A simulation study to investigate runway capacity using TAAM

Abstract

This study outlines a method to evaluate runway layouts using simulation, to aid in the airport planning and decision making process. As a sample study, the maximum throughput capacities of proposed expansion alternatives at Philadelphia International Airport (PHL), constrained at varying levels, are identified. The objective is to compare these ultimate airport capacities achievable for each of the different layouts to estimate their respective efficiencies in terms of runway system utilization. TAAM (Total Airspace and Airport Modeller) is used to simulate each proposed alternative given its capabilities for modeling at a very high level of detail and closely representing reality in terms of applicable separation standards and air traffic control procedures.

Full text

Page 1

Proceedings of the 2002 Winter Simulation Conference
E. Yücesan, C.-H. Chen, J. L. Snowdon, and J. M. Charnes, eds.

ABSTRACT
This study outlines a method to evaluate runway layouts
using simulation, to aid in the airport planning and decision
making process. As a sample study, the maximum
throughput capacities of proposed expansion alternatives at
Philadelphia International Airport (PHL), constrained at
varying levels, are identified. The objective is to compare
these ultimate airport capacities achievable for each of the
different layouts to estimate their respective efficiencies in
terms of runway system utilization. TAAM (Total Airspace
and Airport Modeller) is used to simulate each proposed
alternative given its capabilities for modeling at a very
high level of detail and closely representing reality in terms
of applicable separation standards and air traffic control
procedures.

1 INTRODUCTION

Airports hold a key role in the commercial aviation system
by allowing airlines and their customers to converge.
However, they now face the challenge of meeting the
growing demand for air transport. In fact, a lack of airport
capacity has been forecasted by the FAA to be one of the
most serious constraints to the growth of commercial and
private aviation (Wells, 2000). One main reason for this
lack of capacity is that airport development projects are
enormously capital-intensive and probably some of the
largest infrastructure development projects that are under-
taken. Hence, it is a challenging task for airports to keep
pace with the rapidly growing demand for air transport
(Dempsey, 2000). This fact also accentuates the impor-
tance of thorough analysis of the various options and their
outcomes at the planning stage. Demand-capacity analysis,
therefore, plays a key role in defining the physical re-
quirement of airport facilities to meet future demand.
This simulation study investigates different runway
configurations to evaluate each of the airport layout in
terms of runway system capacity utilization. Indexes are
computed under varying levels of ground and airspace
constraints. The objective is to make a comparison be-
tween these indexes, which are essentially measures of
utilization, computed for each scenario.
From a planning perspective, this would allow more
informed decision making, by providing estimates of effi-
ciency in terms of design functionality, sensitivity to tech-
nological and procedural improvements and overall utiliza-
tion of potential capacity.
As a sample study of the application of the above
evaluation methodology, two proposals for expansion at
Philadelphia International Airport (PHL) were investi-
gated. For each alternative the capacity measures discussed
above were determined to arrive at the runway system
utilization indexes described above. A comparison between
these indexes was made and inferences were drawn with
regard to the best alternative in terms of the factors dis-
cussed above.

2 LITERATURE REVIEW

2.1 Capacity

An airport’s capacity may be broadly defined as its ability
to handle a given volume of traffic (demand). Congestion
occurs when demand approaches or exceeds capacity.
The Airports Council International (ACI) and Interna-
tional Air Transport Association (IATA) guidelines for
airport capacity/demand management (1996) defines the
most significant aspect of an airport’s capacity, Runway
System Capacity, as the hourly rate of aircraft operations
which may be reasonably expected to be accommodated by
a single or a combination of runways under given local
conditions.
The Runway System Capacity is primarily dependent
on the runway occupancy times of, and separation stan-
dards applied to successive aircraft in the traffic mix. Other
key items affecting runway capacity include: availability of
exit taxiways, especially that of high speed exits that help



A SIMULATION STUDY TO INVESTIGATE RUNWAY CAPACITY USING TAAM


Massoud Bazargan
Kenneth Fleming
Prakash Subramanian

Embry Riddle Aeronautical University
600, S. Clyde Morris Blvd.
Daytona Beach, FL 32114, U.S.A.









1235

Page 2

Bazargan, Fleming, and Subramanian

minimize runway occupancy times of arriving aircraft; air-
craft type/performance; traffic mix; Air Traffic Control
(ATC) and wake vortex constraints on approach separa-
tion; weather conditions [Visual Meteorological Condi-
tions (VMC)/Instrument
(IMC)]; spacing between parallel runways; intersecting
point of intersecting runways; mode of operation, i.e., seg-
regated or mixed.
To better explain the capacity measures introduced
here, we may begin with the concept of Practical Capacity.
This is defined as the number of operations that can be ac-
commodated in a given time period, considering all con-
straints incumbent to the airport, and with no more than a
given amount of delay (Wells, 2000). On a typical delay
curve, this may be depicted as in Figure 1 (Raguraman,
1999).












Figure 1: Practical Capacity: λP

Maximum throughput capacity or Saturation capacity
may be measured as the number of operations that can be
accomplished in a given period of time disregarding any
delay that aircraft might experience and assuming that the
aircraft will always be present, waiting to land or take-off
(Wells, 2000, Ashford and Wright, 1992). This may be de-
picted as in Figure 2.












Figure 2: Saturation Capacity: λS

In this study, three measures of capacity are deter-
mined for each scenario. These are essentially saturation
capacities of each layout constrained at varying levels.
Each of these is discussed below.
Meteorological Conditions
Fully constrained capacity (λS1), takes into account all
constraints that exist in an airport environment. These in-
clude both layout/ground factors as well as airspace fac-
tors. Ground constraints include the location of runway ex-
its and taxiway and apron capacity. Airspace constraints
arise from factors such as increased controller workloads
due to the absence of sufficient procedural and technologi-
cal support. This measure of capacity is similar to what is
described by Reynolds-Feighan and Button (1999) as Ul-
timate capacity.
The second measure of capacity (λS2), which may be
called semi-constrained capacity, assumes that technologi-
cal and procedural improvements are in place. These im-
provements aid in maintaining separation standards more
precisely thereby increasing runway throughput. However,
the airport layout constraints discussed above, are still con-
sidered in determining this measure of capacity.
Finally, Unconstrained capacity (λU), assumes away
all constraints except those posed by safety requirements.
In particular, it is assumed that sufficient high speed run-
way exits exist allowing significant reduction of runway
occupancy times, taxiway and apron constraints are absent
and procedures to support high intensity runway operations
are implemented. This concept may be represented dia-
grammatically as in Figure 3. The concept of uncon-
strained capacity has been advanced by IATA and repre-
sents the maximum possible capacity of a given runway
configuration (Pitfield and Jerrard, 1999).












Capacity (movements/hour)
Figure 3: Unconstrained Capacity: λU

2.2 Capacity Estimation Models

A distinction between analytical and simulation models
may be made based on the methodology used to compute
capacity, delay or other such metrics. Analytical models
are primarily mathematical representations of airport and
airspace characteristics and operations and seek to provide
estimates of capacity by manipulation of the representation
formulated. These models tend to have a low level of detail
and are mainly used for policy analysis, strategy develop-
ment and cost-benefit evaluation (Odoni et al., 1997).
Most earlier analytical models generated to estimate
runway capacity such as that proposed by Harris (1972),
x
λPx
Delay (mins)
Capacity (movements/hour)
x
λPx
Delay (mins)
Capacity (movements/hour)
λS
λU
x
λPx
Delay (mins)
λS
1236

Page 3

Bazargan, Fleming, and Subramanian

subsequently extended by Amodeo, Haines and Sinha
(1977) aimed to compute the average interarrival time be-
tween aircraft over the runway threshold given a certain
mix of lead and trail aircraft. The inverse of this would
yield the runway arrival capacity per unit of the interarrival
time, using which, the hourly arrival capacity of the run-
way could be computed.
For mixed operations, the probability of releasing a
departure between arrivals could be factored into the model
for the arrivals only configuration assuming that departures
occur only when permissible by the separation between ar-
riving aircraft. If perfect interleaving of arrivals and depar-
tures was assumed, then the separation between arrivals
would have to be the greater of the minimum separation
required between arrivals and the minimum runway occu-
pancy time of the departure released between the two arri-
vals. Error correction factors were applied to these models
where appropriate. Most computer based models for run-
way capacity estimation in the late 70s and early 80s were
based on this fundamental logic (Weiss, 1978).
The primary analytical models that are used currently to
estimate runway capacity include, The LMI Runway Capac-
ity Model and the FAA Airfield Capacity Model (Odoni et
al., 1997). A hybrid of these two models, with the logic of
the LMI model and the extension to multiple runways fea-
tured in the FAA model has been recommended and is ex-
pected to be very useful in providing quick estimates of
runway system capacity (Odoni et al., 1997).
Simulation of the airport environment has been in-
creasingly used recently to obtain more realistic estimates
of capacity by randomizing the various input parameters.
In fact, meteoric improvements in computer technology,
especially in the areas of computer graphics; human-
computer interaction; computer networks; and the world
wide web, have had a significant impact on modeling and
simulation (Nance and Sargent, 2002). Fishburn and
Stouppe (1997) have suggested that simulation modeling
and analysis be integrated into the airport planning process
rather than being simply used for final evaluations.
Monte-Carlo simulations have been used extensively
to study the airport environment. This tool was used by Pit-
field and Jerrard (1999) to estimate the unconstrained air-
port capacity – taking only safety requirements into con-
sideration, and assuming all other factors such as air traffic
management and control procedures and best pilot prac-
tices as “ideal” - at the Rome Fiumucino International Air-
port. Pitfield, Brooke and Jerrard (1998) have also used
Monte-Carlo simulation to analyze potentially conflicting
ground movements at a new airport proposed in Seoul, Ko-
rea. This is a common simulation tool for sampling from
cumulative distributions using random numbers until a
steady state evolves. Given known or reasonable distribu-
tions, as the number of simulations increase, the results
match the distributions and predict the likely outcome.
In comparison to the above, microscopic simulation
models dedicated to airport or airspace types of simulation
seek to generate traffic flows through the airspace seg-
ments and airports which are modeled and are configured
to represent actual constraints and uncertainties. Observa-
tions from these flows allow appropriate measures of ca-
pacity and/or delay to be computed. Microscopic simula-
tions tend to have a much higher level of detail including
conflict resolution, airport taxiway and gate selection,
pushback maneuvering, etc., to deal with more tactical is-
sues (Odoni et al., 1997).
Microscopic models, can be either node-link or 3-
dimensional (3-D). Node-link models such as SIMMOD
and the Airport Machine separate the airport and airspace
into a number of nodes and links over which aircraft move.
Conflict occurs when more than one aircraft try to pass one
node. 3-D models such as TAAM and HERMES (Heuristic
Runway Movement Event Simulation), allow flight over
random 3-dimensional routes (Odoni et al., 1997).
A detailed compilation of all existing and required
modeling capabilities for ATM systems and concepts is
provided by Odoni et al. (1997). This study also presents
an exhaustive list of airport capacity estimation models to-
gether with extensive insights into and comparisons be-
tween these.
To summarize, a variety of techniques may be used to
evaluate runway capacity. These may range from basic
analytical models, through more sophisticated Monte-Carlo
and other random number probabilistic models, to complex
computer-intensive discrete event models requiring exten-
sive input data. The compromise in the choice of a tech-
nique lies between “the higher reliability of the results of
the higher-order model versus the increased effort and
cost” (Mumayiz, 1997).

2.3 TAAM Review

Developed by The Preston Group (now Preston Aviation
Solutions) in cooperation with the Australian Civil Avia-
tion Authority, TAAM (Total Airspace & Airport Model-
ler) is a large scale detailed fast-time simulation package
for modeling entire air traffic systems. The model is a four
dimensional flight path simulator and allows greater real-
ism than mesh based simulations such as SIMMOD (Odoni
et al., 1997). A number of factors may be randomized in
the simulation to reflect day-to-day fluctuations. A versa-
tile simulation model, TAAM has been used in a wide va-
riety of applications including airport capacity estimation
(gate, taxiway, runway capacity), planning airport im-
provements, extensions, de-icing, noise impact, effect of
severe weather, design of terminal area procedures
(SIDs/STARs) and terminal area ATC sectors, controller
workload assessment, impact of new ATC rules, system
wide delays and cost/benefit studies.
1237

Page 4

Bazargan, Fleming, and Subramanian

Being a large scale simulation of an air traffic system,
TAAM requires comprehensive input data files describing
the entire Air Traffic system. The level of detail, however,
is variable and can be adapted to suit individual project
needs. Typical inputs include, the airport layout, air traffic
schedule, environment description, aircraft flight plans and
air traffic control rules. These are used to investigate the
usage of the airport and airspace, conflict detection and
resolution, and to compute aggregate metrics using
TAAM’s internal algorithms and user specified rules
(Odoni et al., 1997). These aggregated metrics include sys-
tem delay and its distribution; costs: fuel, non-fuel, and to-
tal; airport movements; operations on taxiways and run-
ways; runway occupancy and airspace operation metrics
such as usage of routes, sectors, fixes and coordination.
TAAM has been verified by many users on many dif-
ferent scenarios. TAAM simulation outputs have been
compared with some FAA studies on aspects of new ATM
concepts and have shown comparable results. In fact, the
four dimensional movement of aircraft can be simulated in
TAAM to get within 3 - 4% of the actual aircraft profiles.
Airport movement rates and other characteristics can be
modeled with similar accuracy (Odoni et al., 1997). An op-
erational evaluation of TAAM by the Eurocontrol Experi-
mental Center (Sillard, Vergne and Desart, 2000), has pro-
vided detailed evaluation of the different aspects of the
model. The study identified a number of discrepancies and
limitations, however, experts in the field of airports, whose
opinions were solicited during the course of this study,
were in agreement that the model was responding to par-
ticular events or scenarios in a manner that reflected day-
to-day fluctuations in airport operations.
The evaluation also concluded that TAAM demon-
strates a significant capability to simulate an airport and its
environment in a manner that can be very close to reality.
Besides being recognized by ATC controllers who exam-
ined the baseline, this relative accuracy has been measured
through different sensitivity analyses.

3 AIRPORT LAYOUT EVALUATION

3.1 Airport Layouts in General

Most airport layouts are customized to represent the most
useful configuration given the airport environment. As a
result, the runway dependencies, airspace procedures and
limitations, and other characteristics are usually unique to
every airport. A more generic description of runway con-
figurations and their corresponding dependencies has been
laid out by the FAA. These configurations include the fol-
lowing:

1. Single runway
2. Close parallels (distance between runway center-
lines, less than 2500 feet)
3. Intermediate parallels (distance between runway
centerlines, 2500 – 4300 feet)
4. Far Parallels (distance between runway center-
lines greater than 4300 feet)
5. Dual lane (two pairs of close parallel runways
separated by more than 4300 feet)

Under instrument flight conditions, simultaneous in-
dependent approaches are permissible on far parallels. In-
termediate parallels can employ simultaneous dependent
approaches, requiring a diagonal separation between ap-
proaching aircraft. Close parallels are treated as a single
runway and simultaneous operations are not permitted
(Burnham, Hallock and Greene, 2001).
Airport layouts may correspond with one of the above
configurations or may be a combination of two or more of
them.

3.2 Evaluation Methodology

To begin with, the three capacity measures (λi) described in
section 2.1 are determined for each of the layouts. A stan-
dard assumption in the determination of these measures
was that visual meteorological conditions exist. Also in
each of the configurations studied, only the westerly flows
were considered. Hence we have,

1. λS1: Capacity as influenced by all constraints in-
cumbent at an airport – ground as well as air-
space constraints,
2. λS2: Capacity under procedural and technological
constraints – only Airspace constraints,
3. λU: Capacity in an unconstrained environment–
considering only safety related constraints such as
separation standards.

Based on the above measures of capacity, the follow-
ing ratios are computed for each layout,

1. λS1/λU: indicates the runway system utilization
owing to all constraints incumbent at an airport.
This would show where the layout stands, in
capacity terms, in light of its maximum potential.
Hence, [(λU-λS1) / λS1] indicates the potential for
maximum runway system utilization.
2. λS1/λS2 : provides an estimate of the utilization as a
result of airspace constraints. Therefore, the sensi-
tivity of the layout to technological and proce-
dural changes that improve the traffic flow in and
out of the airport is indicated by [(λS2-λS1) / λS1].
3. λS2/λU: indicates the utilization constrained by the
airport layout design factors affecting taxiing, gate
usage etc., thus throwing light on the layout’s
functionality or what may be called its design ef-
ficiency. Here again, [(λU-λS2) / λS2], shows the

1238

Page 5

Bazargan, Fleming, and Subramanian

potential for runway system utilization by improv-
ing airport design.

Comparison between different layouts are made based
on these indexes to arrive at the best configuration, primar-
ily in terms of,

1. Efficiency in terms of design functionality;
2. Sensitivity to technological and procedural im-
provements and;
3. Overall utilization of potential capacity.

3.3 Sample study: Philadelphia
International Airport

The FAA Capacity Benchmark Report (2001) has esti-
mated the current capacity benchmark at Philadelphia In-
ternational Airport (PHL) to be 100-110 flights per hour in
good weather (VFR conditions) and 91-96 flights (or
fewer) per hour in adverse weather conditions (IFR condi-
tions), which could include poor visibility or low cloud
base. Figure 4 represents a westerly usage of the runways
in VFR conditions. In the figures, the callouts provide the
runway names. The arrows present the usage of the run-
ways. An arrow toward a runway represents arrivals to that
runway while an arrow away from the runway represents
departures from that runway.
One of the current problems faced at PHL is that of
significant delays. For example, in 2000, over 4% of all
flights at Philadelphia experienced significant delay (de-
fined by the FAA as more than 15 minutes of delay). Un-
der IFR conditions, capacity is exceeded for about 3 1/2
hours of the day resulting in about 14% of the flights ex-
periencing significant delay. Moreover, traffic at PHL is
expected to increase by 23% over the next decade, which
will further increase delays. The capacity estimates in the
FAA report assume that the short runways 17/35 and 8/26
provide for 25% of airport traffic operations. The airport’s
capacity stands to decrease if this percentage declines.
Because of these current capacity problems, a number
of enhancement initiatives are being undertaken by the air-














Figure 4: Current West-VFR Operations at PHL

port authorities. Technological and procedural improve-
ments to be implemented include:

•
Automatic Dependent Surveillance-Broadcast /
Cockpit Display of Traffic Information with Local
Area Augmentation System [ADS-B/CDTI (with
LAAS)], which would provide a cockpit display of
the location of other aircraft thus helping pilots
maintain desired separations more precisely;
•
Flight Management System/Area Navigation
(FMS/RNAV) Routes, to enable a more consis-
tent flow of aircraft to the runway;
•
Land and Hold Short Operations (LAHSO), al-
lowing independent arrivals for specific aircraft
types on intersecting runways and
•
Precision Runway Monitor (PRM), a sophisti-
cated radar system that allows simultaneous in-
strument approaches to parallel runways as close
as 3000 feet apart.

According to the Capacity Benchmark Report, these
changes will improve Philadelphia’s capacity in good
weather by 17% (to 117-127 flights per hour) over the next
10 years, while capacity under adverse weather is expected
to increase by 11% (to 101-106 flights per hour).
Besides these, major expansions involving the con-
struction of new and/or expansion of existing runways and
taxiways, improved and/or new terminal area and cargo
handling facilities are being planned. These expansion
plans may be categorized under two broad concepts,

1. The Parallel concept, which is an extension of the
current layout, and
2. The Diagonal concept, which involves a complete
change of the layout including new runway
orientations, new terminal area design, new apron
and taxiway designs.

Under each of these concepts, two proposed full-build
layouts were chosen for purpose of this analysis. The
Parallel concept layouts chosen were:

1. Full-Build Parallel Layout With Crosswind Run-
way (Parallel-1) –The baseline layout altered to
have 09L/27R shifted to the south and west, 17/35
and 08/26 extended and a new runway, 09R/27L
built to the south of the airfield. The existing
09R/27L would also be extended and would now
be called 09C/27C. Figure 5 represents the full
build of this layout and also explains its usage.
2. Baseline Layout with 4th Parallel Runway (Paral-
lel-2)– This configuration is essentially the same
as the Parallel-2 except that here the crosswind
runway, 17/35 is converted to a taxiway in order
to provide for easier taxiing to and from the
09R/27L
09L/27R
08/26
17/35
1239