Joint Analysis of Queuing Delays Associated With Secondary Incidents

Abstract

Incidents cause substantial travel delays on urban freeways. Travel delays due to incidents are typically estimated for each incident independently. However, when multiple incidents occur on the same stretch of a roadway, treating them independently and ignoring their interactions leads to inaccurate results, which ultimately may lead to inefficient incident management strategies. This study examines delays due to primary–secondary incident pairs that occur on the same stretch of a freeway within a short time gap. Depending on their correlation in time and space, they can have varying degrees of impacts on the traffic flow. This article assesses the total delays induced by primary–secondary incident pairs by jointly modeling their occurrences. First, incident data combined with roadway inventory data from Hampton Roads, Virginia, are analyzed to understand the attributes of primary–secondary incident pairs, e.g., durations, lane blockages, and time gaps (between start times of the primary and its secondary incidents). Then using microscopic simulation as a primary evaluation tool, the effects of three critical parameters are investigated: time gap, distance between primary and secondary incidents, and traffic demand level. The historical incident analysis shows that, on average, primary–secondary incident pairs have substantially longer durations than single incidents. The joint analysis of queuing delays induced by primary incident and its secondary incident suggests that time gap and distance between a primary incident and its secondary incident are significantly associated with the total delays after controlling for other factors (e.g., the two associated incidents’ durations and lane blockages). Furthermore, based on the scenarios tested (i.e., two incidents both blocking the right lane in a three-lane basic freeway section), it is found that the sum of delays induced by a primary incident and its secondary incident separately will over- or underestimate the actual delays. For those secondary incidents that end after their associated primary incidents, total delays increase as time gap increases; increasing the distance between primary and secondary incidents is associated with reduction in delays. Other results and the implications of the findings for traffic operations are presented in the article.