A review of uncertainty effect in structural condition assessment

Structural condition assessment and/or monitoring exercises involve many uncertainties, which come from modelling procedures, methodologies employed, environmental variation and measurements. Different uncertainties have different patterns but they all affect the assessment results. Ignoring the effects of uncertainties might make the assessment unsuccessful in terms of structural damage undetectable and/or wrongly identifying intact elements as damaged. This paper reviews various uncertainty sources, quantification, and their influence on the structural condition assessment. Uncertainty types include inaccuracy in the Finite Element (FE) model discretization, uncertainties in geometry and boundary conditions, inaccuracy in material properties, environmental variability, errors associated with measurement, and improper methodologies applied in assessment. Analytical methods and numerical methods for uncertainty analysis, such as variance propagation, Monte Carlo simulation, Bootstrap, and perturbation technique will also be discussed. As examples, a probability based damage detection with consideration of FE modelling error and measurement error will be presented and discussed. Environmental variation effects on dynamic properties of a RC slab, which has been monitored for more than two years, will also be introduced.