Author: Sajan Shrestha

Publisher:

Published: 2015

Total Pages: 61

ISBN-13:

To facilitate construction and maintenance activities on freeways, a common practice is to close a lane of traffic. The lane closure affords the work crew space to work, as well as providing access to the work site and even providing a buffer between the work activities and the traffic. However, closing a lane can have an impact on the traffic flow and cause delays to the traveling public and movement of goods. If the traffic flow is greater than that which can be serviced by the lanes which remain open, then traffic backs up, or queues. A good estimate of the number of vehicles that will queue is needed to make informed decisions about when lanes can be closed, and how many lanes can be closed, while minimizing the impact on traffic. The Highway Capacity Manual provides a recommended methodology to analyze freeway operations. The methodology uses a 15 minute time interval for analysis but recommends a much small interval such as 15 to 60 seconds to analyze a queued condition. In contrast, sketch planning tools used by state departments of transportation to analyze queuing typically use an hour interval with hourly traffic volumes. Both the HCM methodology and the sketch planning tools compare the number vehicles arriving to the maximum number of vehicles that can be serviced over a period of time. Therefore, the queue estimate relies on having a good estimate of the capacity of the work zone with the lane closure as well as good traffic volume data. The purpose of this thesis was to examine the sensitivity of queue estimates to the size of the time interval used to aggregate traffic volumes. When the traffic volumes are aggregated into small time intervals, the variability of the flow is better captured than when large time intervals are used. Therefore, it was expected that queue estimates would improve when smaller time intervals were used. To examine this relationship, field studies and a sensitivity analysis were conducted. The field studies were conducted at two Ohio work zones. At a work zone on I-71 in Columbus, 2 of 3 traffic lanes were closed. At a work zone on I-75 in Dayton, 1 of 3 traffic lanes was closed. Traffic volumes and queueing data were collected. The traffic volume data was aggregated using 5, 10, 15, 20 and 30 minute intervals and used to estimate the number of vehicles in the queue. The sensitivity of this queue estimate to the time interval used for aggregation was examined. The results of the sensitivity analysis were as expected. Regardless of the time interval used for aggregation, the queue is considered to have a constant growth rate between estimates. This means that when larger time intervals are used for aggregation, more details about the variability of the traffic flow are lost and the queue estimate between within each time interval includes an aggregation error. Although the 5 minute time intervals would provide the best detail about the formation of the queue, the sketch planning tools and the traffic data themselves are usually based on hourly volumes. The hourly volumes are often accompanied by the peak hour factor, which describes the variability of the traffic flow as the ratio of the hourly flow rate to the 15 minute peak flow rate. The impact of the peak hour factor on the queue was examined. If the peak 15 minute interval has a volume greater than capacity, and the three non-peak 15 minute intervals are assumed to experience the same flow which is less than capacity, then the queue estimate is a function of the peak hour factor. As the peak hour factor increases, or the variability of the traffic flow decreases, the estimate of the number of vehicles in the queue decreases. This relationship is potentially useful for adjusting queue estimates for volumes that fit this pattern.