Pre-Recorded Sessions

Monitoring the health and working conditions of critical structures and machines can prevent failures and arrange for effective maintenance schedules. Access to vibration data from states with damages is valuable in order to train models that can perform explicit classification of faults in later states. This usually requires expensive and sometimes not feasible destructive tests in various operating conditions. Recent works have employed numerical models in order to simulate such scenarios. However, simulated data carries usually systematic errors in terms of the originally identified structure on the healthy state. In the present work, a test structure of a simple truss is studied with a Finite Element Model for vibration data generation. Training data sets are generated containing different errors in terms of damping, stiffness, and mode shape accordance in order to emulate real application faults. The training data sets are used after in a Convolutional Neural Network for damage classification. It is investigated how different simulation error sources affect the generalization in terms of correct damage case prediction on a reference data set. The results may aid in strategies for focus on specific FE model updating parameters in order to generate reliable simulated training data.

In this study, the adverse effect of an inadequately designed input control force is investigated in exciting the torsional modes of a structural system. The initial formulation is developed based on a generalized symmetric structural system under control input that is designed to regulate only the lateral modes of the system. The system is perturbed by considering an assumed eccentricity that may occur due to the change in the system stiffness. A closed-form expression of the control spillover to the torsional modes of the system is generated considering the first-order perturbation approach. The proposed closed-form expression is verified through a numerical example of a single-story three-dimensional symmetric frame structure with a rigid diaphragm at its top. The structure is subjected to several ground excitations with dominant broadband frequencies. The initial input control is designed considering the PID control strategy to control the lateral modes of the structural system. The influence of different parameters, namely, the stiffness degradation, Eigenvalue spacing, and the absolute natural frequencies are studied in the context of torsional spillover phenomena. In the last part of the study, an experimental investigation is performed on a frame, which is a scaled down model of the structure considered during the numerical study. An input force considering the PID control mechanism is fed back using an Arduino data acquisition system. Sinusoidal excitations with different frequency ratios and white noise excitations are considered as the input during the experiment.

This presentation discusses the basic concept and recent developments in the field of the intermodal targeted energy transfer (IMTET) - a novel suggestion for passive mitigation of mechanical and civil structures in conditions of blast and seismic excitations. This is achieved by introducing strategically placed, local strong nonlinearities, in the form of vibro-impacts of the elements of the structure with a relatively light, yet stiff, auxiliary structure(s). These impacts rapidly, robustly and irreversibly redistribute the input energy within the modal space of the structure through extremely rapid IMTET from low-to-high frequency structural modes, yielding enhanced mitigation of unprecedented effectiveness, right from the very first cycle of the structural response. Moreover, when optimized, this concept can be realized fully passively, without the need to adding any mass to the building, and with minimal increase in the resulting local accelerations and stresses. The presentation will address recent extension of the concept to mutli-dimensional and multi-DOF settings, as well as various possibilities for local/global design of the auxiliary structures.

Robot locomotion has become increasingly common to transport and assist humans in complex environments. Even though legged locomotion does not correspond to the most efficient transportation system, in some circumstances it is the only reasonable option to overcome obstacles, as stairs, or to move on rough routes, as slippery and uneven surfaces. As nature has provided ad-hoc solutions, results of the optimization process of the evolution, engineering and robotics have often found inspiration in biological behaviors, generating an entire new field, known as bio-inspired engineering. Most existing quadrupedal robots has been developed only with a single rigid trunk and the actuation is limited to legs and associated joints in order to reduce the complexity of the model to process during the simulations. However the optimization process carried out by the natural evolution shows that most legged vertebrates use flexible spines and supporting muscles to provide auxiliary power during the dynamic behaviors, resulting in higher speeds and lower cost of transport (CoT) during locomotion. In this context, this paper presents a non-linear optimization process to identify the highest performance gaits parameters of a quadruped robot, with a flexible trunk. The optimization process is done using different techniques such as the trajectories optimization carried out through direct or indirect methods and the evolutionary algorithms that are based on a heuristic search inspired by Charles Darwin’s theory of natural evolution and reflects the process of natural selection. Since the identification of the optimal gaits requires the complete definition of the ground reaction forces coupled with the optimal trajectories of the body and its vibrational modes, it is possible to determine what is the influence of the main parameters linked to the vibration of the trunk on the optimal gait resulting from the optimization process. In the final part of the paper is presented a comparison between the results obtained using different optimization techniques and is presented the comparison of the optimal gaits obtained in the case of flexible spine with respect to the case of a rigid trunk to demonstrate the benefits of the elastic model in term of efficiency. Through the observation of the different optimal gait from the two models it is shown that the elastic model permit to find the gallop, often observe in nature, as the output of the optimization process under certain hypothesis and this permit to demonstrate the benefits of the new elastic model of the trunk.

This work evaluates the nonlinear dynamic behavior and dynamic instability of an imperfect cylindrical panel with a discontinuous unilateral elastic base. The cylindrical panel is described by Donnell’s nonlinear shallow shell theory, being discretized by the Galerkin method, using a previous reduced order model obtained by a perturbation method. It is shown that an efficient modal solution with two degree-of-freedom is sufficient to describe the nonlinear softening behavior of the cylindrical panel with a discontinuous elastic base. The Heaviside and the Signum function are used to describe the elastic base domain and the unilateral contact force respectively. The obtained results, using the fourth order Runge-Kutta method, demonstrate the influence of the discontinuous unilateral elastic base and initial geometrical imperfection on the backbone, bifurcation diagrams, phase portraits and resonance curves of the cylindrical panel. They demonstrate the important changes in the stable and unstable regions of the nonlinear equilibrium paths and dynamics instability regions due to the unilateral contact constraint and the localized foundation.

An elastic layer submerged into a compressible fluid is considered. The results of the low-frequency asymptotic analysis are discussed. The adapted scaling corresponds to the so-called fluid-borne bending type wave. The approximate equations for the leading, first, second and third order are derived. The first order approximation corresponds to the traditional formulation for a thin Kirchhoff plate submerged into an incompressible fluid. It is remarkable that the plate inertia can be neglected at leading (zero) order. Higher-order corrections appear at second and third orders. In particular, the transverse shear deformation has to be taken into consideration at second order along with asymptotic corrections in impenetrability conditions. At the same time, the fluid compressibility has to be incorporated only at the third order. The associated approximate dispersion relations are also demonstrated. Numerical comparisons with the ‘exact’ dispersion relation are also presented.

Concrete is one of the most widely used construction materials in the world. As it is brittle and has a random, heterogeneous, multi-phase structure, it is very susceptible to cracking. Therefore, the detection of cracks at the earliest possible stage is of great importance. In recent years, various non-destructive testing (NDT) and structural health monitoring (SHM) techniques have been investigated to improve safety and control the current state of structures made of concrete. In this study, the emphasis is on micro-crack monitoring in concrete beams. The experimental analysis was carried out on concrete elements with dimensions of 4 cm × 4 cm × 16 cm. The beams were subjected to three-point bending in a testing machine under monotonic quasi-static loading. During the tests, the fracture process was characterized by ultrasonic waves. The monitoring was conducted using a set of PZT transducers. One of them was an actuator while the others acted as sensors. The excitation was a wave packet composed of a 5-cycle sine wave modulated by the Hann window. The procedure assumed excitation and registration of wave signals at selected time intervals during the whole process of mechanical degradation of concrete specimens in laboratory conditions. The recorded signals were further processed by coda wave interferometry (CWI). The technique allowed the detection of cracks using the decorrelation between ultrasonic wave signals collected at different stages of degradation. Different values of excitation frequencies in the range from 100 kHz to 500 kHz were used to investigate the influence of frequency selection on the effectiveness of the damage indication based on the decorrelation of the coda. The CWI monitoring was integrated with the optical tracking of the fracture using the digital image correlation (DIC) technique. The front side of the beam has been covered with the speckle pattern and the photographs of the samples were taken at selected time intervals during mechanical degradation. Then, local strain changes were calculated enabling the observation of growing cracks. The integrated use of CWI and DIC methods allowed effective fracture monitoring and early-stage damage detection. The combination of these two methods appeared to be successful in the detection of micro-cracks and upcoming macro-type damage. The results obtained from the experiment were intended to highlight the effect of applied frequencies on the coda wave interferometry.

Masonry construction has been practiced for thousands of years, being an important part of the world's cultural heritage, containing sociological, economic, cultural and political elements of the place where it is inserted, physically offering us the opportunity to research and learn about the past. The analysis and prediction of the behavior of historical structures over the years constitute a fundamental field of research to perpetuate the constructive methods and the cultural heritage of civilizations. Such analyses, involve not only the static aspect of the constructions, but must also include the seismic action, not only due to the possible damages generated to the structure, but also, to the uncertainties associated with the occurrence of this phenomenon in the patrimonial and social security. In this context, the urban complex of the Italian municipality and capital of Tuscany, Florence, which became a symbol of the Renaissance during the early Medici period and has the greatest concentration of universally renowned works of art associated with events of universal importance, is located in a seismic zone, with episodes of earthquakes in the month of May 2022, with a magnitude of up to 3.7Mw. Thus, the present work aims to numerically analyze the dynamic behavior and safety of the main masonry dome located in the historic center of Florence: Dome of Santa Maria del Fiore’s Cathedral. For this, the structural analysis software by the finite element method ABAQUS is used, and the numerical models are validated using the results of experimental analyzes, present in the literature. The final model, which adequately reproduces the experimental results, can be used as part of a permanent monitoring system for the structural health of the dome and the prevention of possible damage.

Advancements in computer sciences and technology allow for implementation of detailed numerical models of a system such as the Finite Element (FE) or Multibody Dynamics (MBD) models. Complex mechanical systems can easily be modelled in detail, yielding accurate results. This opportunity provided by these high-fidelity numerical models has led to the broad application of such methods in development and prototyping of mechanical systems, their optimization and fault analysis and so on. The capability of detailed modeling however usually comes at a great computational cost, with the simulation time needed for a problem in many cases rising exponentially, rendering these models impractical. This problem becomes even more profound when one considers the recent integration of model-based data in data-driven methods where a large number of datasets is usually required and multiple iterations of the same model must be simulated in order to produce the desired number of samples. To mitigate these shortcomings, surrogate modelling has been extensively used in applications including large systems or repetitive runs in the form of Reduced Order Models (ROMs) to reduce the computations time and render these simulation-driven methods more viable. Use of these ROMs however is limited to cases where low loss of information is ensured and the features lost due to the model simplification are insignificant. The developments in Artificial Intelligence (AI) and its applications have demonstrated its potential to accurately describe the relationships between a model’s inputs and outputs and as such using an AI algorithm as a surrogate model is a promising alternative. A properly trained AI algorithm can usually fit to FE and MBD models and yield accurate results at a fraction of the computational burden. To this end, an AI-based surrogate modeling framework is proposed in this work, with application on an experimental gear drivetrain system. A detailed MBD of the actual system is initially constructed and optimized via a black box optimization method in order to better simulate the physical system. A variety of supervised AI algorithms such as regression models and Convolutional Neural Networks (CNNs) is then examined as a surrogate to the various mechanisms of the system, aiming to replace them with the goal of reducing the simulation time while maintaining the high accuracy and fidelity of the original model. The various algorithms are then compared in terms of time reduction and accuracy both to each other and to the initial MBD model in order to conclude to the best suited for the application. The results are also compared to the measured response data of the physical system to ensure the validity of the models and prove the viability of the proposed method through its use on a relatively complex model. The proposed framework provideς an alternative to the commonly used ROM methods and the presented application acts as a benchmark case for its implementation to more complex systems and different operating conditions.

The wind loads on high-rise buildings can be predicted from the high frequency force balance (HFFB) technique. This experimental technique is widely accepted among wind and structural engineers due to its simplicity and adaptability to obtain wind-induced loads and dynamic responses of tall buildings. But, for the general case of non-linear and combined sway and twist mode shapes, analysis of HFFB measurements gives rise to serious challenges and various techniques have been developed using significant assumptions. In this study, a 1:300 scale model of Building A, a benchmark tall building geometry has been chosen to perform HFFB tests in a closed-circuit boundary layer wind tunnel. It has complex 3D mode shapes combining sway and twist motions. First, the base shears, overturning moments and torques for different wind directions were measured. Subsequently, different HFFB mode shape correction approaches were examined for estimating the dynamic wind effects using modal analysis in the time domain. Furthermore, the high frequency pressure integration (HFPI) wind tunnel technique was utilised to measure surface pressure information in separate tests to obtain the more precise wind effects results for validation of the HFFB predictions. The power spectral densities (PSDs) of aerodynamic base overturning moments and torques determined from both techniques are compared. The PSDs and variance coefficients of generalised wind forces for the first three modes are calculated and compared to each other to investigate the accuracy of the different HFFB-based analysis approaches. Also, the statistical properties including standard deviations and peaks of tip acceleration and tip displacement time histories are computed for comparison. The results show that the HFFB approaches overestimate the dynamic wind-induced responses and have some weaknesses for twist prediction. It is found that the HFFB methods with the same wind loading correlation state generally seem to perform in a similar manner.

Recent studies carried out in the literature have shown that the mechanical properties of bioinspired periodic composite materials can be strongly influenced by finite deformations effects leading to highly nonlinear static and dynamic behaviors at multiple length scales. For instance, microscopic and macroscopic instabilities may occur during macrostrain-driven uniaxial loading processes in elastomeric nacre-like composite materials and wave attenuation properties in lightened periodic nacre-like microstructures may evolve as a function of microstructural evolution, in turn, depending on the level of applied deformations. Consequently, finite deformation effects, in conjunction with specific micro-geometrical arrangements, give the opportunity for obtaining metamaterial properties not available in natural materials. In order to give a better understanding of the complex nonlinear phenomena occurring in the above materials, the nonlinear static and dynamic responses of novel periodic microstructures subjected to biaxial macroscopic compressive loading processes are analyzed. Firstly, the static response was investigated by solving the nonlinear boundary value problem based on the periodic homogenization theory while the onset of primary instabilities with short (microscopic instability) or long (macroscopic instability) wavelength was performed by solving the related frequency domain wave equation by exploiting the Floquet-Bloch theorem. Secondly, the dynamic response was investigated in terms of dispersion relations by solving the eigenvalue problem considering small amplitude motions superimposed on a finitely deformed configuration based on the Floquet-Bloch technique. Then the dispersion curves, relating the wavelength of the propagating waves to their frequency, were analyzed to determine the evolution of the complete bandgaps as a function of the applied strains. The numerical calculations were carried out by means of a finite element discretization performed on the simulation software COMSOL Multiphysics v5.6 together with some Java subroutines written in COMSOL Application Builder environment which can automate specific analysis procedures. To the best of the author’s knowledge, the scientific understanding of the wave attenuation properties in soft bioinspired composite materials subjected to high deformations is still limited. Thus, this work provides a valuable opportunity for researchers to further improve their knowledge in this research field by investigating how large deformations and the onset of instabilities affects the vibration control and band gap phenomena in new advanced bioinspired metamaterials. We proposed a new lightened nacre-like microstructure containing periodically arranged solid and hollow platelets and we investigate the propagation of elastic waves superimposed on a deformed state for increasing strains. The results have shown that wide complete bandgaps can be found also at lower contrast between the platelets and the matrix by suitably setting the percentage value of platelets volume fraction and the percentage of void inclusion given by the insertion of hollow platelets. In addition, was found that the onset of microscopic instability leads to a microstructural pattern transformation which, together with the level of applied deformations, strongly influences (sometimes positively and sometimes negatively) the wave propagation properties of lightened nacre-like composite metamaterials. Definitively, the obtained numerical outcomes provide new opportunities to design bioinspired soft composite metamaterials characterized by high deformability and enhanced waves attenuation capabilities.

The present contribution outlines a general methodological procedure for a consistent signal processing analysis of vibration response data, combining both classic and advanced techniques, toward structural monitoring and identification scopes. A specific case study is considered, for testing and validation, regarding the structural assessment of a historical infrastructure, a road three-span reinforced concrete arch bridge. The innovative contributions of the present research highlight four main points, in an integrated approach: the adoption of a Time Domain Compression technique, the application of a low-pass filter, the development of a Wavelet Analysis and the employment of an ARMA modelling approach. The analysis of acquired response signals through a Time Domain Compression technique allows, for set filtering parameter and threshold, to remove lower quality sub-samples from the full data. After such a step, the modified signal, to be employed in subsequent analyses, displays a higher quality, though being possibly endowed by a non-stationarity character. The information content of such signals can further be improved in quality by the adoption of a low-pass filter, with an intrinsic denoising effect, with respect to the characteristic natural frequencies expected for the structure under consideration. An advantageous contribution is provided by a Wavelet Analysis applied to the response data signals, offering, through a Wavelet transform, a further filtering effect based on frequency localisation within the time domain. A proper managing of the Wavelet family choice and leakage filtering issue may lead to a refined, higher quality, reconstructed signal, specifically useful in sub-sample identification and stationarity characterisation. A further deepening in response signal analysis and understanding is obtained by an ARMA modelling approach, with unknown source input (consistently with unknown loading configuration provided by operational traffic conditions on the bridge of the case study). In particular, the polynomial function applied to a white noise source in the model is interpreted as a filtering term apt to transform the source in a non-white noise configuration, relevant to the application case, therefore allowing for effective deciphering of the transfer function features as properties of the investigated structure. The considered approaches and whole methodology are extensively validated on the specific case study at hand and the presented results allow to derive effective observations regarding the current structural behaviour of the reinforced concrete bridge, while the study of the devised methodology shall highlight its general applicability with reference also to other structural setups. The combination of consolidated statistical signal processing analysis with advanced techniques provides a devoted methodological procedure to extract reliable structural properties of current condition assessment toward Structural Health Monitoring and intervention purposes, outlining a robust and efficient management monitoring platform.

One way to know the state of health of a cable and its force in operation under varying environmental conditions is to estimate its modal properties. In the present work, the vibration of a nonlinear polymer cable is studied in the laboratory using computer visual techniques (videos) and compare traditional sensors. Two types of excitations are used to make the cable vibrate: broad band ambient vibrations and with initial conditions (deformation or impact). Tests are carried out with various cable tensions in order to cover a greater range of fundamental frequencies and to evaluate the relaxation process that occurred in this type of material. The oscillation is measured simultaneously with traditional sensors and video cameras (including smartphones). The traditional sensors correspond to 6 ultrasonic sensors equidistantly located along the cable, a load cell that allows the cable tension to be measured, and an accelerometer near one end. With the measurement of each sensor, the frequencies and modal damping can be estimated, and by joining the 6 channels of ultrasonic sensors, the modal shape can be estimated. Additionally, 8 cameras are installed. 5 frontal, which record the entire cable, but from different perspectives, which mainly allow observing the vertical oscillation of the cable; 3 cameras that record parts of the cable (local measurement), and one camera that records the cable from above (oscillation out of the vertical plane). In each video obtained, the tracking of multiple feature points of the cable is performed using the KLT algorithm without the use of artificial targets. With this, the displacement records are obtained in pixels/second. The modal properties are obtained using parametric techniques like NeXT ERA, SSI-COV and ERA. In order to know the mode shapes perspective correction due to the angle of the camera with respect to the cable must be performed. Various methodologies are studied that allow solving this problem which include calibrated and non-calibrated cameras. The mode shape estimation is studied considering each camera separately, and also considering 2 or more simultaneous cameras. The latter allows a 3D reconstruction of the movement of certain discrete points of the cable when it is possible to synchronize the movement of a point from multiple views. This additionally allows performing an in-plane and out-of-plane vibration study simultaneously. A comparison of the results obtained with each method is made and its advantages and disadvantages are studied. Finally, the problems and difficulties that may arise in camera calibration, video synchronization, video sampling frequency instability, point tracking, among others, are presented.

This work studies vibration and radiated noise from railway systems using the 2.5D Boundary Element Method (BEM) formulation in the Bézier-Bernstein space. The proposed method allows the representation of the exact geometry of the track as it is done in computer-aided design (CAD) models. Thus, it is possible to evaluate problems with complex geometries, which are usually not adequately represented by the standard BEM and Finite Element Method (FEM) formulations. Radiated noise is computed from the normal velocity at the boundary of the rail system according to the integral representation of the sound pressure. Only the track boundary is meshed, as the radiation condition is implicitly satisfied in the BEM fundamental solution. The boundary integral equations are numerically evaluated. Moreover, the methodology allows the use of arbitrary high-order elements, making it efficient for the computation of radiated noise at high frequencies. The performance of the proposed method is shown by studying the radiated noise of different track systems concerning slab track and urban rail systems. The influence of track flexibility levels on the radiated noise is evaluated. Rail, rail pad, embedded system and slab properties are considered. The computed results are compared with some published by other authors.

This paper evaluates several data models for high-volume time-series bridge sensor data. Remote sensor technologies are becoming increasingly more reliable and affordable for ensuring safety and guaranteeing appropriate timing for bridge maintenance. Data returned from these sensors tends to be of a very high volume, with modern sensors returning hundreds of readings per second. When working with such a large volume of data, concerns arise with how efficiently the data can be ingested and retrieved. Inefficient data ingestion can cause the data transportation system to be overwhelmed and incorrect or incomplete data to be pushed to the database. Inefficient retrieval will limit the ability of bridge stakeholders to make real-time data-driven decisions about the safety of bridges. Data structure becomes a critical part of ensuring that these two operations can be performed as fast as possible with the technology available. This paper reviews the top current technologies for managing both structured and semi-structured data through the lens of high-volume time-series data from bridge sensors. We propose several potential structured and semi-structured data models for bridge sensor data and implement them with the appropriate technologies for each. The proposed data models are tested using data collected from actual bridges. We conclude by recommending a semi-structured data model that allows for data to be both more easily collected and ingested, without making any concessions in terms of the speed at which the data can be retrieved for data visualizations and analysis.

In this paper, a flexible strain sensing system for full-scale RC frame structures is presented. The system includes two flexible strain gauges in different configurations, a single full-bridge and a double half-bridge. It also includes a selector that can be switched between measuring axial and bending strains depending on where the sensor is mounted. The first part mainly introduces the development process of the flexible strain system, which is mainly designed to measure the requirement of strain field abnormality caused by structural damage. Then the development of a strain switching selector is introduced, which can switch two different strain measurements according to requirements without changing the sensor installation location and identify any anomalies as structural damage. To verify the sensitivity and performance of the flexible strain sensing system, sensors of both configurations were installed on the beams and columns of a full-size concrete frame structure, and cyclic loads were applied to them to measure the strain trend of critical elements. 14 LVDTs were also installed for monitoring the displacement changes of the structural model. The experimental results show that the flexible strain sensing sheet has high stretchability and sensitivity and can well monitor the strain trend of the full-scale frame model, it also can be used as an integrated Structural Health Monitoring (SHM) system to monitor the strain of local structural damage.

In updated model based structural health monitoring problems, updating of displacement-strain transformation matrix has not garnered much attention over the years, as most studies focus on updating mass and stiffness matrices. In this study, updating of the parametrized displacement-strain transformation matrix is carried out in a Bayesian probabilistic framework by combining the data from both acceleration and strain sensors. The Bayesian framework ensures that the uncertainties, especially modeling uncertainties, are explicitly considered and provides multiple possible estimates of the parameters, unlike a classical estimation framework where only one value is estimated. Samples from the posterior probability density function of the parameters are simulated using the transitional Markov Chain Monte Carlo method. Further, in the formulation part, the estimation of the parameters through classical estimation is also developed, and the relationship with Bayesian estimation is established. The updated displacement-strain transformation matrix can be used to obtain strains with higher fidelity in numerical simulations and to obtain robust response predictions incorporating modeling uncertainties. In addition, once a transformation matrix has been updated, only acceleration data are required for damage localization, i.e., strain measurements are not required post-updating, thereby reducing the overall health monitoring costs and storage burden. The Bayesian updating procedure is illustrated using experimental data from a four-story shear building laboratory model.

The demand for train speed is getting higher and higher. When the train increases to the critical speed, it will cause resonance of track and substructure, which will adversely affect the train driving safety. At present, ballastless track is widely used in high-speed railway, and the vibration characteristics is different from ballasted track. Based on two-and-half-dimensional (2.5D) finite element method, this paper establishes the vehicle-track-subgrade coupling model to analyze the critical speed of ballastless track under different foundation conditions. The relationship between critical speed and shear wave velocity of homogenous and layered subsoil is obtained through parametric analysis. The results show that the critical speed of ballastless track under homogenous subsoil depends on the Rayleigh wave velocity of subgrade. Besides, the presence of subgrade can improve the critical speed to be greater than the Rayleigh wave speed of homogenous subsoil and the critical speed is increased with the subsoil stiffness. Moreover, the critical speed of ballastless track under layered subsoil is influenced by the shear wave speed ratio of adjacent subsoil layers. Finally, the proposed empirical formula provides guiding significance for the construction of ballastless track.

The research and development field of structural health monitoring is developing, and as a result, it now provides choices for disaster prevention and life cycle extension. This can help boost structural safety and infrastructures resiliency. Researchers have taken an interest in the use of machine learning techniques for these applications as a result of the promising outcomes that those approaches have delivered. In spite of this, damage detection applications necessitate the use of sophisticated and reliable data-driven enabled technologies. This is because of the complex behaviour of structures that is notoriously difficult to track. In this investigation, triaxial accelerometers are used for data acquisition purposes in order to monitor and examine the structural behaviour of an experimental four-story building frame when subjected to a variety of structural configurations. The effectiveness of supervised machine learning algorithms was analysed and compared to that of unsupervised learning approaches. In spite of the fact that the latter have demonstrated good performances for the detection of initially established anomalies, they have only partially exhibited performances for newly produced damage patterns, which restricts their application in practical settings. The results obtained by the unsupervised algorithms, on the other hand, were superior in both cases, revealing essential capabilities for damage detection applications.

Eliminating expansion joints in bridge decks is considered a good alternative to address deck joint issues, reduce maintenance, and improve bridge-deck life expectancy. Using a link-slab to make the bridge girders (partially continuous) continuous only for lateral and longitudinal load effects, provides lower cost, improved durability, longer spans, improved seismic performance, better resistance to wind loads and storm wave loads, improved structural integrity, and improved riding quality. The aim of this research paper is to monitor and investigate the performance of the link-slab of a pedestrian bridge that is reinforced with Basalt fiber reinforced bars (BFRP). The research also investigates the concrete simple-span beams that are made continuous by pouring the continuity link-slab between the beam ends. The research provided recommendations for the instrumentation of the link-slab, a suggested monitoring plan, and an instrumentation system to monitor the temperature, strain, and elongation of link-slabs. The bridge has been instrumented with embedded and surface-mounted sensors and has been monitored to evaluate the performance of the new link-slab during concrete casting and after casting. Several types of sensors were used, and a data acquisition system recorded strains/deformations at regular time intervals. The preferred sensor types for this application are vibrating wire sensors with integrated thermistors. The sensors were strategically located on both sides of the midline of the link slab to capture strains in the BFRP bars, strains in the concrete link-slab, and the gap between adjacent beams’ ends. The research team started measuring and monitoring the strains, deformations, and cracks in the link-slab. The team also investigated the performance of the link-slab, evaluated the data from installed instrumentation, analyzed the results, and provided conclusions. All measurements have been corrected for temperature changes. Data has been collected during service over two periods of approximately 3 months each. The data captured was related to the temperature and shrinkage of the concrete and did not include any significant live loading since no load testing was performed. The strains experienced by the sensors indicated small strain levels compared to the BFRP ultimate strain levels. In addition, to live load flexural effects, thermal cycling could contribute to the concrete cracking over time in the link-slab if tension stresses build up due to global shrinkage and creep restraint of the connected FSB spans. After about 90 days over the time of monitoring, the average strain in the mid-joint gauges did not change significantly indicating minimal creep and or shrinkage restraint was experienced to date by the link-slab since the initial casting date. The maximum daily strain change due to thermal effects is about 500 microstrains.

Vortex-induced vibrations (VIV) is a phenomenon of self-exciting oscillations of elastic, or elastically mounted rigid bluff bodies exposed to in a gas or fluid flow. This sort of oscillations can be used to harvest electric energy from the kinetic energy of air or water flow. This fact determines the great practical significance in the study of this phenomenon. In this work we experimentally study a system, consisting of a finite-length circular cylinder elastically supported through a short splitter plate, which considered as a model of a wind generator. A support was a long rectangular steel plate 3x34 (mm). Its length, being a variable parameter, defines the structural bending and torsional frequencies. The plate is rigidly embedded into a massive base at one end and into a rigid cylinder with length 350 mm and external diameter 30 mm at the other end. Experiments were performed in a wind tunnel. As expected, we observed classical quasi-two-dimensional VIVs, at which the cylinder support experienced bending oscillations. However, at larger reduced speed, we discovered a new type of vortex-induced vibrations (VIV), which is caused by resonance with the torsional mode. It is suggested that during these fully three-dimensional torsional oscillations, von Karman vortex street, generated by upper and lower pieces of the cylinder, are shifted in phase by pi/2, and the aerodynamic transition from upper to lower segments occurs at the splitter plate, which prevents the vortex shedding at the cylinder center. For a torsion resonance state with beam length 170 mm, the characteristic reduced speed range is 7.2...7.9. The amplitude of torsional oscillations is investigated and turned out to be essentially larger than of classical quasi-two-dimensional VIVs, which suggests more effective energy harvesting from the torsional oscillations.

Over the past few decades, vibration-based structural health monitoring and damage detection have drawn significant research attentions world-wide. The major motivation for such a research emphasis is the need to develop reliable methods to monitor the performance of structures in real-time. Real-time or continuous monitoring is also beneficial for timely maintenance of structure and to avoid any undesired failure. In this study, the efficiency of a few vibration based damage detection approaches has been investigated for localizing damage in a few wall specimens of brick masonry structures (unreinforced and with partially confining horizontal and vertical reinforced concrete bands). Masonry is one of the most common and oldest material which is being used for constructing buildings for many years. Like concrete, masonry is also a brittle and highly non-linear material with good compressive but poor tensile strength. Hence, often reinforcement is required to make a masonry element strong in tension and ductile in nature. In masonry structures, damage generally occurs in distributed fashion and the pattern of cracks is often very complex as compared to steel structures. In addition, damping changes with the level of damage. An experimental study using shake table has been performed on several half-scale wall-specimens. A suit of ground motions were selected based on Indian Code specified design ground motion spectrum. These motions were then scaled from very low intensity to high intensity by introducing amplitude scale factor with respect to the code specified level. The purpose of varying intensity input was to introduce progressive damage when subjected to the these excitations. In between each scale change, shake table tests with low intensity white-noise motions were performed. The purpose of white noise run in between two consecutive level of excitation was to simulate an ambient vibration condition for system identification. The specimens and the shake table were instrumented with accelerometers and displacement sensors to measure the input and responses. Various vibration-based techniques such as frequency domain decomposition and natural excitation technique (NExT) coupled with Eigen realization algorithm (ERA) techniques have been used for extraction of modal properties. It has been found from the experimental results that damage level can be easily correlated with a few damage identification parameters.

This work focuses on the structural health monitoring of jacket-type foundations used by offshore wind turbines. A vibration-only response mechanism based on accelerometer data is specifically suggested. Classification models (supervised models) are one possible method for achieving the aim; however, it is difficult to collect data on structural damage to wind turbines, resulting in unbalanced data sets. The use of anomaly detection techniques to detect damage to wind turbines has been the subject of much research in an attempt to find a solution to this problem. This work, based on an anomaly detection model, has developed a methodology to detect crack bar deterioration at the wind turbine jacket consisting of two training phases with only healthy data: training of a generative adversarial network (GAN), and encoder training of an autoencoder based on the GAN model that has already been learned. A generator and a discriminator may be obtained through the GAN network training process. This model is used to train an encoder that permits the mapping of healthy pictures into a latent vector. Following encoder training, the encoder places the data at points in latent space that correspond to the input data's healthy state. The mapping of the image space to the latent space through the encoder and the subsequent mapping of the latent space to the image space through the generator should closely resemble the input image in the event of a healthy input image. However, when damage-state input images are used, the model output does not resemble the input. The image reconstruction error and a residual error comparison of the discriminator properties are the final two loss functions used for anomaly identification. The following are the work's key contributions: : i) the proposed strategy is based solely on healthy data; ii) a time-frequency feature extraction preprocess based on the Wigner-Ville (WV) transform is performed, as this transform gives accurate time-frequency representations for non-stationary signals; and iii) a signal-to-image conversion of WV features in multichannel grayscale images with as many channels as there are sensors in the structural health monitoring system. The proposed strategy has been tested through laboratory experiments on a scale model.

Models of multibody mechanical systems subject to motion constraints are employed in several areas of large engineering importance, like automotive, railway, marine and aerospace structures. A better understanding of the dynamics of these models leads to better designs. However, investigation of their dynamics is still a quite challenging engineering task. This is mainly due to the fact that the equations of motion governing the behavior of such systems are represented by a set of differential algebraic equations (DAEs) of high index. Since the treatment of these equations is a delicate task, much research effort has been devoted to the subject, in order to cure the associated problems, like constraint violation, leading to a gradual drift of the numerical solution from the exact response. Essentially, all the previous efforts at tempt to overcome these problems by application of index reduction or coordinate partitioning techniques. The analysis performed in the present work is characterized by three distinct features. First, the equations of motion before enforcement of the motion constraints are linear second order ordinary differential equations (ODEs). This is in contrast to most of the previous studies on the subject, where the underlying set of equations is strongly nonlinear, due to consideration of large rigid body rotations. Moreover, the motion constraints are also expressed in a linear form with respect to the coordinates. Finally, the present analysis is based on an appropriate set of equations of motion, which are expressed as a system of second order ODEs in both the original generalized coordinates and the Lagrange multipliers related to the constraint action. This was achieved in some recent work of the authors, by combining some fundamental concepts of Analytical Dynamics and Differential Geometry. Specifically, this causes a natural elimination of singularities associated with DAE formulations and leads to major advantages compared to previous work in computational Multibody Dynamics. Due to the linear nature of the class of systems examined, a modal analysis becomes applicable. Therefore, the analysis of the eigenvalue problem arising from the set of the equations of motions employed is of vital importance. First, it is shown analytically that this problem possesses two separate sets of eigenvalues. The first corresponds to a set of single eigenvalues, which are shown to coincide with the set of eigenvalues of the reduced system, resulting after elimination of the motion constraints. The second set of eigenvalues is related to the motion constraints directly. In this set, all the eigenvalues are double and possess a single eigenvector (that is, they have algebraic multiplicity two and geometric multiplicity one). In sharp contrast to previous work, all these eigenvalues are bounded, due to the inertia assigned to the La grange multipliers. Based on these results, the solution of the undamped problem is performed first, using a generalized modal analysis. Then, the solution of the corresponding problem with classical damping is also determined. Finally, the accuracy of the results obtained is checked and verified by comparison with results obtained by application of classical numerical methods.

The transmissibility function (TF) has been widely reported to be a damage-sensitive but excitation-insensitive damage feature. However, most TF-based novelty detection approaches fail to accommodate various uncertainties with a proper probabilistic model. Making full use of the multivariate complex-valued Gaussian ratio probabilistic model of TFs, a new damage detection method is proposed in this study by integrating the advantages of TF and hierarchical clustering. Different from hierarchical clustering for damage detection conducted on the basis of deterministic distance as a similarity metric, a multivariate Bhattacharyya distance is used to account for the uncertainty and correlation of multiple TFs. An analytical approximation of Bhattacharyya distance is efficiently derived by applying an efficient asymptotic expansion to avoid numerical integration. A case study is used to validate the performance of the proposed damage detection method. Results show that the method can avoid specifying the number of clusters and show more robust performance compared to the clustering based on deterministic distance of a univariate TF.

3D printing (additive manufacturing, AM) is a promising approach for the production of light and strong structures with many successful applications, e.g., in dentistry and orthopaedics. Numerous types of filaments differing in mechanical properties can be used to produce 3D printed structures, among others, polymers, metals or ceramics. Thanks to the ease of the manufacturing process, the wide use of biodegradable polymers is observed, e.g., polylactide (polylactic acid – PLA) and polyvinyl alcohol (PVA) with a practical application as soluble support for manufacturing complex-shaped elements. The current work dealt with the application of ultrasonic guided waves for non-destructive damage detection and imaging in AM plate elements. Different specimens were considered. The first one had an area of 100 x 180 mm2 and a thickness of 2 mm and was only made of PLA filament. It had structural defects in the form of voids, one being a total lack of material and three with different meshed fillings. The second one had the same surface area but its thickness was equal to 3 mm. The specimen was also made of PLA, however, some PVA inserts were introduced, thus it could be considered a composite structure. Each specimen was tested using the same algorithm. Guided waves were excited using a single PZT actuator and recorded contactless with the use of scanning laser Doppler vibrometry (SLDV) at a set of points spread in a regular grid located on one surface of the sample. Additionally, numerical finite element models of analysed samples were prepared using Abaqus/Explicit software. The real structure of the prepared plates was reflected using volume finite elements with a grid size of 1 mm. The wave excitation was simulated as a concentrated force with varying amplitude. The signals of propagating waves were registered in the same area as in the experiment. The collected signals were processed using the weighted root mean square (WRMS) algorithm. The processing was performed with different values of calculation parameters, namely, time of averaging (time window), determining the part of the signal taken into calculation, and the weighting factor, differentiating the influence of specific parts of the signal due to the time of propagation. The WRMS damage maps for both samples were prepared to differentiate intact and damaged areas. It was observed that the type of defect strongly influenced the efficiency of imaging. The limitations of the proposed approach were characterized. The presented results confirmed that guided waves are promising for non-destructive damage imaging in AM elements.

Model updating bridges the gap between experimentation and simulations, building trust in the model predictions. Stochastic model updating (SMU) gives a better understanding of the effect of uncertainties in the structure due to modelling errors, measurement errors and imprecise assumptions of boundary conditions, material/geometric properties, etc. SMU can identify the variability in nonlinear parameters of dynamical systems and at the same time assist in defining the parameter space. The Bayesian model updating framework of various model updating methods is used in this research to develop such stochastic models for jointed structures. Existing physical joint models such as cubic stiffness, dry friction, Iwan, Valanis, Dahl, Bouc-Wen and lumped models are used to generate samples of backbone curves. Backbone curves can aid in understanding the system's nonlinearity in the absence of any forced excitation. Researchers have used several control-based experimental methods over time to generate these backbone curves, such as phase-locked-loop (PLL) control testing, response-controlled stepped-sine testing, control-based continuation testing and resonant decay testing. This study focuses on the Resonant Decay method to generate the backbone curves where steady-state vibrations of the system are captured once the excitation is removed from the system and the free decay is achieved. Using this free decay, instantaneous frequency and amplitude are extracted to obtain the backbone curve. These backbone curves are used to develop the stochastic model of the system. The Bayesian model referred to here is the Monte Carlo Markov Chain (MCMC) with the Metropolis-Hastings algorithm as the sampler to sample from a high dimensional probability distribution. To reduce the computational time without losing the performance of the MCMC model, a machine learning (ML) model is integrated with the Bayesian model for stochastic modelling of a non-linear system. This integration is performed to make the Bayesian model less computationally exhaustive. The samples of the backbone curves are treated as a database for training, testing and validation of the machine learning (ML) model. Experimental validation of this ML-integrated Bayesian is performed using vibration testing of a non-linear system to generate backbone curves. The resonant decay method is employed by exciting the system at its first natural frequency and then removing the excitation to allow free decay. Backbone curves are extracted from this free decay. These experimentally generated backbone curves are used to validate the ML-integrated Bayesian results and finally develop the stochastic model.

Steel Moment Frames (SMFs) are among the most popular lateral load-resisting systems, especially in seismic zones. Due to their popularity, these systems have been studied extensively over the last decades. With the advances in computational capabilities, sophisticated mathematical models have been developed and used to study SMFs deeply, particularly in the context of performance-based earthquake engineering. Under this context, masonry infills are among the most vulnerable components of SMF buildings. Recent earthquakes (e.g., Ecuador 2016) have caused significant damage to these components. Motivated by this issue, this research investigates the influence of different types of infills on the seismic performance of SMFs. For this purpose, nonlinear mathematical models of SMFs varying in height are developed and subjected to nonlinear time history analysis. Thus, the effect of different types of infills on metrics such as the collapse probability (following the FEMAp695 methodology) is assessed. The infills evaluated in this study are masonry infills attached to the frame structure, light infills (e.g., gypsum), and masonry infills with flexible vertical joints. The mathematical models used to simulate the infills consist of state-of-the-art models validated experimentally in previous studies on the topic. Design recommendations will be suggested based on the numerical simulations. Limitations of the study, as well as lines for future research, will be discussed.

Real-time Hybrid Simulation (RTHS) is an effective method for estimating the dynamic response of structural systems. The system splits into a numerical substructure (NS) and an experimental substructure (ES). The NS is solved numerically, while a servo-hydraulic actuator imposes boundary conditions (typically displacements) over the ES. The command signal is based on the current NS response. This framework scheme has been demonstrated to work very well in different research applications. However, it is limited when the displacement is imposed on the high stiffness ES (e.g., axial DOF of a column). A possible solution is to impose forces over the ES instead of displacements. However, the literature shows that the load cells add considerable noise to the measured signal, causing high tracking errors and could cause instability in the simulation. To sort out this problem, the literature proposes adding a compliance spring between the piston of the actuator and the ES. The compliance properties are known a priori and considered deterministic, adding flexibility to the system. With this scheme, it is possible to command forces to the actuator and control them by measuring the spring elongation and the restoring force from a load cell. However, developing dynamic compensation for force-based RTHS is challenging due to the uncertainty of specimen-compliance-actuator interaction. Therefore, this study proposes a force-based RTHS with robust compliance compensation. The compliance spring and the load cell will have uncertainties in their properties. Robust adaptive model-based compensation will be employed to overcome force-tracking errors between substructures. The proposed methodology will be verified in a virtual RTHS environment, where parametric studies will be considered to check the system's robustness over uncertain compliance. In addition, specimens with high stiffness will be tested to evaluate and verify the proposed scheme, which will be implemented on a force-based RTHS considering the axial DOF of the specimen.

The motivation to accurately model the dynamics of flexible aircraft grew with the development of great aspect ratio aircraft and the great research effort for the development of energy efficient and ecologically correct aircraft, such as those powered by solar energy. These configurations have, as a common characteristic, lower natural frequencies of the first structural modes in relation to more conventional aircraft. This decrease in structural frequencies can result in coupling with the rigid body response. The development of a suitable model that accurately represents the flight dynamics of a flexible aircraft has been pursued by industry and aeronautical research organizations during the last decades. Many groups have proposed a number of different approaches in this direction. And one direction of these approaches is to address the problem of finding a flexible aircraft model using systems identification methods. Recent work on flight tests showed the possibility of adopting the Output Error Method in the time domain to identify and model the dynamics of a flexible aircraft. In this work, the objective is to apply an integrated model containing longitudinal and lateral directional rigid body dynamics, coupled to the first eight flexible body modes, for identification and validation from flight test data. That is, it represents an extension of the current use of the time domain system identification methodology to obtain dynamic models with a wider application frequency range, which can be especially useful for flight control systems projects, in which Structural sway control is intended to improve passenger comfort or reduce structural loads. To carry out this work, an aircraft with a flexible wing, EOLO, from the Aeronautical Systems Laboratory of the Aeronautics Institute of Technology was used. Initially, a finite element structural model (FEM) based on beam elements, concentrated masses, and rigid bars was used. Furthermore, the quasi-stationary panel model based on the Vortex Lattice Method (VLM) was adopted for the aerodynamic model. After obtaining the interpolation between the structural and aerodynamic models, the concept of the mean axes reference system was used in the motion equations of flexible aircraft, and a simulation of the aircraft's dynamics was obtained, consequently, the maneuvers were designed to excite dynamic modes during flight. After the flight test, the identification was produced. For aircraft identification, instead of obtaining stability derivatives, commonly obtained in several identification works, two diagonal matrices were used to correct the matrix of aerodynamic influence coefficients (AIC) obtained via VLM, pre and post-multiplication. The estimation of the main diagonal elements of each matrix was obtained through the Output Error Method in the time domain. After the identification stage, a validation analysis was carried out, which shows the improvements obtained by using the integrated model when compared to the traditional rigid body approach. In addition, it is beneficial to get correction matrices instead of stability derivatives, since they can be used more directly in the correction of aeronautical design stages and also observe the behavior of the aircraft in relation to loads described in space, such as gust loads.

Tsujun bridge is a stone arch bridge in Kumamoto Prefecture, which has been designated as an important cultural property of Japan. Tsujun bridge suffered slight damage during 2016 Kumamoto earthquake. No stones fell down, but wall stones were pushed out in the transverse direction of the bridge. Moreover, cracks were found in the soil which was filled between stone canals and wall stones. Those damages were found not in the central part of the stone arch but between the arch end and the bridge end. This study explains the damage mechanism of Tsujun bridge due to Kumamoto earthquake though microtremor observation and numerical simulations. First, input ground motions for Tsujun bridge during Kumamoto earthquake was estimated by microtremor observations. Next, an analytical model of Tsujun bridge was created based on the observed natural frequencies. Then, the seismic response of Tsujun bridge was simulated using the refined distinct element method. The actual damage situation that observed damage was not found in the central part of the stone arch but between the arch end and the bridge end could be simulated, and its reason was explained. Moreover, a simple countermeasure to reduce the bridge deformation without ruining historic values was considered and its effect was investigated.

This work shows the preliminary results related to the evaluation of the dynamic behavior of a 5x5 m masonry cross vault after undergoing different changes in its structural behavior. It has been built using solid clay bricks and lime and sand mortar. The cross vault has been built in the laboratory specifically for carrying out these tests and evaluating its structural behavior. A device has been designed to be able to carry out a controlled settling of one of its supports without causing an instantaneous collapse of the structure. Dynamic identification using OMA was performed by installing seismic accelerometers at several selected points on the structure, using ARTEMIS software to analyze the results. The evolution of the dynamic characteristics was carried out in four different states: (i) After construction (ii) After the generation of generalized damage through the forced settlement of one of its supports. (iii) After repair using Textile Reinforced Mortars (TRM) (iv) After generating new damage by settling one of its supports in the TRM repaired structures. After the generation of the damage in the structures, important cracks are observed in various points of the structure, although no situation was reached that would foresee a possible collapse. The repair with TRM has consisted of the introduction of fiberglass mesh bands coated in lime and sand mortar. These bands have been arranged forming a cross in the upper part of the structure and in all the edge arches of the structure. In all these phases the OMA was developed under a situation of ambient vibrations, where these vibrations levels were very low due to the situation inside of a laboratory. After each of these phases, their main frequencies, their vibration modes and the structural damping factor of each of the identified modes have been obtained. It is observed how a loss of rigidity of the vault occurs after the production of damage and how it completely recovers by means of reinforcement and repair with TRM. An increase in the damping ratio of practically all the vibration modes has been detected when the structure is damaged, although after repair with TRM these values are reduced again.