Treatment of Uncertainties in Seismic Risk Analysis of Transportation Systems

Transportation network systems are vital for the economy and the smooth functioning of societies. When subjected to earthquakes, transportation networks experience structural damages. As a result, travelers experience delays or cancellation of their trips. This thesis studies the consequences of earthquake events to transportation network systems and develops a probabilistic framework to quantify them.

Estimation of the structural damage of the network bridges due to earthquake events is discussed first. It is based on the methodology suggested in HAZUS (1999). Rather than computing the expected loss values, a methodology to assess the probability density function of the structural loss for single or multiple bridges is presented. Furthermore, two new uncertainties are included in the formulations. These include the replacement cost and the damage factor.

Operational loss is computed based on the fixed demand assumption. The total daily delays of the commuter traffic of the transportation network are used to measure its post earthquake performance. These delays are then multiplied with an assumed passenger value of time equal to $12/Hour in order to be converted to monetary units. The analysis is repeated for 1, 3, 7, 14, 30 and 180 days after the event has occurred in order to find the change of the daily operational loss in time and estimate the total operational loss until full restoration of the damage is made.

Fifty six earthquake hazard analyses and 336 network analyses of events with moment magnitude higher than 6.75 on the San Andreas and the Hayward faults are performed. The structural and the operational losses are computed for each scenario event and the annual risk curve of the San Francisco Bay Area transportation network is generated based on the rate of occurrence of each event and its corresponding loss. The expected value of the annual loss of the transportation network is found to be $13.3 Million. The equivalent present value of this annual risk is defined based on the effective interest rate.

Operational loss is smaller than the structural loss at the low magnitude groundmotions; however, it governs the loss in the higher ground motions. Moreover, the Hayward fault is found to cause higher operational losses than the San Andreas fault.

In the evaluation of the probability density function of the structural loss, consideration of the uncertainty in the replacement cost increased the risk. This was not the case when the damage factor was considered, which actually led to a slight decrease of the risk.