Marco Torbol

Fragility Analysis of Highway Bridges Based on Long‐Term Monitoring Data

This article describes how fragility curves are used to represent the vulnerability of a bridge in seismic regions of highway transportation networks. Because these networks have hundreds or thousands of bridges, it is impossible to study each individual bridge, so bridges with similar properties are grouped together and are represented by the same fragility curve. However, this approach may be inadequate at times for different reasons because bridges with similar geometrical and material properties could have different ages and could deteriorate at different rates. Moreover, certain bridges are unique such as a cable stayed bridge or a suspension bridge. Fragility curves are calculated based not only on the geometry and material properties, but also on vibration data recorded by a structural health monitoring system. The fragility curves are used to track changes of the structural parameters of a bridge throughout its service life. Based on vibration data the fragility curves are updated reflecting a change in structural parameters. Fragility curves based on vibration data, whenever these are available, represent the vulnerability of a bridge with greater accuracy than fragility curves based only on the geometry and material properties. This article demonstrates the applicability of structural health monitoring to generate more reliable fragility curves. This is useful not only for bridges that are unique, which are usually the first to be instrumented, but for every instrumented bridge as well.

Real‐Time Frequency‐Domain Decomposition for Structural Health Monitoring Using General‐Purpose Graphic Processing Unit

In order to analyze the data output of structural health monitoring (SHM) systems, civil engineers use frequency-domain decomposition (FDD) to identify the modal properties of structures. FDD is computationally expensive and it prevents central processing units (CPUs) from achieving real-time performance. The article discusses how SHM systems are becoming larger and shows how a CPU takes seconds to perform FDD of 16 input signals but minutes to perform FDD of hundreds of input signals. A supercomputer can achieve real-time performance but it cannot be installed near a civil structure because it is bulky, expensive, and requires constant maintenance. FDD is performed using a general-purpose graphic processor unit (GPGPU) because a graphic processor unit (GPU) is capable of massive parallel computing. A GPU is energy efficient and does not require the maintenance of a supercomputer and can be installed inside a base station at a structure site. The developed parallel FDD algorithm is up to hundreds of times faster than its serial version on CPU. For SHM of civil structures, where natural frequencies are less than 20 Hz parallel FDD a single GPU achieves real-time performance and the use of GPGPU offers many advantages because the modal properties are tracked in real time.

The directionality effect in the seismic risk assessment of highway networks

This study introduces the directionality effect of the ground motion in the probabilistic seismic risk assessment (PSRA) of lifeline systems. Given an earthquake scenario, the seismic wave strikes each component of the system with a different angle. The angle may vary significantly depending on the shape, the location and the orientation of the structure. An appropriate example of a lifeline system is a highway transportation network, in which under earthquake conditions the bridges are considered the most vulnerable components. The proposed PSRA model requires that the seismic fragility model is a function of a ground motion intensity measure (IM), as in the traditional risk analysis, and the angle of seismic incidence. The model was implemented in a new framework for the PSRA of highway transportation network. In addition, the framework includes new algorithms. One reduces the confidence interval of the results and one increases the computational efficiency. The example used is the highway transportation network serving the Los Angeles area, which has more than 3000 bridges. The results show a considerable difference in the system resilience with or without the seismic directionality taken into consideration. This is important for benefit/cost analysis and it represents a clear departure from the current risk analysis.

Real-time remote monitoring: the DuraMote platform and experiments towards future, advanced, large-scale SCADA systems

The economic and social prosperity of our society depends much on the proper functioning of structures and civil infrastructure systems. Structural health monitoring has been long recognised as a vital tool to preclude and/or mitigate degradation effects and failures of structural systems. Along this line, the DuraMote platform is presented in this paper (named after Durable and Mote) together with real-life applications, laboratory and field experiments, which promote the effort to expand the existing concept of structural monitoring into remote, real-time, continuous and permanent performance monitoring of spatially extended systems. Successful implementation of this technology can improve the resilience and sustainability of large-scale complex infrastructure systems and lead to future, advanced Supervisory Control and Data Acquisition methodology that could be routinely used in a variety of structures and networks.

Vision-Based Natural Frequency Identification Using Laser Speckle Imaging and Parallel Computing

This study focuses on the identification of the natural frequencies of structures through the analysis of the speckle pattern that a laser creates and a camera records. The laser pointer spreads its light over a target area on the structure and creates the speckle pattern. The ambient vibrations affect the pattern and a camera records the changes. The stream of images is fed into a graphics processing unit (GPU). The developed parallel code includes different algorithms: the speckle contrast image (SCI), the speckle flow imaging (SFI), and an innovative application of k-means clustering that uses the geometrical centroid of each cluster as virtual sensors. The tracking of the centroid in time domain through the images creates a vibration signal. The signals from different clusters are processed together to extract the natural frequencies of the structure. This study applies the proposed method to different sample structures both in laboratory and in the field to demonstrate how the obtained signals are reliable and easy to handle. The GPU technology enhances the performance of the entire method and allows the achievement of real-time processing. All these features create an inexpensive, portable, and efficient tool to inspect any structure or its components.

Residual capacity estimation of bridges using structural health monitoring data

In this study, a vibration-based procedure for residual capacity estimation of bridges after damaging earthquake events is proposed. The procedure starts with estimation of collapse capacity of the intact bridge using incremental dynamic analysis (IDA) curves. The collapse capacity is defined as the median intensity level of the earthquakes that cause global or local collapse within the structure. A database of post-earthquake modal properties is created by calculating the analytical modal properties of the bridge after each nonlinear response history analysis performed for generation IDA curves. After the damaging event, experimental modal properties of the bridge are identified from vibration measurements of the bridge. These properties along with the modal properties database are used to find ground motionintensity pairs that can drive nonlinear FE model of the structure to the current damage state of the bridge. The IDA curves corresponding to the damaged FE model of the bridge are subsequently used to estimate amount of loss in collapse capacity of the damaged structure. Estimated loss in capacity of the bridge besides the bridge-site-specific seismic hazard curves are used to update the functionality status of the bridge. Proposed procedure is applied to experimental data from a large-scale shake table test on a quarter-scale model of a short-span reinforced concrete bridge. The bridge was subjected to a series of earthquake ground motions introducing progressive seismic damage to the bridge which finally led to the failure of one of the bents. Residual collapse capacity and functionality status of the bridge are updated at different stages of the experiment using the proposed procedure.

Hybrid networking sensing system for structural health monitoring of a concrete cable-stayed bridge

The purpose of this study is the remote structural health monitoring to identify the torsional natural frequencies and mode shapes of a concrete cable-stayed bridge using a hybrid networking sensing system. The system consists of one data aggregation unit, which is daisy-chained to one or more sensing nodes. A wireless interface is used between the data aggregation units, whereas a wired interface is used between a data aggregation unit and the sensing nodes. Each sensing node is equipped with high-precision MEMS accelerometers with adjustable sampling frequency from 0.2 Hz to 1.2 kHz. The entire system was installed inside the reinforced concrete box-girder deck of Hwamyung Bridge, which is a cable stayed bridge in Busan, South Korea, to protect the system from the harsh environmental conditions. This deployment makes wireless communication a challenge due to the signal losses and the high levels of attenuation. To address these issues, the concept of hybrid networking system is introduced with the efficient local power distribution technique. The theoretical communication range of Wi-Fi is 100m. However, inside the concrete girder, the peer to peer wireless communication cannot exceed about 20m. The distance is further reduced by the line of sight between the antennas. However, the wired daisy-chained connection between sensing nodes is useful because the data aggregation unit can be placed in the optimal location for transmission. To overcome the limitation of the wireless communication range, we adopt a high-gain antenna that extends the wireless communication distance to 50m. Additional help is given by the multi-hopping data communication protocol. The 4G modem, which allows remote access to the system, is the only component exposed to the external environment.

Machine learning and digital image processing for non-contact modal parameters identification of structures

This study introduces an innovative non-contact sensing technique for vision-based displacement measurement. Existing vision-based displacement measurement techniques utilizes physical target panels or physical features to compute relative displacement between the target and the observation point. Instead, the proposed method exploits the optical reference of a speckle pattern. A coherent light that is diffusely reflected on the surface of the target structure creates the speckle pattern. In this study, a camera records the changes in the speckle pattern in real time. Because the speckle pattern is sensitive to small changes of surface, the ambient vibration is enough to affect it. To estimate the displacement of the target from the raw speckle images, speckle contrast imaging (SCI), speckle flow imaging (SFI), and k-means clustering algorithm were used. After SCI and SFI quantifies the blurring effect in each image, the k-means clustering algorithm creates virtual sensing node from each image. The connection of virtual nodes from frame to frame highlights the displacements of the surface in time domain. Because the algorithms are time-consuming and computationally intensive, a GPU executes the entire post-processing operation in parallel and identifies the natural frequencies of the structure.

Visual based laser speckle pattern recognition method for structural health monitoring

This study performed the system identification of a target structure by analyzing the laser speckle pattern taken by a camera. The laser speckle pattern is generated by the diffuse reflection of the laser beam on a rough surface of the target structure. The camera, equipped with a red filter, records the scattered speckle particles of the laser light in real time and the raw speckle image of the pixel data is fed to the graphic processing unit (GPU) in the system. The algorithm for laser speckle contrast analysis (LASCA) computes: the laser speckle contrast images and the laser speckle flow images. The k-mean clustering algorithm is used to classify the pixels in each frame and the clusters’ centroids, which function as virtual sensors, track the displacement between different frames in time domain. The fast Fourier transform (FFT) and the frequency domain decomposition (FDD) compute the modal properties of the structure: natural frequencies and damping ratios. This study takes advantage of the large scale computational capability of GPU. The algorithm is written in Compute Unifies Device Architecture (CUDA C) that allows the processing of speckle images in real time.

Dof-based submatrix scaling factors for damage detection in reinforced concrete bridges

This study focuses on the system identification and the damage detection of reinforced concrete bridges using neural network algorithm, eigenvalue analysis and parallel computing. First, autoregressive coefficients (ARCs) of both temporal output and forced input of the real structure are computed. The ARCs are used for the eigen-system realization algorithm (ERA) to obtain the modal parameters of the structure. Second, the ARCs are utilized as the input variable of the neural network algorithm while the outputs are the submatrix scaling factors that contain information about the degeneration of each element and each mode within the element. However, the neural network algorithm requires training to output reliable results. The training is the most challenging task of this study and finite element analysis is used to compute the modal parameters of the model built around the neural network outputs. The model is compared with the ERA results to update the neural network coefficients. Due to the scale of the neural network used parallel computing is necessary to reduce the computational time to a reasonable amount.