Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

South African

Highway Capacity Research

TRB WORKSHOP 153SOUTH AFRICAN HIGHWAY TECHNOLOGY AND PRACTICE

Christo van As

Alex van Niekerk

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

Need for South African Highway Capacity Research?

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

Need for South African Highway Capacity Research?

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy Need for South African

Highway Capacity Research

�Highway Capacity Manual– Often prescribed as minimum standard for

capacity improvements– Capacity improvement required when

• Level-of-service poorer than • LOS D in 30th Highest hour

�Therefore important to– Validate/Calibrate the HCM– For South African conditions

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

Recent Contributions

�Capacity Analysis of– Two-lane roads– Multilane highways & Freeways– Signalised intersections

South African National Roads Agency

Capacity R

esearchT

wo-Lane H

ighways

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy Two-Lane Highways

Measures of Effectiveness

�HCM Measures of Effectiveness– Percent Time-Spent-Following (PTSF)– Average Travel Speed

�PTSF, however, difficult to observe– HCM Allows a surrogate measure– Percentage vehicles with headways < 3 s– Termed “Percentage followers”

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy Two-Lane Highways

Research Methodology

�Observations using– Double inductive loops

�Traffic loggers– New traffic loggers capability developed– Directly observes percentage followers– Can be used to directly observe LOS

�Observations on– 25 Different two-lane highways

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

Two-Lane HighwaysPercentage Followers

�Observations indicate that HCM– Significantly overestimates

Percentage followers

�It was therefore necessary to develop a new local two-lane model

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy Two-Lane Highways

HCM Model: Typical result

South African National Roads AgencyT

wo-Lane H

ighways

New

Model

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

HCM OverestimationProbably due to shoulder use

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

Appropriateness ofPercentage Followers

as a

Measure of Effectivenessfor establishing

Level of Service

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

Percentage followers as aMeasure of Effectiveness

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

New Measure of Effectiveness“Follower Density”

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

New Measure of Effectiveness“Follower Density”

�Follower Density– Percentage Followers x Traffic density– Units: Followers per km (per lane)

�Advantages– Gives greater weight to traffic flow– Automatically takes speed into account

South African National Roads Agency

New

SA

NR

AL

Highw

ay Traffic M

odel

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

HTM Highway Traffic Model

Vertical queue

Platoon (queue) length modelled every 20m over length of road

Macroscopic “simulation”model based on catching-up

and overtaking rates

South African National Roads Agency

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

Capacity ResearchMultilane Highways & Freeways

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

Multilane Highways/FreewaysResearch Methodology

�Observations using– Double inductive loops– Newly developed traffic loggers

�Observations on– 20 Different freeways/highways

�Scope of research– Restricted to Basic Freeway Sections– Interchanges excluded

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

Multilane Highways/FreewaysSpeed/Flow Relationship

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

Multilane Highways/FreewaysPer Lane Speed/Flow Relationship

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

Multilane Highways/FreewaysLane Distribution

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

FreewaysTypical observed capacities

23001900South Africa24002300HCM

Capacity rangePc/hour/lane

Description

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

Multilane Highways/FreewaysPercentage Followers?

�Multilane Highways/Freeways�Modelled in terms of

– Percentage followers

�Advantage: Uniformity– Follower density used as a MOE for all

types of highways:• Two-lane Highways• Multilane Highways/Freeways

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

Multilane Highways/FreewaysPercentage Followers/Flow

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

Multilane Highways/FreewaysSpeed/Percentage Followers

South African National Roads Agency

Signalised Intersections

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy Signalised Intersections

Delay Models

�HCM delay model found to– Overestimate delay significantly

�New model was therefore developed– University research (Pieter Pretorius)– Improve delay estimates significantly

�Study methodology– Simulation evaluations

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy Signalised Intersections

HCM Delay Model

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy Signalised Intersections

New Delay Model

South African National Roads Agency

Conclusions

Conclusions

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

Conclusions

�Major differences found between USA & South Africa

�A need exists for a South African Highway Capacity Manual

�South Africa has the expertise– That can make a significant contribution– In the field of traffic engineering

Sout

h A

fric

an N

atio

nal R

oads

Age

ncy

Thank You

Images courtesy of SA Tourism