Airfield Capacity

Amedeo R. Odoni

September 10, 2002

Airfield Capacity

• Objective
Q To summarize fundamental concepts re. airfield
capacity

• Topics
Q Definitions of capacity
Q Factors affecting capacity
Q Separation requirements
Q A simple model for a single runway
Q Capacity envelopes and capacity coverage chart
 Capacity Measures

• Maximum-Throughput Rate
• Average number of demands a server can
process per unit of time when always busy
– µ = maximum throughput rate
– E(t) = expected service time
1
E(t) µ=
• Level of Service (LOS) related
capacity
• Number of demands processed per unit of
time while meeting some pre-specified LOS
standards (must know µ to compute)

Capacity Definitions

• Maximum Throughput (or Saturation)

Capacity
The expected (“average”) number of operations (takeoffs
and landings) that can be performed in one hour on a
runway without violating ATC rules, assuming
continuous aircraft demand.

• Practical Hourly Capacity

The average number of operations that can be performed
in one hour on a runway with an average delay per
operation of 4 minutes.

Note: Definitions can also be extended to entire runway system.

 Capacity Definitions (2)
• Sustained Capacity
The average number of operations per hour that can be
“sustained” for periods of several hours; vaguely-
defined, workload-related, limited use; in US typically
set to about 90% of saturation capacity in VMC, near
100% in IMC

• Declared Capacity
The capacity per hour used in specifying the number of
slots available for schedule coordination purposes;
used extensively outside US; no standard method for its
determination; generally set to about 85-90% of
saturation capacity; may be affected by apron capacity
and terminal capacity

Factors Affecting Capacity

• Number and layout of •
Mix and sequencing of
active runways operations (landings,
takeoffs, mixed)
• Separation requirements
• Quality and
(longitudinal, lateral)
performance of ATM
• Weather (ceiling, system (including
visibility) human factor -- pilots
• Wind (direction, strength) and controllers)
• Runway exit locations
• Mix of aircraft
• Noise considerations
 Configuration 22L/27 - 22R/22L

Source: Idris (2000)

Role of ATC Separation

Requirements
• Runway (and airfield) capacities are constrained
by ATC separation requirements
• Typically aircraft are separated into a small
number (3 or 4) of classes
• Example: FAA classification
Q Heavy (H): 255000 lbs < MTOW
Q Large (L): 41000 lbs < MTOW < 255000 lbs
Q Small (S): MTOW < 41000 lbs
• Required separations (in time or in distance) are
then specified for every possible pair of aircraft
classes and operation types (landing or takeoff)
• Example: “arrival of H followed by arrival of S”
 IFR Separation Requirements: Single
Runway (USA)
Arrival-Arrival:
(1) Airborne separations on final approach (nmi):
Trailing aircraft
H L or B757 S
H 4 5 5/6*
Leading B757 4 4 5
aircraft L 2.5 (or 3) 2.5 (or 3) 3/4*
S 2.5 (or 3) 2.5 (or 3) 2.5 (or 3)

* Applies when leading aircraft is at threshold of runway

(2) Leading aircraft must be clear of the runway before

trailing aircraft touches down

IFR Separation Requirements: Single

Runway (USA) [2]
Departure--Departure (approximate, in seconds)
Departure
Trailing aircraft
H L + B757 S
H 90 120 120
Leading B757 90 90 120
aircraft L 60 60 60
S 45 45 45
Arrival-Departure and Departure-
Arrival- Departure- Arrival
• Leading aircraft must be clear of runway at the instant
when trailing aircraft starts takeoff roll or touches down
on the runway, respectively. In D-A case, trailing arrival
must also be at least 2 nmi from runway when takeoff
run begins
 Separation Requirements
(Italy; until recently)
H M /L S

Arrival/Arrival H 5 5 7
5 5 5
(in nautical miles) M /L  
S 5 5 5
Departure/Departure
120 seconds between successive departures

• Departure/Arrival
Q Arrival must be at least 5 n.mi. away
from runway threshold

Parallel Runways (IFR): USA

Separation Arrival/ Departure/ Arrival/ Departure/
between arrival departure departure arrival
runway
centerlines

700-2499 As in single As in single Arrival Departure is

ft runway runway touches clear of
down runway

2500- 4300 1.5 nmi Indep’nt Indep’nt Indep’nt

ft (diagonal)

4,300 ft or Indep’nt Indep’nt Indep’nt Indep’nt

more
 The diagonal separation between two aircraft
approaching medium-spaced parallel runways

Aircraft i

Sij = 1.5 n. mi. d [2,500 ft. ≤ d < 4,300 ft.]

Q
Aircraft j

Staggered parallel runways; the “near” runway

is used for arrivals and the other for departures

“offset” arrivals runway

“near
end”

Q
departures runway
“far
end”
 Typical classification of weather conditions
(ceiling and visibility) at an airport in the United
States

VFR
2500
Ceiling (feet)

MVFR
1000
IFR
800

LIFR
CAT I
200
CAT II
100
CAT III
0 1/4 1/2 1 3 5
Visibility (miles)

Weather Category (%) By

Season -- Boston Logan Airport
Season
Weather
Category
Spring Summer Fall Winter Annual

VFR1 79.19 78.78 80.03 77.94 78.99

VFR2 / IFR1 10.26 13.86 11.73 12.42 12.07

IFR2 7.95 5.83 6.43 6.89 6.78

IFR3 .08 .08 .19 .02 .09

IFR4 2.50 1.45 1.63 2.80 2.07

Two high-capacity configurations with opposite
orientations at Boston/Logan

27/22L-22R/22L 4R/4L-4L/4R/9

North

Configurations: Same Direction,

Different Weather Conditions

LIFR VFR
A, B1, B2
B, C, D
B3, C2, D2

4L 4R
B3, C1, D1
4R

09
B, C, D

A, B1, B2 B3, C, D
 Typical Approach for
Estimating Airside Capacity
1. Compute average time interval for all
possible aircraft class pairs i, j
tij = average time interval between successive
movements of a pair of aircraft of types i and j (i
followed by j) such that no ATC separation
requirements are violated
2. Compute probability for all i, j
pij = probability of occurrence of the pair of aircraft
types i and j (i followed by j)

3. Compute overall average service time

E (t ) = ∑∑ pij ⋅ tij µ = E(t)
1
i j

Numerical Example
Aircraft Types
Type Mix (%) Approach Runway
Speed (kts) Occupancy
Given: Single Runway Time (secs)
(Arrivals Only: IFR) Heavy (1) 20 140 60
Large (2) 50 120 55
n = 5 N. Miles Small (3) 30 100 50

1 2 3
1  4 5 6 * * Applies only with lead aircraft at
[sij]= 2  3 3 4 *
threshold (all other separations
apply throughout final approach).
3  3 3 3 
A simple representation of a runway used for
arrivals only under IFR

Runway Final approach

ç ç
L s LT T

“Gate”

Single Runway Model

Q (Arrivals Only: IFR)

• Separation Requirements
Q Airborne longitudinal separation
• 3 - 6 n. miles, depending on aircraft pair, as
shown in matrix of example

Q Runway occupancy separation

• no two aircraft simultaneously occupy the
runway
 Effect of Airborne
Separation Requirement
QClosing Case
• Second aircraft is faster, and must have
required separation distance from first
aircraft at runway threshold; separation at
merge is greater than minimum.

Q Opening Case
• Second aircraft is slower, and must meet
separation requirement from first aircraft in
merge area when approach is initiated;
separation at runway threshold is greater
than minimum.

Appendix: Single Runway

Model
• Consider two aircraft, i and j
Let n = length of final approach (typically 5-8 n.mi.)
sij = separation in air between i and j
vi, vj = approach speed of i, j
O , O = runway occupancy time of i, j
i j

Ti,j = min. time separation between i and j at runway

Assume vi > vj
• Opening Case: Aircraft i precedes j
 n + sij
 
n 
Tij = max  − , oi 
 vj vi 


• Closing Case: Aircraft j precedes i

 sji 
 
Tji = max  , o j 
 vi 
 
 Matrix of Minimum
Separations
QThe number Tij in row i and column j is the
minimum separation(sec) for the case of
aircraft type i followed by type j
103 171 216 
 
Tij =  77 90 144 
 77 90 108 

• Opening Case
 10 n. mi. 5 n. mi. 
T12 = max  − , 60 sec 
 120 knots 140 knots 

= max (171sec , 60 sec ) = 171 sec

Matrix of Minimum
Separations [2]
• Closing Case
 3 n. mi. 
T31 = max  , 50 sec 
 140 knots 
= max (77 sec , 50 sec) = 77 sec

• Stable Case
 3n. mi. 
T 22 = max  , 55sec 
 120 knots 
= max (80 sec , 55 sec ) = 80 sec

• “Special” Case (also T 23)

 6 n.mi. 
T13 = max  , 60sec 
 100knots 
= max (216 sec , 60 sec ) = 216 sec
 Safety Buffer

• In practice, a safety buffer is added to

the minimum separations between
aircraft, to make up for imperfections in
the ATC system
• Allow a buffer of an additional B = 10
seconds between each aircraft for
safety (10 seconds implies about 1/3 n.
mi. longitudinal separation)

Matrix of Average Time

Separations
QThe number tij is the average separation
(sec) between an aircraft of type i and a
following aircraft of type j.

tij = Tij + b
113 181 226 
tij =  87 100 154 
 87 100 118 
 Matrix of Pair Probabilities
QLet pij = probability that an aircraft of type i
will be followed by one of type j
QAssume first-come, first-served (FCFS) runway
service
 0. 04 0.1 0.06
pij =  0 .1 0.25 0.15
 
Example  0. 06 0.15 0.09
• 20% of aircraft are Type 1, 50% are Type 2
• Therefore, the probability of a Type 1 followed
by a Type 2 is: p 12 = (0.2)*(0.5) = 0.1
• Note: This is only valid for an FCFS system; no sequencing.

Numerical Example [2]

1 2 3
Matrix of average time intervals, tij
(in seconds), for all possible pairs
of aircraft types:
1 113 181 226 
 
[tij] = 2 87 100 154 
3 87 100 118 
Matrix of probabilities, pij, that a
1 2 3
particular aircraft pair will occur:
1 0 .04 0 .1 0 .06 
 
[ pij] = 2  0 .1 0.25 0 .15 
3 0 .06 0.15 0 .09 
 Numerical Example [3]
Q By multiplying the corresponding elements of the
matrices [pij ] and [tij ] we can compute the average
separation (in seconds) between a pair of aircraft
on the runway in question.

That is: Numerically:

E (t ) = ∑∑ pij ⋅ tijE (t ) = (0. 04)(113) + (0. 1)(181) + (0. 06)(226)
i j + (0.1)(87) + (0. 25)(100) + (0. 15)(154)
é E(t) = 124 seconds + (0.06)(87) + (0.15)(100) + (0.09)(118)

Saturation
Capacity = 3600 seconds
124 seconds
= 29 aircraft

Numerical Example [4]

• The variance (a measure of variability) of the service
times (intervals between successive landings in this
case) can also be computed from:

σ t2 = ∑∑ pij ⋅ [ t ij − E (t )]2
i j

• Or,
(0.04)(113-124)2 + (0.1)(181-124)2 + …. + (0.09)((118-124)2
= 1542 sec2
• The standard deviation, σ t = √ 1542 = 39 seconds
 A typical capacity envelope for a
single runway

Departures/hour

Feasible
region
2

O 1 Arrivals/hour

Capacity envelope for two parallel runways, one

used for arrivals and the other for departures

Departures/hour

2
4

0 1 Arrivals/hour
 A hypothetical capacity envelope for a multi-
runway airport with mixed use of the runways

Departures/hour

Arrivals/hour

Runway Configuration
Capacity Envelopes
Runway Configuration Capacity Envelops
(Source: ETMS / Tower Records, 7-9 AM, 4-8 PM, July 1-15
1998 except Saturdays, Logan Airport)
25

4L/4R-9 (reported
Actual Arrival Rate (per 15 minutes)

20 average 68 AAR - 50
DEP)
27/22L-22R (reported
15 average 60 AAR - 50
DEP)
33L/33R-27 (reported
10
average 44 AAR - 44
DEP)

5 Single Runway (January

1999, reported average
34 AAR 34 DEP)
0
0 5 10 15 20 25

Actual Departure Rate (per 15 minutes) Source: Idris (2000)

 Capacity Coverage Chart

• CCC shows how much capacity is

available for what percentage of time
• Assumptions :
• airport will operate at all times with the
highest capacity configuration available for
prevailing weather/wind conditions
• the capacity shown is for a 50% -50% mix of
arrivals and departures
Note: Neither of these assumptions is necessarily true
in practice (e.g., noise may be principal
consideration in selecting configuration during
periods of low demand)

The capacity coverage chart for

Boston/Logan

Number of
movements
per hour

9
120

10
11
80
2 2a, 7, 12, 15
3
3, 3a, 16
40

0
20 40 60 80 100

Percent of time (%)

 Capacity Coverage Chart [2]
• The CCC summarizes statistically the
supply of airside capacity
• CCC requires a capacity analysis for all
weather/wind conditions and runway
configurations
• “Flat” CCC implies predictability and more
effective utilization of airside facilities
Q Operations (takeoffs and landings) can be scheduled with
reference to a stable capacity level
Q Fewer instances of under-utilization and over-utilization of
facilities

Runway configuration usage at Boston/Logan,

January 1999 (from Logan FAA tower logs)

35

30

25
Frequency

20

15

10

0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Time (hour)

4R/L-9 27/22L-22R 33L-27

33L-15R (Best for Noise) Other Good for Noise Other

 Range of Airfield Capacities
• The capacity of a single runway varies greatly
among airports, depending on local ATC rules,
traffic mix, operations mix, local conditions and
the other factors identified earlier (12- 60+
movements per hour is possible)
• At major commercial airports in developed
countries the range is 25-60 movements per hour
for each runway
• Depending on the number of runways and the
airport’s geometric configuration, total airfield
capacity of major commercial airports ranges
from 25 per hour to 200+ per hour

Capacity of Aprons
• Often a tough problem!
• Different stands can accommodate different
sizes of aircraft
• Remote vs. contact stands
• Shared use vs. exclusive use (airlines,
handlers)
• Dependences among stands
• Static capacity: No. of aircraft that can be
parked simultaneously at the stands. (Easy!)
• Dynamic capacity: No. of aircraft that can be
accommodated per hour. (Can be difficult to
compute.)
 Stand Blocking Time (SBT)

• Scheduled occupancy time (SOT) [20

mins – 4 hours, excepting overnights]
• Positioning time (PT) [3 – 10 mins]
• Buffer time (BT) [up to 1+ hour at some
locations]

SBT = SOT + PT + BT

A Simple Case

• Assume n stands; all can accommodate

all aircraft sizes
• Subdivide aircraft into K relatively
homogeneous classes w.r.t. SBT
K
E [SBT ] = ∑ pi ⋅ SBTi
i =1

• Dynamic capacity = n / E[SBT]