Abstract:

In response to the issue of subjective uncertainty in the formulation and implementation of maintenance and strengthening strategies for in-service bridges, this paper proposes a bridge maintenance decision framework that considers the time interval it takes. This framework introduces interval numbers to quantify subjective uncertainty that cannot be described by probabilities. It achieves a direct mapping of the most unfavorable reliability index under probability-interval mixed uncertainty based on surrogate models. Thus, the multi-objective optimization algorithm NSGA-II can be used to make the framework efficient. Taking a typical assembled simply supported T-beam bridge as an example, a probability model of vehicle load effects is established based on WIM system measured data. Subsequently, a time-dependent resistance degradation model is introduced to optimize the maintenance strategy for the T-beam bridge and formulate a maintenance decision library for typical T-beam bridges. The results indicate that strategies with smaller time intervals correspond to smaller Life Cycle Cost (LCC) and lower permissible subjective uncertainty. Strategies with larger time intervals, although resulting in higher LCC, provide more flexibility for construction, decision-making, and other aspects. For simply supported T-beam bridges with spans ranging from 20 to 40 meters, various reinforcement strategies can be employed to meet the service reliability index requirements while minimizing LCC.

Key words: bridge operation and maintenance strategy, time-dependent reliability, Life Cycle Cost, Weigh-in-motion, vehicle load effect