The SIDRA INTERSECTION roundabout capacity model treats a roundabout as a closed system with interactions among roundabout entries. This requires an iterative capacity estimation process that takes into account many parameters that are calculated as a function of the capacities estimated in the previous iteration, and in turn, affect the capacities in the current iteration. This contrasts with the traditional method of roundabout modelling that treats the roundabout as a series of independent T-intersections with no interactions among flow conditions of roundabout approaches.
While the traditional method may be adequate for low to medium flow conditions, the treatment of a roundabout as an interactive system improves the prediction of capacities under heavy flow conditions, especially at multi-lane roundabouts with unbalanced demand flow patterns.
The treatment of a roundabout as a closed system involves the following model elements:
(i) Capacity constraint: Circulating and exiting flow rates at the subject approach are affected by entry flow rates limited to the capacity flow rate when the demand flow rate exceeds the capacity (i.e. when some entry lanes are oversaturated) at upstream approaches.
(ii) Lane balance of circulating flow rates: This is affected by entry lane flow rates of contributing upstream entry flows (including the effect of approach lane underutilisation), and reflected in the bunched headway distribution of circulating road headways. In the SIDRA INTERSECTION model, entry lane capacities also depend on approach lane flow rates. The traditional models that use the total circulating flow rate only cannot take the effect of the balance of approach and circulating lane flow rates on capacity.
(iii) Bunched headway distribution for the circulating flow: Bunching in the circulating flow reflects the bunching of unqueued vehicles in the arrival flow as well as the bunching that represents queued vehicles entering from upstream entry lanes. This allows the effect of the number of circulating lanes and the lane balance of the circulating flow on the entry capacities. This headway distribution also allows the use of extra bunching to model the effect of upstream signalised intersections or pedestrian crossings on roundabout capacity. The simple exponential (random) distribution of headways cannot take any of these parameters into account.
(iv) Unbalanced flow conditions: Under these conditions, circulating flow in front of an approach originates mostly from one approach and is highly queued on the approach before entering the roundabout with uniform queue discharge headways, thus affecting the downstream entry capacities adversely. One of the most important enhancements to the roundabout capacity estimation method introduced in SIDRA INTERSECTION is modelling of these conditions by allowance for the effects of origin-destination pattern of entry flows, proportion queued and any unequal lane use at entry lanes of upstream approaches.
The above capacity model elements affect each other, thus requiring an iterative capacity estimation process. The lack of these model elements indicate the shortcomings of the models that treat a roundabout as a series of independent T-junctions and ignore the interactions among roundabout entries. These models include the Highway Capacity Manual 2010 roundabout capacity model. This also applies to the TRL empirical model used in the ARCADY and RODEL software with the only exception of the use of capacity constraint. It means that this model treats the roundabout as a series of T intersections when all roundabout approaches have demand flow rates below capacity (i.e. when all entry lanes are undersaturated).