Onsite systematic monitoring benefits hazard prevention immensely. Hazard identification is usually limited due to the semantic gap. Previous studies that integrate computer vision and ontology can address the semantic gap and detect the onsite hazards. However, extracting and encoding regulatory documents in a computer-processable format often requires manual work which is costly and time-consuming. A novel and universally applicable framework is proposed that integrates computer vision, ontology, and natural language processing to improve systematic safety management, capable of hazard prevention and elimination. Visual relationship detection based on computer vision is used to detect and predict multiple interactions between objects in images, whose relationships are then coded in a three-tuple format because it has abundant expressiveness and is computer-accessible. Subsequently, the concepts of construction safety ontology are presented to address the semantic gap. The results are subsequently recorded into the SWI Prolog, a commonly used tool to run Prolog (programming of logic), as facts and compared with triplet rules extracted from using natural language processing to indicate the potential risks in the ongoing work. The high-performance results of Recall@100 demonstrated that the chosen method can precisely predict the interactions between objects and help to improve onsite hazard identification.

The mechanical characteristics of steel fiber-reinforced lightweight concrete (SFLWC) under high temperatures are studied in this paper. Different concrete matrices, including all-lightweight concrete (ALWC) and semi-lightweight concrete (SLWC), and different steel fibers with hooked ends and crimped shapes are considered as objects. In addition, normal-weight limestone aggregates concrete (NWC), no-fiber ALWC, and SLWC were tested after high-temperature treatment as a control group. The temperature effects on the splitting tensile strength, ultrasonic pulse velocity, compressive stress–strain curve, elastic module, peak strain, and axial compressive strength of the SFLWC were investigated. The results showed that, with increasing exposure temperature, both the axial compressive strength and the elastic modulus decreased, while the axial peak strain has a certain increase, and hence the stress–strain curves were gradually flattened. The toughness of all the concretes increased first and then reduced with increasing temperature, while the specific toughness of all the concretes increased with the increase in temperature. Compared with NWC and SLWC, ALWC had a better capacity to resist high temperatures, especially temperatures > 400 °C. Adding steel fibers can improve the capacity of energy absorption, specific toughness, and residual splitting tensile strength of lightweight concrete (LWC) before and after it is exposed to high temperatures. Based on a regression analysis, a segmented constitutive equation for LWC and SFLWC under uniaxial compression was derived from fitting the experimental findings, and the fitting curve agrees well with the test results.

The poor safety performance of the construction industry is a global concern. Workers’ unsafe behaviors (WUBs) are viewed as critical causes of construction accidents. Thus, it has been a leading research topic to identify antecedents of WUBs and establish reliable models to explain WUBs. Cognitive biases (CBs) are salient antecedents of WUBs. However, the relationships between CBs and WUBs are not further explored in construction safety management. As such, this study selected three prevalent CBs, i.e., availability bias (AB), confirmation bias (FB), and overconfidence, and investigated the relationships between the three CBs and WUBs and the functions of RPs in these relationships among tunnel construction workers. We established six multiple regression models. Research results show that CBs can directly influence RPs and WUBs, RPs are antecedents of WUBs, and the mediating effect of RPs on the relationships between CBs and WUBs is salient. In detail, AB and FB explain more ERP and overconfidence explains more BRP. AB and FB indirectly influence more WUBs through ERP, while overconfidence indirectly affects more WUBs through BRP. The study provides a novel theoretical paradigm for conceptualizing CBs and their relationships with RPs and WUBs. The research results can also guide managers to take debiasing measures to decrease workers’ unsafe behaviors on the construction site.

High-speed railway construction is extending to mountainous areas, and the harsh environment and complex climate pose various risks to the slope construction. This seriously threatens human lives and causes huge economic losses. The existing research results on the construction safety risks of high cutting slope construction in HSRs are limited, and a complete set of safety risk assessment processes and methods has not yet been formed. Therefore, in this study, we aimed to develop a safety risk assessment model, including factor identification and classification and assessment data processing, to help project managers evaluate safety risks in high cutting slope construction. In this study, comprehensive identification of high cutting slope construction safety risks was carried out from three dimensions, risk technical specification, literature analysis, and case statistical analysis, and a list of risk-influencing factors was formed. Based on the historical data, a high side slope risk evaluation model was established using a BP neural network algorithm. The model was applied to the risk evaluation of HF high cutting slopes. The results show that the risk evaluation level is II; the main risks are earthwork excavation method, scaffolding equipment, slope height, slope rate, groundwater, personnel safety awareness, and construction safety risk management system. Finally, a case study was used to verify the proposed model, and control measures for safety risks were proposed. Our findings will help conduct effective safety management, add to the knowledge of construction safety risk management in terms of implementation, and offer lessons and references for future construction safety management of HSR.

Landslides are one of the most widespread natural hazards that cause damage to both property and life every year. Therefore, the landslide susceptibility evaluation is necessary for land hazard assessment and mitigation of landslide-related losses. Selecting an appropriate mapping unit is an essential step for landslide susceptibility evaluation. This study tested the back propagation (BP) neural network technique to develop a landslide susceptibility map in Qingchuan County, Sichuan Province, China. It compared the results of applying six different slope unit scales for landslide susceptibility maps obtained using hydrological analysis. We prepared a dataset comprising 973 historical landslide locations and six conditioning factors (elevation, slope degree, aspect, lithology, distance to fault lines, and distance to drainage network) to construct a geospatial database and divided the data into the training and testing datasets. We based on the BP learning algorithm to generate landslide susceptibility maps using the training dataset. We divided Qingchuan County into six different scales of slope unit: 4,401, 13,146, 39,251, 46,504, 56,570, and 69,013, then calculated the receiver operating characteristic (ROC) curve, and used the area under the curve (AUC) for the quantitative evaluation of 6 different slope unit scales of landslide susceptibility maps using the testing dataset. The verification results indicated that the evaluation generated by 56,570 slope units had the highest accuracy with a ROC curve of 0.9424. Overelaborate and rough division of slope units may not get the best evaluation results, and it is necessary to obtain the slope units most consistent with the actual situation through debugging. The results of this study will be useful for the development of landslide hazard mitigation strategies.

Railway engineering generates large amounts of construction and demolition waste (CDW). To quantify the amount of CDW generated from railway engineering projects throughout the whole life cycle, a process-based life cycle assessment model is proposed in this paper. The life-cycle CDW is divided into four parts: CDW from off-site transportation of construction materials (OSTCM), CDW from site operation wastage of construction materials (SOWCM), discard ballast from roadbeds, stationyard, bridges and tunnels (DB), and CDW from reparation and renewal of aging components (RRAC). Yun-Gui Railway is selected as a case study to validate the developed model, and an uncertainty analysis is conducted with Oracle Crystal Ball software. The results show that between 175 and 311 million tons of CDW is generated throughout the whole life cycle of Yun-Gui Railway. DB is the largest component of the life-cycle CDW from railway engineering projects. This indicates the negative environmental impacts of railway construction can be significantly mitigated by optimizing the location of ballast disposal sites and developing suitable landfill proposals. Also, the CDW generated by wastage of construction materials during off-site construction and site operation is important in waste management in railway engineering projects, in which rubble, sand, and cement have the high potential for waste reduction. Findings from this study can contribute to the knowledge body as well as the engineering practice in green railways.