Reliability analysis of corrosion initiation in reinforced concrete structures subjected to chlorides in presence of epistemic uncertainties

Abstract

This paper presents a methodology for assessing the corrosion initiation in reinforced concrete (RC) structures subjected to chlorides, with a focus on the influence of epistemic uncertainties inherent to input parameters. Epistemic uncertain parameters were modeled using fuzzy probability theory. The chloride penetration was modeled using a finite element solution of the coupled heat transfer, moisture and chloride transport in concrete. The environmental conditions were modeled using a stochastic process. The reliability analysis was undertaken using a hybrid fuzzy Monte Carlo simulation. As an illustration, the durability of concrete basins of a newly constructed waste water treatment plant, with respect to reinforcement corrosion, was examined. The time to corrosion initiation was found to approximately follow a generalized extreme value distribution with fuzzy parameters. The stochastic modeling of exposure conditions accelerated the corrosion initiation. The time to corrosion initiation is highly sensitive to the epistemic uncertainty in the threshold chloride level in concrete. Reducing this uncertainty will result in an important improvement in accuracy. The proposed method is suitable for durability assessment of recent RC structures with little or no available performance data.