• Smart Structures and Systems
• /
• 제29권1호
• /
• pp.181-193
• /
• 2022

Data-driven structural health monitoring (SHM) of civil infrastructure can be used to continuously assess the state of a structure, allowing preemptive safety measures to be carried out. Long-term monitoring of large-scale civil infrastructure often involves data-collection using a network of numerous sensors of various types. Malfunctioning sensors in the network are common, which can disrupt the condition assessment and even lead to false-negative indications of damage. The overwhelming size of the data collected renders manual approaches to ensure data quality intractable. The task of detecting and classifying an anomaly in the raw data is non-trivial. We propose an approach to automate this task, improving upon the previously developed technique of image-based pre-processing on one-dimensional (1D) data by enriching the features of the neural network input data with multiple channels. In particular, feature engineering is employed to convert the measured time histories into a 3-channel image comprised of (i) the time history, (ii) the spectrogram, and (iii) the probability density function representation of the signal. To demonstrate this approach, a CNN model is designed and trained on a dataset consisting of acceleration records of sensors installed on a long-span bridge, with the goal of fault detection and classification. The effect of imbalance in anomaly patterns observed is studied to better account for unseen test cases. The proposed framework achieves high overall accuracy and recall even when tested on an unseen dataset that is much larger than the samples used for training, offering a viable solution for implementation on full-scale structures where limited labeled-training data is available.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• 제31권3A호
• /
• pp.235-249
• /
• 2011

A new form of I-type PSC bridge girder, which has hole in the web, is proposed in this paper. Three different concepts were combined and implemented in the design. First of all, a girder was precast at a manufacturing plant as divided pieces and assembled at the construction site using post-tensioning method, and the construction period at the site will be reduced dramatically. In this way, the quality of concrete can be assured at the manufacturing factory and concrete curing can be well controlled, and the spliced girder segments can be moved to the construction site without a transportation problem. Secondly, a numerous number of holes was made in the web of the girder. This reduces the self-weight of the girder. But more important thing related to the holes is that about half of the total anchorages can be moved from the girder ends into individual holes. The magnitude of negative moment developed at girder ends will be reduced. Also, since the longitudinal compressive stresses are reduced at ends, thick end diaphragm is not necessary. Thirdly, Prestressing force was introduced into the member through multiple stages. This concept of multi-stage prestressing method overcomes the prestressing force limit restrained by the allowable stresses at each loading stage, and maximizes the magnitude of applicable prestressing force. It makes the girder longer and shallower. Two 50 meter long full scale girders were fabricated and tested. One of them was non-spliced, or monolithic girder, made as one piece from the beginning, and the other one was assembled using post-tensioning method from five pieces of segments. It was found from the result that monolithic and spliced girder show similar load-deflection relationships and crack patterns. Girders satisfied specific girder design specification in flexural strength, deflection, and live load deflection control limit. Both spliced and monolithic holed web post-tensioned girders can be used to achieve span lengths of more than 50m with the girder height of 2 m.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• 제35권4호
• /
• pp.759-768
• /
• 2015

The ground motions of large dimensional structures such as long span bridges at different stations during an earthquake, are inevitably different, which is known as the ground motion spatial variation effect. There are many causes that may result in the spatial variability in seismic ground motion, e.g., the wave passage effect due to the different arrival times of waves at different locations; the loss of coherency due to seismic waves scattering in the heterogeneous medium of the ground; the site amplification effect owing to different local soil properties. In previous researches, the site amplification effects have not been considered or considered by a single-layered soil model only. In this study, however, the ground motion amplification and filtering effects are evaluated by multi-layered soil model. Spatially varying ground motion at the sites with different number of layers, depths, and soil characteristics are generated and the variation characteristics of ground motion time histories according to the correlation of coherency loss function and soil conditions are evaluated. For the bridge system composed of two unit bridges, seismic behavior characteristics are analyzed using the generated seismic waves as input ground motion. Especially, relative displacement due to coherency loss and site effect which can cause the unseating and pounding between girders are evaluated. As a result, considering the soil conditions of each site are always important and should not be neglected for an accurate structural response analysis.