Performance-Based Coastal Engineering Framework

The changing dynamics of coastal regions and climate pose severe challenges to coastal communities around the world. Such challenges are exacerbated by an increase in population over time coupled with the aging of existing infrastructure, rise in property value, and the expected shifts in the frequency and intensity of natural hazards due to climate change. Thus, effective planning of engineering projects and resilience strategies in coastal regions must not only address current conditions but also take into consideration the expected changes in the exposure and multi-hazard risk in these areas. Performance-based engineering methodologies allow for quantification of a system’s response and performance in terms of a decision variable DV, which represents the desired/required performance objectives of the system. However, existing performance-based engineering frameworks generally neglect time-varying factors, do not explicitly consider potential cascading-effects, and miss the opportunity to leverage evidence – in the form of data – as it becomes available. To address these gaps, this thesis proposes the first performance-based framework specific for coastal structures and systems that is flexible enough to accommodate uncertain time-varying factors, multi-hazard conditions, systems with multiple structures, cascading-effects, and different performance metrics. The framework consists of six basic components: performance objectives, hazard analysis, structural characterization, structural analysis, damage analysis, and performance analysis. To compute the marginal probability distribution of the decision variable, this study proposes the implementation of a probabilistic graphical model in the form of Bayesian networks (BNs) and dynamic Bayesian networks (DBNs). BNs and DBNs allow modeling the causal dependence among the variables, offering an efficient sampling strategy and facilitating the incorporation of expert knowledge. Different applications of the proposed PBCE framework at the structural level and at the regional level are presented, making emphasis on the analysis of time-varying factors and cascading-effects. This study also proposes a methodology to expand the PBCE framework to the community scale in support of evaluating resilience and adaptation strategies. Beyond the conceptual and methodological advancement of PBCE, this thesis offers new insights on the spatial and temporal evolution of dynamic processes in coastal settings using residential structural portfolios and housing as a focal point. Moreover, to expose potential disparities in the impact of coastal hazards to different sectors of the community in the short- and long-term, a correlation analysis between immediate damage and social vulnerability factors, as well as between the recovery index and social vulnerability factors is pursued. Finally, the adaptation of the PBCE framework to the analysis of hurricane-induced debris at the regional level is introduced. Debris is one of the most challenging cascading effects posed by hurricane events, causing large financial and logistical burdens to coastal communities. Given the lack of understanding of the debris generation and spreading process, a Bayesian network analysis that leverages domain knowledge and a probabilistic data-driven model to approximate system response is presented. With this aim, a comprehensive database that expands across human-built-natural systems is used to inform and develop a predictive Gaussian process model of debris volume. Test cases for PBCE highlight the potential for the framework to efficiently capture different unique challenges in coastal environments, along with practical insights regarding system performance. For example, the analysis of cascading failure in a two-frame system due to added debris loading showed that neglecting cascading-effects can underestimate the expected economic losses up to 87%. In the case study of building performance in Galveston Island (i.e., application of time-varying factors at the regional scale), results revealed that changing climate conditions exacerbate the probability of failure of the building stock and associated housing recovery. For instance, the percentage change between 2030 and 2050 in average loss per block group in Galveston Island ranges from a minimum of 86% increase to a maximum of 245% increase. Results throughout the case study applications indicate that the incorporation of evidence reduces the uncertainty in the parameter estimations. Finally, this thesis showcased how domain knowledge and data-driven models facilitate the performance evaluation of complex systems, such as hurricane-induced debris, setting the basis for informed decision-making and policy development on resilient debris management. Overall, the proposed PBCE framework allow a comprehensive performance evaluation of coastal systems now and into the future, while analyzing the consequences of their performance on interdependent systems and coastal communities. The resulting PBCE framework can support recovery efforts, decisions on retrofitting and mitigation strategies, integration of domain knowledge in performance analysis of infrastructure systems, as well as the identification of vulnerable regions and critical components in coastal multi-hazard regions.

Name thesis:degree_name

Doctor of Philosophy

Level thesis:degree_level

Doctoral

Discipline thesis:degree_discipline

Engineering

Grantor

Rice University

Year dc:date.issued

2022

Author dc:creator

• Gonzalez Duenas, Catalina

Advisor dc:contributor.advisor

• Padgett, Jamie E.

• performance-based engineering
• coastal engineering
• time-varying factors
• cascading effects
• dynamic Bayesian network
• debris
• Gaussian process

Statement dc:rights

• Copyright is held by the author, unless otherwise indicated. Permission to reuse, publish, or reproduce the work beyond the bounds of fair use or other exemptions to copyright law must be obtained from the copyright holder.

Language dc:language.iso

eng

Handle dc:identifier.uri

https://hdl.handle.net/1911/114155

OAI identifier oai:identifier

oai:repository.rice.edu:1911/114155