Stochastic Capacity Adjustment Considering Secondary Incidents

The onset of queuing upstream of an Incident site is marked by a reduction in discharge flow. Additional delays are expected when a speed reduction caused by the queue and rubbernecking provokes additional secondary incidents. The growing number of secondary incidents on freeways compels the development of models to estimate the delay under the extended time required to clear congestion. We need a new indicator that represents the capacity reduction due to both primary and secondary incidents. In this paper, secondary incident delay is described in a geometric surface area with explicit formulations considering gap/overlap between occurrence and clearance of the primary and secondary incidents. The estimation of capacity reduction is consolidated when a secondary incident occurs. A stochastic extension of a secondary incident delay model provides real-time prediction of delay by considering the dynamic characteristics of the network. The first response unit and a secondary response unit arrival times are considered to obtain location-specific incident duration, which is one of the input parameters for the estimation of capacity reduction. Empirical evidence shows the significant impact of secondary incidents on capacity reduction on freeway segments. Results indicate higher or lower capacity reduction depending on different types (i.e., time series) of the secondary incident occurrences. Without consideration of the time series of incident occurrences, the capacity adjustments of Highway Capacity Manual are either underestimated or overestimated. When real-time information is updated, the proposed model can be easily re-run to estimate the total delay in the system.