Connecting runway and gate capacity for higher capacity utilization at hub airports

(1)

Master’s Thesis

Connecting runway and gate capacity for

higher capacity utilization at hub airports

by Marcel Michelkens

Dual Degree in Operations Management

Rijksuniversiteit Groningen & Newcastle University

12.12.2016

Word Count: 11,408

Abstract

An increasing number of airports worldwide reach their capacity limits and aim for an optimized utilization of existing capacity to avoid or postpone capital intensive expansion projects. Research to date has focused primarily on runway capacity and utilization, as this is the critical capacity component of most airports. However, several hub airports experience shortages in gate capacity. This paper proposes research on a combined analysis of runway and gate capacity. It expands an existing scientific model to show gate capacity under the influence of aircrafts inflow from the runway system. The study shows that the aircraft ground time and thereby the aircraft throughput at gates is highly influenced by the runway system’s capacity resulting in a varying gate capacity over time. The research is set up as a case study at Amsterdam Schiphol Airport using quantitative data as well as expert interviews with various stakeholders.

(2)

Acknowledgments

This thesis marks the end of my studies and I would like to use the opportunity to thank some persons for their help and guidance on that way. I am very thankful to my supervisors Dr. Martin Land and Dr. Gu Pang for their recommendations and guidance from the first research proposal up to this final version. Especially the suggestions and valuable feedback of Dr. Martin Land helped me very much. As case study at Amsterdam Airport Schiphol, this study would not have been possible without Mark van Gaalen acting as an extremely helpful supervisor from Schiphol side also opening several doors for us and giving us complete freedom in our research. The time at Schiphol was also very pleasant due to the work with Marnix Reijgersberg not only by bridging my Dutch skills but also the frequent and constructive discussions. Finally, I would like to thank my girlfriend for her support and my parents for enabling me to study abroad and always supporting me on my way. Thank you!

Marcel Michelkens s3078035 / B5058626

m.michelkens@student.rug.nl

Supervisor & Assessor: Dr. M.J. Land (m.j.land@rug.nl) Second supervisor & Co-assessor: Dr. G. Pang (gu.pang@ncl.ac.uk)

Supervising Universities University of Groningen

Faculty of Economics and Business Nettelbosje 2

9747 AE Groningen The Netherlands

Newcastle University Business School 5 Barrack Road

Newcastle upon Tyne, NE1 4SE, United Kingdom

Case company

Amsterdam Airport Schiphol Evert van de Beekstraat 202 1118 CP Schiphol

(3)

List of Abbreviations

Abbreviation Description

AAG Aircraft at Gate

AAS Amsterdam Schiphol Airport

ADAC Adjusted Dynamic Apron Capacity

ATC Air Traffic Control

BT Buffer Time

BTO Blocktime Overshoot

Buf Buffer

DAC Dynamic Apron Capacity

EU European Union

FAA Federal Aviation Administration

ICAO International Civil Aviation Organization

LCC Low Cost Carrier

LVNL Luchtverkeersleiding Nederland (Dutch Air Traffic Control)

MaxCT Maximum Connecting Time

MCT Minimum Connecting Time

Pr Pier

PT Positioning Time

RASAS Regulation Aircraft Stand Allocation Schiphol

SACN Stichting Airport Coordination Netherlands (Dutch Slot Coordination)

SBT Standard Blocking Time

SOT Scheduled Occoupancy Time

(4)

1. Introduction

Aviation has been a growing industry with yearly increasing passenger, cargo, and movement figures since the start of commercial air services (Doganis, 2010). This trend is expected to hold for the next years: Eurocontrol (2016) forecasts a yearly increase in flights by 2.9% within Europe until 2022. However, airport capacity did not grow proportionally to traffic figures and led to congestion during peak times or the entire day especially at large airports (Gelhausen, Berster, & Wilken, 2013). In addition, capacity constrained airports are seriously affected by delays in their operations (Madas & Zografos, 2010).

In order to manage scarce airport capacity, the European Union (EU) introduced regulation (EC) 95/93 to establish a slot allocation system at European airports (European Commission, 1993). A slot is defined by EC 95/93 as “the permission […] to use the full range of airport infrastructure necessary to operate an air service at a coordinated airport on a specific date and time for the purpose of landing or take-off […]”(European Commission, 1993, p. 1). The current slot allocation system in the EU is still based on EC 95/93 and an amendment in 2004 (Avenali et al., 2015).

Although the mentioned EC 95/93 definition states the full range of airport infrastructure, literature has focused on runway capacity as equivalent for airport capacity. Very limited attention is given to other airport capacity factors like apron or gate capacity, although those can be relevant to determine capacity for some airports (Katsaros & Psaraki, 2012).

In terms of gate capacity, the Gate Assignment Problem has been intensively studied in the operations research field by developing heuristics, models, and systems for a faster or/and more efficient assignment of flights to gates (Diepen, Van Den Akker, Hoogeveen, & Smeltink, 2012; V. P. Kumar & Bierlaire, 2014). However, these studies take a fixed flight schedule as input without suggesting optimizations to the inflow of aircrafts from the runway system.

A joint analysis of both problems has been rarely studied in literature, although hub airports as Frankfurt or Amsterdam Airport Schiphol (AAS) are constrained by both their gate and runway capacity. With the ongoing growth in passenger figures, airports can either respond by capacity expansions or increased utilization of existing facilities (Abeyratne, 2000), which includes runways as well as gates. The runway system determines the in- and outflow of aircrafts to the airport and thereby to the gates. By connecting runway and gate capacity, an optimized in- and outflow from the runway system with regards to the available gates might improve the overall capacity utilization. The aim of this paper is to close this gap and show the influence of the aircraft inflow from the runway to gate capacity for gate constrained hub airports.

(7)

with the exploration of opportunities. Finally, chapter five summarizes the work through a discussion and conclusion.

2. Background / Theoretical Framework

The following section introduces the relevant capacity measures for airports in general, as well as runways and gates in specific. Finally, relevant literature is introduced and discussed leading to the research questions at the end of the chapter.

2.1 Defining Airport Capacity

Before detailing about capacity of specific airport components, the term capacity itself needs to be defined in an airport context. General operations management textbooks define capacity as “the maximum level of value-added activity that an operation, process, or facility is capable of over a period of time” (Slack, Brandon-Jonas, & Johnston, 2013, p. 701). The translation of this definition into an airport context is not obvious. Firstly, airport capacity might be measured in aircrafts (e.g. arrivals, departures, movements) or by their load (passengers, cargo, baggage) per time period depending on the capacity component. Secondly, three capacity measures are commonly used in airport planning.

As a first measure, the maximum throughput, ultimate or design capacity provides the maximum number of movements possible within a period of time without considering delay or any service level (Bazargan, Fleming, & Subramanian, 2002). When demand reaches design capacity, the level of delay is usually unacceptable (Wilken, Berster, & Gelhausen, 2011) and operating at this capacity is not favorable.

Therefore, the second term of practical or sustainable capacity considers a predefined service level and maximum delay which can be achieved over a longer period of time (Bubalo & Daduna, 2011). The practical capacity deviates significantly from the design capacity by only reaching 60-70% in terms of possible movements under the design capacity (Wilken et al., 2011)

Finally, the declared capacity is used as the first step of the slot allocation procedure and defined by the relevant regulation 95/93. It sets the number of movements at an airport, which can be operated at an predefined service level (Wilken et al., 2011). While the practical capacity stays a theoretical concept without direct implication, the declared capacity determines the number of slots per interval of time, which are issued for a coordinated airport. The commonly used time interval are hours but detailed capacities for shorter periods of 20 minutes intervals at AAS or 5 minutes at Brussels are also possible (Belgium Slot Coordination, 2016; SACN, 2016).

(8)

capacity can only be seen as a good estimator of runway capacity, if the runway system is the actual bottleneck of the airport (Wilken et al., 2011).

2.2 Runway System Capacity

Following the differentiation between airport capacity definitions, this section reviews the capacity determination of the critical component of most congested airports: the runway system. The capacity of a runway system is generally stated in movements (arrivals and departures) per time interval. It is hardly possible to give a general number of movements achievable on one runway due to the variety of influencing factors. As runway capacity determination fills entire textbook chapters (De Neufville, Odoni, Belobaba, & Reynolds, 2013; Horonjeff, McKelvey, Sproule, & Young, 2010), only a basic model and additional factors relevant for the case example are discussed.

The US Federal Aviation Administration (FAA) pioneered in runway capacity determination in (1976) by publishing a detailed research. It resulted in a general formula for runway capacity calculation, which can be adapted for more than a hundred different runway configurations. For a general and simplified capacity calculation, this formula is still the leading method to date (Horonjeff et al., 2010). Its components and factors on runway capacity are summarized in Table 1.

Factor Effect on Capacity of Runway System

Number of runways &

geometric location

Capacity generally increases with number of runways. Their geometric location to each other determines, whether or to what extent they can be used independently.

Capacity decreases, if they cannot be operated independently (e.g. no sufficient separation of parallel runways, crossing of runways) Share of landings Capacity decreases for most configurations, when share of landings

increases due to different separation requirements for take-offs and landings

Mix of Aircraft Types

Capacity decreases with increased operations of large (i.e. widebody) aircrafts, as separation requirements increase for them

Number & location of exit taxiways

Capacity decreases, when no or few exit taxiways from the runway are available, as the time to clear the runway increases

Touch-and-go operations

Capacity increases, when share of touch-and-go operations is high. Touch-and-go operations can be neglected for most larger airports Visual vs

Instrument Flight Rules (VFR / IFR)

Capacity increases, when VFR is used. While this is common at American airports when meteorological conditions are met, European airports usually operate completely under IFR (Morisset, 2010) Table 1: Runway Capacity Factors based on (Federal Aviation Administration (FAA), 1976)

(9)

Besides structural and operational factors, government policy influences permitted movement capacity. Some airports like AAS face a so-called “noise-budget”, which fixes the amount of noise produced within a period of time (usually a year). Further policy restrictions include night curfews or limitations for specific noisy aircraft types (Kuiper, Visser, & Heblij, 2012). Finally, all of the above factors are considered for the declared capacity. As mentioned in the slot definition, a slot is not limited to the usage of runways but includes all airport components including gates.

2.3 Gate Capacity

Research by Katsaros & Psaraki (2012) outlined, that the runway capacity can be significantly larger than the capacity of the apron. Besides the Greek regional airports of their analysis, Berlin Tegel airport faces a similar problem: While the runway system can handle 52 movements (arrivals + departures) with a maximum of 30 arrivals per hour, the gate occupancy is limited to 32 aircrafts (Airport Coordination Germany, 2016). In that case, the critical coordination parameters are both runway and gates for the determination of the declared capacity. However, research on apron capacity is limited, as this capacity component is highly specific per airport and not considered as bottleneck in most instances (Mirkovic, 2011).

In general, literature distinguishes between static and dynamic capacity of aprons and gates (De Neufville et al., 2013). When a gate is seen as a terminal near apron position and bus gates are excluded, gate and apron can be seen as equivalents (see definitions in 3.2). The determination of the static capacity is straightforward by summing up the number of all gates at the airport. However, this does not provide detailed insights in the number of aircrafts, which can be handled at these gates within a certain time. As described with runway capacity, the number of runways alone does not indicate the runway system’s capacity. In order to compare and connect runway and gate capacity, a time-based figure for gate capacity is required: The dynamic apron capacity calculates the number of aircrafts which can be served with the static capacity available within a certain time (De Neufville et al., 2013). Different approaches to calculate the dynamic capacity are presented in the following section.

Due to different aircraft types and gate sizes, some gates are not capable of handling certain aircraft types. Amsterdam Airport Schiphol (AAS) groups aircrafts into nine different size categories depending on wingspan and length to check whether a certain aircraft type fits to a specific gate (Schiphol Group, 2016a). A further constraint is set by the origin and/or destination of a flight usually grouped into flights from/to airports within or outside the Schengen area. In addition, airlines prefer certain gates for their flights or even have exclusive usage rights, which is common in the US (Diepen et al., 2012; Idris et al., 1998).

(10)

In conclusion, it is important to derive a capacity figure measured in flights per time interval. A scientific method is introduced in the following section. However, due to the mentioned constraints the calculated gate capacity and the actual capacity under a certain flight schedule are likely to differ.

2.4 Current State of the Literature

As indicated, the strategic connection of runway and gate is not studied in depth in literature so far. The following section introduces the existing literature of runway / gate capacity connection and apron capacity calculation.

The method for dynamic gate capacity calculation shown by De Neufville et al. (2013) remains the basis for most gate and apron capacity calculations or adaptions. The dynamic capacity determination starts with the calculation of the Standard Blocking Time (SBT) of a position by an aircraft of the size i:

SBTi = SOTi + PTi + BTi

with

SOTi = Scheduled Occupancy Time of gate / position

PTi = Positioning Time to/from position from/to runway

BTi = Buffer Time between departure and next arrival at position

After calculating the SBT for every aircraft category, the estimated average SBT for all aircrafts categories is calculated using their share in total arrivals:

E(SBT) = ∑𝐾_𝑖=1SBTi x pi

with

pi = share of arriving aircrafts of category i in total arrivals

Finally, the Dynamic Apron Capacity (DAC) per hour is derived by including the number of available positions (i.e. static capacity):

DAC = N

E(SBT)

with

N = Static Apron Capacity (number of stands available)

(11)

traditional scheduled airlines). While LCCs constantly achieved a BTO of less than 30%, traditional scheduled airlines mostly exceeded 30%. Therefore, the use of actual instead of scheduled ground times might be favorable. Finally, the method assumes a constant arrival rate of aircrafts and does not account for specific arrival patterns (i.e. waves) as experienced at most hub airports.

The first connection in literature between runway and gate capacity was established by Mirković & Tošić (2015). They grouped airports into three categories: only point-to-point traffic, only hub traffic with transfers, or mixed hubs with both kinds of traffic. Their analysis concluded, that a connection between runway and gate capacity exists for airports with a hub function. At hub airports, the transfer airline organizes an arrival wave followed by a certain time to allow passengers to transfer to the next flight and finally a departure wave. The time between the waves is set by the airport’s minimum connecting time (MCT). The total length of both waves and the MCT must not exceed the maximum connecting time (MaxCT), which passengers accept when transferring. The length of the arrival and departure wave depends on the ability of the runway system to allow aircrafts to arrive or depart as fast as possible. Mirković & Tošić conclude that a higher number of possible operations on the runway system during peaks could decrease the ground time of aircrafts. Due to later arrival and/or earlier departure through the increased number of movements, the gate capacity can be increased through the decreased ground time of aircrafts.

However, this model stays idealistic in several aspects. The MaxCT depends among others by customer groups (e.g. continental vs. intercontinental) leading to several MaxCTs for one airport or hub system (Burghouwt & De Wit, 2005). Although the MCT is clearly set by the airport, different MCTs exist depending on the transfer of the passenger. For example, the MCT for passengers at Amsterdam connecting within Europe is 40 minutes, while 50 minutes are required for intercontinental connections (Kusumaningtyas & Lodewijks, 2013). The strict use of MCT also ignores, that a connection might not be necessary for all flights: an arriving flight from Canada does not need a connection to a departing flight to the US at a European airport. In addition, the model does not take slots into account. It assumes that additional slots during the peaks could be used by the hub airline. However, new entrants are favoured in the allocation of new slots by the current slot allocation system (see Appendix 3) (Zografos, Salouras, & Madas, 2012). Moreover, an increase in permitted departures might lead to a decreased number of possible arrivals (or the other way around), as long as the runway system is not expanded. When more slots with grandfather rights exist than the new capacity plans, airlines would have to be forced to give back slots. However, this will create immense resistance by the affected airlines.

(12)

2.5 Research Question

The literature review reveals a gap between the capacity usage of runways and gates, as they have been analyzed separately to date. Therefore, the main aim of this research is to close the gap by the following research questions.

How can runway and gate capacity be linked at a hub airport based on the case example? How can a method account for the influence of the runway system to the gate capacity? How can airports with gate capacity shortages increase their gate utilization?

3. Methodology

In order to derive valid research results, the study has to be based on sound research methodologies and reliable data and sources. The following chapter explains the steps of the underlying research methodology from research design to data collection and analysis. Finally, the measures taken to ensure validity and reliability are introduced.

3.1 Research Design & Case Selection

In Operations Management, case research marks a powerful research method. For explorative research answering “how” and “why” questions, which aim at theory development, case research is the leading method (Voss, 2009). It can be used to identify and describe key variables, analyze their linkages, and find reasons for their linkages (Voss, 2009). When existing literature has not analyzed a phenomenon from a certain perspective, case study research offers the possibility to overcome the literature gap (Eisenhardt, 1989). As this thesis aims to link the capacity of two components namely gates and runways, a case study approach offers the required features.

Case research is usually based on one or few in-depth studied examples. Such as the majority of studies in the field of airport capacity, this study is based on a single airport case. While research on overall factors like average aircraft size (Berster, Gelhausen, & Wilken, 2015) compares several airports, detailed system analyses for runways (Bazargan et al., 2002) or gates (Daniel, 2014; Diepen et al., 2012) are usually conducted at one airport. Due to the complexity of a single, larger airport, detailed studies for several major airports are beyond the scope of most research projects. In addition, case studies for airports can often be generalized (Diepen et al., 2012).

(13)

2016c). A new pier is scheduled to open in 2019 followed by a new terminal in 2023, which will increase capacity significantly (Schiphol Group, 2015a). However, capacity shortages are expected until the opening of the new facilities calling for an optimal utilization of existing terminals including their gates. Due to AAS’s gate shortages and its position as one of the largest worldwide airports, it qualifies as case example. Moreover, the airport provided extensive quantitative data of its operation and established contacts for interviews with internal and external experts.

3.2 Research Setting

Amsterdam Airport Schiphol’s passenger facilities consist of several piers around the main terminal (see Appendix 1 for details). In total, 94 pier gates were in use in 2015 at seven piers (b-h). The piers are able to handle different aircraft sizes and are dedicated to Schengen and/or non-Schengen flights (see Appendix 2). As AAS serves as transfer hub several airlines, a few piers operate as the “Central transfer zone” to group their flights together for shorter transfer connections (see Appendix 2). Besides the piers, several buffer areas are primarily used for aircraft parking in case of longer ground times with the exception of b-buffer handling a high number of regional flights. The runway system consists of six runways, but only five are regularly used (see Appendix 1). Only segregated operations of the runways is allowed meaning that a runway exclusively handles departures or arrivals at a given time (ICAO, 2004). During peak times, up to three runways can be used at the same time in a so-called 2+1 configuration. Due to the segregated operations, either two runways are used for arrivals and the third one for departures or the other way around leading to an unequal number of possible departures and arrivals during peak times. Therefore, arrival and departure peak modes exist in the capacity declaration (Appendix 4). During off-peak times, two runways are in use.

As some terms have a slightly different meaning depending on the organization or literature they are used in, Table 2 clarifies the meaning of the core variables for this study.

Variable Description

Gate Specific location in pier for a flight, where passengers board/de-board the aircraft

Stand / Position Place where an aircraft is parked for turn-around process or to wait for next flight

Pier Buildings directly linked to main terminal where gates are located and passengers wait for their departure

Terminal Main building where security checks, check-in, baggage reclaim etc. take place

Pier Gate/Stand Stand directly at the pier, where passengers can walk to the aircraft mostly through a bridge or by foot (“walk in/walk-out”). As every pier gate belongs to a pier stand, the terms gate/stand/position might be used interchangeably when pier gates are considered

Remote Stand / Buffer Stand

(14)

Schengen/Non-Schengen Gate

Due to different security and immigration regulations, Schengen & non-Schengen passengers have to be separated and gates are dedicated for Schengen or Non-Schengen use. Dual-Use Gate Pier gates, which can be used for both Schengen and

non-Schengen flights by the use of two floors at the pier. Walk-in/walk-out

Gate/Position

Stands directly at the pier, where passengers walk the short distance from the pier to the aircraft by foot. Used at h-pier. Aircrafts at Gate (AAG) Number of aircrafts being parked at piers and/or buffers at a

certain moment in time. Table 2: Variable Definitions

3.3 Data Collection

The quantitative data is mainly sourced from databases of AAS. As this data is collected by the airport itself and used for various internal purposes, the data reliability is assumed to be high. However, different reporting standards within the airport created some complications and required a thorough set-up of a combined database. The main data set included all flights operated for the year 2015 with extensive information for every flight. A second data set provided information on all towing movements of aircrafts from one stand at the airport to another one. All analyses were based on the data for 2015, as this was the latest data set available for a complete year. It was decided against a partial use of 2015 and 2016 data (e.g. summer season 15 & winter season 15/16) to minimize possible errors in combination of data bases and increase possible comparability to other airports, as the publicly available data is usually on a yearly basis.

In order to establish theory triangulation, qualitative data was collected through nine interviews (Croom, 2009). The interviews were conducted with experts from different departments of the airport, an AAS-based handling company, the Dutch slot coordination, a major airline at AAS, and independent academics. The interviews were conducted in person and semi-structured, as this form allows both exploratory and explanatory insights (Saunders, Lewis, & Thornhill, 2009). Thereby, the interviews helped to explore the topic in more depth through the insights of different stakeholders. In addition, they provided the option to verify and explain findings of the quantitative analyses and discuss options of the opportunity exploration. The used interview guide and short summaries of all interviews can be found in Appendix 6.

Finally, internal documents of AAS on runway operating modes, gate characteristics and stand allocation were used as well as the publicly available capacity declaration.

3.4 Data Analysis

(15)

During the macro analysis phase (1), the usage and capacity of the runway system and the gates were analyzed individually. These analyses were mainly based on the quantitative data provided by AAS. Further information and explanation were included and verified through the expert interviews.

Building on the findings of the macro phase, the main gate capacity factor ground time and the main utilization parameter Aircrafts at Gate (AAG) were analyzed in more depth in the detailed analysis phase (2). Again, the expert interviews supported the explanation and verification of the quantitative analyses findings and gave further directions for analyses. In the following, the information of phase 2 allowed a calculation of the gate capacity based on the introduced method of De Neufville et al. (2013) and an enhanced formula.

Finally, opportunities were determined and evaluated through the use of the quantitative analyses, the expert interviews, and the literature.

Figure 1: Research Framework

Descriptive Characteristics on gate usage

Macro Analysis

 Declared & Runway Capacity

 Overall AAG Pattern

 Overall Aircraft Ground Time

 Gate Idle Time

Directions for detailed analyses

Detailed Analysis

 AAG Pattern by Pier

 Detailed Ground Time by Pier

Determination of Apron Capacity

 Standard Method

 Adjusted for Runway Inflow

Directions for Opportuntity Exploration Exploration of Opportunties  Interviews  Quantitative Analyses PHAS E 1 PHAS E 2 PHAS E 3

(16)

3.5 Quality Assurance

Reliability and Validity are of special importance for case study research and can be broken down in several dimensions (Voss, 2009). The relevant dimensions are presented in Table 3.

Criteria Description

Construct Validity  Usage of various sources: primary data from AAS and interviews with experts from different departments at AAS, other stakeholders, and researchers

 Clear outline of the research framework in methodology section

 Review of draft case study by peers

Internal Validity  Usage of multiple research methods (data analysis + interviews)

 Data triangulation through interviews with representatives from different departments within the airport, external stakeholders (slot coordination, handling agent, airline) and independent academics

External Validity  Generalizable research results for airports not limited to case example

 Interviews with experts on generalizability of results to further airports

Reliability  Detailed interview protocols for all conducted interviews and short summaries in appendix

 Thorough description of data analysis

Table 3: Validity and Reliability in Case Research (Croom, 2009; Rowley, 2002; Voss, 2009)

4. Findings

The findings of the research project are presented in the following chapter and organized according to the research framework presented in Figure 1.

4.1 Descriptive Characteristics

The general characteristics and nature of flights at AAS was intensively studied by Visscher (2015) analyzing the share of airlines, aircraft types and traffic regions. This information is also used for this study and this section expands on descriptive characteristics for the gate and pier usage.

The seven piers serve different aircraft categories and thereby also traffic regions and airlines. With the exception of the d-pier, all gates of a pier a relatively homogenous in terms of possible aircraft types and traffic regions. For simplification, this study assumes the homogeneity of gates at one pier and distinguishes only between wide- and narrow-bodies. A detailed description of the piers can be found in Appendix 2.

(17)

and not included in the analysis. Furthermore, one position of the e-pier was not in use during the entire year 2015 and is also excluded from the analysis.

The individual usage characteristics of the piers differ (Table 4). While positions serving narrowbodies handle on median seven arrivals per day (b-, c-, h-pier), widebody positions are only used by a median three arrivals per day (e-, f-pier). Piers handling both widebody and narrowbody flights are located in between with 5 arrivals (d-, g-pier). The buffer platforms excluding b-buffer are only used rarely over the day for long-term parking of aircrafts and are therefore not further considered in detailed analyses.

Arrivals per Day/Pier B-buf b-pr c-pr d-buf d-pr e-buf e-pr f-pr g-buf g-pr h-pr Total Pier Max 117 108 115 9 220 4 51 33 2 63 71 Median 105 89 92 3 169 0 37 24 0 34 50 Per Position Max 8 12 12 4 11 2 8 7 5 12 15 Median 3 7 7 0 5 0 3 3 0 5 7 Number of Positions in use 33 13 14 6 33 3 12 7 4 8 7

Table 4: Pier Usage per Day in 2015

4.2 Macro Analysis of Gate & Runway Component

4.2.1 Analysis of Declared & Runway Capacity

The current capacity declaration of Amsterdam is primarily based on two factors: runway system capacity and legal boundaries. Interestingly, gate capacity is not considered as coordination parameter in the capacity declaration so far. Amsterdam’s capacity declaration differs to the ones of other European hubs like Frankfurt: Most European airports declare a stable number of movements over the day, but Amsterdam declares different figures for departure-, arrival-, and off-peak periods (see declaration in Appendix 4 & 5) (Airport Coordination Germany, 2016; Belgium Slot Coordination, 2016; SACN, 2016). The operating mode changes over the day several times. According to interviewed experts this benefits to the existing wave structure at Amsterdam and allows a better organization of the transfer traffic.

The maximum number of possible runway movements within a certain period of time is set by the Dutch air traffic control LVNL (Luchtverkeersleiding Nederland). Their decision is based on the capacity of runways and airways as well as safety regulations. In addition, a buffer is included so that the actual number of movements might be slightly above the declared capacity in some short time intervals. The resulting maximum arrival and departure capacities are reflected in the peak modes of the capacity declaration (Appendix 4).

(18)

topic: On the one hand, the airport wants to use its capacity to the maximum possible, on the other hand the set movement limit must not be exceeded in any case as this is a very politically sensitive topic.

In combination, the airport could not operate at the peak limits set by LVNL all day and every day, as this would lead to an overshoot of the legal limit. Both factors have to be considered leading to a more complicated slot process.

4.2.2 Overall AAG Analysis

Although gates are not seen as critical for the capacity declaration so far, they operate at capacity limits during some periods of the day. When aircrafts cannot be placed at a pier gate, they have to be handled on a remote position. According to a capacity planner, AAS aims to handle 100% of passenger flights at pier positions with the exception of some regional aircrafts. An airline representative added that only a share of 95% is currently reached.

The number of aircrafts parked at a pier gate or at a buffer varies significantly over the course of the day from less than 40 in the evening to more than 100 in the morning peak (see Figure 2). In general, five peaks can be seen with the largest ones in the morning around 9:00 and in the evening around 20:00. During these peaks, a reasonable share of flights is handled at buffer positions (mainly b-buffer), as the pier gates are nearly completely utilized. At night, only a smaller number of aircrafts stays at AAS, as the hub airline places most narrow-bodies at out-stations to allow early morning flights to AAS. This leads to both a low number of aircrafts staying overnight in Amsterdam and an extreme incoming wave in the morning. However, the high amount of aircrafts at buffers during the third afternoon and the evening wave is surprising. The pier gates do not seem to be completely utilized when comparing the figures to the morning wave.

Figure 2: AAG of Buffers and Piers over the Day for 2015. Median values in bold line and values for 1st and 3rd Quartile in dashes.

0 20 40 60 80 100 120 00:00 02:00 04:00 06:00 08:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00 AAG

AAG Level over Day (Median, Q3, Q1)

(19)

4.2.3 Ground Time Analysis

The ground time of an aircraft at a position might depend on different factors: the aircraft size (i.e. category), the airline’s business model, the time of the day and the overall traffic level (see Table 5). However, Table 6 indicates, that the ground time changes only marginally if traffic level changes. Visscher (2015) reported that the median throughput time for an aircraft was up to 1:43h shorter in summer than in winter by comparing an eight weeks summer to winter period of 2015. From our analysis, only in very low demand periods (Q1), the ground time increases notably at some piers (e.g. e-pier) but stays relatively stable at other piers (e.g. g-pier). However, Visscher’s analysis calculated the overall ground time for aircrafts at all positions and buffers in one, which includes aircrafts parked at several buffers not operating flights. Instead, this analysis concentrates only on piers and buffers frequently used for daily operations neglecting buffers used for long-term parking, as they are not relevant for the gate analysis.

Additionally, Table 6 indicates a tremendous difference in median ground times in 2015 between the piers from 38 minutes at the h-pier to 143 minutes at the f-pier. This may be explained partially by operating airlines and used aircraft types. While the h-pier is exclusively used by narrow bodies of LCCs being known for very short ground times, the f- and e-pier serve mainly widebodies of network carriers. An interviewed handling expert indicated that the airline’s business strategy and the following required services during a turn-around process can be seen as the most important factor for the ground time.

Factor Effect Explanation

Airline Strategy / Business Model

Very high Determines range of services performed during turn-around process. Airlines with transfer passengers might also increase ground times to allow more possible connections Aircraft Size Medium –

High

In general, larger aircrafts require longer ground times. However, the main difference is between wide- vs. narrowbody (see next factor).

Flight Destination (long vs short haul)

High Long-haul flights usually require full range of ground services (cleaning, catering etc.), which are skipped for some short-haul flights

Airport’s Traffic Level

Low Ground time increases only in very low traffic periods (see Table 6)

Time of Day Medium Depends on traffic waves (see section 4.3.2) Pier vs. Remote

Stand

Medium Bussing process requires additional 10-20 minutes Type of Boarding

at Pier (Bridge vs Walk-Boarding)

Medium Boarding through a bridge instead of walk-boarding, where passengers walk from the terminal to the aircraft and board through front and rear door, takes around 10 minutes longer. Only available at the h-pier and used by most LCCs (e.g. EasyJet, Ryanair).

Table 5: Influencing Factors for Turnaround Time (based on interviews & data analysis)

(20)

pier capacity is needed (Schiphol Group, 2015b). Although the minimum parking time at a buffer stand is set at 30 minutes, the median ground times are significantly higher as it can be seen for the d-, e-, and g-buffer. While these buffers are used for complete aircraft turnarounds only in rare occasions, b-buffer handles a high number of regional flights on a daily basis. Therefore, the ground times of b-buffer are more comparable to piers also handling these types of flights (e.g. c-pier) than to the other buffers.

Median ground time b-buffer b-pier c-pier d-buffer d-pier e-buffer e-pier f-pier g-buffer g-pier h-pier Year 77 62 76 283.5 84 166 140 143 152 92 38 Busiest 4 weeks 72 62 74 119 84 144 136 144 146 93 38 Busy (Q3) 4 weeks 73 60 73 176 83 154.5 138 147 123.5 95 39 Median 4 weeks 74 59 74 369 80 161 129 144 161.5 88 37 Low (Q1) 4 weeks 82.5 66 82 244.5 87 160 147 139.5 163 91 40

Table 6: Median ground time (in min) of aircraft at certain Pier or Buffer for periods with different traffic levels in 2015.

4.2.4 Idle Time Analysis

The idle time of the position measures the time between a departure and the next arrival of an aircraft. Compared to the median ground times, the idle times do not vary a lot between the piers (Table 7). The median idle time for nearly all piers is between 45 to 60 minutes. Again the median value stays relatively stable regardless of the traffic level. However, a small tendency for shorter idle times during busy periods can be seen at most piers. As more flights are handled during these periods with a stable ground time, the additional flights can only be handled if idle time decreases.

Table 7: Median idle time in minutes of a position by pier and overall traffic level in 2015

According to RASAS, a minimum time separation has to be considered between the departure and next arrival at a stand. The separation is set at twenty minutes for all piers except the h-pier with ten minutes. This follows the logic of the presented DAC method of De Neufville et al. (2013) to include a certain Buffer Time (BT). In addition, the BT prevents the usage of the

(21)

gates to their design capacity, as delays will be unacceptably high when capacity is used to the design capacity.

Interestingly the idle time increases at some piers for the low demand periods (Q1). Following the finding from the previous section and the reference to Visscher (2015), it can be seen that aircrafts not being operated in low traffic periods are usually not parked at piers. Thereby, they do not block capacity at piers and are parked at buffer positions. If aircrafts would be parked at the piers, the idle time would decrease. An expert also pointed out, that wide-body aircrafts with a ground time above 210 minutes might be towed to a remote position to free up capacity.

4.4.5 Summary of Macro Analysis

The macro analysis pointed out that significant differences exist in ground times between the piers. Due to the wave structure with five peaks, an analysis of the ground time over the course of the day is required. It can also be seen that the no shortages in gate capacity exist during the night. As the main waves exist between 7:00 and 22:00 (see Figure 2), only this period will be analyzed further in detail. In addition, the d-, e-, and g-buffer do not play an integral role in the operation compared to the other piers and b-buffer and are therefore not analyzed in detail.

4.3 Detailed Analysis of Gate Utilization based on Runway Usage

The findings in the macro analysis call for a more detailed view: First, the AAG level needs to be analyzed by pier to see the utilization of every pier separately. Although the analysis is specific to the case airport AAS, a certain generalization to other airports can be assumed: A case study by Wendler (2010) outlined a structure with piers or terminals dedicated to Schengen and/or non-Schengen traffic at seven European hub airports similar to AAS . Secondly, the ground time requires a more in-depth analysis, as already the results of the macro analysis deviated significantly on the ones from Visscher (2015). Finally, the consequences of the declared peak mode capacities need further analysis to see the effects on the aircraft inflow.

4.3.1 AAG Analysis per Pier

The overall AAG analysis posted some questions including the intense use of b-buffer. This section analyzes the median AAG for all piers individually and groups them into piers with Schengen and/or dual-status gates (b-buffer, b-, c-, d-, h-pier) and non-Schengen gates (e-, f-, g-pier).

(22)

100% of flights to pier gates. However, the handling of regional jets (Fokker & Embraer) is accepted at remote position. According to an expert, a main airline prefers a clustered operation of these aircraft at b-buffer instead of assigning them to several piers with free gates. However, this attitude is slightly changing and more bridge positions at the b-pier are requested by the airline. In addition, the airline saves time and costs of pushbacks at these positions, as aircrafts can roll out without a pushback tractor at these positions according to another expert.

The h-pier handling only LCCs without transfer connections shows a rather constant usage over the day with a peak in the morning and the late evening. Experts validated this finding and pointed out the rather constant usage of this pier with a small peak in the morning and evening.

Figure 3: Median AAG for piers handling primarily narrow bodies for Schengen flights (d-pier serves up to 10 wide-bodies and has non-Schengen and dual-usage gates)

A different wave profile can be seen for the wide-body piers e, f, and g ( Figure 4). The wave profile is less extreme and only four waves can be seen over the day, as the second afternoon wave around 14:00 is missing. While all piers operate at or close to full capacity from the morning until the early afternoon, several stands are available in the later afternoon and evening. As only Non-Schengen flights can be handled at these piers, flights from the other piers cannot be relocated.

Figure 4: Median AAG for piers handling primarily wide-bodies for non-Schengen flights 0 5 10 15 20 25 30 00:00 02:00 04:00 06:00 08:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00 AAG

Median AAG Schengen Gates over day

bbuf b-pr c-pr dpr1 h-pr 0 2 4 6 8 10 12 00:00 02:00 04:00 06:00 08:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00 AAG

Median AAG non-Schengen Gates over day

(23)

The individual view on the AAG level points out, that the waves and piers have to be considered separately, as different traffic structures between the waves lead to different capacity utilization at the piers.

4.3.2 Detailed Ground Time Analysis over the Day

The macro analysis of the ground times revealed a large difference among the piers. A capacity planner emphasized that the whole day has to be analyzed in detail due to the traffic structure with peaks. Therefore, the ground time is analyzed by a throughput diagram for the core time of operations taking the median number of cumulative departures and arrivals for 2015 (Figure 5).

A throughput diagram visualizes the cumulative inflow (i.e. arrivals) and cumulative outflow (i.e. departures) of items into a production system. The vertical distance between the in- and outflow graph then reflects the work-in progress (i.e. AAG) at any point in time. The average time of a work in progress item in the system (i.e. ground time of aircrafts) can be measured as the horizontal distance (Nyhuis & Wiendahl, 2009). The thus calculated average ground time of an aircraft arriving at a certain time is depicted as the green curve in Figure 5.

Figure 5: Cumulative Arrivals & Departures and resulting ground time for all piers & b-buffer. Median values based on complete year dataset.

The average ground time of aircrafts deviates over the day by close to half an hour. Also for the ground time, a certain wave structure can be seen with minima before the peaks in AAG and maxima after these peaks. As most aircrafts arriving in the late evening stay overnight, the average ground time increases sharply after 20:30.

0 25 50 75 100 125 0 150 300 450 600 750 07:00 09:00 11:00 13:00 15:00 17:00 19:00 21:00 A ve rag e Gr o u n d Ti m e (m in ) Fl ig h ts Time of Day

Throughput Analysis all piers & b-buffer

(24)

By comparing the ground time to the number of incoming and outgoing flights this pattern becomes more obvious (Figure 6). During arrival peaks the ground time decreases and increases again for the departure peak. The difference is between ten to twenty minutes. This means that aircrafts arriving during an arrival peak have a more optimized ground time than the ones arriving during departure peaks. One could argue that this follows the concept of Mirković & Tošić (2015): Ground time for hub airports decreases, when arrivals and departure peak modes exist for the runway system, as aircrafts can arrive and depart in short periods of times. It can be seen, that a link exists between the aircraft inflow from the runway system and the resulting ground times influencing gate capacity.

Figure 6: Overview of incoming and outgoing flights in 15 minutes intervals and ground times of arriving flights. Median values based on complete year dataset.

The ground time patterns of most piers follow in general the overall pattern (see b-pier example Figure 7). A significant difference can be seen for the e-pier (Figure 8) with the f-pier showing the same pattern: The ground time in the afternoon increases to more than five hours for these piers. This is a consequence of the low demand for non-Schengen widebody positions in the afternoon and evening. While aircrafts with very long ground times are towed to a remote position for parking to free up pier gates in the morning, this is not done due to the low number of these flights in the evening. This could be seen as hidden capacity reserve, as towing could also be done at these gates in the evening.

80 85 90 95 100 105 110 115 120 -20 -15 -10 -5 0 5 10 15 20 07:00 08:00 09:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 A ve rag e Gr o u n d Ti m e ( m in ) Fl ig h ts

Overview of Flight Volume and Resulting Ground Time

(25)

Figure 7: Cumulative Arrivals & Departures and resulting ground time for the B-Pier. Median values based on complete year dataset.

Figure 8: Cumulative Arrivals & Departures and resulting ground time for the E-Pier. Median values based on complete year dataset.

4.4 Gate Capacity Calculation based on inflow from Runway

In order to compare and connect the runway and gate capacity, equal units for both components are required. As the runway capacity is stated in arrivals and departures per time interval (usually per hour), this unit is derived by different methods for gate capacity as well. 4.4.1 Calculation of Stand Capacity per Pier

Following the calculation method of De Neufville et al. (2013) for Dynamic Apron Capacity introduced in section 2.4, it seems that AAS is able to assign nearly all flights to a pier gate during the entire day (Figure 9). The calculation uses the yearly median ground times for

0 20 40 60 80 100 0 20 40 60 80 100 07:00 09:00 11:00 13:00 15:00 17:00 19:00 21:00 A ve rag e Gr o u n d Ti m e (m in ) Fl ig h ts Time of Day

B-Pier

Cum. Arrivals Cum. Departures Ground Time

0 50 100 150 200 250 300 350 0 10 20 30 40 50 60 07:00 09:00 11:00 13:00 15:00 17:00 19:00 21:00 A ve rag e Gr o u n d Ti m e (m in ) Fl ig h ts Time of Day

E-Pier

(26)

every pier (Table 6) plus the twenty minutes separation (ten minutes for h-pier) between two flights as specified in RASAS. Thereby, the DAC for every pier is determined and summed up to derive at the total airport’s capacity. As block times were used for the ground time calculation and experts indicated that taxi times are not considered by the airport for gate calculation, these are not included in contrast to De Neufville et al. (2013). The analysis is done for blocks of twenty minutes, as the capacity declaration also defines peak capacities for this period. Additionally, an arrival or departure peak mode lasts for at least forty minutes. Although the arrivals in the arrival peak mode are close to the DAC, it seems that gate shortages are not a severe challenge. In contrast, the previous analyses and expert interviews suggested a different picture showing the shortcomings of this method. The changing traffic structure over the day is ignored by this method: While traffic in the morning includes a high share of widebodies with long ground times, mainly narrowbodies operate in the afternoon and evening. By taking median ground times of the year (see Table 6), the method assumes both a stable traffic mix with stable ground times over the day and a continuous inflow of aircrafts from the runway system. As shown, both assumptions do not hold for Amsterdam, and a more advanced method is required.

Figure 9: Comparison of median actual arrivals and dynamic apron capacity incl. and excl. b-buffer for twenty minutes intervals

4.4.2 Calculation of Stand Capacity over the Course of Day

In order to account for the changing traffic structure and the varying inflow depending on the runway system, a more advanced method is required. Therefore, the ground times derived from the inflow of aircrafts in the throughput diagram in section 4.3.2 are used, which can be determined for every minute. These adjust for the changing ground times due to the different traffic structure over the day and account for the changing inflow of aircrafts. The results of this Adjusted Dynamic Apron Capacity (ADAC) can be found in Figure 10.

The curves directly indicate that the current schedule with up to 23 arrivals per twenty minutes in arrival peaks can only be operated by using b-buffer during peaks. However, the

0 4 8 12 16 20 24 28 7:00 8:00 9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 A ir cr afts p e r 20 M in u te s Time of Day

DAC vs. Actual Arrivals

(27)

number of arrivals is in three peaks still slightly higher than the ADAC including b-buffer. This calculation also uses a separation of twenty respectively ten minutes between two flights at a position. An apron expert indicated that this separation might be lowered in tactical planning for some flights based on operational experience on punctuality.

Figure 10: Comparison of median actual arrivals and adjusted dynamic apron capacity incl. and excl. b-buffer for twenty minutes intervals with standard separation of 20 and 10 minutes (h-pier) between two flights

By slightly adjusting the separation from twenty minutes to fifteen minutes based on the expert’s information, the shortages can be mostly solved (Figure 11). In the evening peak, a slight mismatch still exists which might be explained by summarizing flights to twenty minutes intervals. The remaining arrivals which can neither be assigned to a pier nor to b-buffer have to be handled at another remote position. Both interviewed airline and airport experts indicated, that this has to be done frequently in the traffic intensive summer season. Furthermore, the number of aircrafts being handled on additional remote positions increased over the last years and reached an all-time high in summer 2016.

In conclusion, the ADAC shows that Amsterdam is operating at its gate capacity limits during the peaks. While the DAC indicated a match of capacity and demand, the ADAC outlined shortages through its consideration of aircraft inflow from the runway system. This was validated through the expert interviews and further analyses.

0 4 8 12 16 20 24 28 7:00 8:00 9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 A ir cr afts p e r 20 M in u te s Time of Day

ADAC vs. Actual Arrivals; 20/10 separation

(28)

Figure 11: Comparison of median actual arrivals and adjusted dynamic apron capacity incl. and excl. b-buffer for twenty minutes intervals with standard separation of 15 and 10 minutes (h-pier) between two flights

4.5 Opportunity Exploration

The previous sections outlined the problem of gate capacity and also the connection to runway capacity. The analyses also showed some potential for improvements, which are explored in the following section.

4.5.1 Increasing the number of towing movements for long ground times

Under the current regulations, widebody aircrafts with a ground time of more than 210 minutes might be towed to a remote position to free up pier gates. An increased usage of this policy might solve some capacity issues in the morning waves at the e-, f-, and g-pier, where the majority of widebody flights is handled. However, an expert raised the concern of shortage of taxiway capacity through the increased number of towing movements in peak times. Another expert indicated that towing movements might cross and thereby reduce the capacity of an active runway, if the number of crossings per hour exceeds five. As taxiways are operating at their capacity limits already during peaks with the regular flights, more towing movements might lead to overcrowding and delays. Additionally, most flights with long ground times are already towed to buffer positions in the morning, for example at the e-pier (Figure 8).

4.5.2 Increasing number of Dual-Status Gates

The detailed analysis of AAG levels showed, that a distinction has to be made between the morning and evening peak. Although the overall number of aircrafts is comparable, the level of flights handled at buffer positions for the evening peak is much higher. The majority of flights in the evening peak are Schengen, which can only be handled at the b-, c-, d- or h-pier and b-buffer. However, the non-Schengen piers e, f, and g have free stands in the evening. Due to the sufficient capacity, several widebodies with long ground times are not towed to a

0 4 8 12 16 20 24 28 7:00 8:00 9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 A rr iv al s p e r 20 M in u te s Time of Day

ADAC vs. Actual Arrivals; 15/10 separation

(29)

buffer in the evening, as capacity does not require this action. Through towing, even more capacity could be realized at these piers in the evening.

The interviews showed that AAS is indeed considering this option: The potential of converting the south part of the d-pier from non-Schengen gates to dual-status gates is evaluated at the moment. Thereby, more Schengen flights could be handled in the evening peak and the currently handled non-Schengen flights could be moved to the e- or f-pier. In general the expert emphasized, that in an ideal situation all capacity would be as flexible as possible. Therefore, the new a-pier will be designed in a very flexible way to allow different kinds of traffic and possible future developments. Although the advantages of flexibilizing existing capacity are seen, the immense investments to restructure piers and terminals as well as space constraints mark a high barrier for this option.

A further constraint is set by the Dutch Air Traffic Control, LVNL, and was posted during an expert interview: From the current control tower, smaller aircrafts cannot be seen at certain pier stands, as the pier building is too high. Increasing the height of the pier by a further floor to enable parallel handling of Schengen and Non-Schengen passengers might lead to more stands which are not in the view of the controllers when used by narrow-bodies. This would then also require a new and higher control tower.

4.5.3 Decreasing Time Separation in Gate Planning

As mentioned, the current separation in gate planning between two flights is ten minutes for the h-pier and twenty minutes for the remaining piers. A relatively simple capacity increase could be achieved by lowering this buffer. Interviewed experts indicated, that on a tactical planning level, shorter separations are already used based on operational experience, as some flights tend to be usually on time or delayed.

As delays and too early arrivals exist (see also Visscher (2015)), a reduced separation on a strategical planning level would only increase the operational problems. If the system-wide on-time performance could be increased, a reduced separation might however be a point of discussion according to the expert.

4.5.4 Changing declared capacity

As shown, the runway system controls the inflow to the apron system and its capacity is mainly set by legal policies and the air traffic control. Experts indicated that an increased number of flights would technically be possible. These changes would be implemented through a modification of the declared capacity leading to new number of issued slots per time interval. Depending on the point of view, an increased or a decreased peak capacity could be favorable, which is analyzed in the following paragraphs.

4.5.4.1 Increasing peak capacity

(30)

be achieved by reductions in other periods. Due to the large share of grandfather rights slots (~95% according to SACN), this reduction could only take place gradually over a long time, when airlines give back slots. A gradual process of not granting a certain share of grandfather rights every year would also be an option, but would create immense resistance of airlines. While airlines operating transfer connections might benefit from this development, point-to-point airlines might neither benefit nor be interested in this change and would not voluntarily change their plans.

Additionally, a decreased ground time might not lead to a better utilization of gates. Currently, narrowbody gates (e.g. b- or c-pier) serve seven arrivals per day. Per peak, only one flight can be handled in order to ensure transfer connections to other flights. Thereby, only two flights are operated outside the five peaks. If the ground time of the five flights within the peaks can be decreased, it is at least questionable whether the gate will be used by additional flights between the peaks. As seen in the analyses, flights are mainly demanded during the peaks and a sharp increase of off-peak flights like in Frankfurt is not expected for Amsterdam by experts.

Furthermore, the required peak gate capacity is mainly determined by the number of the hub airlines’ flights for transfer connections. As nearly all aircrafts within a wave are on the ground at least for a short period, a decreased ground time would not change the total number of gates required, but only the time frame in which they are used. Also experts indicated that a change of the hub airlines’ schedule with fewer flights on the ground would have the largest effect on AAG levels.

4.5.2.2 Decreasing peak capacity

In contrast to the concept of Mirković & Tošić (2015) introduced in section 2.4, it could be aimed for a more even distribution of flights over the day. The current declared peak modes facilitate the existence of peaks with inbound and outbound waves leading to times with a high AAG level. Frankfurt airport operating with a flat declared capacity expierneces a less extreme peak pattern today, as traffic grew especially in off-peaks times. Therefore, the overall airport utilization might stay more constant over the day. Due to the existing peak modes in the capacity declaration, such a flattening is not possible at AAS, although traffic is growing fast.

Several experts indicated that a reduction of slots in the peak time would cause extreme problems for the hub airline, as its business model is based on the existence of transfer connections during peaks. In order to still ensure transfer connections, aircrafts would have to wait longer on the ground longer to allow these transfers. However, the increased ground and thereby transfer time lowers the attractiveness of the transfer connections for passengers. As a result, they might choose a faster connection via another hub airport.

Connecting runway and gate capacity for higher capacity utilization at hub airports

Master’s Thesis