Time-dependent risk associated with deterioration of highway bridge networks

• A methodology for time-dependent expected losses of bridge networks is proposed.
• The methodology is used for assessing the risk associated with bridge deterioration.
• A five-state Markov chain is used for modeling time-dependent performance of bridges.
• Expected losses are identified based on scenarios regarding service states of bridges.
• The methodology is time efficient for large networks.

Risk-based performance indicators offer valuable information on the performance of spatially distributed systems, such as highway bridge networks, whose functionality is vital for economic and social purposes. In this paper, the time-dependent expected losses of deteriorated highway bridge networks are investigated. A five-state Markov model is used to predict the time-dependent performance of bridges. The direct consequences are assessed on the basis of scenarios characterized by individual bridge failures and maintenance shutdowns. The indirect consequences are quantified on the basis of scenarios characterized by single and multiple bridges out-of-service, by solving the traffic assignment problem. The variation of direct, indirect and total expected losses in time is computed. The proposed methodology is suitable for the assessment of losses due to the service loads and the effects of deterioration. The proposed approach is illustrated on a highway bridge network. The results indicate that the maximum total expected indirect loss is much higher than the maximum total expected direct loss. The expected loss profiles show a pattern that first increases and then decreases due to the time-variation of the Markov chain state probabilities.