Health monitoring of civil infrastructures is a key application of Internet of things (IoT), while edge computing is an important component of IoT. In this context, swarms of autonomous inspection robots, which can replace current manual inspections, are examples of edge devices. Incorporation of pretrained deep learning algorithms into these robots for autonomous damage detection is a challenging problem since these devices are typically limited in computing and memory resources. This study introduces a solution based on network pruning using Taylor expansion to utilize pretrained deep convolutional neural networks for efficient edge computing and incorporation into inspection robots. Results from comprehensive experiments on two pretrained networks (i.e., VGG16 and ResNet18) and two types of prevalent surface defects (i.e., crack and corrosion) are presented and discussed in detail with respect to performance, memory demands, and the inference time for damage detection. It is shown that the proposed approach significantly enhances resource efficiency without decreasing damage detection performance. 相似文献
Precast concrete structures are increasingly being adopted by building designers in regions of high seismicity. An unbonded posttensioned (PT) precast split shear wall system (UPPSSW) was proposed by the Precast Seismic Structural Systems (PRESSS). The UPPSSW system is composed of two or more single precast concrete wall panels that are connected together with energy‐dissipating shear connectors and anchored to the foundation with unbonded PT tendons located at the panel center. In this paper, an optimum design program has been developed for designing this system. The objective of the optimum process is to find the optimum combination between PT tendons and shear connectors while keeping the moment capacity of the wall equal to the applied design moment and achieving zero residual drift simultaneously. In addition, MATLAB was employed to explore an optimization program using genetic algorithm. Compared with the existing design methods for the system, the optimum design program proposed in this research is accurate, efficient, and direct. Moreover, it can yield the optimum design automatically and quickly. As a result, the existing lengthy and manual design process of trial and error for the system can be avoided. 相似文献
This paper proposed a new Q690 circular high‐strength concrete‐filled thin‐walled steel tubular (HCFTST) column comprising an ultrahigh‐strength steel tube (yield strength fy ≥ 690 MPa). A quasi‐static cyclic loading test was conducted to examine the seismic behavior, and the obtained lateral load‐displacement hysteresis curves, skeleton curves, and ductility were analyzed in detail. Then, a numerical model based on a nonlinear fiber beam‐column element incorporating the modified uniaxial cyclic constitutive laws for concrete and steel was developed mainly to predict the seismic behavior of the tested Q690 circular HCFTST columns. The effects of the concrete cylinder compressive strength (fc), steel yield strength (fy), axial compression ratio (n), and diameter‐to‐thickness (D/t) ratio on the seismic behavior were investigated through a parametric study. Finally, a simplified hysteretic model incorporating the moment‐resisting capacity and deterioration of the unloading stiffness in addition to a normalized skeleton curve and hysteretic criterion was established. The results indicate the following: the proposed Q690 circular HCFTST columns can display reasonable hysteretic behaviors to some extent; the use of high‐strength steel can lead to a significantly larger elasto‐plastic deformation capacity and delay the appearance of post‐peak behavior, even if a lower ductility capacity is provided; moderately loosening the limitations on the D/t ratio can also result in ideal hysteretic behaviors; and the established numerical model and simplified hysteretic model can satisfactorily predict the experimentally observed load‐displacement hysteretic curves, including the deterioration of the strength and stiffness and can, thus, offer design references for the elasto‐plastic analysis of circular HCFTST columns. 相似文献
The peak shear strength of discontinuities between two different rock types is essential to evaluate the stability of a rock slope with interlayered rocks. However, current research has paid little attention to shear strength parameters of discontinuities with different joint wall compressive strength (DDJCS). In this paper, a neural network methodology was used to predict the peak shear strength of DDJCS considering the effect of joint wall strength combination, normal stress and joint roughness. The database was developed by laboratory direct shear tests on artificial joint specimens with seven different joint wall strength combinations, four designed joint surface topographies and six types of normal stresses. A part of the experimental data was used to train a back-propagation neural network model with a single-hidden layer. The remaining experimental data was used to validate the trained neural network model. The best geometry of the neural network model was determined by the trial-and-error method. For the same data, multivariate regression analysis was also conducted to predict the peak shear strength of DDJCS. Prediction precision of the neural network model and multivariate regression model was evaluated by comparing the predicted peak shear strength of DDJCS with experimental data. The results showed that the capability of the developed neural network model was strong and better than the multivariate regression model. Finally, the established neural network model was applied in the stability evaluation of a typical rock slope with DDJCS as the critical surface in the Badong formation of China.
Bulletin of Engineering Geology and the Environment - Rainfall-induced slope failures in natural terrains are destructive natural disasters. Transport of fine particles may be induced by the... 相似文献
Bulletin of Engineering Geology and the Environment - Since rock mass in many fields of rock engineering usually undergoes a cyclic heating and cooling process, it is very meaningful to investigate... 相似文献
For fatigue damage prognosis of a long-span steel bridge, the dynamic stress analysis of critical structural components of the bridge under the future dynamic vehicle loading is essential. This paper thus presents a framework of dynamic stress analysis for fatigue damage prognosis of long-span steel bridges under the future dynamic vehicle loading. The multi-scale finite element (FE) model of the bridge is first developed using shell/plate elements to simulate the critical structural components (local models) and using beam/truss elements to simulate the rest part of the bridge (global model). With the appropriate coupling of the global and local models, the multi-scale FE model can accurately capture simultaneously not only the global behavior in terms of displacement and acceleration but also the local behavior in terms of stress and strain. A vehicle traffic load model is then developed for forecasting the future vehicle loading based on the recorded weigh-in-motion (WIM) data and using the agent-based traffic flow microsimulation. The forecasted future vehicle loading is finally applied on the multi-scale model of a real long-span cable-stayed bridge for dynamic stress analysis and fatigue damage prognosis. The obtained results show that the proposed framework is effective and accurate for dynamic stress analysis and fatigue damage prognosis. 相似文献