KEYNOTE SPEAKERS
From Dynamic Testing to Smart Dynamic Monitoring of Civil Structures
ABSTRACT
The purpose of this lecture is to briefly characterize three decades of research and innovation of the Laboratory of Vibrations and Structural Monitoring of FEUP in the fields of dynamic testing and vibration based structural health monitoring, taking profit from a wide variety of real-world applications.
In a first instance, the topic of dynamic testing and modal identification of bridges and special structures is addressed, stressing the accuracy and usefulness provided by advanced instrumentation and state-of-art operational modal analysis tools along the structures life-cycle (design, construction, commissioning and service phases).
Subsequently, the evolution from dynamic testing to smart dynamic monitoring is also characterized based on the development of long-term dynamic monitoring programs in large Civil structures with different typologies (e.g. roadway, railway and pedestrian bridges, stadia suspension roofs, concrete dams, high voltage transmission lines and onshore and offshore wind turbines).
These monitoring applications show the efficiency of the developed tools and the usefulness of the monitoring programs implemented, enabling the achievement of different objectives, such as: (i) the development of finite element model correlations and updating; (ii) the vibration serviceability safety checking, particularly in case of lively bridges involving the inclusion of vibration control devices; (iii) the implementation of automated versions of the most powerful methods of operational modal analysis, and their application for tracking the time evolution of modal parameters in long-term dynamic monitoring applications; (iv) the use of statistical methods to remove the influence of environmental and operational factors (e.g. temperature, intensity of traffic, wind) on the modal variability, supporting the development of reliable techniques for vibration based damage detection; (v) the experimental assessment of fatigue, based on the measurement of effects of real traffic loads; (vi) the experimental assessment of aerodynamic problems in bridges based on in-situ measurements; (vii) the characterization of the influence of the water level in the reservoir on the dynamic properties of concrete arch dams or (viii) the tracking of modal parameters in onshore and offshore wind turbines, enabling damage detection and fatigue assessment.
Hybrid physics-knowledge-data approaches for monitoring engineering structures
ABSTRACT
This talk will focus on describing challenges and opportunities for monitoring engineering structures by using hybrid physics-knowledge-data models, and the need of explainable and interpretable machine learning strategies for guiding decision making in engineering applications.
The challenges that make the integration of physics-data-knowledge a very exciting field of research with the potential of providing explainable, interpretable, fast and accurate solutions that can extrapolate to unseen conditions in the data, inform physics, and identify key measurements/experiments to be carried out are going to be detailed.
State-of-the-art hybrid physics-knowledge-data approaches (also known as Physics-Enhanced Machine Learning strategies) developed within the Data, Vibration and Uncertainty group are going to be summarised, and real-world engineering applications explored with outstanding collaborators around the world, spanning from laboratory setups to bridges, wind turbines, quay walls and ferry quays are going to be showcased.
From Vibrations to Structural Intelligence: Bridging Modal Analysis, Artificial Intelligence, and Digital Twins
ABSTRACT
Vibration-based monitoring has become an essential tool for characterizing the dynamic behaviour and assessing the condition of civil engineering structures. Yet, inferring structural integrity, damage, and ultimately safety from vibration measurements acquired under predominantly elastic service conditions remains a fundamental challenge. This lecture will explore the evolution from conventional operational modal analysis towards automated, intelligent, and physics-aware structural monitoring frameworks. Particular attention will be given to advances in modal identification and tracking, the integration of artificial intelligence with physics-based modelling, and the role of digital twins in connecting experimental observations with numerical predictions. Drawing on recent research and applications involving bridges and heritage structures, the lecture will illustrate how these approaches can enhance the interpretation of measured responses, support damage assessment, and reduce reliance on manual intervention. The opportunities and limitations of artificial intelligence will also be examined, emphasizing the importance of physical consistency, uncertainty quantification, and experimental validation. Finally, emerging perspectives on adaptive and interpretable monitoring systems will be discussed, highlighting their potential to transform vibration data into more reliable information for engineering decision-making.