Spatial Degradation in Reliability Assessment of Ageing Concrete Structures

Abstract

Presented paper concerns the difficulty of objective characterization of spatial variability due to advancing degradation as a result of environmental exposure. The main issue is to enhance realism in the prediction of remaining service life of existing concrete infrastructure and thus effectively cover the relatively large sample space of possible future deteriorating states. This is done by introducing a special sampling strategy where artificial realizations of damage scenarios are generated. Here, not only material and geometry of a particular structural member is randomized in a nonlinear 3D finite element context, but advanced evolutionary spatial degradation models are applied to account for the variability of future damage states. These are computed using a non-traditional evolutionary scheme based on cellular automata (CA), which is used here to solve the transport equations in time and space and generate the irregular and heterogeneous structure of concrete. Within the presented example of an ageing bridge, the CA simulation also accounts to complex boundary conditions, e.g. the non-stationary seasonal de-icing salt application, irregular turbulent feed or washout effects. The selected case study of an existing deteriorated bridge serves as an application example with historical evidence and well documented damage profiles. It is further discussed with respect to Monte Carlo based structural reliability assessment, how to infer likelihoods associated to the set of implicit statements on damage, as this concept still offers open questions for research, yet is critical to successful and objective uncertainty quantification.

Aggregate Clustering and Casting Direction Effects in Lattice Discrete Particle Model Simulations

“…Recent developments in sampling schemes for spatial or temporal variability are discussed by [7,8,4], where a feasible sample selection strategy for spatial variability is proposed. With the introduction of random (spatially variable) fields in the MC simulations the question of autocorrelation quickly emerges as the amount of response scattering, e.g.…”

Random and Gradient Based Fields in Discrete Particle Models of Heterogeneous Materials

“…This has prompted research into Monte Carlo (MC) based small-sample simulation methods [3], [4], where, despite the capacity of current computers, and in particular in the context of spatial variability, practical utilization requires the availability of an effective sampling strategy that would dramatically reduce the number of required realizations while maintaining accurate estimates of the response characteristics (low-probability large-consequence events) [5]. Recent attempts addressing the sampling strategy for spatial variability in the MC simulation framework include [6], [7], which is based on the original work of [8], where critical samples of stochastic processes are identified.…”

Implications of spatial variability characterization in discrete particle models

“…This was originally formulated by Podroužek et al [9,19] and presented in the context of seismic risk analysis, where the stochastic process represented ground motion events.…”

“…In order to estimate a multi-decade performance of concrete structures, not only the material heterogeneity have to be taken in to account, but also the spatially variable boundary conditions (orientation and location of the exposed surfaces) and stochastic processes (e.g. weather conditions) involved [19]. Among the typical phenomena of deterioration are salt injury, freezing and thawing damage, carbonation and chemical attack.…”