analysing busway station capacity under mixed stopping ...eprints.qut.edu.au/80209/1/trb15-3949...
TRANSCRIPT
Analysing Busway Station Capacity under Mixed Stopping and Non-
Stopping Operation Rakkitha Widanapathiranage1, Jonathan M Bunker2, and Ashish
Bhaskar3
Paper 15-3949, TRB 94th Annual Meeting
(1) Doctoral Student, (2) Associate Professor, (3) Senior Lecturer Queensland University of Technology, Australia
Bus Rapid Transit Advantages
Dedicated Running-
way
High Station
Inclusions
Premium Bus
Services
Increased Ridership
Speed Reliability Efficiency Identity Customer Satisfaction
Rationale
โข Some BRT lines have โnon-stoppingโ buses passing certain stations
โข Brisbaneโs South East Busway for example
โข This study addresses this phenomenon
All buses stop
Capacity through greatest constriction โข Usually
busiest stop
Transit Line Service Capacity
Some buses non-stopping
?
BRT Line Capacity Estimation Knowledge Gap
Capacity Definitions
TCQSM Service Capacity โข Stipulated repeatable,
safe working conditions โข Operating margin avoids
congested operation โ Average dwell time โ CV of dwell time โ Z variate
Potential Capacity โข No operating margin โข Represents maximum
possible outflow โข all other conditions as
for service capacity โข Degree of saturation = 1
Study Methodology
compare
verify Base deterministic potential capacity โขno operating margin โขactual number loading areas
Base simulation capacity โขCV dwell time = 0 โขtrain-like throughput
AIMSUN microscopic simulation testbed
โขAv dwell time โขAv clearance time
โขHeadway distribution โขDwell time distribution AIMSUN API
Field surveys
Some non-stopping buses
Mixed-Stopping Buses potential
capacity
(TCQSM) potential capacity โขno operating margin โขeffective number loading areas
All-Stopping Buses potential
capacity
Bus-bus interference โขCV dwell time โฅ 0 โขmerging behavior
Buranda Station South East Busway, Brisbane Australia
N
200m (650ft)
Buranda station
Eastern Busway
Eastern Busway
South East Busway
To CBD Cleveland
railway
Buranda Station Simulation Testbed
Inbound platform
Outbound platform
B
A
CBD suburbs
Buranda Station Measured Headway Distributions
0
0.01
0.02
0.03
0.04
0.05
0.06
0 20 40 60 80 100
Prob
abili
ty D
ensit
y
Headway (s)
Exponential(0.055)
Buranda Station Measured Dwell Time Distributions
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0 10 20 30 40 50 60 70
Prob
abili
ty D
ensit
y
Average dwell time (s)
Log-normal(2.718,0.612)
AIMSUN Microscopic Simulation Model Development
Feature AIMSUN default AIMSUN modified using API
Arrival distribution Normal Negative exponential Trajectory Car-following = Dwell time distribution Normal Log-normal Merging Gap acceptance =
Driver reaction time 0.75s moving 1.35s stationary =
Simulation time step 0.15s =
Simulation Model Development Scenarios and Experimental Values
Simulation Model
Percentage Non-stopping
Buses
Av Dwell Time (s)
CV of Dwell Time
Base potential capacity (ASB)
0 Incremental 5s to 90s
0
ASB potential capacity
0 Incremental 5s to 90s
0.4, 0.5, 0.6
MSB potential capacity
10, 20, 30, 40 Incremental 5s to 60s
0.4, 0.5, 0.6
Base Deterministic Potential Capacity (Train-like operation)
๐ต๐ =3600
๐ก๐ + ๐ก๐๐๐๐
Where: ๐ต๐ = potential capacity (bus/h) ๐ก๐ = fixed dwell time (s) ๐ก๐ = fixed clearance time (s) e.g. 16s ๐๐๐ = actual number of loading areas e.g. 3
Base Simulation Model and Base Deterministic Potential Capacity
0
100
200
300
400
500
600
0 10 20 30 40 50 60 70 80 90 100
Pote
ntia
l cap
acity
(bus
/h)
Dwell time (s)
deterministic model with zero dwell time coefficentsimulation results with zero dwell time coefficent
(TCQSM) Potential Capacity with No Operating Margin
๐ต๐ =3600
๐ก๐ + ๐ก๐๐๐๐
โข Where: ๐ต๐ = potential capacity (bus/h) ๐ก๐ = average dwell time (s) ๐ก๐ = average clearance time (s) e.g. 16s ๐๐๐ = effective number of loading areas e.g. 2.65
Proposed ASB Potential Capacity Model
๐ต๐๐๐|๐ =3600
๐ก๐ + ๐ก๐๐๐๐๐๐๐๐
โข Where: ๐ต๐๐๐|๐ = all-stopping potential capacity (bus/h) ๐ก๐ = average dwell time (s) ๐ก๐ = average clearance time (s) e.g. 16s ๐๐๐ = actual number of loading areas e.g. 3 ๐๐๐๐ = station bus-bus interference factor
Bus-bus Interference Factor
โข Accounts for loss of capacity: โ Varying dwell times causes asynchronous bus
movements, constraining lead LAsโ usage โ Shared priority gap acceptance process
โข Alternative approach to โeffective number of loading areasโ
Bus-bus Interference Factor (From Regression on Simulation Data)
00.10.20.30.40.50.60.70.80.9
1
0 10 20 30 40 50 60 70 80 90 100
Bus-
bus i
nter
fere
nce
fact
or
Average dwell time (s)
dwell time coefficent=0.4dwell time coefficent=0.5dwell time coefficent=0.6
Bus-bus Interference Factor (From Regression on Simulation Data)
๐๐๐๐ = 0.90 โ 0.004 ๐๐ฃ๐ก๐
โข Where: ๐ก๐ = average dwell time (s) e.g. 5s to 90s ๐๐ฃ = coefficient of variation of dwell time e.g. 0.4, 0.5, 0.6
Calibrated ASB Potential Capacity against TCQSM (no operating margin)
050
100150200250300350400450
0 10 20 30 40 50 60 70 80 90 100
ASB
pote
ntia
l cap
acity
(bus
/h)
Average dwell time (s)
TCQSM (CV = 0) CV = 0.4 CV = 0.5 CV = 0.6
Calibrated MSB Potential Capacity Model (CV dwell time = 0.4)
050
100150200250300350400450500
0 10 20 30 40 50 60 70
MSB
pot
entia
l cap
acity
(bus
/h)
Average dwell time (s)
0% (sim)10% (sim)20% (sim)30% (sim)40% (sim)0% (mod)10% (mod)20% (mod)30% (mod)40% (mod)
Proposed MSB Potential Capacity Model (From Regression on Simulation)
๐ต๐๐๐|๐ =๐ต๐๐๐|๐
1 โ 0.48 ๐๐๐๐
โข Where: ๐ต๐๐๐|๐ = mixed-stopping potential capacity (bus/h) ๐ต๐๐๐|๐ = all-stopping potential capacity (bus/h) ๐๐๐๐ = proportion of non-stopping buses
Stopping Buses Potential Capacity from MSB Potential Capacity Model
๐ต๐๐|๐ =๐ต๐๐๐|๐ 1 โ ๐๐๐๐ 1 โ 0.48 ๐๐๐๐
โข Where: ๐ต๐๐|๐ = stopping potential capacity under MSB operation (bus/h) ๐ต๐๐๐|๐ = all-stopping potential capacity (bus/h) ๐๐๐๐ = proportion of non-stopping buses
Non-stopping Buses Potential Capacity from MSB Potential Capacity Model
๐ต๐๐๐|๐ =๐ต๐๐๐|๐๐๐๐๐
1 โ 0.48 ๐๐๐๐
โข Where: ๐ต๐๐๐|๐ = non-stopping potential capacity under MSB operation (bus/h) ๐ต๐๐๐|๐ = all-stopping potential capacity (bus/h) ๐๐๐๐ = proportion of non-stopping buses
Stopping and Non-stopping Capacities under MSB (ASB Capacity 100 bus/h)
89 76
22
51
112 126
020406080
100120140
0 0.1 0.2 0.3 0.4
Pote
ntia
l cap
acity
(bus
/h)
Proportion of buses non-stopping
Stopping Non-stopping MSB
Conclusions for Application
โข Microscopic simulation model can replicate deterministic BRT capacity estimation
โข Bus-bus interference factor useful alternative to โeffective loading areasโ โ relates dwell time variation
โข Models developed can estimate BRT line capacity when non-stopping buses operating โ Higher capacity than when all buses stop
Further Research
โข More comprehensive validation using observational data
โข Definition of โpractical capacityโ according to degree of saturation โ Requires model of queuing / delay upstream of
station
With Thanks
โข Queensland Department of Transport and Main Roads, TransLink Division โ data and access
โข Mr Hao Guo โ smartphone data collection app
โข Student survey team