Probabilistic lifetime-oriented multiobjective optimization of bridge maintenance: Combination of maintenance types

In this paper, the performance of deteriorating bridges is analyzed considering the combination of essential and preventive maintenance actions. Multiobjective optimization under uncertainty is used to find the best combinations of condition, safety, and cost; satisfying all performance constraints of deteriorating bridges during a specified time horizon under multiple maintenance types. The thresholds at which essential maintenance actions are applied and the times of application of preventive maintenance actions are considered as design variables. The evolution in time of nondeterministic performance is modeled as probabilistic condition, safety, and cost profiles, and genetic algorithms are combined with Latin hypercube sampling to optimize the maintenance strategies in a full probabilistic context. Owing to the scarce information on the structural performance effects, due to the interaction among different types of maintenance actions, special attention is paid to the assumptions with respect to the combination of effects of different maintenance types. Examples of application, based on data collected on reinforced concrete bridge crossheads in the United Kingdom, are presented. The benefits of the combination of preventive and essential maintenance actions over single maintenance type strategies are clearly indicated; namely, the reduction in cost for similar condition and safety indices. The use of alternative probabilistic indicators in bridge maintenance, such as the lower percentiles of performance, is also analyzed.