Special Sessions
Please click on the corresponding title of each Special Session to get more information
Guided Waves in Structures for SHM
Organizer: Wieslaw Ostachowicz
Keywords: sensors, sensing, SHM, damage detection, signal processing
Scope of Session: There is an excellent potential for model-based approaches in which guided waves are utilised for damage detection, localisation and size estimation. The session covers the main disciplines based on guided wave propagation in both isotropic and anisotropic materials. Authors are encouraged to submit papers that include the elastic waves propagation phenomenon, which spans a wide range from linear and nonlinear, 1D, 2D and 3D, time or frequency, and experimental and numerical approaches in complementary investigations of structures. The proposed novel techniques should allow for efficiently performing both local and global SHM technologies. The above investigations are intended to develop various strategies related to diagnostics (damage size estimation and damage type recognition) and prognostics. The promising combination of investigated techniques should lead to an innovative approach to ensure safe operation.
SHM Technology in Wind Turbines
Organizer: Wieslaw Ostachowicz
Keywords: wind turbines, sensors, sensing, SHM, damage detection, signal processing
Scope of Session: The session covers the main Structural Health Monitoring (SHM) topics on wind turbine structures. The research methodologies here span a wide range of experimental and numerical approaches in complementary investigations of a rotor with blades drive train and support structure. The crucial issue is assessing fibre-reinforced polymer materials because they are widely used for wind turbine blades. The research methodologies should span a wide range of topics from piezoelectric transducers, elastic waves propagation phenomenon, fibre Bragg gratings, structural vibrations analysis, electro-mechanical impedance method, acoustic emission, damage mechanics, 3D laser vibrometry applications and others. The combination of proposed techniques allows for performing efficient both local and global SHM of the structure. It also includes various techniques related to diagnostics (damage size estimation and damage type recognition) and prognostics. The promising combination of selected techniques should lead to an innovative approach to ensure the safe operation of the structure.
Seismic SHM for civil structures
Organizers: Maria Pina Limongelli(1) and Mehmet Celebi(2)
Key words: seismic SHM, civil structures, damage identification, real time monitoring, emergency management
Scope of Session: During the last two decades, due to a need and a growing interest by both researchers and professional, seismic structural health monitoring (SHM) has evolved. Numerous monitoring systems installed in structures in various seismic-prone countries utilize real-time or near-real-time responses recorded during strong earthquakes to make informed decisions related to the health of their structures. These data have strategic importance both for the advancement of knowledge on the behavior and performance of structures under strong seismic actions and for the calibration of realistic and reliable numerical models that are aimed to reproduce the structural behavior and to formulate a diagnosis about possible damages. Furthermore, the possibility to assess the seismic vulnerability based on data recorded on the monitored structure opens new avenues in maintenance policies, shifting from a traditional ‘scheduled maintenance’ to a ‘condition-based maintenance’, carried out ‘on demand' or ‘automatically’, basing on the current structural condition. The aim of this Special Session is to report recent advances in this field and successful applications for civil structures and infrastructures: buildings, bridges, historical structures, dams, wind turbines, and pipelines. The session deals with theoretical and computational issues and applications and welcomes contributions that cover, but are not limited to, seismic SHM algorithms for identification and damage detection, requisite strong motion arrays and real-time monitoring systems and projects, instrumentation and measurements methods and tools, optimal sensors location, experimental tests, integration of seismic SHM in procedures for risk assessment and emergency management. Such a session will provide a venue for the exchange of information on ongoing developments and assess successes and limited successes of SHM.
Recent Advances on Data Processing Techniques for Ultrasonics-based SHM/NDE
Organizer: Prof. Salvatore Salamone
Key words: guided waves, acoustic emissions, signal processing, SHM, NDE, damage detection
Scope of Session: This special session aims to collect and share recent developments in data processing techniques to enhance accuracy and capabilities of ultrasonic wave techniques for the SHM of complex structures. Authors are encouraged to submit papers topics that include but are not limited to: 1) deep learning, 2) data mining, 3) data analytics, 4) sparse matrices for machine learning. Both theoretical contributions and practical applications are welcome.
Distributed and Quasi-distributed Fiber-optic and Electrical Sensors, and Associated Data Analysis and Management
Organizers: Branko Glisic1 and Daniele Zonta2
Key words: Distributed fiber optic sensors; Sensing skins, sheets, and paints; Self-sensing materials; Dense arrays of active wave-propagation sensors; MEMS; Distributed/decentralized data analysis; Wireless nodes for dense arrays of sensors; Power harvesting for dense arrays of sensors
Scope of Session: Damage frequently occurs in form of strain-field anomalies. Strain-sensitive sensors installed at location of damage have unusually high change in their output signal and thus, can detect the damage reliably. However, it is difficult to know the exact location of damage prior to its occurrence. To address this challenge, very dense arrays of sensors could be used. Their “omnipresence” on the structure and their high sensitivity to damage, makes them very promising for reliable and accurate detection, localization, and quantification of damage. Several innovative techniques for enabling distributed and quasi-distributed arrays of sensors emerged in the last decade or so: (i) 1D distributed fiber optic sensors, (ii) 2D distributed sensing skins, paints, and sheets based on nano-technologies, large-area electronics, photonic crystals (nanospheres), conductive polymers, etc., and (iii) 2D and 3D active wave sensing techniques, embedded MEMS, and self-sensing materials. The aims of this special session are (1) to assess the state of the art of the techniques enabling dense arrays of sensors, (2) to identify challenges related to their applicability in real-life settings and (3) to cross-fertilize the research field through an exchange of ideas. In a broader sense, the topic of the session includes data management and power harvesting techniques that can address the challenges related to execution, processing and analysis of large number of measurements performed by very dense arrays of sensors.
Challenges of and Solutions for Long-term Structural Health Monitoring
Organizers: Mauricio Pereira1 and Branko Glisic2
Key words: long-term structural behavior, rheological effects, creep and shrinkage, aging, corrosion, wind loads, thermal loads, machine learning, artificial intelligence
Scope of Session: The long-term behavior of structures, monitored over multiple years, is of high importance in our context of aging infrastructure and anthropogenic climate change, which is associated with shorter return periods of critical events, such as hurricanes and storm surges, and changes in ordinary environmental loads (e.g., temperature and wind load patterns). Further, improved prediction of long-term response of civil infrastructure can benefit maintenance scheduling and early anomaly detection. However, the long-term behavior of real structures has not been extensively studied due to multiple complicating factors, such as uncontrolled environment, uncertain loads and boundary conditions, rheological effects (creep and shrinkage), and the challenges associated with long-term SHM disruptions, such as sensor damage, power and/or data losses, cost, etc. This session invites research and/or application contributions addressing the long-term safety, performance, and serviceability of civil infrastructure using SHM. The session is open to a wide range of contributions, including, but not limited to, evaluation of long-term performance of sensors, algorithms for identification of slowly evolving phenomena in long-term, data management and analysis, and real-life long-term applications. Approaches employing innovative machine learning methods are particularly welcomed.
Multifunctional Materials and Metamaterial Structures
Organizers: Ken Loh 1, Donghyeon Ryu 2, Nathan Salowitz 3
Key words: actuation, bio-inspiration, energy dissipation, energy harvesting, metamaterial, mechanical response, nanocomposite, self-healing, sensing, stimuli-responsive
Scope of Session: Multifunctional materials and metamaterial structures are ones that have been intentionally engineered to exhibit more than one precisely defined property and with properties that can even surpass materials commonly found in nature. Often, encoding of desired properties is achieved through a “bottom-up” design methodology during material manufacturing. While nanotechnology has enabled molecular assembly of a variety of new materials/structures that are then scaled up, the design and engineering of innovative multifunctional structures can occur at any length scale. As a result, this new class of material system can exist in the form of a two- or three-dimensionally structured nanocomposite, composite, coating, and/or multi-phase material. This special session welcomes contributions that showcase the breadth of multifunctional and metamaterial material architectures, nanocomposites, field-responsive metamaterials, mechanical metamaterials, novel and additive manufacturing methods, multi-scale design and characterization, numerical modelling, topology optimization, validation and testing, and technology demonstration, among many others.
Nonlinear Acoustic and Ultrasonic Techniques for Structural Health Monitoring
Organizer: Tribikram Kundu
Key words: sensing, nonlinear ultrasonic, SHM, NDE, detection, damage
Scope of Session: Papers are invited from various aspects of nonlinear acoustic and ultrasonic techniques such as higher harmonic generation, sub-harmonic generation, nonlinear resonant acoustic spectroscopy, sideband peak count - Index (SPC-I), vibro-acoustics and different wave modulation techniques. How these techniques are used for non-destructive evaluation (NDE) and structural health monitoring (SHM) will be the focus of this special session. Papers dealing with the difficulties and shortcomings of various nonlinear techniques and challenges encountered by the investigators in implementing nonlinear techniques using body waves and/or guided waves are of interest for this session. Recent developments of new promising nonlinear techniques that can overcome some of the existing shortcomings are of particular interest. Objective of this session is to give the attendees a broad overview and recent developments of nonlinear acoustic techniques.
Acoustic Emission and Hybrid SHM
Organizers: Prof. Victor Giurgiutiu1 and Prof. Zhenhua Tian2
Key words: acoustic emission, AE, non-destructive evaluation, NDE, structural health monitoring, SHM, passive detection, active detection, fracture, crack growth, composite, fiber breakage, matrix cracking, damage
Scope of Session: This special session will address the topic of acoustic emission and hybrid SHM. Acoustic emission (AE) is a passive SHM technique that relies on ‘listening’ to the elastic waves generated when an incremental crack growth occurs, or impact damage happens in composites. The elastic waves associated with AE events can travel a considerable distance in metallic structures which have a low damping dissipation coefficient. AE waves also travel in composite materials, but their travel distance may be less due to the higher damping dissipation of polymer matrix composites. Hybrid SHM techniques encompass a large class of methods that aim at combining several techniques to increase the probability of damage detection. For example, one may use passive SHM to record a damaging event (such as an impact in a composite structure) and then apply active SHM to try to estimate the magnitude of the resulting damage and its severity. Or one can listen to AE events which indicate that cracks are progressing into the structure and then follow up with the active SHM technique to evaluate the crack size. Or one can use two different active SHM techniques (e.g., pitch-catch wave propagation and electromechanical impedance standing waves) to better detect the damage location and size. But these are just examples. The session is open to all innovative techniques aimed at enhancing SHM capabilities. Contributions that judiciously combine theory and experiments are highly encouraged.
AI-based diagnostics & prognostics of lightweight structures
Organizers: Dimitrios Zarouchas & Theodoros Loutas
Key words: diagnostics, prognostics, AI/ML, digital twins for SHM, surrogate modelling
Scope of Session: This special session will gather the research community active in the area of SHM towards damage diagnostics & prognostics, address the challenges, discuss the present as well as future trends and exchange ideas & experiences across different engineering applications. Studies in the area of damage detection covering the four levels of the SHM hierarchy (anomaly detection, localization, identification and severity of damage) as well as studies aiming the ultimate-level of SHM prognostics of the remaining useful life of lightweight structures subjected to various types loading, i.e. fatigue, impact, using data-driven, physics-based and hybrid models, are welcomed to be presented in this session. Emphasis is given in how AI/ML and digitalization enable real-time damage diagnostics and prognostics by integrating SHM data with numerical simulations and structural degradation rules in the form of digital twins.
Human Performance and Human-Structure Interactions
Organizers: Ken Loh 1, Fernando Moreu 2, Hae Young Noh 3, Liming Salvino 4
Key words: actuators, behavior, data visualization, digital health, human in the loop, internet-of-things, prehabilitation, rehabilitation, sensors, structural performance, wearable
Scope of Session: The optimal performance of complex systems not only requires both the human operator and structure to function at peak performance, but it also depends on how the two can effectively work together. Therefore, monitoring the human operator and how they interact with and control artificial structures is crucial for optimizing system performance and functionality while ensuring safety. Failure to consider the human operator and structure as an integrated system – and the failure of any one of these – can result in mission failure or poor/sub-optimal performance. This special session is soliciting contributions focused on sensing and modelling for human performance and health, as well as the interactions/interfaces between humans and artificial structural systems. Examples of specific topics of interest include: (1) wearable Internet-of-Things (IoT) technologies and feedback mechanisms; (2) bio-marker, biochemical, and bio-molecular sensing; (3) understanding and modelling of structural responses induced by humans or animals; (4) human-structure interfaces that enhance system performance; (5) novel augmented/virtual reality and data visualization methods; (6) human-centric structural management methodologies; and (7) laboratory and field validation studies on human performance assessment and human-structure interactions.
Remote satellite-based structural health monitoring of structures and infrastructure
Organizers: Maria Pina Limongelli1 and Daniel Cusson2
Key words: InSAR, space-borne, radar backscatters, deformation monitoring, damage detection
Scope of Session: Satellite-borne InSAR technology provides an appealing complementary approach to traditional SHM in order to measure mm-accurate displacements over large geographic areas and to follow their evolution over time. The possibility to monitor large areas (e.g., an urban centre) opens new avenues for the development of automatic alerting systems that can flag several single structures with suspected structural integrity issues within a given network. For instance, InSAR measurements can provide information relevant to displacement time history, displacement rate, and thermal deformation, which can all provide useful insight into ongoing deterioration phenomena. Optical-band satellite imagery can nicely complement InSAR-based monitoring with additional information for hydraulic structures, for example, on river current speed and direction, nearby vortex formation, which can help assess the risk of pier scouring for river bridges and marine ports. This special session will provide the venue to present and discuss theoretical developments and field applications in order to foster future research collaborations on the topic. It welcomes contributions on algorithms (including AI-based) to process and analyse satellite imagery and data for damage detection purposes, case studies, measurement and calculation methods, and integration of remote and local sensing data for condition assessment and decision making.
Advances in Data Science, Artificial Intelligence, Machine Learning, and/or Computer Vision for Structural Health Monitoring
Organizers: Mohammad Jahanshahi1
1 Lyles School of Civil Engineering, Elmore Family School of Electrical and Computer Engineering
Purdue University, West Lafayette, IN, jahansha@purdue.edu
Keywords: data analysis, machine learning, artificial intelligence, deep learning, computer vision
Scope of Session: The recent advances in artificial intelligence (AI) have led to ground-breaking innovations in the broad area of structural health monitoring (SHM). In particular, besides computer vision approaches, data science and big data analytics provide an unprecedented opportunity to complement traditional SHM. To this end, this special session will provide the opportunity to discuss recent theoretical, computational and experimental advances in using not only computer vision but also machine learning and big data analytics solutions in general for structural identification, control, damage detection, inspection and health monitoring. Topics relevant to this session include, but not limited to, deep learning, machine learning-based damage assessment, physics-informed machine learning, generative adversarial networks, network pruning, virtual, augmented and extended reality, innovative imaging, image/video data collection and analysis, vision-based displacement and dynamic measurements, 3D LIDAR and depth sensors, vision-based inspection using unmanned aerial vehicles (UAVs) and unmanned ground vehicles (UGVs), robotics integration, and other new emerging machine learning or vision-based technologies.
Probabilistic SHM
Organizers: Daniele Zonta1 and Branko Glisic2
1 CEE, University of Trento, Via Mesiano 77, 38123 Trento, daniele.zonta@unitn.it
2 CEE, Princeton University, E330 EQuad, Princeton, NJ 08544, USA, bglisic@princeton.edu
Key words: Bayesian inference, probabilistic methods, sensor fusion, structural reliability, risk analysis, decision making
Scope of Session: Structural health monitoring aims to understand the condition of a structure based on sensor observations, a process which is typically affected by uncertainties in the model assumptions and in the measurements. Key questions are how to provide a reliable and robust diagnosis, properly accounting for these uncertainties, and how to rationally exploit the monitoring information to make decision on such issues as structural maintenance, repair and replacement. The goal of the session is to bring together researchers working on statistical data interpretation, structural risk assessment, and decision making. Contributions are invited in the fields of structural reliability, probabilistic analysis, Bayesian logic, sensor fusion, risk analysis, including economic and social aspects that affect decisions in SHM applications. Contributions proposing methodological developments and in-field applications are both welcome.
Integrating physics in data-driven methods for SHM
Organizers: Fotis Kopsaftopoulos1 & Dimitrios Zarouchas2
1Intelligent Structural Systems Laboratory, Department of Mechanical, Aerospace & Nuclear Engineering, Rensselaer Polytechnic Institute, Troy, NY, USA
2 Center of Excellence in Artificial Intelligence for Structures, Prognostics & Health Management, Aerospace Engineering Faculty, Delft University of Technology, 2629 HS, Delft, the Netherlands
kopsaf@rpi.edu & d.zarouchas@tudelft.nl
Key words: physics-informed models, data-driven models, Explainable Artificial Intelligence (XAI), machine learning methods
Scope of Session: A successful implementation of SHM relies to a large extend on the quality of data and the damage information that this data contains. While the structure is in operation, multi-sensing techniques are usually employed (i.e., vibration-based, acousto-ultrasound, and optical-based techniques) in order to fulfill the four levels of SHM; damage existence, localization/identification, severity estimation (quantification), and remaining useful service life estimation/prediction, resulting to a vast amount of data that increases the complexity of analysis and, at the same time, reduces the efficiency and interpretability of the SHM system. Data-driven and Machine learning (ML) algorithms have attracted the interest of the community, and although promising results have been produced, they are still facing several significant challenges and their acceptance remains under consideration by the operators of structural assets and certification bodies. To overcome this challenge and provide competent solutions, integrating physics into the data-driven/ML SHM methods, in the form of prior expert knowledge, (semi)-empirical rules, physics laws and constraints, physics informed neural networks (PINN) offers a great potential. This special session welcomes contributions that explore and propose how integrating physics into data-driven methods/ML for SHM has the potential to increase the effectiveness, robustness, reliability, and deployment of SHM systems, as well as the interpretability of the health monitoring data and the explainability of data-driven/ML algorithms. Emphasis is placed on contributions that address the integration of physics- and data-driven methods within statistical and/or probabilistic frameworks as well as highlight comparative analyses using experimental data.
Structural Monitoring for Spacecraft Structures
Organizers: Maria Sakovsky1
1 Department of Aeronautics and Astronautics, Stanford University, Stanford, CA, sakovsky@stanford.edu
Keywords: sensing, SHM, flexible structures, composite materials, extreme environments, multifunctional materials and structures
Scope of Session: Spacecraft structures are subjected to the harsh thermo-mechanical environments of space as well as launch. They can be in operation for decades and offer little opportunity for servicing. In addition, modern spacecraft carry large apertures that are packaged and deployed on orbit. Such deployable structures are subjected to high strains and even when deployed, can be flexible when compared to structures on the ground. Therefore, it is crucially important to understand the state of these structures, whether to measure their shape or detect damage. This session invites contributions focusing on all aspects of structural monitoring specific to spacecraft structures. Topics of interest include but are not limited to: structural health monitoring for thin composite materials, sensing for the space environment, monitoring of large flexible structures, and lightweight, energy-efficient sensing of large structures.
Dynamic Data Driven State Awareness for Intelligent Structural Systems
Organizers: Erik Blasch1, Fotis Kopsaftopoulos2, Fu-Kuo Chang3
1 Air Force Office of Scientific Research, 875 N Randolph Street, Arlington, VA 22203, erik.blasch@gmail.com
2 Department of Mechanical, Aerospace & Nuclear Engineering, Rensselaer Polytechnic Institute, Troy, NY
3 Department of Aeronautics and Astronautics, Stanford University, Stanford, CA 94305
Key words: dynamic data, state sensing, data-driven modelling, state awareness, machine learning, structural awareness
Scope of Session: Future intelligent structural systems will be autonomous, i.e., they will be able to “feel,” “think,” and “react” in real time based on high-resolution ubiquitous sensing, leading to unprecedented self-awareness, own-diagnostic, and intrinsic-healing capabilities. Infusing intelligence through autonomous capabilities requires the continuous acquisition, processing, and interpretation of large amounts of data dynamically collected from heterogeneous sensors and input types both from real-time sensing sources as well as historical sources, such as design/fabrication inputs, and maintenance data. Analysis systems should fuse these modalities, anticipate the availability of new sensor types, as well as integrate multi-scale multi-physics computational models of different levels of fidelity and computational cost. Recent interests in Artificial Intelligence and Machine Learning (AI/ML) promote a data-driven approach to stochastic and deterministic analysis; however for complex and time-dependent scenarios, AI/ML techniques may not meet performance needs. Contrasted to data-driven approaches are model-based approaches based on the first-principles modelling evident in theory, simulations, and analysis. For future data-driven state awareness and structural health monitoring (SHM) methods, a coordination of model-driven and data-driven approaches may be needed under the multidisciplinary research field of Dynamic Data-Driven Application Systems (DDDAS); DDDAS have the ability to dynamically incorporate real-time data into an executing application embedding a model, and in reverse, the ability to steer the data measurement processes based on the system’s dynamic data integration and interpretation. Papers are sought that address the development and understanding sensing, awareness, and monitoring of complex physics-enhanced machine learning (PEML) structural systems. Topics include sensor systems, state and input estimation, self-diagnostics, real-time monitoring and assessment, full-scale simulations, networked sensors, designs towards situational, materials, and state awareness of engineered systems. Applications of interest include aerospace systems, aerial vehicles, mechanical systems, smart and additive manufacturing, novel material systems/designs, civil infrastructure, etc.
Inspecting and Preserving Infrastructure through Robotic Exploration (INSPIRE)
Organizers: Genda Chen1 and Yang Wang2
1 Missouri University of Science and Technology, USA gchen@mst.edu
2 Georgia Institute of Technology, USA yang.wang@ce.gatech.edu
Key words: damage detection, infrastructure inspection, image analytics, robotics,
and structural health monitoring
Scope of Session: This session is focused on advanced robotics and sensing technologies toward structural inspection, image analytics, abnormality detection, and infrastructure planning in the broad area of structural health monitoring (SHM). The field of robotics has been widely explored long before SHM technologies attracted significant attention. However, only in recent years have robotic prototypes been developed to a level of maturity that makes them suitable for realistic applications in SHM. This session welcomes papers that explore the use of robotics and sensing technologies in overcoming contemporary challenges associated with aging physical infrastructure. Examples include but are not limited to vision-based unmanned aerial vehicles (UAVs) for crack detection and other surface condition of structures, UAVs for construction site monitoring towards abnormality detection, crawling robots with non-destructive evaluation for bridge inspection, among others.