Contexto de aplicación normativa

Once the initial BRT corridors are selected, the demand forecasts for these corridors can be used to determine optimum values for factors such as vehicle capacity, vehicle load factors, service frequency, and dwell times. These attributes in conjunction with the desired preferences for service types (trunk-feeder, direct, local, limited-stop, etc.) and the configuration of stopping bays will allow system developers to model different options for meeting the expected passenger capacities.

Vehicle passenger capacity, load factors, and required service frequency are all mutually dependent. The maximum passenger capacity for a given vehicle is in part dependent on assumptions about culturally acceptable levels of customer comfort at peak times. A trade-off exists between the number of seats provided versus the amount of standing space provided. In some cases, a seated passenger consumes as much as twice the space as that required by a standing passenger. However, for long journey times passengers may have a strong preference for seating. The amount of personal space each passenger requires can vary between different cultures. Knowledge of local preferences in conjunction with stated preference surveys can help evaluate the best spatial arrangement.

Some cities with extreme differences in peak and non-peak demand have considered the application of different sized vehicles for the two periods. In this scenario, high-capacity vehicles are operated only during crush peak periods while lower-capacity vehicles are utilised at other times. While this use of different vehicle types can help better match demand and supply, the additional costs and complexity of operating multiple vehicle types usually exceeds the benefits. These additional costs include:

? Higher vehicle costs due to loss of economies of scale in purchasing a single vehicle type

? Difficulty in providing station entry bays for different doorway configurations ? Greater complexity and managerial requirements for dispatching multiple

vehicle types

A typical system will already have at least two vehicle types in operation (i.e., larger vehicles for trunk services and smaller vehicles for feeder services). Adding another layer of complexity in terms of vehicle types is usually not recommended. However, in cases of extreme demand variances between peak and non-peak periods, multiple vehicle types may be an option to consider.

3.4.4.2 Load factors

The vehicle load factor refers actual capacity usage as a percentage of the maximum passenger capacity. For example, if a vehicle has a maximum capacity of 160 passengers and an average capacity of 128 passengers, then the load factor is 80 percent (128 divided by 160). Generally, it is not advisable to plan to operate at a load factor of 100 percent. At a 100 percent load factor there is no room for system delays or small inefficiencies, both of which are likely outcomes of over- crowded conditions. The desired load factor may vary between peak and non-peak periods. In the Bogotá TransMilenio system, typical load factors are 80 percent for peak periods and 70 percent for non-peak periods.

It is also worth noting that it is possible to operate at a load factor exceeding 100 percent. Such a level implies that passengers are more closely packed than the maximum recommended levels. While such extreme capacities can be expected in some unusual circumstances (e.g, immediately after special events such as sporting events or concerts), it is not desirable to regularly overcrowd vehicles.

3.4.4.3 Service frequency

The service frequency refers to the wait time between arriving vehicles. The wait time is also known as the “headway” between vehicles. In general, it is desirable to provide frequent services in order to reduce customer wait times. Customers often perceive waiting times to be much longer than the actual duration. Thus, to provide a car-competitive public transit service, minimising customer waiting is fundamental. The targeted wait times are closely related to the expected load factors. Longer wait periods will tend to increase the load factor as more passengers will arrive at the station.

Service frequency varies between different cities with BRT, but in general, peak frequencies of one minute to three minutes are quite common. Non-peak frequencies are likely to be longer but usually in a range of four minutes to eight minutes. Service during weekends may also tend to follow non-peak frequencies. However, weekend services may also require peak and non-peak schedules, depending upon local circumstances. For example, weekend markets and sporting events may necessitate higher frequency services.

If the wait is too long, a passenger backlog can occur in which insufficient space is available in the arriving vehicle. As load factors approach 100 percent, significant customer dissatisfaction can be expected. Passengers will be quite frustrated if they are not able to board the vehicle. Such backlogs may imply that passengers will have to wait for many vehicles to pass before there is sufficient boarding space. Some passengers may force their way into the vehicle by pushing against the passengers standing near the doorway. This occurrence leads to both discomfort and the flaring of tempers . Further, the amount of time the vehicle sits at the station will likely increase in this scenario. The confusion at the door-to-station interface will likely prevent the closing of the doors in a timely manner. The catching of bags and even limbs within the closing door will not only slow the overall service but again will lead to significant customer dissatisfaction.

3.4.4.4 Dwell time

Another factor impacting feasible operating conditions is the vehicle “dwell time”. The dwell time is the amount of time vehicles are stopped at a station to allow passenger boarding and alighting. The amount of time required depends upon many variables including:

? Passenger flow volumes ? Number of vehicle doorways ? Width of vehicle doorways

? Entry characteristics (stepped or at-level entry)

? Open space near doorways (on both vehicle and station sides)

BRT systems operate with dwell times as low as 20 seconds. Conventional bus services can require over 60 seconds for boarding and alighting. In general, dwell times may be somewhat higher during peak periods than non-peak periods. The increase during peak periods is due to the additional time needed to board and alight the higher customer volumes.

In addition to dwell time, another key performance measure is the “saturation level” at a given stop. The saturation level measures the relative congestion of vehicles at a stop. The parameter is calculated as follows:

Equation 2

Saturation level at a stop = Dwell time (minutes) x Frequency (buses per hr)

For example, if the dwell time is 20 seconds and there are 60 buses per hour, then the saturation level will be:

Saturation level at a stop = (20 seconds / bus) x (60 buses / hour) / (3600 seconds / hour) = 0.33

As the saturation level increases towards a value of 1.0, then the likelihood of bus queuing increases.

3.4.4.5 Stopping bay configurations

Passenger capacities along a corridor can be increased by providing multiple stopping bays at station. A stopping bay is the designated area where a vehicle will stop and align to the platform. In cities such as Curitiba, Kunming, and Taipei, only one stopping bay is provided per station. However, in other systems, allowing multiple vehicles to stop at the same time has proven to dramatically increase system capacity. Cities such as Bogotá and Porto Alegre employ multiple stopping bays within their BRT systems. Each stopping bay represents a different set of services or routes (e.g., local services versus limited-stop services or routes with a different final destination). In Bogotá, there are as many as five different stopping bays at an individual station (Figure 57).

In order for multiple stopping bays to function properly, the appropriate vehicle must have unencumbered access to its designated stopping bay. In Bogotá, the vehicles have this type of flexibility due to the provision of passing lanes at stations. The second set of busway lanes allows vehicles to pass others in accessing the correct bay.

In some instances, such as in Porto Alegre, roadway space may not permit a passing lane. However, Porto Alegre still manages to provide multiple stopping bays by ensuring the correct order of buses along the busway. This technique in which the order of vehicles is controlled is known as the “convoy” technique or the “platooning” of vehicles. In this scenario, two or more buses may run along the busway in a closely bunched pack. The order of the buses is set so that the first bus stops at the first stopping bay and the next bus stops at the subsequent stopping bay. Each stopping bay represents a different service or a different route. Unfortunately, the convoying or platooning of vehicles is quite difficult to manage and control. The buses must enter the busway in the appropriate order or there will be considerable delays and backing up of vehicles (Figure 58). Further, since passenger boardings will vary

for different vehicles, the dwell times will also vary. Some vehicles may needlessly wait behind others while a longer boarding takes place. Thus, in a convoy system the slowest vehicle will likely set the speed for the entire fleet. For these reasons, multiple stopping bays are probably best implemented through the provision of passing lanes at stations.

3.4.4.6 Vehicle velocity

System capacity is actually not strictly dependent upon vehicle velocity. A system can move 20,000 passengers per hour at 20 kilometres per hour as well as at 10 kilometres per hour. Prior to the development of the Bogotá TransMilenio system, the city possessed a median busway that catered to all private bus operators. The uncontrolled system meant that there was considerable congestion on the corridor. The congestion was due to buses stopping at random locations as well as the over- supply of less efficient smaller vehicles . Nevertheless, the previous system moved approximately 30,000 passengers per hour per direction, but it did so at an average speed of less than 10 kilometres per hour. The TransMilenio system moves a similar number of passengers but at an average commercial speed of approximately 27 kilometres per hour. Figures 59 and 60 provide a visual comparison of Bogotá with the previous uncontrolled busway and with the TransMilenio BRT system along the same corridor.

Clearly, from the perspective of minimising travel time and fulfilling customer preferences, a rapid service is more desirable. While velocity and capacity may not be directly dependent, many factors that affect passenger capacity also affect average velocity. The factors that affect average velocity (i.e., “commercial” velocity) are:

? Number of busway lanes ? Dwell times

? Headways

? Vehicle acceleration and deceleration characteristics ? Number of controlled intersections

As the number of vehicles on the corridor increases, the level complexity and opportunity for conflicts also increases. In turn, these conflicts between vehicles lead to reduced velocities and increased travel times. Figure 61 shows the relationship between the frequency of vehicles and the average velocity on the Avenue Caracas Corridor in Bogotá.

Figure 61 Relationship between average velocity and frequency of vehicles in Bogotá

Source: Steer Davies Gleave

Thus, to maintain a system that both achieves high passenger capacities and high average velocities, the BRT system will necessitate the inclusion of design principles that promote unencumbered operation.

3.4.4.7 Capacity calculations

The passenger capacity of a given corridor is calculated based upon the discussed factors of vehicle capacity, load factors, service frequency, dwell times, and stopping bay configurations. Quite often a software model will assist in calculating the expected capacity and flow rates based on these factors. In general, though, the overall corridor capacity can be calculated from the following equation:

Equation 3