Plenary lectures
Abstract of the presentation: Steel exoskeletons offer a versatile and minimally invasive solution for the seismic retrofit of existing buildings. This keynote lecture presents a conceptual classification of exoskeleton systems based on three main criteria: the selected construction material, the structural arrangement—with resisting systems positioned either parallel or orthogonal to the building façades—and the typology of their constituent elements. The lecture also introduces a performance-based design methodology centred on the N2 method, as incorporated into current international seismic codes, for evaluating the coupled response of the existing building and the external retrofit system. Particular attention is devoted to the distribution of stiffness, strength and energy-dissipation capacity, as well as to the interaction between the two structural systems. The crucial role of the exoskeleton connections—both to the existing structure and to the foundations—is also examined, highlighting their influence on load transfer, global seismic performance and the feasibility of the retrofit intervention.
Abstract of the presentation: Augmented perception refers to technological enhancements to interact with digital information integrated into the real-world environment to process complex data and/or sense subtle or imperceptible information that our raw sensory systems may not be able to capture. The field is rapidly developing and is poised to transform civil engineering by extending human sensing capabilities through the integration of artificial intelligence (AI), advanced sensor networks, digital twins, robotics, extended reality (XR), edge computing, and ubiquitous connectivity. Extending beyond conventional structural health monitoring and automation, augmented perception enables infrastructure systems to continuously interpret, predict, and respond to changing environmental and operational conditions with unprecedented situational awareness. The presentation explores the emerging paradigm in which physical infrastructure is coupled with intelligent cyber-physical systems that fuse multimodal data from distributed sensors, uncrewed aerial and ground systems including satellite observations, Building Information Modeling (BIM), Geographic Information Systems (GIS), and Internet of Things (IoT) platforms. The presentation further discusses how foundation models, multimodal AI, physics-informed machine learning (PIML), federated learning, and neuromorphic computing may redefine the design, construction, operation, and maintenance of resilient and sustainable infrastructure. The presentation examines the role of augmented perception in enabling predictive maintenance, autonomous inspection, disaster resilience, adaptive construction, climate-aware infrastructure management, and human–AI collaboration. Special attention is devoted to explainable AI, cybersecurity, digital trust, interoperability, and ethical governance as essential enablers for its widespread adoption. Future developments, including self-learning infrastructure, cognitive digital twins, quantum sensing, distributed intelligence, and collaborative human–machine decision-making, are identified as key research frontiers. By presenting an integrated vision of intelligent infrastructure that perceives, reasons, and acts in real time, this presentation outlines a roadmap toward next-generation civil engineering systems that are safer, more resilient, resource-efficient, and responsive to the challenges of urbanization, aging infrastructure, and climate change.
Abstract of the presentation: HSLA steels use in heavy loaded constructions is presented and analysed, starting with the first generation (YS up to 500 MPa), continuing with the second generation (YS up to 700 MPa) and ending with the third generation (YS above 960 MPa). For the first two generations of HSLA steels case studies are used to provide more detailed insight into their behaviour, especially regarding crack resistance and related problems that occurred in their welded joints. As a representative of the third generation of HSLA steels, also called Ultra HSLA), SIMAXX 1000 (YS around 1,000 MPa) was examined under static, impact and amplitude loading. It was shown that its structural integrity is not jeopardized in any of these loading conditions, making it suitable choice for all welded constructions, including large offshore platforms, pipelines, cranes, pressure vessels and bridges.
Abstract of the presentation: The distribution of buildings within a city provides only part of the information needed to assess the consequences of an earthquake. The number of occupants, building use and structural vulnerability jointly influence the scale of potential losses. This study examines these relationships within a selected area of the Râșcani district in Chișinău, based on an inventory and visual survey of more than 700 buildings.
The building stock is characterised by structural typology, number of storeys and use, alongside the estimated number of occupants. Buildings are classified according to the RISK-UE methodology, and the data are organised in a geographic information system to examine the spatial distribution of vulnerability and population exposure. The results reveal substantial differences between the proportion of buildings belonging to each structural typology and the share of the population they accommodate. Although low-rise buildings account for approximately half of the surveyed stock, most occupants are concentrated in residential buildings of medium and greater height. Four- and five-storey masonry apartment buildings and taller reinforced concrete buildings therefore warrant particular attention when assessing potential earthquake consequences.
The analysis is supplemented by an indicative casualty estimate for an earthquake scenario originating in the Vrancea seismic zone, using predicted building damage and occupancy assumptions. The results are discussed in light of the limitations of empirical vulnerability assessment and uncertainties in building occupancy. The study provides a basis for prioritising detailed seismic assessments by considering both structural vulnerability and the number of people exposed.
Abstract of the presentation: Predicting the nonlinear mechanical response of masonry under external loading and inferring the properties and local conditions of existing structures from observed responses are challenging direct and inverse problems. These challenges arise from masonry’s vast design space, heterogeneous and anisotropic behaviour, brittle response, evolving damage, and complex boundary conditions. Experimental testing remains essential but is constrained by cost, scale, and the number of configurations that can be investigated. Numerical modelling enables virtual experimentation, yet high-fidelity nonlinear simulations are computationally demanding, require specialist expertise, and are difficult to deploy across extensive design spaces, uncertainty quantification, and real-time assessment. Inverse problems are particularly challenging because in-situ investigations often rely on sparse, local, or low-fidelity measurements. This talk presents a generative artificial intelligence framework designed to address both direct and inverse mechanics problems. Conditional generative adversarial networks (cGANs), originally developed for image-to-image translation, are employed to establish bidirectional mappings between visual representations of masonry and its mechanical behaviour. Masonry configurations, mechanical parameters, and corresponding full-field nonlinear strain maps are encoded as colour images, without explicitly providing the underlying constitutive laws. This formulation enables the model to learn complex nonlinear relationships directly from data. Despite training on relatively limited datasets, the framework demonstrates predictive capabilities within the training design space and generalises to previously unseen material properties and local heterogeneities. Furthermore, continual-learning strategies enable the model to extend its predictive capabilities into new regions of the design space using limited additional training data. Beyond masonry, this research establishes a pathway towards generative AI-based mechanical prediction and assessment of heterogeneous, brittle materials and composites. By moving beyond conventional pattern recognition towards bidirectional, mechanics-oriented generation, the framework offers potential for accelerating virtual experimentation, supporting inverse characterisation, and enabling more efficient structural assessment and design under uncertainty.
Abstract of the presentation: This presentation investigates how collaborative, technology enhanced and data-driven research, based on a comprehensive approach, can contribute to earth-quake preparedness and response.
The first part discusses the use of ICT tools and smartphone applications which deliver real-time data, to support efficient earthquake response. These tools en-hance situational awareness by delivering reliable earthquake damage assessment (buildings, landslides and liquefactions included), thus leading to efficient plan-ning of both preventive and response actions.
The second part emphasizes the vital role of public engagement-particularly in schools-in enhancing resilience. It demonstrates how scientific knowledge can be converted into actions which can promote a culture of preparedness among citi-zens while at the same time, improving their response capacity.
The final section discusses the deployment of low-cost monitoring devices in ur-ban areas, which provide continuous data streams related to both ground motion and building response, thus supporting a more precise, data-driven understanding of seismic risk.
The overall approach presented is based on the results of two successfully com-pleted, EU co-funded projects, the “Rapid Earthquake Danage Assessment con-sortium” (REDACt) and the “Earthquake Resilient Schools” (EReS), and offers a model for improving existing practices, placing data-driven information to sup-port competent authorities and the Public, at the core of earthquake risk mitigation actions.
Abstract of the presentation: The keynote speech discusses the main features of a novel Computerized Upper Bound Limit Analysis approach aimed at the accurate prediction of the ultimate limit state behavior of historical masonry structures exhbiting high geometric complexity. Particular attention is given to the estimation of the seismic vulnerability.
The approach bases on a full 3D discretization of the masonry structure, which allows plastic dissipation exclusively at the interface between contiguous elements. After the derivation of a standard linear programming problem to predict collapse multiplier, active failure mechanism and -from the dual problem- internal stress distribution at failure, an efficient iterative procedure is discussed to simulate the problem in case of interfaces purely frictional.
A series of paradigmatic case studies is reviewed, focusing in particular on building aggregates and churches. A final extension of the approach presented is proposed, which relies in a fast algorithm equipped with a surrogate optimization kernel, to automatically provide that disposition of tie rods that maximizes the load carrying capacity.
as a sustainable resource and comparative analysis among alternative solutions
Abstract of the presentation: The research focused on the experimentation of components made with natural fibres, with the aim of assessing their mechanical performance and potential fields of application. In this context, the study analyzed the physical and mechanical properties of a lime-based mortar reinforced with randomly distributed hemp fibres within the mixture, as well as an FRCM system (Fabric-Reinforced Cementitious Matrix) composed of a lime mortar combined with a hemp fiber mesh. For the fiber reinforced plaster, initial tests included fiber water absorption and workability (flow) of the fresh mixture. Subsequently, mechanical characterization was carried out through flexural and compressive strength tests on specimens. As for the meshes, tensile tests were performed on hemp yarns, followed by compressive strength tests on masonry panels (50x50 cm), both with and without FRCM reinforcement, to evaluate the static benefits provided by the proposed retrofit system. The experimental results highlighted the excellent performance of plasters reinforced with hemp fibres and meshes, confirming their effectiveness as a strengthening system for existing structures.
Other natural fibres, such as coconut and jute, also proved to be valid alternatives to hemp to produce eco-sustainable plasters and blocks. For this reason, a research activity was undertaken on mortar samples reinforced with coconut, jute, and hemp fibers, aiming to evaluate their physical and mechanical properties. Once the optimal mix design was defined for each mixture, experimental laboratory tests were conducted to characterize them. The comparison of the results allowed for the identification of the best-performing material and, consequently, the selection of the most suitable type of natural fibres for the development of the fibre-reinforced mortars investigated in this study.