reliability two indexes representing explaining metrics results. In end, outcomes proposed observed also real-life study. discussion opens scenarios application these supporting road management companies public organizations evaluation networks state.

Язык: Английский

---

Artificial intelligence in structural health management of existing bridges

Automation in Construction, Год журнала: 2024, Номер 167, С. 105719 - 105719

Опубликована: Авг. 23, 2024

Язык: Английский

---

Seismic fragility and risk assessment of isolated bridges subjected to pre-existing ground displacements

Soil Dynamics and Earthquake Engineering, Год журнала: 2025, Номер 194, С. 109335 - 109335

Опубликована: Март 4, 2025

Язык: Английский

---

Advancing bridge resilience: a review of monitoring technologies for flood-prone infrastructure

Open Research Europe, Год журнала: 2025, Номер 5, С. 26 - 26

Опубликована: Янв. 27, 2025

Floods pose a critical threat to bridge infrastructure, which plays an essential role in transportation networks and economic resilience. This review examines state-of-the-art Structural Health Monitoring (SHM) technologies tailored mitigate flood risks, focusing on their real-world applications flood-prone bridges. A central feature of this is the extensive use case studies, illustrating diverse SHM methods applied globally monitor challenges such as debris accumulation, hydrodynamic forces, scour—primary causes failures. These examples provide detailed insights into like sonar-based devices, scour probes, photographic monitoring, rotation- vibration-based techniques. By showcasing specific studies—such bridges monitored using smart magnetic rocks, Interferometric Synthetic Aperture Radar (InSAR), fibre optic sensors—the highlights practical outcomes, demonstrating how systems enhance resilience through early detection predictive maintenance. It also explores implementing these systems, including environmental sensitivity, cost, data complexity, while identifying gaps integrating hydraulic structural for holistic risk assessments. advocates multidisciplinary collaboration advanced data-driven solutions, AI-based maintenance, address climate change impacts increasing risks. bridging cutting-edge research with applications, article provides actionable scalable, adaptive inspiring engineers researchers develop more resilient infrastructure changing world.

Язык: Английский

---

Advancing bridge resilience: a review of monitoring technologies for flood-prone infrastructure

Open Research Europe, Год журнала: 2025, Номер 5, С. 26 - 26

Опубликована: Март 17, 2025

Floods pose a critical threat to bridge infrastructure, which plays an essential role in transportation networks and economic resilience. This review examines state-of-the-art Structural Health Monitoring (SHM) technologies tailored mitigate flood risks, focusing on their real-world applications flood-prone bridges. A central feature of this is the extensive use case studies, illustrating diverse SHM methods applied globally monitor challenges such as debris accumulation, hydrodynamic forces, scour-primary causes failures. These examples provide detailed insights into like sonar-based devices, scour probes, photographic monitoring, rotation- vibration-based techniques. By showcasing specific studies-such bridges monitored using smart magnetic rocks, Interferometric Synthetic Aperture Radar (InSAR), fibre optic sensors-the highlights practical outcomes, demonstrating how systems enhance resilience through early detection predictive maintenance. It also explores implementing these systems, including environmental sensitivity, cost, data complexity, while identifying gaps integrating hydraulic structural for holistic risk assessments. advocates multidisciplinary collaboration advanced data-driven solutions, AI-based maintenance, address climate change impacts increasing risks. bridging cutting-edge research with applications, article provides actionable scalable, adaptive inspiring engineers researchers develop more resilient infrastructure changing world.

Язык: Английский

---

Bridge-specific flood risk assessment of transport networks using GIS and remotely sensed data

The Science of The Total Environment, Год журнала: 2022, Номер 850, С. 157976 - 157976

Опубликована: Авг. 11, 2022

A novel framework for the expedient assessment of flood risk to transportation networks focused on response most critical and vulnerable infrastructure assets, bridges, is developed, validated applied. Building upon recent French guidelines scour (CEREMA, 2019), this paper delivers a thorough methodology, that incorporates three key, parameters: (i) hydrodynamic loading, hazard component equal significance scour, hazard; (ii) correlation select indicators with new index relating flow velocity, primary measure adverse impacts flow–structure interaction, enabling more accurate automated, bridge susceptibility scour; (iii) use new, comprehensive indicator, namely Indicator Flood Hazard Intensity (IFHI) which incorporates, in simple yet efficient way, key parameters controlling severity impact floodwater height, obstruction, sediment type. The implemented analysis case study area, considering an inventory 117 bridges diverse construction characteristics, were affected by major impacted Greece September 2020. reliability method against extensive record inspected documented damages. Regional scale facilitated adoption Multi-Criteria Decision-Making indexing, geomorphological, meteorological, hydrological, land use/cover data, based processing remotely sensed imagery openly available geospatial datasets GIS.

Язык: Английский

---

Probabilistic-based assessment of subsidence phenomena on the existing built heritage by combining MTInSAR data and UAV photogrammetry

Structure and Infrastructure Engineering, Год журнала: 2024, Номер unknown, С. 1 - 16

Опубликована: Ноя. 4, 2024

Among the natural risks to which structures and infrastructures are subjected, action of geohazards is often ignored. Several efforts required predict geological effects on existing built heritage, such as by employing sensor-based systems or performing periodical visual inspections. Alternatively, novel cost-effective techniques could be used. The paper presents a probabilistic-based approach evaluate occurrence subsidence phenomena affecting infrastructures, combining information provided multitemporal interferometry via synthetic aperture radar (MTInSAR) data unmanned aerial vehicle (UAV) photogrammetry. Given structure monitor period observation, idea consists retrieving MTInSAR about UAV flight surveys surrounding area at beginning end considered period. From both surveys, statistical distributions spatiotemporal velocities can processed evaluated toward predefined limit state. obtained results classified under different scenarios, reveal possible phenomena. proposed methodology was firstly validated real landslide (the vertical component assessed) and, subsequently, tested another case for risk prediction, showing satisfying capacity providing warnings employ mitigation plans.

Язык: Английский

---

Study on scour simulation and boundary condition conversion technology for a shallow foundation bridge

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Фев. 7, 2025

Bridges with shallow foundations are highly susceptible to flood scouring due their limited embedment depth and small contact area between the soil foundation. This can lead foundation voids, posing a serious threat bridge safety. To prevent mitigate risks, this paper investigates riverbed characteristics of bridges under different hydrological conditions.The study found that high water levels flow velocities, scour significantly increased.Under extreme conditions, horseshoe vortex forms at base front end pier, causing pits on both sides upstream face foundation, which is main cause voids first appear 2580 s maximum -2.51 m void 0.5%, continuing increase over time.Based simulated data, proposes method for converting boundary conditions from model mechanical model. utilizes point cloud reverse engineering technology generate surface three-dimensional coordinate matrix import it into structural analysis field. Hydraulic effects calculated using CFD transferred domain through fluid-structure interaction technology, achieving multi-physical field coupling among flow, soil, structure. addresses current limitations in simulating complex damage research, providing reliable technical support subsequent studies behavior conditions.

Язык: Английский

---

Data-driven lightning-related failure risk prediction of overhead contact lines based on Bayesian network with spatiotemporal fragility model

Reliability Engineering & System Safety, Год журнала: 2022, Номер 231, С. 109016 - 109016

Опубликована: Ноя. 29, 2022

Язык: Английский

---

Seismic risk-informed prioritisation of multi-span RC girder bridges considering knowledge-based uncertainty

Bulletin of Earthquake Engineering, Год журнала: 2023, Номер 22(2), С. 693 - 729

Опубликована: Окт. 21, 2023

Abstract In earthquake-prone countries, transport network managers need to perform extensive seismic risk assessments coping with a considerable number of bridges characterised by an unsatisfying knowledge level and designed in the past without anti-seismic requirements. This study proposes framework for efficient assessment multi-span girder considering knowledge-based uncertainties. The is intended be applied risk-informed prioritisation bridge portfolios. It based on subsequent modules that involve input data, simulation uncertainties, simplified analysis, fragility loss assessment. vulnerability given represented ratio percentiles related intensity measure which can used quantify expected annual losses corresponding variability due influence uncertainty. A case-study section demonstrates widespread category simply supported girder-reinforced concrete bridges. addresses issues such as use optimal measures, required model realisations discrepancies respect accurate nonlinear time-history analysis. Finally, illustrative example proposed eight Southern Italy its applicability critical directing in-depth data collections where needed.

Язык: Английский

---