On the Use of Hydrogen As the Future Aviation Fuel Aerospace Engineering

On the Use of Hydrogen as the Future Aviation Fuel

Martim Novais Cálão

Thesis to obtain the Master of Science Degree in Aerospace Engineering

Supervisors: Prof. Fernando José Parracho Lau Dr. Frederico José Prata Rente Reis Afonso

Examination Committee Chairperson: Prof. Filipe Szolnoky Ramos Pinto Cunha Supervisor: Prof. Fernando José Parracho Lau Member of the Committee: Prof. Afzal Suleman

December 2018 ii Acknowledgments

First and foremost, I would like to thank my supervisor at Airbus, Mr. Matthieu Meaux, for his guid- ance, expertise, trust and endless enthusiasm throughout the internship. I would also like to extend my gratitude to the Team-X for their experience and valuable advices, and the whole TPR Department of Airbus CTO in Toulouse for receiving me so well and making these six months such a pleasant experience. I would also like to address a big thank you to my supervisors at Instituto Superior Tecnico,´ Prof. Fernando Lau and Dr. Frederico Afonso for their constant availability and support in preparing this thesis, and to Prof. Inesˆ Esteves Ribeiro for her insights in Life Cycle Assessment. Last but not least, for the unconditional support of my family, friends and girlfriend, Mafalda, throughout my academic life I can only be grateful. Thank you!

iii iv Resumo

A presente dissertaçao˜ foi realizada durante um estagio´ de seis meses no CTO da Airbus e descreve o projeto conceptual de aeronaves regionais com propulsao˜ a hidrogenio,´ atraves´ de uma plataforma de Otimizaçao˜ Multidisciplinar. A implementaçao,˜ na referida plataforma, dos modelos de alguns compo- nentes da aeronave e´ detalhada, nomeadamente da unidade propulsiva, dos tanques de hidrogenio,´ e da respetiva integraçao˜ na fuselagem do aviao.˜ Sao˜ estudadas diferentes configuraçoes˜ de aeronave e alguns cenarios´ alternativos de modo a explorar um maior numero´ de poss´ıveis soluçoes˜ para o grande problema abordado: o impacto ambiental da aviaçao.˜ A performance das aeronaves a hidrogenio´ e´ com- parada com a da aeronave de referenciaˆ movida a querosene, atraves´ de indicadores como o consumo energetico´ e as emissoes˜ de poluentes. A avaliaçao˜ do ciclo de vida de cada tipo de combust´ıvel e uma discussao˜ preliminar relativa a` viabilidade economica´ da utilizaçao˜ de hidrogenio´ revelam a mudança de paradigma inerente a` transiçao˜ energetica.´ Os resultados deste estudo provam a viabilidade do projeto de aeronaves a hidrogenio´ e os seus benef´ıcios de um ponto de vista ambiental, desde que produzido a partir de energias renovaveis.´ O metano l´ıquido surge como uma alternativa muito interessante, na medida em que permite vislumbrar um cenario´ de transiçao˜ tendo em conta os atuais desafios economicos´ associados a` produçao˜ de hidrogenio.´

Palavras-chave: Otimizac¸ao˜ Multidisciplinar, Projeto de Aeronaves, Avaliac¸ao˜ do Ciclo de Vida, combust´ıveis alternativos, hidrogenio´

v vi Abstract

The present work was performed during a six-month internship at Airbus CTO and describes the conceptual design of hydrogen-fueled regional aircraft using a Multidisciplinary Design Optimization (MDO) tool. The implementation of the aircraft models inside the tool is detailed, notably with regard to the propulsion plant and the hydrogen tanks design and integration. Distinct hydrogen aircraft architectures are studied and alternative scenarios are modeled so as to explore a greater number of potential solutions for the main issue addressed: the environmental impact of aviation. The main aircraft performance ﬁgures to be compared with the kerosene-powered reference are energy consumption and pollutant emissions. The life cycle assessment of the different fuel types and a preliminary discussion on the economic viability of hydrogen as a fuel reveal the paradigm shift inherent to the energy transition. The results prove the feasibility of hydrogen aircraft from a design perspective and its beneﬁts from an environmental point of view as long as produced using renewable energies. Liquid methane appears as a very interesting candidate to enable a potential transition scenario given the current economic challenges related to the hydrogen production.

Keywords: Multidisciplinary Design Optimization, Overall Aircraft Design, Life Cycle Assess- ment, alternative fuels, hydrogen.

vii viii Contents

Acknowledgments...... iii Resumo...... v Abstract...... vii List of Tables...... xi List of Figures...... xiii Nomenclature...... xv Glossary...... xvii

1 Introduction 1 1.1 Motivation and Topic Overview...... 1 1.2 Background...... 2 1.3 Company Presentation...... 4 1.3.1 Airbus...... 5 1.3.2 Airbus CTO...... 5 1.3.3 Technology Planning and Roadmapping...... 5 1.3.4 Airbus Group Team-X...... 6 1.4 Objectives...... 6 1.5 Methodology...... 6 1.6 Thesis Outline...... 7

2 Hydrogen as an Aviation Fuel9 2.1 Hydrogen Properties...... 9 2.2 Safety...... 10 2.3 New Aircraft Concept...... 10 2.3.1 Propulsive Unit...... 11 2.3.2 Hydrogen Tanks...... 12 2.4 Environmental and Economic Impacts...... 14

3 Multidisciplinary Design Optimization 17 3.1 MDO Overview...... 17 3.2 XMDO Platform...... 19 3.2.1 Mission and Vehicle Description...... 19

ix 3.2.2 Operation Modes...... 21

4 Implementation 25 4.1 Aircraft Conﬁgurations...... 25 4.2 Mission Requirements...... 26 4.3 Physical Models...... 26 4.3.1 Turboshaft...... 27 4.3.2 Hydrogen Tanks...... 27 4.3.3 Fuselage...... 34 4.4 New XMDO Functionalities...... 35 4.5 Life Cycle Assessment...... 35 4.5.1 Goal and Scope Deﬁnition...... 36 4.5.2 Inventory Analysis (LCI)...... 37

5 Results 41 5.1 Problem Description...... 41 5.2 Solution...... 42 5.3 Alternative Scenario Exploration...... 43 5.3.1 Methane...... 44 5.3.2 Kerosene for the Reserves...... 45 5.4 Life Cycle Assessment...... 45 5.4.1 Impact Assessment (LCIA)...... 45 5.4.2 Interpretation...... 45 5.5 Fuel Pricing...... 46

6 Conclusions and Future Work 49 6.1 Conclusions...... 49 6.2 Future Work...... 50

Bibliography 51

x List of Tables

2.1 Liquid hydrogen and kerosene properties...... 10 2.2 Liquid hydrogen and synthetic kerosene combustion properties...... 11 2.3 Foam and MLI advantages and disadvantages...... 14

3.1 Advantages and disadvantages of the L-BFGS-B and CMA-ES algorithms...... 23

4.1 Top-level requirements...... 27 4.2 Cryogenic tank model input parameters...... 28 4.3 Tank wall materials properties...... 29 4.4 Empirical coefﬁcients for different MLIs...... 32 4.5 Inventory analysis for two kerosene stream types...... 38 4.6 Baseline GHG emissions from jet fuel production...... 38

4.7 Inventory analysis for two H2 production pathways...... 39 4.8 Hydrogen and kerosene combustion products...... 40 4.9 GWP values for a 100-year time horizon deﬁned in the IPCC Fifth Assessment report [31] 40

4.10 Global warming potential (in g CO2eq/MJ fuel) of kerosene and hydrogen...... 40

5.1 MDO results for all the aircraft conﬁgurations studied...... 43 5.2 Methane properties...... 44

xi xii List of Figures

1.1 In-ﬂight CO2 emissions forecast for international aviation, from 2005 to 2050...... 2 1.2 Airbus A350XWB’s ”sharklets”...... 3 1.3 First hydrogen-fueled aircraft (Tupolev Tu-155)...... 4 1.4 Team-X’s hierarchy branch...... 4 1.5 TPR valuation chain...... 5

2.1 Fuel leak simulation...... 11 2.2 Hydrogen tanks concept...... 13 2.3 Tank wall structures studied in this work...... 13 2.4 Rear and forward hydrogen tanks with a catwalk between the cockpit and the cabin.... 14

3.1 Aircraft design disciplines...... 17 3.2 Conventional vs. optimal design process...... 19 3.3 Twin engine aircraft deﬁned in XMDO...... 20 3.4 B-spline example...... 21 3.5 Vehicle and mission coupling with failure scenarios...... 21 3.6 Basic forces applied to an aircraft during ﬂight...... 22 3.7 Flight point calculation process...... 22

4.1 Aircraft configurations overview...... 26 4.2 Aircraft’s mission profile...... 27 4.3 Influence of the filling and venting pressures on the liquid volume fraction...... 28 4.4 Gravimetric index for different tank dimensions and insulation types...... 30 4.5 Schematic distribution of the temperature and thermal resistances...... 33 4.6 Gravimetric index in function of the time before venting for different insulation types and thickness...... 34 4.7 Hydrogen tanks integration and external fairing...... 35 4.8 LCA framework based on ISO 14040...... 36 4.9 System boundaries for the Life Cycle Assessment...... 37

5.1 Payload-range diagram of the kerosene- and the hydrogen-fueled aircraft conﬁgs..... 43 5.2 Assessment of the design mission GWP...... 46

xiii 5.3 Inﬂuence of the carbon pricing policies on the fuel price...... 47

xiv Nomenclature

Greek symbols

α Angle of attack.

Thermal emittance.

ηint Cryogenic tank integration factor.

γ Flight path angle.

µ Dynamic viscosity.

Ω Design space.

Ωfeasible Feasible design space.

φ Energy derivative.

ρ Density.

σ Steffan-Boltzmann constant.

υ Aluminum welding efﬁciency.

ε Engine setting angle (relative to the aircraft longitudinal axis).

Roman symbols

ˆc Vector of equality constraints of the MDO problem. c Vector of inequality constraints of the MDO problem. cp Speciﬁc heat at constant pressure. di Internal diameter of the cryogenic tank. f Objective function of the MDO problem.

F oS Safety factor to be applied to the cryogenic tank structure sizing. g Acceleration of gravity absolute value. h Heat transfer coefﬁcient.

xv hlg Heat of vaporization of hydrogen. k Thermal conductivity.

K Limited stress of the cryogenic tank internal wall material (aluminum alloy 2219).

L Cryogenic tank characteristic length. m Mass. m˙ out Cryogenic tank outlet ﬂow rate. nj Number of equality constraints of the MDO problem. nk Number of inequality constraints of the MDO problem.

Nu Nusselt number. p Pressure.

P r Prandtl number.

Q˙ Heat ﬂux through the cryogenic tank structure.

R Set of real numbers.

R Cryogenic tank structure thermal resistance.

Re Reynolds number.

S Wetted surface area.

T Temperature. t Thickness of the cryogenic tank walls. u Speciﬁc internal energy.

V Aircraft ground speed.

Vlayer Volume of a certain structural layer of the cryogenic tank.

Vtank Cryogenic tank internal volume. x Vector of design variables of the MDO problem. xlg Quality of the fuel (xlg=0 for saturated liquid and xlg=1 for saturated vapor).

Subscripts i, j, k Computational indexes. ref Reference condition. vent Cryogenic tank venting condition.

xvi Glossary

ACARE Advisory Council for Aeronautics Research in Eu- rope CASK Cost per Available Seat Kilometers CEO Chief Executive Ofﬁcer CMA-ES Covariance Matrix Adaptation Evolution Strategy CORSIA Carbon Offsetting and Reduction Scheme for Inter- national Aviation CTO Chief Technology Ofﬁce DAM Double Aluminized Mylar DSN Double Silk Net EADS European Aeronautic Defence and Space Company EEA European Economic Area EI Emission Index ETS Emissions Trading System EU European Union FL Field Length FoM Figures of Merit GHG Greenhouse Gas GMBM Global Market-Based Measure GWP Global Warming Potential HTP Horizontal Tailplane IATA International Air Transport Association ICAO International Civil Aviation Organization IPCC Intergovernmental Panel on Climate Change

xvii ISO International Organization for Standardization L-BFGS Limited-memory Broyden Fletcher Goldfarb Shanno LCA Life Cycle Assessment LCI Life Cycle Inventory LCIA Life Cycle Impact Assessment LNG Liquefied Natural Gas MDF Multidisciplinary Feasible MDO Multidisciplinary Design Optimization MLI Multilayer Insulation MTOW Maximum Take-Off Weight MZFW Maximum Zero Fuel Weight NASA National Aeronautics and Space Administration NIST National Institute of Standards and Technology NREL National Renewable Energy Laboratory R&T Research and Technology RPK Revenue Passenger Kilometers SAFe Scaled Agile Framework SE Specific Energy SFC Specific Fuel Consumption SMR Steam Methane Reforming STP Standard Temperature and Pressure TET Turbine Entry Temperature TPR Technology Planning and Roadmapping UHC Unburned Hydrocarbons USD United States Dollar VTP Vertical Tailplane

xviii Chapter 1

Introduction

The present thesis details the work developed during the author’s ﬁnal year internship at Airbus’ Chief Technology Ofﬁce (CTO), from April 9 to September 28, 2018, in Toulouse. As a pioneer in the aeronautical industry, Airbus wants to lead the way towards a greener future in aviation. Therefore, the aim of this internship was to conduct research on sustainable propulsion technologies at large, and hydrogen-fueled aircraft in particular.

1.1 Motivation and Topic Overview

Aircraft engines’ emissions contribute to the greenhouse effect and lead to a loss of local air quality, mainly in the vicinity of airports. By 2007, it was acknowledged that global aviation was responsible for 2% of anthropogenic CO2 emissions. This number seemed, however, doomed to rise, according to the air traffic forecast, that anticipated passenger and freight traffic annually growths of 4.9% and 5.8%, respectively, during the following twenty years (Airbus Global Market Forecast 2007, [1]). Thus, the aviation industry came together in 2009 under the lead of the ICAO (International Civil Aviation Organization), during the United Nations Climate Change Conference in Copenhagen and set itself three goals to minimize its environmental footprint: to improve fuel efficiency by 1.5% per year to 2020, to stabilize CO2 emissions through carbon-neutral growth from 2020 and to halve CO2 emissions by 2050 compared to 2005 [2]. Some ambitious goals addressing the aviation’s environmental impact were also set by the European Comission, that aims at reducing CO2 emissions by 75%, NOx emissions by 90% and noise pollution by 65% in 2050 compared to 2000 (Flightpath 2050, [3]). In order for these goals to be achieved, a paradigm shift towards sustainable alternative fuels is of paramount importance, which is further reinforced by the depletion of oil resources. Liquid hydrogen appears as an interesting candidate to replace kerosene as energy carrier, since it has a very high specific energy and a very low environmental impact (its combustion only releases water vapor and small amounts of NOx). Nevertheless, as this compound cannot be found naturally, when accounting for the whole product life cycle, its environmental impact may increase depending on its production method. Furthermore, liquid hydrogen has a low volumetric mass density and needs to be

1 stored at cryogenic temperatures (approx. 20 K), which poses serious challenges to transportation and storage, since large tanks with good insulation properties must be used. Adopting hydrogen as a fuel will not only lead to a considerable change in ground infrastructures but to different aircraft conﬁgurations as well, which will ultimately increase costs. All these issues being identiﬁed, in-depth investigation needs to be carried out and hydrogen’s viability has to be discussed.

1.2 Background

The impact of aviation on the climate change is not a recent issue but has been raising more and more awareness in the past few decades. In 1999, the IPCC (Intergovernmental Panel on Climate Change) published a report on the ”Aviation and the Global Atmosphere” [4] that analyses the impact of aircraft’s emissions on the atmospheric composition. According to this report, aviation contributes to global climate change approximately in proportion to its contribution to radiative forcing, which is a term that has been employed by the IPCC to denote ”an externally imposed perturbation in the radiative energy budget of the Earth’s climate system” [5]. Aviation was already responsible for 3.5% of the total radiative forcing in 1992 but according to [4] its share is expected to be 2.6 to 11 times higher than the 1992 value (0.05 W.m−2) by 2050. Aviation is reportedly one of the fastest-growing sources of GHG

(greenhouse gas) emissions, as ICAO forecasted that by 2050 in-ﬂight CO2 emissions could grow by up to 450% compared to 2005 (ﬁgure 1.1).

Figure 1.1: In-ﬂight CO2 emissions forecast for international aviation, from 2005 to 2050 (in [6])

Such a dark scenario gave birth to several activities aiming at mitigating the environmental impact of aviation, not only comprising technological improvements but also political agenda. In 2001, a report entitled ”European Aeronautics: A Vision for 2020” was published by a group of personalities invited by European Commissioner Philippe Busquin ”to agree on how aviation could better serve society’s needs”

2 [7]. This group of personalities established the ACARE (Advisory Council for Aeronautics Research in Europe) to help achieve the goals of Vision 2020, among which are the environmental targets of reducing

CO2 and NOx emissions by 50% and 80%, respectively [8]. Environmental goals regarding aviation’s

CO2 and NOx emissions were also set by the ICAO in 2009 and the European Comission in 2011 as stated in section 1.1. In 2008, the European Comission decided to include emissions from aviation in the European Union Emissions Trading System (EU ETS) starting from 2012. Airlines have thus to monitor and report emissions from flights within the EEA (European Economic Area) and to surrender allowances against those emissions [9]. More recently, in 2016, the ICAO agreed on a Resolution for a global market-based measure (GMBM) scheme for international aviation addressing CO2 emissions from 2021 on. Carbon Offsetting and Reduction Scheme for International Aviation (CORSIA) is the name of this global scheme that will require airlines to offset the growth of their CO2 emissions after 2020. The International Air Transport Association (IATA) also introduced a Carbon Offset Program that offers airlines’ customers the possibility to compensate for their flights’ environmental footprint by paying extra fees that are invested in carbon reduction projects. In order to accomplish the environmental goals set by the aviation industry, many improvements on aircraft technology have also been achieved throughout the years. The use of lighter materials on the airframe and the design of aircraft with enhanced aerodynamic properties result in less fuel burn and, consequently, in less pollutant emissions, by reducing the aircraft’s weight and drag, respectively. Ex- amples of such advances are the Airbus A350XWB and the Boeing 787, planes on which approximately half of the airframe is made of composite materials. Wingtip devices such as the A350XWB blended winglets (fig. 1.2) reduce lift-induced drag by increasing the effective wing-aspect ratio, which is a good example of improved aerodynamic design. In addition, aircraft engines are becoming more and more fuel-efficient. The Airbus A320neo family (new engine option), for instance, uses high-bypass turbofan engines, being approximately 15% more fuel-efficient than the original A320 family aircraft.

Figure 1.2: Airbus A350XWB’s ”sharklets”

Although these technological improvements help mitigating the impact of aviation on the climate change, they fall short of the ambitious goals set by the ICAO and the European Comission. Thus, the urge to ﬁnd an effective solution led to the exploration of alternative fuels. Kerosene is produced from crude oil and is currently used as the standard fuel in aviation. However, it can alternatively be synthetically produced from coal or natural gas, resulting in the so-called synthetic fuels that mitigate the dependence on crude oil resources rather than the environmental footprint of aircraft, as explained

3 in [10]. Biofuels, in turn, derive from biological feedstocks such as plants, and even if their combustion is still polluting, they are considered as sustainable alternatives since the CO2 released by the aircraft in the atmosphere is used by the plants in the process of photosynthesis, closing a carbon cycle. Hydrogen is even a greener alternative than biofuels since its combustion only produces water vapor and very small amounts of nitrogen oxides. The potential of liquid hydrogen for aircraft propulsion has been a matter of interest for some decades. Daniel Brewer published, in 1991, the book ”Hydrogen Aircraft Technology” [11] that summarizes the joint studies on hydrogen-fueled subsonic aircraft conducted by NASA and Lockheed Martin in the 1970s. However, the ﬁrst aircraft operating on liquid hydrogen was the russian

Tupolev Tu-155 that first flew in 1988, with one of its three engines running on LH2 (fig. 1.3). From 2000 to 2002 several European research partners (including Airbus) also studied the viability of using liquid hydrogen as the future aviation fuel under the Cryoplane Project framework. Having reinforced the interest on liquid hydrogen, these projects paved the way to more recent worldwide research activities on hydrogen, some of which will be used as bibliographical sources throughout this thesis.

Figure 1.3: First hydrogen-fueled aircraft (Tupolev Tu-155, in [12])

1.3 Company Presentation

The internship resulting in this work was carried out within the Team-X, in the Technology Planning and Roadmapping department (TPR) of the Airbus’ Chief Technology Office (CTO). As Airbus is such a large and stratified group, a brief description of each level of the Team-X’s hierarchy branch (fig. 1.4) will be provided to give a better perception of its role within the company.

Figure 1.4: Team-X’s hierarchy branch

4 1.3.1 Airbus

Airbus is a European aerospace company operating in the commercial aircraft, helicopters, defence and space sectors. The group was established in 2000 under the name European Aeronautic Defence and Space Company (EADS) as a result of the merger of a French, a German and a Spanish company: Aerospatiale-Matra,´ DaimlerChrysler Aerospace AG and Construcciones Aeronauticas´ SA , respectively. EADS was renamed Airbus Group in 2014 and ﬁnally merged with its own dominant division Airbus SAS in 2017 adopting its present name Airbus. The group’s CEO is Thomas Enders since 2012. The company has three divisions: Airbus Commercial Aircraft, Airbus Defence and Space and Airbus Helicopters. The former generates the largest share of the company’s total revenue and forms the commercial aircraft duopoly together with its American competitor Boeing. Airbus Defence and Space has strong capabilities in military aircraft, unmanned aerial systems, intelligence/security and space technologies. Finally, Airbus Helicopters is the world’s largest helicopter manufacturer providing both civil and military helicopter solutions.

1.3.2 Airbus CTO

Airbus CTO is a Company-wide function responsible for guiding all Research and Technology (R&T) activities of the group, including budgets and roadmaps. It ensures the company’s product and service portfolio is competitive in the medium- and long-term, and drives technology integration within the group. Major social issues such as the environment, mobility, defence, safety and security are addressed by the R&T activities that look forward to propose cutting-edge solutions.

1.3.3 Technology Planning and Roadmapping

The TPR department manages R&T investments by establishing technology roadmaps in the context of current and potential future Airbus products and services. The budget for the R&T projects and demonstrators can thus be allocated aiming at maximizing the value added to the company.

Figure 1.5: TPR valuation chain

Figure 1.5 illustrates the valuation chain of an R&T investment scenario. The level of maturity of a

5 given technology depends on the degree of investment on its development and affects its impact when integrated in the company’s concerned products. The performances of each product with the enhanced technologies can thus be analyzed through conceptual studies and quantiﬁed through Figures of Merit (FoM). Finally, analyzing how each product would perform in the corresponding market allows a valuation of the investment scenario. The role of the TPR department is to ﬁgure out the technologies that have to be developed by predicting the investment scenario that maximizes the added value.

1.3.4 Airbus Group Team-X

Team-X’s ultimate mission is then, in light of the TPR department’s role, to perform R&T portfolio optimization. In order to do so, this team has the ability to run quick (but high quality) concept and mission studies and to elaborate speciﬁc business cases for quantitative valuation of the R&T investments. The targeted headcount is only eight members, making the team very dynamic given its challenging mission. There was also a high level of interaction with other teams not only within the CTO but also from other Airbus’ functions. The work environment was thus very enriching and collaborative. My contribution to the team consisted mainly of performing concept studies regarding greener aircraft solutions.

1.4 Objectives

From a purely academic point of view, this work is the result of a six-month internship at Airbus with which the author plans to obtain his Aerospace Engineering Degree. Following the research axes of the company, the internship’s aim is to study the potential of liquid hydrogen as an aviation fuel. Bearing in mind this main purpose, some more speciﬁc objectives may be derived:

• Discuss the main advantages and disadvantages of using liquid hydrogen for propulsion and the changes it would lead to in aviation;

• Establish a fair comparison between hydrogen and standard kerosene conﬁgurations, both in terms of aircraft design and performance;

• Assess the environmental impact and carry out a preliminary discussion on the economical viability of hydrogen-fueled aircraft.

1.5 Methodology

An applied research methodology was followed since it focused on solving the particular issue of the harmful impact of aviation. This is the most common type of research conducted within a company, addressing practical problems rather than aiming at the enhancement of knowledge. In order to attain the aforementioned goals, concept studies were carried out using a Multidisciplinary Design Optimization (MDO) approach thanks to a platform called XMDO developed at Airbus by Matthieu

6 Meaux. My work consisted both of further developing XMDO and applying it to the studies. This tool played a major role in testing, validating and comparing different aircraft conﬁgurations. A Life Cycle Assessment (LCA) methodology was also applied in this work in order to quantify the potential environmental impact of the different fuel options throughout their entire life cycle, with special focus on the global warming potential. During the internship, the work was developed under the Scaled Agile Framework (SAFe) which is a set of workﬂow patterns intended to implement Lean-Agile practices in a company. The SAFe methodology aims at value delivery in the shortest sustainable lead time [13]. Therefore, it is based on an iterative development principle that involves the division of a project into several short duration phases (usually two weeks) at the end of which the goals previously set must have been attained and the results presented. This requires an ability to plan and prioritize work without ever losing sight of the long-term objectives.

1.6 Thesis Outline

This thesis is composed of six main chapters whose contents are brieﬂy described hereafter. Chapter 1: Introduction The motivation, objectives and methodology behind this work are explained. A brief overview on the thesis’ subject background is provided and Airbus is presented. Chapter 2: Hydrogen as an Aviation Fuel Summarizes the most important hydrogen properties and discusses the main changes on aviation resulting from the energy transition. Chapter 3: Multidisciplinary Design Optimization General introduction to MDO and presentation of the platform XMDO used to perform Overall Aircraft Design within this project’s framework. Chapter 4: Implementation Describes the aircraft architectures studied, the physical models implemented on the XMDO platform and the approach used to assess the overall fuel’s life cycle. Chapter 5: Results Presentation and discussion of the results obtained. Chapter 6: Conclusion and Future Work Derives the main conclusions from the work developed and maps out future work suggestions.

7 8 Chapter 2

Hydrogen as an Aviation Fuel

Hydrogen seems a very interesting energy carrier to replace the standard kerosene fuel from an environmental point of view. Several studies on this compound have already been published and their promising results show that further work is worth being developed. Adopting hydrogen as an aircraft fuel would imply a paradigm shift due to its properties fundamentally different from those of kerosene. This section essentially based on literature review contains information on the hydrogen properties and presents an overview of the changes it would lead to in aviation.

2.1 Hydrogen Properties

Some hydrogen properties and the corresponding kerosene values are summarized in table 2.1 to allow a comparison between the two fuel options. The hydrogen property that makes it such a promising fuel alternative is obviously its speciﬁc energy (energy per unit mass) which is about three times higher than that of kerosene. On the other hand, the hydrogen state at STP conditions (Standard Temperature and Pressure - 0◦C and 1 bar) is gaseous and its density is 0.0899 kg/m3. Such a low value makes liquid hydrogen the only viable option for aviation. Nevertheless, not only liquid hydrogen has still a very low density when compared to kerosene but it must be stored in cryogenic conditions (approximately 20K) as well. If we compare the fuel masses of same energy content we realize liquid hydrogen is almost three times lighter than kerosene but occupies a volume four times bigger. Therefore, the installation of large hydrogen tanks with special insulation properties on the aircraft presents a serious challenge. This topic will be discussed in detail later on.

Despite its simple structure, molecular hydrogen can exist in two different isomers according to its nuclear spin: ortho-H2 with spin unpaired, or para-H2 with spin paired. The equilibrium composition of the two forms is highly dependent on the temperature. At the boiling temperature, the equilibrium composition is 0.2 vol% of ortho-H2 and 99.8 vol% of para-H2.

9 Table 2.1: Liquid hydrogen and kerosene properties

Property Unit Kerosene Hydrogen Boiling temperature ◦C 150-300 -252.87 Melting temperature ◦C -40 -259.14 Speciﬁc energy MJ/kg 42.8 122.8 Density kg/m3 775-840 70.8 * * In liquid state

2.2 Safety

The safety aspects of using hydrogen as a fuel must be clearly discussed not only because they will be a key factor in certifying this technology but also because public acceptance is crucial in order for the transition to hydrogen to happen. A detailed safety analysis was conducted in the framework of the Cryoplane study developed by Airbus Deutschland GmbH and its final technical report [14] states that ”there is no fundamental problem, which would prevent the successful operation of a commercial aircraft running on liquid hydrogen”. Nevertheless, safety measures must obviously be adopted because hydrogen gas is very flammable and yields explosive mixtures with air and oxygen. During flight, hydrogen burns at concentrations considerably below the limits for detonation which is rather reassuring. The hydrogen tanks should nevertheless be designed to prevent leakages and should not be located in impact areas such as the rotor disk burst cones. Structural strengthening may also be needed depending on the tanks location and dimensions. Hydrogen accumulation on ground infrastructures is a potential risk linked to this technology as it may lead to massive explosions. Hydrogen leakage should then be minimized by improving the tanks’ insulation properties but cannot be completely avoided. A further solution to prevent hydrogen accumulation should be proposed. Figure 2.1 shows a fuel leakage simulation conducted at University of Miami and proves a clear safety advantage of hydrogen compared to kerosene since the gasoline vehicle was severely damaged while the hydrogen vehicle was undamaged. The images show that hydrogen rises fast in the atmosphere causing a standing flame that is quickly extinguished instead of detonating. This suggests that a good safety measure to prevent catastrophic events on the ground would be to burn the hydrogen whenever it is leaked. This should however be managed in a way that would not raise acceptance issues among aircraft’s passengers.

2.3 New Aircraft Concept

As previously mentioned, using hydrogen as a fuel implies some changes regarding aircraft design and lead to new concepts whose performances should be assessed. For now, we shall focus on identifying in which aspects a hydrogen-fueled plane would differ from a standard kerosene vehicle. The propulsive unit and the fuel tanks were the components identiﬁed as critical and subject to change the most.

10 Figure 2.1: Fuel leak simulation (in [15])

2.3.1 Propulsive Unit

As suggested by Vestraete, 2009 [16] hydrogen is a very attractive fuel alternative because it allows a very stable combustion over a wide range of operating conditions. In this study, a comparison between synthetic kerosene (synjet) and hydrogen’s combustion characteristics is made. Hydrogen’s ﬂammability limits are wider, the ﬂame velocity is higher and the minimum ignition energy is lower comparatively to synjet (table 2.2).

Table 2.2: Liquid hydrogen and synthetic kerosene combustion properties (values extracted from [16])

Combustion property Unit Synjet Hydrogen Flammability limits in air vol% 0.8-6.0 4.0-75.0 Minimum ignition energy in air mJ 0.25 0.02 Laminar burning velocity in air cm/s 43 265

Storing the hydrogen at cryogenic temperatures requires the fuel storage and supply systems to have good insulation properties. In addition, a heat exchanger has to be installed in the engine to heat hydrogen from the tank temperature to injection conditions as explained in [17]. The heat exchanger vaporizes the hydrogen heating it up to temperatures between 150 K and 250 K. Minimizing hardware changes should nevertheless be kept in mind when adding this component. The comparison between a hydrogen- and a kerosene-fueled engine running at the same net thrust and using the same fuel injection temperature (250 K) shows that the TET (Turbine Entry Temperature) is about 40 K lower when burning

11 hydrogen which may lead to an important increase in turbine life [17]. In addition, the SFC (Speciﬁc Fuel Consumption) of the hydrogen-fueled engine is nearly three times lower. Previous studies in Airbus concerning the Cryoplane showed this SFC reduction can be estimated as follows:

SEkerosene SFCLH2 = SFCref × (2.1) SELH2 where SE stands for the speciﬁc energy per unit mass of fuel. Reducing the fuel injection temperature would allow a further reduction of the TET but implies a slight increase of the speciﬁc fuel consumption. Fuel injection temperature was, for instance, decided to be kept the same as that of kerosene in the Green Freighter Project due to contradictory trends of previous studies regarding this parameter [10].

2.3.2 Hydrogen Tanks

The hydrogen tanks design and integration in the aircraft is certainly the most challenging task of this study and on which a considerable part of the internship was spent. Liquid hydrogen as an energy carrier is only used for space applications for the time being and the information on the fuel supply system technology applied to aviation comes down to very few scientiﬁc publications. The available literature on the topic was however crucial to develop a physical model of the tanks needed to assess the performance of the aircraft through an MDO approach. The present section only presents an introduction on the topic while the details on the developed model are explained later (section4). a. Design

As the hydrogen is stored at cryogenic temperatures, both the mechanical and the thermal designs are very important when modelling the fuel tank. Figure 2.2 helps illustrating the behaviour of a liquid hydrogen tank. A filling valve is used to inject the hydrogen into the tank at a given filling pressure. The difference of temperatures between the inside and the outside of the tank generates a certain heat input that causes the hydrogen pressure to rise during flight and holding-phases on the ground. If the maximum allowable pressure (venting pressure) inside the tank is reached, venting the hydrogen is required to keep or decrease the pressure level [18]. Therefore, an amount of gaseous hydrogen is needed in the case of venting. The choice of the vessel’s shape and materials is of crucial importance since the overall system should be as light as possible while keeping good mechanical and thermal properties. A cylindrical shape with hemispherical end caps is adopted because, according to [18], it minimizes the tank wall surface-to-volume ratio making it lighter. From a mechanical point of view, the tank wall should prevent hydrogen permeation and its material’s properties should include a high strength, stiffness and fracture toughness. In further accordance with [18], the tanks can be classified as integral or non-integral from a load perspective. Integral tanks are a ”structurally-integrated part of the airframe and must be capable of withstanding loads to which the supporting structure is exposed” contrarily to non-integral tanks. Moreover, the vessels must have good insulation properties in order to reduce the heat input and retard

12 Figure 2.2: Hydrogen tanks concept the venting of hydrogen, minimizing the waste of fuel. In this work’s scope, non-integral double-walled hydrogen tanks were chosen and two different tank wall structures were modeled and compared. These structures can be observed in ﬁgure 2.3.

(a) Structure of a foam-based insulation (in [18]) (b) Structure of an MLI-based insulation

Figure 2.3: Tank wall structures studied in this work

The ﬁrst corresponds to the one proposed in [18] that consists of a foam-based insulation with an in- ner aluminum alloy wall, an external epoxy composite fairing and two MAAMF vapour barriers. MAAMF is a multilayer sandwich composed by aluminum foils, Mylar layers and dacron or glass net fabric. In this structure, the insulation material (Rohacell foam) is still combined with a vacuum jacket to prevent air from permeating into the foam once it could condensate or even solidify at such low temperatures working as a thermal bridge. The second structure is a double aluminum alloy wall structure with a multilayer insulation (MLI). Vacuum is created between the two aluminum walls to avoid thermal conduction and the MLI is used to reduce the radiative heat transfer in the vacuum. The selected aluminum alloy was the 2219 series whose properties were proven to be superior than other materials’ by Brewer [11]. Both insulation technologies are well established and their advantages and disadvantages are summarized in table 2.3 in accordance to Verstraete [16]. b. Installation

The installation of large hydrogen vessels in the aircraft is a difficult task that needs special attention. Several possible aircraft configurations regarding the tanks location are identified in the Cryoplane final technical report [14] and will be briefly summarized. First of all, the simplest solution would be to install a single tank behind the aft pressure bulkhead. This could however pose some issues regarding the center of gravity of the whole system that would move backwards due to the fuel supply system’s weight and would constantly change during flight as

13 Table 2.3: Foam and MLI advantages and disadvantages according to [16]

Advantages Disadvantages Low cost Relatively high thermal conductivity Foam Easy to implement Low resistance to thermal radiation Light weight and low density Potential damage from environmental hazards Very low thermal conductivity High vacuum required Very low radiation heat transfer Heavy tank walls required MLI Extremely low density Costly to implement and maintain Near catastrophic failure upon loss of vacuum the fuel would progressively be consumed. This problem could be avoided if a second hydrogen tank would be installed in the forward part of the plane, between the cockpit and the cabin. Although it helps ﬁxing the center of gravity issue, this solution also has a problem: the forward tank would disconnect the cockpit and the cabin. A catwalk could be installed like illustrated in ﬁg. 2.4 but would change the tank’s cylindrical shape. Further detailed studies on the tank shape should be developed to assess the structural impact of the catwalk. It is also important to mention that installing the tanks in the fuselage would either lead to a reduction of the number of passengers or to a considerable stretching of the aircraft.

Figure 2.4: Rear and forward hydrogen tanks with a catwalk between the cockpit and the cabin (in [14])

Another possible solution is to install the tanks over the fuselage resulting in a shape similar to the Beluga that creates a weight and drag penalty. This solution does not create center of gravity or cabin problems like the previous ones. Special attention has however to be paid to the rotor disk burst that can damage the tanks leading to a catastrophic event. A different alternative is still suggested in the Green Freigher Project ﬁnal results [19] that consists of installing external hydrogen tanks under the wings.

2.4 Environmental and Economic Impacts

To conclude about the environmental impact of hydrogen as an alternative fuel, its whole life cycle needs to be analyzed. Hydrogen’s production, transportation and storage on the ground can be even more important than its combustion during ﬂight when it comes to assessing its overall impact.

14 As previously mentioned in chapter1, hydrogen combustion only releases water vapor and very small amounts of NOx. However, although the hydrogen combustion is carbon-free, its current production processes are not. According to [20] 99% of hydrogen is currently produced through fossil fuel reforming since it is the most economic pathway. Even though fossil fuel reforming generates CO2, if during the whole life cycle carbon dioxide is only released in specific point sources (hydrogen production sites), carbon capture and storage may be an interesting solution since it prevents CO2 from being released into the atmosphere. Capturing carbon increases, however, the hydrogen production cost. Hydrogen can alternatively be obtained through the electrolysis of water and, in order to eliminate carbon dioxide emissions, this process can be powered by renewable energy sources. This option is currently expensive but the Hydrogen Council [20] expects its costs to decrease by 50% with increasing application (from 700-850 USD/kW to 450-550 USD/kW by 2050). The National Renewable Energy Laboratory (NREL) [21] still mentions biogas reforming as another possible hydrogen production option. After being produced, hydrogen still has to be liquefied, transported and stored and there are also several options regarding these operations that must be studied. Hydrogen can first be liquefied and then transported or the other way round. Hydrogen transport options include truck/ship transport both for gaseous and liquid hydrogen and pipelines for gaseous hydrogen. In terms of storage, the most common options include cryogenic liquid or compressed gas vessels, or even metal hydride storage that consists of storing hydrogen by chemically bonding it to metal or metalloid elements and alloys. It can then be concluded that in order for this energy transition to happen, considerable investments regarding the deployment of hydrogen infrastructures on the ground are necessary. These investments must be carefully planned through a rigorous assessment of the economical viability of the different pathways along with the analysis of their contribution to the environmental footprint. Besides the costs related to the hydrogen operations on the ground, the hydrogen aircraft architectures and their maintenance will also be more costly than the kerosene-powered aircraft as suggested by [19]. A trade-off between emissions and costs will then allow us to understand how much should we be willing to pay to become greener.

15 16 Chapter 3

Multidisciplinary Design Optimization

This section consists of a general introduction to the Multidisciplinary Design Optimization (MDO) methodology and to the XMDO platform which was used to carry out the concept studies under the internship’s scope of work.

3.1 MDO Overview

MDO is an engineering method that aims at automating the design process of a certain multidisciplinary system through optimization techniques. The largest number of applications have been in the ﬁeld of aerospace engineering because it studies highly multidisciplinary and complex systems. Figure 3.1 presents the disciplines involved in aircraft design and illustrates how an airplane could look like in order to achieve the best possible performance from each discipline point of view.

Figure 3.1: Aircraft design disciplines (in [22])

MDO is still not widely used in industry because an aircraft design usually involves several different teams specialized in a single discipline and many human decisions in the loop. The interaction between

17 all the teams can however be very difficult to manage and time-consuming. In this traditional way of working, partial optimization is commonly performed often neglecting the interdependence between the different disciplines and an iterative procedure is needed to try to reach convergence between the work developed independently in each team. This sequential optimization does not guarantee the finding of the true optimum of the system. MDO can thus help improving the convergence towards an overall optimized system by coupling several disciplines even though it increases the complexity of a problem. According to [22], ”optimization is the process of choosing the design variables that yield an optimum design” respecting certain equality and/or inequality constraints. The most challenging part of the design process is normally the definition of the optimization problem under the form:

minimize f(x)

n with respect to x ∈ Ω, Ω ⊂ R (3.1) subject to cˆj(x) = 0, j = 1, 2, ..., nj

ck(x) ≥ 0, k = 1, 2, ..., nk where:

• f is the objective function that enables a quantitative assessment of the system design (problem output);

• x is the vector of design variables which are the parameters that are allowed to vary during the design process (problem inputs);

• Ω is the design space deﬁned by the bounds set for the design variables;

• cˆ and c are the vectors of equality and inequality constraints, respectively.

We can still define the feasible design space as Ωfeasible = {x ∈ Ω|cˆ(x) = 0, c(x) ≥ 0} and the output space as {f(Ω), cˆ(Ω), c(Ω)}. Choosing the models that relate the constraints and the objective to the design variables is a very important part of the designer’s job. The main differences between the conventional and the optimal design processes can be observed in figure 3.2. While in the conventional process the performance of the system is evaluated until reaching satisfying results, in an optimal approach it is the objective and the constraints that are evaluated aiming at the optimal design. As MDO is an automated method that uses optimization algorithms to perform systems design, it makes it easier and faster to explore more product-related scenarios and more complex solutions. That is the main reason why MDO is used to develop concept studies within Team-X. Introducing cutting-edge technologies in the conventional design process loop may be difficult given the lack of knowledge both on the technology itself and on the way it interacts with other sub-systems of a certain product. On the other hand, preliminary physical models concerning a technology can first be developed and integrated on the MDO process and then be improved as the related knowledge expands.

18 Figure 3.2: Conventional (left) vs. optimal (right) design process (in [22])

3.2 XMDO Platform

The XMDO platform is a valuable asset for Team-X that enables the development of rapid (but high quality) concept studies integrating disruptive technologies. It is an MDO tool programmed in Python under the object-oriented paradigm and it has been being developed by Matthieu Meaux since 2011. XMDO performs both trajectory and vehicle optimization by coupling the vehicle design to the mission description and it can operate either in mono- or multi-mission mode. The result of a sizing loop using XMDO is then a vehicle with optimal performance either for a single optimal mission or a set of optimal missions. Saying the vehicle and the mission description are coupled means its average performance is only optimal if the optimal trajectory is followed for each of the deﬁned missions.

3.2.1 Mission and Vehicle Description

In order to describe the mission(s) and the vehicle, a set of data is used as an input for the physical models integrating the platform. The input data is divided into three categories:

• Design variables, that can vary within a speciﬁed range during an optimization;

• Variables, that have a speciﬁed value and cannot vary during an optimization;

• Parameters, that are deﬁned as a function of other input variables or parameters.

The number of design variables is the dimension of the optimization problem and deﬁnes its complexity. XMDO is composed of several parametric models that couple the input data to the vehicle performance throughout the ﬂight.

19 a. Vehicle

The description of the vehicle is achieved by decomposing it into different ”components” that represent its sub-systems. Some sub-systems can still be decomposed in their own sub-systems. The power plant, for instance, is a component of the aircraft that can still be decomposed in its own components such as the turbofan, the nacelle, the fuel supply system, etc. Such a vehicle description allows to easily build different configurations by simply assembling the different vehicle parts. It is therefore possible to quickly explore and trade different vehicle configurations. An example of a twin engine aircraft defined in XMDO is shown in fig 3.3.

Figure 3.3: Twin engine aircraft deﬁned in XMDO

Components are modeled by discipline meaning that for each component exist several models, one for each concerned discipline. There is then a dual view of the vehicle, per components and per disciplines. The disciplines modeled in the platform are: geometry, mass, aerodynamics, propulsion, energy, thermal management and costs.

b. Mission

Unlike some conventional concept studies where only some sizing points are analyzed (take-off, top of climb, cruise, approach), the entire mission is evaluated in the XMDO platform. Just as a vehicle is decomposed in several components, a mission is decomposed in several segments that correspond to the different flight phases. The segments composing a mission are represented by B-splines with a set of control points describing altitude and distance. The whole trajectory is then constructed as a linear combination of splines that are piece-wise polynomial functions. A visual example of a B-spline can be observed in figure 3.4. Top level requirements concerning the mission can either be defined as input variables (range, cruise speed, etc.) or as constraints (take-off field length, approach speed, etc.). The overall mission profile can be optimized by setting values like the speed, the altitude and the distance at different control points

20 Figure 3.4: B-spline example (in [23]) as design variables. As previously mentioned, the vehicle and its trajectory are coupled and its performance can be optimized for a set of different missions, making it more versatile which can be interesting from an airline’s point of view. This coupling is illustrated in figure 3.5 that consists of a schematic view of the platform’s working principle. The vehicle’s performance figures throughout each mission or even during specific failure cases are accounted in the objective and constraints’ functions that are evaluated at each iteration of the sizing loop until the optimal set of design variables is reached, in accordance to the optimal design process outlined in figure 3.2.

Figure 3.5: Vehicle and mission coupling with failure scenarios

3.2.2 Operation Modes

The platform can either run in a simple performance evaluation or in the optimization mode. Both these modes are described hereafter.

a. Performance Evaluation

The evaluation of the aircraft’s performance figures during flight is done by progressively solving the longitudinal trajectory equations. These equations can be derived from figure 3.6 that illustrates the

21 basic forces applied to an aircraft at a given ﬂight point. The forces’ application points are not taken into account as the moments generated would add another level of complexity to the problem.

Figure 3.6: Basic forces applied to an aircraft during ﬂight

According to the Newton’s second law one can write the longitudinal trajectory equations:

  m.ax = T hrust.cos(α + ε) − Drag − m.g.sin(γ)

 m.ay = T hrust.sin(α + ε) + Lift − m.g.cos(γ)

where m is the mass of the vehicle, ax and ay are the accelerations in the flight path direction and its normal, respectively, V the ground speed, α the angle of attack, ε is the engine angle relatively to the aircraft longitudinal axis, and γ is the flight path angle. The process used to solve each flight point is described in figure 3.7. After setting the flight point data, the geometry and the mass properties of the vehicle are computed. The angle of attack and the rating (power setting defined by the pilot) at that flight point are then calculated by numerically solving the trajectory equations using the Newton-Raphson method. Finally the energy and thermal management disciplines are analyzed.

Figure 3.7: Flight point calculation process

22 After computing an entire mission, a post-processing analysis including the cost discipline is conducted. The vehicle’s performance ﬁgures throughout the mission and the associated costs are presented as an output. b. Optimization

The optimization mode makes use of the performance evaluation capabilities of the platform to compute the objective function and the constraints at each iteration. Among the available optimization algorithms, only the L-BFGS-B [24] and the CMA-ES [25] provide satisfying results. While other algorithms try to manage the problem’s constraints before performing the optimization, these two are not able to handle constraints. Therefore, an exterior penalty function is added to the objective function, imposing a penalty for violating the constraints as explained by [22]. A penalty parameter of 2 was considered in this formulation. The vehicle performance figures and the problem constraints are then aggregated into one single objective function. This formulation enables the exploration of the entire design space rather than only the feasible design space. The advantages and disadvantages of each algorithm are summarized in table 3.1. The L-BFGS-B algorithm was predominantly used during this work because it is faster and best suited for large dimension problems. The projected gradient tolerance was set at 1×10−8 and the precision goal is defined by a factor of 10.0 multiplied by the machine floating-point precision (2×10−16), which represents an extremely high level of accuracy. Furthermore, a Multidisciplinary Feasible (MDF) architecture was used to solve the MDO problem, meaning that the optimizer always returns a vehicle design that satisfies the consistency constraints [22].

Table 3.1: Advantages and disadvantages of the L-BFGS-B and CMA-ES algorithms

Algorithm L-BFGS-B CMA-ES Gradient-based Gradient-free Type Quasi-Newton method Evolutionary Uses a limited amount of computer memory Wide exploration Advantages Faster Result is a global optimum Explores less Slower Disadvantages Result is a local optimum that depends on the initial point

23 24 Chapter 4

Implementation

The study of the potential of hydrogen as a fuel using the XMDO platform required the implementation of some physical models and functionalities that had not yet been developed. In this chapter, the aircraft configurations that were studied in this internship’s framework will first be outlined to understand the need for new physical models. The aircraft’s mission profile will then be described and finally the new XMDO models and functionalities developed will be explained.

4.1 Aircraft Conﬁgurations

In order to study hydrogen as an aviation fuel, a kerosene-fueled aircraft was chosen as a reference from which the hydrogen-fueled vehicle conﬁgurations will be derived and that will enable a comparison between the two fuel types. The reference plane is a regional twin turboprop aircraft that is illustrated in ﬁgure 4.1(a).

Following the discussion in section 2.3.2 b., two aircraft configurations with different hydrogen tank locations were studied and are presented in figures 4.1(b) and 4.1(c). The first one is a double tank configuration with both hydrogen vessels being installed over the fuselage and will be referred to as our baseline solution. The fact that this configuration does not imply major issues regarding the center of gravity and that the cabin properties rest unchanged is very attractive and was the main driver to choose it as our baseline. The length of both tanks can vary according to our storage needs and to the center of gravity location. Figure 4.1(b) is just an example in which the forward tank is the smallest. The second aircraft configuration is an architecture with a single rear hydrogen tank. This configuration is expected to pose center of gravity problems but its choice was mostly due to the fact that it results in an aircraft shape more similar to the standard shape we are used to, which is very interesting from a public acceptance point of view.

The exact dimensions of each aircraft will be an output of the optimization process whose results are detailed in section5.

25 (a) Kerosene-fueled reference aircraft

(b) Double hydrogen tank conﬁguration (c) Single hydrogen tank conﬁguration

Figure 4.1: Aircraft conﬁgurations overview

4.2 Mission Requirements

The aircraft’s mission profile including the reserves phase is outlined in figure 4.2 and its top-level requirements are specified in table 4.1. It is important to stress that the reserves requirements include both a 45-minute continued cruise phase (for possible holding periods) and a 100NM range extension to fly to an alternate airport for some unforeseen circumstances such as extreme weather conditions, terrorist activity, crash, etc. The optimization was launched in multi-mission mode due to the two mission ranges specified in table 4.1. A take-off engine failure case was added as well since it may be sizing for the engines along with the take-off field-length constraint. The mission segments defined in XMDO both for the block time and for the reserves phase can be derived from figure 4.2:

  Taxi in     Continued cruise  Take-off       Climb  Climb   Block time Reserves phase Cruise Cruise    Descent    Descent    Landing    Landing   Taxi-out

4.3 Physical Models

As previously mentioned some new physical models had to be implemented in XMDO in order to define and size the hydrogen-fueled aircraft configurations. These models concern aircraft components that were already part of the kerosene-fueled reference but were modified (fuselage and turboshaft) and components that did not exist at all in the reference aircraft (hydrogen tanks). Therefore, some models

26 Figure 4.2: Aircraft’s mission proﬁle

Table 4.1: Top-level requirements

Requirement Unit Value Max range (at maximum number of passengers) Mission Range * NM 30% of max range (at maximum payload) Cruise Mach number - 0.5 Cruise altitude ft 25000 Take-off/landing field length (FL) m < 1400 Approach speed kt < 110 * The maximum range and number of passengers cannot be specified for confidentiality reasons were just updated from the previous kerosene version while others were built from scratch. The models of each component were coded per discipline as explained in section 3.2.1. The new fuselage model was just used for the baseline double tank configuration while the new turboshaft and hydrogen tank models were applied to both hydrogen-fueled aircraft configurations.

4.3.1 Turboshaft

The kerosene-powered turboshaft model was updated in order to enable the use of hydrogen as its fuel. This model uses the instant power/thrust command (rating) to compute the shaft output power and the fuel mass flow rate at each time step. The only modification that was made to this model concerned the calculation of the specific fuel consumption. The SFC of the hydrogen-powered turboshaft is computed from that of the kerosene-powered using the simple relation previously introduced (eq. 2.1).

4.3.2 Hydrogen Tanks

The implementation of a model concerning a cryogenic tank was one of the most important and long tasks of the present work given the MDO approach used to tackle the subject. It was thus essential to have a parametric model in order to deﬁne the sizing parameters as the problem’s design variables

27 to be changed during the optimization loop. This model was built based on the academic publications available that were previously introduced in section 2.3.2.

This model allows us to compute the tank dry mass, the tank capacity (maximum liquid hydrogen mass) and the time before venting due to the pressure rise inside the vessel during holding periods on the ground. The six input parameters can be divided into three categories and are presented in table 4.2:

Table 4.2: Cryogenic tank model input parameters

Unit Parameters m Diameter Dimensions m Length - Material (Foam/MLI) Insulation properties m Thickness bar Filling pressure Characteristic pressures bar Venting pressure

As previously mentioned, the vessel must contain a certain amount of gaseous hydrogen in case venting hydrogen is necessary. Therefore, the amount of liquid hydrogen inside the tank is limited and that is why the tank capacity should be carefully computed. In order to do so, the maximum liquid volume fraction of the container has to be determined and, according to Verstraete 2009 [16], it is a function of the tank sizing pressures as it can be observed in ﬁgure 4.3. The data from the curves plotted in this ﬁgure was subjected to a polynomial regression and a bivariate polynomial function was formulated enabling the computation of the liquid volume fraction using the input pressures. Knowing the liquid fraction and the tank dimensions one can easily calculate the liquid hydrogen mass that we refer to as the tank capacity.

Figure 4.3: Inﬂuence of the ﬁlling and venting pressures on the liquid volume fraction (in [16])

28 a. Mechanical Design

The tank wall structure is chosen between the two options studied (figure 2.3) according to the insulation type specified in the input parameters of the model. Once the structure is defined, the thickness of each wall layer has to be specified in order to compute the tank wall mass as follows:

X mtank wall = ρi.Vlayeri (4.1) i

where ρi represents the volumetric mass density of each material and Vlayeri the volume of the respective layer. For both wall structures the insulation thickness is speciﬁed as an input parameter while the thickness of the internal aluminum wall depends on the maximum allowable pressure, being calculated according to the following relation:

pvent.di ti = (4.2) υ(2K/F oS − pvent) where pvent is the venting pressure, di the internal diameter, υ the weld efficiency, K the limited stress and F oS the safety factor. Since the hydrogen tanks are classified as critical structures due to the high pressure levels, a safety factor of 2 has to be considered. Information on the aluminum properties needed to calculate the internal wall thickness is provided in table 4.3. It is still important to mention that the wall thickness had to be limited to a minimum allowable value for manufacturing reasons. Finally the MAAMF vapor barriers and the external fairing of the foam-insulated structure have a fixed thickness defined by Winnefeld et al. [18] (1.524 × 10−5 m and 1.57 × 10−2 m, respectively). The fairing has a considerable thickness in order to protect the insulation from external damage. For the exact same reason, the aluminum external wall in the MLI-insulated structure is computed by multiplying the internal wall thickness by an additional safety factor. Table 4.3 summarizes the materials’ properties used to compute the tank wall mass.

Table 4.3: Tank wall materials properties (values extracted from [18])

Material Property Unit Value Volumetric density kg/m3 2825 Aluminum alloy 2219 Limited stress MPa 172.4 Weld efﬁciency - 0.8 Rohacell foam Volumetric density kg/m3 35.24 MAAMF vapor barrier Surface density kg/m2 0.225 Epoxy composite fairing Surface density kg/m2 1.304

The tank dry mass is ﬁnally calculated as the sum of the wall mass and the auxiliary systems mass (piping and valves). Using the tank’s capacity and its dry mass, we can deﬁne the tank gravimetric index that we want to maximize:

m grav. index (%) = LH2 × 100 (4.3) mdry tank + mLH2

29 Figure 4.4 is a plot from the tank model that shows the gravimetric index as a function of the normalized liquid hydrogen mass for different tank dimensions and insulation types. For each curve plotted, the diameter is kept constant and the length of the tank is changed to increase its capacity. First of all, we can easily conclude that for a given tank capacity the MLI insulation results in lower gravimetric indexes than the foam insulation, due to the need for a heavy aluminum external wall. It is also remarkable that low capacity tanks have very low grav. indexes (left-hand side of the curves). Finally, we conclude that for the same insulation type and hydrogen mass, reducing the diameter (and increasing the length) leads to higher grav. indexes.

Figure 4.4: Gravimetric index for different tank dimensions and insulation types

b. Thermal Design

The last output of the tank model is the time before venting which is the amount of time the pressure inside a filled tank takes to rise from the filling to the maximum allowable level causing the venting valve to open and the gaseous hydrogen to start being released in the atmosphere. The thermal behaviour of the tank has to be modelled in order to estimate this pressure change once it depends on the heat transfer through the tank wall as previously detailed. In this thermal model, the properties of para-hydrogen will be used because, according to [16], it allows a more conservative design due to the slightly greater pressure fluctuations. Moreover, hydrogen is considered to be in saturation state, with both phases in thermodynamic equilibirum. Lin et al. 1991 [26], introduce a model assuming an homogeneous mixture of liquid and gaseous hydrogen inside the tank, even if in reality there exists a certain degree of stratification. Using the first law of thermodynamics and the conservation of mass the pressure change is given by:

" #−1 dp φ h ∗ i ∂u = Q˙ − m˙ out.hlg. (xlg + ρ ) with φ = ρ. (4.4) dt Vtank ∂p ρ

30 where φ is the energy derivative that represents the pressure rise per volume per energy input, Vtank the volume of the tank, Q˙ the heat flux, m˙ out the outlet flow rate, hlg the heat of vaporization at the tank pressure, xlg the quality of the fuel (xlg=0 for saturated liquid and xlg=1 for saturated vapor) and ∗ ρ = ρg/(ρl − ρg). Since the homogeneity assumption under-predicts the pressure change, the tank model of this study uses twice the pressure change rate given by equation 4.4, as suggested by Lin et dp φ.Q˙ al. [26]. When the outlet valve is closed, m˙ out = 0 and the previous expression simplifies dt = V .A function of the pressure and density was formulated for the energy derivative using the thermodynamic data from NIST [27]. To calculate the heat that flows between the atmosphere and the hydrogen inside the tank, different thermal mechanisms were considered and the corresponding heat transfer coefficients were calculated according to [16].

External heat transfer

Two different mechanisms take place in the heat transfer between the atmosphere and the external aircraft surface: convection and radiation. The convective heat transfer coefﬁcient writes:

Nu.k h = air (4.5) conv ext L where Nu is the Nusselt number, kair the air thermal conductivity and L the characteristic length (length of the cylindrical part of the tank). The correlation for convection on a ﬂat plate is used to compute the Nusselt number [16]

0.43 0.8 µair.cp ρair.V.L Nu = 0.03625.P r ReL with P r = and ReL = (4.6) kair kair where P r and ReL are the Prandtl and the Reynolds numbers, respectively, µair is the dynamic viscosity of air, cp the specific heat at constant pressure, ρair the air density and V the flight speed. The flight speed is zero on the ground but a non-zero value was taken in this case to account for the natural convection. On its turn, the radiative heat transfer coefficient is calculated as follows:

2 2 hrad = σ..(Tskin + Tatm).(Tskin + Tatm) (4.7) with σ representing the Stefan-Boltzmann constant, the emittance of the skin, and Tskin and Tatm the temperatures of the aircraft skin and the atmosphere, respectively. The total external heat transfer coefﬁcient is simply given by the sum of both contributions:

hout = hconv ext + hrad (4.8)

Since there is a gap between the aircraft fairing and the hydrogen tanks there is still thermal conduction in the air outside the tank. The conductive coefﬁcient is given by hair = kair/tair, with tair representing the thickness of the gap.

Tank wall heat transfer

31 The heat transfer mechanisms through the tank wall depend on the choice of the insulation material. While in the foam-insulated tank the heat transfer happens through thermal conduction, in the MLI it happens predominantly through radiation since there is vacuum between the tank walls. As previously mentioned, the conductive heat transfer coefﬁcient is given by h = k/t, with k representing the thermal conductivity of the material and t its thickness. The Rohacell foam possesses a thermal conductivity between 5 × 10−3 and 35 × 10−3 W/mK, that rises with increasing temperatures. For the heat transfer through the MLI, Vestraete 2009 [16] gives a relation to compute the heat ﬂux in mW/m2 based on some thermal properties reported in literature:

4.67 4.67 Q˙ (T + T ).(T − T ) T − T p n n = C N C2 . H C H C + C . H C + C T g − T g (4.9) S 1 t 2(N + 1) 3 N g N H C where C1, C2 and C3 are coefﬁcients empirically determined for each MLI type (table 4.4), Cg is a 4 coefﬁcient that depends on the interstitial gas in the insulation (1.46×10 for GN2), ng an exponent depending on the interstitial gas (0.53 for GN2), p the pressure representing the vacuum level and TH and TC the temperatures of the hot and cold boundaries of the insulation, respectively. N is the total number of insulation layers and Nt the layer density of the MLI (in layers/cm). Finally, for both tank wall structures, the heat is transferred by thermal conduction both in the external and internal tank walls.

Table 4.4: Empirical coefﬁcients for different MLIs (values extracted from [16])

MLI C1 C2 C3 DAM / Dacron 8.95 × 10−8 2.56 5.39 × 10−10 DAM / Tissuglas 4.43 × 10−11 3.91 8.03 × 10−10 DAM / DSN 2.11 × 10−9 3.56 5.39 × 10−10 Perforated DAM / DSN 2.98 × 10−8 2.84 5.86 × 10−10

Internal heat transfer

An internal natural convection phenomenon occurs between the internal tank wall and the hydrogen inside the tank. The corresponding heat transfer coefﬁcient is calculated using hint = Nu.kH2 /L.

Nu represents the Nusselt number (a constant value was adopted) and kH2 represents the thermal conductivity of the hydrogen.

These heat transfer coefﬁcients and the wetted surface area S of the respective material layers allow us to calculate the successive thermal resistances R between the hydrogen and the atmosphere using the simple relation:

1 R = P (4.10) hi.Si i Finally, the heat ﬂux is computed by solving the following system of equations that assumes one-

32 dimensional heat transfer:

    TH2   −1 1 0 0 0   R   in    TC     0 −1 1 0 0   R       ins ˙   .  TH  −   .Q = 0 (4.11)  0 0 −1 1 0    R       ex    Tskin   0 0 0 −1 1   Rout Tatm

Figure 4.5 presents a schematic view of the problem, identifying the material layers that are accounted in each thermal resistance.

Figure 4.5: Schematic distribution of the temperature and thermal resistances

The temperature of the atmosphere was fixed at a certain value and the temperature of the hydrogen is exclusively function of its pressure. Therefore, the unknowns of the problem are the heat flux (Q˙ ), the temperature of the aircraft skin (Tskin) and the temperature of the hot and cold boundaries of the insulation (TH and TC respectively). The Newton-Raphson method already programmed in the XMDO platform was used to solve this system of equations at each time step until the venting pressure level was reached. The gravimetric index was plotted in function of the normalized time before venting for different insulation types and thickness and the results can be observed in figure 4.6.

It becomes clear that despite the lower gravimetric indexes the MLI is much more effective than the foam insulation since it can hold for longer periods on the ground without releasing hydrogen. We can also conclude that increasing the insulation thickness leads to longer holding periods before venting with no major drawbacks in terms of mass. Another interesting solution to increase the time before venting is to widen the gap between the ﬁlling and the venting pressures, but always keeping in mind that higher venting pressures require thicker and heavier tank walls.

After reaching the venting pressure, we use the equation 4.4 once again, but this time to compute the boil-off rate, i.e., the instant hydrogen mass ﬂow rate m˙ out that is required to keep or decrease the dp pressure level inside the tank ( dt ≤ 0).

33 Figure 4.6: Gravimetric index in function of the time before venting for different insulation types and thickness

4.3.3 Fuselage

The single rear tank configuration’s fuselage is longer but holds the same shape of the reference aircraft. On the other hand, the integration of two hydrogen tanks over the cabin leads to a slightly different fuselage shape in our baseline solution. The integration of the tanks is achieved thanks to an external fairing illustrated in figure 4.7 that leads to mass and drag penalties. A new fuselage model was thus derived from the standard one in order to compute the new fuselage wetted area and mass depending on the hydrogen tanks dimensions. Even if in reality they are detached, the external fairing is assumed to be tangent both to the fuselage and the hydrogen tanks (tangency points A and B, respectively) in the new fuselage model. The length of the fairing cross-section (red contour in figure 4.7) is then estimated through basic trigonometry principles and used to calculate the fairing surface in the constant cross-section zone of the aircraft’s fuselage. An additional fairing surface is estimated for the forward and rear zones of the fuselage by connecting the tangency points to the fuselage ends. The new fuselage wetted surface area can finally be computed by adding the fairing surface to the old fuselage area that is not covered by this external fairing. The wetted area is then used to compute the friction drag within the aerodynamics discipline of the platform. The mass of the new fuselage is derived from the previous one by considering not only the fairing mass but also an integration factor to account for the brackets and local structural strengthening:

mfus = mref .(1 + ηint) + Sfairing.ρAfairing (4.12)

where mref represents the reference fuselage mass, ηint the integration factor, Sfairing and ρAfairing the fairing surface and surface density, respectively. The mass penalty is simply estimated in the new fuse-

34 Figure 4.7: Hydrogen tanks integration and external fairing lage model. A detailed structural analysis should be carried out if a rigorous calculation was required.

4.4 New XMDO Functionalities

If part of the work consisted of implementing physical models concerning specific components of the aircraft, another part involved the development of some new functionalities in the XMDO platform. First of all, the main motivation of this work being to explore sustainable alternatives to kerosene, a routine to calculate the harmful emissions on the ground and throughout the flight was implemented in the post-mission analysis module of the platform. This is made by multiplying the fuel burned during flight by the respective emissions indexes introduced in the next section 4.5. Another feature implemented concerned the single rear tank configuration. The propulsion and energy disciplines of the platform were related in a way that did not allow the use of a single fuel tank to feed more than one propulsive units. This capability was thus added to the platform by rethinking the way the energy source was linked to the propulsion train. Finally, it was developed the capability of running multi-fuel aircraft configurations on XMDO. As it will later be explained in section5, following the results of the first optimizations, some interesting alternative solutions to our hydrogen configurations were explored, one of which involving an aircraft powered by multiple energy sources. Therefore, as it was not yet implemented, a multi-fuel mode was developed allowing a vehicle to carry and use different fuel options.

4.5 Life Cycle Assessment

LCA is a technique to assess the potential environmental impacts throughout a product’s life cycle from raw material acquisition through production, use, end-of-life treatment, recycling and ﬁnal disposal [28]. This ”cradle-to-grave” analysis is an extremely powerful tool, since the complete assessment of a product’s environmental performance can provide valuable support both to industrial activities and to public policy making. As deﬁned by the ISO 14040 standard (International Organization for Standard- ization), an LCA study is a systematic approach that consists of four phases:

35 1. Goal and scope definition - defining the purpose of the study enables the definition of its scope, including the system boundaries, the functional unit and the level of detail;

2. Life cycle inventory analysis (LCI) - collection of input/output data regarding the product’s life cycle, such as energy and raw material requirements, emissions and wastes;

3. Life cycle impact assessment (LCIA) - assessment of the environmental signiﬁcance of the LCI;

4. Interpretation - the results of the LCI and LCIA are summarized and discussed to draw the main conclusions and recommendations.

The LCA framework according to ISO 14040 and its direct applications are schematized in ﬁgure 4.8.

Figure 4.8: LCA framework based on ISO 14040 [28]

In order to be assessed, the environmental impact of a product’s life cycle is divided into different categories. Global warming potential, land use, stratospheric ozone depletion, human toxicity, ecotoxic- ity, resource scarcity, acidiﬁcation, eutrophication, photochemical ozone formation and ionizing radiation are some of the most common ones. This particular LCA analysis will focus on the assessment of the global warming potential of aircraft fuels. Global warming potential (GWP) is a relative measure of the contribution of a certain GHG to the radiative forcing and, consequently, to the temperature increase at the earth’s surface. It compares the amount of energy the emissions of a certain mass of a GHG absorb over a given period of time, to the amount of energy absorbed by a similar mass of CO2 emissions. The larger the GWP, the greater the contribution of the gas to the earth’s warming compared to CO2 over the speciﬁed time period. The most commonly used time period is 100 years and the unit indicator for GWP is kg CO2eq (equivalent carbon dioxide).

4.5.1 Goal and Scope Deﬁnition

The purpose of this analysis is to quantify and compare the global warming potential of kerosene and liquid hydrogen used as aircraft fuels. Two different hydrogen production pathways will be looked at, namely steam methane reforming (SMR) and electrolysis of water powered by wind energy. The LCA

36 will be carried out using data available in the literature and the functional unit of the study is revenue passenger kilometers (RPK), as used in [2], meaning that it will be assessed the impact of transporting a paying passenger over a distance of one kilometer. The system boundaries include the fuel production and its combustion during ﬂight. They are outlined in ﬁgure 4.9 along with the respective inputs and outputs that are going to be considered in this LCA analysis.

Figure 4.9: System boundaries for the Life Cycle Assessment

4.5.2 Inventory Analysis (LCI)

Kerosene production

Jet fuel is one of the major products from processing crude oil in a reﬁnery. It is typically a blend of different kerosene streams: straight-run kerosene directly obtained from the atmospheric distillation process and hydroprocessed kerosene that results from hydrotreating and/or hydrocracking of primary streams. Producing the former generates fewer harmful emissions than the latter, but the two cases will be presented to provide the low and high emissions scenario. Since kerosene is not the only product from the reﬁning process, an allocation criteria must be chosen to assign life cycle emissions between the co-products. The data used in the present work was derived from [29] that adopts an energy allocation criteria, assuming that emissions are allocated by the energy content of the different fuels. The GHG emissions reported in [29] are expressed in carbon dioxide equivalent assuming the global warming potential (GWP) values from the IPCC Fourth Assessment Report [30]. These GWP values were used to express the absolute values of each gas emissions in g per kg of kerosene in order to compute the GWP using the Fifth Assessment Report [31] values as it will later be explained. The inventory data regarding the two kerosene stream types is presented in table 4.5. According to [29], the baseline GHG emissions scenario can be obtained by considering an average of the two processing techniques. The corresponding values are presented in table 4.6.

Hydrogen production

Two hydrogen production pathways are going to be assessed, as previously explained. The ﬁrst and most common method of producing hydrogen is steam methane reforming, that consists in reacting steam with methane in the presence of a nickel catalyst at high temperatures. The water-gas shift reaction can then be applied to obtain additional hydrogen. These two chemical reactions write as follows:

37 Table 4.5: Inventory analysis for two kerosene stream types

Unit Straight-run Hydroprocessed Process fuel Electricity J/MJ kero. 0.423 9.137 Natural gas J/MJ kero. 5.772 82.683 Reﬁnery gas J/MJ kero. 9.669 30.713 Coke J/MJ kero. 3.548 11.294 Residual oil J/MJ kero. 0.643 2.067 Process energy consumption J/MJ kero. 20.055 135.894 GHG emissions

Carbon dioxide (CO2) g/kg kero. 299.6 980.1

Methane (CH4) g/kg kero. 0.86 22.26

Nitrous oxide (N2O) g/kg kero. 0.014 0.014

Table 4.6: Baseline GHG emissions from jet fuel production

GHG Unit Baseline emissions

Carbon dioxide (CO2) g/kg kero. 509.3

Methane (CH4) g/kg kero. 3.94

Nitrous oxide (N2O) g/kg kero. 0.014

CH4 + H2O CO + 3H2 (4.13)

CO + H2O CO2 + H2 (4.14)

The second production process is the electrolysis of water, that consists in decomposing water into oxygen and hydrogen by passing an electric current through it. This yields an oxidation-reduction reaction, with hydrogen being produced at the cathode. Wind energy will be used to power this production process in order to make it as green as possible.

− − Cathode (reduction): 2H2O + 2e → H2 + 2OH (4.15)

− − Anode (oxidation): 2OH → 1/2O2 + H2O + 2e (4.16)

The inventory analysis of these two hydrogen production alternatives, including resources and energy consumption, emissions and solid wastes is summarized in table 4.7. This data was reported by the National Renewable Energy Laboratory [32, 33] and is referenced by several authors. The energy consumed to process natural gas into hydrogen (SMR pathway) was computed from [32] by subtracting the energy contained in the natural gas feedstock and the energy consumed by all upstream processes required to operate the hydrogen plant from the total energy consumed within the system. The process energy consumption in the renewable hydrogen production concerns the electrolysis share of the total energy consumption reported by [33]. It can be observed, from the inventory data, that hydrogen

38 production is more energy consuming than kerosene production regardless of the chosen pathway. Re- newable hydrogen production is particularly energy consuming, which is one of the reasons that makes SMR economically more attractive. The feedstock energy-consumption is however very high in the SMR process [32] due to the amount of natural gas fed into the system. Despite generating slightly more solid wastes, the electrolysis pathway releases far less harmful emissions into the atmosphere than SMR.

Table 4.7: Inventory analysis for two H2 production pathways

Unit SMR Wind electrolysis Resources

Coal g/kg H2 159.2 214.7

Iron (Fe, ore) g/kg H2 10.3 212.2

Iron scrap g/kg H2 11.2 174.2

Limestone g/kg H2 16.0 366.6

Natural gas g/kg H2 3642.3 16.2

Oil g/kg H2 16.4 48.3

Water l/kg H2 19.8 26.7 Emissions

Benzene (C6H6) g/kg H2 1.4 -

Carbon dioxide (CO2) g/kg H2 10620.6 950.0

Carbon monoxide (CO) g/kg H2 5.7 0.9

Methane (CH4) g/kg H2 59.8 0.3

Nitrogen oxides (NOx) g/kg H2 12.3 4.7

Nitrous oxide (N2O) g/kg H2 0.04 0.05

Non-methane hydrocarbons (NMHCs) g/kg H2 16.8 4.4

Particulates g/kg H2 2.0 28.7

Sulfur oxides g/kg H2 9.5 6.1 Solid Waste

Miscellaneous non-hazardous waste g/kg H2 201.6 223

Process energy consumption J/MJ H2 480.5 3557.0 Reference [32][33]

Fuel combustion during ﬂight

The substances produced during the combustion of kerosene and hydrogen during ﬂight are summarized in table 4.8 along with their respective emission indices (EI) per kg of fuel. These EI are extracted from [34] and they are highly dependent both on the engine technological level and on the speciﬁc fuel used. It is nonetheless clear that the hydrogen combustion produces fewer harmful substances. Com- pared to the energy content of 1kg of kerosene, an energy-equivalent amount of hydrogen produces only 3.24kg of water vapor and about 1.5g of nitrogen oxides. However, the large quantities of water vapor released by the hydrogen combustion may be concerning due to the possible formation of contrails (depending on the altitude) that have been proven to contribute to climate change [35].

Overall GWP

39 Table 4.8: Hydrogen and kerosene combustion products (values extracted from [34])

Unit Kerosene Hydrogen

Carbon dioxide (CO2) g/kg fuel 3150 -

Water vapor (H2O) g/kg fuel 1250 9000

Nitrogen oxides (NOx) g/kg fuel 14 4.3

Sulfur oxides (SOx) g/kg fuel 1 - Carbon monoxide (CO) g/kg fuel 3.7 - Unburned hydrocarbons (UHC) g/kg fuel 1.3 - Soot g/kg fuel 0.04 -

As previously detailed, this LCA analysis is intended to study the environmental impact of each fuel type from a global warming potential perspective. The GWP of the whole fuel life cycle is a combination of the CO2, CH4 and N2O emissions expressed in equivalent CO2. The GWP values of these compounds for a 100-year time horizon are deﬁned in the IPCC Fifth Assessment Report [31] and presented in table 4.9. Table 4.9: GWP values for a 100-year time horizon deﬁned in the IPCC Fifth Assessment Report [31]

GHG GWP value for a 100-year time horizon

Carbon dioxide (CO2) 1

Methane (CH4) 28

Nitrous oxide (N2O) 265

Multiplying the emissions of each of these compounds listed in tables 4.6, 4.7 and 4.8 by the respective GWP value and adding the three contributions results in the overall GWP for each fuel type and production pathway. These GWP values are presented in table 4.10 in g CO2 eq. per MJ of the respective fuel type to enable the comparison of the environmental impact per energy content.

Table 4.10: Global warming potential (in g CO2eq/MJ fuel) of kerosene and hydrogen

GWP (g CO2 eq./MJ fuel) Production Combustion Total Kerosene 14.5 73.6 88.1 Hydrogen (via SMR) 100.2 - 100.2 (via wind electrolysis) 7.9 - 7.9

In spite of being generally perceived as a green fuel, hydrogen can have a greater environmental impact than kerosene. Consuming 1 MJ of hydrogen produced via steam methane reforming results in a global warming potential higher than burning the mass of kerosene with the same energy content. A very low GWP can however be reached in case hydrogen is obtained through the electrolysis of water using wind power. The GWP values of each fuel and production pathway and the aircraft energy consumption ﬁgures will be used to assess the overall environmental impact of performing the design mission. The results of this analysis are presented in the following chapter.

40 Chapter 5

Results

This chapter ﬁrst details the optimization problem to be solved using XMDO and then presents the corresponding results. Alternative scenarios derived from our hydrogen baseline solutions are ﬁnally explored and compared.

5.1 Problem Description

In order to size the hydrogen aircraft configurations presented in chapter4 using an MDO approach, the optimization problem must be defined under the previously specified form (eq. 3.1). The objective chosen to be minimized was the Maximum Take-Off Weight (MTOW) which is the maximum weight at which the pilot is allowed to attempt to take-off. The MTOW is an indirect measure of the cost per available seat kilometres (CASK). The CASK is a commonly used measure of unit cost in the airline industry that is determined by dividing the operating costs by the passenger carrying capacity (number of seats available multiplied by number of kilometres flown). The design variables of the problem can be divided in two types: those concerning the mission profile and those concerning the aircraft design. Most of the design variables are common to both aircraft configurations. The only difference is that, at the aircraft level, in the baseline configuration the length of the two tanks is fixed and the diameter is a design variable, while in the single tank configuration the diameter is fixed and the length is a design variable. The aircraft fuselage’s wetted surface area and weight will be minimized by the optimizer by reducing the diameter of the tanks in the baseline configuration and the length of the tank in the rear tank configuration. In the baseline configuration the length of the two hydrogen tanks was chosen to be the same and it is constrained by the blade burst cones. Following the results of the previous section, the MLI-insulated structure was chosen over the foam-insulated due to the longer times before venting that minimize the waste of fuel. The other design variables concerning the aircraft dimensions are the MTOW, the Maximum Zero Fuel Weight (MZFW), the carried fuel mass, the wing span and root chord, and the diameter of the main components of the propulsive train (turboshaft, gearbox and propeller). At the mission level, the design variables are the altitude, the longitudinal position and the speed at the control points of the splines used to define the

41 aircraft trajectory. Several inequality constraints were deﬁned. The ﬁeld length and the approach speed were set to respect the missions requirements (table 4.1). The aircraft take-off and zero fuel weights were constrained to be lower than the MTOW and the MZFW, respectively. The rating is limited by the maximum available power. Finally, the fuel tanks capacity must be higher than the fuel needed to complete the mission. No equality constraints were added to the problem.

5.2 Solution

The results of the optimizations regarding the hydrogen configurations using XMDO are summarized in the first two columns of table 5.1. The remaining columns contain results regarding alternative scenarios explored in the following section and are presented in the same table to allow an overall comparison between the configurations studied. These are relative values expressed in percentage with respect to the kerosene reference aircraft. The energy consumption and the pollutant emissions concern the design range at maximum payload without considering the reserves as these are only used in exceptional circumstances. Only the CO2 and the NOx emissions are displayed since these are the compounds that raise more concern within the aviation industry (given their climate impact), being specifically addressed by the ICAO and the European Comission’s goals (section 1.1). It can be observed that despite being heavier and consuming more energy than the kerosene-fueled architectures, the hydrogen configurations drastically reduce the harmful emissions throughout the flight, as expected. According to [14] similar results regarding the energy consumption of hydrogen-fueled aircraft were obtained in the Cryoplane project as ”Due to the bigger wetted surface the energy consumption would increase by 9% to 14%”. The rear tank configuration presents a higher MTOW due the heavier fuel supply system when compared to our baseline. However, due to the integration of the tank inside the fuselage and the absence of an external fairing, the fuselage wetted area is smaller, resulting in a lower friction drag and consequently in a lower overall energy consumption. Figure 5.1 shows the normalized payload-range diagrams of the reference and the hydrogen-fueled aircraft. The hydrogen aircraft is capable of flying a slightly longer range than the reference aircraft at maximum payload. On the other hand, at maximum fuel, the range achievable by the H2 plane is considerably shorter than that of the kerosene plane. For ranges lower than the range at maximum fuel, the hydrogen payload-range diagram runs very flat, meaning that increasing the range does not bring considerable losses in terms of payload. However, when the maximum fuel capacity is reached, the payload drops drastically for small range increases. It can therefore be concluded that the hydrogen aircraft is optimized to fly short ranges carrying great payloads. This result regarding the payload range diagram is in agreement with the conclusions from the green freighter project [10]. To sum up, despite the positive results regarding the environmental impact of the fuel combustion, using hydrogen requires very large and heavy fuel supply systems and fuselages that finally result in higher energy consumptions, making it costly to fly these greener aircraft. Alternative solutions aiming at reducing the fuel supply system dimensions and the overall energy consumption will be looked at in

42 Figure 5.1: Payload-range diagram of the kerosene- and the hydrogen-fueled aircraft conﬁgs. the following section.

Table 5.1: MDO results for all the aircraft conﬁgurations studied. Relative deviation (in %) with respect to the kerosene reference aircraft ﬁgures

Fuel type Hydrogen Methane Hydro. + Kero. A/C conﬁg. Baseline Rear tank Baseline Rear tank Baseline MTOW +7.5% +7.8% +5.3% +5.6% +8.6% Energy cons.* +12.9% +10.4% +9.4% +7.8% +11.7% CO -100% -100% -26.6% -27.7% -100% Emissions* 2 NOx -87.6% -87.9% -86.8% -87.0% -87.8% Mass +30.7% +18.7% +22.4% +11.1% +28.2% Fuselage Wetted area +21.3% +15.7% +9.5% +8.8% +16.9% Qty. 2 1 2 1 2 Mass [-] 6.7 17.0 2.4 9.7 5.0 Fuel tank** Length [-] 1.90 1.02 1.90 0.57 1.90 Diameter [-] 0.45 0.98 0.26 0.98 0.38 * For the mission at the design range and maximum payload (without using reserve fuel). ** These values concern one single fuel tank. The mass is normalized by the mass of the kerosene fuel system and the dimensions are normalized by the reference fuselage diameter.

5.3 Alternative Scenario Exploration

Two alternative scenarios were explored under the same MDO approach in order to widen the scope of the study and present different solutions always aiming at reducing the environmental impact of aviation.

43 5.3.1 Methane

The use of liquid methane instead of hydrogen was the ﬁrst alternative that seemed potentially interesting. Methane’s chemical formula is CH4 and some of its properties are summarized in table 5.2. It can be seen that although methane is less energetic, it is much denser than hydrogen and it is still more energetic than the standard jet fuel. Despite being higher than those of hydrogen, the boiling and melting temperatures are still very low and thus, special insulated vessels are also required.

Table 5.2: Methane properties

Property Unit Methane Boiling temperature ◦C -161.5 Melting temperature ◦C -182.5 Speciﬁc energy MJ/kg 55.6 Density kg/m3 422.6 * * In liquid state

CH4 does not seem to be an alternative solution as green as hydrogen since it still emits large quantities of carbon dioxide when burned. On the other hand, it is very cheap even when compared to kerosene, which is a very positive point in favor of this alternative fuel. Furthermore, methane is an abundant element on Earth because besides being naturally produced through the digestion and decay of biological organisms, it is also a byproduct of several industries. Capturing and using it as a fuel would then decrease the amount of methane in the atmosphere that signiﬁcantly contributes to the green house effect. Nonetheless, in order to enable a fair comparison between all the modeled scenarios, methane production’s global warming potential will be computed. Actually, the GWP of methane production can be derived from the hydrogen inventory analysis (section 4.5) since methane is naturally the main feedstock used by the SMR technique. According to [32], natural gas production and transport are responsible for 25% of the hydrogen production GWP via SMR. Knowing the natural gas consumption required to produce 1 kg of hydrogen (4.7), it can be calculated that 0.85 kg CO2eq are released during the production of 1kg of natural gas. Given that 94.5% of the natural gas used in the referenced analysis consists of methane, this value is a fair estimation of the methane production GWP.As for its combustion, according to [36], burning 1kg of natural gas produces around 2.8kg of CO2, 2.2g of NOx and very small amounts of other compounds such as carbon monoxide and sulfur dioxide. As carbon dioxide is the only

GHG released by methane combustion, the GWP throughout methane’s entire life cycle is 65.6 g CO2eq per MJ of methane used. In order to analyze the potential of methane as a fuel using the XMDO platform, the physical models concerning the hydrogen configurations that were detailed in the previous chapter were reused. The single difference were the properties of the fuel, impacting the specific fuel consumption of the turboshaft and the dimensions of the fuel tanks, in particular. Both the baseline and the single tank configurations powered by liquid methane were sized and its results are also summarized in table 5.1.

It is clear that the CH4 tanks are smaller than the hydrogen vessels, resulting in a smaller and lighter fuselage. The MTOW and the energy consumption are therefore reduced when compared to

44 the H2 conﬁgurations but remain above the reference values. As expected, the CH4 is less ecological reducing carbon dioxide emissions only by 25% when compared to kerosene. Methane is nevertheless as effective as hydrogen when it comes to ﬁghting nitrogen oxides.

5.3.2 Kerosene for the Reserves

The second scenario that could also lead to a reduction of the cryogenic tanks size and, consequently, to a reduction of the mass and drag penalties consists of using hydrogen as the main fuel for the block mission while keeping kerosene for the reserve segments. As the kerosene is stored inside the wings it does not have a significant impact on the fuel supply system mass comparatively to hydrogen. On the other hand, carrying less hydrogen means having smaller vessels and a lighter fuselage. If we look exclusively at the environmental impact throughout the block mission, this solution is, evidently, as friendly as the full H2 aircraft. However, looking at the whole mission including the reserves naturally shows this is a worse alternative from an environmental point of view. In order to size this alternative aircraft, a multi-fuel feature was implemented on XMDO as it was previously detailed in section 4.4. Only the baseline configuration was sized using this solution and the results are presented in table 5.1 as well. This configuration presents the highest MTOW among all the alternatives because besides the cryogenic tanks there is a considerable amount of kerosene stored in the aircraft wings. Nevertheless, if compared with the hydrogen baseline architecture it can be observed that reducing the hydrogen tanks size results in a considerable reduction of the fuselage wetted surface area and, consequently, the overall energy consumption is slightly lower. Moreover, as the fuel reserve is only used for exceptional circumstances, this alternative can be interesting from an environmental point of view.

5.4 Life Cycle Assessment

5.4.1 Impact Assessment (LCIA)

Knowing the aircraft consumption and the global warming potential per kg of each fuel type, the overall environmental impact of performing a given mission can now be assessed. The GWP per passenger- kilometer was then computed for the design mission at maximum payload and the different fuel alternatives. Figure 5.2 presents the results of the GWP assessment in relative deviations with respect to the kerosene reference. These values concern the double tank aircraft conﬁguration and the block fuel consumption. The rear tank conﬁguration results are not displayed since they are very close to the baseline solution (+25.6%, -90.1%, -19.7% for the H2 via SMR, H2 via electrolysis and CH4 scenarios, respectively).

5.4.2 Interpretation

By simply looking at the emissions levels during ﬂight (table 5.1), it seemed that, with the exception of the methane-fueled aircraft, the modeled scenarios could easily help achieving the CO2 reduction targets

45 Figure 5.2: Assessment of the design mission GWP. Relative deviation (in %) with respect to the kerosene reference aircraft figures set by the European Comission and the ICAO. However, the complete life cycle assessment shows that achieving these goals depends greatly on the fuel production pathway. Powering aircraft with hydrogen produced through steam methane reforming, for instance, results in a greater GWP than using standard kerosene because of the very high level of GHG emissions released during the SMR process. Using liquid methane, on the other hand, despite also falling short of the CO2 reduction targets would reduce the impact of aviation on the climate change because although being less energy-efficient, methane has a lower GWP per unit energy than kerosene over its entire life cycle. It can finally be concluded that using hydrogen produced through the electrolysis of water powered by renewable energy sources is the only alternative that would not just meet but push beyond aviation’s environmental targets, which is not surprising given the very low GWP of this production pathway. Choosing the rear tank aircraft architecture and carrying kerosene reserves are two solutions that could still bring small gains in GWP reduction due to the higher energy-efficiency of these alternatives (table 5.1).

5.5 Fuel Pricing

Although LCA does not involve cost assessment, its outcome may have a great inﬂuence on the economic domain. In this particular case, the fuel price depends on its global warming potential due to the carbon pricing initiatives aiming at mitigating the climate change. Figure 5.3 shows the impact of carbon taxes on fuel prices based on their whole life cycle GWP. The baseline prices of kerosene and liquid methane were extracted from the IATA Jet Fuel Price Monitor [37] and the Bluegold Research LNG price online data [38], respectively, and represent their wholesale prices in November 2018. Hydrogen baseline prices regarding the two production pathways were derived from [21] and include the cost of production and liquefaction. As the wholesale price should still include the manufacturer’s proﬁt margin, a mark-up on the hydrogen production cost was considered. This mark-up was assumed to be the same as

46 the 7% profit margin of the oil refinery industry estimated from the U.S. Energy Information Administration (EIA) online data [39, 40]. According to the EIA, 68% of gasoline’s retail price on September 2018 covered crude oil purchase and refining costs. The difference between gasoline’s wholesale price (also reported by the EIA) and these costs is a good estimation of the refinery profit margin. Although the fuel economy is extremely volatile, this assumption enables a preliminary discussion on the economic viability of hydrogen comparatively to the other fuel alternatives.

Figure 5.3: Inﬂuence of the carbon pricing policies on the fuel price

It can be seen that although renewable hydrogen production is very expensive, carbon pricing can make it competitive with the other fuel alternatives due to its low environmental impact. Hydrogen via electrolysis becomes cheaper than hydrogen via SMR and kerosene from carbon taxes of approximately

200USD and 400USD per tonne of CO2eq, respectively. The current price of the EU ETS allowances is only 18USD/tCO2eq but if the implementation of CORSIA follows the accelerated rate of increase of the carbon taxes worldwide [41] electrolysis can soon become economically competitive with SMR, or even kerosene. Methane’s low price is, on the other hand, more difﬁcult to overcome.

47 48 Chapter 6

Conclusions and Future Work

The objectives set in the introductory chapter were accomplished. It is now necessary to draw the main conclusions of this study and map out the options for future work.

6.1 Conclusions

The feasibility of using hydrogen as an aviation fuel, for regional transport aircraft in particular has been proven from an aircraft performance perspective. The cryogenic tanks design and integration were clearly identified as the main challenge related to the use of liquid hydrogen. It has been shown that finding the right balance between mechanical and thermal properties while aiming at minimizing the overall fuel supply system mass is far from being an easy job. Using the foam insulation results in lighter vessels which can be a good opportunity to explore. However, in the present work, the MLI technology was chosen over the foam to enable a more conservative design since there might be exigent constraints regarding holding periods on the ground. The choice of the fuel storage system location in the aircraft was proven to be of great importance because it drives the mass and drag penalties inherent to the larger fuselage and, ultimately, the energy consumption. When compared to standard kerosene-powered planes, hydrogen aircraft are more energy-consuming but drastically reduce the harmful emissions during flight. Nonetheless, in order to achieve the environmental targets set by the ICAO, hydrogen production must be powered by renewable energy sources. Otherwise, if produced through steam methane reforming, the global warming potential of hydrogen throughout its life cycle is greater than that of kerosene. From the payload-range diagrams it could be understood that hydrogen aircraft are best suited to fly shorter ranges with greater payloads. Although the impact on the aircraft’s center of gravity was not analyzed, the single rear tank configuration was proven to be more energy-efficient than the baseline architecture. Two alternative solutions were also studied and led to interesting results. First of all, carrying kerosene reserves slightly reduces energy consumption while keeping the emissions very low during the block mission. The second alternative to hydrogen is using liquid methane as fuel that has been

49 proven to have a lower global warming potential than kerosene over its life cycle despite still generating carbon dioxide emissions during flight. Furthermore, liquid methane is very cheap and significantly more energy-efficient than liquid hydrogen when used as an aircraft fuel. Methane could provide a transition scenario while the technological and economic challenges regarding hydrogen production remain difficult to overcome. Finally, it is important to conclude that the MDO approach was very well suited for this work since it enabled a quick exploration of many different aircraft configurations providing satisfying and coherent results.

6.2 Future Work

Some future work suggestions can now be made in light of the work’s conclusions. First of all, a complete study regarding the economic viability of using hydrogen can be carried out by assessing the costs related not only to the new aircraft architectures but also to infrastructure, hydrogen production and transport, and maintenance operations. Some improvements can still be made in what concerns aircraft design. The rear tank conﬁguration is expected to raise some issues regarding the center of gravity as previously explained. Therefore, building up a scissor chart and calculating the plane’s center or gravity for different payload cases would enable sizing the horizontal tailplane (HTP) and repositioning the wing in order to meet control and stability requirements. The baseline conﬁgurations may also pose some problems regarding the aircraft’s control surfaces since a big part of the vertical tailplane (VTP) will be hidden by the hydrogen tanks over the fuselage. For that reason, further studies concerning the sizing of the VTP should also be carried out. A detailed structural analysis of the cryogenic tanks and respective integration in the fuselage can help maturing this technology.

50 Bibliography

[1] Airbus. Flying by nature - Global Market Forecast 2007-2026, 2007.

[2] Environment Branch of the International Civil Aviation Organization. On board a sustainable future. Technical report, ICAO, 2016.

[3] High Level Group on Aviation Research. Flightpath 2050 Europe’s Vision for Aviation. Technical report, European Comission, 2011.

[4] A. Vedantham. Aviation and the Global Atmosphere: A Special Report of IPCC Working Groups I and III. Technical report, 1999.

[5] V. Ramaswamy, O. Boucher, J. Haigh, D. Hauglustaine, J. Haywood, G. Myhre, T. Nakajima, G. Shi, and S. Solomon. Radiative Forcing of Climate Change. In J. Houghton, Y. Ding, D. G. M. Noguer, P.van der Linden, X. Dai, K. Maskell, and C. Johnson, editors, Climate Change 2001: The Scientiﬁc Basis - Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change, chapter 6, page 349. Cambridge University Press, Cambridge, 2001.

[6] Environment Branch of the International Civil Aviation Organization. ICAO Environmental Report 2013 - Aviation And Climate Change, 2013.

[7] About ACARE. [Accessed on November 2018]. URL https://www.acare4europe.org/ about-acare.

[8] P. Arguelles,¨ J. Lumsden, M. Bischoff, D. Ranque, P. Busquin, B. Droste, S. R. Evans, W. Kroll,¨ J.-L. Lagardere,` A. Lina, S. Rasmussen, P. Reutlinger, S. R. Robins, H. Terho, and A. Wittlov.¨ European Aeronautics: A Vision for 2020, 2001.

[9] European Comission. EU ETS Handbook, 2015.

[10] K. Seeckt. Conceptual Design and Investigation of Hydrogen-Fueled Regional Freighter Aircraft. Master’s thesis, KTH School of Engineering, 2010.

[11] G. D. Brewer. Hydrogen Aircraft Technology. CRC-Press, 1991.

[12] Tupolev - Cryogenic Aircraft [Accessed on November 2018]. URL https://web.archive.org/web/ 20120213180146/http://www.tupolev.ru/English/Picture.asp?PubID=1788.

51 [13] Scaled Agile, Inc. Introducing SAFe 4.6 for Lean Enterprises (PowerPoint presentation) [Accessed on July 2018]. URL https://www.scaledagileframework.com/videos-and-presentations/.

[14] Airbus Deutschland GmbH. Liquid Hydrogen Fuelled Aircraft – System Analysis. Technical report, 2003.

[15] M. R. Swain. Fuel Leak Simulation. In DOE Hydrogen Program Review, 2001.

[16] D. Verstraete. The Potential of Liquid Hydrogen for long range aircraft propulsion. PhD thesis, Cranﬁeld University, 2009.

[17] G. Corchero and J. L. Montan˜es.´ An approach to the use of hydrogen for commercial aircraft engines. Proceedings of the Institution of Mechanical Engineers, Part G: Journal of Aerospace Engineering, 219(1):35–44, 2004.

[18] C. Winnefeld, T. Kadyk, B. Bensmann, U. Krewer, and R. Hanke-Rauschenbach. Modelling and Designing Cryogenic Hydrogen Tanks for Future Aircraft Applications. Energies, 11(1), 2018.

[19] K. Seeckt, W. Heinze, and D. Scholz. Hydrogen Powered Freighter Aircraft – The Final Results of the Green Freighter project. 2010.

[20] Hydrogen Council. How hydrogen empowers the energy transition. Technical report, 2017.

[21] T. Ramsden, M. Ruth, V. Diakov, M. Laffen, and T. Timbario. Hydrogen Pathways - Updated Cost, Well-to-Wheels Energy Use, and Emissions for the Current Technology Status of Ten Hydrogen Production, Delivery, and Distribution Scenarios. Technical report, National Renewable Energy Laboratory, 2013.

[22] J. R. R. A. Martins. A Short Course on Multidisciplinary Design Optimization. University of Michigan, 2012.

[23] B-spline. [Accessed on September 2018]. URL https://en.wikipedia.org/wiki/B-spline.

[24] R. H. Byrd, P. Lu, J. Nocedal, and C. Zhu. A Limited Memory Algorithm for Bound Contrained Optimization. Technical report, Northwestern University - Department of Electrical Engineering and Computer Science, 1994.

[25] N. Hansen. The CMA Evolution Strategy: A Tutorial. Computing Research Repository, abs/1604.00772, 2016.

[26] C.-S. Lin, N. V. Dresar, and M. Hasan. A Pressure Control Analysis of Cryogenic Storage Systems. Journal of Propulsion and Power, 20(3):480–485, 1991.

[27] National Institute of Standards and Technology. Thermophysical Properties of Fluid Systems. [Ac- cessed on May 2018]. URL https://webbook.nist.gov/chemistry/fluid/.

[28] International Organization for Standardization (ISO). ISO 14040 Environmental Management – Life Cycle Assessment – Principles and Framework, 2006.

52 [29] R. W. Stratton, H. M. Wong, and J. I. Hileman. Life Cycle Greenhouse Gas Emissions from Alter- native Jet Fuels. Technical report, Massachusetts Institute of Technology (MIT), 2010.

[30] F. P., V. Ramaswamy, P. Artaxo, T. Berntsen, R. Betts, D. Fahey, J. Haywood, J. Lean, D. Lowe, G. Myhre, J. Nganga, R. Prinn, G. Raga, M. Schulz, and R. V. Dorland. Changes in Atmospheric Constituents and in Radiative Forcing. In S. Solomon, D. Qin, M. Manning, M. Marquis, K. Averyt, J. Melinda M.B. Tignor Henry LeRoy Miller, and Z. Chen, editors, Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, chapter 2, page 212. Cambridge University Press, Cambridge, 2007.

[31] G. Myhre, D. Shindell, F.-M. Breon,´ W. Collins, J. Fuglestvedt, J. Huang, D. Koch, J.-F. Lamarque, D. Lee, B. Mendoza, T. Nakajima, A. Robock, G. Stephens, T. Takemura, and H. Zhang. Anthro- pogenic and Natural Radiative Forcing. In T. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex, and P. Midgley, editors, Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovern- mental Panel on Climate Change, chapter 8, page 349. Cambridge University Press, Cambridge, 2013.

[32] P.L. Spath and M. K. Mann. Life Cycle Assessment of Hydrogen Production via Natural Gas Steam Reforming. Technical report, National Renewable Energy Laboratory, 2001.

[33] P. L. Spath and M. K. Mann. Life Cycle Assessment of Renewable Hydrogen Production via Wind/Electrolysis. Technical report, National Renewable Energy Laboratory, 2004.

[34] K. Seeckt and D. Scholz. Jet versus Prop, Hydrogen versus Kerosene for a Regional Freighter Aircraft. DGLR: Deutscher Luft- und Raumfahrtkongress 2009 : Tagungsband - Ausgewahlte¨ Manuskripte, 2009.

[35] N. Stuber, P. Forster1, and G. R. . K. Shine. The importance of the diurnal and annual cycle of air trafﬁc for contrail radiative forcing. Nature, 441:864–867, 2006.

[36] Natural Gas 1998 - Issues and Trends. Technical report, Energy Information Administration, Ofﬁce of Oil and Gas - U.S. Department of Energy, 1999.

[37] IATA - Jet Fuel Price Monitor [Accessed on November 2018]. URL https://www.iata.org/ publications/economics/fuel-monitor/Pages/index.aspx.

[38] Bluegold Research - Global LNG Prices [Accessed on November 2018]. URL https:// bluegoldresearch.com/global-lng-prices.

[39] Petroleum & Other Liquids - Gasoline and Diesel Fuel Update [Accessed on November 2018], . URL https://www.eia.gov/petroleum/gasdiesel/.

[40] Petroleum & Other Liquids - Spot Prices [Accessed on November 2018], . URL https://www.eia. gov/dnav/pet/PET_PRI_SPT_S1_D.htm.

53 [41] C. Ramstein, R. Goyal, S. Gray, A. C. Kallhauge, L. Lam, N. Klein, L. Wong, M. Quant, S. Nierop, T. Berg, and P. Leuschner. State and Trends of Carbon Pricing 2018. Technical report, World Bank Group and Ecofys, 2018.