Random fields for the modeling of deteriorated structures

Abstract

The assessment of existing structures is generally a complex task, due to the uncertainties related mainly to the deterioration processes and the consequent mechanical behaviour. When dealing with reinforced concrete structures subject to corrosion of the reinforcement, the prediction of the evolution of the safety level is very important for the planning of maintenance and repair interventions. Clearly, the effectiveness of these interventions depends on the analytical models adopted for the prediction of the time dependent corrosion of reinforcement and the probabilistic models of the model parameters. Models for the prediction of the initiation of corrosion due to chloride ingress are available in literature, as well as guidance to describe the randomness of the material and environmental parameters involved in the diffusion process. Hence the structural effects of corrosion (reduction of the cross section of the rebars and mechanical properties of reinforcement, cracking of concrete with subsequent reduction of the compressive strength and loss of bond strength, etc.) have to be described in probabilistic terms. The paper is focused on the probabilistic analysis of the effects of the corrosion process, when the spatial variability of the relevant parameters is considered. The attention is focused on the random field models and their application for the assessment of the existing structures. For the purpose, a numerical discretization procedure based on Karhunen-Loeve series expansion and the finite element method is applied. development of more realistic probabilistic models (Lay & Schiesl 2003, fib 2006). Extensive research has been conducted in order to estimate the effect of the randomness of the main parameters (i.e. diffusion coefficient, surface chloride concentration and depth of the concrete cover) on corrosion initiation (Tikalsky et al. 2005, Val & Trapper 2008, Ann et al. 2009). Other papers are more focused on the analysis of spatial variability of corrosion (Engelund & Sorensen 1998, Li et al. 2004, Frier & Sorensen 2007). As a result of the available knowledge, the reliability of existing concrete structures and service life optimization have been deeply investigated (Enright & Frangopol 1998, Thoft-Christensen 1998, Val et al. 2000, Kirkpatrick et al. 2002, Duprat 2007, Stewart & Mullard 2007, Stewart & Al-Harthy 2008, Stewart & Suo 2009, Ferreira 2010). In the present paper, a probabilistic approach is adopted for the assessment of the structural effects of chloride-induced corrosion. In a first step, the ingress of chlorides into the concrete structure is modeled by a 2D numerical model, which takes into account the point to point variability of material properties and environmental conditions. In a second step, the effect of corrosion in terms of reduction of the mechanical properties of the reinforcing steel and cracking and spalling of the concrete cover are investigated.