Reliability-based condition assessment and life prediction of concrete structures

Concrete structures may be exposed to aggressive environmental effects that cause their load capacity to decrease over an extended period of service. If a safety assessment of such structures were to be performed to evaluate the possibility of continued service, a major concern would be to ensure that in their current condition they are able to withstand future extreme load events during the intended service life with a level of reliability sufficient for public safety. Methodologies to perform this assessment currently do not exist. A methodology is proposed to facilitate quantitative assessments of current and future structural reliability and performance of concrete structures. This methodology takes into account the nature of past and future loads, randomness in strength and in degradation resulting from environmental factors. Changes in engineering properties of steel and concrete over the service life are identified. An adaptive Monte Carlo simulation procedure is proposed to evaluate time-dependent system reliability. The sensitivity of the reliability to various parameters describing load occurrence and strength degradation is illustrated. The reliability is sensitive to the choice of initial strength and strength degradation models. Less sensitivity is found to correlation in component strengths within a system. It is important to identify the critical components before performing system reliability analysis so that the size of the analysis can be reduced.The evaluation of time-dependent reliability and deterioration of concrete structures provides a basis for selecting appropriate periods for continued service and/or for determining optimum intervals and extent of inspection and maintenance. Inspection/maintenance strategies are identified that minimize the expected future cost keeping the failure probability at or below an established target failure probability during the lifetime of a structure. Needs for additional data to allow the reliability models to be fully implemented as a decision tool are identified.