A new framework for the analysis of infrastructure performance and the planning of inspection and maintenance and rehabilitation (M&R) activities is described. The main facets of this framework are (a) treating facility performance, the key variable in the process, as a latent variable that manifests itself through measured condition indicators; (b) explicitly analysing and including in the decision process the errors and uncertainties in infrastructure condition measurement and performance analysis; and (c) accounting for the interactions between the two decisions, inspection and maintenance strategies, by jointly selecting them within an optimisation algorithm. A model system relating the latent facility performance to explanatory variables and to observed indicators is developed and parametric studies using this framework are presented. The approach recognises the errors inherent in the measurement of the condition indicators; these errors are quantified through the use of measurement error models, which are estimated using a rigorous statistical method. The M&R activity planning model accounts for the presence of uncertainty in the output of the inspection of facility condition and the forecasting of facility performance. Inspection activity decisions are addressed jointly with the M&R activity decisions through a common stochastic optimisation algorithm, which leads to the selection of cost-effective inspection activities. A parametric study investigating the effects of uncertainty in condition measurement and forecasting on life cycle costs is also presented.
Abstract