Capacity and level of service analysis of signalized intersections

In this section the capacity and level of service analyses of signalized intersections are described. The description presents the principles on which the analyses are based and does not go into the specifics. This is because the Indian codes' (see IRC Special Publication 41 [#!ircsp41!#] and IRC: 93-1985 [#!irc93!#]) description of the process is somewhat incomplete and in any case the principles of the analysis process is more important than the exact details of the process. The section is divided into two parts, the first part discusses the capacity analysis and the second part the level of service analysis.

Capacity analysis

The capacity of signalized intersections as such is not meaningful. What is meaningful is capacity of an approach or a lane or a lane group² of an intersection. The capacity, $c_i$, of lane, $i$, to a signalized intersection, is defined as

$$c_i=s_i\times \frac{g_i}{C}$$

where, $s_i$ is the saturation flow on lane $i$ in vehicles per hour of green per lane, $g_i$ is the green time for the lane, and $C$ is the cycle length at the intersection.

Thus, the primary factor which needs to be determined is the saturation flow for the lane. Typically, saturation flow on a lane or lane group depends on a number of factors such as (i) the number of lanes in the lane group and width of lanes or alternatively the width of the lane group, (ii) the gradient of the lane, (iii) percentage of turning traffic, (iv) vehicle mix, (v) the number of parking maneuvers, and (vi) the number of bus stoppings. The HCM [#!hcm98!#] provides a detailed and complete overview of how the effect of these factors can be and should be incorporated in the calculation of saturation flows. The manual assumes a set of ideal conditions and assumes an ideal saturation flow for these conditions. This saturation flow is then reduced by multiplying it with a series of correction factors. Generally, each correction factor quantifies the detrimental effects of the non-ideal conditions with respect to each of the factors mentioned above. These correction factors are provided in a series of easy-to-follow tables.

The IRC Special Publication 41's [#!ircsp41!#] suggestion on determining the saturation flow does not incorporate all the features that affect the quantity. For example, though the code acknowledges the effect of vehicle mix on saturation flow it does not give any procedure to quantify this effect. Similar is the case for turning traffic's effect on the saturation flow of a lane or lane group (although the code attempts at quantifying the effect of right-turning traffic, the procedure is poorly explained and never illustrated). However, the code makes clear suggestions on the effect of width and gradient on saturation flow. For example, the code states that the saturation flow from a lane group of width $w$ (for $w \ge 5.5$ m) is $525\times w$ passenger cars per hour of green (pcphg); a table is provided for lesser values of $w$. The code also states that this value of saturation flow is for flat roads and should be reduced by 3 percent for every 1 percent uphill slope and increased by 3 percent for every 1 percent downhill slope. One may refer to [#!ircsp41!#] for more details on these aspects.

Level of service analysis

The level of service of different lanes and lane groups at signalized intersections should be determined through a measure which directly gives the level of discomfort (or comfort) of drivers using these lane or lane groups at the intersection. One such measure is the average delay to vehicles of different lanes and lane groups. There are various equations, each with certain shortcomings, which can give the average delay (see the section on delay analysis provided earlier). Surprizingly, the IRC codes are silent on the matter of level of service at signlaized intersections. However, other codes like the HCM [#!hcm98!#] of USA does provide a relation between delay to vehicles and level of service.