A random field based technique for the efficiency enhancement of bridge network life-cycle analysis under uncertainty

Studies associated with distributed civil infrastructure systems are usually very demanding from a computational point of view, especially when they involve life-cycle analysis, uncertainty, and optimization. For this reason, computational tools that enhance the efficiency of the analysis and make it feasible for complex practical applications are of utmost importance. In this paper, a computational technique for the efficiency enhancement of bridge network life-cycle analysis under uncertainty is presented and its impact in terms of CPU time reduction is investigated.The proposed technique consists in the joint use of random field theory and probabilistic reliability models for the simulation of the individual bridge service states over the life-cycle of the infrastructure. This random field based approach is extremely efficient and takes simultaneously into account the deterioration in time of the bridge reliability and the correlation in space of the service states of bridges belonging to the same transportation network. Compared to other techniques previously used to perform the same task, the proposed methodology is theoretically more solid and improves the computational efficiency by more than two orders of magnitude.A numerical example is provided to validate the proposed technique. Moreover, a second example involving the life-cycle performance analysis of a complex bridge network in Santa Barbara, CA, is presented.

► Random field based technique for life-cycle analysis of bridge networks is proposed.
► Efficiency enhanced approximately 250 times compared to previous approach.
► Correlation among bridge service states, uncertain deterioration, and maintenance.
► Superior scalability compared to previous approach.
► Application to a 38-bridge network located in California.