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

Published standards may be incomplete because they provide no guidance on wind load factors appropriate for use with the wind tunnel procedure. The purpose of this paper is to contribute to such guidance. Based on a classical definition of wind load factors as functions of uncertainties in the micrometeorological, wind climatological, aerodynamics, and structural dynamics elements that determine wind loads, the paper presents a simple, straightforward approach that allows practitioners to use appropriate wind load factors applicable when those uncertainties are either the same as or different from those assumed in the standard. Illustrations of the approach are presented for a variety of cases of practical interest. In estimating design wind loads, the various uncertainties should not be accounted for in isolation, for example, by specifying peak pressure coefficients with percentage points higher than those corresponding to their expected values. Rather, to achieve risk-consistent designs, the uncertainties should be accounted for collectively, in terms of their joint effect on the design wind loading. The design wind effect is equal to the estimated expectation of the peak wind effect times a load factor that, in most cases, is not significantly different from the load factor explicitly or implicitly specified in the standard. Notably, the load factor is not affected significantly by errors associated with interpolations required in typical database-assisted design applications. However, if the available wind speed records are several times shorter than 20 to 30 years, for example, the wind load factors increase by approximately 15%.

This paper presents an approach for updating the robust structural reliability that any particular response of a structural dynamic system will not reach some specific failure or unfavorable state when it is subjected to future stochastic excitation. In particular, the updating approach is based on incomplete modal data identified from the structural system. Uncertainties arising from structural modeling and modeling of the stochastic excitation that the structure will experience during its lifetime are considered. The proposed approach integrates the Gibbs sampler for Bayesian model updating and subset simulation for failure probability computation. A new efficient approach for conditional sampling called a constrained multigroup Metropolis within Gibbs (CMMG) sampling algorithm is developed by the authors. Another appealing feature of the proposed method is that it provides not only the exceedance probability estimates but also conditional samples that allow in-depth failure analysis in a single simulation run. The proposed method provides a substantial improvement in efficiency over estimators based on crude Monte Carlo simulation (MCS) for the updated robust reliability and is robust to the number of random variables and uncertain parameters and the amount of modal data involved in the problem. The effectiveness and efficiency of the proposed approach are shown by two illustrative examples.

Hurricanes are a way of life in South Florida, where owning a home without windstorm insurance is almost an impossibility. Insurance premiums for windstorm loses are computed through the use of complex mathematical models called catastrophe (cat) models. When they were first developed, cat models focused on the calculation of average annual loss due to wind. However, with exposure increasing rapidly along the coast of South Florida, it is imperative for insurance companies to protect themselves from the “once-in-100-year event”; in other words, probable maximum loss (PML). Gulati et al. (2014) investigated the computation and distribution of probable maximum loss in the case of personal residential structures for version 5.0 of the Florida Public Hurricane Loss Model using parametric and nonparametric methods. Here, the authors investigate the computation of probable maximum insured losses for personal and commercial residential buildings in version 6.1 of the model using the same methods. The authors also compare how PML values for total insured loss have changed between the two versions.

External pressure coefficients specified in the ASCE 7-10 Standard, used to determine design wind pressures for the components and cladding of buildings, are developed from wind tunnel test data that date back 30–50 years. In recent decades, advances in pressure measurement and computer technology have made it possible to obtain simultaneous pressure records, with high sampling rates, at many more wind tunnel pressure taps than was the case in the past. This paper proposes a method to calculate external pressure coefficients using aerodynamic wind tunnel databases such as Tokyo Polytechnic University’s large, publicly available database. Voronoi diagrams are used to assign tributary areas to irregularly spaced pressure taps. User-defined grids of various sizes and shapes are placed at various offsets over the building surface to perform area-averaging of the pressure time series. Considering all wind directions for which measurements are obtained in the wind tunnel, the peak pressures are determined assuming a Gumbel distribution, and are extrapolated to a standard storm duration. The external peak pressure coefficients are then plotted as functions of their corresponding area for various zones of the building enclosure to produce plots similar to the ASCE 7-10 specifications on components and cladding. Results for three gable buildings analyzed in the paper show that the current ASCE 7-10 specifications can severely underestimate the external pressure coefficients for components and cladding of low-rise buildings.

Concrete spalling in prefabricated shield tunnels is a well-recognized issue in underground engineering that impacts the performance and reliability of a tunnel. This paper presents a novel multivariate probabilistic assessment methodology to estimate failure risk, damage growth, and residual useful life of a tunnel structure in terms of its concrete spalling, with the ultimate purpose of facilitating predictive maintenance for prefabricated shield tunnel structures. The multivariate hierarchical approach, a proportional hazard modeling technique, is developed to model the spalling accumulation in a given tunnel segment. The probabilistic model is developed from historically inspected data and characterized to take into account various influencing factors. The most broadly used lognormal distribution is investigated in the development of probabilistic damage accumulation modeling for concrete spalling, with its distribution parameters modeled as a multivariate function of multiple influencing factors. A generic procedure combining multiple qualitative and quantitative techniques is proposed to systematically verify and validate the probabilistic model, considering uncertainties. Advanced lifing analysis approaches, including conditional risk estimation, spalling growth projection, and residual life estimation, are derived from the established empirical model. The probabilistic methodology associated with statistical data analysis techniques is demonstrated by using the spalling data collected from real-world subway tunnels.

In this paper two methodologies are investigated that contribute to better assessment of risks related to extreme rainfall events. Firstly, one-parameter bivariate copulas are used to analyze rain gauge data in the Netherlands. Out of three models considered, the Gumbel copula, which indicates upper tail dependence, represents the data most accurately for all 33 stations in the Netherlands. Seasonal variability is noticeable, with rank correlation reaching maximum in winter and minimum in summer as well as other temporal and spatial patterns. Secondly, an expert judgment elicitation was undertaken. The experts’ opinions were combined using Cooke’s classical method in order to obtain estimates of future changes in precipitation patterns. Experts predicted mostly an approximate 10% increase in rain amount, duration, intensity and the dependence between amount and duration. The results were in line with official national climate change scenarios, based on numerical modelling. Applicability of both methods was presented based on an example of an existing tunnel in the Netherlands, contributing to better estimates of the tunnel’s limit state function and therefore the probability of failure.

The enhanced-Fujita scale (EF-scale) is used to identify tornado intensity. It uses several damage indicators (DIs), each of which has descriptions of the degrees-of-damage (DOD) along with associated wind speeds. Recent research has indicated that for wood-frame, one-family and two-family houses, differences in the structural details result in significant variations in the wind speeds estimated to cause specific levels of damage, particularly with respect to the performance of roofs. This suggests that a single damage indicator for this class of structure may be inadequate. In order to examine this point in detail, the paper focuses on failures of the roof-to-wall-connections (RTWCs) in wood-frame houses, which are frequently damaged in tornadic wind events. Fragility analyses were conducted using an extensive wind-tunnel-based dataset for the determination of the statistics of wind loads and full-scale house test data for the toe-nailed RTWC resistances. The wind load data came from wind tunnel simulations of the atmospheric boundary layer, which are likely to provide upper-bound failure wind speeds for tornadoes. The results indicate that the roof shape and the capacity of RTWCs (i.e., number/type of connections) are the primary factors affecting the failure winds for houses with dominant openings. Recommendations for modifications to the EF-scale to account for these are provided.

Stochastic descriptions and simulations of oceanographic variables are essential for coastal and marine engineering applications. In the past decade, copula-based approaches have become increasingly popular for estimating the multivariate distribution of some variables at the peak of a storm along with its duration. The modeling of the storm shape, which contributes to its impact, is often simplified. This article proposes a vine-copula approach to characterize hourly significant wave heights and corresponding mean zero-crossing periods as a random process in time. The model is applied to a data set in the North Sea, and time series with the duration of an oceanographic winter are simulated. The synthetic wave scenarios emulate storms as well as daily conditions. The results are useful, for example, as input for coastal risk analyses and for planning offshore operations. Nonetheless, selecting a vine structure, finding appropriate copula families, and estimating parameters is not straightforward. The validity of the model, as well as the conclusions that can be drawn from it, are sensitive to these choices. A valuable byproduct of the proposed vine-copula approach is the bivariate distribution of significant wave heights and corresponding mean zero-crossing periods at the given location. Its dependence structure is approximated by the flexible skew-t copula family and preserves the limiting wave steepness condition.

Limitations with the current design using partial factors in Eurocode 7 have been identified. Uncertainties in the material properties are incorporated in both the cautious estimate of the characteristic value and the partial factor. Furthermore, the partial factor is fixed, which limits the opportunities to update the design when additional information is available. A more rational procedure of managing uncertainties in design with partial factors is proposed based on the Bayesian methodology referred to as the extended multivariate approach. The benefits of the approach are illustrated with a case study in which uncertainties of undrained shear strength are characterized for a Swedish clay. The characteristic value and design value is calculated in accordance with the Swedish national annex to Eurocode 7 by adjusting the otherwise fixed partial factor with a conversion factor allowable through EN 1990. The study highlights major benefits in managing uncertainties in a quantifiable and rational way.

The economic input-output interdependency model is a quick decision tool for stakeholders of critical infrastructure to understand the economic impact of a disruptive event. To demonstrate its applications, two real case studies (Pulau Bukom refinery fire in Singapore in 2011 and Tohoku earthquake in Japan in 2011) are presented in the paper. The paper will describe how the model is used to compute the effects on other economic sectors. It also seeks to compare and analyze the computed cascading effects. Understanding the severity and extent of the disruptive event is very important and the input-output interdependency model serves as a quick and cost-effective decision deployment tool for use by relevant stakeholders.

The water supply system (WSS) is a lifeline of the modern city. Transmission and distribution pipes, spatially distributed components of the WSS, are most often vulnerable to failure (leakage, breakage, or burst). Many factors contribute to pipe failure. These factors can be categorized as pipe physical attributes, operational practices, and environmental factors (i.e., climatic factors and soil corrosivity). The impact of failure factors can be static or dynamic (time dependent) in nature. This study quantifies the impact of time-dependent factors on the annual and monthly trends of pipe failures. It considers a dynamic Bayesian network (DBN) as an alternative model for prediction of pipe failures. The DBN extends the static capability of a Bayesian belief network to model dynamic systems. The developed DBN model was trained using annual and monthly data categorized based on combined (metallic pipes) and specific pipe material. The annual model was considered to predict annual pipe failure trends, whereas the monthly model was used to evaluate seasonal variations and trends of pipe failure. Model performance evaluation results show that the proposed models are effective in predicting the trends and expected total number of annual and monthly pipe failures. The DBN model result indicates that the models trained on specific pipe material performed well compared to the model trained using combined metallic pipe data.

The financing of the 2011 Fukushima disaster and the U.K. Hinkley nuclear power plant investment, respectively by the Japanese, and U.K. and Chinese governments and the private sector, provide a strong motivation for this paper to explore deeper the concept of modeling and pricing nuclear catastrophe (N-CAT) risk bonds. Because of the magnitude of the potential liabilities and reinvestments needed, the demand to develop a dependable liability coverage product that can be triggered in a case of emergency is required more than ever, and it should be considered thoroughly. Thus, in the present paper, under a semi-Markov structure environment to model the relationship between claims severity and intensity, the N-CAT risk bond is further explored under various scenarios supporting further the bond sponsors, allowing them to appreciate more their significance. Consequently, the new version of the N-CAT risk bond includes several absorbing and transit states to make it more suitable for practitioners. Additionally, this paper uses the two most commonly used interest rate models and considers four types of payoff functions. Finally, two numerical examples illustrate the main findings.

Stratification in geologic profiles is one of the most significant uncertainties in geotechnical site characterization. In this paper, a three-level probabilistic framework is proposed for geotechnical stratification modeling considering stratigraphic uncertainty. The framework consists of model parameter identification, conditional simulation, and stratigraphic uncertainty quantification. Both boundary-based and category-based stratigraphic models are adopted in the framework, and a heuristic combination model is further recommended to combine the advantages of the boundary-based and category-based models. The geological stratification at a construction site in Hong Kong is characterized to illustrate the probabilistic framework. Results indicate that probabilistic stratification modeling quantifies stratigraphic uncertainty in a rigorous manner. Additionally, the heuristic combination model has the ability to generate almost arbitrary geotechnical strata and to account for material spatial distribution trends and engineering judgment to a certain degree.

In this paper, the challenge of quantifying the uncertainty in stochastic process spectral estimates based on realizations with missing data is addressed. Specifically, relying on relatively relaxed assumptions for the missing data and on a kriging modeling scheme, utilizing fundamental concepts from probability theory, and resorting to a Fourier-based representation of stationary stochastic processes, a closed-form expression for the probability density function (PDF) of the power spectrum value corresponding to a specific frequency is derived. Next, the approach is extended for also determining the PDF of spectral moments estimates. Clearly, this is of significant importance to various reliability assessment methodologies that rely on knowledge of the system response spectral moments for evaluating its survival probability. Further, utilizing a Cholesky decomposition for the PDF-related integrals kept the computational cost at a minimal level. Several numerical examples are included and compared against pertinent Monte Carlo simulations for demonstrating the validity of the approach.

This paper investigates the possibility of consistently identifying a site-specific geotechnical trend function $t(z)$ in the presence of spatial variability $ϵ(z)$, where $z$ denotes depth. The trend function $t(z)$ is represented as the linear combination of prescribed basis functions (BFs), whereas $ϵ(z)$ is modeled as a zero-mean stationary Gaussian stochastic process with unknown standard deviation and scale of fluctuation. It is found that $t(z)$ can be unidentifiable if the single exponential (SExp) autocorrelation model is adopted for $ϵ(z)$. Two mechanisms, overfit and poor fit, that cause unidentifiability are explored. The overfit happens when part of $ϵ(z)$ is falsely fitted by the BFs, whereas the poor fit happens when part of $t(z)$ is erroneously interpreted as spatial variability. Nonetheless, identifiability is also affected by the sounding depth (or data record length) and the number of (statistically independent) soundings. An important feature for the SExp autocorrelation model is that it produces rough realizations with local jitters. If an autocorrelation model that produces smooth realizations is adopted, the trend function can become identifiable. The reason the identifiability for $t(z)$ is related to the smoothness of the realizations is explored. Finally, it is found that the sparse Bayesian learning framework can effectively alleviate overfit but cannot alleviate poor fit.

Earthquakes kill approximately 20,000 persons per annum, with the peak annual loss rate in the last 120 years estimated at 280,000 humans. There are four broad categories of earthquake of interest, ones that have no impact on humans or the human infrastructure, Type I; ones that impact on the infrastructure, but not humans, Type II; ones that impact the infrastructure and humans, but do not cause fatalities, Type III; and ones that kill humans and severely damage infrastructure, Type IV. The fatal earthquake loss model for humans has been statistically demonstrated to be a generalized Poissonian distribution, which is a nonstationary process, governed by the complete set of Type IV earthquakes. This paper investigates some aspects of the impact of world’s seismic risk on the rapidly increasing human population in terms of economic and societal risks. The research work is based on the concept of a systems approach to the current problem, addressing the specific issues of risk, reward, and insurance. New York is used as a model. The results demonstrate the potential and tragic losses, both human and economic, from a major earthquake event striking a major world city.

Nonductile reinforced concrete (RC) frame buildings were found to be vulnerable to earthquakes. Thus retrofitting of these buildings is necessary, as addressed in the recent retrofit program of Los Angeles. A potential retrofit strategy is base isolation, which has many advantages but is not yet routinely used in the United States because of its high cost. To determine whether base isolation is economical when the advantages and costs are compared, seismic performance assessment and cost-benefit analysis are performed on two nonductile RC frame buildings before and after being numerically retrofitted with base isolation, which makes consideration of uncertainties. Location of the buildings is assumed to be Los Angeles, CA. Aftershock ground motions are generated to examine the influence of aftershock hazard. Direct loss, downtime, fatalities, and total loss for each of the buildings are investigated. Results suggest that base isolation can reduce the seismic loss, downtime, and fatalities effectively. Besides, the benefits can outweigh the retrofit costs. The effect on the results caused by indirect losses from downtime and fatalities is quantified and shown to be substantial. Aftershock hazard is also found to be at a relatively significant level.

Different variabilities and uncertainties (e.g., inherent variability, measurement errors, statistical uncertainty, and transformation uncertainty) in site characterization are usually lumped together as total variability and used subsequently in geotechnical reliability-based design or probabilistic geotechnical analysis. However, only the inherent variability affects the observed performance of geotechnical structures. Knowledge uncertainties (i.e., measurement errors, statistical uncertainty, and transformation uncertainty) have no impact on performance of geotechnical structures, although they may significantly affect the calculated failure probability or risk. In this paper, a comparative study is performed using the total variability approach, which lumps various uncertainties together, and the Bayesian inverse analysis approach, which explicitly characterizes inherent variability, to perform probabilistic characterization of effective friction angle from cone penetration test (CPT) results. The probabilistic characterization results from the two approaches are used in a Eurocode 7 evaluation example of foundation design to explore their impacts on reliability-based design of foundation. It is found that the total variability approach leads to much larger uncertainty in geotechnical properties than the Bayesian inverse analysis approach. Consequently, the total variability approach leads to larger calculated failure probabilities for the foundation and subsequently produces a foundation design that is much more conservative than those from the Bayesian inverse analysis approach.

Resilience, described in this work as a function of vulnerability and recoverability dimensions, is an increasingly important concept in preparing for and responding to disruptions in many kinds of cyber-physical-social systems, including networks. This paper focuses on the vulnerability dimension of resilience, proposing (1) a bi-objective optimization formulation to devise defense strategies across a range of diverse attack scenarios, and (2) a three-step solution approach of approximating Pareto-optimal defense strategies for each attack scenario and aggregating characteristics of strategies across attacks to identify a robust defense strategy. An example network illustrates the formulation and solution approach, identifying contributions to enhance network resilience analytics.

The stationary Markov process model is widely used to predict the geological conditions in tunnel excavation projects. However, the validity of the stationary assumption made in the model is questionable. The prediction error caused by the assumption has not been investigated in previous studies. In this study, the significance of a stationary Markovian zone detection is evaluated by comparing the predicted geological conditions with the real soil layer distributions in boreholes. A new method is proposed to detect the stationary Markovian zones in a tunnel-covered area. Borehole data from Perth, Australia are collected to illustrate the significance of the stationary Markovian zone detection and the effectiveness of the proposed method. The results show that the stationary assumption leads to considerable errors in the predicted number, location, and thickness of soil layers. The proposed method is robust with respect to the starting point of a detection.

This paper proposes a framework to evaluate hurricane resilience of residential communities. Building resilience often refers to community resilience instead of individual resilience. Few efforts have been made to relate individual resilience to community resilience in a quantitative manner subjected to hurricane events. The proposed framework includes disaggregating hurricane community resilience to the individual level and quantifying the resilience for individual residential buildings. The framework not only consists of hurricane damage assessment including hurricane fragility analysis, reliability analysis, and loss analysis, but also analysis of recovery time, which makes the assessment framework more comprehensive and accurate. Sources of uncertainties in the framework include (1) structural modeling uncertainty (e.g., changes in roof panel resistance due to effects of corrosion on metal fasteners), (2) load uncertainty (e.g., hurricane wind characteristics), (3) uncertainty in direct and indirect loss estimation, and (4) recovery time modeling uncertainty. It is found that a community could meet certain community resilience goals as long as the individual buildings in the community could meet the individual building resilience goals.

Estimating properties of the distribution of the peak of a stochastic process from a single finite realization is a problem that arises in a variety of science and engineering applications. Furthermore, it is often the case that the realization is of length $T$ whereas the distribution of the peak is sought for a different length of time, $T_1>T$. The procedure proposed here is based on a peaks-over-threshold extreme value model, which has an advantage over classical models used in epochal procedures because it often results in an increased size of the relevant extreme value data set. For further comparison, the translation approach depends upon the estimate of the marginal distribution of a non-Gaussian time series, which is typically difficult to perform reliably. The proposed procedure is based on a two-dimensional Poisson process model for the pressure coefficients $y$, above the threshold $B$. The estimated distribution of the peak value depends upon the choice of the threshold. The threshold choice is automated by selecting the threshold that minimizes a metric that captures the trade-off between bias and variance in estimation. Two versions of the proposed new procedure are developed. One version, denoted FpotMax, includes estimation of a tail length parameter with a similar interpretation of the generalized extreme value distribution tail length parameter. The second version, denoted GpotMax, assumes that the tail length parameter vanishes.

Vegetation enhances the strength characteristics of soil by inducing suction in soil through root water uptake. This in turn affects the stability of a wide variety of geotechnical structures, including slopes. In order to properly assess the stability of vegetated slopes, it is essential to address the hydrological (suction, $\psi$, and volumetric water content, ${\theta }_w$) and mechanical (cohesion, $c$, and angle of internal friction, $\varphi$) properties of vegetated soil. Both hydrological and mechanical parameters share negative correlations among each other, by virtue of which the stability of vegetated slopes is significantly affected. This study addresses the individual and combined effects of both $\psi -{\theta }_w$ and $c-\varphi$ correlations on the reliability of vegetated slopes in a probabilistic framework. The aforementioned dependencies are incorporated utilizing copula theory, which establishes a unique relationship between the individual bivariate density functions and their respective marginal distributions. The bivariate data of the aforementioned hydrological and mechanical parameters are obtained from a field monitoring program conducted on a homogeneously compacted vegetated slope. To better comprehend the unsaturated behavior of vegetated slopes, the effect of $\psi -{\theta }_w$ correlation on the soil water characteristic curve (SWCC) of vegetated soil is established. It is found that vegetation induces uncertainties in $c-\varphi$ parameters to a greater extent than does $\psi -{\theta }_w$ correlation. The randomness induced by vegetation is found to be more in the saturation zone than in the desaturation zone. Moreover, the impact of mechanical correlations on slope stability is found to be more significant than that of hydrological correlation.

The effective pretress of a strand is very important for prestressed concrete (PC) bridges because it can influence cracking and deflections. At present, the effective prestress is obtained by means of calculating the theoretical prestress loss, which is a theoretical value. However, because of uncertainty in estimating prestress losses and tensioning prestress tendons, the actual effective prestress can be very different from the theoretical values. In this paper, the transverse-tensile-incremental method (TTIM) is applied to investigate the effective prestress of prestressing tendons for PC bridges. First, a ${\mathrm{\varphi }}^{j}15.2$-type strand was tensioned with a fixed force on a machine, and the strand tensile force was tested by a specific instrument manufactured according to the TTIM. An error analysis between the fixed force and tested force was conducted to certify the instrument’s accuracy. The results showed that the error range was from 0.03 to 3.71% when the prestressing tendons tension was between 60 and 200 kN (for the ${\mathrm{\varphi }}^{j}15.2$-type strand, the tensile stress mean was from 428.6 to 1,428.6 MPa). Therefore, it can be accepted that the TTIM could be used in actual testing. Second, a total of 141 on-site measurements were collected and analyzed. A statistical parameter $K_{\mathrm{EP}}$ was introduced to be the statistical objective for the effective prestress of a strand. This parameter was defined as the ratio of mean effective prestress from the testing value to that from theoretical calculation. A statistical analysis was carried out to determine an appropriate probability distribution for the $K_{EP}$ using commercially available software. Based on the collected specimens, a normal distribution was found to be an appropriate probability distribution, and the distribution parameter $K_{EP}$ is $\sim N(0.983,0.066)$. Finally, an example of a reliability assessment was demonstrated using a simply supported T-beam. The anticrack reliability index for the section bottom of the midspan is $\beta =1.223$ when the uncertainty of effective prestress was considered, whereas $\beta =2.488$ if the uncertainty is neglected. Therefore the anticrack reliability would be overestimated if the uncertainty of effective prestress is not included. The method proposed in this paper can provide a way to properly consider the uncertainty of effective prestress and subsequently be used in anticrack reliability assessments for PC bridges.