ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems, Part A: Civil Engineering

From the literature, time domain and modal domain data are commonly used in system identification and damage detection of various systems. This paper focuses on the comparison between time and modal domain system identification of a rail–sleeper–ballast system, which is modeled with the beam-on-springs theory. Linear elasticity is assumed in modal domain analyses, while the ballast layer is considered elastoplastic, in line with the behavior of ballast under large amplitude vibration in time domain analyses. A simple nonlinear model—the modified Ludwik model—was utilized to capture the strain-hardening behavior of ballast in the tensionless ballast springs. An enhanced Markov chain Monte Carlo (MCMC)-based Bayesian algorithm is utilized to handle the uncertainties associated with the identified system parameters from a probabilistic sense. This algorithm caters for cases that are unidentifiable and where the posterior probability density functions (PDF) are possibly nonGaussian. System identification was carried out using measured data obtained from impact hammer tests under laboratory conditions. Analysis results prove the applicability of the Bayesian algorithm in accurately identifying the severity and location of ballast damage in ballasted tracks. The results also showcase the limitations and merits of system identification of a highly damped system in the time and modal domains. It is concluded that the time domain is more favored than the modal domain for system identification of the considered rail–sleeper–ballast system owing to the effects of ballast nonlinearity under large amplitude vibration.

Civil structures and infrastructures are often subjected by design to the impacts of natural and human-caused hazardous events, and accordingly may suffer from damages, functionality loss, and failure. In order to quantitatively measure the associated likelihood and consequences for quantifying risks, an appropriate measure of structural reliability and resilience is essentially required. This paper presents an explicit measure for the time-dependent resilience of repairable structures as a natural extension of time-invariant and time-dependent structural reliability concepts, taking into account the effects of structural performance deterioration and nonstationary external loads. First, the resilience measure associated with a single load event is discussed, and an integral-free expression is developed for a special case of structural robustness and recovery. Subsequently, the resilience measure for a planning horizon, which is a function of the duration of considered service period, is proposed in a closed form. Remarkably, the time-dependent resilience can be treated as a generalized form of the time-dependent reliability. Motivated by this, the resilience-based design and cost optimization of structures are also discussed. A numerical example is presented to demonstrate the accuracy and applicability of the proposed resilience measure.

A seismic risk assessment is a guarantee to keep the integrity of the infrastructure under an uncertain earthquake action. This paper carried out seismic risk assessment of a novel bridge-aqueduct cobuilt structure (BACS) under uncertain parameters, and the vulnerability was evaluated by the fragility of the pier and waterproof plate. The drift rate of the pier and the crack width ratio of the waterproof plate were defined as the damage index of the fragility curves. A back propagation (BP) neural network model was constructed, and the fragility data of the two simulated components were grouped into training and testing. The average deviation of the two models was within 5%, which can provide a methodology for BACS seismic risk assessment under other conditions.

Service life extensions of bridge components and systems due to maintenance can be correlated when the maintenance actions are applied to multiple bridge components. This paper deals with the probabilistic optimum bridge maintenance management of deteriorating bridge components and systems. The proposed maintenance management integrates two types of correlation: (1) the correlation among the safety margins of bridge components, and (2) the correlation among service life extensions due to maintenance actions. The effects of these two types of correlation on the extended service life, probability of failure, and optimum maintenance planning for bridge components and systems are investigated. This paper uses series-parallel models to represent a bridge system. The extended service life of a bridge system is estimated based on the extended service lives of its components considering (1) independent and perfectly correlated safety margins, and (2) independent, partially correlated, and perfectly correlated service life extensions. The optimum maintenance plannings for bridge components and systems are based on maximizing the time to reach their target probability of failure. The effects of the correlation associated with both the safety margins and the service life extensions on the optimum maintenance application times and the time to reach the target probability of failure are investigated. The proposed approach is illustrated on an existing reinforced concrete bridge under corrosion.

Construction is a key pillar in the global economy, but it is also an industry that has one of the highest fatality rates. The goal of the current study is to employ machine learning in order to develop a framework based on which better-informed and interpretable injury-risk mitigation decisions can be made for construction sites. Central to the framework, generalizable glass-box and black-box models are developed and validated to predict injury severity levels based on the interdependent effects of identified key injury factors. To demonstrate the framework utility, a data set pertaining to construction site injury cases is utilized. By employing the developed models, safety managers can evaluate different construction site safety risk levels, and the potential high-risk zones can be flagged for devising targeted (i.e., site-specific) proactive risk mitigation strategies. Managers can also use the framework to explore complex relationships between interdependent factors and corresponding cause-and-effect of injury severity, which can further enhance their understanding of the underlying mechanisms that shape construction safety risks. Overall, the current study offers transparent, interpretable and generalizable decision-making insights for safety managers and workplace risk practitioners to better identify, understand, predict, and control the factors influencing construction site injuries and ultimately improve the safety level of their working environments by mitigating the risks of associated project disruptions.

Many bridge management systems (BMSs) plan future maintenance and inspection based on deterioration models derived from probabilistic analysis of field inspection data. Such analysis considers the aleatoric but not the epistemic uncertainty arising from subjective or imprecise data. This raises questions regarding the efficiency and safety of maintenance and inspection decisions. Several methodologies have been proposed to address both uncertainties; however, they tend to be taxing in terms of inspection data requirements. Thus, this work proposes a new BMS-compatible methodology to derive deterioration models using logistic regression to capture aleatoric uncertainty and fuzzy set theory to capture epistemic uncertainty. To formulate the models, subjective or imprecise data, such as bridge condition rating, is modeled using membership functions, rather than discrete values, and then integrated into logistic regression analysis. This results in logistic models with fuzzy coefficients. The proposed fuzzy-logistic models can be used to predict a range of possible future bridge conditions, rather than a discrete condition and hence lead to a range of possible management strategies that can be then optimized using life-cycle cost analysis. The application of the proposed framework is demonstrated through a case study.

The performance of retaining walls is affected by spatial variability in the properties of the retained backfill. In practice, the conventional approach for designing retaining walls is deterministic and is based on ensuring acceptable factors of safety for different limit states of wall failure. One exception is the design method in Eurocode 7, where partial load and resistance factors are recommended to ensure a target level of safety in the design. The main limitation of partial load and resistance factors is that the recommended partial safety factors may not realistically incorporate the impact of spatial variability of the backfill properties, because the calibration studies that were conducted to determine the safety factors were not based on realistic random field modeling of the soils involved. Another limitation is that existing geotechnical reliability-based design approaches for analyzing retaining walls focus on failure in the soil and do not include design aspects of the structural behavior of the wall itself. The main objective of this study is to quantify the probability of failure associated with the design of a cantilever retaining wall using approaches that are based on partial factors of safety (Eurocode 7). The objective is achieved by utilizing random fields that represent the variability in the backfill in the finite difference software FLAC 2D. The effect of the properties of the random field on the design of the retaining wall are investigated to provide recommendations that can aid in the design of cantilever walls supporting cohesionless backfill. Results indicate that the partial load factors recommended by Eurocode 7 may not meet safety requirements, particularly for high uncertainties in the backfill friction angle and for large vertical correlation lengths.

Due to the expanding availability of high-fidelity weigh-in-motion (WIM) data in recent years, various local agencies have modified bridge design and rating procedures to best reflect state-specific traffic loads. However, accurate procedures for load model revision are often accompanied by high computational cost and implementation complexity. To address this concern, a reliability-based approach is proposed for vehicular live-load model development that involves selecting an actual vehicle configuration from the WIM database to serve as the basis for the load model. The approach first determined the required live-load factor for each potential vehicle configuration such that all considered structures would meet a minimum level of reliability. Next, the set of potential models was screened by imposing a limit on the level of design or rating conservatism allowed for any individual structure. Finally, an optimal load model was selected from the remaining set based on a penalty point approach that accounted for the deviation of results for any single structure as well as the overall deviation across all structures. Relative to an ideal reliability-based design optimization (RBDO) solution, the proposed method requires low computational cost, is straightforward to implement, results in a realistic vehicle configuration for the live-load model, and provides reasonable accuracy. The method was found to be slightly superior to the existing best-selection approach for large databases, but significantly better for small databases.

This paper presents the risk assessment of a footing-on-slope system that uses the random finite element method. The effect of a soil autocorrelation structure on the probability of failure and the associated risk is quantified. In this study, the anisotropic spatial variability of the soil is described with a major principal scale of fluctuation, a minor scale of fluctuation, and a rotation angle, and the spatial variability is modeled using the rotated random field. The generated random field is mapped onto a finite element model, which can quantify the bearing capacity of a footing on a slope. Further, the K-means cluster method is adopted to calculate the sliding area of the soil mass. Following Monte Carlo simulation, the probability of failure and the corresponding risk for footings on slopes are evaluated for various soil spatial variability scenarios. The results show the importance of considering the anisotropy of soils when attempting to identify the worst-case scenarios for risk.

The highly nonlinear and uncertain driver-vehicle-road (DVR) traffic system will be unstable and cause potential risks under certain conditions. This paper analyzes the attributes and interactive characteristics of DVR, and constructs the physical and mathematical models of the DVR microtraffic system. We first consider the potential accident consequences caused by vehicle-road interaction, the behavior uncertainty of driver-vehicle interaction and the risk sensitivity of driver-road interaction, construct a DVR system model to characterize the influence of DVR interaction process on the system safety. Further, by revealing the risk generation mechanism of a DVR system, we can achieve system risk identification and early warning. Experimental results verified by naturalistic driving dataset show that the risk trend generated by the DVR system model is consistent with the output of traditional risk indicators (THW, TTC). Compared with TTC, the proposed model is more universal to different traffic participants and road topology scenarios and can assist DVR systems to make early warning and control strategy.

Fragility functions are used in the vulnerability analysis of structures considering different sources of uncertainties. In this research, a framework to develop time-dependent fragility functions for circular tunnels embedded in soft soils is proposed considering the impact of corrosion on the lining reinforcement. Typical shallow and deep circular tunnel sections in soft soils of Shanghai City are used as case studies. The seismic response of the tunnel lining was obtained based on a series of nonlinear dynamic analyses of the soil-tunnel system. The aging effect due to corrosion of the reinforcement bar is considered by decreasing the strength properties of the tunnel lining. Time-dependent fragility curves as a function of free-field peak ground velocity ($PGV$), as well as fragility surfaces in terms of $PGV$ and service time $t$, are proposed for minor, moderate, and extensive damage states. The main sources of uncertainty are linked with the input motion and frequency content, the soil properties and response, the tunnel embedment depths, and the estimation of the damage levels. Results show an overall increase in the seismic fragility for both the shallow and deep tunnels over time, emphasizing the significant impact of aging effects on the performance of tunnels. The findings of this study provide an improved understanding of the performance of tunnels exposed to diverse hazards and hence facilitate the life-cycle seismic risk assessment and resilient designs of transport infrastructure.

Given the special geometric features of curved bridges, they show a complex behavior during an earthquake. Considering the fact that bridges have shown vulnerability during past earthquakes, vulnerability evaluation and retrofitting of bridges has become of significant importance. Therefore, this work addressed the impact of different radii of curvature on the probability of seismic damage for curved multiframe concrete box-girder bridges and compared the results with those of straight bridge. The bridge piers were then retrofitted with carbon fiber–reinforced polymer (CFRP). Subsequently, bridge systems were modeled in OpenSEES software subjected to 80 near-fault ground-motion records, considering three components of the earthquake in three dimensions and including all the existent uncertainties. Numerical results indicate that with decreasing bridge radius, the probability of bridge vulnerability substantially increases. Moreover, strengthening bridge piers with CFRP increases median fragility by approximately 32% on average, and thus has a significant impact on reducing the vulnerability of the bridge system.

The digital twin (DT) is a virtual replica of real-world buildings, processes, structures, people, and systems created and maintained to answer questions about its physical part, the physical twin (PT). In the case of the built environment, the PT is represented by smart buildings and infrastructures. Full synchronization between the DT and the PT will allow for a perpetual learning process and updating between the two twins. In this work, we introduce a novel concept of DT called risk-informed digital twin (RDT). In the DT the model predictions are developed through data-driven tools and algorithms. However, multiple sources of uncertainty during the lifecycle challenge our understanding and ability to effectively model the performance of the modeled systems. The RDT’s importance lies in its integration of the methods and tools of statistics and risk analysis with machine learning. To this aim, the platform incorporates a novel framework of data-driven uncertainty quantification and risk analysis rooted in information theory. At the core of the RDT is a framework of sustainable and resilient based engineering (SRBE), introduced in this work and considered the first step toward the extension of performance-based engineering (PBE) approaches to socioecological-technical systems under uncertainty. A risk-informed multicriteria decision support tool able to incorporate social aspects is also included, and it can be used for sustainable and resilient design in the early stage, or management under uncertainty of smart buildings and infrastructure systems.

This paper establishes a risk control method that ensures the construction safety of a comprehensive underground pipe gallery in a complex surrounding environment. Through risk identification, risk analysis, and the application of numerical simulation, the construction safety of an underground pipe gallery crossing in close range of an existing bridge was evaluated and monitored. Construction risk control measures are formulated based on the geological conditions, engineering experiences, numerical simulation, and risk assessment. By using an information monitoring system during the whole construction process, the construction parameters are adjusted and optimized. The implementation shows a good control effect. The numerical analysis indicates the key risk control indicators that are verified from the monitoring results at the construction site. Both the displacement of the ground and the structure all meet the control requirements. The good implementation effect suggests that the construction risk can be controlled and that the suggested risk control method is reasonable and applicable. This research and case study results provide design, construction, and risk management experience for subsequent similar projects.

The importance of developing accurate seismic analysis for dams becomes crucial, especially in active seismic zones. In this research, a probabilistic seismic hazard analysis (PSHA) for 48 Spanish dams in the Andalusian region (southern Spain) has been carried out. The selected dams (25 concrete + 23 embankment dams) fulfill the double requirement of being higher than 61.0 m and having a reservoir capacity greater than $13.60{\mathrm{hm}}^3$. Seismogenic parameters have been retrieved from the new seismogenic zones established in 2015, considering the soil homogeneously as rocky. Furthermore, return periods and structural periods range from 500.0 to 5,000.0 years and from 0 to 4.0 s, respectively. Results are plotted in terms of peak grand accelerations (PGAs) and pseudospectra accelerations (PSAs) for each dam. A disaggregation analysis was carried out to provide the contribution of hazard for magnitude and distance pair. The results allowed a possible PGA and return period in any point in Andalusian region to be obtained by linear interpolations. A double dimension for results is provided: first, uniform hazard spectra (UHS) are provided indicating the probability of exceeding the spectral acceleration in each spectral period is constant at a certain site; then, disaggregated results indicate the right combination of magnitude and distance that contribute the more to the seismic hazard at a specific site, for a given intensity measure, and at a certain level of intensity.

The objective of this paper is to investigate how the spatially varied soil stiffness affects the ultimate lateral load and failure mechanism of monopiles embedded in soft clay. Responses of a laterally loaded monopile in spatially varied soils were simulated by a three-dimensional finite-element analysis combined with random field theory. The influence of the coefficient of variation (${\mathrm{COV}}_E$) and the correlation distance (${\delta }_z/l_L$) regarding the stiffness parameters was investigated. Results revealed that the spatial structure has an impact on stiffness distribution and leads to a significant discrepancy between the ultimate load of the spatially random soil and the uniform soil. The assumption of uniformity in stiffness yields an overestimation of the ultimate load and an excessively high cost of construction. Both the ${\mathrm{COV}}_E$ and ${\delta }_z/l_L$ intensify the discrepancy and increase the probability of the foundation exceeding the ultimate limit states. Due to the spatial pattern of soil, the failure plane exhibits an unsymmetrical pattern and an irregular shape in random soil instead of a symmetrical shape in uniform soil. The results provide a better explanation of the failure mechanisms of monopile foundations and reveal the importance of considering the reliability indices in the designing process.

Recent data trends and analysis have highlighted the need to incorporate more imprecise, ambiguous, and unreliable data into uncertainty analysis traditionally handled by probability theory. However, data fraught with potential error and missing information are not well suited for analysis using probability theory due to high epistemic uncertainty. Evidence theory offers an alternative method of assessing epistemic uncertainty and is well suited for expanded use in engineering applications. Unfortunately, a unified approach to the application of evidence theory is lacking. To address this gap, we developed a protocol for engineering applications of evidence theory. The protocol proposes a logical procedure for defining the frame of discernment, the initial assignment of belief mass, the selection of combination rule, and sensitivity analysis. A literature review of prevailing methods related to the application of evidence theory highlighted concepts and considerations to address. The steps of the protocol were explored and discussed using an example problem including several rule combinations in order to highlight differences in the results and implications of making different analytical decisions. The protocol proposed herein is intended to facilitate engineering applications of evidence theory and to promote more widespread use of the theory in the field of civil engineering.

To assess tsunami evacuation risk accurately in order to guide effective evacuation planning, various uncertainties (including the aleatory and epistemic uncertainties) associated with the evacuation process need to be quantified properly. Reducing the epistemic uncertainties associated with the evacuation process (e.g., through data collection) can facilitate more-accurate estimation of tsunami evacuation risk. To guide such reduction or prioritize data collection, this study performed sensitivity analysis of tsunami evacuation risk (i.e., risk sensitivity analysis) with respect to epistemic uncertainty. An agent-based tsunami evacuation model was used to simulate the evacuation within a simulation-based risk assessment framework, which incorporated various uncertainties associated with the evacuation process. The aleatory uncertainty in the input random variable was quantified by probability distribution models, and the epistemic uncertainties were quantified by distribution parameters that also were modeled by probability distributions. Sensitivity analysis of tsunami evacuation risk with respect to the epistemic uncertainty was performed to evaluate the impact of various epistemic uncertainties on the variability of the evacuation risk and identify those that have relatively large impacts. An augmented sample-based approach was used to calculate efficiently the variance-based sensitivity indexes (i.e., Sobol’ indexes) for all distribution parameters. The sensitivity information can be used to prioritize the data collection for effective epistemic uncertainty reduction, and for a more accurate risk assessment to support more-effective evacuation planning. As an illustrative example, sensitivity analysis of tsunami evacuation risk of Seaside, Oregon with respect to epistemic uncertainty was performed under different risk measures.

Power distribution systems, composed of wood poles and wires, are susceptible to damage caused by strong winds during hurricanes or winter storms. Power outages induced by failed poles result in millions of revenue and restoration costs and impact millions of customers. Failure of aged wood poles is a major reason for power outages during hurricanes or winter storms. Replacing aging poles is an easy and effective approach but with a high cost to enhance radial power distribution poles. In this paper, the optimal enhancement strategy of the power distribution system was obtained through the Bayesian decision network (BDN), which is an ideal tool to deal with trade-off problems. Besides the economic benefit of the electricity utility company, the benefits of residential customers, namely the power supply reliability, were also integrated into the reference of decision makers through BDN. To determine priorities of the pole replacements, the component importance index of individual poles, instead of the commonly used component importance index of a line, was explored using physics-based reliability analysis rather than an identical mathematical model applied to all poles to describe their vulnerabilities. A surrogate model established by Bayesian regularization neural network (BRNN) was employed to save the computational cost in the Monte Carlo simulation of reliability analysis. The cost of the hardening strategy using the component importance index of each individual pole was about 80% of that with the component importance index of a line, which indicates that the component importance index of each individual pole is more effective. The optimal choice of strategies would vary with the utility function that describes the decision-making references because benefits between the utility company and customers might conflict with each other. When the utility company and customers are given the same priority in the hardening process against Category 5 hurricanes, the optimal pole replacement percentage was 30%.

This paper proposes a probabilistic framework for updating the governing parameters in the hysteretic constitutive model for tubular steel with strength degradation. The hysteretic constitutive model is formulated to track the strength degradation due to the local buckling of square hollow steel beam-columns imposed by cyclic loadings with large elastoplastic deformation. Despite various hysteretic laws that have been proposed to model the steel tubular strength degradation, limitations for determining parameter values remain in numerical analysis. The parameters are generally obfuscated by the inevitable epistemic uncertainties from material and geometric properties. The updating process of the material parameters is performed within the Bayesian framework employing the Markov chain Monte Carlo algorithm. The epistemic uncertainty involved in the computational procedure is initially represented as predefined intervals of the uncertain parameters. The proposed Markov chain Monte Carlo (MCMC) algorithm can generate samples from the posterior distributions of the parameters according to the experimental results. The epistemic uncertainty is hence significantly reduced by the Bayesian updating process such that the updated model is feasible to predict the degradation behavior of square hollow steel beam-columns subjected to cyclic loadings. The benchmark example indicates that the proposed framework can find the optimal path for updating key parameter values to accurately assess the condition of steel tubular structures in terms of the degradation behavior.

The epistemic uncertainty of geological condition, hydrological condition, and human factors causes the risk assessment result of metro construction to be imprecise. To reveal the epistemic uncertainty, a probabilistic risk assessment approach for metro construction is proposed based on the credal network. The epistemic uncertainty of risk factors is described by prior probability and conditional probability. Based on the scarcity of available data, prior probability is quantified by interval probability and probability box, and conditional probability is qualified by imprecise leaky noisy-OR model and imprecise Dirichlet model. Taking the shield tunnel projects of Beijing Metro Line 14 as the research object, the proposed approach and the conventional approach without considering epistemic uncertainty were adopted to predict the risk probability of surface settlement caused by shield tunnel construction. The predicted results were compared with the statistical probability derived from Peck formula and the field monitoring data, which verifies the rationality and effectiveness of the proposed approach.

Abstract

The primary concern in the residential projects industry is to implement a project with the optimal quality within the planned schedule and planned budget. Simulation is considered an effective tool for analyzing the construction process in residential projects because it considers behavior of feedback as well as the frequent changes with time. This study aimed to develop a system dynamics (SD) model to estimate the overall cost overrun risks in addition to allowing decision makers to anticipate the feedback of indefinite numbers of what-if questions considering highly dynamic uncertainty and dynamic risk interactions influencing residential project performance that traditional methods may miss. Major risks that can cause cost overruns in residential projects were initially identified through a review of related literature, a questionnaire survey, and several workshop rounds. Then, the model boundaries’ variables were identified, and a qualitative model was created through a causal loop diagram (CLD). Afterward, mathematical equations and relationships between variables were established in a stock-flow diagram. The proposed model consists of five subsystems to simulate how changes in dynamics of inspecting the completed works from the contractor’s and the consultant’s points of view; the probability of scope change; the material resources management process; the workforce hire-quit cycle; and the work-rework cycle can influence the cost performance of a residential project. The model was tested and validated using data derived from a residential project located in Egypt, and testing results showed that the simulated behavior of the model is the same as the actual behavior of the project. The sensitivity analysis results showed that increasing workforce more than the desired, scope changes, and rework rate are the parameters that cause the largest cost performance regression in residential projects.

Practical Applications

Simulation is considered an effective tool for analyzing the construction process in residential projects because it considers the behavior of feedback as well as frequent changes with time. In this research, a SD model is presented to estimate the overall cost overrun risks in addition to allowing decision makers to anticipate the feedback of indefinite numbers of what-if questions. The whole model is divided into five subsystems, which simplified the whole system into some more understandable ones. The model provides a framework for uncertain events that may influence cost performance in the construction industry such as the impact of global events (e.g., Russia–Ukraine war) on the supplies of building materials and their prices. The proposed system dynamics model provided a material-resource subsystem to simulate those kinds of global events and their influence on cost performance on construction projects.

Rapid urbanization due to fast economic development has put tremendous stress on urban infrastructure and has caused an increase in impervious surface area and runoff, resulting in flash floods in many cities in the world. In developing countries such as India, urban floods have severe impacts due to highly vulnerable population density. The normal functioning of the urban region is disrupted due to the effect of floods on the urban road network. This paper examined vehicular mobility on urban road networks under nonflood and flood conditions and identified an affected region under the flood condition. Multiple centrality assessment (MCA) was used with multiple classical metrics, such as betweenness centrality and closeness centrality, to assess the status of nodes using distance and travel time under nonflood and flood conditions. The road network under floods was generated by the Soil Conservation Service curve number (SCS-CN) and time area histogram methods using flooding depth over road network and betweenness centrality. Weighted overlay centrality layers were created to identify the regions with severely affected mobility under the flood condition by generating a ward-level vulnerability map. The Pearson’s correlation analysis of centrality maps showed that network centrality is correlated strongly with the road network availability at a 95% confidence level. The study also indicates that under flood condition, high betweenness centrality decreases at the city center and closeness centrality exhibits contrasting spatial patterns. According to the ward-level vulnerability map, mobility on road networks was affected severely in 29.35% of the areas, most of which were centrally located. This research suggests that during urban planning, major roads that have greater impedance during floods should be distributed throughout the region so that mobility is less affected by floods.