The United States has a national inventory of almost 600,000 highway bridges, many of which have deteriorated substantially and will require large expenditures to repair. Given that funds and resources are scarce, system reliability and optimization techniques have the potential to more efficiently maintain an adequate level of safety in these highway bridges while minimizing the total expected life-cycle cost. This study develops a reliability-based methodology for optimizing the inspection and repair of highway bridges.; An optimization method for inspection is proposed which relies on an event tree and the probability of making a repair to determine the optimal timing and number of inspections over the life of a structure. The method is developed and illustrated using a hypothetical structure, deterioration model, cost data, and inspection techniques. The method is later applied to an actual bridge deck using actual cost data, a chloride penetration model, and the half-cell potential method.; First-Order Reliability Method theory for component and system reliability is covered. A computer program RELSYS (RELiability of SYStems) is developed which computes the reliability index and probability of failure for any structural system which can be modeled as a series-parallel combination of components. The program is completely described through numerical examples and a comprehensive user's manual.; Using a system reliability approach, a hypothetical deterioration model is applied to statically determinate and indeterminate truss structures. Repairs are made whenever the system reliability index of the truss falls below a prescribed minimum level. A component reliability threshold is established to determine which components are repaired. The optimum component threshold reliability is based on minimizing total life-cycle cost from which a repair strategy is developed.; This system reliability approach is applied to the existing Colorado State Highway Bridge E-17-AH. All relevant bridge failure modes are identified and their component reliabilities are computed. The bridge is modeled as a series-parallel combination of these failure modes. Both the load rating and the system reliability of the index of the bridge are computed and compared. Using realistic live-load and deterioration models, the system reliability index of the bridge is evaluated over time. Considering several realistic repair bridge is repaired every time its system reliability falls below a options, the prescribed minimum value. The optimum lifetime repair strategy is based on all feasible combinations of repair options and cost data developed in consult with the Colorado Department of Transportation. Since the system reliability model is entirely strength-based, serviceability issues are addressed through serviceability flags.; A method for updating the bridge repair strategy over time is introduced based on the results of inspections. Updating the reliability of a bridge using data from the PONTIS Bridge Management System is investigated. The repair strategy is updated based on the hypothetical results of specifically selected non-destructive evaluation tests.
展开▼