Polya Urn Model for Assessment of Prestress Loss in Prestressed Concrete (PSC) Girders in a Bridge System using Limited Monitoring Data

Abstract

A procedure based on Polya urn model is proposed for the assessment of prestress loss of prestressed concrete (PSC) girders in a bridge system using data from monitoring of strains in a limited number of girders. The procedure integrates: (i) the Polya urn model (which is a stochastic pure birth process model), (ii) prestress loss estimation method of AASHTO LRFD 2012 specifications and (iii) prestress loss assessment at different times using monitored strain data. It is assumed that all the girders in the bridge system considered are exposed to nominally similar environment, and the number of girders in the entire bridge system is much greater than the number of girders inspected/monitored. While Markov chain models have been proposed in the literature for condition assessment of bridge system based on limited inspection data, these models can be used only when inspection data on adjacent girders are available. However, it may not be always feasible to obtain information from the adjacent girders. The proposed Poly urn model based procedure can be used when the inspection/monitoring data is from randomly selected girders exposed to nominally similar environment. The usefulness of the proposed procedure is illustrated by considering a bridge consisting of 100 PSC bridge girders. Three different scenarios, namely, (I) 5, (II) 15 and (III) 20 PSC girders that are inspected/monitored for strain, are considered. The different probability values obtained for the different scenarios suggest that the proposed model takes into consideration the value of information available. The efficacy of the proposed procedure is demonstrated by considering a practical case study of prestress loss in nominally similar PSC girders exposed to nominally similar environment.