Reliability of Civil Intrastructure Systems Subject to Multiple Hazards
Speaker: Paolo Gardoni, Texas A&M University
Series: CEE Departmental Seminars
Location: Bowen Hall Auditorium
Date/Time: Monday, April 18, 2011, 4:30 p.m. - 5:30 p.m.
Transportation networks are critical lifeline systems. Their functionality is of primary importance to promote the well-being and economic growth of a community, yet they can be severely damaged by multiple hazards during their service life. In particular, earthquakes experienced over the last 20 years have resulted in the failure of many bridges. While the failures due to earthquakes are significant because they typically affect several bridges at the same time and thereby hinder – or completely shut down – the transportation infrastructure system at the very moment it is most needed, other causes of failure for a single bridge are more frequent. Scour and collision of vehicles have been identified as two of the most probable causes of bridge failure. Furthermore, bridges are often subject to deterioration mechanisms during their service life like low-cycle fatigue due to their load history, corrosion, alkali-silica reaction (ASR), and delayed ettringite formation (DEF). Such deterioration mechanisms are likely to reduce the structural integrity of reinforced concrete structures and thereby negatively affect their serviceability and safety. Finally, bridges are part of a transportation network. Therefore, their importance as links in the network should be accounted for to optimally allocate resources for preventive maintenance or retrofitting actions.
A comprehensive Bayesian framework is formulated to construct probabilistic models for the seismic demands on reinforced concrete (RC) bridges and the corresponding structural capacities. To facilitate the use in practice of the models, they are constructed by developing correction terms to existing deterministic demand and capacity models that are derived from first principles, e.g., rules of mechanics. The approach takes into account information from scientific/engineering laws, observational data from laboratory experiments or field investigations, engineering experience, and subjective judgment. The probabilistic models correct the conservatism inherent in deterministic models and explicitly account for the most relevant uncertainties. The probabilistic demand and capacity models are used in a formulation to assess the seismic reliability of bridge components and systems. Predictive estimates and confidence intervals are developed to properly represent and quantify the inherent epistemic and aleatory uncertainties. The demand and capacity models are then extended with probabilistic models for chloride-induced corrosion and time-dependent corrosion rate to assess the reliability of bridges over time. Models are also presented to capture the deterioration of RC bridges due to low-cycle fatigue, ASR, and DEF. To improve the accuracy of the reliability assessment, information about the current aging and deterioration condition of a bridge is incorporated. Nondestructive testing (NDT) is used to evaluate the actual conditions of bridges and to avoid the use of deterioration models that bring additional uncertainties into the reliability assessment. Probabilistic demand and capacity models are also developed to assess the reliability of bridges subject to scour and vehicle impacts. Finally, the bridge reliability estimates are used to assess the reliability of bridge networks using a novel Matrix-based System Reliability (MSR) method.
Although the methodology presented is aimed at developing probabilistic demand and capacity models, and reliability estimates for RC bridges and bridge networks, the approach is general and is currently applied to assess the reliability of other engineering systems including buildings, offshore systems, and wind turbines.