DEVELOPMENT OF CAPACITY ANALYSIS PROCEDURE FOR URBAN INTERSECTION

Abstract

At-grade intersections are the most common type of road junctions found in India and other developing countries. Since these intersections are required to process all converging movements through a common area, a scientific approach is needed for their design. Capacity of a signalised intersection is based on the concept of saturation flow which is defined as the maximum rate of flow that can pass through a given intersection approach or lane group under prevailing conditions of roadway, traffic and signalisation assuming that the lane had 100 per cent of real time as the effective green. Traffic flow parameters required to estimate ' capacity of a roadway facility have been studied extensively in western countries. In India, there has not been any major effort to develop capacity norms except a few scattered ones to identify some parameters affecting the capacity. It might be due to the chaotic traffic conditions and poor regulatory systems prevailing. In the analysis of traffic flow, the various categories of vehicles in the stream are converted into a uniform measure which is considered to be an equivalency for the vehicle type. The most accepted such equivalency factor as adopted in the west is the common unit of passenger car. Such common unit for each vehicle type is designed necessarily with due consideration of static and dynamic characteristics of vehicles, and thus, all vehicles of heterogeneous traffic stream are converted to a homogeneous equivalent in passenger car unit (PCU). Traffic engineers normally (ML) employ PCU values available as national standards to calculate capacity of a facility. Various agencies in India and abroad have recommended PCU values for different types of vehicles. These values are fixed, and therefore, their applicability to all traffic and control conditions is doubtful. A number of studies have been carried out all over thi world to determine the most realistic PCU value for a given type of vehicle and each of these studies has resulted in a different set of PCU values. This can be considered as the sufficient basis to argue that PCU for a vehicle is not a constant value but varies from one location and operating condition to another. Its value depends on the method of derivation as well. The concept of dynamic PCU has been examined in the present research. Data collected by video filming of a series of signalised intersections handling representative mix of traffic in Delhi have been analysed. The vehicles were divided into 5 different categories namely; car, bus, 3-wheeler, 2-wheeler and bicycle. However, bicycle traffic was present at two Intersections only, and hence, detailed study on its PCU could not be conducted. Variation in PCU for a given type of vehicle with parameters like composition of traffic stream, saturation flow rate, approach width, and ))))N turning radius have been modelled independently as well as jointly. Graphical exposition of the interdependencies clearly shows the variation in PCU values with the composition of traffic stream. It is also indicated that PCU of a vehicle decreases with increase in its own proporton in the traffic stream. This is in conformity with the findings of Branston (1979), Miller (1969), Smeed and Hillier (1965), Webster and Scraggs (1964) who have also reported the same trend for PCU values. It is observed that PCU for a given type of (tv) /* vehicle is controlled by the degree of homogeneity or heterogeneity of the traffic stream. Further, for a given proportional composition of traffic stream, PCU of a vehicle decreases with increase in the saturation flow rate. This is attributed to the fact that at higher value of saturation flow all vehicles are forced to move at uniform speed resulting in lesser speed differential, and thereby having lower PCU values. The effect of approach width on PCU values has also been studied in this research, and it is revealed that for all other conditions of traffic remaining unchanged, PCU of a vehicle increases with the increase in approach width. This is again attributed to the freedom of movements experienced by individual vehicles at wider approaches. PCU of a turning vehicle was found to be dependent on turning radius alongwith other associated factors. The curves drawn to show the variation in PCU for a given type of vehicle with influencing parameters Indicate that it depends on all the traffic and geometric parameters of the intersection. While it is easy to relate the influence of a single factor to the variation of PCU, the multiple effect can be traced only through a mathematical relationship. A mathematical model has been developed to relate the PCU of a vehicle type with the composition of traffic stream. These equations were established for the data collected at individual approaches and the coefficients were found to be dependent on the total number of vehicle categories in the traffic stream and the approach width. Accordingly, a single set of coefficients cannot be suggested to estimate PCU of a given type of vehicle. To neutralise the effect of the approach width on PCU, » data for various intersections were normalised on the basis of per (v) lane width (3.5 metres). The pooled data from all the approaches was analysed by varying the number of vehicle categories in the traffic stream. PCU equations were established for each type of vehicle under different combinations of vehicle categories. Nomograms have been developed to estimate the coefficients of PCU equations for a given condition of traffic stream. In these nomograms the presence or absence of a particular category of vehicle in the traffic stream is specified through the dummy variable of 0 and 1 type. Having established the dynamic nature of PCU, saturation flow was estimated in terms of PCU per hour of green time (pcuphg). Straight through and right turn movements were analysed separately to estimate the effect of right turning traffic on the total saturation flow for the approach. Saturation flow values for approaches were calculated using the static PCU values suggested in IRC-106, 1990 and the dynamic PCU values developed in this research. It is observed that the saturation flow is overestimated upto 130 per cent when static PCU values were employed. A linear relationship has been developed between saturation flow and the approach width. Right turn movements were also analysed In the same manner and the saturation flow was found to be related with the approach width allocated to the right turners and the turning radius. A mathematical relationship between per cent reduction in saturation flow due to presence of right turning traffic and the turning radius has also been developed. The equivalency factors for right turners (defined as the ratio of straight through saturation flow and the right turn saturation flow) were found to be related to the turning radius. Another important aspect in the analysis of any roadway facility including intersection is the Level of Servive (LOS) which is related solely to measures and characteristics affecting the quality of service provided to the driver. Currently, level of service at signalised intersection is defined in terms of average stopped time delay caused to individual vehicles. Delay is a complex parameter and depends upon a large number of variables, including the quality of progression, cycle length, the green ratio, and the v/c ratio for the approach under consideration. Further, delay is a more relative measure than'v/c ratio; what is unacceptable in a major metropolis may be quite acceptable in a small urban area. Exact relationship is still not very well known between delay and degree of frustration of a driver. It is not only more travel time or time lost, but the range and frequency of speed changes that annoy the driver. Coupled with many other problems associated with the interpretation and measurement of delay, parameters like degree of saturation (DOS), saturated green ratio, and the per cent vehicles required to stop in the approach have been used in this study to define level of service at signalised intersections under mixed traffic conditions. These parameters are easy to measure in the field and directly related to the users' satisfaction. Stopped time delay was also measured in the field and it was observed that the delay limits given in HCM (1985) for different levels of service cannot be applied to mixed traffic conditions prevailing in India because the measured average delay in most of the cases was more than 60 sec/veh. This is the limit for LOS F as suggested in HCM (1985). The extrapolated value of stopped delay for the degree of saturation equal to 1.0 was (VLL) obtained as 140 sec/veh from a plot between these two parameters. Degree of saturation value greater than 1.0 indicates a situation when demand volume is higher than capacity, and hence corresponds to LOS F. Taking delay greater than 140 sec/veh for LOS F, and increasing the limits for other levels of service as given in HCM (1985) proportionately, breakpoints in the complete range of DOS were determined for six levels of service. These limits of DOS for levels of service were then used to determine ranges in other parameters for levels of service such as saturated green ratio and the per cent vehicles stopped in the approach. The methodology developed in this research has been formulated into stepwise procedure to design and analyse a signalised intersection for its use by practicing engineers. The example problems of design and performance analysis of an intersection have also been included in this work for illustration. The results provided in the form of mathematical relationships and nomograms can directly be used by the designer. Further study is recommended to extend the concept of dynamic PCU to other categories of vehicles also and to develop the nomograms for direct estimation of PCU for all such vehicle types. The methodology can be made more general by taking data from different cities, and therefore, making graphical and mathematical relationships geographically transferable. (