1,720,972 research outputs found
Civil infrastructure defect assessment using pixel-wise segmentation based on deep learning
Full text linkNowadays, the number of aging civil infrastructures is growing world-wide and when concrete is involved, cracking and delamination can occur. Therefore, ensuring the safety and serviceability of existing civil infrastructure and preventing an inadequate level of damage have become some of the major issues in civil engineering field. Routine inspections and maintenance are then required to avoid leaving these defects unexplored and untreated. However, due to the limitations of on-field inspection resources and budget management efficiency, automation technology is needed to develop more effective and pervasive inspection processes. This paper presents a pixel-wise classification method to automatically detect and quantify concrete defects from images through semantic segmentation network. The proposed model uses Deeplabv3+ network with weights initialized from pre-trained neural networks. The comparison study among the performance of different deep neural network models resulted in ResNet-50 as the most suitable network for applications of civil infrastructure defects segmentation. A total of 1250 images have been collected from the Internet, on-field bridge inspections and Google Street View in order to build an invariant network for different resolutions, image qualities and backgrounds. A randomized data augmentation allowed to double the database and assign 2000 images for training and 500 images for validation. The experimental results show global accuracies for training and validation of 93.42% and 91.04%, respectively. The promising results highlighted the suitability of the model to be integrated in digitalized management system to increase the productivity of management agencies involved in civil infrastructure inspections and digital transformation
Automated classification of civil structures defects based on Convolutional Neural Network
Full text linkToday, the most used method for civil infrastructure inspection is based on visual assessment performed by certified inspectors following prescribed protocols. However, the increase in aggressive environmental and load conditions, coupled with the achievement for many structures of the end life-cycle, highlighted the need to automate damage identification to satisfy the number of structures that need to be inspected. To overcome this challenge, the current paper presents a method to automate the concrete damage classification using a deep Convolutional Neural Network (CNN). The CNN is designed after an experimental investigation among a wide number of pretrained networks, all applying the transfer learning technique. Training and Validation are performed using a built database with 1352 images balanced between “undamaged”, “cracked”, and “delaminated” concrete surface.
To increase the network robustness compared to images with real-world situations, different
configurations of images has been collected from Internet and on-field bridge inspections.
The GoogLeNet model is selected as the most suitable network for the concrete damage
classification, having the highest validation accuracy of about 94%. The results confirm that
the proposed model can correctly classify images from real concrete surface of bridges, tunnel
and pavement, resulting an effective alternative to the current visual inspection
Structural Assessment of 50-year-old Prestressed Concrete Bridge Girders
Full text linkL'abstract è presente nell'allegato / the abstract is in the attachmen
A new approach for displacement and stress monitoring of tunnel based on iFEM methodology
Full text linkStructural monitoring plays a key role for underground structures such as tunnels. Strain readings are expected to report structural conditions during construction and at the final delivery of the works. Furthermore, it is increasingly requested an extension to long-term monitoring from contractors with possible use of the same system in service during construction. A robust and efficient monitoring methodology from discrete strain measurements is the inverse finite element method (iFEM), which allows to reconstruct the structural response without input data on the load pattern applied to the structure as well as material and inertial properties of the elements and therefore it is interesting for structural configurations affected by uncertain loading conditions, such as the tunnel. The formulation presented in this paper, based on the iFEM theory, is improved from the previous work available in literature for both the shape functions used and the computational procedure. Indeed, the approach allows to overcome inconsistencies related to structural loading conditions and a pseudo-inverse matrix preserve all the rigid body modes without imposing specific constraints which is typical for tunnels. Numerical validation of the iFEM procedure is performed by simulating the input data coming from a tunnel working in a heterogeneous soil under different loading conditions with direct FEM analysis
Shape sensing with inverse Finite Element Method for slender structures
No full textThe methodology known as "shape sensing" allows the reconstruction of the displacement field of a structure starting from strain measurements, with considerable implications for structural monitoring, as well as for the control and implementation of smart structures. An approach to shape sensing is based on the inverse Finite Element Method (iFEM) that uses a variational principle enforcing a least-squares compatibility between measured and analytical strain measures. The structural response is reconstructed without the knowledge of the mechanical properties and load conditions but based only on the relationship between displacements and strains. In order to efficiently apply iFEM to the most common structural typologies of civil engineering, its formulation according to the kinematical assumptions of the Bernoulli-Euler theory is presented. Two beam inverse finite elements are formulated for different loading conditions. Depending on the type of element, the relationship between the minimum number of required measurement stations and the interpolation order is defined. Several examples representing common applications of civil engineering and involving beams and frames are presented. To simulate the experimental strain data at the station points and to verify the accuracy of the displacements obtained with the iFEM shape sensing procedure, a direct FEM analysis of the considered structures is performed using the LUSAS software
Dynamic response of damaged precast bridge girders
Full text linkThis paper presents a comprehensive dynamic test campaign conducted on prestressed concrete bridge beams sourced from a 50-year-old decommissioned viaduct in Turin, Italy, as part of the BRIDGE 50 research project. Individual beams were subjected to dynamic testing to assess the impact of varying levels of damage on their dynamic properties. Vibration data was collected before applying static loads, after reaching the first cracking condition, and after reaching maximum load capacity, and analysed to identify principal modal components. The findings underscore the correlation between damage progression and dynamic response, demonstrating the effectiveness of vibration tests in detecting and tracking damage evolution. Specifically, this study presents the experimental results pertaining to the tested I-shape beams
Automated corrosion surface quantification in steel transmission towers using UAV photogrammetry and deep convolutional neural networks
Full text linkCorrosion in steel transmission towers poses a challenge to structural integrity and safety, requiring efficient detection methods. Traditional visual inspections are unsustainable due to the complexity and volume of structures. Their manual, qualitative, and subjective nature often leads to inconsistencies in maintenance planning. This study proposes a deep learning-based approach for semantic segmentation of corroded areas on steel towers. Using the DeepLabv3+ model, the network was trained and validated on 999 field photographs. MobileNetV2, serving as the feature extractor, was chosen for its optimal balance between accuracy and computational efficiency, achieving a validation accuracy of 90.8% and a loss of 0.23. The trained network was applied to real-world inspections using orthomosaics derived from photogrammetric reconstructions of the South-East tower at the Torino Eremo broadcasting center. These photogrammetric products not only enabled precise segmentation of corroded areas but also provided the foundation for corrosion quantification with metrical accuracy, a critical advantage for maintenance planning. Unlike traditional image segmentation methods, which lack a spatial reference and precise scaling, the photogrammetric approach ensures that the corrosion extent and distribution are quantified in exact physical dimensions, enhancing the reliability of the analysis. The results show that deep learning-based inspections can automate detection, providing reliable data and reducing reliance on manual inspections, enhancing efficiency, safety, and accuracy
Dynamic response of PC bridge beams under different damages
Full text linkThe present paper describes the dynamic test campaign on prestressed concrete bridge beams taken from a dismantled viaduct in Turin, Italy after a service life of 50 years in the framework of BRIDGE|50 research project. Dynamic measurements were previously performed on the decks from which the 29 beams were taken to characterize the behaviour of the viaduct in service condition. Successively the single beams are tested to analyse and evaluate the effects of the different damage levels on the dynamic properties. The vibration data have been collected before the application of static load, after the first cracking condition and after the maximum load applied on the beam to extract the principal modal components. The results highlight the correlation among the evolution of the damage and the dynamic response of the beam and then the effectiveness of vibration tests to identify the occurrence of damages and follow their evolution. The experimental findings could be used in future works to explore the effects of damages of the single beams on the global response of this bridge typology. This work presents the results of the experimental tests on the first eight beams tested
Experimental tests for mechanical characterization of prestressed concrete bridge deck beams
Full text linkThis paper presents the preliminary results of extensive experimental activities for mechanical characterization of 50-year-old prestressed concrete bridge deck beams within the BRIDGE|50 research project (http://www.bridge50.org). The experimental campaign includes non-destructive diagnostic tests (e.g. sclerometer and ultrasonic tests) carried out on several prestressed concrete deck beams and laboratory mechanical tests on concrete cores extracted from both precast beam and cast-in-situ slab. The results of these activities have been used for calibration and validation of structural analysis models and to support a proper planning of full-scale load tests up to collapse to be performed on the deck beams
- …
