Journal of Performance of Constructed Facilities

Surface settlement is one of the key engineering issues during shield construction process. In order to accurately predict surface settlement, this paper proposes a new machine learning method based on relevance vector machine (RVM), principal component analysis (PCA), and particle swarm optimization (PSO). Taking Beijing Metro Line 6 as a case study, the PCA-PSO-RVM model is used to make the prediction and compared with the prediction results of the RVM model using the same samples. In order to evaluate the reliability of the model, three evaluation indexes including mean relative error (MRE), root mean square error (RMSE), and Theil inequality coefficient (TIC) were calculated, and sensitivity analysis was carried out on them. The results show that the minimum relative error between PCA-PSO-RVM and the actual value is only 0.06%. The calculated MRE, RMSE, and TIC are 0.17%, 0.0714 mm, and 0.027%, respectively, which shows that PCA-PSO-RVM model has higher prediction accuracy, smaller deviations, and higher reliability compared with the three other models. Through sensitivity analysis, it is found that the weighted average internal friction angle ($\phi$) has the most significant impact on the surface settlement, which should be focused on in relevant research.

To support bridge management systems (BMSs) in establishing strategic maintenance plans to preserve the condition of aging bridges, it is important to estimate a bridge component’s future condition reliably. To this end, many data-driven studies have attempted to apply diverse algorithms and explore major factors influencing the condition of specific components. Despite these efforts, it is still difficult to construct a robust and generally applicable condition estimation model for bridge components regardless of the characteristics of the BMS data because BMS data become heterogeneous and complex by period, region, or country. Therefore, the objective of this study is to develop an optimized condition estimation model for bridge components using data-driven approaches. To achieve the main objective, the proposed model included the following elements: (1) outstanding algorithm selection by comparing the performance of diverse algorithms; and (2) influential variable identification by utilizing the recursive feature elimination (RFE) method according to the permutation variable importance. Based on a case study to estimate the condition grades of decks on concrete-girder bridges using the Korean Bridge Management System (KOBMS) data, extreme gradient boost (XGBoost) was selected as the optimal algorithm, and influential variables were identified such as “bridge age” and “first past condition grade of deck.” Finally, the optimized model based on the integrated results of the algorithm selection and the influential variables identification showed good performance with an average weighted average F1 score of 0.876. The outcome of the research will contribute to reliably estimating the future condition of bridge components by constructing an optimal model suitable for each BMS data and supporting strategic maintenance decisions based on the expected components’ condition for proactive bridge management.

Steel bridge decks are prone to fatigue cracking due to long-term wheel loading with high cyclic stress. Fatigue cracks threaten the safety of the bridge and even cause collapse, making crack detection very important. Crack detection with little influence on traffic is difficult due to the pavement. To confirm the feasibility of eddy current testing (ECT) for steel bridge fatigue cracks, the key parameters were calculated. The validity of the parameters was verified preliminarily using the finite-element method. An eddy current movement detection device was set up, and a prefabricated crack test was conducted. The influences of crack size and barrier thickness on the detection signal were discussed. The results show that the sampled signal of magnetic induction intensity is more obvious than voltage, and can be used for crack detection. When the direction of a crack is perpendicular to the direction of detected motion, the amplitude of magnetic induction intensity is positively correlated with the width and depth. Crack detection with a 20-mm-thick barrier layer is possible. The barrier layers weaken the strength of the sampled signal, but the variation law of the intensity with crack parameters is consistent with that of no barrier.

High-early-strength (HES) concrete with polypropylene fibers has been used in several parts of bridges with precast concrete girders by the Idaho Transportation Department (ITD). This material is only slightly more expensive than conventional concrete, but speeds up the construction. ITD is interested in the suitability of HES concrete in the longitudinal closure pour connections between bridge deck bulb-T girders. In the laboratory phase of this project, an optimum HES concrete mix was identified through a series of laboratory experiments. The field performance of the HES concrete mix was examined by placing the material in the closure pour joints between deck bulb-T girders in a bridge in Idaho. The main focus of this paper is to report on the 20-month field performance of the bridge closure pour concrete. Tasks included six visits to the bridge site every 4 months to (1) measure strain data under ITD’s Under the Bridge Inspection Truck (UBIT), (2) measure strain data under commercial truck traffic, and (3) inspect the closure pour concrete. Based on the results, it was concluded that there were no changes in the closure pour concrete in the last 20 months.

Fire imposes negative effects on reinforced concrete (RC) members, shifting postfire structures onto dangerous situations. This study aims at carbon fiber-reinforced polymer (CFRP) retrofit of postfire RC beams to recover their mechanical properties, providing technical information for structural engineers in practice. Toward this aim, experiments were performed for 16 RC beams, which were divided into four groups exposed to 30, 45, 60, and 75 min of ISO 834 fire. After exposure to fire, these beams were strengthened by external bonded (EB) and near-surface mounted (NSM) techniques using a similar amount of CFRP for comparison. These specimens were then loaded until failure. The results are compared with one another and compared with control specimens. It is found that both EB and NSM CFRP retrofitted beams failed in the form of peeling off of concrete, whereas no rupture of CFRP was observed. CFRP retrofit significantly changes the behavior of postfire RC beams, being brittle or exhibiting low ductility. The effectiveness of EB and NSM CFRP techniques is similar and it decreases with the increase in fire duration. CFRP retrofit increases the yield load-carrying capacity by $27.5\%-40.9\%$ for $30-60‐\min$ postfire RC beams and recovers the yield load-carrying capacity for 75-min postfire RC beams. CFRP retrofit significantly increases the ultimate load-carrying capacity, whereas its increasing levels depend not only on the CFRP retrofit but also on the fire duration. Both CFRP retrofitting techniques successfully recovered the yield stiffness of postfire RC beams. The moment capacity of CFRP retrofitted postfire RC beams was theoretically analyzed, and a prediction model was proposed. The proposed model is simple, but it accurately predicts the moment capacity of retrofitted postfire beams; thus, it can be useful for structural engineers in practice.

Concrete cracks that cause durability problems and bearing capacity failure are critical for evaluating the in-service performance of concrete bridges. It is still challenging for deep-learning methods to accurately detect and quantify concrete cracks. To improve detection precision, a computer-vision (CV)-based crack detection framework is presented in this study. The proposed framework introduces redesigned deep convolutional generative adversarial networks (DCGANs) and extends the dataset of concrete crack images by generating synthetic examples from collected crack images. Based on the enlarged dataset and the you-only-look-once version 5s (YOLOv5s) algorithm, model training and crack detection are conducted through backbone, bottleneck, and prediction part. Subsequently, two algorithms, including the Ostu method and the medial axis algorithm, are combined to calculate the length and width of concrete cracks. Experimental tests compared the YOLOv5s model with YOLO series algorithms, region-based fast convolutional neural network (faster R-CNN), and single shot multibox detector (SSD). The dataset augmentation by redesigned DCGANs increased mean average precision (mAP) by 3.7%, and YOLOv5s outperformed in detection speed with $43.5\mathrm{frames}/\mathrm{s}$. The crack size measurement of concrete members in the laboratory demonstrates that the calculation of crack size is achieved at pixel level. The proposed concrete crack detection framework can meet precision requirements and provide a promising measure for patrol inspection.

Abstract

Ensuring reliable facility services with aging infrastructure and growing deferred maintenance backlogs in water and fuel systems requires properly managing the maintenance of these critical systems. Dormitories safeguard the comfort and health of their occupants, but they rely on proactive management of and response to infrastructure failures. This study employed a data and text-mining approach to analyze 43,704 water and fuel work orders queried from a military maintenance database to identify the most common and costly failure types and the components that most frequently fail in dormitories. Statistical and correlation analysis indicated that leaks, clogs, and temperature regulation are problems that cost the most and occur most frequently, and toilets, sinks, and showers are the components that most often need repair. This study provides an intellectual framework for analyzing past maintenance work orders to reveal trends in failure types and components. The study contributes to the practice of infrastructure management by revealing that design and maintenance management should focus on preventing and repairing these common and costly failures.

Practical Applications

This study provides infrastructure maintenance managers with an objective understanding of the most common failure types and associated costs for water, sewer, and fuel supply infrastructure in dormitories to better inform resourcing decisions. By using readily available data from a work order database and incorporating data and text-mining techniques, this research highlights for maintenance managers the components that fail most often, the type of failure they experience, and costs associated with the repair of these failures. This research found that leaks, clogs, and problems with temperature regulation are the most common problems, and they occur most frequently in toilets, sinks, and showers in priority order. This research found that 74% of costs are associated with these common failure modes. Understanding common failure modes allows maintenance managers to efficiently determine parts on-hand levels, the composition of repair crews, and expected costs over a budget cycle. These findings allow maintenance managers to make better-informed budget and resourcing decisions using objectively identified maintenance trends so that they do not have to rely on experience or gut feelings.

In soft-soil areas, excessive transverse deformations have occurred in numerous shield tunnel linings accompanied by serious structural defects. However, to date, a method to effectively diagnose the health status of these poor-performing structures is unknown. This study focuses on health diagnosis for the typical stagger-joint-assembled shield tunnel linings in soft-soil areas of China. First, based on theoretical analyses, it is assumed that diametral deformation can be a significant indicator for the structural performance of a lining; subsequently, this is verified by abundant field data. In particular, a strong positive correlation between the diametral deformation and development of defects is observed. Subsequently, a full-scale test is performed to reveal further quantitative correlations. Besides the service limit values determined in the range of 0.24%–0.79% by specifications and prior studies, the test reveals detailed structural performance deterioration when the deformation is getting larger, where 1.5% is considered a threshold for the failure the entire structure. Additionally, the values of 0.5%, 0.8%, and 1.1% are observed to be key points indicating the gradual degradation of waterproofing, cracks, and overall mechanical performance of the lining. Based on these key points, a five-level health-degree model is constructed to determine the health status of a ring that can be regarded as a supplement of the previous limit values, which is more detailed and practical for the common large-deformed in-service linings in soft-soil areas. These findings can also serve as benchmarks for related studies on linings with different structural parameters in the future.

Demands to incorporate blast loading in building structural designs have increased recently. Focus has been placed on far-field detonations based on large charge weights in the development of approaches to blast mitigation designs. However, more recently, frequent attacks using lighter and portable devices detonated very closely to targets have occurred. There are increasing concerns about the destruction of a few critical structural members due to near-field detonations, which potentially can lead to a more substantial collapse of a building. Although there are approaches to analyzing structural frames after removal of a few columns, there currently are no design codes or specifications to determine whether the columns actually would collapse when subjected to near-field detonations. Therefore this study identified the failure or damage characteristics of standard wide-flange steel columns subjected to near-field detonations, performed a series of numerical analyses to develop simple design tools and charts, and provides details of nonlinear explicit finite-element analysis that can be followed by practicing engineers. A total of 369 numerical models were built to cover a wide variety of standard wide-flange steel columns subjected to a range of portable charge weights detonated at realistically close proximity with and without service loads. The developed design charts can be used by practicing engineers for preliminary design, cost-estimate, or quality control purposes prior to performing detailed finite-element analyses. The numerical models used to develop these design charts were verified using experimental results available in the literature.

The propagation of construction vibration to nearby buildings is a highly complex process. Real-time vibration monitoring is a straightforward means to minimize vibration impact on nearby buildings. Although traditional monitoring systems are typically designed for structural safety, safeguarding the normal operation of vibration-sensitive ultraprecision equipment represents an emerging but increasing need. By developing a tailor-made vibration monitoring system for a hospital, this paper presents the monitoring and assessment of vibration impact on ultraprecision equipment throughout the whole construction period of a hospital expansion project, which, to the best of the authors’ knowledge, has not been reported in the literature. The developed system could monitor construction-induced vibration levels, assess the impact on ultraprecision equipment, and issue alarm signals, if necessary, in a continuous and real-time manner. The application of the developed system in a hospital expansion project together with vibration mitigation measures guaranteed the normal operation of ultraprecision equipment throughout the whole construction period. Recorded vibration data were processed and analyzed to evaluate the vibration impact in different construction stages. The empirical prediction formulas for rotary minipiling and demolition work based on statistical analysis results can provide valuable guidance on corresponding construction operations to mitigate excessive environmental impact.