Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
The challenges and potential benefits of electrified propulsion for aircraft
(USC Thesis Other)
The challenges and potential benefits of electrified propulsion for aircraft
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
THE CHALLENGES AND POTENTIAL BENEFITS OF ELECTRIFIED PROPULSION FOR AIRCRAFT by Michael Kruger A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulllment of the Requirement for the Degree DOCTOR OF PHILOSOPHY AEROSPACE ENGINEERING May 2022 Copyright 2022 Michael Kruger Dedicated to Zhamilya... for your unwavering support ii Acknowledgments Completing my doctorate without a strong support structure would have been nearly im- possible, and I would like to acknowledge the following people for their support. I would like to thank my mother Aritha and father Nico for shaping me into the person I am today and for encouraging me to always be curious about the world around me. I would like to thank my sister Marilet for her support throughout the years and the very happy childhood we spent together. I would like to thank my uncle Tom, who has now passed on to the next life. I have yet to meet a person more curious about the natural world than uncle Tom, and our many conversations played a big part in the shaping of my mind. I would like to acknowledge my grandmother Hester for her support over the years and, being an academic herself, for encouraging me to pursue my doctorate. I would like to thank my academic advisor at the University of Pretoria, Dr Lelanie Smith, for everything she facilitated for me during my graduate studies. I would like to ac- knowledge my academic advisor at the University of Southern California, Professor Alejandra Uranga, for her guidance throughout my doctoral studies. I have experienced a level of rigor working with you that I can only strive for. I think it's safe to say that no one `on the outside' really understands what completing a doctorate is like, and this creates a very real feeling of camaraderie between a cohort of doctoral candidates. Among my cohort, I would especially like to thank my good friend Arturo Cajal for his support. In a very real sense I do not know if I would have been able to get this far without his support. I would also like to thank Saakar Byhahut, whom I worked with closely throughout this time. And lastly, I would like to thank my beautiful ance Zhamilya for her unwavering support over the past years. Leaving everything I knew behind to move to a faraway land was not easy, but having you here as a rock to lean on has made the process not just easier, but enjoyable. I look so much forward to discovering the beauties of life with you. iii Contents Dedication ii Acknowledgement iii List of Tables vii List of Figures x Nomenclature xiv Abstract xvii 1 Introduction 1 1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.3 Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.4 Thesis Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 2 Background 6 2.1 A Brief History of Aircraft Electrication . . . . . . . . . . . . . . . . . . . . 6 2.2 Propulsion System Architecture Terminology . . . . . . . . . . . . . . . . . . 8 2.3 Previous Design Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.4 Existing Vehicles and Sub-Systems . . . . . . . . . . . . . . . . . . . . . . . 16 2.4.1 Trainer Market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2.4.2 Technology Demonstrators . . . . . . . . . . . . . . . . . . . . . . . . 18 2.4.3 Commuter Airlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2.4.4 Technology Testbeds . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 2.4.5 Propulsion Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 2.5 Synergistic Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 2.5.1 Distributed Propulsion . . . . . . . . . . . . . . . . . . . . . . . . . . 21 2.5.2 Boundary Layer Ingestion . . . . . . . . . . . . . . . . . . . . . . . . 24 2.6 Component Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 2.6.1 Batteries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 2.6.2 Electrical Machines . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 2.6.3 Power Electronics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 2.6.4 Thermal Management . . . . . . . . . . . . . . . . . . . . . . . . . . 30 2.7 Other Paths Towards Improved Eciency . . . . . . . . . . . . . . . . . . . 30 iv 3 Methodology 33 3.1 Framework Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 3.2 Aero-Propulsive Coupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 3.2.1 Flow Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 3.2.2 Compressibility Correction . . . . . . . . . . . . . . . . . . . . . . . . 41 3.2.3 Net Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 3.2.4 Dissipation Buildup . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 3.2.5 Eciencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 3.3 Propulsion System Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 3.3.1 Unied Model of Propulsion System Architectures . . . . . . . . . . . 44 3.3.2 Electrication Design Space . . . . . . . . . . . . . . . . . . . . . . . 48 3.3.3 Propulsion System Sizing . . . . . . . . . . . . . . . . . . . . . . . . . 51 3.3.4 Propulsor Distribution and Integration . . . . . . . . . . . . . . . . . 55 3.3.5 Fan Sizing, Distribution and Boundary Layer Ingestion . . . . . . . . 56 3.3.6 Estimating Fuel Burn and Battery Energy . . . . . . . . . . . . . . . 58 3.3.7 Drag/Dissipation Buildup . . . . . . . . . . . . . . . . . . . . . . . . 59 3.4 Integration Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 3.4.1 Mission Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 3.4.2 Aircraft Sizing and Optimization . . . . . . . . . . . . . . . . . . . . 60 3.4.3 Performance Constraints . . . . . . . . . . . . . . . . . . . . . . . . . 62 3.4.4 Weight, Balance, and Wing Placement . . . . . . . . . . . . . . . . . 62 3.5 Performance Metric . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 3.6 Electrical Technology Level Assumptions . . . . . . . . . . . . . . . . . . . . 64 3.7 Baseline Missions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 4 Results 69 4.1 Baseline Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 4.2 Commuter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 4.2.1 Design Range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 4.2.2 Extended Range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 4.3 Regional . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 4.3.1 Design Range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 4.3.2 Extended Range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 4.4 Transcontinental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 4.4.1 Design Range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 4.4.2 Extended Range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 4.5 Sensitivity Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 4.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 5 Conclusions 106 5.1 Summary and Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 5.2 Electric Propulsion Fundamental Truths . . . . . . . . . . . . . . . . . . . . 108 5.3 Further Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 Bibliography 111 v A Unied Propulsion System Model 121 A.1 Linear System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 B Performance Constraints 123 B.1 Maximum Takeo and Landing Field Lengths . . . . . . . . . . . . . . . . . 123 B.2 Maximum Rate of Climb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 B.3 Minimum Fuel and Battery Volume . . . . . . . . . . . . . . . . . . . . . . . 124 B.4 Minimum Stall and Approach Speeds . . . . . . . . . . . . . . . . . . . . . . 125 C Propulsion System Mass 126 C.1 Propulsion System Secondary Masses . . . . . . . . . . . . . . . . . . . . . . 126 D Wing Placement 128 D.1 Wing Placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 E Baseline Input Parameters 130 vi List of Tables 3.1 Components accounted for in static margin estimation. The wing location is a design variable used to meet the static margin requirement. MAC is the mean aerodynamic chord. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 3.2 Technology level parameters for 2035 entry into service (Kruger et al., 2018). 65 3.3 Data lters used to infer aircraft categories from T-100 data. . . . . . . . . . 66 3.4 Flight distances chosen for design and extended range missions. . . . . . . . 68 4.1 Characteristics of reference conventional aircraft. Data for Do 228 and E175 from manufacturer datasheets. Data for CSR-01 from Risse et al. (2016). . . 71 4.2 Calibration factors used for modeling reference conventional aircraft in LUCAS. 72 4.3 Comparison of model outputs for conventional reference and LUCAS results. 72 4.4 Performance constraints and related parameters used in the current work. Data for Do 228 from RUAG Aerospace Services GmbH (2021) and Finger et al. (2020). Data for E175 from manufacturer brochure, with maximum lift coecients assumed to be the same as CSR-01. Data for CSR-01 from Risse et al. (2016). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 4.5 Comparison of current and advanced technology baseline aircraft. PSEC is calculated assuming a fuel specic energy of 43 MJ/kg. The advanced tech values are used as baseline values to compare the performance of the electried concepts to. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 vii 4.6 Eect of distributing propulsors while keeping total propulsive area xed to that of baseline (22.5m diameter propellers) for design mission commuter with optimistic technology, f BLI = 0. . . . . . . . . . . . . . . . . . . . . . . . 75 4.7 Eect of distributing propulsors while varying total propulsive area for design mission commuter with optimistic technology. . . . . . . . . . . . . . . . . . 76 4.8 Eect of distributing propulsors while varying total propulsive area for extended mission commuter with optimistic technology. All congurations experience 24% BLI. 370 km alternate removed from reserve requirement to make analysis feasible (for this analysis only). . . . . . . . . . . . . . . . . . 80 4.9 Impact onM TO andPSEC for commuter when either using batteries for the block and reserve mission or only for the block mission. . . . . . . . . . . . . 86 4.10 Segment-varying electrication for design mission regional with optimistic technology. Congurations could use batteries for the block mission, but use only fuel for reserve mission. . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 4.11 Eects of distributed propulsion and boundary layer ingestion for turbo- electric design mission regional with conservative technology. Top and middle show results for a xed total propulsive area without and with BLI. Bottom shows results with propulsors auto-sized to cover 60% of the wing span for maximum distribution and BLI. Shaded row shows optimal design. . . . . . 92 4.12 Comparison of dierent technology levels for turbo-electric design mission re- gional with auto-sized propulsors to maximize BLI. Shaded rows show optimal designs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 4.13 Comparison of dierent technology levels for turbo-electric extended mission regional architecture with auto-sized propulsors to maximize BLI. Shaded rows show optimal designs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 viii 4.14 Comparison of dierent technology levels for turbo-electric design mission transcontinental architecture with auto-sized propulsors to maximize BLI. Shaded rows show optimal designs. . . . . . . . . . . . . . . . . . . . . . . . 99 4.15 Comparison of dierent technology levels for turbo-electric extended mission transcontinental architecture with auto-sized propulsors to maximize BLI. Shaded rows show optimal designs. . . . . . . . . . . . . . . . . . . . . . . . 101 4.16 Major modelPSEC sensitivities evaluated around design mission with inter- mediate technology assumptions. . . . . . . . . . . . . . . . . . . . . . . . . 103 ix List of Figures 2.1 Energy conversion eciency chains of dierent propulsion systems. Adapted from Hepperle (2012). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.2 Various electried aircraft design concepts. . . . . . . . . . . . . . . . . . . . 10 2.3 Various existing electried aircraft projects. . . . . . . . . . . . . . . . . . . 17 2.4 Cube-square scaling relating propulsor thrust and weight, going from one to four propulsors (Kruger et al., 2018). . . . . . . . . . . . . . . . . . . . . . . 22 2.5 Cube-square scaling eects showing change in system weightW and propulsive areaA versus number of propulsors. ForA 2 =A 1 curve, weight is kept constant. For W 2 =W 1 curve, propulsive area is kept constant. . . . . . . . . . . . . . . 23 2.6 Principle behind BLI benet: Ingestion and re-energizing of viscous wake leads to reduction in wasted kinetic energy. Figure from Uranga et al. (2017). 25 2.7 Airbus Cryoplane concept. Compressed hydrogen tanks stored in bulbous appendage on top of fuselage. Figure from Klug (2000). . . . . . . . . . . . . 32 3.1 Extended Design Structure Matrix (XDSM) (Lambe and Martins, 2012) for LUCAS, showing connections and data ow between subsystems. . . . . . . 34 3.2 Control volume around fan modeled as actuator disk with swirl. . . . . . . . 38 3.3 Unied propulsion system model (Kruger et al., 2018). The subscriptsM and E refer to mechanical and electrical components, respectively, and the powers P indicate the power ow between components. . . . . . . . . . . . . . . . . 45 3.4 Schematic of f S {f L design space showing regions for the dierent propulsion system architectures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 x 3.5 Unique propulsion system architectures that can be modeled with unied model by adjusting f S and f L . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 3.6 Parameter denitions used to compute propulsor fan diameters and extent of distribution and BLI. Propulsors could either be unducted (left) or ducted (right). Unducted propulsors are assumed for aircraft that y in the low subsonic regime, and ducted for transonic aircraft. . . . . . . . . . . . . . . . 57 3.7 Integration of pyOptSparse and SUAVE into LUCAS framework. . . . . . . 61 3.8 Distribution of commuter, regional and transcontinental ight distances in the US in 2019. Source: US BTS T-100 segment data. . . . . . . . . . . . . . . . 67 4.1 Chapter 4 layout. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 4.2 Conventional aircraft used as reference. Center: Do 228, Left: E175, Right: CSR-01. Models created with OpenVSP (Hahn, 2010). Human shown next to Do 228 for scale. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 4.3 Change inPSEC andM TO across electrication design space for design mis- sion commuter with intermediate technology. Congurations shown here have 4 electrical propulsors for f L 6= 0 and 2 mechanical propulsors for f L 6= 1, f S < 1. Hatched region represents infeasible design space. Optimum lies at f S = 0:42, f L = 0:26. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 4.4 Change inPSEC andM TO across electrication design space for design mis- sion commuter with optimistic technology. Congurations shown here have 4 electrical propulsors for f L 6= 0 and 2 conventional propulsors for f L 6= 1, f S < 1. Optimum lies at f S = 1, f L = 0:33. . . . . . . . . . . . . . . . . . . . 79 4.5 Change in PSEC and M TO across electrication design space for extended mission commuter with optimistic technology. Congurations shown here have 4 electrical propulsors for f L 6= 0 and 2 conventional propulsors for f L 6= 1, f S < 1. Hatched region represents infeasible design space. Optimum lies at f S = 0:72;f L = 0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 xi 4.6 Baseline conventional commuter category aircraft. Aircraft uses fuel and gas turbine powered turbo-prop engine to power two propellers with 2.5 m diam- eter propellers; this gives a total propulsive area A prop;tot of 9.82 m 2 . . . . . . 83 4.7 All-electric commuter category aircraft. Aircraft uses battery powered electric motors to power four propellers with 2.14 m diameter propellers; this gives an A prop;tot of 14.4 m 2 . This 47% increase in A prop;tot compared to the baseline shown in Figure 4.6 leads to a reduced disk loading and improved propulsive eciency. The primary eciency gains, however, come from the more ecient electrical components compared to the gas turbine system. The propulsors are integrated into the wing trailing edge to allow for BLI, which leads to further benets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 4.8 Parallel hybrid-electric commuter category aircraft. Aircraft uses both fuel and batteries, with an electric motor coupled to the drive shaft of a gas turbine powered turbo-prop engine. Such a conguration can use battery power to supplement the gas turbine, or can even operate on batteries only for the block mission, and switch to turbine power for the reserve mission, leading to further gains. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 4.9 Sweep across mission range for commuter showing optimal f S , i.e. optimal split between energy carried in batteries vs fuel. All missions include full reserve mission with 370 km diversion, 45 min loiter and 5% contingency. . . 86 4.10 Cumulative eects of electrication for commuter. All cases assume interme- diate technology, except for rightmost bar. 370 km alternate removed from reserve, with 45 min loiter still included. . . . . . . . . . . . . . . . . . . . . 88 4.11 Mass breakdowns of various commuter architectures, from left to right: baseline conventional, optimistic tech all-electric, intermediate tech hybrid- electric, optimistic tech hybrid-electric. All aircraft sized for design mission. 89 xii 4.12 Partial turbo-electric regional category aircraft. Aircraft uses fuel only to power turbo-fan like engines that have turbo-generators mounted to their output shafts. The turbo-generators are used to power an array of distributed electric propulsors integrated into the wing trailing edge, facilitating BLI. The addition of the 40 electric propulsors increases the overall propulsive area by 91%, increasing propulsive eciency in addition to the BLI benet. . . . . . 97 4.13 Mass breakdowns of various regional architectures, from left to right: base- line conventional, conservative tech turbo-electric, intermediate tech turbo- electric, optimistic tech turbo-electric. All aircraft sized for extended mission. 98 4.14 Mass breakdowns of various transcontinental architectures, from left to right: baseline conventional, conservative tech turbo-electric, intermediate tech turbo-electric, optimistic tech hybrid-electric. All aircraft sized for extended mission. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 xiii Nomenclature Acronyms AE, HE, TE All-, hybrid- and turbo-electric BLI Boundary Layer Ingestion BSE Battery Specic Energy DP Distributed Propulsion EIS Entry Into Service LUCAS Library for Unied Conceptual Aircraft Synthesis MAC Mean Aerodynamic Chord PSEC Productivity Specic Energy Consumption SUAVE Stanford University Aerospace Vehicle Environment TMS Thermal Management System Symbols surf Change in surface dissipation due to BLI propulsor installation _ h Aircraft rate of climb _ m Mass ow rate Component eciency p Propulsive eciency, (P K jet )=P K f L Load electrication factor, f L =P K;E =(P K;E +P K;M ) L D Lift to drag ratio f S Source electrication factor, f S =P bat =(P bat +P turb ) f wake Dissipation occurring in wake, 1f surf jet Jet dissipation surf Surface dissipation wake Wake dissipation P K Mechanical ow power Air density V jet Propulsive jet velocity xiv A Area AR Wing aspect ratio C L Lift coecient C L;max Aircraft maximum lift coecient D Drag or Diameter D 0 Total drag of equivalent aircraft without BLI D 0 p Prole drag of equivalent aircraft without BLI f BLI Boundary layer ingestion fraction F X Net stream-wise force F Y Net transverse force k s Propulsor swirl parameter L Lift M Mach number m Mass N Number of components (specic component given in subscript) P Power p Air pressure T Thrust V Velocity W Weight Subscripts () 1 Property in freestream ow () Tangential component () E Electrical () M Mechanical () p Propulsor () x Stream-wise component () bat Battery () EM Electrical Machine (motor or generator) () f Fan face () nace Nacelle () PE Power Electronics (inverter or rectier) () TO Takeo () turb Turbine xv Superscript () 0 Quantity of non-BLI conguration xvi Abstract This work presents a study of the potential benets and limitations of electric propulsion for aircraft. Components such as chemical batteries, electrical machines and power electronics benet from high power conversion eciencies and high specic powers (power conversion ca- pability per unit weight). These components could also enable higher levels of aero-propulsive integration through distributed propulsion and boundary layer ingestion, leading to further benets. Their use in aircraft propulsion has thus been proposed as a potential path towards improving aircraft energy eciency. The primary challenge lies in the fact that batteries have very low specic energies (energy stored per unit weight). A direct swapping of fuel for bat- teries would result in a ight range reduction of almost two orders of magnitude. Thus, for aircraft electrication to be benecial, solutions have to be found that take advantage of the benets whilst mitigating potential penalties. The design space of missions, propulsion system architectures and their technology levels is large. Understanding and quantifying the potential benets of electrication needs to be done on the basis of tools that can capture the multiple trade-os with sucient level of delity. This work presents LUCAS, a computational library that was developed around the SUAVE framework, which is able to simulate ight missions and aircraft with propulsion system architectures making use of varying degrees of electrication. This includes all-electric architectures storing all their energy in batteries, hybrid-electric architectures storing energy in both fuel and batteries, as well as turbo-electric architectures storing energy in fuel only, but converting power using generators to power electric propulsors. Using LUCAS, the on-board energy requirement of these architectures can be compared to that of baseline conventional aircraft on an equal basis and the design space extensively explored. Various ight missions are considered, assuming an entry into service in the 2035 time frame. The missions include commuter aircraft carrying 19 passengers over a distance as low as 180 km, up to trans-continental aircraft carrying 180 passengers over a distance as high as 2 400 km. It is found that all-electric architectures only become feasible for aircraft as large as the commuter category if very large improvements in battery specic energy are made, but could then lead to on-board energy savings of up to 36%. Hybrid architectures have the potential to increase the feasible range and reduce the weight penalties of all-electric architectures, albeit with lower energy benets. Such battery technology is, however, not expected to be realized soon. For the larger aircraft classes, it was found that turbo-electric architectures could be xvii benecial even with today's technologies, leading to benets of up to 6%, and if technology progresses further, benets of up to 11% could be attained. It was found that aero-propulsive benets due to distributed propulsion and boundary layer ingestion play a role in the poten- tial gains, although the benets primarily come from the high eciency and specic powers of the electrical propulsion components. Based on the work performed here, two candi- date concepts that seem best suited for electrication and most warrant further work are hybrid-electric commuter and turbo-electric regional and transcontinental aircraft. xviii Chapter 1 Introduction 1.1 Motivation In 2018, commercial passenger and freight aviation was responsible for around 2.4% of global carbon dioxide emissions due to the burning of fossil fuels (Graver et al., 2019). Even though this is not a very large contribution to global emissions, the gure represents a 32% increase over the preceding ve years. Thus, if commercial aviation continues to grow at the expected rate without a signicant improvement in the global eet eciency, CO2 emissions due to commercial aviation will end up representing a signicant fraction of global emissions. Additionally, because most commercial operations take place in the upper troposphere, the emissions are deposited in a region of the atmosphere where their eects might be magnied compared to fossil fuels being burnt at ground level (Schumann, 1993; Hidalgo and Crutzen, 1977). One technology that could lead to an increase in overall energy eciency is the use of electrical components in aircraft propulsion systems. The motivation for this is three- fold. First, electrical propulsion components (motors, generators and power electronics) are inherently more ecient than internal combustion engines. Second, these components have high specic powers, meaning that even lightweight components can convert large amounts of power. Third, electrical components can be scaled down while still remaining ecient, 1 with potential integration benets when distributing smaller propulsors over the airframe, whereas doing so with conventional propulsors is inherently more complex. The challenge, however, is that even the most modern chemical batteries have much lower specic energies than hydrocarbon fuels. This means that a battery storing the same amount of energy as fuel will weigh many times that of the fuel: with current technology the dierence is roughly two orders of magnitude. The primary goal of this research is to investigate whether electrifying the propulsion system could lead to improved aircraft energy eciency, and if so for which class of ight missions (payload and range) and with how much benet. Furthermore, the propulsion system components can be arranged in various ways, resulting in fundamentally dierent architectures. Another goal is to investigate whether certain architectures are best suited for specic missions. Lastly, it is hypothesized that electried aircraft performance depends strongly on the assumed technology level of the electrical components; a nal goal is to quantify the eects of assumed technology level on performance. 1.2 Approach This work introduces LUCAS, the Library for Unied Conceptual Aircraft Synthesis, devel- oped at the Aerodynamic Design & Research Laboratory (ADRL) of the University of South- ern California. LUCAS is a multi-disciplinary design, analysis and optimization (MDAO) framework that can be used to model advanced subsonic transport aircraft with an array of dierent propulsion system architectures: from conventional turbofan or turboprop sys- tems, to hybrid-, turbo- and all-electric aircraft 1 . A unied propulsion system model is implemented to represent the wide variety of propulsion system architectures via just two electrication parameters. The framework also models the eects of boundary layer inges- tion (BLI) and distributed propulsion (DP), which are technologies that are enabled by electrication and could lead to synergistic benets. 1 Architecture nomenclature will be introduced in following chapter. 2 At its core the framework uses the Stanford University Aerospace Vehicle Environment (SUAVE) (Lukaczyk et al., 2015) for mission simulation, which in turn uses pyOptSparse (Wu et al., 2020) for sizing and optimization. Using this framework, a trade-space analysis is conducted at a level of delity that will allow for meaningful conclusions to be drawn regard- ing which architectures are best|in terms of on-board energy usage|for which missions, with a low degree of uncertainty. 1.3 Scope This work focusses on conventional takeo and landing passenger-carrying aircraft, though the results could easily be extended to freight aircraft. The missions, however, dier fun- damentally from those performed by small unmanned aerial vehicles and those proposed for urban air mobility (sometimes called air taxis). Additionally, the scope is limited to aircraft carrying between 19 and 180 passengers, and the conclusions might not hold for smaller or larger aircraft. The missions considered here account for up to 48% of global commercial aviation emissions (Graver et al., 2019), and eciency improvements in these areas could thus have a signicant eect on overall aviation emissions. Analyzing both smaller and larger aircraft categories than considered here could be interesting avenues for further work. The scope of the current research is further limited to the use of relatively low modeling delity for subsystem components. Considering the unique choices of (1) aircraft category, (2) mission range, (3) technology level assumptions and (4) propulsion system architecture, this work looks at a design space with four dimensions. Within each of these, there are three aircraft categories, two ight mission ranges, three technology level assumptions and four propulsion system architectures. This translates to an exploration of 72 unique architectures. This large design space justies the use of relatively low component modeling delity. How- ever, once promising candidates out of this space have been identied, it might be benecial to improve the modeling delity to better quantify their performance. Lastly, only on-board energy usage is considered, meaning well-to-wake eects like en- 3 ergy required to produce and transport fuel and to manufacture and charge the battery are not considered and is beyond the scope of this research. 1.4 Thesis Organization Chapter 2 presents a general background on electried aircraft propulsion. First, a brief history of important developments and literature that led to the modern concept is presented. This is followed by the terminology that is used in this work to unambiguously refer to dierent propulsion system architectures. Following this, a summary of both theoretical design studies as well as real-world successes that have been achieved is given; this summary covers some of the early developments, spanning back around a decade. The potential aero-propulsive benets due to distributed propulsion and boundary layer ingestion form an important part of the current work, and the required knowledge to understand these principles is then presented. Similarly, it is expected that electrical component technology level assumptions will be in uential in determining the feasibility and potential benet of electried propulsion systems. A summary of the key parameters of interest is presented, which will form the basis of the modeling assumptions made in later chapters. Lastly, this chapter presents other potential paths towards more ecient aircraft that relate indirectly to electric propulsion. Chapter 3 presents the modeling methodology used in the current work. This starts with a high level description of the LUCAS framework. In order to model the eects of boundary layer ingestion, LUCAS makes use of the power balance method (Drela, 2009), which is described next. Following that, the unied propulsion system model is described. This model was created as part of previous work in which the author was involved (D. Hall et al., 2019; Kruger et al., 2018). A unique aspect of the unied model is that it allows for a simple parameterization of the propulsion system design space and thus forms an important part of LUCAS. Next, this chapter includes: methodology to size the propulsion system, approach to model distributed propulsion, modeling of performance constraints, description 4 of technology level assumptions, and a description of the metric used to compare the relative performance of electried aircraft. The chapter concludes with the analysis of market data that is used to select baseline mission denitions. Chapter 4 presents the results of applying the LUCAS framework to study the design space for electried aircraft. This starts with the calibration and validation of three baseline conventional aircraft representative of commuter, regional and transcontinental categories. These aircraft transport between 19 and 180 passengers over distances of between 180 and 2 400 km. Modeling these baselines with LUCAS is crucial to ensure a fair comparison with the electried aircraft. Next the mission, architecture and technology design space is explored by simulating electried versions of the baselines performing the same missions, and computing their energy eciency. For each aircraft category, both a short-range design mission and a longer range extended mission is considered, and the full set of propulsion system architectures (all-, hybrid- and turbo-electric) investigated for each mission. Further, for each mission and architecture, three dierent technology level scenarios are considered. Finally, Chapter 5 presents the conclusions drawn from this work and indicates potential avenues for further work. 5 Chapter 2 Background 2.1 A Brief History of Aircraft Electrication In 1973 a modied Brditschka HB-3 motor glider performed a 15 minute ight (17 km range) on battery power alone, in what was likely the rst all-electric manned ight in history. The modication included swapping the internal combustion engine for Ni-Cd batteries and an o-the-shelf forklift motor. If this ight were repeated today, while swapping the Ni- Cd battery with current state of the art lithium-ion batteries and the forklift motor for a modern purpose-designed one, the ight duration would be extended to over two hours and the range to 261 km (Hepperle, 2012). Even though some skepticism for a 15 minute ight time might have been warranted at the time, technology has progressed to a point where, at the time of writing, there are on the order of ten all-electric motor gliders on the market 1 . This application of electric propulsion has been facilitated primarily by the improvements in battery energy capacity and lightweight, high power electric motors and power electronics. It is the work of Hepperle (2012) that was most in uential to this author in highlighting the potential benets of electrication, through the gure adapted here in Figure 2.1. Internal combustion engines are strongly aected by the fact that nearly half of the useful energy 1 Models include Lange Antares 20E, Lange Antares 23E, Schempp-Hirth Arcus E, Pipistrel Taurus Electro G2, Silent 2 Targa LE, Air Energy AE1 Silent, Yuneec Apis 2, Yuneec EViva, Alpaero Exel, A eriane Swift and Alatus AL12 6 stored in fuel is lost through waste heat, as quantied by the cycle thermodynamic eciency. This results in an overall system eciency of between around 30% and 40% (turboprop engines are slightly more ecient that turbofans). These values will increase for future generations of gas-turbine engines, but the technology is very mature and is believed unlikely to exceed values of around 45% (Torenbeek, 2013). Due to the high eciency of all the components in an electrical system, the overall eciency of all-electric propulsion systems could reach values of up to 75%. Kerosene Kerosene Battery 100% 100% 100% Propeller =80% Turboprop Turbofan Electric =39% =33% =73% Thermo. cycle =50% Gearbox =98% Propeller =80% Thermo. cycle =50% Fan + nozzle =65% Gearbox =98% Electric motor =95% Controller =98% Propeller =80% Figure 2.1: Energy conversion eciency chains of dierent propulsion systems. Adapted from Hepperle (2012). Both industry and academia have taken notice of these technological improvements and the possible eciency gains, and a market study performed in 2018 showed that, at the time, there were around 170 electric aircraft programs in development, and the number was predicted to be more than 200 by the end of 2019 (Roland Berger, 2018). However, around half of those programs were related to urban air mobility, a mission class dominated by vertical take-o and landing vehicles that is outside the focus of the current work. Among the more traditional, conventional takeo and landing aircraft, it seems that there is not 7 yet a consensus on what missions would benet most from electrication, or which specic propulsion system architecture would be most appropriate to which mission. Although an aircraft ight mission is typically dened by numerous parameters, a mission is dened in this work as a combination of aircraft payload (number of passengers) and design range. Major questions to consider are thus: (1) Given the current state of the art in component technology, does a mission and architecture exist for which electrication can lead to a reduction in on-board energy requirement? (2) What are the required technology levels that need to be reached to enable electrication, and for which missions might this not only be feasible, but benecial? In this chapter, some of the studies that have been done in an attempt to answer these questions will be presented. The chapter will be divided into conceptual design studies, as well as real-world projects where hardware has already been created and tested, or are in the process of being introduced. Brelje and Martins (2019) give a good review of such projects, and is recommended reading for a more in-depth analysis. Another means in which electrication can be benecial is by enabling technologies that cannot be implemented to the same degree without electrication. Such technologies include distributed propulsion (SP) and boundary layer ingestion (BLI). Some of the requisite knowledge required to understand these technologies will also be discussed in this chapter. Lastly, it is noted that electric propulsion is one of a number of technologies that could be used to improve the eciency of aircraft. The nal section of this chapter will look into another path towards improved eciency, namely the use of alternative fuels. Although not the direct focus of the current work, the author would be remiss not to touch on this related topic. 2.2 Propulsion System Architecture Terminology Before looking at the existing work that has been done, it is necessary to dene the terminol- ogy used here to refer to dierent electric propulsion system architectures. A conguration where all energy is stored in hydrocarbon fuels and all propulsive power comes from internal 8 combustion engines is here called a conventional conguration. When all energy is stored in electro-chemical batteries and all propulsive power comes from electrically powered fans, the conguration is called all-electric. When energy is stored in both fuel and batteries, the conguration is referred to as a hybrid, and a distinction is made between series-hybrid and parallel-hybrid architectures. For series hybrids, power is rst converted via electrical generators, and then used to supplement battery-powered electrical fans for propulsion. For parallel hybrids, mechanical propulsive power is supplemented with battery-powered motors mounted directly to the mechanical fan shafts. Lastly, turbo-electric architectures are ones where all the energy is stored in fuel, but propulsive power can be provided entirely from electrically powered fans via generators (fully turbo-electric) or a split between electrically and mechanically powered fans (partial turbo-electric). Note that an aircraft could operate in dierent modes during dierent mission segments, for example as an all-electric during takeo and climb, and as a conventional or hybrid conguration during cruise. 2.3 Previous Design Studies In this section, some important design studies that have looked into electrication are pre- sented. The word `studies' here implies that the work hasn't resulted in any actual ight hardware at the time of writing. This section is not necessarily exhaustive, as this eld is expanding rapidly, and many of the existing studies by commercial ventures are likely not widely published. Some of the concepts that will be discussed are shown in Figure 2.2. In the past decade or so, there have been multiple large-scale, primarily government- funded, research eorts into the potential benets of electrication. In the US, these pro- grams have mostly been funded by NASA in partnership with industry and academia. Promi- nent among these projects is the Subsonic Ultra Green Aircraft Research (SUGAR) study (Bradley et al., 2015), which was funded under a NASA N+3 program and performed by Boeing, in partnership with General Electric and Georgia Tech. This study looked at the 9 Boeing Concept Artwork Boeing Sugar Volt. © Boeing. NASA STARC-ABL. © NASA. NASA PEGASUS. © NASA. ONERA DRAGON. © ONERA. Bauhaus Luftfahrt Ce-Liner. © Bauhaus Luftfahrt. Bauhaus Luftfahrt Centerline. © Bauhaus Luftfahrt. ES AERO ECO-150-300. © ES AERO. Airbus E-fan X. © Airbus. Figure 2.2: Various electried aircraft design concepts. parallel hybridization of a Boeing 737 sized airplane carrying 154 passengers in a duel class layout, which was termed the SUGAR Volt. Assuming a battery specic energy (BSE) of 750 Wh/kg, this study found that signicant fuel burn and emissions reductions could be 10 achieved for a 900 nmi economic mission, although an overall energy reduction (from bat- teries and fuel) was not predicted. The fuel burn reduction due to electrication was found to be just more than 10%. A NASA in-house project was the turbo-electric STARC-ABL concept (J. Welstead et al., 2017; J. R. Welstead and Felder, 2016). The STARC-ABL is a single aisle concept that makes use of generators embedded in the conventionally-mounted turbofans, which power a tail-mounted electrically powered fan that ingests the fuselage boundary layer. Those researchers found that the concept could result in around a 15% reduction in fuel burn for the 3 500 nmi design mission and around 9% for the 900 nmi economic mission. The modeling of the tail-cone thruster integration on the fuselage for this study was done at a relatively low delity, and a signicant amount of work has been done since then by Gray et al. to develop a method to optimize the aerodynamic design of a tail-mounted boundary layer ingesting propulsor (J. S. Gray and Martins, 2019; J. S. Gray et al., 2018; J. Gray et al., 2017). Although these higher delity analyses predict that the tail-mounted BLI propulsor could lead to power savings of up to 20% if the fuselage and tail-mounted propulsors are seen in isolation, it was found that the actual performance will strongly depend on the power transmission eciency from the turbo-generators mounted in the under-wing turbofans to the tail-mounted fan. If the power transmission is 95% ecient, power savings of only 2.5% would be seen, and if it is 98% ecient, 4.6% power savings could be attained. The NASA PEGASUS study is currently looking at a similar electrical distribution architecture, but applied to a regional turboprop, using the ATR-42 as a baseline, with some promising initial results. The PEGASUS, however, also uses batteries for energy storage and is thus a hybrid-electric concept, whereas the STARC-ABL is turbo-electric. NASA also funded the LEARN3 project (D. Hall et al., 2019), in which this author was involved. That project was performed by the Massachusetts Institute of Technology, in partnership with the University of Southern California and Aurora Flight Sciences in 2017. Two distinct studies were performed as part of that work: (1) A low delity trade-space 11 analysis of electric propulsion system architectures and missions (Kruger et al., 2018), upon which this thesis builds. (2) A validation study of the NASA STARC-ABL concept (David K. Hall et al., 2018). Both studies used the GPKit geometric programming framework (Burnell et al., 2020). The former of these two studies modeled a ight mission for a single cruise-only operating point, and did not model the full mission. The study did, however, cover a large range of payloads (20 to 350 passengers) and ranges (500 to 6 000 nmi). That study found that the largest benets from electrication can be attained by all- and hybrid-electric smaller aircraft ying short range missions, and that turbo-electric architectures become optimal for longer ranges. It was, however, acknowledged that a higher delity approach that models the entire mission should be implemented to draw more meaningful conclusions. The latter study (David K. Hall et al., 2018), which had the goal of validating the STARC-ABL results found by NASA, found that a conguration like the STARC-ABL would actually lead to a slight increase in fuel burn due to a signicant increase in propulsion system mass. Various studies into electrication have also been carried out in Europe under the Clean Sky program (Atkinson et al., 2017) which ended 2017, and the ongoing Clean Sky 2 program, set to end in 2024. These are large programs with many government, industry, university and private research institutes as participants. The main participants in Clean Sky 2 that are looking into electric propulsion are the French ONERA, the German DLR and the Dutch NLR and TU Delft. The work of ONERA has culminated in the DRAGON concept, which is a 150 passenger aircraft with a design range of 2750 nmi. This turbo-electric concept (no batteries) focuses on distributed propulsion, and predicts a fuel burn reduction of around 7% compared to a conventional aircraft, both designed for entry into service (EIS) in 2035 (P. Schmollgruber et al., 2019). The work of the DLR has culminated in a parallel hybrid concept that they call the `Boosted Turbofan' (Hecken et al., 2020). The performance of this concept was compared to that of an advanced technology baseline aircraft for EIS in 2035, and it was found that the concept doesn't result in any signicant reduction in fuel burn. This concept didn't make use of any technologies enabled by electrication, such as BLI or 12 DP. As part of the Dutch eort, de Vries et al. (2018) has published a methodology for the analysis of hybrid-electric aircraft that includes a method for estimating the aero-propulsive eects of distributed propulsion. This methodology was applied to an aircraft carrying 70 passengers with a design range of 850 nmi (de Vries et al., 2019), and found no benet for that mission while predicting a slight increase in energy requirement. A later study that opened up the design space and looked at mission ranges from 500 to 2000 nmi and 50 to 200 passengers found that a 5% energy reduction can be attained if an 11% improvement in aero- propulsive eciency ( p L D ) could be facilitated by distributed propulsion, and this occurs if 20% of propulsive power comes from batteries (de Vries et al., 2020). Similar work done by Hoogreef et al. (2020) investigated various levels of electrication of regional turboprops, and found little or no benet due to electrication. That work, however, considered only turbo-electric architectures, with no batteries. One of the most active participants in this eld has been the German Bauhaus Luftfahrt, rst with their Ce-Liner concept (Hornung et al., 2013), and more recently the Centerline consortium under the European Union Horizon 2020 project (Seitz et al., 2018). The Ce- Liner is an all-electric concept, designed to carry around 190 passengers over a range of 1 666 km. In terms of conguration, the concept uses a tubular fuselage, but has a non-planar `c-wing' wing design. The study assumed EIS for 2035 and an optimistic 2000 W-hr/kg BSE. Even with this assumption, the study found that takeo mass compared to a conventional aircraft would grow signicantly (almost 50%), although the energy requirement decreases by around 26%. The follow-up Centerline project, which is being done by Bauhaus Luftfahrt in collaboration with various industry, government and academic partners, is looking at a turbo-electric conguration similar to the NASA STARC-ABL concept, although the concept is larger, designed for a 340 passengers, 12 000 km (6 500 nmi) mission, and uses the Airbus A330-300 as a baseline conventional aircraft. Preliminary wind-tunnel testing have shown a 14% reduction in required power, and their design framework showed a 11.3% reduction in block fuel burn, compared to an advanced technology version of the baseline conventional 13 aircraft. Also under the Horizon 2020 project, the IMOTHEP project, which will be lead by the French ONERA, will look at various electric propulsion aircraft concepts. ONERA has re- leased an initial assessment of a turbo-electric concept that they call DRAGON, which uses aft-fuselage mounted turbo-generators to power 40 fans distributed along the wing. They found that this 150 passenger concept can reduce fuel burn by 8.5% compared to a conven- tional conguration for a 1 482 km (800 nmi) mission, assuming moderate improvements in electrical component technology for an entry into service in 2035 (Peter Schmollgruber et al., 2019). Another concept that has seen much work and numerous iterations is the ES Aero ECO-150. There is a large body of literature on the evolution of the concept, which is summarized by Freeman and Schiltgen (2019). In its latest variation, the ECO-150-300 carries 150 passengers over a design range of 6 500 km, or an economic range of 1 666 km, while cruising at Mach 0.785 (increased from M 0.7 in previous iterations). The concept uses a turbo-electric architecture, with turbo-generators powering an array of distributed propulsors embedded in a `split-wing' design. For a 2035 EIS, those authors report an 11% fuel burn reduction compared to an advanced technology conventional baseline aircraft. These benets reportedly come from a 5% increase in transonic eciency (M L D ) due to aero-propulsive benets of the array of propulsors integrated into the wing, as well as a 45% increase in maximum aircraft lift coecient (C L;max ) due to powered lift and blown aps. A hybrid-electric architecture was also considered, assuming a BSE of 400 Wh/kg. It was found that such a system would not be benecial, and the latest ECO-150-300 concept thus remains a fully turbo-electric one. ES Aero is also the prime contractor for the NASA X-57 concept that will be discussed in the next section. Other projects that are also worth mentioning are Zunum Aero, which proposed a small regional hybrid-electric concept, and even though the company received much attention, at the time of writing it does not seem like the project is still being pursued. The company 14 Wright Electric is developing electric propulsion system components and have partnered with the British low cost airline easyJet to develop a 186-passenger airliner with 500 km range, but full details of their chosen architecture and battery specic energy assumption is unknown. Another project is the Eviation Alice, which is a 9-seater all-electric commuter aircraft concept that has also received much attention, and has reportedly received many orders, although the rst prototype has not yet been own. Airbus is working with Daher and Safran on the EcoPulse, retrotting a Daher TBM turboprop aircraft with a turbo- generator coupled to the existing main engine in the nose, a battery system and an array of electrically powered fans along the leading edge, which would require a smaller wing thanks to blown lift augmentation. The rst ight for this aircraft is expected in 2022. Lastly, the author of this thesis was also involved in a conceptual design study of a 19 passenger commuter aircraft where the eects of technology level, mission range and propulsion system architecture were investigated (Kruger and Uranga, 2020; D. Hall et al., 2019). That study looked at all-, hybrid- and turbo-electric architectures, and found that signicant improvements in battery technology would be required to make the mission feasible for all-electric designs. It was, however, found that if improvements were made then such architectures could lead to signicant reductions in on-board energy requirements. It was further found that hybrid architectures could improve the feasible range of electried commuter aircraft and make them feasible at lower, more realistic, technology levels. Turbo- electric architectures were found to not be benecial for that aircraft class. That work forms the foundation of the commuter category analysis that will be presented later in this thesis. A key addition to that work is the consideration of a full reserve mission, which includes energy storage for a diversion to an alternate airport, a loiter segment and an additional 5% contingency requirement. In the previous work, only the loiter segment was considered, which in hindsight lead to somewhat optimistic conclusions. 15 2.4 Existing Vehicles and Sub-Systems In this section, some important projects that have produced ight hardware are presented. The distinction between this and the previous section is that these projects are not only conceptual, but have shown real-world results. As in the previous section, the list of projects given here is not exhaustive, but rather the most prominent projects that can be found in the literature at the time of writing are presented, while the eld is changing rapidly. Where the design studies of Section 2.3 focus mainly on large aircraft, the projects that will be discussed here focus mainly on smaller aircraft. This is likely due to the fact that many of the costs related to aircraft development scale with aircraft weight (Raymer, 2012), and also because of the lower technology readiness level (TRL) of the high-power electrical components that will be required for large aircraft. Smaller aircraft thus represent an easier short-term route to commercial applications. Some of the projects that will be discussed are shown in Figure 2.3. 2.4.1 Trainer Market The trainer market represents an ideal application for all-electric aircraft because it could signicantly reduce the barrier to entry for ight by reducing operating costs. Using a eet of all-electric trainers would eliminate fuel costs, and since grid electricity for battery charging is much cheaper than the cost of fuel, 2 this will result in a net saving above that due to the inherently more ecient all-electric architectures. By also reducing maintenance costs because of the higher reliability of electrical components, schools could signicantly reduce the costs of training. Even though all-electric aircraft are currently heavily range-limited because of the low specic energy of batteries, circuit (or pattern) training doesn't require long mission endurances, and constitutes a large part of initial training. For this market, the Slovenian company Pipistrel markets the Velis Electro, which is the rst type-certied all- 2 At this time, grid electricity in Los Angeles, California, costs around $0.2/kWh, whereas 100LL aviation gas, typically used for trainers, costs around $2.50/kWh: an order of magnitude more expensive. 16 Pipistrel Velis Electro. © Pipistrel. Bye Aerospace eFlyer 2. © Bye Aerospace. Pipistrel T aurus G4. © Pipistrel. Airbus E-fan. © Airbus. Harbour Air + MagniX. © Bloomberg. Ampaire Electric EEL. © CNN. Figure 2.3: Various existing electried aircraft projects. electric aircraft, and was certied in Europe by EASA in 2020. The Velis has an endurance of up to 50 min, plus reserves. Another small aircraft that is aimed at the ight training market is the Bye Aerospace eFlyer 2. The eFlyer 2 is a 2-place aircraft that is currently in the process of being certied under FAA Part 23. This aircraft is said to have and endurance 17 of 3.5 hours and uses an electrical motor from Rolls-Royce (initially designed by Siemens, whose electrical propulsion division was acquired by Rolls Royce). There are plans to develop a 4-place version, the eFlyer 4. 2.4.2 Technology Demonstrators The Pipistrel Taurus G4 was a twin fuselage version of their Taurus motor glider, which in 2011 won the Comparative Aircraft Flight Eciency (CAFE) Foundation Green Flight Challenge by ying 320 km in less than two hours, using less than the energy-equivalent of one gallon of fuel. The runner-up for the challenge was also a motor glider: the e-Genius, developed by the University of Stuttgart. The e-Genius was later converted into a series hybrid-electric aircraft (Gei et al., 2018). Another much publicized all-electric demonstrator was the Airbus E-fan, which in 2015 performed a crossing of the English Channel. The E-fan was a small single-seat aircraft, with an endurance of around one hour. At the time of that project, there were talks of commercially producing a 2-seat production version, the E-fan 2.0, and a 4-seat hybrid- electric variant, the E-fan 4.0. Although these commercial projects never realized, Airbus did create a hybrid-electric demonstrator version of the E-fan, which they called the E-Fan 1.2. The demonstration project which has arguably seen the most eort, at least from what is published, is the NASA X-57 Maxwell. The X-57 is more of a demonstrator platform than one specic aircraft, and uses a Tecnam P2006T as a baseline aircraft and plans on systematically electrifying the aircraft through a series of modications. The nal goal is to have an all-electric aircraft with a much smaller wing than the original, with small electric fans distributed over the leading edge providing `blown lift'. It is unclear exactly where the program stands at the moment, but at the time of writing, it seems that Mod II ground tests have been performed, in which the baseline piston engines are replaced with electric motors, while keepingthe original wing. Mod III will see the motors moved to the wing-tips, 18 and for Mod IV the wing will be replaced completely by a custom-built, high aspect ratio wing with leading edge distributed propulsors. As mentioned earlier, ES Aero is the primary contractor for this project. The X-57 was initially scheduled for rst ight in 2020, but at the time of writing this has not yet happened. 2.4.3 Commuter Airlines The other area where electrication is being considered is for aircraft belonging to the com- muter category, which typically y short `island-hopping' style missions. Key amongst exist- ing projects in this eld is the collaboration between Harbour Air and MagniX. Harbour Air is a Vancouver-based operator of sea planes, primarily de Havilland Canada (DHC) models, and the operator is planning to replace the conventional propulsors of its existing eet with all-electric propulsion systems developed by MagniX. A rst and successful test ight was performed with a retrotted DHC-2 Beaver in 2019. Because of the short distances between Harbour Air's operating bases, the limited ranges allowed by all-electric aircraft might still be able to cover a signicant number of their routes, a mission segment that could benet from electrication even with today's technology. Another company that is envisioning electrication for commuter missions is Ampaire with their Electric EEL. This is a modied Cessna 337 Skymaster, a twin-engine aircraft, with one engine in front of the fuselage and one behind it, both providing center-line thrust. For the modication, the rear engine was replaced with an electric motor. The conguration thus uses a hybrid architecture, and the mechanical and electrical propulsors operate completely independently from each other. The Electric EEL has already performed extensive test campaigns, and performed a pilot project in Hawaii in 2020 with scheduled cargo missions between towns on the Maui island. 19 2.4.4 Technology Testbeds Signicant investments have been made into ground testing equipment, or so-called `copper birds', for electric propulsion systems. The most widely published testbed is the NASA Electric Aircraft Testbed (NEAT), which can reportedly test propulsion systems of up to 24 MW and at simulated pressure altitudes of up to around 36 km (120 000 ft). At the time of writing, only preliminary testing results of a 500 kW system has been published (Dyson, 2018; Jansen et al., 2017), but a facility like NEAT may prove to be a very useful testbed for the future development of electric propulsion. 2.4.5 Propulsion Systems Apart from the airframe development, a signicant amount of research has also been devoted to developing the propulsion systems themselves. This includes the development of electrical machines (motors and generators) and power electronics (rectiers, inverters and converters). The primary companies who have done commercial work in this eld are Siemens, MagniX, SAFRAN and Rolls-Royce. Other large names among conventional engine manufacturers, such as General Electric and United Technologies Corporation (UTC, now Raytheon Tech- nologies) have also looked into integrating electric motors into existing turbofan engines to create hybrid systems, but it is unclear what hardware has been built. General Electric was a partner on the NASA SUGAR Volt project (Bradley et al., 2015). Also Raytheon Technologies (through subsidiaries Pratt & Whitney Canada and Collins Aerospace), and De Havilland Canada are working together on retrotting a DHC-8 with a hybrid-electric propulsion system which, at the time of writing, is scheduled to y in 2024 (Polek, 2021). Siemens (whose electric propulsion division was acquired by Rolls-Royce in 2019) has already developed and is selling a series of motors developed specically for electric aircraft, and was used to power the E-Fan and a retrotted Extra 300 that set a world record for shortest time to climb to 3 000 m and a speed record for electric aircraft. The Siemens 20 motors were also selected as the primary motors to power the Eviation Alice. At the time of writing, it seems that the biggest competitor to the Siemens motors is MagniX, which has also developed motors of similar size of that of Siemens. Theey have successfully own one of their systems in a demonstration ight of a retrotted Harbour Air DHC-2 Beaver, as discussed earlier. MagniX was also selected as an alternative propulsion system option for the Eviation Alice. As far as more fundamental research, NASA has been working on various projects to scale electric propulsion to larger power levels. Apart from the NEAT testbed that was mentioned above, there is also the Hybrid Electric Integrated System Testbed (HEIST), which was used to test a stand-alone wing with an array of leading-edge propulsors for the LEAPTech project (Clarke et al., 2015), which later merged into the X-57 project. Jansen et al. (2017) gives a good overview of NASA involvement in these projects, which includes: work with Ohio State University to develop a 2.7 MW induction electrical machine, work with the University of Illinois on a 1 MW permanent magnet machine (Haran et al., 2017) and at NASA Glenn on a 1.4 MW wound eld machine (Jansen et al., 2017). As an eort to develop power electronics, NASA is working with General Electric to develop a 1 MW Silicon Carbide inverter (Zhang, 2017), with Boeing to develop a 1 MW cryogenically cooled inverter (Liu, 2017), and with the University of Illinois to develop a 200 kW gallium nitride inverter (Pilawa, 2017), which is reportedly scalable to the MW level. 2.5 Synergistic Technologies 2.5.1 Distributed Propulsion Distributed Propulsion (DP) refers to the distribution of multiple propulsors over the air- frame, allowing for total thrust to be distributed over a large number of propulsors, rather than coming from few (typically less than four) propulsors, as in conventional aircraft. There is no formal denition or convention for how many propulsors are required for an aircraft 21 Thrust ∝ Area ∝ D 2 Weight ∝ Vol ∝ D 3 Figure 2.4: Cube-square scaling relating propulsor thrust and weight, going from one to four propulsors (Kruger et al., 2018). to qualify as a DP aircraft, but here it is assumed that a number higher than four qualies, since most existing aircraft typically have between one and four propulsors. DP has various theoretical benets, including scaling eects due to the cube-square relationship between en- gine thrust and weight, redundancy due to larger number of propulsors and reduced power requirement per component. The basic principle behind the cube-square scaling benet mentioned above is a reduc- tion in weight per fan face area, or an increase in fan area if weight is kept xed, as illustrated in Figure 2.4. If one assumes that propulsor weight scales with a given length scale (say, it's diameter) cubed, and propulsor thrust scales with the length scale squared. Thus, going from one to four propulsors would lead to a 1.6 times increase in propulsive area if total weight is kept constant, or equivalently a halving in system weight if propulsive area is kept constant. When weight is kept constant, the increase in propulsive area would lead to an increase in propulsive eciency for a given thrust. Figure 2.5 shows the general trend of propulsive area and system weight scaling. It can be seen that, especially for weight, improvements get diminished for more than around 10 propulsors, whereas area (and thus eciency) benets still scale to higher numbers. How much of the potential benets are actually achieved would depend on the strategy of how 22 Figure 2.5: Cube-square scaling eects showing change in system weight W and propulsive area A versus number of propulsors. For A 2 =A 1 curve, weight is kept constant. For W 2 =W 1 curve, propulsive area is kept constant. DP is implemented in an actual design: because of the quadratic relationship between a propulsors diameter and its propulsive area (A p / D 2 p ), a reduction in diameter would lead to a disproportionate reduction in area. It might thus be dicult to maintain the same propulsive area (and eciency) without requiring a very large number of distributed propulsors. It might be noted that the trend in propulsor count has in fact gone from more propulsors to less, with the latest wide-body airliners having only two engines; here the author would argue that this relates more to maintenance benets of having less gas turbine engines to maintain and the fact that gas turbine cores become inecient when they get too small (Rolt et al., 2017). Electrical propulsors might negate the maintenance issues because of their inherent reliability and the fact that high eciencies can still be maintained by small electric motors. Distributed propulsion also has the benet of reducing the power requirements on a per-component basis, requiring lower current levels to power individual components. This might have the benet of making a wider range of o-the-shelf components usable, since lower power components are typically more widely available, and since electrical losses scale 23 with current squared (P loss / I 2 R). For a given total power level required, less propulsors would then lead to lower current levels per propulsor, and thus lower electrical losses overall. The electrical distribution in DP aircraft architectures are often compared with micro- grids, since they require power generated from few sources to be distributed to multiple power sinks (propulsors), while ensuring synchronization between all electrical signals. Elec- tried aircraft making use of DP could leverage existing methods and techniques developed for microgrids to improve power distribution eciency, as well as ensuring robustness and fault-tolerance. Loder and Armstrong (2018) present a useful distribution model that is applicable to both AC and DC systems. The model uses two turbo-generators to power an array of electrical propulsors, and ensures that all propulsors can still be powered if one of the turbo-generators fail. 2.5.2 Boundary Layer Ingestion Boundary Layer Ingestion (BLI) is a technology that leverages the closer integration be- tween airframe and propulsor to reduce the required system-level power to produce a given streamwise force. This is done by ingesting and re-energizing the viscous wake, or boundary layer, of the aircraft. One of the earlier treatments of the potential performance benets of BLI was done by Smith (1993), who notes that this technology is already widely used in marine applications, such as for submarines and torpedoes. BLI is also a key technology used in the MIT D8 transport conguration (Mark Drela, 2011), for which the power balance method was developed (Drela, 2009). In contrast to more conventional formulations that model force and momentum ow, the power balance method models mechanical power and kinetic energy ow. A key dierence is that where momentum-based formulations balance thrust and drag, the power balance method balances power and dissipation. Although used in marine applications, BLI for aircraft is a relatively new technology and is not currently in use in any existing aircraft. Its merit has, however, been demonstrated through extensive wind-tunnel testing for the D8 conguration by Uranga et al. (2017), which showed a ben- 24 wake, or “draft” Wasted Kinetic Energy Zero Net Momentum combined wake and jet propulsor jet + + + + + + + - - - Figure 2.6: Principle behind BLI benet: Ingestion and re-energizing of viscous wake leads to reduction in wasted kinetic energy. Figure from Uranga et al. (2017). et due to BLI, as quantied by a reduction in required cruise power, of up to 9% for that conguration, when ingesting around 40% of the fuselage boundary layer. The fundamental principle behind BLI is shown in Figure 2.6. By integrating the propulsor closer to the aircraft, the slow-moving boundary layer air can be ingested and re- energized. This has a wake-canceling eect, and for the same net streamwise force (analogous to drag minus thrust), the total momentum defect in the boundary layer is reduced. This reduction implies that the aircraft would need to add less power to the ow to overcome total dissipation (or wasted kinetic energy), which would translate directly to a reduction in fuel burn or battery power draw. An inherent complexity of BLI is that thrust{drag bookkeeping becomes ambiguous since it is not clear whether the reduced momentum defect should be handled as an in- crease in available thrust, or a decrease in drag. The power balance method avoids this ambiguity by analyzing the aerodynamics and propulsion integration problem in terms of mechanical power, kinetic energy and dissipation ows. The application of this method to aircraft performance analysis has been demonstrated by Uranga et al. (2018) and Hall et al. (2017). Others have also created methodologies to quantify the eects of BLI, and often rely on the adjustment of the aerodynamic eciency, or lift-to-drag ratio L=D. An in-depth review of dierent performance bookkeeping methods for BLI congurations is presented by Habermann et al. (2019). In that publication, the bookkeeping methods are divided into 25 momentum-based methods, kinetic (or mechanical) energy-based methods or exergy-based methods, and an exhaustive list of publications in which the various methods were developed and applied are given. As mentioned above, the power balance method used here falls under the category of kinetic energy-based methods. 2.6 Component Technologies Ultimately, the feasibility and potential benet of electric propulsion for a specic ight mission will depend on the technology levels of the electrical components used. The primary parameters of interest are the battery specic energy and specic power, in addition to the specic powers of the electrical machines (motors and generators) and of the power electronics (rectiers, inverters and converters), as well as the power conversion eciencies of all the components. This section will provide a brief summary of the current state of the art of the above-mentioned technologies, as well as looking into some of their potential future developments. This section will draw in part on an earlier summary, as presented in Kruger et al. (2018) and D. Hall et al. (2019). 2.6.1 Batteries Currently, the greatest limitation for electric aircraft is the specic energy (typically mea- sured in Wh/kg) of batteries, compared to that of hydrocarbon fuels. State of the art lithium-ion (li-ion) cells store around 250 Wh/kg, and after packaged into a battery pack, store around 175 Wh/kg (Kruger et al., 2018). Compared to Kerosene, which has a specic energy of around 12 000 Wh/kg, batteries store almost 70 times less usable energy per unit weight. As given by the Breguet range equation, the range an aircraft can y is directly proportional to its fuel source specic energy (Raymer, 2012), and to rst order this dier- ence represents a factor of 70 reduction in range for an all-electric aircraft compared to a similarly sized conventional aircraft. 26 Much research is currently being conducted into various technologies that could signif- icantly increase battery specic energy (BSE), and good summaries of the current state of the art, as well as existing research eorts and targets have been presented by Misra (2017a, 2017b, 2018). The focus of most of the work is still on lithium chemistries, but with various dierent materials used in the anode, cathode, and as electrolytes. Current investments are focussing on various chemistries that use nickel manganese cobalt in the cathodes, and various dierent materials, such as graphite, silicon-carbide, silicon and lithium-metal in the anode. These investments seem to have secured a path towards a pack level BSE of up to around 300 Wh/kg, although the exact time-line is uncertain. A chemistry using lithium- metal in the anode and sulfur in the cathode is also being researched, and could lead to pack level BSE values of almost 500 Wh/kg, but the path towards commercialization is much less certain. Another technology that could bring signicant improvements is so-called solid- state batteries, that use solid electrolytes as opposed to the liquid or polymer gel electrolytes currently used by lithium chemistries. These cells have the benet of higher packaging ef- ciencies going from cell to pack, and are much safer than conventional li-ion cells. The research into solid-state batteries is lagging behind the chemistries listed above, but could be feasible in the 2030 time frame. Pack-level BSE's above 500 Wh/kg will only be possible through more novel chemistries, such as lithium-air, advanced solid-state lithium-sulfur, or ow batteries. These are more reminiscent of fuel cells than conventional batteries. In terms of ultimate potential, lithium-air batteries are the most promising, and based on the redox reaction alone has a theoretical specic energy of almost 11 000 Wh/kg. A more realistic pack-level value that takes into account the mass increase of the battery with discharge is 900 Wh/kveg (D. Hall et al., 2019; Kruger et al., 2018). Another concept that could lead to signicant system-level benets is the use of struc- tural batteries, where aircraft structural elements are constructed from batteries. This will combine the aircraft structural weight and battery weight, allowing more energy to be stored. This has obvious challenges from an aircraft design perspective, as well as maintenance im- 27 plications, since current batteries have nite cycle-lives and will need to be replaced after a certain amount of charge-discharge cycles, which could be dicult if the batteries are incorporated into the aircraft structure. Apart from specic energy, specic power (typically measured in kW/kg) is also a major parameter of interest for batteries, and depending on the application might be the limiting factor. The relationship between battery specic energy and specic power is know as the Ragone relation, and it implies that there is a trade-o between the power required from a battery and the amount of useful energy it can provide (Brodd, 2012). The usable energy available from a battery can be modeled through the battery eciency and the battery power draw compared to the total power available. A model like this is used in the current work, and will be discussed in Chapter 3, where the component models are presented in more detail. Lastly, a battery will also require space, or volume, to be installed in an aircraft, and unlike fuel that can conform to the shape of its container, batteries might be subject to more stringent packaging constraints. To rst order and for conceptual design purposes, battery volume can be estimated through an assumed specic volume, and then calculated once the total required battery energy is known. Little research is available to guide the value to pick for the specic volume of batteries, and in this work a value of 500 Wh/` is used 3 . 2.6.2 Electrical Machines Electrical machines, which include motors and generators, form an important part of elec- trical propulsion system architectures, and it is important that their performance be well understood for accurate conceptual design estimates. The parameters of interest are the machines' specic powers, and their electrical conversion eciencies. Power ratings are usu- ally given as maximum continuous power and peak power that can only be sustained for a 3 This value is obtained from the data sheet of the Panasonic NCR18650GA cell, which has a cell specic volume of 693 Wh/` (Panasonic, 2018), and is reduced by 75% and rounded down to give a conservative estimate of packaging density as recommended by Byahut (2021). 28 limited time. The higher peak powers are a result of the thermal mass of the system, which allows the components to heat up temporarily without getting damaged. Misra (2017; 2017) has also presented some work on the state of the art and potential developments of electrical machines, which mainly cites the work of Jansen et al. (2017). Ongoing projects were touched upon in Section 2.4.5, and include research into induction, permanent magnet and wound eld machines. Commercial electrical machines designed specically for aviation are currently sold by companies such as Rolls-Royce and MagniX, and have specic power ratings of around 5 kW/kg. The projects mentioned above are looking at signicantly improving this number within the 2035 time frame, and values as high as 16 kW/kg might be reached for machines rated to almost 3 MW with 99% eciency (Jansen et al., 2017). If superconducting technologies are used, values approaching 25 kW/kg could be attained (Misra, 2017a; Misra, 2017b), but this might be a longer time o and would require auxiliary systems such as cryo-coolers that would add additional system weight and complexity. The values listed above should be compared with conventional gas-turbine engines, which have specic powers of around 4 kW/kg for engines in the 1 MW power range (Raymer, 2012). 2.6.3 Power Electronics Power electronics components such as inverters (converting DC to AC), rectiers (converting AC to DC) and converters (converting AC to AC or DC to DC) will also form an important part of electrical propulsion system architectures, and their improvement depends on the improvement of capacitors, magnetics and semiconductors (Misra, 2017a; Misra, 2017b). As also discussed by Misra, potential improvements are driven primarily by wide band gap semiconductors such as silicon carbide and gallium nitride, which allow higher frequency of operation and would thus reduce the size requirements of capacitors and inductors, also allowing for higher operating temperatures. As mentioned in Section 2.4.5, ongoing research 29 might see specic power values for power electronics reach values as high as 19 kW/kg in the 2035 time frame with eciencies as high as 99% (Jansen et al., 2017). 2.6.4 Thermal Management Although conventional engines that burn fuel generate a lot of waste heat, they have the ad- vantage of shedding most of it directly into the freestream with the exhaust gases. Electrical systems don't have this advantage and the waste heat from electrical components needs to be dealt with more directly via a thermal management system (TMS). Thermal management systems are essentially heat exchangers that remove heat from components to avoid over- heating. Even though electrical systems are very ecient, the waste heat from high power level systems could represent hundreds of kW of waste heat that needs to be managed. For example, a 20MW electrical system (which is what would be required for large electric air- craft) operating at 99% eciency will generate 200 kW of heat due to the 1% ineciency. At 98% eciency, it will generate 400 kW of heat, which would, to rst order, require a twice as powerful (and heavy) TMS. There is, unfortunately, not much research available to help with the modeling of TMS's for conceptual design, so a lumped parameter model is used in the current work, which sums the waste heat of all electrical components and uses an assumed specic power to nd the weight of the TMS. More details will be given in Chapter 3. 2.7 Other Paths Towards Improved Eciency Electric propulsion is only one potential avenue towards improving the eciency of the global air transport system. Without moving away from internal combustion engines, using alter- native fuels could also lead to signicant improvements in aircraft eciency. Although not explicitly part of the scope of the current work, some considerations regarding alternative fuels will be mentioned here. Alternative fuels can conceptually be divided into two groups: 30 (1) drop-in alternative fuels that can be used with existing engine technology (possibly with some minor modications to the engines), and (2) alternative fuels that require dierent en- gine technologies. An in-depth review of the former category is given by Blakey et al. (2011). A concise but practical review of the latter category is given by Raymer (2012). Primary drop-in alternative fuels include biofuels, liqueed natural gas (LNG) primarily consisting of methane, and liquid hydrogen. Biofuels are created directly from biomass, and so are considered a renewable energy source, but they produce essentially the same emissions as kerosene, the primary fuel used in aviation today. LNG is abundant in the Earth's crust and might thus be more economically attractive, but it is not renewable and also generates carbon dioxide (CO2) during combustion. Hydrogen can be used as both a drop-in fuel and as a means to power alternative propulsion systems like fuel cells. Hydrogen has a very high specic energy, and also has the advantage of being renewable and clean when combusted. It does, however, produce oxides of nitrogen, or NOx, when oxidized with atmospheric air. Since it does not occur naturally on Earth, it has to be synthesized. The process of steam reforming natural gas to produce hydrogen, which is currently the most common and cheapest method of synthesis (Santhanam et al., 2017), has a CO2 byproduct of between 9 and 12 tons for every ton of hydrogen produced (Collodi, 2010), and requires high temperature steam as an input, which also requires energy to produce. Unless the production of hydrogen can be done eciently and the CO2 captured during production, its use in aviation could be counterproductive. Another challenge of designing aircraft to use hydrogen is its low volumetric energy density, even when stored under high pressure (Hepperle, 2012). When the higher specic energy is taken into account, a given amount of compressed hydrogen that stores the same amount of energy as a given amount of kerosene will have around four times the volume of the kerosene (Raymer, 2012). This necessitates novel aircraft design solutions, which would likely reduce the eciency at an aircraft level compared to an aircraft designed for the same mission but that uses kerosene. In particular, the requirement for high pressure tanks and cryogenic 31 Figure 2.7: Airbus Cryoplane concept. Compressed hydrogen tanks stored in bulbous ap- pendage on top of fuselage. Figure from Klug (2000). storage systems would likely be required and would add additional mass and wetted surface. An example of such a design compromise, the Cryoplane, is shown in Figure 2.7. Another possible advantage of hydrogen, when used in a fuel cell system, is that it would also benet from the high specic powers and eciencies of electrical propulsion components, since fuel cells produce electricity. In this sense, hydrogen is relevant to the current work since the batteries assumed here could be exchanged for hydrogen and fuel cells used to power distributed electric propulsors, although that was not part of the scope. 32 Chapter 3 Methodology 3.1 Framework Overview The LUCAS framework adds advanced models (including the power balance method, mod- eling of BLI, and a unied propulsion system model) into and functions as a wrapper around the Stanford University Aerospace Vehicle Environment (SUAVE) (Lukaczyk et al., 2015), which is a Python-based framework for the analysis and optimization of aircraft. At its core SUAVE performs mission simulation, and has been wrapped within an optimizer to allow for sizing and optimization (Botero et al., 2016). Although SUAVE is a fully function- ing framework on its own, LUCAS provides an interface that parses input and output les and performs post-processing on SUAVE analyses to provide a smoother experience when working with large sets of input data. Figure 3.1 shows the basic layout of LUCAS and the data ow between LUCAS, SUAVE and the optimization wrapper. This is visualized using an Extended Design Structure Matrix (XDSM), as introduced by Lambe and Martins (2012). 33 x guess Geometry Input Initial Guesses Flight Profile Initial Guesses x ∗ PyOptSparse with SNOPT Analysis Design Variables Mission Design Variables PB Design Variables y ∗ 1 Geometry y 1 y ∗ 2 Discipline Analyses y 2 Mission y ∗ 3 R Mission y 3 Power Balance y ∗ 4 R Power Balance df/dx, dc/dx Gradients Figure 3.1: Extended Design Structure Matrix (XDSM) (Lambe and Martins, 2012) for LUCAS, showing connections and data ow between subsystems. 34 In the semantics of XDSM diagrams, dierent elements have the following meanings: Output parameters are listed in parallelogram boxes to the left Input parameters are listed in parallelogram boxes at the top Optimizers are represented by blue rounded boxes Functions are represented by green rectangular boxes Solvers are represented by orange rounded boxes Thick gray bands represent data ow between components Thin black lines represent processes Stacked components represent processes that can be executed in parallel As can be seen in Figure 3.1, at the highest level is the pyOptSparse Python-based op- timization framework (Wu et al., 2020), which in this work uses the SNOPT (Gill et al., 2005) non-linear quadratic programming optimizer. Note that dierent optimizers can be used with pyOptSparse, but that SNOPT was found to work well with SUAVE (Wendor et al., 2016). The pyOptSparse solver takes as an input the initial guesses of all the prob- lem design variables, x guess , and acts as a driver that sequentially runs the SUAVE analysis until the optimization problem converges. SUAVE is set up in this work to use the opti- mizer decomposition approach by Kroo et al. (1994) for sizing, meaning that sizing is done by implementing sizing constraints, and the optimizer enforces these constraints along with performance constraints to size the vehicle. An optimizer iteration starts with the parsing of user inputs, including baseline aircraft geometry. SUAVE is then used to perform various discipline analyses and simulate the aircraft performing its predened ight mission. The mission solver is iterative, and thus functions as a subprocess between the mission process (orange block) and its function analysis (green block). In this case SciPy (specically scipy.optimize.root) is used to drive the mission residuals to zero, using the hybr algorithm from MINPACK-1 (Mor e et al., 1980). A key addition to SUAVE introduced by LUCAS is the use of the power balance method, which is also solved iteratively using scipy.optimize.root. Once the power balance and 35 then the mission solution processes have converged, a cost function can be computed within SUAVE, and pyOptSparse can compute cost and constraint gradients, which are used to determine a step direction for the next iteration. Note that the `Gradients' function and outputs are shown as stacked boxes, implying that gradients can be computed in parallel, which reduces computational time. This process repeats until user-specied residual limits on problem feasibility and optimality are met. A detailed description of the discipline analyses, including aero-propulsive modeling and weight buildup, is the focus of the remainder of this chapter. 3.2 Aero-Propulsive Coupling The LUCAS framework is designed to capture interactions between the airframe aerody- namics and the propulsion system. As such, it uses the power balance method developed by Drela (2009). As was mentioned in Section 2.5.2, this method models mechanical power and kinetic energy ow instead of the force and momentum ow used in the more traditional force balance approach. The aero-propulsive problem can thus be viewed as a balance of power and dissipation (rather than thrust and drag). By doing so, the power and dissipation of each sub-component remains clearly dened and unambiguous in the presence of strong airframe-propulsor coupling as with BLI, where the denitions of thrust and drag become unclear. We follow the approach introduced by David K. Hall et al. (2017) and Uranga et al. (2018), via the relations presented next. 3.2.1 Flow Power The power P K added to the ow by a propulsor is given by P K = 1 2 (V 2 jet V 2 1 ) _ m +f BLI 0 surf ; (3.1) 36 where V jet is the propulsive jet velocity, V 1 is the freestream velocity, and the mass ow through the propulsor is given by _ m = fan V fan A fan (3.2) via the density at the fan face fan , the fan face velocity V fan and the fan area A fan . For an array of propulsors, the mass ow can be aggregated over all the propulsors, all having the same fan area, as _ m tot = X propulsors 1 to N fan fan V fan A fan ; (3.3) where N fan is the total number of propulsors. Note that `fan' here refers generally to a propulsor which can either be a propeller, a ducted fan or an unducted fan. In the current work, a distinction is made between mechanically and electrically driven propulsors, where for each type the general geometrical design and fan area is xed, independent of the number of propulsors. As such, the total mass ow across all propulsors can be computed as _ m tot =N fan;M fan;M V fan;M A fan;M +N fan;E fan;E V fan;E A fan;E ; (3.4) where N fan;M and N fan;E are the total number of mechanical and electrical propulsors, re- spectively. The method for determiningV fan depends on the type of propulsor. For a ducted fan, it is assumed that ow straightening vanes would be used to remove any swirl induced by the fan, in other words we assume a rotor + stator fan stage. For such fans, the classical solution from actuator disk theory (McCormick, 1979) can be used to compute the fan velocity as V fan no swirl = V 1 +V jet 2 : (3.5) It is assumed here that both conventional turbofan style engines, as well as electrical propul- sors used on larger aircraft classes that y at high subsonic speeds will make use of ducted 37 fans. Smaller aircraft classes typically use unducted turboprop engines and it is assumed here that electrical propulsors used on such aircraft will also be unducted. In order to determine V fan for such propulsors, actuator disk theory is adjusted to account for the swirl introduced into the ow, as follows. Figure 3.2 shows the control volume around the fan, which is here modeled as an actuator disk. Bernoulli's equation can be applied between the freestream and a station just upstream of the fan, as well as between just aft of the fan and the jet exit, giving p 1 + 1 2 1 V 2 1 =p 1 + 1 2 1 V 2 1 () p 1 =p 1 + 1 2 1 V 2 1 V 2 1 ; (3.6) p 2 + 1 2 1 V 2 2 =p j + 1 2 1 V 2 jet () p 2 =p 1 + 1 2 1 V 2 jet V 2 2 : (3.7) Subtracting (3.6) from (3.7) gives p 2 p 1 = 1 2 1 (V 2 jet V 2 2 V 2 1 +V 2 1 ): (3.8) The classic actuator disc formulation is now extended to account for radial swirl added to the ow by the rotating fan blades. Because this radial ow does not provide useful ow Freestream Jet Fan (actuator disk) A 1 V 1 p 1 1 2 e Thrust V jet A jet p jet = p 1 CV f , 1 Figure 3.2: Control volume around fan modeled as actuator disk with swirl. 38 power, but still requires energy input, it has to be accounted for. A ratio for radial to axial ow velocity V =V x is assumed, which is taken to be constant from station 2 downstream to the exit. The velocity magnitude at any axial point downstream of the fan can then be written as V 2 k 2 s V 2 x ; (3.9) where V is the velocity magnitude, k 2 s is a parameter calculated via V =V x , and V x is the axial velocity. The terms are squared to ensure positive magnitudes. Then from Pythagoras: V 2 x +V 2 k 2 s V 2 x ; (3.10) and k 2 s can be solved as k 2 s = 1 + V V x 2 : (3.11) By conservation of mass, the axial velocity does not change over the fan, and thusV x;1 =V x;2 if density changes across the fan are neglected. Since V fan is the axial fan face velocity, V x;1 =V fan , V fan =V x;2 , and from (3.9) it can be seen that V fan = 1 k s V 2 () V 2 =k s V fan (3.12) which can be used to rewrite (3.8) as p 2 p 1 = 1 2 1 V 2 jet (k 2 s 1)V 2 fan V 2 1 : (3.13) Fan thrust can be written equally as T =A fan (p 2 p 1 ) and T = _ m(V x;j V 1 ); (3.14) and thus _ m(V x;j V 1 ) =A fan (p 2 p 1 ); (3.15) 39 and by substituting (3.2), (3.9) and (3.13) this can be written as 1 V fan A fan ( 1 k s V jet V 1 ) =A fan 1 2 1 V 2 jet (k 2 s 1)V 2 fan V 2 1 : (3.16) This relation can be simplied to get V fan 1 k s V jet V 1 = 1 2 V 2 jet (k 2 s 1)V 2 fan V 2 1 ; (3.17) which amounts to an implicit equation for computing the fan velocity. Noting that this is a quadratic equation inV fan , the fan velocity can be solved for using the quadratic formula as V fan = V 1 V jet =k s + q V 2 jet =k 2 s V jet V 1 =k s +V 2 1 (k 2 s 1)(V 2 1 V 2 jet ) k 2 s 1 ; (3.18) using the positive root solution since the negative root would not satisfy the relation V 1 < V fan <V jet . This equation is used in the current work, with a value of V =V x = 0:4, and thus k 2 s = 1:16. The value of 0.4 is chosen in order to assume that every unit of axial (useful) velocity comes with 0.4 units of swirl. The total power required to produce a unit of axial velocity is thus 1.4 times the power needed if there were no swirl. Note that (3.18) is valid for open propellers that don't have any ow correcting stator vanes. In the case of a ducted propulsors that do have correcting vanes (stators), (3.5) is used instead. The second term in (3.1), f BLI 0 surf , is the total surface dissipation in the boundary layer that is ingested by the propulsor, and amounts to the aerodynamic benet of BLI following Uranga et al. (2018). It is determined by the fraction of the total kinetic energy defect ingested and re-energized by the BLI propulsors, f BLI , and via the surface dissipation evaluated as 0 surf = (1f wake )D 0 p V 1 ; (3.19) where D 0 p is the prole drag acting on an equivalent non-BLI aircraft, and f wake is the fraction of the airframe's non-vortex dissipation (or equivalently fraction of prole+wave 40 drag) occurring in the wake as opposed to along the surface. In this work, we assume a value of f wake = 0:1 as is typical of attached ows over aerodynamic bodies, and determine D 0 p from the classical component drag buildup in SUAVE. For more detailed denitions of the variables used here, the original source by Uranga et al. (2018) should be consulted. Inn the absence of BLI, the mechanical ow P K is equivalent to the change in kinetic energy across the propulsor that more classical approaches use (Torenbeek, 2013). 3.2.2 Compressibility Correction When the ow through the propulsors reaches a suciently high Mach number, the ow diuses before reaching the fan face at M fan = 0:6. Given the known quantities M 1 , p 1 , 1 , M fan , p 0;fan =p 0;1 and T 0;fan =T 0;1 , the ratio of fan to freestream pressure and density can be calculated as p fan p 1 = p fan p 0;fan p 0;fan p 0;1 p 0;1 p 1 = p 0;fan p 0;1 p 0;1 =p 1 p 0;fan =p fan = p 0;fan p 0;1 " 1 + 1 2 M 2 1 1 + 1 2 M 2 fan # 1 (3.20) and fan 1 = fan 0;fan 0;fan 0;1 0;1 1 = 0;fan 0;1 0;1 = 1 0;fan = fan = 0;fan 0;1 " 1 + 1 2 M 2 1 1 + 1 2 M 2 fan # 1 1 (3.21) after using the isentropic relations between stagnation and static pressure (Anderson, 1982) p 0 p 1 = 0 1 = 1 + 1 2 M 2 : (3.22) Note that the term p 0;fan =p 0;1 captures the inlet losses. Next, using the equation of state at stagnation, we can write 0;fan 0;1 = p 0;fan =RT 0;fan p 0;1 =RT 0;1 = p 0;fan =p 0;1 T 0;fan =T 0;1 : (3.23) 41 The speed of sound at the fan can then be computed directly as a fan = r p fan fan = r 1 fan 1 1 p fan p 1 p 1 = s p fan =p 1 fan = 1 p 1 1 ; (3.24) thanks to equations (3.20) and (3.21). Knowing a fan , the fan face mach number can be calculated as M fan = V fan a fan : (3.25) In this work, if it is found that M fan > 0:6, its value is constrained to 0.6, which is a typical value to ensure ecient fan operation (Greitzer et al., 2010). 3.2.3 Net Forces As shown in Uranga et al. (2018), the net streamwise force acting on the aircraft, F X , is given by F X = D 0 f BLI D 0 p surf =V 1 (V jet V 1 ) _ m + _ h W V 1 : (3.26) Here D 0 is the total drag force experienced by an equivalent non-BLI aircraft, surf is the change in surface dissipation due to the installation of the BLI propulsor (assumed to be negligible in this work), _ h is the rate of climb and W is the aircraft weight. The last term in (3.26) was not included in Uranga et al. (2018) who only considered level ight, but was included in the original formulation by Drela (2009). In the transverse direction, the net force is F Z = L dR dt W V 1 ; (3.27) where L is the lift force which is assumed to be unchanged by BLI (as justied by Uranga et al. (2018)) and dR=dt is the horizontal speed (rate of range change). For steady ight, F X and F Z are zero. Equations (3.1){(3.27) are assembled into a system of equations together with dissipa- tion buildup quantities, and the system is solved numerically. The variables that are actually 42 solved for are P K;M , P K;E , _ m M , _ m E , V jet;M and V jet;E . 3.2.4 Dissipation Buildup For an aircraft with no BLI, the surface dissipation is directly related to the viscous drag acting on its surface, with dissipation being equal to drag times ight velocity,D 0 V 1 . It can thus be calculated using a standard friction and form drag buildup for conceptual design, like the classical method presented by Raymer (2012). Such a method is already implemented in SUAVE, and it is used here as is. The only exception is the drag caused by the propulsors, which was added to the framework as described in the next paragraph. For an aircraft with BLI, the surface dissipation can still be estimated in this manner, and the eect of BLI modeled using the power balance equations previously described. Propulsor drag is calculated using the methods described in Raymer (2012). It is as- sumed that the mechanically powered propulsors are either turbofan engines fully immersed in the air ow, or turboprop engines attached directly to the wing with only the turbine cores covered with fairings. Electric propulsors can be either ducted or unducted. In the former case, they look like turbofans, in that the motor and fan is shielded in a pod. Unducted propulsors have open rotors with the electrical motors covered with fairings that sit in the air stream. The LUCAS framework models the electrical propulsors as either located in fairings along the leading edge of the wing, or distributed along the wing's trailing edge, as will be discussed in Section 3.3.4. 3.2.5 Eciencies In the power balance method, the propulsive eciency can be calculated as propulsive = P K jet P K ; (3.28) 43 where jet is the total dissipation occurring in the propulsive jet, given by Uranga et al. (2018) as jet = 1 2 (V jet V 1 ) 2 _ m: (3.29) In the absence of BLI and swirl, (3.28) reduces to the inviscid Froude eciency i = 2=(1 +V jet =V 1 ). The chain of eciencies needed to go from the total shaft power P fan to the total ow power P K can be written as P K = fan;v P fan ; (3.30) in which fan;v accounts for the viscous prole drag on the fan blades. A value of fan;v = 0:9 is assumed in the current work (Mattingly et al., 2000). Depending on the architecture, the total shaft power can be supplied by the electric motors (with power P fan;E ), and the turbine and/or generator shaft (with power P fan;M ), which are respectively linked to the electrically-driven and the mechanically-driven propulsor fans: P fan =P fan;M +P fan;E . An aircraft can have multiple propulsors and thus multiple propulsive streams, The above equations are implemented separately for the mechanical and electrical streams. The total mass ow of each stream is then divided among the propulsors composing that stream. Thus, there is one overall mass ow, jet velocity and propulsive eciency associated with the mechanical stream, and another set of those parameters for the electrical stream. 3.3 Propulsion System Modeling 3.3.1 Unied Model of Propulsion System Architectures In order to determine the suitability of aircraft electrication, we employ a propulsion system model that represents the various types of propulsion system architectures|from conven- tional (internal combustion) systems to all-electric, hybrid-electric, and turbo-electric sys- tems, including all the serial and parallel variations|via two electrication design variables. 44 Mechanically-Powered Propulsors Gas Turbines N fan,E N fan,M N turb bat Electrically-Powered Propulsors turb fan M P K,M P K,E Mechanical Source P turb P bat Electrical Source Battery System Source (Consumed Power) Load (Useful Power) Mechanical Load Electrical Load gen= inv=rect mot P inv P mot P fan,E mot inv fan E P fan,M P gen=mot P inv=rect Link P link Figure 3.3: Unied propulsion system model (Kruger et al., 2018). The subscriptsM andE refer to mechanical and electrical components, respectively, and the powers P indicate the power ow between components. It is thus able to treat the dierent architectures in a unied manner, rather than as separate instances. This unied propulsion system model was introduced in previous works (D. Hall et al., 2019; Kruger et al., 2018), and originally developed as part of an MIT-USC-Aurora- NASA LEARN3 program 1 It is implemented here within SUAVE as part of the LUCAS additions. A schematic diagram of the unied model is shown in Figure 3.3. The left-hand side shows the possible sources of power: mechanical sources in the form of gas turbines, and a battery system as electrical source. Power ows from these sources to the load side shown on the right, which is composed of mechanically- and electrically-powered propulsors. These propulsors in turn transfer useful propulsive power to the ow in the form of ow power P K . Electrication of the propulsion system is uniquely dened based on two distinct elec- 1 Final report published December 2019 under NASA/CR|2019-220382. 45 trication variables: the source electrication factor f S P bat P bat +P turb ; (3.31) and the load electrication factor f L P K;E P K;E +P K;M : (3.32) The source electrication factor, f S 2[0; 1], denes the split between power coming from a chemical battery system and that coming from gas turbines, P bat and P turb , respectively. The load electrication factor, f L 2[0; 1], denes the split between ow power imparted by the electrical and mechanical propulsive branches, P K;E and P K;M , respectively. Here P K;E and P K;M amount to the total ow power from all the propulsors in each branch. Thus, with N fan;M mechanically-driven propulsors, each of those propulsors supplies a ow power of P K;M =N fan;M , while each electrically-driven propulsor supplies P K;E =N fan;E ow power. The mechanical and electrical subsystems shown as the top and bottom branches in Figure 3.3 can be connected via a power conversion device, or link, that functions as ei- ther a motor or a generator system. This would typically be comprised of one or more motors/generators coupled to a mechanical shaft, and so there are several such couplings inside the link, one for each coupled component (propulsor and/or turbine depending on the architecture). When power is sent from the electrical branch towards the mechanical branch, with power owing from the lower branch to the upper branch in Figure 3.3, the link is a motor and P link > 0. Conversely, when the power direction is reversed, P link < 0 and the link becomes a generator. Each component in the unied propulsion system is modeled as having input power and an output power, the ratio of which is the component's eciency. Following the power ow 46 directions indicated in Figure 3.3 from load to source, we thus have P K;M = fan nacelle P fan;M P K;E = fan nacelle P fan;E P fan;M = P turb + EM PE P link P fan;E = EM P mot P mot = PE P inv (3.33) Note that a propulsor is characterized by its fan and nacelle eciencies, fan and nacelle . The term fan here should be understood to be general, and include both ducted fans and propellers. If the propulsor is unducted, then nacelle = 1. Power electronics components (whether inverters or rectiers) are all set to have an eciency PE , and electrical machines (motors or generators) an eciency EM . All powers are positive, except for the link power P link which can be positive or negative. If the power output from a component is zero, that component is not present in the architecture. From the output of the battery, power is sent to inverters towards the electrical propul- sors and/or the link. Thus, just as current is split at the node of an electric circuit, we have P bat = P inv +P link ; (3.34) which holds whether the link power is positive or negative. At the node in the mechanical branch, whenP link >0 the power output by the link, PE EM P link , is added to the power out- put by the turbine,P turb , before it is sent to the mechanically-driven propulsors. Conversely, when P link < 0, the link receives the excess power from the turbine that is not sent to the mechanically-driven propulsors. Thus, 8 > > < > > : P turb + PE EM P link =P fan;M when P link >0 ; PE EM (P turb P fan;M ) =P link when P link <0 : (3.35) 47 Equation (3.34) is an expression of the power balance at the node in the electrical branch (bottom node in Figure 3.3), while the power balance at the mechanical branch's node is expressed by (3.35) and depends on the direction of the power ow in the link. Equations (3.33){(3.35) can be cast as a linear system for the unknown component powers, using as parameters the total required ight ow power P K;tot , the electrication factors f L andf S , and a set of component eciencies. Values for the component eciencies are set based on the technology assumptions that will be presented in Section 3.6. The linear system in matrix form is given in Appendix A. 3.3.2 Electrication Design Space Each choice of the pairf S andf L denes a unique propulsion system, and the unied model can thus model any of the dierent propulsion systems architectures, whether conventional or electried. The power required by each component in the propulsion system is uniquely determined by the total required ow powerP K;tot =P K;M +P K;E , the values off S andf L and the component eciencies. Conversely, given an aircraft-level ow power requirement P K;tot and a load electrication factor f L , the ow power from mechanically-driven propulsors is P K;M =(1f L )P K;tot and that from electrically-driven propulsors is P K;E =f L P K;tot . Figure 3.4 shows the f S {f L design space. Conventionally powered aircraft that use internal combustion engines correspond to f S =f L = 0. As f S increases, the fraction of source power provided by the battery increases, and as f L increases, the fraction of load power provided by electrically powered propulsors increases. The dashed line in the gure corresponds to P link = 0 and divides the region of hybrid-electric architectures: below it are parallel hybrid congurations for which the link acts as a motor and the battery boosts the power to the mechanically-powered propulsors. Above the line, the link is a generator and the architecture is series hybrid. Along the vertical axis withf S =0 and 0<f L <1 are partial turbo-electric congurations, with all energy stored in fuel but ow power being provided by both conventionally powered propulsors and electric motors. With fully turbo-electric 48 f S f L Parallel hybrid Series hybrid 1 1 0 0 P link = 0 Conventional Fully turbo-electric Partial turbo-electric All-electric Figure 3.4: Schematic of f S {f L design space showing regions for the dierent propulsion system architectures. systems (f S = 0 and f L = 1) the source power comes from one or several turbo-generators and is passed through the link so that all the ow power is imparted by electrically-powered fans. When all the energy is stored in a battery system, f S = 1 and the conguration is all- electric. In this case, all values off L result in the same architecture: the link just completes the inverter-motor-fan system to form an extra propulsor set, and all the propulsors are powered electrically. However, f L can be used in this framework with an f S 6= 1 to provide some exibility as to the placement of the propulsors. Specically, we assume that all the N fan;E electrically-driven propulsors are distributed and perform BLI (being integrated into the wing), but that theN fan;M are placed in freestream ow, so we can adjust f L in order to model dierent propulsor arrangements even though they are all driven by electric power. Figure 3.5 shows the distinct architectures that can be modeled using the unied model, along with their associated electrication parameters. The key advantage, however, is that the design space can be modeled in a continuous sense as visualized in Figure 3.4. It is important to note that, for a givenf S andf L , the direction of the power ow in the 49 Mechanically-Powered Propulsors Gas Turbines N fan,E N fan,M N turb bat Electrically-Powered Propulsors turb fan M P K,M P K,E Mechanical Source P turb P bat Electrical Source Battery System Source (Consumed Power) Mechanical Load Electrical Load gen= inv=rect mot P inv P mot P fan,E mot fan E P fan,M P gen=mot P inv=rect Link P link inv Load (Useful Power) (a) Conventional, f S = 0;f L = 0 Mechanically-Powered Propulsors Gas Turbines N fan,E N fan,M N turb bat Electrically-Powered Propulsors turb fan M P K,M P K,E Mechanical Source P turb P bat Electrical Source Battery System Source (Consumed Power) Mechanical Load Electrical Load gen= inv=rect mot P inv P mot P fan,E mot fan E P fan,M P gen=mot P inv=rect Link P link inv Load (Useful Power) (b) All-Electric, f S = 1 Mechanically-Powered Propulsors Gas Turbines N fan,E N fan,M N turb bat Electrically-Powered Propulsors turb fan M P K,M P K,E Mechanical Source P turb P bat Electrical Source Battery System Source (Consumed Power) Mechanical Load Electrical Load gen= inv=rect mot P inv P mot P fan,E mot fan E P fan,M P gen=mot P inv=rect Link P link inv Load (Useful Power) (c) Parallel Hybrid, 0<f S < 1;f L = 0 Mechanically-Powered Propulsors Gas Turbines N fan,E N fan,M N turb bat Electrically-Powered Propulsors turb fan M P K,M P K,E Mechanical Source P turb P bat Electrical Source Battery System Source (Consumed Power) Mechanical Load Electrical Load gen= inv=rect mot P inv P mot P fan,E mot fan E P fan,M P gen=mot P inv=rect Link P link inv Load (Useful Power) (d) Series Hybrid, 0<f S < 1;f L = 1 Mechanically-Powered Propulsors Gas Turbines N fan,E N fan,M N turb bat Electrically-Powered Propulsors turb fan M P K,M P K,E Mechanical Source P turb P bat Electrical Source Battery System Source (Consumed Power) Mechanical Load Electrical Load gen= inv=rect mot P inv P mot P fan,E mot fan E P fan,M P gen=mot P inv=rect Link P link inv Load (Useful Power) (e) Partial Turbo-Electric, f S = 0; 0<f L < 1 Mechanically-Powered Propulsors Gas Turbines N fan,E N fan,M N turb bat Electrically-Powered Propulsors turb fan M P K,M P K,E Mechanical Source P turb P bat Electrical Source Battery System Source (Consumed Power) Mechanical Load Electrical Load inv=rect mot P inv P mot P fan,E mot fan E P fan,M P gen=mot P inv=rect Link P link inv Load (Useful Power) gen= (f) Fully Turbo-Electric, f S = 0;f L = 1 Figure 3.5: Unique propulsion system architectures that can be modeled with unied model by adjusting f S and f L . 50 link (and thus the kinds of feasible architectures) is constrained by the eciency values for power electronics and electrical machines, more specically the product PE EM . This can be seen by combining the power split at the two nodes expressed in (3.34) and (3.35), and making use of (3.31) and (3.32), to show that in order to have a parallel hybrid architecture (P link >0) it is necessary that EM PE f S (1f L ) > (1f S )f L : (parallel hybrid) (3.36) If, however EM PE f S (1f L ) < (1f S )f L ; (series hybrid or turbo-electric) (3.37) only series hybrid and turbo-electric architectures are possible (P link <0) since battery power will not suce to power all the electrically-driven propulsors. 3.3.3 Propulsion System Sizing Complementing the power-based unied propulsion system model is a mass buildup. The overall propulsion system mass is equal to the sum of its component masses, so that the total system mass is m prop;tot = m prop;prim +m prop;sec +m prop;trans : (3.38) The terms on the right hand side of (3.38) are the bare mass of primary components m prop;prim , the additional mass of secondary components m prop;sec and the mass of electri- cal transmission lines m prop;trans . The primary component masses are calculated as m prop;prim = (N fan;M m prop;fan;M )+(N fan;E m prop;fan;E )+(N turb m prop;turb )+m TMS ; (3.39) whereN fan;M andN fan;E are the number of mechanical and electrical propulsors, respectively, 51 and N turb is the number of turbine cores in the system. The terms m prop;fan;M , m prop;fan;E and m prop;turb include component masses that scale directly with the number of mechanical and electrical fans and turbines in the system. Lastly, m TMS is the mass of the thermal management system. All components are sized based on power and if a component is deter- mined to use no power for a specic architecture, the component mass becomes zero and it is eectively removed from the architecture. The subsystem masses in (3.39) are calculated as m prop;fan;M =m fan;M ; (3.40) m prop;fan;E =m fan;E +m mot +m mot;fair +m inv ; (3.41) m prop;turb =m turb +m gen=mot +m inv=rect +m turb;fair : (3.42) In these equations, the term fan is used as a general term that could refer to either a mechanically or electrically powered fan or propeller. In (3.40) the fan mass is given per Torenbeek (1982) as m fan = 0:108 D propeller P propeller p N blades 0:782 ; (3.43) where the fan diameterD is specied in feet and powerP in horsepower, the resulting mass is in pounds, and N blades is the number of blades on the propeller. The electrical motor mass is calculated simply as its peak power multiplied by a specic power as m mot =P mot P m 1 mot ; (3.44) where [ P m ] mot is an assumed specic power for the motor in W/kg. This value depends on the assumed technology level and will be given later. These motors are covered by fairings, and estimates of the fairing and motor dimensions are required. The diameter and length of 52 the electric motor is estimated as D mot =k 1 (P mot ) 1=3 ; (3.45) ` mot =k 2 D mot : (3.46) These scaling laws are given by Krishnan (2017) and suitable constants are k 1 = 0:0065 m=W 1=3 and k 2 = 0:72, as estimated from the Siemens SP260D motor 2 . The motor fairing diameter is assumed to be 1:2D mot , and the length 1:2` mot . The fairing mass is then estimated as m mot;fair = 0:6724K ng ` 0:10 fair D 0:294 fair N 0:119 z m 0:661 mot S 0:224 fair : (3.47) Input values are in English units and the result is given in pounds. Here K ng is 1.017 for a pylon-mounted fairing, and 1.0 otherwise, ` fair and D fair are the fairing length and width, N z is the ultimate load factor, m mot is the mass of the engine or motor inside the fairing in and S fair is the fairing wetted surface. The inverter mass m inv is calculated similarly to the motor mass with an assumed specic power [ P m ] inv . For (3.42), the turbine core mass is estimated per Raymer (2012) as m turb = 1:67P 0:803 turb ; (3.48) where P turb is the turbine core power in horsepower and the resulting mass is in pounds. The electrical machines and power electronics in the link are sized as m gen=mot =P gen=mot P m 1 gen=mot ; (3.49) m inv=rect =P inv=rect P m 1 inv=rect ; (3.50) where [ P m ] gen=mot and and [ P m ] inv=rect are assumed specic power values. 2 This motor has a total power output of 264 kW, a diameter of 0.418m and a length of 0.3m, which is used to compute constants k 1 and k 2 . 53 If acting purely as a generator, the turbine can be housed internally in the fuselage without a fairing. If, however, the turbine is mounted externally, it has a fairing that is sized based on the core dimensions. The core dimensions are estimated following Raymer (2012) as D core = 0:25(P turb =1000) 0:120 ; (3.51) ` core = 0:12(P turb =1000) 0:373 ; (3.52) with the turbine power P turb in kW. Again, fairing diameter is assumed to be 1:2D core , and the length 1:2` core . The fairing mass m turb;fair is estimated using (3.47). The model accounts for the weight of a thermal management system (TMS) which removes waste heat from electrical components. The mass of this system scales with the heat ow and a specic power, so its mass is m TMS = _ Q P m 1 TMS ; (3.53) where _ Q is the total waste heat from all electrical components, computed as _ Q = X i (1 i )P i ; (3.54) where the sum is carried over the components that generate waste heat, i is the electrical eciency of the component and P i is the peak power drawn by the component. The values used for these component eciencies and specic powers are given in Section 3.6. Lastly, an important modeling consideration is the fact that batteries deliver power at an eciency that depends on the power level at which it is operated. This relationship between eciency and power is known as the Ragone relation, and in this work is modeled 54 following Kuhn and Sizmann (2012) as bat = 1 2 + 1 2 1 P bat P bat;max ; (3.55) where bat is the battery eciency, P bat is the instantaneous power being delivered by the battery, and P bat;max is the absolute maximum power the battery can deliver. The remaining terms in (3.38) are m prop;sec and m prop;trans . The former includes engine accessories as well as the exhaust and oil system and is estimated using correlations from Torenbeek (1982). The latter is the mass of transmission lines and is estimated as a fraction of the total mass of all components that require electrical power transmission. The details of these estimates are given in Appendix C. 3.3.4 Propulsor Distribution and Integration Although the goal is to make LUCAS as general as possible, some assumptions regarding distribution are made, as described here. In order to model both conventional turboprop and turbofan engines, the mechanical propulsors modeled by LUCAS can be either ducted or unducted. It is assumed that these propulsors are mounted to the leading edge of the wings, as is common for conventional aircraft. Electrical propulsors can also be either ducted or unducted. It is assumed that these propulsors are placed along the trailing edge of the wing, where they can ingest the wing's boundary layer. For the smaller commuter aircraft class used in the present work, it is assumed that the electrical propulsors are unducted, since this will lead to less weight and drag compared to ducted propulsors, and since commuter aircraft cruise at low subsonic speeds such that ducts are not necessary for diusion. Conversely, it is assumed that electrical propulsors used by the larger aircraft classes (that cruise at transonic speeds) are ducted, since diusing ducts are needed to reduce the fan face Mach number to a value lower than the freestream for better fan performance. 55 When a distributed propulsion arrangement is employed, we assume that only electrically-driven propulsors are integrated into the wing trailing edge; BLI thus only applies to electric propulsors. Furthermore, these electric propulsors are assumed to span both upper and lower surfaces of the wing, maximizing the amount of boundary layer that can be ingested. It is further assumed that the propulsors can be distributed across the entire wing span, even through the ap and aileron regions, and considerations of how such an arrangement could best be achieved are outside the scope of this work. This is consistent with our goal of determining the potential benets of electrication within a conceptual design approach. 3.3.5 Fan Sizing, Distribution and Boundary Layer Ingestion As mentioned in the previous section, BLI is done only on the wings with electrically- driven propulsors. The amount of BLI is set via the fraction, f BLI , of boundary layer kinetic energy defect that is ingested and re-energized by the propulsors, out of the total airframe's defect 0 surf , with the product of these two appearing in the power balance accounting of Equation (3.1), as discussed in Section 3.2. This fraction is linked to the geometry of the propulsor installation, as illustrated in Figure 3.6. In the case of unducted propulsors, the electrical fan diameter is D fan;E , and the fans are placed with a gap between blade tips, fan;gap , dened as a fraction of the fan diameter. Thus, for a wing of span b, the fraction of wing span covered by fans is f wing = N fan;E (D fan;E + fan;gap )D fuse b : (3.56) The fuselage diameter D fuse is subtracted here since it is assumed that propulsors are not distributed over the fuselage area. In the case of ducted electrical propulsors, the electrical fans are placed within nacelles with diametersD nac;E , with no gaps between the nacelles. In 56 D fuse b N fan,E ... ... D fan,E fan,gap b dist D nac,E D fan,M D nac,M D fan,M Unducted Ducted Figure 3.6: Parameter denitions used to compute propulsor fan diameters and extent of distribution and BLI. Propulsors could either be unducted (left) or ducted (right). Unducted propulsors are assumed for aircraft that y in the low subsonic regime, and ducted for transonic aircraft. this case, the fraction of wing span covered by propulsors is f wing = N fan;E D nac;E D fuse b : (3.57) In order for the distribution to remain physical, the numerator in the above equation is constrained to be less than the wing span b, so that f wing 2 [0; 1]. It is assumed here that 50% of the total aircraft prole drag (i.e. non-vortex dissipation) is caused by the wing; a reasonable estimate for tube-and-wing congurations. It is also assumed that both the upper and lower wing boundary layers can be ingested. The total fraction of ingested boundary layer is thus estimated as f BLI = 0:5f wing : (3.58) Combined with the airframe dissipation 0 surf estimated from classical methods, this gives the f BLI 0 surf term in the power balance equations used in the aero-propulsive performance calculations. 57 The electrical propulsor fan diameters can either be specied or computed to maximize distribution (and thus BLI). For the latter approach, the fans are distributed over some spec- ied wing span and their diameters determined so as to fully cover this span, while allowing for the gaps between fan tips and excluding the fuselage width, as shown in Figure 3.6. In the case of unducted propulsors, the fan diameters are estimated as D fan;E = (f dist b)D fuse (1 + fan;gap )N fan;E : (3.59) For ducted propulsors, the nacelle diameters are calculated as D nace;E = (f dist b)D fuse N fan;E ; (3.60) and the fan diameter calculated to allow for clearance between the fan tips and nacelle, with D nace;E = 1:2D fan;E assumed here. The parameter f dist b dist =b is the fraction of total wingspan over which propulsors are distributed. In this work this parameter is assumed to be equal to 0.6 unless otherwise stated, meaning propulsors are only distributed over 60% of the span. This is done because the outer wing sections are typically lightly loaded and most of the viscous dissipation thus occurs inboard, while having propulsors mounted too far outboard could also lead to aeroelasticity problems and additional structural weight. We also assume fan;gap = 0:1, meaning there is a gap between fan tips of 10% the fan diameters. 3.3.6 Estimating Fuel Burn and Battery Energy The total required mission fuel and battery energy are calculated as integrated quantities of the instantaneous fuel mass ow rate _ m fuel and battery power P bat along the mission. The former is given as _ m fuel =P turb PSFC; (3.61) 58 where P turb is the core power and PSFC is the power specic fuel consumption. The value of PSFC used in this work will be presented along with the application of the framework in Chapter 4. The instantaneous battery power is P bat = P inv +P link bat ; (3.62) where P inv +P link is the power demanded from the battery by the inverter and link (see Figure 3.3), and bat is the battery eciency, as calculated through the Ragone relation of Equation (3.55). 3.3.7 Drag/Dissipation Buildup For an aircraft with no boundary layer ingestion, the surface dissipation is directly related to the viscous drag acting on its surface. The dissipation can thus be calculated using a standard prole drag buildup, such as is presented in Raymer (2012). Such a method is already implemented in SUAVE, and it used as is. The only exception is the drag caused by the propulsors, added as part of this work. For an aircraft with BLI, the surface dissipation can still be estimated in this manner, and the eect of BLI can be modeled using the method that will be described in the following section. In the current work, the propulsor drag is calculated using the methods described in Raymer (2012). As mentioned above, it is assumed that the mechanically powered propulsors are podded turbofan engines, which are fully immersed in the freestream, or turboprop engines that are attached directly to the wing. The electrically powered propulsors are embedded in the wings such that only half of their surface area are exposed to the freestream. An electrically powered propulsor will thus experience half the drag (and thus, dissipation) of a mechanically powered one with a similar fan diameter. Note that this assumption might underestimate drag for large propulsors that cannot be fully integrated into the wing, but should be a reasonable estimate for a large number of small-diameter fans. In fact, this 59 assumption might even be conservative, since it is likely that the total wetted surface of such an integrated array of propulsors could be made less than that of the sum of the surface areas of the individual propulsors. 3.4 Integration Framework 3.4.1 Mission Analysis The Stanford University Aerospace Vehicle Environment (SUAVE) (Lukaczyk et al., 2015) is used in the current work for the purpose of mission analysis. The built-in SUAVE models for the wings (main and tail), fuselage, landing gear and for non-structural components (furnishings and cabin systems) are used, but the custom models for propulsion system weight and performance presented in Section 3.3 were implemented. Traditionally, mission analysis programs integrate the equations of motion of the air- craft along a specied mission to determine mission quantities such as fuel burn. SUAVE takes a dierent approach by dening all parameters that vary over a mission as Chebyshev polynomials and modeling the mission as a system of equations. An array of unknown mis- sion parameters is created with a user-specied number of control points along the mission, and a root nding algorithm solves for the unknowns. This approach has the benet that any number of unknown parameters can be added to the system, as long as the problem is still well-posed and a solution can be found. This extensibility makes SUAVE suitable for modeling aircraft with complex propulsion systems. 3.4.2 Aircraft Sizing and Optimization The LUCAS framework described here is built around SUAVE and uses the pyOptSparse optimization framework (Perez et al., 2012) with the SNOPT (Sparse Nonlinear OPTimizer) solver (Gill et al., 2005) to converge the sizing process and perform optimization. The pyOptSparse wrapper developed for SUAVE is documented in Botero et al. (2016) and 60 DesignVariables Sizing/Optimization Outputs f L,i f S,i ` fuse d fuse R mis M PL V cruise P K,tot ... Design Parameters SUAVE pyOptSparse Analyses: Mission Weight & Balance Aerodynamics M TO M OE M fuel M bat S wing ` TO ... ... Constraints ` TO <= x m ` L <= y m V stall <= z m=s W=S M fuel M bat M TO ˙ m tot Propulsion x wing E bat,end >= 0 J Stability Figure 3.7: Integration of pyOptSparse and SUAVE into LUCAS framework. Wendor et al. (2016). Figure 3.7 shows schematically how SUAVE and pyOptSparse are integrated into the LUCAS framework, which is here represented by the whole gure. Note that where Fig- ure 3.1 showed the framework at a higher level, focussing on data ow, Figure 3.7 shows the framework at a lower level. SUAVE is used to analyze a given aircraft and contains modules for mission, weight and balance, aerodynamics, propulsion and stability. The user decides which parameters should be optimized and which ones should be xed, and supplies bounds for all the design variables, as well as constraints. Examples of design variables to be optimized in the current work are electrication levelsf L;i ,f S;i and wing loadingW=S. Here, the subscriptsi denote that electrication level can vary between dierent mission segments. Design parameters that are kept xed during sizing and optimization include fuselage length and width,` fuse andd fuse , mission rangeR mis , payload massM PL , cruise speedV cruise , and altitude h cruise . Design parameters can be changed to design variables if the user wants to optimize them, although care has to be taken to ensure the models contain the required 61 sensitivities to the design variable. In this framework, sizing is performed using an optimizer decomposition approach with compatibility constraints of the form y = y 0 , as presented by Kroo (1994), in which y is the design variable, and y 0 is the actual value of the variable calculated in the framework. The aircraft sizing equations are thus written as a set of constraints of the form y=y 0 = 1, where is a user specied tolerance. The sizing variables are set to be takeo mass M TO , total propulsor mass ow _ m tot , fuel mass M fuel , battery mass M bat , total ow power P K;tot and wing location x wing . A sizing tolerance of = 0:001, or 0.1%, is used. In the optimizer decomposition method, the optimizer itself makes no distinction between the sizing constraints discussed here and the performance constraints that will be presented next, with both sets of constraints being enforced simultaneously. 3.4.3 Performance Constraints In order to ensure that the aircraft can meet certain performance requirements, the following constraints are implemented: Maximum takeo and landing eld lengths Maximum rate of climb Minimum stall and approach speeds via an imposed maximum lift coecient Minimum volume required for fuel and batteries The actual values that will be used for the constraints in this work depend on the aircraft class. The dierent aircraft classes that are analyzed will be presented in Section 4.1, along with the values of the performance constraints. 3.4.4 Weight, Balance, and Wing Placement Due to its high weight, the location of the batteries in the aircraft can have a strong eect on the aircraft's stability and control (S&C) characteristics. Here we assume that batteries are stored in the wing, and the location of the wing along the fuselage, x wing , is determine 62 based on a specied static margin. Although a full S&C analysis is beyond the scope of the current work, this approach ensures that the aircraft should at a minimum be statically longitudinally stable. This method is also extendable to placing the battery elsewhere in the aircraft, although that was not considered here. Calculation of the static margin requires knowing the locations of the aircraft center of gravity (CG) and of its neutral point. The optimizer sets the wing position relative to the fuselage to ensure the specied static margin is met. In this work a static margin of 10% is used, which is statically stable and a reasonable value to use for commercial passenger aircraft (Raymer, 2012). Table 3.1 lists the components that are accounted for in this analysis and the method by which the component center of gravity is estimated. The `Other' row in the table includes electrical systems, furnishings, air conditioner, auxiliary power unit, and hydraulics and control systems. Note that many of the components are located relative to the wing CG, and therefore, as the optimizer moves the wing, the CG of those components move along with it. The tail starts at 95% of the fuselage. The propulsion system's link components are placed with the mechanically-driven propulsors if they are present or with the turbine assemblies otherwise. Either way, the link ends up in the same place. The equations used to estimate the location of the CG and neutral point are given in Appendix D. 3.5 Performance Metric In order to compare aircraft with dierent propulsion system architectures across various missions, the performance metric selected for this work is the Productivity Specic Energy Consumption (PSEC). This is dened as the energy used during ight divided by the product of payload weight and range, namely PSEC [ on board energy ] [ payload weight ] [ range ] : (3.63) 63 Table 3.1: Components accounted for in static margin estimation. The wing location is a design variable used to meet the static margin requirement. MAC is the mean aerodynamic chord. Component CG location Fuselage 40% of fuselage length Wing (design variable) 40% of wing MAC Horizontal tail 40% of tail MAC Vertical tail 40% of tail MAC Mechanically-drive propulsor sub-system Wing leading edge Electrically-drive propulsor sub-system Wing trailing edge Passengers 50% of passenger cabin Fuel Wing CG Battery Wing CG Main landing gear Wing CG Nose landing gear End of fuselage nose section Avionics End of fuselage nose section Other Lumped at 50% of fuselage length This non-dimensional metric is comparable to the liters of fuel consumed per passenger- kilometer (or gallons per seat-mile) that is typically used as a measure of the eciency of transport aircraft. Being more general, however,PSEC allows for the comparison of aircraft with dierent energy storage sources, such as hydrocarbon fuel and/or batteries. The present work considers only on-board energy consumption, rather than including a full `well-to-wake' analysis. As such, any consideration related to the energy required to produce jet fuel and manufacture an charge batteries before ight is beyond the scope of this work. 3.6 Electrical Technology Level Assumptions The modeling of electried aircraft in LUCAS requires as input some parameters for elec- trical components that are dependent on an assumed technology level. Examples of such parameters are the specic energies and powers of the components used in the propulsion system, as well as their eciencies. The set of values given to these parameters are referred to as the technology level. 64 Table 3.2: Technology level parameters for 2035 entry into service (Kruger et al., 2018). Parameter Unit Conservative Intermediate Optimistic Pack-level battery specic energy Wh/kg 250 575 900 Pack-level battery specic power kW/kg 0.745 1.7 2.7 Electrical machine specic power kW/kg 9 12 16 Power electronics specic power kW/kg 9 14 19 Thermal management system specic power kW/kg 10 15 25 Electrical component eciencies - 0.98 0.99 0.99 We employ technology levels estimated for an entry into service of 2035, and use the values given in Table 3.2 following the analysis in previous work (D. Hall et al., 2019; Kruger et al., 2018), as well as information presented in Section 2.6 of this thesis. The conser- vative level assumes incremental improvements on current technologies with no signicant breakthroughs. The optimistic level assumes a signicant breakthrough in battery technol- ogy, such as the commercialization of lithium-air batteries, and signicant improvements in electrical machine and power electronics. A third level, the intermediate set, is added in between the conservative and optimistic values, and assumes the commercialization of a battery chemistry such as lithium-sulfur. Note that all values for batteries are at battery pack level, and already account for the packaging eciency of combining cells to form packs, with a system to prevent thermal runaway. The value used for the optimistic battery specic energy is taken as the geometric mean of the fully charged and fully discharged states of the predicted specic energy of lithium-air batteries. This is done because the weight of these batteries increases as they are discharged. For the thermal management system specic power, results by Rheaume et al. (2019) are used to set the optimistic level, and the conservative and intermediate values are then set at 40% and 60% of that, respectively. More details on how the technology level parameters were selected can be found in Kruger et al. (2018). 65 Table 3.3: Data lters used to infer aircraft categories from T-100 data. Category Filter Commuter Turboprop with N pax 19 Regional Twin-engine turbofan with N pax 90 Transcontinental Twin-engine turbofan with 150N pax 200 3.7 Baseline Missions Due to the range penalty of using batteries, it is expected that aircraft that use batteries to store some or all of their energy would be designed to have less range capability than conventional aircraft. When deciding which mission ranges to model, it is thus more in- structive to consider the missions that will actually be own. For this, use is made of the publicly accessible Air Carrier Statistics database published by the US Bureau of Trans- portation Statistics (BTS) 3 . This data is commonly known as the T-100 or Form 41 data, and publishes statistics on all commercial ights performed in the US. Using this data it is possible to isolate the number of ights and the distance covered during each ight for the dierent categories. The categories selected for study in this work are commuter, regional and transcontinental 4 and are dened in Tables 3.3 and 3.4. The T-100 data is categorized in broad groups and it is necessary to infer the category to which ights belong by applying the appropriate lters. The lters used here are listed in Table 3.3. Furthermore, in the regional and transcontinental categories ights with distances of less than 50 miles are rejected as outliers. Figure 3.8 shows the frequency of ights performed in 2019 for the dierent categories by applying these lters. It can be seen that for all categories the majority of ights have rel- atively low ranges. The metric used here for selecting mission range is to choose ranges that capture specied fractions of the total number of ights. The shorter range design mission is dened as the range that covers 50% of all ights, and the longer range extended mission is 3 Database accessible at www.bts.gov/airline-data-downloads 4 This category is traditionally called narrow-body or single-aisle, but aircraft could be designed to perform the same missions that do not t these descriptions. The term transcontinental is thus adopted here since it is a more general description of the missions performed by these aircraft. 66 50% of flights < 180 km 90% of flights < 550 km (a) Commuter 50% of flights < 800 km 90% of flights < 2000 km (b) Regional 50% of flights < 830 km 90% of flights < 2400 km (c) Transcontinental Figure 3.8: Distribution of commuter, regional and transcontinental ight distances in the US in 2019. Source: US BTS T-100 segment data. dened as the range that covers 90% of all ights. Based on these limits, Table 3.4 lists the mission ranges selected here for the dierent aircraft classes. These ranges will be used later to compare the energy eciency and performance of conventional and electried aircraft. These ranges are for the block mission, but aircraft must also meet reserve/contingency re- quirements dictated by regulations. All aircraft are sized for a 370 km (200 nmi) alternate mission. In addition, the commuter is sized for a 45 min loiter, while the larger aircraft are sized for a 30 min loiter. Finally, all aircraft are sized with an additional 5% contin- gency based on the block mission energy requirement. For aircraft that store energy in fuel 67 Table 3.4: Flight distances chosen for design and extended range missions. Category Design range mission Extended range mission Commuter 180 km 550 km Regional 800 km 2 000 km Transcontinental 830 km 2 400 km and batteries, the reserve energy split is taken to be the same as during the block mission unless otherwise stated. These reserve requirements have a profound impact on the over- all sizing process and ultimately the feasibility and benet of electrication, especially for the commuter category, where the reserve segment accounts for more than twice the energy requirement as the block mission. It is crucial that these reserve requirements be met to ensure a realistic assessment of the potential benets and feasibility of electrication. 68 Chapter 4 Results In previous chapters the LUCAS framework was presented, along with the methodology used to model aircraft and ight missions with varying degrees of electrication. In this chapter the framework is used to model and study the design space for electried aircraft in the commuter, regional and transcontinental classes. Figure 4.1 shows a roadmap for this chapter. Chapter 4 Baseline Models Commuter Design Range Extended Range Regional Design Range Extended Range Transcontinental Design Range Extended Range 4.1 4.2 4.3 4.4 Summary 4.6 Sensitivity Study 4.5 Figure 4.1: Chapter 4 layout. 69 In section 4.1, conventional existing aircraft will be modeled with LUCAS. Following the calibration of these congurations, advanced technology versions of them will be modeled, which will serve as baselines for comparison in later sections. In sections 4.2 to 4.4 electried aircraft will be modeled and their performance compared with the baselines. Subsections will systematically cover dierent aircraft categories, missions and technology levels. Sub- section 4.2 will cover the commuter category, looking at design and extended range missions. Subsections 4.3 and 4.4 will follow the same format for the regional and transcontinental cat- egories. Finally, the outcomes of a sensitivity study will be presented in section 4.5, followed by a summary of the results presented here in Section 4.6. 4.1 Baseline Models The baseline aircraft are chosen to represent an array of commercial aviation missions that could benet from electrication. Three dierent aircraft categories were selected: commuter, regional and transcontinental. For the commuter category, the Dornier Do 228 was selected as a reference. For the regional, the Embraer E175 was selected, and for the transcontinental, the CeRAS CSR-01. The Central Reference Aircraft data System (CeRAS) is an open database designed at the University of Aachen with the goal of providing reference data for conceptual design and technology assessment. The CSR-01 is an aircraft design similar in size to the Airbus A320 or Boeing B737, but since the CeRAS database provides much more data than what is known for the A320 or B737, it is well suited for the current work. For more information on it, see Risse et al. (2016). Table 4.1 lists some relevant information on these reference aircraft, and Figure 4.2 visually compares their sizes. In addition to the mission parameters specied in Table 4.1, all aircraft are also sized to meet reserve and contingency fuel requirements required by regulations, as discussed in Section 3.7. The rst step of the analysis is to model the reference aircraft using LUCAS. This is done by gathering the required aircraft parameters, including geometric information from 70 Table 4.1: Characteristics of reference conventional aircraft. Data for Do 228 and E175 from manufacturer datasheets. Data for CSR-01 from Risse et al. (2016). Parameter Unit Do 228 E175 CSR-01 Maximum takeo mass M TO kg 6 400 38 790 74 102 Operating Empty mass M E kg 3 900 21 886 42 100 Wing span m 16.97 26.01 34.10 Wing area m 2 32.0 72.7 122.4 Wing aspect ratio - 9.0 9.3 9.5 Maximum number of passengers - 19 78 180 Design range km 1 280 3 980 5 000 Cruise altitude m (ft) 3 050 (10 000) 9 100 (30 000) 10 000 (33 000) Cruise speed m/s (kts), Mach nr M 115 (223) M 0.72 M 0.78 Stall speed m/s (kts) 35 (67) 56 (109) 84 (163) Fuselage width and height m 1.5, 1.9 3.0, 3.4 4.0, 4.1 Fuselage length m 16.6 31.7 37.6 Turbine core PSFC kg/hr/kW 0.325 0.186 0.161 Figure 4.2: Conventional aircraft used as reference. Center: Do 228, Left: E175, Right: CSR-01. Models created with OpenVSP (Hahn, 2010). Human shown next to Do 228 for scale. 71 Table 4.2: Calibration factors used for modeling reference conventional aircraft in LUCAS. Do 228 E175 CSR-01 PSFC calibration factor f Cp 1.29 1.07 1.10 M E calibration factor f M E 1.14 1.09 1.05 M core calibration factor f Mcore 0.93 0.91 0.98 Table 4.3: Comparison of model outputs for conventional reference and LUCAS results. Do 228 E175 CSR-01 Actual LUCAS Dierence Actual LUCAS Dierence Actual LUCAS Dierence Takeo mass [kg] 6 400 6 413 0.2 % 38 790 39 547 2.0 % 74 102 78 177 5.5 % Wing area [m 2 ] 32.0 34.2 6.9 % 72.7 72.1 -1.0 % 122.4 130.14 6.3 % Wing aspect ratio [-] 9.0 9.0 0.0 % 9.3 9.3 0.0 % 9.48 9.48 0.0 % Wing span [m] 16.97 17.55 3.4 % 26.01 25.95 -0.23 % 34.10 35.10 3.1 % Initial cruise C L [-] 0.33 0.33 0.0 % 0.48 0.49 2.1 % 0.53 0.53 0.0 % manufacturer data and extracted from the OpenVSP models. The input data used to dene the reference aircraft are tabulated in Appendix E. Due to incomplete information available, the models were calibrated by adjusting their core power-specic fuel consumptions (PSFC), the empty weight correlations, as well as the turbine core weights iteratively to ensure that the fuel burn, empty weight and turbine core weights match the data available for each reference. These parameters were chosen since they represent the major performance drivers of the conventional aircraft. Table 4.2 shows the values of the nal calibration factors used. Heref Cp adjusts the corePSFC,f M E adjusts the empty mass andf Mcore adjusts the turbine core mass. Note that these adjustments are not simply applied to the nal sized vehicle, but are rather applied to the correlations used to size the aircraft, and are thus an integral part of the sizing process. Table 4.3 compares various model outputs generated by LUCAS to the actual values of the reference aircraft. The goal here was not to get perfect agreement since that would require exact knowledge of the mission denition and design constraints used to size the actual aircraft. The goal was rather to get reasonable reference design models that can be used as a sound basis for comparison with the electried architectures. It can be seen that most of the compared values match within less than a few percent, which is deemed accurate enough for the work to be performed here. 72 Table 4.4: Performance constraints and related parameters used in the current work. Data for Do 228 from RUAG Aerospace Services GmbH (2021) and Finger et al. (2020). Data for E175 from manufacturer brochure, with maximum lift coecients assumed to be the same as CSR-01. Data for CSR-01 from Risse et al. (2016). Parameter Do 228 E175 CSR-01 Takeo distance [m] 793 1 724 2 200 Landing distance [m] 451 1 259 1 850 Maximum lift coecient C L;max;landing [-] 2.7 2.8 2.8 Maximum lift coecient C L;max;takeo [-] 2.4 2.2 2.2 Maximum rate of climb [m/s] 8.0 4.6 7.6 The nal aspect to be discussed regarding the baselines is the performance constraints used in their sizing process. These constraints are shown in Table 4.4. In order to ensure a fair comparison, the same constraints are used for all of the electried aircraft, unless otherwise stated. In some cases it was deemed necessary to relax some of the most stringent constraints; in those cases, the constraints were also relaxed for the conventional aircraft to keep the comparison fair. The results presented up to here served the purpose of comparing and validating the model outputs to the known values of the conventional reference aircraft. The reference aircraft are, however, relatively dated designs and the state of the art conventional aircraft are likely to be improved by the 2035 entry into service assumed for the electried aircraft of this work. Thus, the following assumptions are made to model the assumed technological advances that will aect both conventional and electried aircraft. The resulting conventional aircraft are termed baselines, and will act as baselines for comparison of electried aircraft in later comparisons. To model the advanced technology baselines we assume a 20% reduction in engine core PSFC, a 20% reduction in turbine core mass (for a given power output) and a 15% reduction in aircraft empty mass, as recommended in Raymer (2012). Table 4.5 shows the resulting performance improvements when these assumptions are applied to the reference aircraft. Separate results are used for the design and extended range missions to give a fair comparison between the conventional and electried aircraft. 73 Table 4.5: Comparison of current and advanced technology baseline aircraft. PSEC is calculated assuming a fuel specic energy of 43 MJ/kg. The advanced tech values are used as baseline values to compare the performance of the electried concepts to. Design Range Mission Extended Range Mission Current Tech Advanced Tech Dierence Current Tech Advanced Tech Dierence Commuter Takeo Mass [kg] 6 234 4 910 -21% 6 762 5 209 -23% Empty Mass [kg] 3 749 2 592 -31% 4 008 2 714 -32% Fuel Mass [kg] 518 360 -31% 775 529 -32% PSEC [-] 0.899 0.630 -33% 1.173 0.790 -33% Regional Takeo Mass [kg] 30 715 25 453 -17% 33 724 27 448 -19% Empty Mass [kg] 19 912 15 486 -22% 20 533 15 807 -23% Fuel Mass [kg] 2867 2073 -28% 5 141 3 667 -29% PSEC [-] 0.706 0.497 -30% 0.853 0.600 -30% Transcontinental Takeo Mass [kg] 60 472 50 401 -17% 67 060 54 451 -19% Empty Mass [kg] 38 299 29 754 -22% 39 600 30 267 -24% Fuel Mass [kg] 4 909 3 458 -30% 9 942 6 826 -31% PSEC [-] 0.516 0.353 -32% 0.657 0.445 -32% Table 4.5 shows that the advanced technology assumptions for turbofan eciency, tur- bine core weight and empty weight lead to PSEC reductions of between 30% and 33% across all aircraft categories and missions. These advanced technology values will be used as baselines to measure the performance of the electried architectures in the following sections. 4.2 Commuter 4.2.1 Design Range We start by discussing the predicted performance of electried commuter aircraft sized for the design mission with a range of 180 km (97 nmi). Comparison is always made to the conventional baseline aircraft sized for the same mission and with the same technology level. 4.2.1.1 All-Electric Performance The most extensive electrication strategy is the all-electric architecture with f S = 1. By xing the mission (range and payload) and electrication level we can look specically at the 74 Table 4.6: Eect of distributing propulsors while keeping total propulsive area xed to that of baseline (22.5m diameter propellers) for design mission commuter with optimistic technology, f BLI = 0. N prop D prop [m] A prop [m 2 ] M TO [kg] M bat [kg] M TO [%] PSEC [%] 2 2.50 9.82 9 305 3 401 90 -33 4 1.77 9.82 9 314 3 431 90 -33 6 1.44 9.82 9 345 3 466 90 -33 8 1.25 9.82 9 377 3 494 91 -32 10 1.12 9.82 9 408 3 521 92 -32 eects of electrical component technology level, distributed propulsion (DP) and boundary layer ingestion (BLI). Table 4.6 shows some model outputs when increasing the number of electrical propulsors, while keeping the total propulsive area constant. Due to the very large battery weight required when using either the conservative or intermediate technology assumptions (see Table 3.2), this architecture is not feasible, so Table 4.6 shows the results for optimistic technology only. Also, f BLI is set to zero in this case in order to isolate the eects of distribution from those of BLI. The conguration with just two propulsors leads to the largest reduction in PSEC, and as the number of propulsors are increased, the benet diminishes slowly. Nevertheless, all cases show a signicant energy usage benet of up to 33% with optimistic technology. This reduction, however, comes at the cost of a large growth in takeo mass. All these congurations end up having a battery mass fraction (M bat =M TO ) of 37%. It should be noted that the optimistic technology assumptions are very strong and thus might represent an upper bound of what can be achieved even beyond the 2035 time frame. The reason why thePSEC benet diminishes with an increase in propulsor count is due to the increase in total weight and drag of the fairings covering the electric motors, which doesn't scale benecially with distribution. This eect is not captured in the theoretical cube-square law, but is signicant enough to negate the predicted benet: In the case of 2 propulsors, the fairing weighs 11% of the total motor + fan + fairing assembly, where for the 10 propulsor case it is 16%. Furthermore, the total wetted surface of the motor pods increases by 72% between the two cases, leading to increase propulsion system drag for the 75 Table 4.7: Eect of distributing propulsors while varying total propulsive area for design mission commuter with optimistic technology. N prop D prop [m] A prop [m 2 ] f BLI M TO [kg] M bat [kg] M TO [%] PSEC [%] Without BLI 4 2.13 14.2 0 9 063 3 268 85 -37 6 1.45 9.90 0 9 343 3 464 90 -33 8 1.10 7.50 0 9 439 3 556 92 -31 10 0.88 6.10 0 9 565 3 657 95 -28 12 0.74 5.20 0 9 707 3 763 98 -25 14 0.64 4.50 0 9 860 3 873 101 -23 With BLI 4 2.14 14.3 23.6 9 025 3 212 84 -38 6 1.43 9.6 23.6 9 034 3 270 84 -36 8 1.07 7.2 23.6 9 092 3 337 85 -34 10 0.86 5.9 23.6 9 176 3 410 87 -33 12 0.72 4.9 23.6 9 274 3 487 89 -31 14 0.63 4.4 23.7 9 410 3 577 92 -29 case with 10 propulsors. The results in Table 4.6 assume that the total propulsive area is kept constant. Dis- tributed propulsion, however, allows for the total propulsive area to be adjusted. This changes the fan disk loading, which eects propulsive eciency. In order to show the eects of using distribution to adjust propulsive area, Table 4.7 shows the results of distributing propulsors over the wing with their diameters set to fully span 60% 1 of the wing. The results are repeated for cases both without and with BLI. When the propulsors are sized to maximize distribution in this manner, the total propul- sive area increases when fewer propulsors are used, leading to a higher PSEC benet. The benecial eect of BLI can also be seen here and is shown to lead to an additional benet of between 1 and 6% depending on the number of propulsors. When the total propulsive area A prop is large, there is relatively little jet dissipation and the eect of BLI is small. However, whenA prop is small, the relatively high jet dissipation associated with the high disk loading gets reduced signicantly by BLI, leading to larger benets. 1 See Section 3.3.5 for the reasoning behind why 60% is used. 76 4.2.1.2 Electrication Design Space The all-electric architecture could lead to substantial energy benets, but these come at the cost of a substantial growth in weight. This penalty can be reduced while still experiencing energy benets by using a hybrid-electric architecture. Figures 4.3 and 4.4 show the changes in PSEC and M TO for the entire f S f L design space, assuming the intermediate and optimistic technology levels, respectively. All the aircraft shown that use electrical propulsion have 4 electrical propulsors, as it was identied in the previous section that this leads to the largest benet by increasing propulsive area and allowing for BLI. The aircraft that use conventional propulsors have two propellers with the same diameter as the baseline conventional aircraft (2.5 m), and all architectures that have electrical distributed propulsors (f L > 0), benet from BLI. The hatched region in Figure 4.3 shows that the aircraft are not feasible for designs with f S > 0:6 as the performance constraints can't be met. However, the PSEC-optimum design that stores 42% of its energy in batteries (f S = 0:42) and is powered by both me- chanical and electrical propulsors (f L = 0:26) could lead to a PSEC benet of 6.5%, with a weight growth of 45%. If the optimistic technology levels could be achieved, Figure 4.4 shows that the PSEC-optimum design is all-electric (f S = 1;f L = 0:33) with a design that experiences a 43% reduction in energy with a weight increase of 80%. A slightly PSEC- compromised design with e.g. f S = 0:5;f L = 0:33 could still give a 22% energy benet with a more acceptable weight increase of 26%. It was found here that turbo-electric congurations (f S = 0, 0<f L < 1) do not lead to any performance benets for the commuter category design mission. 4.2.2 Extended Range We now discuss the performance of electried commuter aircraft sized for the extended range mission with a range of 550 km (297 nmi). Again, the comparison is always made to the conventional baseline aircraft sized for the same mission and with the same technology. 77 Figure 4.3: Change inPSEC andM TO across electrication design space for design mission commuter with intermediate technology. Congurations shown here have 4 electrical propul- sors for f L 6= 0 and 2 mechanical propulsors for f L 6= 1, f S < 1. Hatched region represents infeasible design space. Optimum lies at f S = 0:42, f L = 0:26. 4.2.2.1 All-Electric Performance Because of the range of this mission, the all-electric conguration was found to be infeasible for any of the technology level assumptions if the standard reserve mission requirement is 78 Optimum 0.0 0.2 0.4 0.6 0.8 1.0 f S 0.0 0.2 0.4 0.6 0.8 1.0 f L 0 15 30 45 60 75 90 Change in M TO [%] Figure 4.4: Change inPSEC andM TO across electrication design space for design mission commuter with optimistic technology. Congurations shown here have 4 electrical propulsors for f L 6= 0 and 2 conventional propulsors for f L 6= 1, f S < 1. Optimum lies at f S = 1, f L = 0:33. enforced. It was therefore decided to drop the 370 km alternate requirement from the reserve requirement. This will have implications for certication, but it is still worth exploring the electried aircraft performance for the longer range mission. Note that the mission still 79 Table 4.8: Eect of distributing propulsors while varying total propulsive area for extended mission commuter with optimistic technology. All congurations experience 24% BLI. 370 km alternate removed from reserve requirement to make analysis feasible (for this anal- ysis only). N prop D prop [m] A prop [m 2 ] M TO [kg] M bat [kg] M TO [%] PSEC [%] 4 2.14 14.3 9023 3211 73 -43 6 1.42 9.6 9029 3266 73 -42 8 1.07 7.2 9086 3333 74 -41 10 0.86 5.8 9170 3407 76 -39 12 0.72 4.9 9268 3484 88 -38 14 0.62 4.3 9375 3563 80 -36 includes a 45 min loiter reserve on top of the primary mission. Even with the relaxed reserve requirement, only the optimistic assumptions lead to a feasible design, and Table 4.8 shows the performance of all-electric architectures for the ex- tended range mission along with the eects of distribution. Here we see similar trends to the all-electric design range mission with the optimistic technology levels: fewer propulsors lead to the best designs because of the increased propulsive area and thus propulsive eciency. Even with these strong assumptions, in particular a battery specic energy of 900 Wh/kg, the aircraft still experiences a large increase in takeo weight, as much as 73% for the most energy ecient conguration with 4 propulsors. 4.2.2.2 Electrication Design Space Figure 4.5 shows the change in PSEC and M TO for the optimistic technology assumptions as a function of the source and load electrication factorsf S andf L . It can be seen that the optimal design is a parallel hybrid withf S = 0:72;f L = 0, and gives a 21% energy reduction at the cost of a 76% increase in weight. The designs in the hatched region (f S > 0:8) are not feasible. A slightly PSEC-compromised design with f S = 0:4;f L = 0 could, however, give much of the benet at a 15% PSEC reduction with a 29% increase in weight. This result suggests that longer range missions for this aircraft class will only be opened up to electric propulsion if technology advances signicantly above that of the intermediate assumptions. 80 In other words, breakthroughs in battery technology are needed for commuter aircraft to be feasible for missions beyond roughly 180 km range. As was the case for the design range mission, it was found here that turbo-electric commuter congurations do not lead to any PSEC improvement. Figures 4.6 to 4.8 show a comparison of the general propulsion system congurations of the conventional baseline commuter, the all-electric version and the parallel hybrid. 4.2.2.3 Electrication Architecture and Range Figure 4.9 shows the optimal source electrication factor f S as mission range is increased for various technology levels. Note that the ights start at 93 km, which is required for the climb to reach cruise altitude; the rst data point thus consists of only a climb plus reserves. As range is increased, the climb segment range stays constant while the cruise segment range increases. It can be seen that the optimal conguration is initially hybrid-electric for the intermediate technology assumption, with f S = 0:43. Since this mission uses a constant electrication level over the entire ight, this implies that 43% of the total energy is stored in batteries, and the rest in fuel. As range increases, the optimal f S decreases linearly up to a range of 665 km, after which a conventional conguration becomes optimal. If the optimistic technology is assumed, electried architectures remain optimal for much higher ranges. An all-electric architecture is optimal up until 200 km, after which optimal f S decreases almost linearly up to a range of 1665 km. Note that this range is much higher than the range the original conventional baseline (Do 228) can y with 19 passengers, and represents a much larger aircraft. With the optimistic technology assumptions electrication remains benecial up to unusually long ranges for a commuter aircraft of as much as 1700 km. 4.2.2.4 Segment-Varying Electrication For the results shown up to now, the electrication level (f S -f L -combination) has been con- stant throughout the mission, including the reserve segments. The eect of using dierent 81 Figure 4.5: Change in PSEC and M TO across electrication design space for extended mis- sion commuter with optimistic technology. Congurations shown here have 4 electrical propulsors for f L 6= 0 and 2 conventional propulsors for f L 6= 1, f S < 1. Hatched region represents infeasible design space. Optimum lies at f S = 0:72;f L = 0. electrication strategies during dierent segments, such as climb, cruise, descent and reserve was also investigated. It was found here that no benet results from varying electrication through the block mission, but using only fuel to power the aircraft during the reserve mis- 82 Fuel Gas turbine 2.5 m Figure 4.6: Baseline conventional commuter category aircraft. Aircraft uses fuel and gas turbine powered turbo-prop engine to power two propellers with 2.5 m diameter propellers; this gives a total propulsive area A prop;tot of 9.82 m 2 . 83 Electrical motor + controller Battery 2.14 m Figure 4.7: All-electric commuter category aircraft. Aircraft uses battery powered electric motors to power four propellers with 2.14 m diameter propellers; this gives an A prop;tot of 14.4 m 2 . This 47% increase in A prop;tot compared to the baseline shown in Figure 4.6 leads to a reduced disk loading and improved propulsive eciency. The primary eciency gains, however, come from the more ecient electrical components compared to the gas turbine system. The propulsors are integrated into the wing trailing edge to allow for BLI, which leads to further benets. 84 Fuel Battery Figure 4.8: Parallel hybrid-electric commuter category aircraft. Aircraft uses both fuel and batteries, with an electric motor coupled to the drive shaft of a gas turbine powered turbo- prop engine. Such a conguration can use battery power to supplement the gas turbine, or can even operate on batteries only for the block mission, and switch to turbine power for the reserve mission, leading to further gains. 85 Intermediate technology Optimistic technology Figure 4.9: Sweep across mission range for commuter showing optimal f S , i.e. optimal split between energy carried in batteries vs fuel. All missions include full reserve mission with 370 km diversion, 45 min loiter and 5% contingency. Table 4.9: Impact on M TO and PSEC for commuter when either using batteries for the block and reserve mission or only for the block mission. Block Mission Reserve Mission Mission Range Tech. Level f S f L f S f L M TO [%] PSEC [%] Design (180 km) Intermediate 0.5 0.5 0.5 0.5 +60 -6 Design (180 km) Intermediate 0.5 0.5 0 0.5 +29 -15 Extended (550 km) Optimistic 0.5 0.5 0.5 0.5 +43 -17 Extended (550 km) Optimistic 0.5 0.5 0 0.5 +32 -20 sion could be benecial. Table 4.9 shows the eects of assuming moderate electrication during the primary mission (storing half of the energy in the battery), but using only fuel to power the reserve mission. For the design mission, it can be seen that an additional 9%PSEC benet (from 6% to 15% benet) can be attained while reducing the weight growth penalty by 31% (from 60% to 29% penalty). For the extended range mission, an additional 3% PSEC benet can be attained while reducing the weight growth penalty by 11%. These benets occur because the aircraft doesn't carry the reserve energy as battery weight, but rather fuel which is much lighter per unit energy. The eect on the design mission is larger because the reserve mission 86 is xed and thus constitutes a larger fraction of the overall requirement for the design range mission. For the design mission, the reserve mission endurance is 75% of the total endurance, where for the extended mission, it is 53%. 4.2.2.5 Cumulative Benets Finally, we look at the cumulative eects of electrication and some model sensitivities. Fig- ure 4.10 shows the eects on PSEC and M TO when systematically implementing dierent technologies. These representative results are for the design range mission with the interme- diate technology assumptions. In order to make this mission feasible for the all-electric case, the 370 km alternate was removed from the reserve mission, to keep only the 45 min loiter. Starting from the baseline conventional architecture on the left, the bars represent the systematic introduction of: Converting to all-electric power with two propellers with the same diameter as the baseline: leads to 26% energy benet at the cost of 92% growth in mass. Increasing to 4 propulsors distributed across the wing: leads to an additional 2% energy benet with negligible change in weight. Enabling BLI: leads to an additional 5% energy benet with a 3% reduction in weight. Converting to hybrid (f S = 0:4,f L = 0:2) and using only fuel for the reserve mission: leads to 24%PSEC increase, but 42% reduction in weight. Still provides a 14%PSEC benet compared to conventional baseline. Lastly, assuming optimistic technology instead of intermediate. Optimum design op- erates as all-electric during mission (f S = 1,f L = 0:33) and turbo-electric (f S = 0,f L = 0:33) for reserves: leads to an additional 34% energy benet with 17% increase in weight. 87 -26% -2% -5% +24% -34% (a) Change in PSEC. +92% -0.1% -3% -42% +17% (b) Change in M TO . Figure 4.10: Cumulative eects of electrication for commuter. All cases assume intermediate technology, except for rightmost bar. 370 km alternate removed from reserve, with 45 min loiter still included. 88 4.2.2.6 Mass Breakdown It is instructive to compare the mass breakdowns of various architectures. Figure 4.11 com- pares such breakdowns for dierent electried commuter aircraft compared to the baseline conventional system; all results are for the design range mission. Here the takeo mass is dened as M TO =M PL +M E +M prop +M fuel +M bat ; (4.1) where M PL is the payload mass, M E the empty mass exclusive of the propulsion system, M prop the propulsion system mass, M fuel the fuel mass and M bat the battery mass. The left `Base' set shows the baseline conventional aircraft. It has fuel on board, but no battery. Next the `OptTech AE' shows the all-electric with optimistic technology. It has no fuel on board, but a very heavy battery; it can be seen here that the battery mass is 8:9 times that of the fuel mass of the baseline, and is the largest single component for this conguration. Next the `IntTech HE' shows the hybrid-electric architecture with intermediate technology. - - - - - - Figure 4.11: Mass breakdowns of various commuter architectures, from left to right: baseline conventional, optimistic tech all-electric, intermediate tech hybrid-electric, optimistic tech hybrid-electric. All aircraft sized for design mission. 89 Even though this conguration is still 30% heavier than the baseline, it is 30% lighter than the all-electric. Lastly, the `OptTech HE' shows the same conguration but with the optimistic technology. Its breakdown looks similar, except for a slight reduction across all components, except for payload which is kept constant. Although any increase in mass should be considered along with the potential energy savings of that conguration, mass in its own right is also important since many costs scale with aircraft mass 2 . 4.3 Regional 4.3.1 Design Range 4.3.1.1 All- and Hybrid-Electric It was found that an all-electric regional aircraft is not feasible for any of the technology levels due to the high battery weight required. The aircraft weight grows to a level where a feasible design could not be attained that satises all performance constraints. Furthermore, hybrid architectures that use the battery throughout the full mission are also not feasible. Segment-varying hybrids that use some batteries for the block mission but fuel only for the reserve mission were found to be feasible for low source electrication values (f S < 0:1) but provide no benet above what turbo-electric architectures could oer, as shown in Table 4.10. From these results it can be seen that the turbo-electric design withf S;block =f S;reserve = 0 leads to a PSEC reduction of 10%. When the conguration is converted into a hybrid that uses some battery energy for the block missions only and fuel for the reserve mission (f S;block = 0:1,f S;reserve = 0), the energy benet diminishes with a benet of only 5:6%, with almost three times the weight penalty (from 5.4% to 15% weight growth compared to the 2 For conventional aircraft, manufacturing, acquisition and maintenance costs are known to scale with M E (Raymer, 2012). Landing fees, which accounts for a non-negligible part of airline operating costs, also depends on aircraft mass (Belobaba et al., 2015). For aircraft that use batteries, it is expected that battery mass will also be a driver for both acquisition and maintenance costs, although operational data to justify this statement is not yet available. 90 Table 4.10: Segment-varying electrication for design mission regional with optimistic tech- nology. Congurations could use batteries for the block mission, but use only fuel for reserve mission. f S;block f S;reserve M TO M bat M TO PSEC 0.0 0.0 26 835 0.0 +5.4% -10% 0.1 0.0 29 137 1 395 +15% -5.6% baseline conventional). The conguration is thus feasible and benecial, but less so than the more simple turbo-electric architecture. For larger source electrication factors f S the conguration becomes infeasible due to the heavy battery. The remainder of this section will thus focus on turbo-electric architectures. 4.3.1.2 Turbo-electric In contrast to the smaller commuter category aircraft, it was found that turbo-electric re- gional aircraft could lead to signicant energy benets. The magnitude of the benet de- pends on the assumed technology level. Since turbo-electric architectures don't use batteries, the technology assumptions eect only the electrical machine, power electronics and thermal management system specic powers and eciencies. In this section, the aects of distributed propulsion (DP) and boundary layer ingestion (BLI) will be investigated for turbo-electric architectures. Independent variables aecting the potential benets of distribution include the size and number of distributed propulsors, which in turn eects the total propulsive area. The extent of distribution also determines the amount of boundary layer that can be ingested. First, we will look at the eects of DP and BLI when the total propulsive area is kept xed at the value of the conventional baseline, and then resized to cover a signicant portion of the wing to maximize BLI. Table 4.11 shows the change in M TO and PSEC for turbo- electric architectures employing these dierent distribution strategies. Note that all the designs shown here are partial turbo-electric, using both conventional turbofan and distributed electrical propulsors. Most of the ow power, however, comes 91 Table 4.11: Eects of distributed propulsion and boundary layer ingestion for turbo-electric design mission regional with conservative technology. Top and middle show results for a xed total propulsive area without and with BLI. Bottom shows results with propulsors auto-sized to cover 60% of the wing span for maximum distribution and BLI. Shaded row shows optimal design. N prop;E [-] D prop;E [m] A prop;E [m 2 ] BLI [%] f L;opt [-] M TO [%] PSEC [%] Fixed A prop;E without BLI 2 1.17 2.15 0 0.22 6.1 -2.2 4 0.83 2.15 0 0.22 6.1 -2.4 6 0.68 2.15 0 0.22 6.1 -2.4 8 0.59 2.15 0 0.22 6.1 -2.4 10 0.52 2.15 0 0.22 6.0 -2.4 12 0.48 2.15 0 0.22 6.0 -2.4 14 0.44 2.15 0 0.22 6.0 -2.4 16 0.41 2.15 0 0.22 6.0 -2.4 Fixed A prop;E with BLI 2 1.17 2.15 6.5 0.22 6.4 -3.2 4 0.83 2.15 9.1 0.23 6.4 -3.4 6 0.68 2.15 11.2 0.24 6.5 -3.8 8 0.59 2.15 12.9 0.24 6.5 -4.0 10 0.52 2.15 14.4 0.24 6.6 -4.0 12 0.48 2.15 15.8 0.25 6.6 -4.2 14 0.44 2.15 17.1 0.25 6.6 -4.4 16 0.41 2.15 18.3 0.25 6.7 -4.6 Auto-sized propulsors with BLI 10 0.94 6.93 26.4 0.31 13 -1.2 20 0.48 3.60 26.5 0.29 9.1 -4.4 30 0.32 2.44 26.6 0.28 7.5 -5.2 40 0.24 1.84 26.6 0.26 6.4 -5.6 50 0.19 1.46 26.6 0.24 5.7 -5.8 60 0.16 1.22 26.6 0.23 5.1 -5.8 70 0.14 1.05 26.6 0.22 4.7 -6.0 80 0.12 0.92 26.6 0.21 4.3 -6.0 92 from the turbofans, with f L only going as high as 0.31. All cases shown here show a small increase in takeo mass M TO , and a varying amount of energy benet PSEC. These changes occur because the addition of distributed propulsors increase propulsive area and thus eciency, reducing PSEC, but increases empty weight, leading to a small net weight gain. In the absence of BLI and when xing the propulsive area to that of the baseline, the top dataset shows that increasing the number of propulsors leads to a small increase inM TO and a slight decrease inPSEC. Independent of the number of propulsors, the conguration optimizes to a load electrication factor of f L = 0:22. When BLI is enabled, the middle dataset shows that distribution becomes more bene- cial, leading to aPSEC benet of up to 4.6% with 16 propulsors. Comparing the top and middle sets, it can be seen here that BLI is more benecial for the congurations with more propulsors since a larger fraction of the wing is covered and more boundary layer can be ingested. For the bottom set of data in Table 4.11 the xed-area constraint is removed and many more propulsors can be t onto the wing. In this case, more propulsors give higher energy benet, leading to a benet of 6.0% with 80 propulsors. This benet is due primarily to the weight and drag reduction of the smaller propulsors, as well as the benet of BLI. Increasing propulsor count further leads to no additional benet. Note that even though this is the theoretical optimum at 80 propulsors the fan diameters are already small at 0.12 m (12 cm), and using such a high number of propulsors might be impractical to implement due to both the high number and small size of the propulsors. For the intermediate and optimistic technology assumptions, the general trends are the same, but savings are higher due to the lower weight and higher eciencies of the electrical components. Table 4.12 shows the auto-sized propulsor cases for these technology assump- tions, where the fan diameters are automatically determined to distribute the propulsors over 60% of the wing span. The results are to be compared to the bottom set in Table 4.11. It can be seen here that increasing the number of propulsors leads to a reduction in energy usage, up to a point. However, as the technology level increases, the optimal number of 93 Table 4.12: Comparison of dierent technology levels for turbo-electric design mission re- gional with auto-sized propulsors to maximize BLI. Shaded rows show optimal designs. N prop;E [-] D prop;E [m] A prop;E [m 2 ] BLI [%] f L;opt [-] M TO [%] PSEC [%] Intermediate technology 10 0.95 7.07 26.5 0.50 12 -6.0 20 0.49 3.71 26.6 0.46 8.7 -8.5 30 0.33 2.50 26.6 0.41 7.0 -8.9 40 0.25 1.89 26.6 0.38 6.0 -8.9 50 0.20 1.51 26.6 0.15 5.2 -8.9 60 0.16 1.25 26.6 0.33 4.6 -8.7 70 0.14 1.08 26.6 0.31 4.1 -8.7 80 0.12 0.94 26.6 0.29 3.7 -8.5 Optimistic technology 10 0.94 6.93 26.4 0.57 9.6 -8.0 20 0.48 3.66 26.6 0.51 6.8 -9.9 30 0.33 2.49 26.6 0.46 5.4 -10.0 40 0.25 1.89 26.6 0.42 4.6 -10.0 50 0.20 1.51 26.6 0.39 3.9 -9.9 60 0.16 1.25 26.6 0.36 3.4 -9.7 70 0.14 1.08 26.6 0.34 3.0 -9.7 80 0.12 0.94 26.6 0.32 2.6 -9.5 propulsors decreases. If the optimistic assumptions are reached, the combination of DP and BLI could lead to an energy benet of up to 10.0% with only a 5.4% increase in weight for the design range mission. 4.3.2 Extended Range 4.3.2.1 All- and Hybrid-Electric Like for the design range mission, it was found that all- and hybrid-electric architectures are not feasible for the extended range mission. 4.3.2.2 Turbo-electric Architecture Performance results for the extended range mission show similar trends as for the design mission, as can be seen in Table 4.13. As technology improves, the optimal number of 94 propulsors changes: from 80 to 40 when going from conservative to intermediate technology, and then increases to 50 propulsors going to optimistic. However, the actual values ofPSEC indicate that this i a relatively at optimum. Note that the predicted benets for dierent technology assumptions are similar going from the design to the extended range missions. Figure 4.12 conceptually shows what the partial turbo-electric regional aircraft would look like. This conguration is partial turbo-electric because it uses both conventional turbo- fan like propulsors and distributed electric ones. 4.3.2.3 Mass Breakdown Figure 4.13 shows mass breakdowns for turbo-electric architectures with dierent technology assumptions, compared to the baseline conventional. The overall change when going from conventional to turbo-electric is much more subtle than for the commuter aircraft when congurations using batteries were considered, due to the dominant eect of the battery mass. For turbo-electric regional aircraft, the main dierence is a slight increase in maximum takeo mass, due to a slight increase in empty mass. As technology level increases (going from `ConsTech TE' to `IntTech TE'), empty mass increases slightly, while fuel mass decreases slightly. The reason for the increase in empty mass is because the congurations optimize to higher load electrication factors f L as technology level increases, which results in a slightly heavier, but more ecient propulsion system. This leads to a net fuel burn reduction of 4.1% going from conservative to optimistic assumptions. 4.4 Transcontinental 4.4.1 Design Range Where the regional aircraft class was substantially larger than the commuter class (M TO;reg 6 M TO;com ), the jump from the regional to transcontinental class is less substantial (M TO;transc 2M TO;reg ). Furthermore, the typical distances own by these two 95 Table 4.13: Comparison of dierent technology levels for turbo-electric extended mission regional architecture with auto-sized propulsors to maximize BLI. Shaded rows show optimal designs. N prop;E [-] D prop;E [m] A prop;E [m 2 ] BLI [%] f L;opt [-] M TO [%] PSEC [%] Conservative technology 10 0.98 7.51 26.6 0.30 13 -1.2 20 0.49 3.77 26.7 0.30 9.2 -4.7 30 0.33 2.53 26.7 0.29 7.5 -5.7 40 0.25 1.90 26.7 0.27 6.5 -6.0 50 0.20 1.52 26.7 0.26 5.8 -6.2 60 0.16 1.27 26.7 0.25 5.2 -6.3 70 0.14 1.09 26.7 0.24 4.7 -6.3 80 0.12 0.95 26.7 0.23 4.4 -6.5 Intermediate technology 10 0.97 7.41 26.6 0.50 12 -6.0 20 0.49 3.82 26.7 0.46 8.3 -8.7 30 0.33 2.60 26.7 0.42 6.7 -9.3 40 0.25 1.95 26.7 0.39 5.7 -9.5 50 0.20 1.56 26.7 0.36 4.9 -9.3 60 0.17 1.30 26.7 0.34 4.3 -9.3 70 0.14 1.11 26.7 0.32 3.8 -9.2 80 0.12 0.97 26.7 0.31 3.4 -9.2 Optimistic technology 10 0.96 7.18 26.5 0.57 9.2 -8.0 20 0.49 3.77 26.7 0.51 6.2 -10.0 30 0.33 2.57 26.7 0.47 4.9 -11.0 40 0.25 1.93 26.7 0.43 4.1 -11.0 50 0.20 1.56 26.7 0.40 3.4 -11.0 60 0.17 1.28 26.7 0.37 2.9 -10.0 70 0.14 1.09 26.7 0.35 2.5 -10.0 80 0.12 0.97 26.7 0.33 2.2 -10.0 96 Fuel 0.25 m 1.17 m Bus To propulsor array Generator Motor Figure 4.12: Partial turbo-electric regional category aircraft. Aircraft uses fuel only to power turbo-fan like engines that have turbo-generators mounted to their output shafts. The turbo- generators are used to power an array of distributed electric propulsors integrated into the wing trailing edge, facilitating BLI. The addition of the 40 electric propulsors increases the overall propulsive area by 91%, increasing propulsive eciency in addition to the BLI benet. 97 - - - - - Figure 4.13: Mass breakdowns of various regional architectures, from left to right: baseline conventional, conservative tech turbo-electric, intermediate tech turbo-electric, optimistic tech turbo-electric. All aircraft sized for extended mission. classes are also similar, as shown in Figure 3.8. Due to these factors, the performance of electried versions of these two classes are similar. Like for the regional class, it was found that all-electric architectures are not feasible for the transcontinental class due to excessive battery weight. Even though segment-varying hybrids could be feasible with optimistic technology, they are not benecial and require signicantly more energy than the conventional baselines. Turbo-electric architectures could however lead to signicant benets, as discussed next. 4.4.1.1 Turbo-electric Architecture As for the regional class, turbo-electric architectures are optimized when the propulsors are sized to maximize distribution. The trends are similar to the regional class, and Table 4.14 shows the full result set for the dierent technology levels, including BLI. For the conservative technology, energy benet increases as propulsor count increases and plateaus at 80 propulsors. Increasing N fan;E further could lead to slightly lower weight 98 Table 4.14: Comparison of dierent technology levels for turbo-electric design mission transcontinental architecture with auto-sized propulsors to maximize BLI. Shaded rows show optimal designs. N prop;E [-] D prop;E [m] A prop;E [m 2 ] BLI [%] f L;opt [-] M TO [%] PSEC [%] Conservative technology 10 1.35 14.3 26.9 0.36 13 0.3 20 0.69 7.44 27.0 0.32 9.4 -3.4 30 0.46 5.03 27.1 0.30 7.6 -4.5 40 0.35 3.78 27.1 0.27 6.5 -4.8 50 0.28 3.03 27.1 0.26 5.8 -5.1 60 0.23 2.51 27.1 0.24 5.2 -5.1 70 0.20 2.16 27.1 0.23 4.8 -5.4 80 0.17 1.88 27.1 0.22 4.4 -5.4 Intermediate technology 10 1.36 14.5 27.0 0.55 12 -4.8 20 0.69 7.50 27.1 0.48 8.5 -7.7 30 0.47 5.12 27.1 0.44 7.0 -8.2 40 0.35 3.87 27.1 0.39 6.0 -8.2 50 0.28 3.10 27.2 0.36 5.3 -8.2 60 0.23 2.58 27.1 0.34 4.7 -7.9 70 0.20 2.20 27.1 0.32 4.3 -7.9 80 0.18 1.92 27.1 0.30 3.9 -7.9 Optimistic technology 10 1.35 14.2 26.9 0.61 9.5 -7.1 20 0.69 7.52 27.1 0.54 6.8 -9.1 30 0.47 5.12 27.1 0.48 5.6 -9.6 40 0.35 3.87 27.1 0.44 4.8 -9.4 50 0.28 3.10 27.1 0.40 4.1 -9.4 60 0.23 2.58 27.1 0.37 3.7 -9.1 70 0.20 2.20 27.1 0.35 3.3 -8.8 80 0.17 1.90 27.1 0.32 2.8 -8.8 99 penalties, but the fan diameters will become impractically small (smaller than 17 cm). For the intermediate and optimistic levels using 50 propulsors is optimal and leads to the greatest energy benet for the least weight gain. Energy benets range between 5.4 and 9.4% at a weight cost of between 4.4 and 4.1% for conservative and optimistic levels,respectively, while intermediate technology results fall in between. 4.4.2 Extended Range 4.4.2.1 Turbo-electric Architecture For the extended range mission the turbo-electric architecture is again the only one that leads to a reduction inPSEC for this aircraft class. Table 4.15 shows the same data as Table 4.14, but for the extended range mission. Depending on the technology level, energy benets range between 5.8 and 10.0% at a weight cost of between 4.5 and 4.3% for conservative and optimistic levels, respectively. 4.4.2.2 Mass Breakdown Figure 4.14 shows mass breakdowns for turbo-electric architectures with dierent technology assumptions, compared to the baseline conventional. The general trends shown here are the same as for the regional aircraft in Figure 4.13, except for the magnitudes of the values. Again, as technology level increases, the cong- uration optimizes to a higher f L , which increases propulsion system mass, whilst reducing fuel burn thanks to an increase in propulsive eciency. When going from conventional to turbo-electric, this leads to a small net gain in takeo mass, which then stays approximately constant as technology level improves. Fuel burn, however, continuously improves. Essen- tially, increased propulsion system mass is traded for reduced fuel burn. 100 Table 4.15: Comparison of dierent technology levels for turbo-electric extended mission transcontinental architecture with auto-sized propulsors to maximize BLI. Shaded rows show optimal designs. N prop;E [-] D prop;E [m] A prop;E [m 2 ] BLI [%] f L;opt [-] M TO [%] PSEC [%] Conservative technology 10 1.37 14.8 27.0 0.33 13 0.5 20 0.70 7.65 27.1 0.32 9.3 -3.6 30 0.47 5.14 27.1 0.30 7.6 -4.7 40 0.35 3.87 27.1 0.28 6.5 -5.2 50 0.28 3.10 27.1 0.26 5.8 -5.4 60 0.23 2.58 27.1 0.25 5.3 -5.6 70 0.20 2.20 27.1 0.24 4.8 -5.6 80 0.18 1.92 27.1 0.23 4.5 -5.8 Intermediate technology 10 1.38 14.9 27.0 0.53 12 -4.7 20 0.70 7.74 27.1 0.48 8.3 -7.9 30 0.47 5.23 27.2 0.44 6.7 -8.5 40 0.35 3.94 27.2 0.40 5.7 -8.8 50 0.28 3.15 27.2 0.37 4.9 -8.8 60 0.24 2.62 27.2 0.35 4.5 -8.8 70 0.20 2.24 27.2 0.33 4.0 -8.4 80 0.18 1.95 27.2 0.33 3.7 -8.5 Optimistic technology 10 1.36 14.6 27.0 0.60 9.3 -6.7 20 0.70 7.65 27.1 0.53 6.4 -9.4 30 0.47 5.18 27.2 0.49 5.1 -9.9 40 0.35 3.91 27.2 0.44 4.3 -10.0 50 0.28 3.15 27.2 0.41 3.7 -9.9 60 0.24 2.62 27.2 0.38 3.3 -9.7 70 0.20 2.24 27.2 0.36 2.9 -9.7 80 0.18 1.97 27.2 0.34 2.6 -9.4 101 - - - - - Figure 4.14: Mass breakdowns of various transcontinental architectures, from left to right: baseline conventional, conservative tech turbo-electric, intermediate tech turbo-electric, op- timistic tech hybrid-electric. All aircraft sized for extended mission. 4.5 Sensitivity Study Lastly, Table 4.16 shows model sensitivities obtained by perturbing various parameters by 10% and computing the change in PSEC due to the perturbation. For the commuter mis- sion, the baseline used before perturbing parameters is a hybrid architecture (f S =f L =0:5) that has 10 distributed electric propulsors and 2 conventional propellers with the intermedi- ate technology assumptions. For the regional and transcontinental categories, the baselines are partial turbo-electric (f S =0;f L =0:5) that have 40 distributed electric propulsors and 2 conventional turbofan engines, also with the intermediate technology assumption. Noting that the sensitivities are listed in descending order, from most to least sensitive, Table 4.16 shows that all architectures are most sensitive to cruise speed, where increasing speed would lead to an increase in energy requirement. The transcontinental aircraft is espe- cially sensitive to this due to its high cruise speed, just below the transonic drag rise where wave drag starts to increase rapidly. All categories show a strong sensitivity to the propulsion 102 Table 4.16: Major model PSEC sensitivities evaluated around design mission with inter- mediate technology assumptions. Parameter Parameter PSEC Commuter V cruise +10% +13% EM +10% -8.5% PE +10% -8.3% C p +10% +8.2% BSE +10% -4.2% M PL +10% -4.2% DOD bat +10% -4.2% E reserve +10% +3.4% R mis +10% +3.3% h cruise +10% -2.3% Regional V cruise +10% +12% C p +10% +11% EM +10% -11% PE +10% -11% M PL +10% -5.8% h cruise +10% -2.1% R mis +10% +1.6% Transcontinental V cruise +10% +28% C p +10% +11% EM +10% -11% PE +10% -11% M PL +10% -5.9% R mis +10% +1.7% system parameters C p , EM and PE , which are the power-specic fuel consumption of the turbines, and electrical machine (motors and generators) and power electronics (inverters and rectiers) eciencies. The commuter, which uses batteries for energy storage, is sensi- tive to battery specic energy (BSE) and the maximum allowable battery depth of discharge (DOD bat ), which was assumed to be 80% in this work. The commuter also shows a high sensitivity to the reserve mission endurance E reserve , since it needs to carry the additional reserve energy in the heavy battery. Finally, all categories show some sensitivity to block mission range R mis and cruise altitude h cruise . All other parameters that were considered 103 showed sensitivities of less than 1% and were omitted. 4.6 Summary It was demonstrated here that the prospect for electrication depends on several factors, primarily: aircraft size category, mission range, reserve mission requirements and the as- sumed component technology level. For the smaller commuter class with the conservative technology assumptions, it was found that all electrication architectures lead to either in- feasible architectures or no energy benet. If the intermediate technology level is reached, in particular a pack-level battery specic energy of 575 Wh/kg, hybrid architectures become feasible and could lead to energy benets of between 7% and 15%PSEC, depending on the specic electrication strategy. If the optimistic technology level is reached, in particular a battery specic energy of 900 Wh/kg, large energy benets could be attained: all-electric ar- chitectures could experience between 38% and 43% benet, depending on the mission range, with aircraft between 84% and 73% heavier than a conventional aircraft sized for the same mission. Hybrid architectures could experience up to 20% benet, with much less weight growth than the all-electric case (32% higher than conventional baseline). Note, however, that 900 Wh/kg is a very high value, especially at pack (versus cell) level and this optimistic assumption is unlikely to be realized by a 2035 entry into service, if ever. For the larger regional and transcontinental classes it was found that only turbo-electric architectures lead to energy benets, and that these benets depend on the assumed tech- nology level, the amount of distribution and the extent of boundary layer ingestion. For the regional class design range mission, turbo-electric architectures could lead to 6.0% energy benets if the conservative technology estimates are achieved, 8.9% for the intermediate es- timates and 10.0% for the optimistic estimates. The corresponding improvements for the extended range mission are 6.5%, 9.5% and 11.0%. In the case of the larger transcontinental class design range mission, the corresponding improvements are 5.4%, 8.2% and 9.4%, and for the extended range mission they are 5.8%, 8.8% and 10.0%. 104 Although the electrication of commuter and regional aircraft could be environmentally and economically benecial, it is perhaps their role as incubation platforms for maturing the technologies for use in the larger transcontinental category that will be most important. Since the transcontinental category represents by far the largest environmental and economic impacts, electrication of single-aisle replacement aircraft could play the biggest part in shaping the future of aviation. 105 Chapter 5 Conclusions 5.1 Summary and Contributions This work presents a study of the potential benets and limitations of electrifying the propul- sion systems of transport aircraft. The work has been performed by numerically simulating, at the conceptual level, the ight missions of a range of aircraft classes employing various propulsion system architectures with varying extents of electrication. This includes all- electric architectures storing energy in batteries, hybrid-electric architectures storing energy in both fuel and batteries, and turbo-electric architectures storing energy in fuel but using electrical generators to convert some or all of the turbine power to drive electrically pow- ered fans. Apart from mission and architecture, the assumed component technology level also plays an important role and adds another dimension to the design space. The analysis was performed using the LUCAS framework, which was introduced in this work. LUCAS is essentially an extension to the existing SUAVE framework, with various additions related primarily to propulsion system modeling and using the power balance method. The primary motivation for the research was reducing the environmental impact of aviation and improving the operating economics of airlines by leveraging the high eciency and specic powers of electric propulsion components. Additionally, electric propulsion could also enable technologies like distributed propulsion and boundary layer ingestion. In order 106 to model these, LUCAS makes use of the power balance method to predict aero-propulsive performance and a unied propulsion system model that is capable of modeling the entire spectrum of propulsion system architectures. Together, these models allow for the on-board energy consumption of all architectures to be estimated and compared to the energy required by conventionally powered aircraft to perform the same mission. The modeling approach was presented here, as well as the methodology used to integrate all the subsystem models and mission analysis into the LUCAS framework. It was found that the potential energy benets of electric propulsion depend strongly on the component technology levels assumed. Three dierent sets of technology levels were used, based on an expected entry into service of roughly 2035. For 19-passenger commuter aircraft, it was found that all-electric architectures are only benecial for highly optimistic technology assumptions, but could then lead to energy benets as high as 43% in terms of PSEC. With more realistic intermediate technology assumptions, hybrid architectures could provide higher feasible ranges and benets of up to around 15% PSEC. For the larger regional and transcontinental categories, carrying between 78 and 180 passengers, respectively, it was found that the weight growth due to the massive batteries make all- and hybrid-electric aircraft either infeasible, or detrimental, independently of assumed technology level. It was, however, found that turbo-electric architectures could lead to signicant energy benets for these larger aircraft, in particular partial turbo-electric ones that convert only some of the turbofan engine power to power an array of distributed propulsors. Although the benet depends on factors like the amount of distribution, mission range and component technology levels, the regional aircraft could see energy benets as high as 11% when the optimistic technology levels are reached. The transcontinental aircraft could see benets as high as 10% with the optimistic technology levels. Lastly, it was found that any benets due to distributed propulsion are amplied by boundary layer ingestion, which could lead to additional benets of up to 6% when 27% of the boundary layer is ingested, whilst also mitigating weight growth. 107 A unique contribution of this work is the extent of the design space explored. This includes missions with three dierent aircraft classes, propulsion systems with four distinct architectures, technology levels with three dierent scenarios, along with the range of dier- ent manners in which distributed propulsion and BLI can be implemented for each aircraft class and propulsion architecture. Furthermore, each case was modelled at a relatively high delity, including a full mission prole with reserves, as well as a full set of design constraints. An outcome of the design space exploration performed here are two candidate concepts that were identied as being best suited for electrication: hybrid-electric commuter, and turbo- electric regional and transcontinental aircraft. 5.2 Electric Propulsion Fundamental Truths This work demonstrated that electric propulsion might have the potential to decrease the onboard energy requirement for dierent aircraft categories and missions. There are, how- ever, some misconceptions regarding what benets might ultimately be achievable, and the goal of this section is to present some fundamental truths that are supported by the present work. These truths are: 1. Battery technology is currently a major limiting factor, and is likely to remain so for the foreseeable future 2. Adding batteries will make aircraft substantially heavier if meaningful ranges are to be own 3. Reserve requirements have a major impact on the design of aircraft with batteries 4. Batteries are unlikely to buy themselves onto large aircraft 5. Turbo-electric architectures add additional power conversion steps, with the rst order eect of reducing eciency Explanation of these points follow: (1) The specic energy of current state-of-the-art bat- teries is roughly 50 times less than that of kerosene. The exponential increase in specic energy over the past decades is not guaranteed to continue, and even if the optimistic value 108 of 900 Wh/kg is reached, it will still be more than an order of magnitude less than that of kerosene. (2) Independant of progress in battery technology, the upper bound of battery specic energy is limited by battery chemistry to values still lower than that of hydrocar- bon fuels. Thus, irrespective of technology improvements, aircraft using batteries sized for a similar range will be heavier than conventional ones, with implications for cost. (3) Re- serve requirements will likely remain a part of the aircraft certication requirements, and the fact that the heavy battery needs to be carried on the reserve mission places an addi- tional burden on top of the actual energy requirement for the reserve mission. For short missions, the reserve mission could require much more energy (and thus weight) than the mission itself. (4) Due to the limitations of battery specic energy, the energy required for large aircraft (78-180 passengers) leads to such high battery weights that the designs are not able to meet the necessary performance constraints and are thus infeasible even under the most optimistic technology development scenarios. (5) Since these architectures require additional power conversion steps, electrication needs to `buy itself onboard' by improving overall eciency. Distributed propulsion and boundary layer ingestion are potential tech- nologies for this. These might act as high level points for aircraft designers to be aware of when considering using electried propulsion system architectures. 5.3 Further Work In order to allow for the ecient exploration of the design space, some of the sub-system component models were of relative low delity, in particular the models for the electrical machines (motors and generators), power electronics (inverters and rectiers) and thermal management systems. These models used constant eciencies to determine their power conversion performance and specic powers to determine their weights. The sensitivity study of section 4.5 demonstrated that the prediction of energy requirement is particularly sensitive to these eciencies. Now that the design space has been suciently explored and promising regions identied, it is recommended that candidate designs be analyzed with 109 higher delity component models. Such models have already been developed (Byahut, 2021; Byahut and Uranga, 2020a; Byahut and Uranga, 2020b) and the process of integrating them into LUCAS is under way, though no results had been published at the time of writing. Also, in this work on-board energy was used as a proxy for concept tness. To improve on this, a well-to-wake approach could be used which will help to quantify upstream energy costs. Work towards this end has been performed internally (Byahut et al., 2021), but has not yet been implemented into LUCAS. Furthermore, the current work did not include aspects like detailed layout design, although some design constraints were enforced as a means to improve design realism. It is recommended that further work include more detailed analysis and layout design, along with the higher delity component models, as a next step in determining the potential benets of the candidate concepts identied here. 110 Bibliography Anderson, John D. (1982). Modern Compressible Flow: With Historical Perspective. Vol. 12. McGraw-Hill New York. isbn: 0070016542. | (2016). Introduction to Flight. Mc Graw Hill Education. Atkinson, Cheryl, Helge Pfeier, Piotr Doerer, et al. (2017). Final evaluation of the Clean Sky Joint Undertaking (2008-2016) operating under FP7. Tech. rep. European Comission. Belobaba, Peter, Amedeo Odoni, and Cynthia Barnhart (2015). The Global Airline Industry. John Wiley & Sons. isbn: 9780470740774. Blakey, Simon, Lucas Rye, and Christopher Willam Wilson (2011). \Aviation gas turbine alternative fuels: A review". In: Proceedings of the Combustion Institute 33.2, pp. 2863{ 2885. issn: 15407489. doi: 10.1016/j.proci.2010.09.011. Botero, Emilio et al. (2016). \SUAVE: An open-source environment for conceptual vehicle design and optimization". In: 54th AIAA Aerospace Sciences Meeting. AIAA 2016-1275, 2016 AIAA SciTech, San Diego, CA, 4{8 January. isbn: 9781624103933. doi: 10.2514/ 6.2016-1275. Bradley, Marty K, Timothy J. Allen, and Christopher K Droney (2015). \Subsonic Ultra Green Aircraft Research: Phase II { Volume II { Hybrid Electric Design Exploration". In: NASA Technical Report CR{2015-21, p. 76. Brelje, Benjamin J. and Joaquim R.R.A. Martins (2019). \Electric, hybrid, and turboelectric xed-wing aircraft: A review of concepts, models, and design approaches". In: Progress in Aerospace Sciences 104, pp. 1{19. issn: 03760421. doi: 10.1016/2018.06.004. 111 Brodd, Ralph J (2012). Batteries for Sustainability - Selected Entries from the Encyclopedia of Sustainability Science. Springer, p. 489. Burnell, Edward, Nicole B Damen, and Warren Hoburg (2020). \GPkit: A Human-Centered Approach to Convex Optimization in Engineering Design". In: Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems. doi: 10.1145/3313831. 3376412. Byahut, Saakar (Mar. 2021). \Modeling and analysis of propulsion systems and components for electried commercial aircraft". PhD thesis. University of Southern California. Byahut, Saakar, Arjun Pundarika, and Alejandra Uranga (2021). \Modeling Well-to-Wake Life-Cycle CO2 Emissions for Electried Aircraft". In: AIAA Aviation, p. 2410. doi: 10.2514/6.2021-2410. Byahut, Saakar and Alejandra Uranga (2020a). \Power Distribution and Thermal Manage- ment Modeling for Electried Aircraft". In: 2020 AIAA/IEEE Electric Aircraft Tech- nologies Symposium (EATS). AIAA, IEEE. doi: 10.2514/6.2020-3578. | (2020b). \Propulsion System Component Modeling for an All-Electric Commuter Aircraft Mission". In: AIAA Scitech 2020 Forum. AIAA. doi: 10.2514/6.2020-0015. Clarke, Sean et al. (2015). \LEAPTech HEIST Power Architecture and Testing Roadmap Power Level for Electrical Propulsion". In: 2015 AIAA Aviation Technology, Integration, and Operations Conference, Collodi, Guido (2010). \Hydrogen production via steam reforming with CO2 capture". In: Chemical Engineering Transactions 19, pp. 37{42. issn: 22839216. doi: 10.3303/ CET1019007. de Vries, Reynard, Maurice F.M. Hoogreef, and Roelof Vos (2019). \Preliminary sizing of a hybrid-electric passenger aircraft featuring over-the-wing distributed-propulsion". In: AIAA Scitech 2019 Forum, p. 1811. isbn: 9781624105784. doi: 10.2514/6.2019-1811. 112 de Vries, Reynard, Maurice Hoogreef, and Roelof Vos (2020). \Aero-Propulsive Eciency Re- quirements for Turboelectric Transport Aircraft". In: AIAA Scitech 2020 Forum, p. 502. doi: 10.2514/6.2020-0502. Drela, M (2009). \Power balance in aerodynamic ows". In: Collection of Technical Papers - AIAA Applied Aerodynamics Conference 47.7, p. 1761. issn: 10485953. doi: 10.2514/ 6.2009-3762. Drela, Mark (June 2011). \Development of the D8 transport conguration". In: 29th AIAA Applied Aerodynamics Conference, pp. 1{14. isbn: 9781624101458. doi: 10.2514/6. 2011-3970. Dyson, Rodger W. (2018). \NASA Electric Aircraft Testbed (NEAT) Single-Aisle Trans- port Air Vehicle Hybrid Electric Tail-Cone Thruster Powertrain Conguration and Test Results". In: 2018 AIAA/IEEE Electric Aircraft Technologies Symposium, EATS 2018. IEEE, pp. 1{19. isbn: 9781624105722. doi: 10.2514/6.2018-5004. Finger, D Felix et al. (2020). \A comparison of hybrid-electric aircraft sizing methods". In: AIAA Scitech 2020 Forum, p. 1006. doi: 10.2514/6.2020-1006. Freeman, Jerey L. and Benjamin T. Schiltgen (2019). \Eco-150-300 design and performance: A tube-and-wing distributed electric propulsion airliner". In: AIAA Scitech 2019 Forum, pp. 1{15. isbn: 9781624105784. doi: 10.2514/6.2019-1808. Gei, Ingmar et al. (2018). \Optimized operation strategies for serial hybrid-electric aircraft". In: 2018 Aviation Technology, Integration, and Operations Conference. doi: 10.2514/6. 2018-4230. Gill, Philip E., Walter Murray, and Michael A. Saunders (2005). \SNOPT: An SQP algo- rithm for large-scale constrained optimization". In: SIAM Review 47.1, pp. 99{131. issn: 00361445. doi: 10.1137/S0036144504446096. Graver, Brandon, Kevin Zhang, and Dan Rutherford (2019). CO2 emissions from commercial aviation, 2018. 113 Gray, Justin S. and Joaquim R.R.A. Martins (2019). \Coupled aeropropulsive design op- timisation of a boundary-layer ingestion propulsor". In: Aeronautical Journal 123.1259, pp. 121{137. issn: 00019240. doi: 10.1017/aer.2018.120. Gray, Justin S. et al. (2018). \Aeropropulsive design optimization of a turboelectric bound- ary layer ingestion propulsion system". In: 2018 Aviation Technology, Integration, and Operations Conference, p. 3976. isbn: 9781624105562. doi: 10.2514/6.2018-3976. Gray, Justin et al. (2017). \Approach to modeling boundary layer ingestion using a fully coupled propulsion-RANS model". In: 58th AIAA/ASCE/AHS/ASC Structures, Struc- tural Dynamics, and Materials Conference, 2017, p. 1753. isbn: 9781624104534. doi: 10.2514/6.2017-1753. Greitzer, EM et al. (2010). N3 Aircraft Concept Designs and Trade Studies, Final Report. Tech. rep. NASA/CR-2010-216794. NASA, MIT. Habermann, A. L. et al. (2019). \Performance bookkeeping for aircraft congurations with fuselage wake-lling propulsion integration". In: CEAS Aeronautical Journal, pp. 1{23. issn: 18695590. doi: 10.1007/s13272-019-00434-w. Hahn, Andrew (2010). \Vehicle sketch pad: a parametric geometry modeler for conceptual aircraft design". In: 48th AIAA Aerospace Sciences Meeting Including the New Horizons Forum and Aerospace Exposition, p. 657. doi: 10.2514/6.2010-657. Hall, David K. et al. (2017). \Boundary layer ingestion propulsion benet for transport aircraft". In: Journal of Propulsion and Power 33.5, pp. 1118{1129. issn: 07484658. doi: 10.2514/1.B36321. Hall, David K. et al. (2018). \Assessment of a boundary layer ingesting turboelectric aircraft conguration using signomial programming". In: 2018 Aviation Technology, Integration, and Operations Conference, p. 3973. isbn: 9781624105562. doi: 10.2514/6.2018-3973. Hall, DK et al. (2019). Feasibility of Electried Propulsion for Ultra-Ecient Commercial Aircraft Final Report. Tech. rep. NASA/CR|2019-220382. NASA, MIT, USC. url: https://ntrs.nasa.gov/citations/20190033478. 114 Haran, K et al. (2017). \High Speed High Frequency Air-core Machine and Drive". In: NASA HGEP NRA Annual Review. Hecken, Tobias et al. (2020). \Conceptual Design Studies of \Boosted Turbofan" Congu- ration for short range". In: AIAA Scitech 2020 Forum. doi: 10.2514/6.2020-0506. Hepperle, Martin (2012). \Electric Flight-Potential and Limitations". In: Energy Ecient Technologies and Concepts of Operation. NATO Science and Technology Organization. Hidalgo, Henry and PJ Crutzen (1977). \The tropospheric and stratospheric composition perturbed by NOx emissions of high-altitude aircraft". In: Journal of Geophysical Re- search 82.37, pp. 5833{5866. doi: 10.1029/JC082i037p05833. Hoogreef, Maurice et al. (2020). \Synthesis of Aero-Propulsive Interaction Studies applied to Conceptual Hybrid-Electric Aircraft Design". In: AIAA Scitech 2020 Forum, p. 0503. Hornung, Mirko et al. (2013). \Ce-Liner - Case Study for eMobility in Air Transportation". In: 2013 Aviation Technology, Integration, and Operations Conference, p. 4302. isbn: 9781624102257. doi: 10.2514/6.2013-4302. Jansen, Ralph H. et al. (2017). \Overview of NASA electried aircraft propulsion research for large subsonic transports". In: 53rd AIAA/SAE/ASEE Joint Propulsion Conference, 2017. doi: 10.2514/6.2017-4701. Klug, Heinz G. (2000). Cryoplane - Hydrogen Fueled Aircraft - Submission for the Energy Globe Award 2001 Category \Transport". Tech. rep. Airbus. Krishnan, Ramu (2017). Switched reluctance motor drives: modeling, simulation, analysis, design, and applications. CRC press. Kroo, Ilan et al. (1994). \Multidisciplinary optimization methods for aircraft preliminary design". In: 5th Symposium on Multidisciplinary Analysis and Optimization, 1994. AIAA 1994-4325, Symposium on Multidisciplinary Analysis and Optimization, Panama City Beach, FL, 7{9 September, pp. 697{707. doi: 10.2514/6.1994-4325. 115 Kruger, Michael and Alejandra Uranga (2020). \The Feasibility of Electric Propulsion for Commuter Aircraft". In: AIAA Scitech 2020 Forum, p. 1499. doi: 10.2514/6.2020- 1499. Kruger, Michael et al. (2018). \Electried aircraft trade-space exploration". In: 2018 Avia- tion Technology, Integration, and Operations Conference. AIAA 2018-4227, 2018 AIAA Aviation, Atlanta, GA, 25{29 June. isbn: 9781624105562. doi: 10.2514/6.2018-4227. Kuhn, Holger and Andreas Sizmann (2012). Fundamental Prerequisites for Electric Flying. Deutsche Gesellschaft f ur Luft-und Raumfahrt-Lilienthal-Oberth eV, pp. 1{8. doi: 10. 13140/2.1.1168.9282. Lambe, Andrew B and Joaquim R.R.A. Martins (2012). \Extensions to the Design Struc- ture Matrix for the Description of Multidisciplinary Design, Analysis, and Optimiza- tion Processes". In: Structural and Multidisciplinary Optimization 46, pp. 273{284. doi: 10.1007/s00158-012-0763-y. Liu, S (2017). \Ultra-Light Highly Ecient MW-Class Cryogenically-Cooled Inverter". In: 55th AIAA Aerospace Sciences Meeting. AIAA (American Institute of Aeronautics & Ast). Loder, David C., Andrew Bollman, and Michael J. Armstrong (2018). \Turbo-electric Dis- tributed Aircraft Propulsion: Microgrid Architecture and Evaluation for ECO-150". In: 2018 IEEE Transportation and Electrication Conference and Expo, ITEC 2018, pp. 488{ 493. doi: 10.1109/ITEC.2018.8450180. Lukaczyk, Trent et al. (2015). \SUAVE: An open-source environment for multi-delity con- ceptual vehicle design". In: 16th AIAA/ISSMO Multidisciplinary Analysis and Optimiza- tion Conference. AIAA 2015-3087, 2015 AIAA Aviation, Dallas, TX, 22{26 June. isbn: 9781624103681. doi: 10.2514/6.2015-3087. Mattingly, Jack D, William H Heiser, and David D Pratt (2000). Aircraft Engine Design. 2nd ed. American Institure of Aeronautics and Astronautics. isbn: 9781600860164. 116 McCormick, Barnes W. (1979). Aerodynamics, Aeronautics, and Flight Mechanics. John Wiley & Sons, Inc. Misra, Ajay (2017a). \Evolution of Fundamental Technologies for Future Electried Air- craft". In: Electric and Hybrid Aerospace Technology Symposium. | (2017b). \Technical Challenges and Barriers Aecting Turbo-electric and Hybrid Electric Aircraft Propulsion". In: Energytech. | (2018). \Energy Storage for Electried Aircraft". In: IEEE Electrication Magazine 6.September, pp. 54{61. doi: 10.1109/MELE.2018.2849922. Mor e, Jorge J, Burton S Garbow, and Kenneth E Hillstrom (1980). User guide for MINPACK-1. Tech. rep. CM-P00068642. Panasonic (2018). Specications for NCR18650GA. Tech. rep. 1. Panasonic, pp. 65{68. Perez, Ruben E., Peter W. Jansen, and Joaquim R.R.A. Martins (2012). \PyOpt: A Python- based object-oriented framework for nonlinear constrained optimization". In: Structural and Multidisciplinary Optimization 45.1, pp. 101{118. issn: 1615147X. doi: 10.1007/ s00158-011-0666-3. Pilawa, A (2017). \Modular and Scalable High Eciency Power Inverters for Extreme Power Density Applications". In: 55th AIAA Aerospace Sciences Meeting. AIAA (American Institute of Aeronautics & Ast). Polek, Gregory (July 2021). Pratt & Whitney Hybrid-electric system to y in a dash 8 in 2024. url: https://www.ainonline.com/aviation-news/air-transport/2021- 07-15/pratt-whitney-hybrid-electric-system-fly-dash-8-2024 (visited on 12/28/2021). Raymer, D P (2012). Aircraft Design: A Conceptual Approach. 5th ed. American Institute of Aeronautics & Astronautics. Rheaume, Jonathan M, Malcolm MacDonald, and Charles E Lents (2019). \Commercial Hybrid Electric Aircraft Thermal Management System Design, Simulation, and Op- 117 eration Improvements". In: AIAA Propulsion and Energy 2019 Forum, p. 4492. doi: 10.2514/6.2019-4492. Risse, Kristof et al. (2016). \Central Reference Aircraft data System (CERAS) for research community". In: CEAS Aeronautical Journal 7.1, pp. 121{133. doi: 10.1007/s13272- 015-0177-9. Roland Berger (2018). Aircraft Electrical Propulsion{Onwards and Upwards. Tech. rep. Sederanger 1, 80538 Munich, Germany. Rolt, Andrew et al. (2017). \Scale eects on conventional and intercooled turbofan engine performance". In: Aeronautical Journal 121.1242, pp. 1162{1185. issn: 00019240. doi: 10.1017/aer.2017.38. RUAG Aerospace Services GmbH (2021). Dornier 228 Advanced Commuter (AC) Facts & Figures. Santhanam, Kalathur SV et al. (2017). Introduction to Hydrogen Technology. John Wiley & Sons. Schmollgruber, P. et al. (2019). \Multidisciplinary exploration of dragon: An onera hybrid electric distributed propulsion concept". In: AIAA Scitech 2019 Forum, p. 501. isbn: 9781624105784. doi: 10.2514/6.2019-1585. Schmollgruber, Peter et al. (2019). \Multidisciplinary Exploration of DRAGON: an ONERA Hybrid Electric Distributed Propulsion Concept". In: AIAA Scitech Forum, p. 1585.doi: 10.2514/6.2019-1585. Schumann, Ulrich (1993). \On the eect of emissions from aircraft engines on the state of the atmosphere". In: doi: 10.1007/s00585-994-0365-0. Seitz, Arne et al. (2018). \Concept validation study for fuselage wake-lling propulsion in- tegration". In: 31st Congress of the International Council of the Aeronautical Sciences, ICAS 2018. Smith, Leroy H. (1993). \Wake ingestion propulsion benet". In: Journal of Propulsion and Power 9.1, pp. 74{82. issn: 07484658. doi: 10.2514/3.11487. 118 Torenbeek, Egbert (1982). Synthesis of Subsonic Airplane Design. Delft University Press & Martinus Nijho Publishers. doi: 10.1007/978-94-017-3202-4. | (2013). Advanced Aircraft Design: Conceptual Design, Analysis and Optimization of Sub- sonic Civil Airplanes. John Wiley & Sons, pp. 1{410. isbn: 9781118568118. doi: 10. 1002/9781118568101. Uranga, Alejandra et al. (2017). \Boundary layer ingestion benet of the D8 transport aircraft". In: AIAA Journal 55.11, pp. 3693{3708. issn: 00011452. doi: 10.2514/1. J055755. Uranga, Alejandra et al. (2018). \Analysis of the aerodynamic benet from boundary layer ingestion for transport aircraft". In: AIAA Journal 56.11, pp. 4271{4281.issn: 00011452. doi: 10.2514/1.J056781. Vries, Reynard de, Malcom T. Brown, and Roelof Vos (2018). \A preliminary sizing method for hybrid-electric aircraft including aero-propulsive interaction eects". In: 2018 Aviation Technology, Integration, and Operations Conference, p. 4228. isbn: 9781624105562. doi: 10.2514/6.2018-4228. Welstead, Jason R. and James L. Felder (2016). \Conceptual design of a single-aisle tur- boelectric commercial transport with fuselage boundary layer ingestion". In: 54th AIAA Aerospace Sciences Meeting. isbn: 9781624103933. doi: 10.2514/6.2016-1027. Welstead, Jason et al. (2017). \Overview of the NASA STARC-ABL (rev. B) advanced concept". In: One Boeing NASA Electric Aircraft Workshop. Wendor, Andrew D., Emilio Botero, and Juan J. Alonso (2016). \Comparing dierent o- the-shelf optimizers' performance in conceptual aircraft design". In: 17th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference. AIAA 2016-3362, 2016 AIAA Aviation, Washington, DC, 13{17 June. isbn: 9781624104398. doi: 10.2514/6.2016- 3362. 119 Wu, Neil et al. (Oct. 2020). \pyOptSparse: A Python framework for large-scale constrained nonlinear optimization of sparse systems". In: Journal of Open Source Software 5.54, p. 2564. doi: 10.21105/joss.02564. Zhang, D (2017). \NASA SiC Light-Weight Inverter for MW-Power (SLIM)|Phase I". In: 55th AIAA Aerospace Sciences Meeting. AIAA (American Institute of Aeronautics & Ast). 120 Appendix A Unied Propulsion System Model A.1 Linear System As mentioned in Section 3.3.1, the equations used to describe the unied propulsion system model can be cast into a linear system and written in the matrix forms given here. When the power ow in the link is upwards (P link > 0), the link acts as a motor and the system can be written as 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 fan 0 0 0 1 0 0 0 0 0 1 fan 0 0 0 1 0 0 0 0 0 0 0 0 1 EM 1 0 0 0 0 0 0 0 0 1 PE 1 0 0 0 1 0 1 0 0 0 EM PE 0 0 0 1 0 0 0 1 1 0 0 f S 1f S 1 0 0 0 0 0 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 P K;M P K;E P turb P bat P fan;M P fan;E P mot P inv P link 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 = 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 (1f L )P K;tot f L P K;tot 0 0 0 0 0 0 0 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 (A.1) 121 When ow in the link is downwards (P link < 0), the link acts as a generator and the system can be written as 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 fan 0 0 0 1 0 0 0 0 0 1 fan 0 0 0 1 0 0 0 0 0 0 0 0 1 EM 1 0 0 0 0 0 0 0 0 1 PE 1 0 0 0 1 0 1 0 0 0 1 EM PE 0 0 0 1 0 0 0 1 1 0 0 f S 1f S 1 0 0 0 0 0 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 P K;M P K;E P turb P bat P fan;M P fan;E P mot P inv P link 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 = 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 (1f L )P K;tot f L P K;tot 0 0 0 0 0 0 0 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 (A.2) Note that the only dierence is the term in the seventh row and nal column of the coecient matrices. Equation (A.1) is used when the link acts as a motor. This is true when EM PE f S (1f L ) > (1f S )f L : (A.3) Conversely, when the left-hand-side is smaller than the right-hand-side, the link acts as a generator and (A.2) is used. 122 Appendix B Performance Constraints B.1 Maximum Takeo and Landing Field Lengths A maximum takeo eld length constraint is implemented using the methods given by Raymer (2012) as ` TO ` G +` T +` C ; (B.1) where ` G=T=C represent the ground distances covered during the ground roll, transition and obstacle clearance segments of the takeo segment, respectively. A maximum landing eld length constraint is implemented, also using the methods from Raymer as ` L ` A +` F +` G ; (B.2) where ` A=F=G represent the ground distances covered during the approach, are and ground roll segments of the landing segment, respectively. The equations used to estimate these segment lengths can be found in Raymer (2012). 123 B.2 Maximum Rate of Climb To ensure the aircraft has enough power to meet a specied maximum rate of climb, the thrust-to-weight ratio is rst calculated as (T=W ) climb = 1 C L =C D + ROC V 1 ; (B.3) where C L =C D is the ratio of aircraft lift to drag in the climb conguration, ROC is the specied rate of climb and V 1 is the freestream velocity. The power constraint is then written as P K;tot (T=W ) climb W TO V 1 ; (B.4) where W TO is the aircraft takeo weight. B.3 Minimum Fuel and Battery Volume It is assumed that both fuel and batteries are stored in the aircraft wings. A 50% fraction of the wing volume is allocated to fuel or battery storage, a conservative estimate compared to guidelines given in Raymer (2012) and Torenbeek (1982). The wing volume is estimated per Torenbeek (2013) as V wing = k Q r p 1 + S r S w AR ; (B.5) where r is the ratio of wing frontal to planform area, is the wing taper ratio, S w is the wing planform area and AR the aspect ratio, and k Q is set to 0.95 (Torenbeek, 2013). The fuel and battery volume constraints are then given as V fuel +V bat 0:5V wing : (B.6) 124 B.4 Minimum Stall and Approach Speeds In order to ensure the wing is of adequate size, wing loading constraints are implemented to ensure adequate stall and approach speeds. The stall speed is estimated from an assumed maximum aircraft lift coecient per Raymer (2012) as V stall = s 2M TO g SL S w C L;max ; (B.7) where SL is sea level air density, is the air density at the altitude where the equation is evaluated, S w is the wing planform area and C L;max is the wing maximum lift coecient. The approach speed is taken as 1.22 times the stall speed (Torenbeek, 2013) calculated at landing weight: V app = 1:22 s 2 (M TO M fuel )g SL S w C L;max : (B.8) The wing loading constraint for stall speed is then W=Sj stall 1 2 SL V 2 stall C L;max ; (B.9) and for approach it is, following Torenbeek (2013), W=Sj app 0:335 SL V 2 app C L;max : (B.10) In this work, the wing loading constraints are evaluated at sea level conditions, so in Eqs. (B.7) and (B.8) are substituted with SL . The specic values assumed for the various performance constraints can be found in Section 3.4.3. 125 Appendix C Propulsion System Mass C.1 Propulsion System Secondary Masses In Section 3.3.3, Equation (3.38) gave the total propulsion system mass as m prop;tot = m prop;prim +m prop;sec +m prop;trans ; (C.1) and the estimation method form prop;prim was given in that section. The secondary component masses are estimated as m prop;sec =m acc +m exhaust +m oilsys +m thrustrev : (C.2) For turboprop engines the right-hand-side terms are computed using correlations from Toren- beek (1982) as m acc = N turb 0:181 (P turb =1000) 0:8 ; (C.3) m exhaust = N turb 0:146S fair=nacelle ; (C.4) m oilsys = N turb 0:07m turb ; (C.5) m thrustrev = 0: (C.6) 126 For turbofan engines Eqs. (C.3) and (C.6) become m acc = N turb 36 _ m fuel ; (C.7) m thrustrev = N turb 0:18m turb : (C.8) HereN turb is the number of turbine cores in the system,P turb is the maximum power delivered by the turbine,S fair=nacelle is the wetted surface of the fairing or nacelle covering the turbine, m turb is the turbine core mass and _ m fuel is the fuel mass ow though the core at its sizing point. Implicit in m exhaust is that the engine exhaust wetted surface is 10% of that of the total fairing or nacelle. Lastly, the total electrical transmission line mass is estimated as m prop;trans =K trans [N turb (m gen=mot +m inv=rect ) +N fan;E (m mot +m inv )]; (C.9) whereK trans is an assumed factor that represents the transmission line mass as a fraction of the mass of all components that convert electrical power, here taken as 0.2. The nomenclature adopted here is the same as in Figure 3.3, wherem gen=mot is the mass of the electrical machine (generator or motor) in the link, m inv=rect is the mass of the power electronics (inverter or rectier) in the link, N fan;E is the number of electrical propulsors, m mot is the mass of an individual electrical propulsor motor, and m inv is the mass of an inverter accompanying an individual electrical propulsor motor. Note that the assumption K trans = 0:2 implies that the total transmission line mass is 20% of the mass of all components requiring electrical power transmission. 127 Appendix D Wing Placement D.1 Wing Placement The location of the center of gravity (CG) along the longitudinal axis of the aircraft relative to some reference point is calculated as x CG = P M comp M TO ; (D.1) where P M comp is the sum of mass moments due to each component listed in Table 3.1 about some arbitrary reference point, and M TO is the takeo mass of the aircraft. The CG along the vertical axis is not used in the current work, and the aircraft is assumed to be symmetrical about the lateral axis, so only longitudinal stability and control eects are considered. The aircraft neutral point is calculated from a simplied approach adapted from An- derson (2016). The location of the neutral point is calculated as ` np =` ac +V h a t a 1 @ @ ; (D.2) where ` np is the neutral point location relative to the mean aerodynamic chord (MAC) 128 leading edge, and normalized by the MAC, V h is the horizontal tail volume coecient, ` ac is the location of the wing-body combination aerodynamic center, a is the lift curve slope of the wing-body combination, a t is the lift curve slope of the tail and @=@ is the change in wing downwash angle with change in angle of attack. The value used for V h depends on aircraft category. For the commuter, a value of 0.8 is used, and for the larger regional and transcontinental categories, a value of 1.0 is used, as recommended by Raymer (2012). The location of ` ac is assumed to be 25% of the wing MAC, a t is assumed to be 0.1 per degree (Anderson, 2016), and a is calculated as a = a 0 1 + 57:3a 0 =(eAR) ; (D.3) witha 0 = 0:1097 per degree,e the Oswald eciency factor taken to be 0.9, andAR the wing aspect ratio. The parameter @=@ is usually determined using wind tunnel testing and is dicult to know during the conceptual design stage, and a value of 0.35 is used here based on an example value given by Anderson (2016). Finally, the static margin is estimated as SM =` np x CG MAC : (D.4) Note that since ` np is already normalized by the MAC, the static margin is calculated as a fraction of the MAC. In this work, the SM is specied by the user, and the optimizer places the wing (and components that are xed relative to the wing) at a suitable location to set the SM. 129 Appendix E Baseline Input Parameters Parameter Commuter Regional Transcontinental Fuselage Seats abreast 2 4 6 Seat pitch [m] 0.762 0.813 0.81 Nose neness ratio 1.6 1.3 1.6 Tail neness ratio 2.0 3.7 3.2 Total length [m] 16.6 31.7 37.6 Nose length [m] 3.48 4.13 6.50 Tail length [m] 5.35 11.7 13.1 Cabin length [m] 7.08 21.3 27.5 Cabin origin [m] 3.77 4.75 5.68 Width [m] 1.47 3.01 3.95 Height [m] 1.85 3.35 4.14 Dierential pressure [Pa] - 50000 50000 Wing Span [m] 16.97 26.0 34.1 Aspect ratio 9.0 9.3 9.5 Quarter chord sweep [deg] 0 24.0 24.5 Average thickness to chord 0.16 0.12 0.12 Average taper 0.70 0.26 0.17 Span eciency 0.9 0.9 0.9 130 Parameter Commuter Regional Transcontinental Root chord [m] 2.17 5.10 7.14 Tip chord [m] 1.65 1.23 1.23 Airfoil zero lift angle of attack [deg] -4.0 -2.5 -2.5 Twist (washout) [deg] 2 2 2 Flaps start location (as fraction of span) 0.10 0.10 0.12 Flaps end location (as fraction of span) 0.50 0.50 0.78 Horizontal tail Aspect ratio 5.0 4.3 5.0 Quarter chord sweep [deg] 0 36.5 28.0 Average thickness to chord 0.08 0.08 0.08 Average taper 1 0.32 0.29 Span eciency 0.9 0.9 0.9 Volume coecient 0.93 1.0 1.0 Location (as fraction of fuselage length) 0.90 0.82 0.84 Vertical tail Aspect ratio 1.5 1.6 1.7 Quarter chord sweep [deg] 0 40 33 Average thickness to chord 0.08 0.08 0.08 Average taper 0.80 0.19 0.32 Span eciency 0.9 0.9 0.9 Volume coecient 0.08 0.09 0.09 Location (as fraction of fuselage length) 0.90 0.69 0.77 Landing gear Main gear strut length [m] 1.5 1.75 2.0 Nose gear strut length [m] 1.5 1.31 2.0 Mission Design range [km] 2360 3980 4630 131 Parameter Commuter Regional Transcontinental Nr of passengers 7 78 180 Cruise altitude [m] 3048 9144 9144 Cruise speed 115 m/s M 0.72 M 0.78 Climb speed [m/s] 77 154 154 Rate of climb and descent [m/s] 2.54 4.57 4.57 Performance constraints Takeo eld length [m] 793 1724 2200 Landing eld length [m] 451 1259 1850 Stall speed [m/s] 35 56 84 Obstacle clearance height [m] 15.2 15.2 15.2 Max rate of climb [m/s] 8.0 11.7 7.6 Static margin 0.1 0.1 0.1 Propulsion Number of propeller blades 5 - - Fan bypass ratio - 5.0 4.8 Number of core stages (comp+turb) 3 6 12 PSFC [kg/hr/kW] 0.325 0.186 0.161 Eta fan 0.9 0.9 0.9 Eta noz 0.99 0.99 0.99 Fan diameter [m] 2.5 1.2 1.6 Number of engines 2 2 2 Number of fuel tanks 2 2 2 k s 2 1 + 0:4 2 1 1 Calibration factors f Cp 1.29 1.07 1.10 f Wempty 1.14 1.09 1.05 f Wcore 0.93 0.91 0.98 Flight envelope Ultimate load factor 3.5 3.5 3.5 Limit load factor 1.5 1.5 1.5 Max lift coecient (landing) 2.7 2.8 2.8 132 Parameter Commuter Regional Transcontinental Max lift coecient (takeo) 2.4 2.2 2.2 Fuselage pressure altitude [m] - 1800 1800 133
Abstract (if available)
Abstract
This work presents a study of the potential benefits and limitations of electric propulsion for aircraft. Components such as chemical batteries, electrical machines and power electronics benefit from high power conversion efficiencies and high specific powers (power conversion capability per unit weight). These components could also enable higher levels of aero-propulsive integration through distributed propulsion and boundary layer ingestion, leading to further benefits. Their use in aircraft propulsion has thus been proposed as a potential path towards improving aircraft energy efficiency. The primary challenge lies in the fact that batteries have very low specific energies (energy stored per unit weight). A direct swapping of fuel for batteries would result in a flight range reduction of almost two orders of magnitude. Thus, for aircraft electrification to be beneficial, solutions have to be found that take advantage of the benefits whilst mitigating potential penalties. ❧ The design space of missions, propulsion system architectures and their technology levels is large. Understanding and quantifying the potential benefits of electrification needs to be done on the basis of tools that can capture the multiple trade-offs with sufficient level of fidelity. This work presents LUCAS, a computational library that was developed around the SUAVE framework, which is able to simulate flight missions and aircraft with propulsion system architectures making use of varying degrees of electrification. This includes all-electric architectures storing all their energy in batteries, hybrid-electric architectures storing energy in both fuel and batteries, as well as turbo-electric architectures storing energy in fuel only, but converting power using generators to power electric propulsors. Using LUCAS, the on-board energy requirement of these architectures can be compared to that of baseline conventional aircraft on an equal basis and the design space extensively explored. ❧ Various flight missions are considered, assuming an entry into service in the 2035 time frame. The missions include commuter aircraft carrying 19 passengers over a distance as low as 180 km, up to trans-continental aircraft carrying 180 passengers over a distance as high as 2 400 km. It is found that all-electric architectures only become feasible for aircraft as large as the commuter category if very large improvements in battery specific energy are made, but could then lead to on-board energy savings of up to 36%. Hybrid architectures have the potential to increase the feasible range and reduce the weight penalties of all-electric architectures, albeit with lower energy benefits. Such battery technology is, however, not expected to be realized soon. ❧ For the larger aircraft classes, it was found that turbo-electric architectures could be beneficial even with today’s technologies, leading to benefits of up to 6%, and if technology progresses further, benefits of up to 11% could be attained. It was found that aero-propulsive benefits due to distributed propulsion and boundary layer ingestion play a role in the potential gains, although the benefits primarily come from the high efficiency and specific powers of the electrical propulsion components. Based on the work performed here, two candidate concepts that seem best suited for electrification and most warrant further work are hybrid-electric commuter and turbo-electric regional and transcontinental aircraft.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Modeling and analysis of propulsion systems and components for electrified commercial aircraft
PDF
Biomimetics and bio-inspiration for moderate Reynolds number airfoils and aircraft
PDF
High temperature latent heat thermal energy storage to augment solar thermal propulsion for microsatellites
PDF
Development and applications of a body-force propulsor model for high-fidelity CFD
PDF
Numerical and experimental investigations of ionic electrospray thruster plume
PDF
An investigation into magnetically induced extractor-less electrospray propulsion devices
PDF
Nano-engineered devices for display and analog computing
PDF
A joint framework of design, control, and applications of energy generation and energy storage systems
PDF
Reinforcement learning in hybrid electric vehicles (HEVs) / electric vehicles (EVs)
PDF
Hydrogen peroxide vapor for small satellite propulsion
PDF
Systems engineering and mission design of a lunar South Pole rover mission: a novel approach to the multidisciplinary design problem within a spacecraft systems engineering paradigm
PDF
Aerodynamic investigation of a novel aircraft configuration
PDF
Integration of energy-efficient infrastructures and policies in smart grid
PDF
Utility – customer interface strategy and user application for electric vehicles and distributed generation management
PDF
Hardware and software techniques for irregular parallelism
PDF
Environmental effects from a large-scale adoption of electric vehicle technology in the City of Los Angeles
PDF
Integration of truck scheduling and routing with parking availability
PDF
Noise benefits in Markov chain Monte Carlo computation
PDF
Flame ignition studies of conventional and alternative jet fuels and surrogate components
PDF
Models and algorithms for pricing and routing in ride-sharing
Asset Metadata
Creator
Kruger, Michael
(author)
Core Title
The challenges and potential benefits of electrified propulsion for aircraft
School
Viterbi School of Engineering
Degree
Doctor of Philosophy
Degree Program
Aerospace Engineering
Degree Conferral Date
2022-05
Publication Date
01/12/2022
Defense Date
11/19/2021
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
aircraft design,aircraft propulsion,all-electric aircraft,boundary layer ingestion,distributed propulsion,electric aircraft propulsion,electrified aircraft propulsion,hybrid-electric aircraft,OAI-PMH Harvest,turbo-electric aircraft
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Uranga, Alejandra (
committee chair
)
Creator Email
michaelkrugerus@gmail.com,mkruger@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-oUC110454556
Unique identifier
UC110454556
Legacy Identifier
etd-KrugerMich-10336
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Kruger, Michael
Type
texts
Source
20220112-usctheses-batch-907
(batch),
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Repository Email
cisadmin@lib.usc.edu
Tags
aircraft design
aircraft propulsion
all-electric aircraft
boundary layer ingestion
distributed propulsion
electric aircraft propulsion
electrified aircraft propulsion
hybrid-electric aircraft
turbo-electric aircraft