The great availability of measurement systems permits to acquire quite easily data related to structural oscillations in operational conditions. This occurrence may permit to enhance our capability to data-driven computing using directly experimental data and pertinent constraints and conservation laws, such as compatibility and equilibrium, surely certain. In the paper, a methodology will be presented to furnish an analytical mechanical model of a suspension bridge in which the main parameters can be derived from vibration measurements. In this respect, Polymax and Enhanced Frequency Domain Decomposition identification procedures are used to determine a complete modal model which is used to evaluate an error function. Optimization algorithms are used to evaluate the function minima in the fundamental parameter space. The procedure will be validated by results coming from a sophisticated finite element model for which geometric measurements are included through a 3D point cloud geometrical model and a consequent Building Information model (BIM) constructed with images acquired by unmanned aerial vehicle (UAV). The case study of the pedestrian cable suspension Polvorines bridge (100 meters of span) is considered to demonstrate the procedure, due the test campaign conducted on March 2020.

Precise knowledge of dynamic characteristics and data-driven inference of material properties of existing buildings are key for assessing their seismic capacity. While dynamic measurements on existing buildings are typically extracted under ambient conditions, masonry, in particular, exhibits nonlinear behavior at already very low shaking amplitudes. This implies that material properties, inferred via data-driven model updating under ambient conditions, may be inappropriate for predicting behavior under seismic actions. In addition, the relative amount of nonlinearity arising from structural behavior and soil-structure interaction are often unknown. In this work, Bayesian model updating is carried out on field measurements that are representative of increasing levels of shaking, as induced during demolition, on a pre-code masonry building. The results demonstrate that masonry buildings exhibit nonlinear behavior as the elastic modulus drops by up to 18% in the so-called equivalent elastic range, in which the observed frequency drop is reversible, prior to any visible sign of damage. The impact of this effect on the seismic assessment of existing structures is investigated via a nonlinear seismic analysis of the examined case study, calibrated under dynamic recordings of varying response amplitude. While limited to a single building, such changes in the inferred material properties results in a significant reduction of the safety factor, in this case by 14%.

The deformability of floor diaphragms plays a primary role in the structural behavior of existing buildings. Nonetheless, few structural identification procedures are available to investigate this matter from in-situ experimental measurements. Ambient vibration tests can be very useful to the purpose, allowing to assess the importance of the floor deformability in operational modal analyses through model-driven approaches. This information is particularly valuable for unreinforced masonry buildings, often characterized by deformable diaphragms whose effective stiffness is commonly unknown and hard to be evaluated. Based on these motivations, in this paper, a discrete linear model of deformable diaphragm is formulated in a novel fashion. The modal properties governing the free undamped dynamics are analytically determined through a fully general perturbation technique (direct problem). Therefore, a model-based structural identification procedure is proposed to analytically assess the inertial and elastic properties of the deformable diaphragm (inverse problem), assuming the outcomes of experimental modal analyses as known input. Consistently with the perturbation approach, explicit formulas are determined for low-order minimal models and higher-order model updating, accounting for mass and inertial eccentricities. Among the other identifiable mechanical parameters, the focus is put on the first and second-order identification of the in-plane shear stiffness of the diaphragm. The theoretical developments are successfully verified on pseudo-experimental and experimental bases, by applying the identification procedure to (i) the computational model of a prototypical steel frame structure, (ii) the large scale laboratory model of a two-story composite structure with mass eccentricities, (iii) a permanently monitored masonry building recently struck by the 2016-2017 Central Italy earthquake sequence.

Miter gates are water-control structures used as the damming surface on river locks and allow the water levels in the lock to raise or lower as needed. Miter gates have channel-like cross sections and are thus prone to torsional deflection due to gravity loads. To counter-act the tendency for torsional deflection and to add torsional rigidity to the gate, slender steel members termed diagonals are added across the diagonal dimension of the gate and pre-tensioned. To maintain appropriate tension in the diagonals over their lifetime, the tension in the diagonals should be monitored; however, no such monitoring is utilized. Vibration based methods to obtain an estimate of the tensile loads in the diagonal are attractive because they are simple, inexpensive, and do not require continuous monitoring. However, employing vibration-based methods to estimate the tension in the diagonals is particularly challenging because the diagonals are subjected to varying levels of submersion in water. Finding a relationship between the frequency of vibration and applied pretension that adequately addressed the effects of submersion on diagonals is difficult. This paper proposes an approach to account for the effect of submersion on the estimated tension in miter gate diagonals. Laboratory tests are conducted using scale-model diagonal specimens subjected to various levels of tension and submersion in water. The frequency of the diagonal specimens is measured and compared to an approximation using an assumed modes model. The effects of submersion on the frequency of vibration for the partially submerged diagonals are largely explained by added mass on the diagonals. Field validation is performed using a previously developed vision-based method of extracting the frequency of vibration in conjunction with the proposed method of tension estimation of an in-service miter gate diagonal that is also instrumented with load cells. Results for the proposed method show excellent agreement with load cell measurements.

This paper proposes a new methodology to address the image quality problem encountered as the use of an unmanned aerial vehicle (UAV) in the field of bridge inspection increased. When inspecting a bridge, the image obtained from the UAV was degraded by various interference factors such as vibration, wind, and motion of UAV. Image quality degradation such as blur, noise, and low-resolution is a major obstacle in utilizing bridge inspection technology based on UAV. In particular, in the field of bridge inspection where damages must be accurately and quickly detected based on data obtained from UAV, these quality issues weaken the advantage of using UAVs by requiring re-take of images through re-flighting. Therefore, in this study, image quality assessment (IQA) based on local blur map (LBM) and image quality enhancement (IQE) using the variational Dirichlet (VD) kernel estimation were proposed as a solution to address the quality issues. First, image data was collected by setting different camera parameters for each bridge member. Second, a blur map was generated through discrete wavelet transform (DWT) and a new quality metric to measure the degree of blurriness was proposed. Third, for low-quality images with a large degree of blurriness, the blind kernel estimation and blind image deconvolution were performed to enhance the quality of images. In the validation tests, the proposed quality metric was applied to material image sets of bridge pier and deck taken from UAV, and its results were compared with those of other quality metrics based on singular value decomposition (SVD), sum of gray-intensity variance (SGV) and high-frequency multiscale fusion and sort transform (HiFST) methods. It was validated that the proposed IQA metric showed better classification performance on UAV images for bridge inspection through comparison with the classification results by human perception. In addition, by performing IQE, on average, 26% of blur was reduced, and the images with enhanced quality showed better damage detection performance through the deep learning model (i.e., mask and region-based convolutional neural network).

Damages on lights and utility poles mounted on the elevated highway or railway bridges were observed in the past several large earthquakes. The damages could have serious consequences to public safety, travelling vehicles or trains, and nearby properties. A previous study shows that the damages were caused by buckling and yielding of the pole due to excessive response amplification during large earthquake. Such amplification occurs when the bridge's natural frequency is close to the light pole's fundamental frequency. An investigation of the seismic performance of existing light pole mounted on elevated highway bridges is needed to avoid the response amplification. This includes the identification of the light pole's natural frequency and damping ratio. Vibration testing of the light pole using conventional contact sensors individually would require enormous effort and is time-consuming. Moreover, such vibration testing on a highway bridge deck would require traffic disruption to provide access. Video camera-based non-contact vision sensing is seen as a promising alternative to the conventional contact sensors for this purpose. The objective of this paper is to explore the use of non-contact vision sensing for operational modal analysis of light pole on highway viaduct. The phase-based video motion magnification method is implemented to obtain the light pole response in an ambient condition. Using this method, small and invisible displacement is magnified for a certain range of frequency of interest. Based on the magnified video frames, structural displacement is extracted using the image processing technique. The natural frequency and damping ratio of the light pole are estimated using the random decrement technique. The method is verified in a laboratory-scale experiment and implemented to practical field measurements of a light pole on a highway viaduct in Kanagawa, Japan. The results are compared with measurement by Laser Doppler Vibrometer. Both experiments suggest that the method could effectively obtain the natural frequency and damping ratio of the structures under the ambient condition where vibration amplitudes are very small and invisible with reasonable accuracy.

In bridge weigh-in-motion (BWIM), dynamic bridge response is measured during traffic and used to identify overloaded vehicles. Most past studies of BWIM use mechanics-based algorithms to estimate axle weights. This research instead investigates deep learning, specifically the recurrent neural network (RNN), toward BWIM. In order to acquire the large data volume to train a RNN network that uses bridge response to estimate axle weights, a finite element bridge model is built through the commercial software package LS-DYNA. To mimic everyday traffic scenarios, tens of thousands of randomized vehicle formations are simulated, with different combinations of vehicle types, spacings, speeds, axle weights, axle distances, etc. Dynamic response from each of the randomized traffic scenarios is recorded for training the RNN. In this paper we propose a 3-stage Bidirectional RNN toward BWIM. Long short-term memory (LSTM) and attention mechanism are embedded in the BRNN to further improve the network performance. Additional test data indicates that the BRNN network achieves high accuracy in estimating axle weights, in comparison with a conventional moving force identification (MFI) method.

Since it may be hard to obtain the exact external load in practice, damage identification of bridge structures using only structural responses under unknown seismic excitations is an important but challenging task. Since structural responses are determined by both structural properties and seismic excitation, it is necessary to remove the effects of external excitation and only retain the structural information for structural damage identification. In this paper, a data-driven approach using structural responses only is proposed for structural damage alarming and localization of bridge structures. The transmissibility functions (TF) of structural responses are used to eliminate the influence of unknown seismic excitations. Moreover, the inverse Fourier transform of TFs and wavelet packet transform are used to reduce the influence of frequency bands and to extract the damage-sensitive feature, respectively. Based on Support vector machines (SVM), structural responses under ambient excitations are used for training SVM. Then, structural responses under unknown seismic excitations are also processed accordingly and used for damage alarming and localization by the trained SMV. The numerical simulation examples of beam-type bridge and a cable-stayed bridge under unknown seismic excitations are studied to illustrate the performance of the proposed approach.

Condition monitoring of railway tracks is essential in guaranteeing the running safety of railways. Track profiles are the primary source of external excitation for a train system. While Track Recording Vehicle is often utilized for maintenance purposes, this particular vehicle is expensive and difficult to use for small railway operators. Therefore, track profile estimation through in-service vehicle response measurements, which potentially provides efficient and frequent measurement, has been studied. However, the quantitative evaluation of the vertical and lateral track profile irregularities is still challenging as the inverse analysis solutions are sometimes inaccurate and even unstable. In this paper, numerical analyses are first carried out to evaluate track profiles from acceleration and angular velocity responses measured on a train car body. For the inverse analysis, an Augmented State Kalman Filter is utilized to solve the problem using 4 degrees of freedom observable train models. The sensor installation locations are investigated through observability rank condition analysis with different measurement layout. Secondly, a field experiment is carried out in a local Japanese in-service railway network to estimate track profile from car body motions. Smartphones are utilized for the field test measurements as prevalent sensing devices. The effectiveness of the proposed approach is demonstrated with the observable train model. Numerical analyses and field experiments clarify the proposed track profile estimation's capability using only one on-board sensing device.

This work deals with structural analysis and health monitoring (SHM) of a valuable structure of the twentieth-century cultural heritage: the Flaminio Stadium in Rome. The Flaminio is one of the iconic reinforced concrete sport facilities designed and built by Pier Luigi Nervi for the 1960 Olympic Games of Rome. In view of the foreseen SHM activity, the structural analysis of the Flaminio Stadium is firstly reported by presenting either preliminary analyses, aimed at studying the stadium response under different modeling hypotheses, and a three-dimensional Finite Element (FE) model of the entire structure. It turns out that the main grandstand canopy plays a pivotal role in the Flaminio's structural response to seismic excitation; in addition, its state of conservation raises some concern. Therefore, the structural modeling and dynamic characterization of the canopy is deepened in the paper. Its unusual features, such as geometry, material characteristics and dynamic interplay with the hosting main reinforced concrete frames are thoroughly assessed. To validate the FE results, characterized by a high modal density, and investigate the response of the structure, dynamic tests carried out under operating conditions are presented. The output-only collected data are used to calibrate the initial FE model. The predicted static and dynamic responses of the canopy are eventually exploited to guide the design of a tailored monitoring system. The relevant data management is framed in a heritage building information modeling (HBIM) context. This study draws a viable process for a proactive structural conservation strategy of twentieth-century heritage buildings and infrastructures.

One of the most remarkable structural elements characterizing the Milan Cathedral is its Main Spire, built in Candoglia marble and completed in 1769. The Main Spire, reaching the height of about 108 m and supporting the statue of the Virgin Mary, is about 40 m high and stands on the octagonal tiburio erected around the main dome. The structural arrangement of the spire includes a central column which is connected through a spiral staircase to 8 perimeter columns and each column is stiffened by inverse flying buttress. Metallic clamps and dowels connect the marble blocks and metallic rods connect the perimeter columns to the central core. A large monitoring system was recently installed in the Milan Cathedral, including seismometers and temperature sensors at 3 levels of the Main Spire as well as a weather station at the top of the spire. After a concise historic background on the Main Spire and the description of the sensing devices installed in this structure, the paper focuses on the dynamic characteristics of the spire and their evolution during a time span of about 16 months. The presented results highlight that: (a) a high density of vibration modes is automatically detected in the frequency range 1.0-7.0 Hz; (b) the lower identified modes correspond to global modes of the cathedral; (c) the normal evolution in time of the resonant frequencies is characterized by clear fluctuations induced by the environmental effects (temperature and wind); (d) especially the dependence of resonant frequencies on temperature is very distinctive and reveals the key role of the metallic elements in the overall dynamic behavior; (e) notwithstanding the remarkable effects exerted by the changing environment on the resonant frequencies, output-only removal of environmental effects and novelty analysis allow an effective monitoring of the structural condition.

The structural health of masonry towers can be monitored by installing few accelerometers (or seismometers) at the top of the building. This cost-effective setup provides continuous and reliable information on the natural frequencies of the structure and allows to detect the occurrence of structural anomalies; however, to move from anomaly detection to localization with such a simplified distribution of sensors, a calibrated numerical model is needed. The paper summarizes the development of a Structural Health Monitoring (SHM) procedure for the model-based damage assessment in masonry towers using frequency data. The proposed methodology involves the subsequent steps: (i) preliminary analysis including geometric survey and ambient vibration tests; (ii) FE modeling and updating based on the identified modal parameters; (iii) creation of a Damage Location Reference Matrix (DLRM) from numerically simulated damage scenarios; (iv) detection of the onset of damage from the analysis of the continuously collected vibration data, and (v) localization of the anomalies through the comparison between the experimentally identified variations of natural frequencies and the above-defined DLRM matrix. The proposed SHM methodology is exemplified on the ancient Zuccaro tower in Mantua, Italy. Pseudo-experimental monitoring data were generated and employed to assess the reliability of the developed algorithm in identifying the damage location. The results show a promise toward the practical applications of the proposed strategy for the early identification of damage in ancient towers.

This paper proposes a novel method suitable for vibration-based damage identification of civil structures and infrastructures under ambient excitation. The damage-sensitive feature employed in the presented algorithm consists of a vector of multivariate autoregressive parameters estimated from the vibration responses collected at different locations of the analyzed structure. Outlier analysis and statistical pattern recognition are exploited for damage detection and localization. In particular, the Mahalanobis distance between a set of reference (i.e., "healthy") and inspection parameters is evaluated. A threshold is then selected to determine whether the inspection vectors refer to damaged or undamaged conditions. The effectiveness of the proposed approach is proved using numerical simulations and experimental data from a benchmark test. The analysis results show that the largest values of Mahalanobis distance can be found in the proximity of those sensors closest to the damaged elements. Thus, the Mahalanobis distance applied to vectors of multivariate autoregressive parameters has proven to be a robust indicator for damage detection and localization.

Recent developments in the field of smart sensing systems enable performing simple onboard operations which are increasingly used for the decentralization of complex procedures in the context of vibration-based structural health monitoring (SHM). Vibration data collected by multiple sensors are traditionally used to identify damage-sensitive features (DSFs) in a centralized topology. However, dealing with large infrastructures and wireless systems may be challenging due to their limited transmission range and to the energy consumption that increases with the complexity of the sensing network. Local DSFs based on data collected in the vicinity of inspection locations are the key to overcome geometric limits and easily design scalable wireless sensing systems. Furthermore, the onboard pre-processing of the raw data is necessary to reduce the transmission rate and improve the overall efficiency of the network. In this study, an effective method for real-time modal identification is used together with a local approximation of a damage feature, the interpolation error, to detect and localize damage due to a loss of stiffness. The DSF is evaluated using the responses recorded at small groups of sensors organized in a decentralized topology. This enables the onboard damage identification in real time thereby reducing computational effort and memory allocation requirements. Experimental tests conducted using real data confirm the robustness of the proposed method and the potential of its implementation onboard decentralized sensor networks.

The large use of glass in buildings, and especially the presence of fenestrations and facade systems, represents a potential critical issue for people safety. The brittle nature of glass (with limited elastic deformation and resistance) is often enforced by its use in combination of several secondary components, whose reciprocal interaction and potential damage should be properly assessed. In the case of windows, accordingly, a special care should be spent for glass members but also for the framing system and possible adhesive or mechanical connections. This study aims at exploring the dynamic response and damage sensitivity of traditional glass window systems, based on the experimental derivation of few key material properties and mechanical parameters. To this aim the attention is focused on traditional, in-service windows that belongs to existing residential buildings and are typically sustained by timber frames, through a linear flexible connection. In doing so, major advantage is taken from experimental analysis, both in the static and dynamic field, for whole window assemblies of single components. For comparative purposes, selected window specimens including plastic (PVC) frame members and Insulated Glass Units (IGUs) are also taken into account in the paper. The static characteristics of the windows components are first preliminary derived. The dynamic performance of such a kind of systems is then experimentally explored with the support of modal analysis techniques and hard body impact procedures, including the experimental derivation of stiffness parameters for the frame members and the glass panels. Further assessment of experimental outcomes is finally achieved with the support of Finite Element numerical analyses.

In this numerical study, an optimal energetic control model applied to local heating sources to prevent black-ice occurrence at transport infrastructure surface is addressed. The heat transfer Finite Element Model developed and boundary conditions hypothesis considered are firstly presented. Several heat powering strategies, in time and space, are then introduced. Secondly, control laws are presented with the objective of preventing ice formation while avoiding excessive energy consumption by taking also into account weather forecast information. In particular, the adjoint state method is adapted for the case of an operation without some continuous properties (discontinuous time heat sources). In such case, a projection from the space of continuous time functions to a piecewise constant one is proposed. To perform optimal control, the adjoint state method is addressed and discussed for the different powering solutions. To preserve some specific technical components and maintain their lifetime, operational constraints are considered and different formulations for the control law are proposed. Time dependent convecto-radiative boundary conditions are introduced in the model by extracting information from existing weather databases. Extension to updated inline weather forecast services is also presented and discussed. The final minimization problem considered has to act on both energy consumption and non-freezing surface temperature by integrating these specific constraints. As a consequence, the final optimal solution is estimated by an algorithm relying on the combination of adjoint state method and gradient descent that fits mathematical constraints. Results obtained by numerical simulations for different operative conditions with various weather conditions are presented and discussed. Finally, conclusion and perspectives are proposed.

Cable-driven robots are parallel manipulators in which rigid links are replaced by actuated cables. The end-effector is then supported by a set of cables commanded by motors that are usually placed in a fixed frame. By varying the cables length, it is possible to change the end-effector position and/or orientation. Among the advantages presented by cable robots are they light-weight structure, high energy efficiency and their ability to cover large workspaces since cables are easy to wind. When high-speed operation is not required, a safer solution is to design cable-driven suspended robots, where all vertical components of cables tension are against gravity direction. Cable-driven suspended robots present limited workspace due to the elevated torque requirements for the higher part of the workspace. In this paper, the addition of a passive carriage in the top of the frame is proposed, allowing to achieve a much greater feasible workspace than the conventional one, i.e., with the same size as the desired inspection area while maintaining the same motor requirements. In the opposite, this new scheme presents non-desired vibration during the end-effector maneuvers. These vibrations can be removed by means of a more complex control strategy. Kinematics and dynamics models are developed in this paper. An analysis of sensor system is carried out and a control scheme is proposed for controlling the end-effector pose. Simulation and experimental results show that the feasible workspace can be notoriously increased while end-effector pose is controlled. This new architecture of cable-driven robot can be easily applied for automated inspection and monitoring of very large vertical surfaces of civil infrastructures, such as facades or dams.

Inspection and maintenance of civil structures are important issues for sustainability of existing and new infrastructures. Classical approach relies on large human activities eventually performed in unsafe conditions. This paper proposed a non-invasive solution for inspecting horizontal surface such as decks of bridges. The proposal presented here is based in cable-driven robots and allows to inspect large surfaces maintaining a very low vertical occupancy in comparison to the conventional architecture of this kind of robot. Using closed cables loop instead of a set of cables a device with low motorization power and very large workspace is designed and prototyped. As example of control an inverse dynamics technique is applied to control the end-effector where inspection tool is located, e.g., a vision system. Experimental results demonstrate that this novel device allows to inspect large horizontal surfaces, with low motorization and low vertical occupancy.