Incorporating Non-destructive Field Measurements in Stochastic Finite Element Analysis of Reinforced Concrete Structures

Abstract

Unlike deterministic analyses where the sensitivity to material inputs cannot be captured, stochastic simulations with the addition of established material distributions and damage prediction models can provide a higher level of confidence in structural behavior. In this study, the use of stochastic analysis as a way to better simulate the uncertainty effects often encountered in structural evaluation was investigated. Stochastic analysis capability was implemented into VecTor2, a nonlinear finite element analysis program for reinforced concrete structures, and simulations were performed on a series of reinforced concrete beams and shear walls. The simulations were compared against deterministic and experimental results and found to be adequate substitutes whilst providing a level of confidence consistent with each specimen. To achieve narrower confidence intervals, early stage field measurements were incorporated in the stochastic analysis. Using early stage measurements such as deflections, crack widths, and reinforcement strains, the stochastic output was recalibrated to produce more accurate results.