Models for estimating saturation flow and maximum demand at closely spaced intersections

Abstract

This thesis describes models for saturation flow and maximum demand at closely spaced intersections. The effects of queue interaction between these two intersections are taken into account in both models. The saturation flow model is based on the Prosser-Dunne model. The presence of queues in the inter-signal link causes a reduction in saturation flow and capacity. The analytical model on which the methodology is based assumes that upstream movements discharge at their normal saturation flow rate or arrival flow rate until the downstream queue extends back to the upstream intersection and blocking occurs. The model calculates the capacities of movements at the upstream intersection as a reduced effective green period. The model can be used to estimate capacities at paired intersections with multiple upstream and downstream green periods. The results from the model are compared with TRAF-NETSIM simulation results. The results of this comparison show that the model predicts throughput better when movements at the upstream intersection (for which throughput are being calculated) are oversaturated. This thesis recommends that the capacity of movements be calculated using the reduced effective green period rather than the reduced saturation flow. The second model developed as a part of this research predicts the maximum demand at the downstream intersection. The through movement at the upstream intersection is assumed to be oversaturated and cross street movements are not considered. The analysis shows that either the upstream capacity, downstream capacity or storage capacity becomes critical and influences the maximum demand depending on the different combinations of upstream and downstream green and storage spacing considered. The demand from the models is used as input to the 1994 Highway Capacity Manual (HCM) delay equation and the delay compared with that simulated by TRAF-NETSIM for various cases. The comparison shows that the models developed predict values that compare favorably with results from TRAF NETSIM. It is recommended that the models be used to compute the upper bound for the HCM delay equation for the cases analyzed.