Electrification for Aircraft Propulsion:

Modeling and Potential Benefits

7th International Workshop on Aviation and Climate Change

University of Toronto Institute for Aerospace Studies
May 19 – 21, 2021

Alejandra Uranga
Gabilan Assistant Professor
University of Southern California

ADRL Aerodynamic
Design & Research
adrl.usc.edu Laboratory
Acknowledgments A. Uranga

Introduction
USC PhD students Methodology
Saakar Byahut Unified Propulsion
System Model

Michael Kruger LUCAS Framework

Technology
Component Modeling

Cruise-Only
NASA LEARN3 Project 2016-2018 Approximation

NASA/MIT Collaborative Agreement NNX16AK25A Full-Mission

Commuter
PI: E. Greitzer (MIT) Modeling
Participating organizations: MIT Lessons

USC Open Questions

Aurora Flight Sciences (now a Boeing Company)

A. Uranga (USC) 1 / 47
Summary A. Uranga

I Propulsion system electrification may help reduce aviation’s emissions Introduction

Methodology
I Electrification is synergistic with distributed propulsion (DP) Unified Propulsion
System Model

and boundary layer ingestion (BLI) LUCAS Framework

Technology
Component Modeling

I Improvements in battery technology are crucial to making electrification Cruise-Only

Approximation
viable and beneficial (other electrical component tech already sufficient) Full-Mission
Commuter
I Each electrified propulsion architecture has its sweet spot Modeling
I Fully electric aircraft best for short ranges and small payloads Lessons

I Turbo-electric seems better suited for high ranges Open Questions

I Hybrid-electric has niche at intermediate ranges

I Electrification can reduce on-board energy requirements

A. Uranga (USC) 2 / 47
Outline A. Uranga

Introduction

1 Introduction Methodology
Unified Propulsion
System Model
LUCAS Framework

2 Methodology Technology
Component Modeling

Cruise-Only
Approximation
3 Cruise-Only Approximation with Low-Fidelity Models Full-Mission
Commuter

Modeling
4 Full-Mission Commuter Results Lessons

Open Questions

5 Lessons Learned from Modeling Fidelities

6 Open Questions

A. Uranga (USC) 3 / 47
Demand for Air Transportation A. Uranga

I Aviation represents ∼3% of global emissions Introduction

Methodology
I Heavily reliant on fossil fuels Unified Propulsion
System Model
LUCAS Framework
I Air transportation projected to grow 4 – 5% annually for next 20 years Technology
Component Modeling
doubles every 15 years FAA, Airbus, Boeing
Cruise-Only
Approximation

Full-Mission
Commuter

Modeling
Annual Lessons
Traffic
Open Questions
(RPK)

Image source: Airbus Current Market Forecast 2019

A. Uranga (USC) 4 / 47
Controlling Emissions from Aviation Requires Radical Changes A. Uranga

2010 ICAO Assembly Commitments Introduction

Methodology
Unified Propulsion
System Model
LUCAS Framework
Technology
Component Modeling

Cruise-Only
Approximation

Full-Mission
Commuter

Modeling
Lessons

Open Questions

Source: “The Right Flightpath to Reduce Aviation Emissions”, position paper, Air Transportation
Action Group, UNFCCC Climate Talks, Cancun, Mexico, 2010
A. Uranga (USC) 5 / 47
Candidate Technology: Propulsion System Electrification A. Uranga

Aircraft electrification Introduction

Methodology
(= use electrical components as major elements in propulsion system ) Unified Propulsion

can help reduce emissions from aviation System Model

LUCAS Framework
Technology

I Alleviate dependence on fossil fuels: batteries carry electrical energy Component Modeling

Cruise-Only
Approximation
I Electricity production from variety of sources (some low-emissions)
Full-Mission
Commuter
I Leverage high efficiency levels of electrical components Modeling
Lessons

I Can facilitate use of beneficial technologies: Open Questions

distributed propulsion (DP) and boundary layer ingestion (BLI)

A. Uranga (USC) 6 / 47
Electrification Benefit: Higher Conversion Efficiency A. Uranga

Introduction

Methodology
Unified Propulsion
System Model
LUCAS Framework
Technology
Component Modeling

Cruise-Only
Approximation

Full-Mission
Commuter

Modeling
Lessons

Open Questions

Image source: Hepperle 2012

A. Uranga (USC) 7 / 47
Electrification Benefit: Distributed Propulsion (DP) A. Uranga

Introduction
DP provides potential weight reduction or larger fan area Methodology
3/2 Unified Propulsion

(cube-squared scaling: mprop ∼ ṁprop ) System Model

LUCAS Framework
Technology
Component Modeling
Distributed Propulsion:
Same weight, 1.6× more fan area Cruise-Only
Approximation

Full-Mission
Commuter

Modeling
Lessons
Single
propulsor Distributed Propulsion: Open Questions
Same fan area, half the weight

Assumption: Electrification facilitates DP

Kruger et al., AIAA 2018-4227, AIAA Aviation Forum 2018

A. Uranga (USC) 8 / 47
Electrification Benefit: Boundary Layer Ingestion (BLI) A. Uranga

Introduction
BLI reduces wasted kinetic energy in combined wake+jet Methodology
Unified Propulsion
⇒ Lower power requirement for a given forward force System Model
LUCAS Framework
Technology
Zero Net Wasted Component Modeling
Momentum Kinetic Energy
Cruise-Only
Approximation
- wake, or “draft” +
Full-Mission
+ propulsor jet + Commuter

Modeling
Lessons
- + Open Questions
+ combined wake and jet +
- +

Assumption: Electrification facilitates BLI

Uranga et al., AIAA Journal 2017

A. Uranga (USC) 9 / 47
Electrification Challenges A. Uranga

The weight of energy Introduction

Methodology
I Battery specific energy is ∼2 orders of magnitude lower than hydrocarbon fuel Unified Propulsion
System Model
e fuel ' 12 000 W·h/kg versus e bat ' 175 – 250 W·h/kgtoday LUCAS Framework
Technology

Aircraft re-design needed Component Modeling

Cruise-Only
I Electrified propulsion not beneficial if just swapped in place of conventional systems Approximation

Full-Mission
Commuter
Nascent technologies
Modeling
Lessons
I Low TRLs not “aerospace-ready”
Open Questions
I Operational and safety challenges
I Well-to-wake analysis may not yield emissions benefit for electrified aircraft,
but trends are promising and electicity generation is becoming greener (oil is not)

A. Uranga (USC) 10 / 47
Terminology A. Uranga

Introduction
I Conventional: propulsion achieved with hydrocarbon fuel as
Methodology
sole energy source and mechanically-driven fans Unified Propulsion
System Model
LUCAS Framework
I Turbo-electric: powered only from gas turbines burning fuel Technology
Component Modeling
I Fully turbo-electric: all fans electrically driven
Cruise-Only
I Partial turbo-electric: fans both electrically- and mechanically-driven Approximation

Full-Mission
Commuter
I Hybrid-electric: propulsion energy comes from both fuel via
Modeling
gas turbines and a battery Lessons

Open Questions
I All-electric: all propulsion energy provided by a battery

Electrified ≡ using electrical components as major elements in propulsion system

(electric motors, maybe batteries, . . . )

A. Uranga (USC) 11 / 47
Performance Metric: On-Board Energy Usage A. Uranga

On-board energy required to bring passengers from point A to point B Introduction

Methodology
Unified Propulsion
Mission energy: Productivity-Specific Energy Consumption System Model
LUCAS Framework
Technology

[total on-board energy] Etot Component Modeling

PSEC = = [ kJ/kg·km ] Cruise-Only

[payload mass] × [range] mPL × R Approximation

Full-Mission
Commuter

where on-board energy Etot = mfuel e fuel + mbat BSE Modeling

| {z } | {z } Lessons
fuel energy battery energy Open Questions

Considerations related to production of electrical energy to charge batteries on the

ground and corresponding chain for kerosene fuel delivery are outside the scope

A. Uranga (USC) 12 / 47
Outline A. Uranga

Introduction
1 Introduction
Methodology
Unified Propulsion

2 Methodology System Model

LUCAS Framework

Unified Propulsion System Model Technology

Component Modeling

LUCAS Framework Cruise-Only

Approximation
Technology Scenarios
Full-Mission
Modeling of Electrical Components Commuter

Modeling
3 Cruise-Only Approximation with Low-Fidelity Models Lessons

Open Questions

4 Full-Mission Commuter Results

5 Lessons Learned from Modeling Fidelities

6 Open Questions

A. Uranga (USC) 13 / 47
Electrified Propulsion System Architectures A. Uranga

Many flavors... Introduction

Methodology
Unified Propulsion
System Model
LUCAS Framework
Technology
Component Modeling

Cruise-Only
Approximation

Full-Mission
Commuter

Modeling
Lessons

Open Questions

Image source: NAE report 2016

A. Uranga (USC) 14 / 47
Unified Propulsion System Model A. Uranga

Source Load
(Consumed Power) (Useful Introduction
Mechanical Load Power) Methodology
Mechanically-Powered Unified Propulsion
Mechanical Source Propulsors System Model
Gas Turbines LUCAS Framework
Pturb PfanM Technology
PKM Component Modeling
turb Pgen fanM Cruise-Only
Link ×NfanM Approximation
×Nturb
Full-Mission
gen/ Commuter
mot
Modeling
Pconv Lessons

conv Open Questions

Electrical Source Electrical Load ×NfanE
Electrically-Powered Propulsors
Battery System
Plink Pinv Pmot PfanE
bat inv mot PKE
Pbat fanE

Pbat PKE
fS ≡ fL ≡
Pbat + Pturb PKE + PKM

A. Uranga (USC) 15 / 47
Unified Propulsion System Model: Electrification Metrics A. Uranga

Level of electrification defined based on propulsion-system power splits Introduction

Methodology
Unified Propulsion
At the source: Power supplied by gas turbine(s) (mechanical source) System Model
LUCAS Framework

and/or battery (electrical source) Technology

Component Modeling

Pbat fL Fully Cruise-Only

Source electrification factor fS ≡ turbo-electric
Approximation
Pbat + Pturb 1
Full-Mission
Commuter

Series hybrid Modeling

At the load: Flow power delivered via

Partial turbo-electric
Lessons

All-electric
mechanically-driven propulsors 0
Open Questions
=
and/or electrically-driven props. P lin
k

PKE Parallel hybrid

Load electrification factor fL ≡
PKE + PKM
0 fS
0 1
Conventional

A. Uranga (USC) 16 / 47
Outline A. Uranga

Introduction
1 Introduction
Methodology
Unified Propulsion

2 Methodology System Model

LUCAS Framework

Unified Propulsion System Model Technology

Component Modeling

LUCAS Framework Cruise-Only

Approximation
Technology Scenarios
Full-Mission
Modeling of Electrical Components Commuter

Modeling
3 Cruise-Only Approximation with Low-Fidelity Models Lessons

Open Questions

4 Full-Mission Commuter Results

5 Lessons Learned from Modeling Fidelities

6 Open Questions

A. Uranga (USC) 17 / 47
Methodology: LUCAS Framework A. Uranga

LUCAS : Library for Unified Conceptual Aircraft Synthesis Introduction

Methodology
I General framework for analysis and design of conventional and novel aircraft Unified Propulsion
System Model
LUCAS Framework

I Power-balance approach for aero-propulsive performance estimation Technology

Component Modeling

Cruise-Only
I Unified propulsion system model Approximation

Full-Mission
I Models for wide variety of aircraft configurations, Commuter

propulsion system architectures Modeling

Lessons
technologies (DP, BLI) Open Questions

I Relies on SUAVE (mission analysis, aerodynamics, weights)

and pyOptSparse for optimization
SUAVE references: Lukaczyk et al. AIAA 2015-3087, 2015; Botero et al. AIAA 2016-1275, 2016

A. Uranga (USC) 18 / 47
Outline A. Uranga

Introduction
1 Introduction
Methodology
Unified Propulsion

2 Methodology System Model

LUCAS Framework

Unified Propulsion System Model Technology

Component Modeling

LUCAS Framework Cruise-Only

Approximation
Technology Scenarios
Full-Mission
Modeling of Electrical Components Commuter

Modeling
3 Cruise-Only Approximation with Low-Fidelity Models Lessons

Open Questions

4 Full-Mission Commuter Results

5 Lessons Learned from Modeling Fidelities

6 Open Questions

A. Uranga (USC) 19 / 47
Battery Specific Energy A. Uranga

Introduction
Theoretical Cell-level Pack-level
BSEpack = ηpack BSEcell Methodology
BSEth ηcell BSEcell = ηcell BSEth ηpack Unified Propulsion
= ηcell ηpack BSEth System Model

electrochemical reactants
∼ 30% electrodes, electrolyte,
60−80% cells, wiring, casing,
LUCAS Framework
Technology
separator, collectors, thermal management Component Modeling
cell structure
Cruise-Only
Approximation

Full-Mission
Theoretical and cell-level BSE Commuter

Modeling
I Cell-level BSE is too often quoted Lessons
in literature and sales pitches Open Questions

I Specific energy at pack level is

the relevant parameter for
aircraft-level performance
BSEpack ≈ 0.2 BSEth
[mature] [mature] [mature] [Boeing 787 APU] [Airbus E-Fan] [Tesla Model S]

A. Uranga (USC) 20 / 47
 Motors
Electric Component Technology Trends 25
Current Aircraft A. Uranga
Automobiles
20 COTS Motors

Specific Power (kW/kg)

Batteries Introduction
Aircraft Studies
2200 In development (NASA Funded) Methodology
15
In Development *
Data From Literature NAE Report Unified Propulsion
2000 Cryogenic System Model
Boeing SUGAR Volt
10
Tesla Motors, Rate 5%/ year Basis for Technology Scenarios LUCAS Framework
1800
Tesla Motors, Rate 8%/ year Technology
Friedrich 2015, Rate 5.5%/ year 5 Component Modeling
1600
Cruise-Only
BSE (W-hr/kg)

1400
40% of theoretical for Li-air 0 Approximation
10 100 1000 10000
1200 Continuous Power (kW) Full-Mission
Commuter
1000
Converters Modeling
800 30
Lab demonstrated for Li-air Lessons
600 25 Open Questions

Specific Power (kW/kg)

In Development *
400 20

40% of theoretical for Li-ion

200 15
Automobiles
0 Aircraft Studies
10
2010 2015 2020 2025 2030 2035 2040 In development (NASA funded)
EIS Year NAE Report
5
Cryogenic
Basis for Technology Scenarios
0
100 1000 10000
Continuous Power (kW)
A. Uranga (USC) 21 / 47
Technology Scenarios A. Uranga

Introduction

Methodology
Current Conser- Interme- Optimistic Unified Propulsion
vative diate 2035 System Model
LUCAS Framework
2035 2035 Technology
Component Modeling

Battery specific energy (pack) [W·h/kg] 175 250* 575 900** Cruise-Only
Approximation
Battery specific power [kW/kg] 0.52 0.745 1.7 2.7
Full-Mission
Electrical machines s.p. [kW/kg] 2 9 12 16 Commuter

Modeling
Power electronics s.p. [kW/kg] 2.3 9 14 19 Lessons

Thermal management syst. s.p. [kW/kg] 13.2 13.2 13.2 13.2 Open Questions

Electrical component efficiencies [-] 0.95 0.98 0.99 0.99

s.p. = specific power

* Current Li-ion chemistries
** Novel Li-ion (Li-S or Li-air) chemistries
A. Uranga (USC) 22 / 47
Outline A. Uranga

Introduction
1 Introduction
Methodology
Unified Propulsion

2 Methodology System Model

LUCAS Framework

Unified Propulsion System Model Technology

Component Modeling

LUCAS Framework Cruise-Only

Approximation
Technology Scenarios
Full-Mission
Modeling of Electrical Components Commuter

Modeling
3 Cruise-Only Approximation with Low-Fidelity Models Lessons

Open Questions

4 Full-Mission Commuter Results

5 Lessons Learned from Modeling Fidelities

6 Open Questions

A. Uranga (USC) 23 / 47
Electric Propulsion System Components A. Uranga

Introduction

Methodology
Battery Converter Motor Fan Unified Propulsion
𝜂𝜂conv 𝜂𝜂mot System Model
LUCAS Framework

Specific Energy, Specific Power Technology

Specific Power Component Modeling

Component Low Fidelity Higher Fidelity Cruise-Only

Approximation
Battery Energy, power, mass Capacity, discharge behavior
Full-Mission
⇒ specific energy/power (current, voltage, power) Commuter
Motor Power, mass, efficiency Angular speed, voltage, current,
Modeling
⇒ specific power physical dimensions, magnetic properties Lessons
Power Distribution Lumped Safe operating voltages Open Questions
Converter Power, mass, efficiency Switching topologies, resistances,
⇒ specific power voltage, current
Wiring Lumped Length, cross-sectional area
voltage, current, efficiency
Thermal management Energy flow rate, mass Temperature, heat exchanger
system (TMS) ⇒ specific power (density, thermal properties)

Component masses calculated based on specific energy/power

A. Uranga (USC) 24 / 47
Battery Model A. Uranga
Capture non-linear battery discharge dynamics under variable power loads
Introduction
I Constant current model: NASA X-57 Maxwell Cell Discharge Profile Methodology
Unified Propulsion
V = V0 − KQ − RI − GIQ System Model

I Calculate V0 , K , R, G from curves LUCAS Framework

Technology
I But flight segments discharge at (piecewise) Component Modeling

constant power ⇒ constant-power model (P = IV ) Cruise-Only

Approximation
V = V0 − KQ − RP/V − GQP/V
Full-Mission
Commuter
V = Vn − K̃ (Q − Qn )
Modeling
K + GP/Vn Lessons
K̃ =
1 − RP/Vn2 − GPQn /Vn2 V0 : Open source (no load) voltage Open Questions
  K: Defines primary dependency of
1
q
Vn = (V0 − KQn ) + (V0 − KQn )2 − 4(RP + GPQn ) voltage and capacity discharged
2 R: Internal resistance
G: Change in slope of discharge curve
I Energy (dE = V dQ): area under the V vs. Q line due to current
V: Voltage under load
I Use initial & final points: (Qi , Vi ) & (Qf , Vf ) I: Current
   Q: Discharge capacity
Qi + Qf
∆E = Vn − K̃ − Qn (Qf − Qi )
2

A. Uranga (USC) 25 / 47
Motor Model: Switched-Reluctance Motor A. Uranga
Model changes in efficiency at different power levels in various flight segments
Introduction
φa φa
φa 2 Methodology
2 +
FP2 − F
Unified Propulsion
System Model

2RSP1 LUCAS Framework

Technology
Component Modeling
RSY1 2RG1 RSY1
Cruise-Only
2RRP1
Approximation
FP1

RRY1
Full-Mission
RRY1 Commuter

Modeling
I Represent magnetic flux φ paths as a network of reluctances R and magnetomotive forces F Lessons
I Use total flux to calculate motor inductance L (⇒ average torque Tavg ) Open Questions
I Output power: Pout,mot = Tavg ωmot
I Copper losses due to resistance of the stator winding wire: PCu = I 2 Rcoil
I Core losses due to hysteresis, eddy-current, excess losses
X X
Pk mk = Phys,k + Peddy,k + Pexs,k ρk Vk


Pcore =

I Motor input power: Pmot,in = Pmot,out + PCu + Pcore

I Motor efficiency: ηmot = Pmot,out /Pmot,in
A. Uranga (USC) 26 / 47
Converter Model A. Uranga
Model changes in efficiency at different power levels in various flight segments
Introduction
(Can be replaced with low-fidelity model with minimal loss in accuracy)
Methodology
Unified Propulsion
System Model
LUCAS Framework
Diode kd0 VD Technology
L RL kd0 I
+ − Component Modeling
+ − +
VL IC Vconv,in RL kd Ron kd0 RD +

+ + +
Vconv,out Cruise-Only
Vconv,in +
Switch C Rload Vconv,out Switch Diode Rload Approximation
−
−
−
−
kd0 Vconv,out
− − Full-Mission
Commuter

Modeling
I Use equivalent circuit model to find relationship between input and output voltage (⇒ power) Lessons
" #
Open Questions
1 kd02 Rload
Vconv,out = 0 Vconv,in − kd0 VD


kd kd02 Rload + RL + kd Ron + kd0 RD

I Duty cycle: k 0 = 1 − kd
d
I Input and output power: Pconv,in = Vconv,in I , Pconv,out = Vconv,out k 0 I
d
I Efficiency ηconv = Pconv,out
Pconv,in

A. Uranga (USC) 27 / 47
Power Distribution and Paschen’s Law A. Uranga

Set a safe operating voltage based on physical limits Introduction

Methodology
I Breakdown voltage: voltage required to create discharge Unified Propulsion
or electric arc between electrodes 10 6 System Model
LUCAS Framework
B pd Technology
Vbd = " # Component Modeling

[V]
10 5
ln A pd  Cruise-Only

bd
ln 1+ γ1
Approximation

Breakdown Voltage, V
I Absolute worst case for air: Full-Mission
10 4
Commuter
Vbd = 327 V at pd = 0.760 Pa.m
I Current aircraft voltages fall below this Modeling
Lessons
10 3
I For insulated conductors Open Questions
d
I Fraction of voltage across air gap: fV =
ti
d+ εr 10 2
10 -1 10 0 10 1 10 2 10 3
Vbd p.d [Pa.m]
I Safe operating voltage: SOV =
fV Vbd : Breakdown voltage
I SOV can be much higher than breakdown voltages both p: Ambient pressure
d: Separation between electrodes
at ground-level and at cruise A, B, γ: Constants specific to air
ti : Insulation thickness
εr : Dielectric constant of insulation

A. Uranga (USC) 28 / 47
Cables and Wiring A. Uranga
Determine if wiring mass is important in propulsion system modeling
Introduction
⇒ Wiring mass not significant with distributed propulsion
Methodology
Unified Propulsion
I Wires supply power to the motors: Pmot,in = I V = I · SOV System Model
LUCAS Framework

I Select diameters from American Wire Gauge (AWG) based on current Technology
Component Modeling

needed D y prop clear,fuse

Cruise-Only
Lcable Propulsor
Approximation
I Cable resistance: Rcable = Wing

σ Acond y clear,prop
Full-Mission
hfuse Fuselage Commuter
I Cable length calculated from aircraft geometry
wfuse Wiring Modeling
I Power dissipated in the cable: Pcable = I 2 Rcable Lessons
I Wiring mass: mass of conductors + insulation Battery Open Questions
π π h i 
2
mwiring = 1.2 ρcond Dcond Lcable + ρins (2 ti + Dcond )2 − Dcond
2
Lcable
4 4 I: Current
SOV: Safe operating
I Factor of 1.2 for mass of fixing components (clamps, mounts, ...) voltage
Lcable : Cable length
Dcond : Conductor diameter
ρ: Density
σ: Conductivity
ti : Insulation thickness

A. Uranga (USC) 29 / 47
Thermal Management System (TMS) A. Uranga
Size heat exchanger to reject wasted heat from electrical components
Introduction

Methodology
I Energy wasted from electrical components: Unified Propulsion
Q̇comp = (1 − ηcomp ) Pcomp,in System Model
LUCAS Framework

I Total heat to be rejected: Q̇ =

P η: Efficiency
i Q̇comp,i
Technology
Component Modeling
I TMS: heat exchanger with hot side and cold side Pcomp,in : Input power Cruise-Only
I Modeled with number of thermal units (NTU) approach to component Approximation

Full-Mission
Waste heat from components Hot air Commuter
Wall
Th,in Tc,out Modeling
Lessons
Hot side Cold side
Open Questions
Electrical
components do di Di
Rh Rw Rc

Th,out Tc,in
Coolant to components Cross section for double pipe heat exchanger
Cold air

A. Uranga (USC) 30 / 47
Outline A. Uranga

Introduction
1 Introduction
Methodology
Unified Propulsion

2 Methodology System Model

LUCAS Framework

Unified Propulsion System Model Technology

Component Modeling

LUCAS Framework Cruise-Only

Approximation
Technology Scenarios
Full-Mission
Modeling of Electrical Components Commuter

Modeling
3 Cruise-Only Approximation with Low-Fidelity Models Lessons

Open Questions

4 Full-Mission Commuter Results

5 Lessons Learned from Modeling Fidelities

6 Open Questions

A. Uranga (USC) 31 / 47
Low-Fidelity Framework Overview A. Uranga

Introduction
I Generalized range equation for cruise-only mission
Methodology
I Energy- and power-to-mass scalings for propulsion system components Unified Propulsion
System Model

I Structural sizing based on correlations LUCAS Framework

Technology
Component Modeling

Cruise-Only
Aero-Propulsive Propulsion Approximation
Performance System
Full-Mission
Commuter
Geometric Programming
using GPKit Modeling
Lessons

Open Questions
Aero-Structural Mission
Sizing Integration

GPKit: Burnell, Damen, and Hoburg, “GPkit: A Human-Centered Approach to Convex Optimization in
Engineering Design”, 2020 CHI Conference on Human Factors in Computing Systems

A. Uranga (USC) 32 / 47
Mission Determines Optimal Architecture A. Uranga

Optimistic 2035 Tech

Introduction
(900 W·h/kg BSE)
I All-electric best for very short ranges Methodology
Unified Propulsion
System Model
(∼ 200 nmi) LUCAS Framework
Technology

I Turbo-electric seems best for long ranges Component Modeling

Cruise-Only
(>400 nmi) Approximation
... but higher-fidelity results may show Full-Mission
Commuter
otherwise
Modeling
Lessons
I Hybrid-electric may have a niche at
Open Questions
intermediate ranges, e.g. to extend range
of otherwise all-electric option
I All-electric for thin-haul missions is
unfeasible with current and conservative
2035 tech

A. Uranga (USC) 33 / 47
Range Limitations of All-Electric Architectures A. Uranga

Introduction
All-electric optimistic 2035 Tech
Methodology
Unified Propulsion
I Large sensitivity to range System Model
LUCAS Framework
Technology
Component Modeling
I All-electric could give large
Cruise-Only
benefits at low ranges, Approximation

but advantage quickly drops with Full-Mission

Commuter
range due to battery mass growth Modeling
Lessons

Open Questions

A. Uranga (USC) 34 / 47
Electrification Benefits: Technology Effects A. Uranga

All-electric 100 nmi mission Introduction

8
7 Methodology
6 Conventional, 2 fans, no BLI I BSE improvements crucial to making Unified Propulsion
System Model

all-electric feasible
PSEC [kJ/kg km]

5 LUCAS Framework

4 Technology
All-electric, 2 fans, no BLI
Component Modeling
3
I Diminishing returns past 900 W·h/kg BSE
Conservative 2035

Intermediate 2035

All-electric, 100 fans,50% BLI

2 Cruise-Only
Optimistic 2035
Current Tech

Approximation
1
0
0 250 500 750 1000 1250 1500 1750 2000 I Component powers less limiting than BSE Full-Mission
Commuter
Battery Specific Energy, BSE [W h/kg]
with diminishing returns past 5 kW/kg Modeling
8
Lessons
7
6 Conventional, 2 fans, no BLI I DP beneficial, but diminishing returns past Open Questions

∼ 6 – 8 propulsors
PSEC [kJ/kg km]

5
4
All-electric, 2 fans, no BLI
3 All-electric, 100 fans, 50% BLI
I DP and BLI facilitate electrification:
Conservative 2035

Intermediate 2035

2
Optimistic 2035
Current Tech

1 larger missions become feasible

0
0 5 10 15 20 25
Component Specific Powers [kW/kg]

A. Uranga (USC) 35 / 47
Outline A. Uranga

Introduction
1 Introduction
Methodology
Unified Propulsion

2 Methodology System Model

LUCAS Framework

Unified Propulsion System Model Technology

Component Modeling

LUCAS Framework Cruise-Only

Approximation
Technology Scenarios
Full-Mission
Modeling of Electrical Components Commuter

Modeling
3 Cruise-Only Approximation with Low-Fidelity Models Lessons

Open Questions

4 Full-Mission Commuter Results

5 Lessons Learned from Modeling Fidelities

6 Open Questions

A. Uranga (USC) 36 / 47
Commuter Aircraft Mission: 19 Pax, 250nmi range A. Uranga

Baseline for comparisons: advanced* technology version of DHC-6 Twin Otter Introduction

Methodology
* (-15% empty mass; -20% in engine SFC and power-to-mass ratio) Unified Propulsion
System Model
LUCAS Framework
Technology
Component Modeling

Cruise-Only
Number of passengers Npax 19 Approximation

Full-Mission
Payload mpay 1 842 kg (3 979 lb) Commuter

Modeling
Mission range R 250 nmi (463 km) Lessons

Open Questions
Cruise speed Vcruise 94 m/s

Takeoff mass minit 5 670 kg

Cruise altitude hcruise 10 000 ft (3 050 m) Image ©Viking Air

A. Uranga (USC) 37 / 47
Electrification Design Space - Optimistic Tech Assumptions A. Uranga

Introduction

Methodology
Unified Propulsion
System Model
LUCAS Framework
Technology
Component Modeling

Cruise-Only
Approximation

Full-Mission
Commuter

Modeling
Lessons

Open Questions

I Energy requirement drops quickly as source electrification (fS ) increases

(more energy in battery vs fuel)
I Parallel hybrids slightly more efficient than series hybrids
I Aircraft weight grows rapidly as electrification increases

A. Uranga (USC) 38 / 47
Hybrid-Electric: 500 nmi Extended Range Mission A. Uranga

Mission-fixed hybrid: Constant electrification throughout mission. Introduction

Methodology
Mission-varying hybrid: Different electrification levels for different mission Unified Propulsion
System Model

segments. LUCAS Framework

Technology
Component Modeling

Cruise-Only
Configuration Tech Climb fS Cruise fS ∆MTO ∆PSEC Approximation

Full-Mission
Fixed hybrid Interm. 0.55 0.55 +59% -27% Commuter
Opt. 0.9 0.9 +52% -63% Modeling
Lessons
Varying hybrid Opt. 1 0 +15% -3.7%
Open Questions

I Potential energy benefits even at intermediate technology

I Fixed hybrid leads to largest energy benefits, but with large weight gains
I Varying hybrids mitigate weight gains at the expense of energy benefits

A. Uranga (USC) 39 / 47
Conclusions of Full Mission Commuter Study A. Uranga

All-electrics are fundamentally range limited and best suited for very short missions Introduction

Methodology
while hybrids can be more versatile Unified Propulsion
System Model
LUCAS Framework

If battery specific energy improves to intermediate tech levels (BSE∼600 W·h/kg) Technology
Component Modeling

I All-electric architectures could lead to large energy reductions Cruise-Only

Approximation
for very short range missions (100 nmi) but with big aircraft weight gains Full-Mission
Commuter
I Hybrid-electric architectures
Modeling
I Give large energy reductions with high electrification levels Lessons
I Can significantly increasing feasible mission range compared to all-electrics Open Questions
I Keep aircraft weight gains relatively low

Electrification benefits are primarily due to high efficiency of electrical components

⇒ higher electrification is more beneficial at the expense of weight gains

A. Uranga (USC) 40 / 47
Outline A. Uranga

Introduction
1 Introduction
Methodology
Unified Propulsion

2 Methodology System Model

LUCAS Framework

Unified Propulsion System Model Technology

Component Modeling

LUCAS Framework Cruise-Only

Approximation
Technology Scenarios
Full-Mission
Modeling of Electrical Components Commuter

Modeling
3 Cruise-Only Approximation with Low-Fidelity Models Lessons

Open Questions

4 Full-Mission Commuter Results

5 Lessons Learned from Modeling Fidelities

6 Open Questions

A. Uranga (USC) 41 / 47
Fidelity Levels A. Uranga

I Level 1: Range equation Introduction

I One-liner, hand calculations Methodology
I Order of magnitude checks with minimal info Unified Propulsion
System Model
LUCAS Framework
Technology
I Level 2: Cruise-only with low-fidelity models (specific powers and efficiencies) Component Modeling

I Need to determine appropriate ‘full-mission average’ values a priori Cruise-Only

Approximation
I Missing sizing effects (since components not always sized for cruise)
Full-Mission
Commuter
I Level 3: All mission segments with low-fidelity models
Modeling
I Realistic sizing and constraints Lessons

I No need to define ‘mission average’ values Open Questions

I Relies on a priori values for component characteristics

I Level 4: All mission segments and detailed component models

I No need to guess component characteristics
I Reduces uncertainty in final answers
I Helps ensure resulting aircraft are realistic
A. Uranga (USC) 42 / 47
Modeling Lessons A. Uranga

Higher-fidelity mission analysis (Level 3) is needed Introduction

Methodology
for proper sizing (i.e. components meet the full mission requirements) Unified Propulsion

... otherwise propulsion system size may be underestimated System Model

LUCAS Framework

... snow-balls into significantly undersized airplane Technology

Component Modeling

... and conclusions on feasibility and benefits may be way off Cruise-Only
Approximation

Higher-fidelity component models teach us that Full-Mission

Commuter
I Battery dominates propulsion system mass for all electrics (∼ 60%) Modeling
Lessons
I Higher safe operating voltages may be possible than currently certified Open Questions

I Non-battery component efficiencies matter mostly as they affect battery mass

A. Uranga (USC) 43 / 47
Outline A. Uranga

Introduction
1 Introduction
Methodology
Unified Propulsion

2 Methodology System Model

LUCAS Framework

Unified Propulsion System Model Technology

Component Modeling

LUCAS Framework Cruise-Only

Approximation
Technology Scenarios
Full-Mission
Modeling of Electrical Components Commuter

Modeling
3 Cruise-Only Approximation with Low-Fidelity Models Lessons

Open Questions

4 Full-Mission Commuter Results

5 Lessons Learned from Modeling Fidelities

6 Open Questions

A. Uranga (USC) 44 / 47
Open Questions A. Uranga

Aircraft electrification can definitely reduce on-board energy requirements Introduction

but whether it is environmentally beneficial or not ultimately depends on Methodology

Unified Propulsion
System Model
I Electricity production and distribution: strongly dependent on how/where LUCAS Framework
Technology
e.g. electric flight from Seattle to Indianapolis may be green, Component Modeling

but return flight might pollute more than on conventional aircraft Cruise-Only
Approximation

I Environmental impact of battery and other electronics production, disposal Full-Mission

Commuter

I Battery life-cycle considerations e.g. replace every 100 versus 1000 flights Modeling
Lessons

I ... etc ... Open Questions

Analysis of broader life-cycle emissions impact is crucial

and ultimately needs to be entered into the design equation

(well-to-wake emissions modeling in progress)

A. Uranga (USC) 45 / 47
References A. Uranga

Downloads available from the ADRL’s site at https://adrl.usc.edu/publications/ Introduction

Methodology
M. Kruger, S. Byahut, A. Uranga, J. Gonzalez, D.K. Hall, and A. Dowdle, “Electrified Aircraft Unified Propulsion
System Model
Trade-Space Exploration”, AIAA 2018-4227, 2018 AIAA Aviation, Atlanta, GA, June 25–29. LUCAS Framework

2018. doi: 10.2514/6.2018-4227 [low-fideity trade-space exploration] Technology

Component Modeling

Cruise-Only
S. Byahut and A. Uranga “Propulsion Powertrain Component Modeling for an All-Electric Approximation
Aircraft Mission”, AIAA 2020-0015, AIAA SciTech, Orlando, FL, Jan. 6–10. 2020 Full-Mission
doi: 10.2514/6.2020-0015 [component modeling part 1] Commuter

Modeling
Lessons
M. Kruger and A. Uranga “The Feasibility of Electric Propulsion for Commuter Aircraft”, AIAA
Open Questions
2020-1499, AIAA Scitech, Orlando, FL, Jan. 6–10. 2020 doi: 10.2514/6.2020-1499
[commuter study]

S. Byahut and A. Uranga “Power Distribution and Thermal Management Modeling for Electrified
Aircraft”, AIAA 2020-3578, AIAA/IEEE Electric Aircraft Technologies Symposium (EATS),
Virtual Event, Aug. 26–28. 2020. doi: 10.2514/6.2020-3578 [component modeling part 2]

A. Uranga (USC) 46 / 47
Contact A. Uranga

Introduction
Alejandra Uranga, Ph.D. Methodology
Unified Propulsion
Gabilan Assistant Professor System Model
LUCAS Framework
Aerospace & Mechanical Engineering Dept. Technology
Component Modeling
University of Southern California
Cruise-Only
Approximation
Email: [email protected]
Full-Mission
Phone: (213) 821 - 0846 Commuter

Web: uranga.usc.edu Modeling

Lessons

Lab site: adrl.usc.edu Open Questions

A. Uranga (USC) 47 / 47