A4.5 Benefits from improved trip time reliability

Introduction

This section outlines how likely variations in journey time can be assessed and the benefits from improvements to trip time reliability incorporated into project evaluation. Trip time reliability is measured by the unpredictable variations in journey times, which are experienced for a journey undertaken at broadly the same time every day. The impact is related to the day-to-day variations in traffic congestion, typically as a result of day-to-day variations in flow. This is distinct from the variations in individual journey times, which occur within a particular period.

Travel time reliability is in principle calculated for a complete journey and the total network variability is the sum of the travel time variability for all journeys on the network. In practice, models may not represent the full length of journeys and this is accounted for in the procedure.

Travel time variability is expressed as the standard deviation of travel time. The sources of variability are road sections and intersections. Reduced variability arises from a reduction in congestion on links and at intersections along a route. For a single section or intersection approach the standard deviation of travel time can be calculated using that section or intersection movement's VC ratio:

where: the VC ratio is represented by s, s0, b and a are taken from table A4.5

Major incidents

The travel time variability that may result from major incidents on the road network is not accounted for in this procedure. For example, where there are high levels of congestion on motorways, a major incident will produce large travel time delays. These delays are not included in the day-to-day variability calculations.

The effect of a major incident will be related to the amount of spare capacity at the location. A specific analysis should be undertaken to determine the economic cost of delays from major incidents.

Reliability benefits calculation

The claimable benefits from improving trip reliability are calculated as:

0.9 × travel time value ($/h) (table A4.1, A4.2 or A4.3)

× reduction in the network variability (in min) / 60

× traffic volume for time period period (veh/h)

× correction factor (table A4.6)

where the reduction in network variability is the difference between the sums of the variability for all journeys in the modelled area for the do minimum and project option. The 0.9 factor is the value of reliability based on a typical urban traffic mix. For projects with a significantly different vehicle mix, evaluators should use 0.8 for cars and 1.2 for commercial vehicles.

Table A4.5 Coefficients to calculate standard deviation of travel time

Note: Evaluations of small retail areas on 50 km/h sections of a rural highway should use the urban other (50 km/h) context.

Adjustment factor for variability calculations

In many cases, a project evaluation will consider a defined area which does not represent the full length of most journeys. As a result, the changes in journey time reliability will be overestimated. In these cases the variability estimates need to be adjusted. Table A4.6 gives some illustrative contexts where different factors might apply. An estimation of the variance of journey times which occurs outside of the evaluation area must be made and the appropriate correction factor intable A4.6 applied.

The trip time reliability benefit is adjusted by multiplying the calculated variability benefit by the factor.

Table A4.6 Adjustment factors to apply to variability calculations

Process for evaluating reliability benefits

1. Calculate standard deviation of travel time on each link between intersections and for each intersection movement or approach.
2. Square the standard deviations to produce variances.
3. Sum variances along each origin-destination path to obtain the total variance for journeys between each origin and destination.
4. Take the square root of the aggregated variance for a journey to give the standard deviation of the journey time.
5. Multiply the total trips for each origin-destination pair by the standard deviation of travel time and sum over the matrix to give the network-wide estimate of the variability.
6. Calculate the difference in variability between the project and do minimum networks.
7. Assess the percentage of variance occurring outside of the selected study area and select the adjustment factor.
8. Calculate the benefit from improving trip reliability using the formula provided above, namely: 0.9 × travel time value × reduction in the network variability/60 × traffic volume for time period period (veh/h) ×adjustment factor.

Network models with origin-destination information

Intersections should be analysed by movement at traffic signals and by movement or by approach for roundabouts and priority intersections. Variability for the uncontrolled movements at priority intersections should be set to zero.

For road sections, the calculation of the standard deviation of travel time assumes there is only one link between junctions or between changes in link context. If the model has more than one link between junctions then variability associated with such artificial network nodes should be set to zero.

Network skims compatible with the assigned flows should be used to aggregate travel time variances (square of standard deviation) along paths to create a matrix (or matrices where multiple paths are used) of journey time variance for origin-destination pairs. The square root of each cell in the resulting matrix will provide the variability (standard deviation) of travel time for that journey.

The total network variability is the sum of the products of the number of journeys between O/D pairs and the standard deviation of travel time for that journey.

It is important to note that the process above produces estimates of travel time variability as a function of VC ratio, reflecting the impact of day-to-day variations in travel demand. This is not the same as variations in individual journey times within a modelled period, a possible output of micro-simulation models. The variation in individual journey times from such models will be influenced by the driver, vehicle type, and generation factors used in the stochastic processes used in the model.

Evaluations without origin destination information

For individual intersection upgrades, the turning movements can be used as a proxy origin-destination matrix with the movement-weighted standard deviation being calculated for the intersection.

For project areas with more than a single congested intersection or link, an estimate of the proportion of trips that travel through more than one of these sources of variability must be made in order to approximate the total study area variability.

Two sources of variability

For two sources of variability, the reliability estimate for each trip direction is the sum of:

Variability for trips which travel only through source x:

and, for trips travelling through both source x and y:

where: F_x = trips that travel through only source x

F_x.y = trips that travel through both x and y

SD_x = standard deviation of travel time for trip at source x

SD_y = standard deviation of travel time for trip at source y

Note: The result of the above method should be multiplied by a correction factor from table A4.6.

Three sources of variability

For each of the three sources of variability, the reliability estimate is the sum of the individual components below:

Through source x only:

Through sources x and y only:

Through sources x and z only:

Through sources x, y and z only:

Where: F_x,y,z = trips that travel through all three sources x, y and z.

The result should be multiplied by a correction factor from table A4.6.

If traffic passes through more than three sources of significant congestion in the modelled area then evaluators must estimate the trip matrix and perform the calculation using the aggregation of journey variance method (with the correction factor from table A4.6).

Rural 2 lane roads

Table A4.7 contains travel time variability values for rural 2-lane roads of varying terrain and the volume to capacity ratio (see appendix A3.17). The time period used to calculate the VC ratio must contain a relatively constant level of traffic volume.

Table A4.7(a) Travel time variability - rural 2 lane road, level terrain

Table A4.7(b) Travel time variability - rural 2 lane road, rolling terrain

Table A4.7(c) Travel time variability - rural 2 lane road, mountainous terrain