Browsing by Author "Ellingwood, Bruce R., committee member"
Now showing 1 - 9 of 9
Results Per Page
Sort Options
Item Open Access A resilience-based decision framework to determine performance targets for the built environment(Colorado State University. Libraries, 2018) Masoomi, Hassan, author; van de Lindt, John W., advisor; Ellingwood, Bruce R., committee member; Mahmoud, Hussam N., committee member; Senior, Bolivar, committee memberCurrent design codes and standards focus on the design of individual facilities. A typical building is designed with the objective of the life safety of occupants. Even performance-based design approaches assess the required physical performance of an individual structure in order to satisfy prescribed criteria for that structure individually. Thus, even these performance objectives are likely not sufficient for a broad view of community-resilience goals. A modern community is made up of highly coupled networks, and disruptions within one or more networks may lead to disruptions to other networks. If a large number of buildings within a community become non-functional for a long time following an event, either because of physical damage or loss of utilities such as electric power and/or water, the consequences may affect other parts of the community such that, eventually, significant socioeconomic losses occur. Therefore, the current approach for designing individual physical components within a community can be reimagined such that it not only takes into account the performance of a component individually after a catastrophic event but also considers the consequences its design has on a community. The main purpose of this dissertation is to develop a methodology that links the performance of components within the built environment to community-level resilience goals by considering the dependencies and cross-dependencies between components and networks. Therefore, ultimately, this methodology enables disaggregation of the community-level objectives into a set of performance targets for the components of the built environment, which leads itself to the needs of policymakers and community leaders in order to make long-term planning decisions for a community.Item Open Access Damage analysis and mitigation for wood-frame structures subjected to tornado loading(Colorado State University. Libraries, 2016) Standohar-Alfano, Christine Diane, author; van de Lindt, John W., advisor; Ellingwood, Bruce R., committee member; Heyliger, Paul R., committee member; Schumacher, Russ S., committee memberTornadoes are one of the most devastating natural hazards that occur in the United States. While there is an average of approximately 1200 tornadoes per year across the country, the annual likelihood of experiencing a tornado at a particular location is quite small due to their relatively small size. However, the high consequence of a tornado strike necessitates the determination of geographic tornado hazard. A methodology to estimate the annualized probabilistic tornado hazard over the contiguous U.S. was developed and used the most recent 38 years of climatological tornado data. Furthermore, with the use of detailed damage surveys after the April 3-4, 1974 and April-May, 2011 tornado outbreaks, an empirical method was developed and applied to account for the gradient of wind speed along a tornado’s path length and path width. From this, a probabilistic tornado hazard index was developed across the United States which quantified the annual probability of experiencing a tornado of any strength on the Enhanced Fujita scale. Tornado hazard curves were developed from the tornado hazard analysis at six illustrative locations which varied as a function of location-specific occurrence rates. Five different residential wood-frame building archetypes were designed at each of the locations based on current residential building code and/or practice. Fragilities for the roof sheathing, truss to wall top-plate, and wall-to-foundation connections were developed for each archetype. At each of the six locations, the fragility curves for the locally adopted residential building code were convolved with the tornado hazard curve at that specific location in order to compute annual failure probabilities for select components along the vertical load path. This was one of the first times unconditional risk of component failure due to tornadoes has been computed since the tornado hazard curve was convolved with the fragility curves. These probabilities quantify failure probabilities of residential wood-frame construction components to tornado winds. In addition, the more wind-resistant Florida residential building code is applied to other locations in the U.S., fragilities are developed and convolved, and failure probabilities for these modified buildings are computed. This resulted in a quantitative measure of risk reduction from tornadoes by using strengthened construction at various locations across the country. The convolved failure probabilities were first developed for individual components. The system level behavior of the entire structure was also assessed and included the correlated dependencies between individual components. Results indicate that stricter building codes may be beneficial in areas with a high annual tornado risk, such as Tornado Alley. The final portion of this work used a simplified property loss model applied to the April 25-28, 2011 tornado outbreak. This was one of the largest tornado outbreaks in U.S. history and resulted in over $5B in property loss. In order to determine property loss over a broad area, census data regarding household income and home market value was utilized. The performance of manufactured homes had to be considered in conjunction with wood-frame residential construction since the tornado outbreak impacted the southern U.S. which has a high number of manufactured homes. Using the system level fragility analysis, property loss was estimated based on both locally adopted residential codes and the stricter guidelines described in the Florida Residential Building Code. Results indicate that using strengthened construction methodologies would reduce property loss up to 40% as compared to current design guidelines.Item Embargo Development of a prediction model for windborne debris damage assessment of coast communities under hurricanes(Colorado State University. Libraries, 2023) Dong, Yue, author; Guo, Yanlin, advisor; van de Lindt, John W., committee member; Ellingwood, Bruce R., committee member; Gao, Xinfeng, committee memberUrban high-rise building envelopes may suffer severe damage induced by windborne debris during hurricanes. Breaches in urban building envelopes may lead to cascading building content loss due to rain intrusion and building functionality loss for long periods of time, thus disrupting the resilience of urban coastal communities. Such a societal disruption may become worse due to the rapid growth of population and development of economy in coastal communities. Objective of this dissertation is to develop an efficient prediction model for assessing windborne debris damage to building envelopes and apply it to the hurricane scenarios identified through a de-aggregation process to enable risk-informed resilience assessment of coastal communities for windborne debris impact. A set of scenario hurricanes corresponding to a stipulated return period (RP) for resilience assessment of coastal communities for debris impact is systematically identified using the de-aggregation approach proposed in this dissertation. The existing building envelope damage assessment models for urban high/mid-rise buildings often neglect the geometry of building clusters or simply assume a homogeneous configuration, which can introduce errors and uncertainties for urban building clusters with varying geometries and layouts. This dissertation proposes a new fragility modeling approach for urban buildings envelopes, which explicitly considers geometric configurations of urban buildings, to improve the accuracy of risk assessment for urban buildings. Estimating the urban wind field that drives the debris flight is critical for constructing fragilities for debris damage. The traditional tools for simulating urban wind fields, such as wind tunnel tests and Computational Fluid Dynamic (CFD), are usually expensive or time-consuming for complex urban environments and cannot offer an efficient prediction of the urban wind field that is sufficient for risk assessment at a community level. Therefore, an efficient machine learning (ML) based prediction model of wind fields around building clusters is proposed using conditional Generative Adversarial Networks (cGANs). Uncertainty analysis is conducted on the models and parameters involved in this procedure of debris damage assessment. The uncertainty in the final prediction of windborne debris damage is evaluated through uncertainty propagation among models and parameters. Sensitivity analysis with some critical factors is also conducted to identify the dominant contributors to the uncertainty in the estimation of debris damage. In the end, windborne debris damage on building envelopes in virtual communities under synthetic hurricanes scenarios is investigated to illustrate the damage assessment procedure for windborne debris impact at the community level.Item Open Access Dynamic assessment of the long-span cable-stayed bridge and traffic system subjected to multiple hazards(Colorado State University. Libraries, 2016) Zhou, Yufen, author; Chen, Suren, advisor; Ellingwood, Bruce R., committee member; Mahmoud, Hussam N., committee member; Sakurai, Hiroshi, committee memberCritical infrastructure systems, such as long-span bridges, offer the underlying foundation for many aspects of modern society, such as national security, quality of life and economy. Although the total number of long-span bridges is relatively small compared to short-span and medium-span bridges, long-span bridges often serve as backbones for critical interstate transportation corridors and also evacuation routes. Any traffic disruption due to bridge damage, failure, retrofitting or even major traffic accidents following some hazards can become disastrous to local community and emergency response efforts, underscoring the importance of the continued integrity, functionality and resilience following hazardous conditions. Wind and traffic are the major service loads for long-span bridges. The extreme loads may include those caused by various natural or man-made hazards, such as earthquake, hazardous winds (hurricane, tornado), fire, blast, vehicle and barge collision etc. Compared to other hazards, hazardous wind and earthquake are particularly critical for long-span bridges, primarily due to their significant threats to the global structure performance and challenges of appropriately modeling the dynamic coupling effects between the bridge, traffic and hazards. In addition, there is another disastrous event: cable loss, which is very unique and critical for cable-supported bridges and could be caused by various natural and man-made hazards. There exist major challenges in the current state of the art on rationally predicting the long-span bridge performance subjected to multiple service and extreme loads. These challenges include realistic load characterization, methodological limitations and considerations of uncertainties. A suite of holistic analytical frameworks of long-span cable-stayed bridges subjected to various service and hazardous loads are developed, with which insightful numerical analyses of the bridge performance subjected to these loads are carried out in this dissertation. Firstly, two general dynamic assessment frameworks are developed based on the mode superposition and finite element methods respectively for a long-span cable-stayed bridge and traffic system subjected to multiple threats, such as stochastic traffic, wind and some hazardous loads. Although developed based on a long-span cable-stayed bridge, the frameworks can be readily applied to long-span suspension bridges as well as bridges with shorter spans. In both simulation platforms, the bridge model and all individual moving vehicles in the stochastic traffic flow are directly coupled under multiple excitations from bridge deck roughness and other external dynamic loads. Through the established simulation platforms, the global dynamic responses of the bridge and each individual vehicle subjected to various service and extreme loads can be rationally predicted in the time domain. Secondly, built on the proposed general simulation platforms, a novel dynamic safety assessment model and a vehicle ride comfort evaluation model for the bridge-traffic system are further developed. Thirdly, also extended from the proposed simulation platforms, both deterministic and reliability-based assessment frameworks for long-span cable-stayed bridges subjected to breakage of stay cables are established by considering more rational service load conditions as well as cable-breakage characterizations. Lastly, in addition to the in-house programs focusing on research purposes, a hybrid simulation strategy for the bridge under traffic and seismic excitations and a time-progressive simulation methodology for cable breakage events are also developed by taking advantage of the strength offered by commercial finite element software, e.g., SAP2000. These SAP2000-based strategies are expected to facilitate design engineers to more easily understand and conduct the related analyses in future engineering practices.Item Open Access Numerical simulations of binary mixtures under gravity deposition using the discrete element method(Colorado State University. Libraries, 2021) Jiang, Chao, author; Heyliger, Paul, advisor; Bareither, Christopher, committee member; Ellingwood, Bruce R., committee member; Venayagamoorthy, Karan, committee member; McGilvray, Kirk, committee memberBinary granular mixtures are frequently used in manufacturing, geotechnical engineering, and construction. Applications for these materials include dams, roads, and railway embankments. The mixing process requires dealing with particles with varying sizes and properties, and the complex composite nature of these mixtures can bring unpredictable results in overall performance. At present, there are no specifications for mixing these materials that can be used to quantify the levels of mixing and give estimates of the overall bulk properties. In this study, the Discrete Element Method (DEM) is used to examine the mechanics of the mixing process and give guidelines on how to achieve a well-mixed aggregate. A comprehensive non-linear visco-elastic damping collision model was developed to better represent the interactions between two dissimilar particles. A general Hertz model was applied for describing the normal force but a refined non-linear spring model was generated to imitate the friction force behavior without having to consider the entire loading history. A transition zone revealing the interactions between static and dynamic friction forces was shown in our numerical results. A moment resistance model was also added to capture the behavior of particle surface asperities and the damping force was calculated using relative motion. An alternative condition was applied to determine the end of a collision. Excellent agreement was found with well-established benchmark solutions and new results are also provided for future comparisons. Using this new DEM model, the mixing process of binary unbonded particles was studied using the effects of the number and position of geometric mixing obstacles and the number of mixing iterations. It was found that the mixing degree can be best quantified by measuring the spatial variation of the volume ratio φv. It was also found that small adjustments in the geometric position of the mixing obstacles could have a significant impact on the final mixing parameters. Surprisingly, the results indicate that two mixing iterations provided almost identical levels of mixing regardless of the number and nature of mixing obstacles. Estimates of the bulk elastic constants were provided and showed a high level of anisotropy as measured by the Poisson ratios for the horizontal versus vertical planes of the control volume. Particle crushing is a typical characteristic of many granular materials and can influence the mixing process, and it is possible to model non-particulate materials by bonding individual spheres together. The particle interactions and possibly impact with mixing barriers can result in the fracture of these solids as the allowable bond strength is exceeded. Therefore, the strength of the bond between individual particles that can be part of the mixing process is a critical parameter. The parallel bond model of Potyondy and Cundall (2004) was extended with the present DEM model was used to study the effects of bond strength on the mixing and mechanical properties of binary mixtures. Three types of particle blocks were studied for this purpose: unbonded, weakly bonded, and strongly bonded particles. The bonded particles result in a wider range of reflection angles as the particles interact with geometric mixers and simultaneously change and improve the level of mixing. Overall, these simulations serve to established specific guidelines and provide a basis for field-level mixing operations. They also provide some levels of expectation for the final mixing and bulk elastic behavior for the final aggregates.Item Open Access Predicting ductile fracture in steel connections(Colorado State University. Libraries, 2016) Wen, Huajie, author; Mahmoud, Hussam N., advisor; Atadero, Rebecca A., committee member; Chen, Suren, committee member; Ellingwood, Bruce R., committee member; Puttlitz, Christian M., committee memberSeparation of material, known as fracture, is one of the ultimate failure phenomena in steel elements. Preventing or delaying fracture is therefore essential for ensuring structural robustness under extreme demands. Despite the importance of fracture as the final stage during inelastic response of elements, the underlying mechanisms and the factors influencing the onset and progression of fracture have not been fully investigated. This is particularly the case for ductile fracture where significant pre-crack deformations are present. Existing approaches geared at predicting brittle fracture, marked by little to no plastic deformation, have been proven inadequate for capturing ductile fracture. Ductile fracture is dependent on two stress state parameters, the stress triaxiality and Lode parameter, which correspond, respectively, to two kinds of work hardening damage – that is hydrostatic and deviatoric stress components. The role of stress triaxiality on ductile fracture has been well defined and implemented in various models over the past several decades. Only until recently, however, has the role of Lode parameter been identified as an important factor for accurate prediction of ductile fracture. In general, no reliable fracture prediction methods are present that are consistent throughout the whole range of stress states, where the stresses are dominated by either tension loading, shear loading, or a combination of both. In this study, a new ductile fracture criterion based on monotonic loading conditions is first developed based on analysis and definitions of the two stress state parameters and subsequently extended to the reverse/cyclic loading conditions. The extension from monotonic to cyclic loading is based fundamentally on the fact that as long as large pre-crack plastic strain fields exist, the inherent mechanism in both loading cases can be viewed to be the same. Although the inherent mechanism is the same for both loading cases, extending the model to the reverse loading conditions required the inclusion of the effects of nonlinearity of the damage evolution rule as well as the loading history. The two criteria, monotonic and cyclic, are then validated on the coupon specimen level through comparisons between predicted fracture strains and their experimental equivalents for various metal types and steel grades that are available in the literature. The newly developed models offer improvements to existing known ductile fracture criteria in terms of both accuracy and practicality. Following the validation of the fracture model on the coupon specimen level, the model is employed on the connection level, up to and including failure, to evaluate block shear failure for gusset plate and coped beam connections under monotonic loading and shear links under cyclic loading. The chosen connection types are dependent on stress triaxiality (tension) and Lode parameter (shear) and are therefore appropriate for the validation of the ductile fracture model. For the block shear failure, prediction accuracy is verified through comparisons with results from corresponding laboratory tests, in the perspective of load versus displacement curves, fracture profiles, and fracture sequences. Some underlying mechanism of block shear is also explored and explained for the first time. Following the same modeling procedure, parametric studies on geometric effects on block shear failure is conducted. Three different block-shear failure modes and one bolt hole tear out mode are captured in the simulations and suggestions on design code changes are provided. For the shear links, which are typically employed in Eccentric Braced Frames, simulation of fracture under reverse/cyclic loading is also conducted and verifications are performed through comparisons with their previous experimental results. The fracture-associated variables are included in the cyclic loading analysis through deriving an implicit integration algorithm for the material constitutive equations with combined hardening, which was integrated in the simulation using a user-defined material subroutine VUMAT.Item Open Access Resilience of healthcare and education networks and their interactions following major earthquakes(Colorado State University. Libraries, 2021) Hassan, Emad Mohamed Shafik, author; Mahmoud, Hussam N., advisor; Ellingwood, Bruce R., committee member; van de Lindt, John W., committee member; Zahran, Sammy, committee member; McCabe, Steven, committee member; Cerato, Amy, committee memberHealthcare and education systems have been identified by various national and international organizations as the main pillars of communities' stability. Ensuring the continuation of vital community services such as healthcare and education is critical for minimizing social losses after extreme events. A shortage of healthcare services could have catastrophic short-term and long-term effects on a community including an increase in morbidity and mortality, as well as population outmigration. Moreover, a shortage or lack of facilities for K-12 education, including elementary, middle, and high schools could impact a wide range of the community's population and could lead to impact population outmigration. Despite their importance to communities, there are a lack of comprehensive models that can be used to quantify recovery of functionalities of healthcare systems and schools following natural disasters. In addition to capturing the recovery of functionality, understanding the correlation between these main social services institutions is critical to determining the welfare of communities following natural disasters. Although hospitals and schools are key indicators of the stability of community social services, no studies to date have been conducted to determine the level of interdependence between hospitals and schools and their collective influence on their recoveries following extreme events. In this study, comprehensive frameworks are devised for estimating the losses, functionality, and recovery of healthcare and educational services following earthquakes. Success trees and semi-Markov stochastic models coupled with dynamic optimization are used to develop socio-technical models that describe functionalities and restorations of the facilities providing these services, by integrating the physical infrastructure, the supplies, and the people who operate and use these facilities. New frameworks are proposed to simulate processes such as patient demand on hospitals, hospitals' interaction, student enrollment, and school administration as well as different decisions and mitigation strategies applied by hospitals and schools while considering the disturbance imposed by earthquake events on these processes. The complex interaction between healthcare and education networks is captured using a new agent-based model which has been developed in the context of the communities' physical, social, and economic sectors that affect overall recovery. This model is employed to simulate the functional processes within each facility while optimizing their recovery trajectories after earthquake occurrence. The results highlight significant interdependencies between hospitals and schools, including direct and indirect relationships, suggesting the need for collective coupling of their recovery to achieve full functionality of either of the two systems following natural disasters. Recognizing this high level of interdependence, a social services stability index is then established which can be used by policymakers and community leaders to quantify the impact of healthcare and educational services on community resilience and social services stability.Item Open Access Simulation-based tsunami evacuation risk assessment and risk-informed mitigation(Colorado State University. Libraries, 2021) Wang, Zhenqiang, author; Jia, Gaofeng, advisor; Ellingwood, Bruce R., committee member; Mahmoud, Hussam N., committee member; Quinn, Jason C., committee memberEarthquake-induced tsunami can be very destructive involving significant loss of life. Evacuation to safety zones is regarded as one of the most effective ways to save lives from the tsunami strike due to the limited effectiveness of structural countermeasures. However, it is extremely challenging to successfully evacuate many people under the multi-hazard environment within a condensed time frame, especially under the near-field tsunami. Proper evacuation planning is crucial to support effective evacuation and reduce casualty. For effective evacuation planning, it is important to better understand the complex evacuation behavior for recommending proper response and behavior in an emergency. Also, it is important to have a clear picture of evacuation risk for informing policy and decision-making. Furthermore, it is important to identify effective pre-event mitigation strategies for effective risk reduction. Important limitations exist in current research on the above aspects. Tsunami evacuation simulation using the agent-based model has been used to investigate the complex evacuation behavior; however, existing agent-based evacuation models usually neglect or simplify many important factors and/or mechanisms associated with the evacuation. The neglect or simplification would make the evacuation simulation less realistic and hence a good understanding of evacuation behavior challenging. For the quantification of tsunami evacuation risk, a systematic framework that can address complex evacuation models and uncertainty (including aleatory and epistemic uncertainties) models is needed; however, no such framework has been developed for the quantification of tsunami evacuation risk. Also, some important uncertainties such as that in the seismic damage to the bridge are usually neglected or the uncertainty quantification is simplified. In this case, it would be difficult to assess the evacuation risk accurately and provide a clear picture of the evacuation risk. For effective pre-event evacuation risk mitigation, the effectiveness of different mitigation strategies needs to be quantitatively evaluated to identify more effective strategies. However, the effectiveness of the mitigation strategy is usually evaluated more qualitatively than quantitatively. Furthermore, the evaluation is typically conducted without systematically considering various uncertainties, which makes the identified strategies not robust to uncertainties. In tsunami evacuation risk assessment and mitigation, risk evaluation using general stochastic simulation techniques (e.g., Monte Carlo simulation) typically entails significant computational challenges. Efficient algorithms are needed to alleviate such computational challenges and facilitate such tasks. To bridge the above knowledge gaps, this research proposes a generalized framework for simulation-based tsunami evacuation risk assessment and risk-informed mitigation. The framework is built layer by layer through integrating tsunami evacuation simulation using agent-based modeling (ABM) technique, simulation-based evacuation risk assessment, sensitivity analysis of evacuation risk, and risk-informed evaluation of mitigation strategies. An improved agent-based tsunami evacuation model is developed for more realistic tsunami evacuation simulation by incorporating many of the typically neglected or simplified but important factors and/or mechanisms in the evacuation. Using the proposed agent-based evacuation model, a simulation-based framework is proposed to quantify the evacuation risk, in which various uncertainties (including aleatory and epistemic uncertainties) associated with the evacuation are explicitly considered and modeled by proper selection of probability distribution models. Sensitivity analysis of evacuation risk with respect to the epistemic uncertainty is performed, and the sensitivity information can be used to guide effective epistemic uncertainty reduction and hence for more accurate risk assessment. Also, sensitivity analysis is performed to identify critical risk factors, and the sensitivity information can be used to guide effective evacuation modeling and selection of candidate risk mitigation strategies. Risk-informed evaluation of different types of candidate mitigation strategies (including infrastructural and non-infrastructural strategies) is conducted to identify more effective strategies that are robust to uncertainties. Efficient sample-based approaches are developed to alleviate the computational challenges in evacuation risk assessment, sensitivity analysis, and risk-informed evaluation of mitigation strategies. As an illustrative example, the proposed framework is applied to tsunami evacuation risk assessment and risk-informed mitigation for the coastal community of Seaside, Oregon.Item Open Access Thickness effects in the free vibration of laminated magneto-electro-elastic beams and plates(Colorado State University. Libraries, 2016) Jiang, Chao, author; Heyliger, Paul R., advisor; Ellingwood, Bruce R., committee member; Holland, Troy B., committee memberA semi-analytical discrete-layer approach is used to evaluate thickness effects in the free vibration of laminated magneto-electro-elastic beams and plates under various lateral boundary conditions. To match the primary physical phenomenon and simplify the study, piecewise continuous approximations are used through the thickness direction and either continuous global polynomial or trigonometric functions are used to simulate the deflection in axial or planar displacement fields. Thin plate models can be recovered to predict frequency estimation for various boundary conditions and compared with continuum-based theories using more complex approximations. Based on symmetry the natural vibratory modes can be grouped to optimize computation. Numerical examples are used to show the thickness effects, with non-dimensional frequencies computed to multiple plates under six lateral boundary conditions: simply supported, clamped, and four different combinations of free and clamped/simply-supported edges. As the out-of-plate dimension becomes small and two opposite sides are free, this methodology can also be applied to beams under simply-supported, fixed-fixed and cantilever support conditions. Along with the influence of electro-elatstic and magneto-elastic coupling, the results of these analyses clearly illustrate the thickness effects within laminated plates by showing how the results vary with length/thickness ratio. Finding the accurate ratio varied with thickness is expected to provide useful specifications for the further study and design of multilayered magneto-electro-elastic beams and plates.