Identifier

etd-05312006-205307

Degree

Master of Science in Civil Engineering (MSCE)

Department

Civil and Environmental Engineering

Document Type

Thesis

Abstract

This study compares link flows of evacuation traffic using Static Traffic Assignment (STA) with those obtained from Dynamic Traffic Assignment (DTA) and observed traffic counts. The analysis is conducted on a sample of households from South Carolina who evacuated in the face of hurricane Floyd in 1999. The results indicate that traditional static models have the potential to significantly underestimate congestion levels in the network, and DTA models account for nonuniform demand and traffic dynamics in hurricane evacuation much more realistically. From observed traffic counts, traffic volumes vary considerably during an evacuation and the DTA flow estimates generally reproduced these variations. More importantly, DTA was able to capture the delay caused by such peaking since the total vehicle hours of travel estimated by DTA was 3.3 million vehilcle hours while it was only 2.5 million vehicle hours estimated by STA. From the survey data, the estimated total vehicle hours of travel was 2.9 million, suggesting that the DTA procedure may have overestimated the delay. Speed calculated using STA and DTA was 23.3 and 30.9 miles per hour during evacuation from hurricane Floyd. The average speed during evacuation calculated from reported travel time and distance was 25 miles per hour. The clearance time calculated using STA and the traditional method of a response curve and the time taken to pass through critical links was 27 hours. On the other hand, the observed time between the issuing of the evacuation order and the last vehicle from the survey clearing the danger area was 56 hours. Using the loading behavior (i.e. time-dependent evacuation demand) shown in the survey data, the DTA process estimated a clearance time of 58 hours.

Date

2006

Document Availability at the Time of Submission

Release the entire work immediately for access worldwide.

Committee Chair

Chester G. Wilmot

DOI

10.31390/gradschool_theses.3881

Share

COinS