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

This study presents frameworks to build nonintrusive polynomial chaos expansion (PCE) models with regression and Smolyak sparse-grid quadrature to perform high-dimensional uncertainty propagation and uncertainty quantification (UQ) for response estimated with the hybrid data + model-based submodeling (HDMS) method. The HDMS method drives the finite-element submodel containing a critical location of a structure using the measured response of the real structure at the preselected submodel boundaries to estimate a refined response distribution around the critical location. The proposed UQ frameworks are implemented on an experimental case study of a plate with holes as critical locations under tensile loading. The UQ results at the critical locations from regression-based PCE models built using different sampling methods and the Smolyak sparse-grid quadrature-based PCE models are compared with the UQ results from the traditional Monte Carlo simulation (MCS) method. The regression-based PCE model with Smolyak sparse-grid sampling demonstrated significantly higher accuracy in distribution parameters and probability density functions (pdf) compared to the other regression-based PCE models. While the Smolyak quadrature-based PCE model with a considerably small experimental design showed slightly lower accuracy, it still outperforms regression-based PCE models with MC-based sampling.

We propose a method for simultaneously estimating the trend and random components of soil properties using the least absolute shrinkage and selection operator (LASSO) for sparse modeling without assuming any basis functions and Gaussian random field theory, respectively. The uncorrelated observation noise is also estimated at the same time. A two-step algorithm is introduced to avoid the shrinkage problem of LASSO. Numerical examples with random realizations show that the method avoids shrinkage. The proposed method requires four parameters, namely, the variances of the random component and observation error, autocovariance distance, and regularization parameter, for one-dimensional problems. We propose a method that uses Akaike’s information criterion or Bayesian information criterion and particle swarm optimization to determine these four parameters. It is shown that the detection ratio of the layer boundary is determined by the number of observation data and the difference between trend values. The proposed method is applied to actual cone penetration test data to estimate the trend component of the soil property.

Increasingly, decision makers and society at large are focusing on ways of conserving natural resources. This change of perspective will certainly impinge on the construction industry. Several recent studies have developed material composition indicators (MCI) in order to assess the quantities of construction materials required for various building types. One major limitation of the MCIs used thus far is that they ignore locational factors and the associated site-specific boundary conditions such as building statics in seismic zones as well as the form of subsoil, which will dictate the needed level of earthquake resistance and thus the materials used in the design. The aim of the exploratory study presented here is to determine the extent to which site-specific parameters affect the composition of MCIs for certain building types. To this end, the influence of these parameters is estimated for various regions in Afghanistan. The results show significant correlations between the material consumption and site-specific boundary conditions.

Data-driven material models can capture the constitutive relationship directly from the data without involving any material-dependent mathematical expressions. But most data-driven approaches, such as artificial neural networks, only estimate the deterministic relations and do not consider the material uncertainty. In this paper, the constitutive relation is taken as a stochastic function following the Gaussian process, where a probability-based nonparametric method, called Gaussian process regression (GPR), is used to capture the constitutive relation with the uncertainty being included. Both one-dimensional (1D) and two-dimensional (2D) material data are used to validate the GPR-based constitutive model (GPR model). The obtained GPR model shows higher accuracy than other data-driven approaches, particularly when the data set size is small. When compared with the assumed true model, the GPR-based model has an average relative error of $<2.3\%$. Finally, with the help of the material uncertainty identified by the GPR-based model from the material data, a data-driven stochastic structural analysis procedure is developed. The relative errors of the expected deflection and probability of failure given by the GPR model are smaller than 2% and 3%, respectively.

Reliability-based design optimization (RBDO) is a well-known design strategy in engineering. However, RBDO usually requires uncertainties to be modeled by statistical distributions. This requires the availability of sufficient sample size so that these variables can be represented accurately by probabilistic distributions. In the design of new systems and structures, usually there is a lack of information about some uncertain variables or parameters and only a reduced set of samples might be available. This prevents their treatment as probability distributions. This type of uncertainty is called epistemic uncertainty. This paper proposes two effective multiobjective evolutionary algorithms to solve design problems under both types of uncertainty: aleatory and epistemic. Two objective functions, namely the cost of the structures and the probability of failure, are considered. The results are Pareto fronts with a trade-off between cost and reliability associated with a specified level of confidence. Pareto fronts show minimum achievable values for the probability of failure for a given cost. The effect of the epistemic uncertainty on the solution is also investigated. An analytical example and two structural examples are solved to show the applicability of the approach and how epistemic uncertainty may affect the results.

The uninterrupted performance of the infrastructure systems of a community is essential to its welfare and prosperity. Infrastructure assets are continuously subject to degradation, which can compromise their performance and increase their seismic vulnerability. Measures such as seismic retrofits should be adopted to ensure the uninterrupted performance of infrastructure assets. Proper timing of these interventions can reduce the total costs incurred by the communities. To this end, an appropriate investment valuation method is needed. This study proposes an investment valuation framework that addresses this need. At the core of this framework is a real options model that determines the optimal time to implement seismic vulnerability reduction measures considering various uncertainties about the state of the asset, the probable future hazards, and the costs incurred by the community. The main contribution of this study to the body of knowledge is the creation of a framework based on real options analysis (ROA) that is tailored to the context of investment decision-making and timing of seismic vulnerability reduction measures. Practitioners can use this framework to plan the implementation of seismic vulnerability reduction measures, avoid over- or underinvestment in these interventions, enhance their flexibility in using limited available financial resources, and achieve resilience enhancement objectives.

The multimodal transportation network in which freight rail plays an essential role continues to enhance the United States’ contributions in the global market. For years, track geometry defects data have been often gathered by visual inspections. However, automated track vehicles are now deployed for the same purpose to save time and cost. One of the limitations of an automated vehicle is the likelihood of nonstationarity of the gathered data due to external influence. The effect of nonstationarity may lead to the wrong representation of track conditions and thereby increases the possibility of false model output. This study applies supervised machine learning (ML) methods to detect the nonstationarity of the geometry data. The methods include random forest (RF), logistic regression (LR), and support vector machine (SVM). The authors vary the train test and validation ratio in phases to ascertain each machine learning method’s accuracy and adaptability to different distributions. In the first phase, the random forest and the support vector machine show an accuracy of 97.1%, while the logistic regression reveals 96% accuracy. In the second and third phases, the random forest method gives a better result than other supervised learners with accuracies of 97% and 97.1%, respectively. Similarly, for validation, the random forest performs better than other classifiers. Conclusively, the developed models’ application indicates that the random forest is a more effective approach to detecting the nonstationarity of track geometry data.

Sinkhole studies of Kuwait have involved the investigation and risk assessment of sinkholes in specific areas. This study assessed sinkhole hazard susceptibility within the geographic borders of Kuwait by integrating the analytical hierarchy process (AHP) with geographical information system (GIS) software. As a result, eight criteria were selected: water table depth within soil cover, fluctuation of the water table within bedrock, proximity of faults, the existence of domes (surface anticline) within bedrock, earthquake zones, drainage class of the soil cover layer, soil cover depth, and distance between the top of the bedrock and the water table. Each criterion was further classified into five subcriteria. The developed sinkhole susceptibility hazard map (SSHM) classified Kuwait into five regions in terms of potential sinkhole susceptibility: very low, low, moderate, high, and very high potential. The SSHM was verified using data from sinkhole incidents in Kuwait and published records and literature regarding karst cavity development in the region. The SSHM can be used as a database for informing engineering decision-making processes, future risk assessment, and emergency planning.

In this study, a two-step approximate Bayesian computation (ABC) updating framework using dynamic response data is developed. In this framework, the Euclidian and Bhattacharyya distances are utilized as uncertainty quantification (UQ) metrics to define approximate likelihood functions in the first and second steps, respectively. A new Bayesian inference algorithm combining Bayesian updating with structural reliability methods (BUS) with the adaptive Kriging model is then proposed to effectively execute the ABC updating framework. The performance of the proposed procedure is demonstrated with a seismic-isolated bridge model updating application using simulated seismic response data. This application denotes that the Bhattacharyya distance is a powerful UQ metric with the capability to recreate wholly the distribution of target observations, and the proposed procedure can provide satisfactory results with much reduced computational demand compared with other well-known methods, such as transitional Markov chain Monte Carlo (TMCMC).

The conventional sparse grid (SG) integration method treats all considered random variables equally for the numerical integration of performance functions. The dimension-equal treatment inevitably causes a considerable demand of computation devoted to unimportant variables. To overcome this shortcoming, an advanced dimension-adaptive sparse grid (ADASG) integration method is proposed by introducing a numerical indicator quantifying the importance levels of tensor products of difference quadrature formulas. By achieving a target number of function evaluations, only several important tensor products of difference quadrature formulas are retained while the unimportant ones are removed. The proposed ADASG integration method is firstly employed to estimate the first four moments of performance functions. The estimated moments are then applied to structural reliability analysis by generating the probability density function of performance functions using the maximum entropy method. The advantage of the proposed ADASG integration method is demonstrated over four examples. The reliability index calculated by the proposed ADASG integration method shows favorable consistency to that obtained by the direct Monte Carlo simulation. Compared with conventional SG integration methods, the proposed ADASG integration method exhibits a better performance in terms of both accuracy and efficiency.

Civil infrastructure may degrade due to the adverse effects of continuous damage (e.g., reinforcement corrosion) and sudden shocks (e.g., earthquakes) during its service life. Many studies have been conducted in the field of reliability-informed life-cycle assessment, but there is still a need for a general and efficient method to assess the time-dependent performance of aging structures by considering different deterioration scenarios and maintenance actions in a unified manner. Some of the traditional methods may have difficulties in handling multiple deteriorations, nonlinear models, a large number of uncertainties, scenarios of nondifferentiable performance functions, and combined effects of deterioration and maintenance. This paper develops a novel approach for a time-dependent reliability analysis based on the proposed point-evolution kernel density estimation (PKDE) method and equivalent extreme performance function. The proposed approach allows consideration of various uncertainties (e.g., external loads, deterioration scenarios, and maintenance models) and the associated correlation effects. In the proposed approach, both the progressive deterioration and sudden damages are considered in the modeling of the performance function. Besides, different types of maintenance schemes are assessed. The equivalent performance function is established, and the proposed PKDE method is used to address the first passage problem and nondifferentiable performance function within a time-dependent reliability analysis. An illustrative example is made to demonstrate the feasibility and accuracy of the proposed PKDE method. The computational results using the proposed method are verified by comparison with those from Monte Carlo simulations (MCS).

For large-span hybrid roof structures, the construction uncertainties of cable tension usually have significant influences on the roof’s mechanical performance and should be considered in reliability evaluation. An effective approach to quantify uncertainties of cable tensions and evaluate structural reliability is proposed to carry out the studies by combining finite-element simulation with the multiple response surfaces method. Taking a hybrid roof structure with cables and steel trusses as an example, the main procedures on this issue are illustrated. First, a finite-element model is established for the hybrid roof structure considering construction deviations, such as the deviations of cable force between the design values and the real measured values. The ultimate bearing capacity of the structure is calculated for models with and without deviations, and the effects of construction deviations on structural bearing capacity are analyzed. Then, an uncertainty model of cable tension for structural reliability evaluation is proposed by establishing the statistics of initial strain in a structural analysis based on the monitored deviations. With subspace division and limit state sample (or sample pair), the multiple response surfaces method is developed to solve reliability for examples with complex failure functions. It is found that the hybrid roof structure has a high reliability index about 6.76; and the uncertainties of cable tensions have a large impact on the reliability, especially the uncertainties of the upper suspension cable tensions and the back cable tensions.

In situ soil properties exhibit spatial variability, which is often described using a three-dimensional (3D) random field. With site investigations, soil properties at some specific locations are available. The corresponding data can be incorporated by a conditional random field to update the uncertainty parameters so that a more realistic or refined model can be achieved. Two algorithms, the Kriging and patching algorithms, are introduced for generating a 3D conditional random field. The conditional random field is linked with finite-element modeling, within the framework of Monte Carlo, to evaluate the performance of these two approaches in slope stability analyses. Sparsely distributed borehole data and cone penetration test (CPT) data are considered. The results indicate that for cases with limited sampled data, the patching algorithm gains an advantage over the Kriging algorithm in terms of prediction accuracy and uncertainty reduction. Data near the sliding surfaces of a slope remarkably affect the stability; thus, with sufficient ground information near the sliding surfaces, a conditional random field can provide better guidance for slope design.

Knowledge of the uniaxial compressive responses of rock mass is crucial for evaluating the stability of rock pillar in underground engineering. The inherent spatial variabilities of rock mass properties are generally ignored in the traditional stability evaluation of rock pillars. In this study, probabilistic analysis of rock pillar was carried out to investigate the compressive responses of spatially variable rock pillars with three-dimensional (3D) random fields. Three-dimensional random fields of rock mass properties were generated using the Cholesky decomposition method, which was implemented into the ${\mathrm{FLAC}}^{3\mathrm{D}}$ using the Fish code. The failure mechanism of spatially variable rock pillar and influence of spatial variability of rock mass properties on uniaxial compressive strength and Young’s modulus were investigated on the basis of the Hoek–Brown criterion with Latin hypercube sampling (LHS). Results indicated that the spatial variability of rock mass properties significantly influences the uniaxial compressive strength and Young’s modulus of rock pillar. The expansion of the plastic region has obvious directivity, which is closely related to the direction of the random field. The vertical scale of fluctuation has a greater effect on the mean value of the uniaxial compressive strength and Young’s modulus of rock pillar than the horizontal scale of fluctuation.

With the rapid urbanization in this era, it has become an economic imperative to make use of underground space in many countries. Achieving long-term durability goals of lining structures is a challenging task to guarantee the normal serviceability of underground space, which may be affected by many threatening factors such as water seepage. In this paper, an explicit method is developed to guide the design of lining structure thickness against water seepage in a probability-based framework. For a two-dimensional surface, the mean value and variance of the average seepage scenario are first derived in a closed form, taking into account the spatial correlation and uncertainty of the lining structure thickness and water seepage depth. The gamma distribution is used to describe the probabilistic behavior of the average seepage scenario, based on which an explicit link is established between the lining structure thickness and structural failure probability. A numerical example is presented to demonstrate the accuracy and applicability of the proposed method.

This study follows the fuzzy analytic hierarchy process to develop a semiquantitative technique to assess the safety risk of construction projects based on effective parameters affecting the probability and severity rates. Various studies targeting safety risk assessment in construction projects together with their research characteristics were investigated. In the second step, the initial structure of risk assessment was designed based on a semiquantitative technique of safety risk assessment. In the third step, a three-round Delphi method was employed, and an expert panel comprising 41 project managers and senior authorities of construction projects determined the existing components of the two factors of risk probability and severity. Finally, the weights of the components of risk probability and severity were determined so that their roles in determining the level of final risk could be specified. The normalized weights of probability and severity factors for calculating the semiquantitative safety were 0.48296 and 0.51704, respectively. Moreover, the incident probability factor included the four components of technical inspection, human reliability, incident learning, and hazard detection, with the following weights, respectively: 0.303, 0.364, 0.088, and 0.245. The incident severity factor included the four components of human injury, financial loss, operational interruption, and reputation damage, with the following weights, respectively: 0.369, 0.233, 0.191, and 0.207. The utilization of this technique can help reduce the risk of occupational incidents in construction projects.

In urban watersheds, floods directly affect mobility on road networks. The present study proposes a methodology for analysing the impact of road construction or reconstruction on topography and, ultimately, on causing floods. The study first proposes a methodology to estimate the vulnerability to mobility on road networks due to flooding, illustrated on a road network in Nagpur city in central India. The study proposes the flood vulnerability road mobility index (FVRMI) through hydrologic modeling using parameters such as precipitation trends, the land use function, elevation models generated by satellite data, flood inundation depth calculated using the soil conservation service curve number (SCS-CN) method, a time area histogram, and the flood-safe driving speed of a road network. Then, FVRMI was estimated for the changed topography observed through elevation models generated by satellite data due to road reconstruction from 2010 to 2018. Results show that this has affected the flow regime of the Nagpur urban watersheds. The estimated FVRMI was 20% higher for the changed topography in 2018 (Model II) than the topography scenario of the base year, 2010 (Model I). The results show that about 63% of road segments had a higher FVRMI value in 2018.

Interdependent infrastructure systems consist of several sectors, including gas, power, transportation, and health systems, and are important in modern societies because they provide essential services for the continuous functioning of communities and society. This study presents a framework for conducting vulnerability analyses and key sector identification in interdependent infrastructure systems during and after the occurrence of an external perturbation. To analyze the correlation of sectors and identify the key sectors at different stages of a hazard, this study introduces input-output models, demand-driven inoperability input-output models (IIMs), and supply-driven IIMs. Moreover, an interdependency matrix is used to construct the network in an interdependent infrastructure system, and new indicators of importance and vulnerability indexes are defined to analyze changes in key sectors in different periods of an external hazard. The proposed model is illustrated through a hypothetical four-sector interdependent infrastructure system based on a flash flood event. However, the model is applicable to any physically interdependent system, such as a transportation system, power supply system, and so on. The importance and vulnerability attributes of interdependent infrastructure systems at different stages of the risk management process also provide guidance for decision makers involved in dealing with natural hazards such as flash flood events.

Steep ramps are often unavoidable in route design, leading to adverse effects on freight transportation organization. For operators, how to safely and effectively run trains on steep downhill lines is a challenging task. In response to the requirement, firstly, the control principles of electric freight trains, including safety demands, rapid running, use of driving modes, and periodic braking, are analyzed. Then, a fast train operation simulation algorithm (FTOSA) is proposed to realize the control principles. Finally, a real case study is carried out, and FTOSA is compared with existing simulation software in China. The results showed that FTOSA could generate a high-efficiency speed profile on long steep downhill lines under the constraints of train safety control. This work can provide some references for the operation of electric freight trains.

Risk response is a core element of project risk management, and is perhaps the area that needs the most improvement. Common causes of construction delay risks have been identified worldwide, and various traditional measures used for delay risk control have been critically reviewed. However, many delays in construction projects have been occurring in recent years throughout the world because of unexpected events such as financial crises and the COVID-19 pandemic. Accordingly, six case studies from five companies were chosen for this work, and were supplemented with face-to-face interviews, documentation, and direct observations. This strategy allowed the research evidence to be triangulated, and thus garnered more confidence in testing a particular concept or theory. Hence, a maturity model for risk response development has been outlined and validated to achieve construction success in preconstruction and during execution. Based on the validation, it is anticipated that by developing the risk response model, the process itself will become more objective, particularly in delay risk control.

The probability density distribution of soil parameters is of ultimate importance in reliability analysis of a geotechnical structure. Constructing the probability density function often needs a large number of experimental data, which is not easily obtained. The dependence structure between two different soil parameters may provide quantitative correlation between the parameters. This study investigated the probability density distribution and the dependence structure of several soil parameters (i.e., cohesion, friction angle, water content, compression modulus, and void ratio) using a large number of data obtained in a construction site in Shenzhen, China. The dependence structure of three random variables is investigated for the first time. An example on estimating bearing capacity and settlement of a shallow foundation is demonstrated using the constructed dependence structures for reliability analysis. The results show that the optimal probability distributions for cohesion and friction angle are normal distributions. The compression modulus and the void ratio follow lognormal distribution. The water content follows a gamma distribution. The Frank copula is the best dependence structure of cohesion and friction angle. The Clayton copula is the best dependence structure of compression modulus, void ratio, and water content. The results show that different dependence structures of soil parameters can significantly affect the probability of failure. The traditional Gaussian copula may overestimate the failure probability in terms of bearing capacity.