Evaluation of power quality and common design concept for AC-DC converters in aircraft

UPTEC E 17 011

Examensarbete 30 hp Januari 2018

Evaluation of power quality

and common design concept for AC-DC converters in aircraft

Andreas Brolund

Teknisk- naturvetenskaplig fakultet UTH-enheten

Besöksadress:

Ångströmlaboratoriet Lägerhyddsvägen 1 Hus 4, Plan 0

Postadress:

Box 536 751 21 Uppsala

Telefon:

018 – 471 30 03

Telefax:

018 – 471 30 00

Hemsida:

http://www.teknat.uu.se/student

Abstract

Evaluation of power quality and common design concept for AC-DC converters in aircraft

Andreas Brolund

This master thesis has been carried out in collaboration with Saab, Avionics Systems in Jönköping, Sweden, during the spring of 2017. The thesis investigates unidirectional rectifier topologies in aircraft and the focus has been

on evaluating the power quality requirements according to the aircraft standards, in the course of the More Electric Aircraft concept. Both passive and

active power factor correction topologies are considered, discussed and compared.

Simulation models are designed in MATLAB/Simulink and the procedures are presented. A modular concept regarding components is discussed where different power supplies and loads are considered. The simulations present both a passive 12-pulse auto-transformer rectifier unit and an active Delta-switch rectifier fulfilling requirements for aircraft such as the total harmonic distortion of the supply current. In addition, the input power factor

is close to unity and an efficiency greater than 97% is obtained. Lastly, future aspects of each topology are discussed and necessary improvements to obtain realistic simulation models are presented.

ISSN: 1654-7616, UPTEC E 17 011 Examinator: Mikael Bergkvist Ämnesgranskare: Markus Gabrysch Handledare: Ingemar Thörn

Popul¨arvetenskaplig sammanfattning

Detta examensarbete ¨ar utf ¨ort p˚a och i samarbete med Saab, Avionics Systems, J ¨onk ¨oping, Sverige under v˚aren 2017.

I examensarbetet utv¨arderades enkelriktade (kraftfl ¨odet kan bara g˚a i en rikt- ning, fr˚an k¨alla till last) topologier f ¨or likriktning av v¨axelstr ¨om f ¨or anv¨andning i elektriska kraftsystem i flygplan. Likriktningen skedde fr˚an 115-230 V v¨axelsp-

¨anning med en frekvens p˚a 360-800 Hz till 270-540 V liksp¨anning med en re- sistiv last p˚a 2-10 kW. En passiv topologi (ej kontrollering av likriktningsproces- sen) baserad p˚a en 12-pulsig automatiserad transformator samt en aktiv topo- logi (kontrollering av likriktningsprocessen) baserad p˚a en delta-konfigurerad halvledare-anordning simulerades och utv¨arderades. Fr˚an litteraturstudier och simuleringsresultat drogs slutsatsen att den passiva topologin i dagsl¨aget ¨ar den b¨asta topologin f ¨or likriktning d˚a den erbjuder h ¨og p˚alitlighet, robusthet och effektivitet. Den aktiva topologin visade p˚a goda resultat n¨ar det kom till el-kvalit´e. I och med p˚ag˚aende framsteg och utvecklingar inom kraftelektro- nik och mikro-processorer, vilket ¨ar en stor del inom den aktiva topologin, ¨ar det ber¨attigat att s¨aga att den aktiva topologin ¨aven kommer vara aktuell f ¨or likriktning inom en snar framtid. Simuleringsmodellerna var f ¨orenklade upp- skattningar av hur det verkliga systemet s˚ag ut och fungerade. Detta ledde till att det tyv¨arr var sv˚art att dra n˚agon slutsats ang˚aende huruvida ett modul¨art koncept f ¨or likriktning givet olika in- och uteffekter kunde uppn˚as eller ej. Det modul¨ara konceptet skulle inneb¨ara att trots givet olika applikationer med oli- ka tekniska specifikationer skulle kunna anv¨anda en likriktnings-modul som fungerade f ¨or samtliga fall, vilket ¨aven var en utav fr˚agest¨allningarna. F ¨or att kunna f˚a mer realistiska simuleringsfall beh ¨ovdes fler parametrar och variab- ler tas med i modellen. Exempel p˚a dessa parametrar var temperaturskillna- der, kosmisk str˚alning och lastprofiler som i samtliga simuleringsfall borts˚ags ifr˚an. Framtida passiva och aktiva topologier som kommer vara av intresse ¨ar 18 och 30-pulsig automatiserad transformatorer och Vienna-likriktare baserad p˚a en konfiguration av sex transistor. Slutligen drogs slutsatsen att mjukvaran, MATLAB/Simulink, som anv¨andes f ¨or simuleringar ¨ar ett utav de b¨asta verk- tygen p˚a marknaden n¨ar det kommer till analys, bed ¨omning, validering och f ¨orb¨attring av b˚ade kraftelektronik och elektriska system.

Utifr˚an utv¨arderingen av enkelriktade likriktnings-topologier i detta examens- arbete kan slutsatser dras om vilka topologier som kan vara intressanta att g˚a vidare med samt vilka f ¨orb¨attringar som m˚aste g ¨oras f ¨or att kunna anv¨anda de framtagna simuleringsmodellerna f ¨or mer realistiska fall. En ¨overblick kan f˚as p˚a framtida topologier v¨arda att forska mer inom vid intresse.

Att f ¨orb¨attra denna typ av enkelriktad likriktning var av intresse f ¨or att mins- ka st ¨orningar och f ¨or att h ¨oja el-kvalit´en i det elektriska systemet p˚a flygplan.

Detta d˚a elektriska st ¨orningar och l˚ag el-kvalit´e ger f ¨ors¨amrad prestanda och sliter mer p˚a komponenterna vilket i sin tur ger l¨agre h˚allbarhet och f ¨ors¨amrad p˚alitlighet. Slutligen leder detta till ¨okade underh˚allskostnader och ¨okad vikt p˚a grund av att filtrering och blockering av st ¨orningar beh ¨ovs. F ¨or att ta re-

da p˚a om en likriktare har bra prestanda finns det olika krav och riktlinjer att f ¨olja f ¨or el-kvalit´e p˚a flygplan. Dessa ¨ar sammanfattade i standard DO-160F och MIL-STD-704F d¨ar den f ¨orstn¨amnda oftast anv¨ands f ¨or civilt flyg och den sistn¨amnda oftast anv¨ands f ¨or milit¨art flyg. Dessa standarder ¨ar sammanfattade i rapporten d˚a de anv¨andes f ¨or validering och utv¨ardering av simuleringsmo- dellerna.

F ¨or att minska vikt, underh˚allskostnader, f ¨orb¨attra p˚alitligheten samt prestan- da av flygplan introducerades ett koncept ben¨amnt More Electric Aircraft (mer elektriska flygplan) p˚a tidigt nittiotal. Konceptet inneb¨ar att ers¨atta tunga me- kaniska, pneumatiska och hydrauliska system mot elektriska system. Detta led- de till ¨okad efterfr˚agan p˚a l ¨osningar av elektriska system som kan driva och f ¨ors ¨orja olika laster p˚a flygplan som tidigare varit drivna av mekaniska, pneu- matiska och hydrauliska system. Det ledde ¨aven till utvidgning av det elekt- riska systemet p˚a flygplan, vilken i sin tur ledde till att god el-kvalit´e blev ¨an viktigare.

Ut ¨over m ¨ojlig f ¨orb¨attring av el-kvalit´e ville Saab ¨aven unders ¨oka och utv¨ardera enkelriktade likriktare p˚a flygplan f ¨or att de olika in- och uteffekterna som anv¨andes i dagsl¨aget i elkraftsystemet p˚a flygplan resulterade i olika l ¨osningar f ¨or olika fall, vilket ledde till ¨okade kostnader. Om ett modul¨art koncept kunde fastst¨allas skulle det vara av intresse f ¨or Saab.

Tillv¨agag˚angss¨attet var att f ¨orst utf ¨ora en genomg˚aende litteraturstudie och marknadsunders ¨okning av aktuella topologier f ¨or likriktning, men ¨aven f ¨or att f˚a en ¨overgripande f ¨orst˚aelse av den elektriska systemet p˚a flygplan. Informa- tion h¨amtades fr˚an rapporter, artiklar, journaler och informativa webbplatser men ¨aven en del fr˚an interna dokument p˚a Saab. Utifr˚an detta presenterades och utv¨arderades de mest trov¨ardiga och intressanta topologierna d¨ar ¨aven spe- cifikationerna, givna av Saab, och framtida aspekter f ¨or likriktare var med och p˚averkade. Utifr˚an utv¨arderingen av de olika topologierna valdes tv˚a vidare f ¨or ytterligare utv¨ardering i simuleringsmilj ¨o. Modellerna simulerades f ¨or olika in- sp¨anningar, utsp¨anningar, frekvenser och laster i hopp om att n˚agon slutsats ang˚aende det modul¨ara konceptet kunde dras samt f ¨or att validera modellerna med avseende p˚a el-kvalit´e. Slutligen utv¨arderades och diskuterades simule- ringsresultaten samt presenterades framtida koncept och slutsatser.

U PPSALA U NIVERSITY M ASTER T HESIS

Evaluation of power quality and common design concept for AC-DC

converters in aircraft

Author:

Andreas B

Examiner:

Mikael B

Supervisor:

Ingemar T

Subject reader:

Markus G

A thesis submitted in fulfillment of the requirements for the degree of Master of Science at the

Department of Engineering Sciences

October 21, 2017

i

“Happiness only real when shared”

Christopher McCandless

ii

Uppsala University

Abstract

Disciplinary Domain of Science and Technology Department of Engineering Sciences

Master of Science

Evaluation of power quality and common design concept for AC-DC converters in aircraft

by Andreas B^ROLUND

This master thesis has been carried out in collaboration with Saab, Avionics Systems in Jönköping, Sweden, during the spring of 2017. The thesis inves- tigates unidirectional rectifier topologies in aircraft and the focus has been on evaluating the power quality requirements according to the aircraft stan- dards, in the course of the More Electric Aircraft concept. Both passive and active power factor correction topologies are considered, discussed and com- pared. Simulation models are designed in MATLAB/Simulink and the proce- dures are presented. A modular concept regarding components is discussed where different power supplies and loads are considered. The simulations present both a passive 12-pulse auto-transformer rectifier unit and an active

∆-switch rectifier fulfilling requirements for aircraft such as the total har- monic distortion of the supply current. In addition, the input power factor is close to unity and an efficiency greater than 97% is obtained. Lastly, fu- ture aspects of each topology are discussed and necessary improvements to obtain realistic simulation models are presented.

iii

Acknowledgements

I would like to thank my supervisor Ingemar Thörn for always taking the time to listen when I had questions and thoughts. Markus Gabrysch, my subject reader but also my second supervisor, for all the help with the simu- lation models. Jonas Dahlqvist for making sure I was integrated in into my new environment, for always checking up on me so that everything was okay and for being a role model when it comes to being a leader. I would also like to thank Mattias Johansson for all the help with troubleshooting simulation models and for the great discussion we had which made me think clearer.

Last but not least, Aldin Avdic for making my time in Jönköping more enjoy- able by being my friend and training partner day in and day out.

iv

Contents

1 Introduction 1

1.1 Background . . . 1

1.2 Objective and Goal . . . 5

1.3 Methodology . . . 5

1.4 Delimitations . . . 5

1.5 Disposition . . . 6

2 Power Quality Requirements of Aircraft 7 2.1 Standard MIL-STD-704F and DO-160G . . . 7

2.2 Voltage requirements . . . 8

2.3 Power and Frequency requirements . . . 11

2.4 Current and EMC requirements . . . 13

3 Rectifier Topologies in Aircraft 16 3.1 Auto-transformer rectifier unit . . . 17

3.2 Active power factor correction . . . 18

3.2.1 ∆-switch and Y-switch rectifier . . . 18

3.2.2 Vienna rectifier . . . 20

3.3 Summary and Comparison . . . 22

4 Design of Simulation Models 25 4.1 Power supply . . . 25

4.2 Load . . . 26

4.3 DC-link . . . 26

4.4 12-pulse ATRU . . . 27

4.4.1 Auto-transformer . . . 27

4.4.2 Diode bridge . . . 30

4.4.3 Interphase transformer . . . 30

4.5 ∆-switch rectifier . . . 33

4.5.1 Boost inductor . . . 33

4.5.2 Diode bridge and Switches . . . 33

4.5.3 Control of switches . . . 33

5 Result and Discussion 39 5.1 12-pulse ATRU . . . 39

5.2 ∆-switch rectifier . . . 44

6 Conclusion and Future Work 50

v

A Appendix 53

A.1 Defintions and Equations . . . 53

A.1.1 Average and Root Mean Square . . . 53

A.1.2 Efficiency . . . 54

A.1.3 Per-unit . . . 54

A.1.4 Mutual inductance . . . 54

A.2 Power Quality Requirements . . . 55

A.2.1 Conducted and radiated emission . . . 55

A.3 Rectifier topologies . . . 57

A.3.1 AT design topologies . . . 57

A.4 Simulation parameters, values and models . . . 58

Bibliography 61

vi

List of Figures

1.1 Electrical system evolution in aircraft . . . 1

1.2 Power distribution of aircraft . . . 2

1.3 Electrical power distribution of B767 and B787 . . . 3

1.4 Electrical power distribution system in B787 and A380 . . . . 4

2.1 Phase sequence and phase shift of supply voltage . . . 8

2.2 Distortion spectrum for 270 VDC systems . . . 10

2.3 Unity, lagging and leading power factor . . . 13

3.1 12-pulse ATRU . . . 17

3.2 ∆and Y-switch rectifier . . . 19

3.3 Switch-pair setup . . . 19

3.4 The Vienna rectifiers . . . 21

4.1 AT topology used in the design of simulation model . . . 27

4.2 Simulation model of the AT with connections . . . 28

4.3 Graphical representation and simulation model of the B6Us and IPTs . . . 31

4.4 Block diagram of the control of the switches . . . 34

4.5 Sector detection with and without phase shift . . . 37

5.1 12-pulse ATRU: Supply waveforms and FFT analysis of case 1 42 5.2 12-pulse ATRU: Supply waveforms and FFT analysis of case 1 43 5.3 ∆-switch: FFT analysis and over-modulation of case 4 . . . 47

5.4 ∆-switch: Supply waveforms and FFT analysis of case 6 . . . 48

5.5 ∆-switch: Supply waveforms and FFT analysis of case 8 . . . 49

A.1 Conducted emission limits . . . 55

A.2 Ratiated emission limits . . . 56

A.3 CM and DM filter . . . 56

A.4 Different 18-pulse AT topologies . . . 57

A.5 Zig-Zag AT topology . . . 57

A.6 Simulation model of voltage loop . . . 59

A.7 Simulation model of carrier voltage generation . . . 59

A.8 Simulation model of control loop . . . 60

vii

List of Tables

2.1 Voltage requirements for ACVF, ACCF and DC . . . 9

2.2 Power & frequency requirements for ACVF and ACCF . . . . 12

2.3 Definition of EMC ranges . . . 14

2.4 Current THD limits . . . 14

3.1 Comparison of ∆-switch, six-switch Vienna and 12-pulse ATRU 23 3.2 Comparison of B6U, 12-pulse ATRU and ∆-switch rectifier . . 23

4.1 Rating of coils in the AT . . . 27

4.2 Rating of the IPT . . . 32

4.3 Clamping action of the switches . . . 36

5.1 12-pulse ATRU: Simulation cases . . . 39

5.2 12-pulse ATRU: Simulation result . . . 40

5.3 12-pulse ATRU: Simulation parameters . . . 41

5.4 ∆-switch: Simulation cases . . . 44

5.5 ∆-switch: Simulation result . . . 44

5.6 ∆-switch: Simulation parameters . . . 45

5.7 ∆-switch: Optimized system parameters . . . 46

A.1 Rectifier bridge (B6U) parameters . . . 58

A.2 Switch parameters . . . 58

A.3 Default pu values . . . 58

viii

List of Abbreviations

AC Alternating Current

ACVF Alternating Current Variable Frequency ACCF Alternating Current Constant Frequency APU Auxiliary Power Unit

APFC Active Power Factor Correction Unit AT Auto-Transformer

ATRU Auto-Transformer Rectifier Unit

B6U Uncontrolled Three-phase Diode Bridge Rectifier

B6C Controlled Three-phase Thyristor/Transistor Bridge Rectifier circuit CE Conducted Emissions

CF Crest Factor

CM Common Mode

DC Direct Current DM Differential Mode

EMC Electromagnetic Compatibility EMI Electromagnetic Interference IPT Interphase Transformer IDG Internal Drive Generator L-L Line-to-Line

L-N Line-to-Neutral MEA More Electric Aircraft P Proportional

PF Power Factor

PI Proportional Integral PLL Phase Locked Loop P-P Peak-to-Peak

P-V Peak-to-Valley

PWM Pulse Width Modulation RE Radiated Emissions RF Ripple Factor RMS Root-Mean-Square

THD Total Harmonic Distortion TRU Transformer Rectifier Unit VAC Volt Alternating Current VDC Volt Direct Current

ix

List of Symbols

A_dist(f ) distortion RMS value dB V

f supply frequency Hz

f_sw switching frequency Hz

F_{P I} PI controller

CFV voltage crest factor

C_o DC-link capacitance F

C_s,D diode’s snubber capacitance F

Cs,S switch snubber capacitance F

D_{F +} positive Vienna rectifier bridge diode D_{F −} negative Vienna rectifier bridge diode Din negative ∆-switch bridge diode D_ip positive ∆-switch bridge diode

D_{M +} positive Vienna rectifier midpoint switch diode DM − negative Vienna rectifier midpoint switch diode D_{N +} positive Vienna rectifier mains switch diode D_{N −} negative Vienna rectifier mains switch diode

g^∗ reference conductance 1/Ω

i_s,p measured L-N supply current phase p A

i^∗_s,p reference L-N supply current phase p A

i_s,A measured L-N supply current phase A A

is,B measured L-N supply current phase B A

i_s,C measured L-N supply current phase C A

i(t) instantaneous current A

Iˆ₁ fundamental peak current A

Iˆ_ipt peak current of IPT coil A

Io output current A

I_{RM S} RMS current A

I_{RM S,n} RMS current of coil n A

IRM S,ipt RMS current of IPT coil A

I_s RMS L-N supply current A

Iˆ_s peak L-N supply current A

j imaginary unit

k coefficient of coupling k_boost ripple current factor

k_ipt peak current of IPT related to output current k_n number of coils of type n

K_I,v integral coefficient of voltage controller K_P,i proportional coefficient of current controller K_P,v proportional coefficient of voltage controller

x

L_a inductance of coil a H

Lb inductance of coil b H

L_base base inductance H

L_c inductance of coil c H

Ld inductance of coil d H

L_e inductance of coil e H

L_f inductance of coil f H

Lipt inductance of IPT coil H

L_m,AT magnetizing inductance of AT H

L_m,pu magnetizing inductance pu value

LN i boost inductance H

L_on,D conduction inductance diode H

L_on,S conduction inductance switch H

Lpu inductance pu value

m modulation index

M midpoint of Vienna rectifier

ML mutual inductance H

M_L,ipt mutual inductance of IPT H

M_R,ipt mutual resistance of IPT H

n type of coil in AT

p notation of phase of voltage and current

P active power W

P F power factor

P_in input active power W

P_o output active power W

Q reactive power VAr

R_a resistance of coil a Ω

R_b resistance of coil b Ω

Rbase base resistance Ω

R_c resistance of coil c Ω

R_d resistance of coil d Ω

Ript resistance of IPT coil Ω

R_m,AT magnetizing resistance of AT Ω

R_m,pu magnetizing resistance pu value RFV output voltage ripple factor

R_load load resistance Ω

R_on,D conduction resistance diode Ω

Ron,S conduction resistance switch Ω

R_pu resistance pu value

R_s,D diode’s snubber resistance Ω

Rs,S switch snubber resistance Ω

S complex power V A

|S| apparent power V A

SAT power rating of AT V A

S_base base complex power V A

S_ij switch ij

Si switch i

xi

S_i+ positive Vienna switch i Si− negative Vienna switch i

t time s

T period (reciprocal value of f ) s

Ts sample time s

T HD_I current total harmonic distortion T HD_V voltage total harmonic distortion

vo measured output voltage V

v_o^∗ reference output voltage V

v_r P-P voltage ripple V

v_rs,p^∗ reference L-N voltage at switch-pair phase p V v_r,ij^∗ reference L-L voltage over switch Sij V v_r,ji^∗ reference L-L voltage over switch Sji V

v_s,p measured L-N supply voltage phase p V

v_s,A measured L-N supply voltage phase A V

vs,B measured L-N supply voltage phase B V

v_s,C measured L-N supply voltage phase C V

v(t) instantaneous voltage V

vtri triangular carrier voltage V

Vˆ peak voltage V

V₁ fundamental RMS voltage V

V_a(t) instantaneous voltage phase a V

V_b(t) instantaneous voltage phase b V

V_base base voltage V

V_c(t) instantaneous voltage phase c V

V_f,D forward voltage diode V

V_f,S forward voltage switch V

V_n n^th harmonic RMS voltage V

V_{N i} supply voltage APFCs V

V_s RMS L-N supply voltage V

Vˆ_s peak L-N supply voltage V

V_o output voltage V

Vo,AC output AC voltage V

V_{AV G,o} AVG output voltage V

V_{RM S} RMS voltage V

VRM S,a RMS voltage of coil a V

V_{RM S,b} RMS voltage of coil b V

V_{RM S,ipt} RMS voltage of IPT coil V

V_{RM S,n} RMS voltage of coil n V

V_{RM S,o,AC} RMS output AC voltage V

X_base base reactance Ω

z complex variable of z-transform

Z_base base impedance Ω

Z_base,a base impedance of coil a Ω

Z_base,b base impedance of coil b Ω

Z_base,n base impedance of coil n Ω

Zbase,primary base impedance primary side of AT Ω

xii

Zbase,secondary base impedance secondary side of AT Ω

δ_a phase of voltage in phase a ^◦

δ_b phase of voltage in phase b ^◦

δc phase of voltage in phase c ^◦

φ phase shift between sinusoidal voltage and current ^◦

ω angular frequency rad/s

1

Chapter 1

Introduction

1.1 Background

The power of an aircraft is obtained from the turbines (the engines of the air- craft) and can be divided in two separate parts where the first is the propul- sive part and the second is the non-propulsive part. The non-propulsive part is used for different kinds of applications such as lights, cabin pressure, avionics, flight control and fuel pumps. The distribution of the power to these applications is on a conventional aircraft, such as Airbus A330 and Boe- ing B767, carried out by a combination of hydraulic, electric, pneumatic and mechanical systems.

FIGURE1.1: Evolution of the electrical power system in aircraft [1].

In recent years the development of the electrical power system have made much progress as seen in figure 1.1. A reason for this is the More Electric Air- craft (MEA) concept, introduced in the early nineties [2]. The first intention of

Chapter 1. Introduction 2

the concept was to reduce the overall weight, lower the maintenance costs, increase the reliability and to improve the performance of military aircraft by increasing the use of electronic equipment instead of heavy mechanical, pneumatic or hydraulic driven elements. With the increased capacity and rating of generators of civil aircraft the MEA concept also applied there. To- day, the objective is to completely replace the systems that in earlier aircraft (both military and civil) have been hydraulic, pneumatic or mechanical sys- tems with electrical systems [1]. An overview of the power distribution of the non-propulsive power of a conventional aircraft and a MEA is presented in figure 1.2a and 1.2b, respectively. In conventional aircraft, bleed air (com- pressed air) is taken from the turbine and used for pneumatic applications (e.g. heating and pressurization of the cabin) and considerably reduces the efficiency of the turbine [2]. Hence, in the course of the MEA concept, a no bleed system is preferable which in turn result in a higher demand of new electrical systems to supply power to these applications and loads.

(A)

(B)

FIGURE1.2: Power distribution of the non-propulsive power of (A) a conventional aircraft and (B) a MEA [2].

Chapter 1. Introduction 3

The electrical power system in both civil aircraft and in military aircraft is often based on a parallel connection of turbine-driven generators. In con- ventional (not MEA) and military aircraft the turbines are connected to inte- grated drive generators (IDG) via a mechanical gearbox resulting in a three- phase AC voltage with a constant frequency (ACCF) of 400 Hz and a root mean square (RMS) line-to-normal (L-N) voltage of 115 V as output to the main AC bus. The latest electrical power distribution system based on the MEA concept is used in Airbus A380 and Boeing B787. An overview of a con- ventional (B767) and a MEA (B787) electrical power distribution system is presented in figure 1.3 and in figure 1.4 a block representation of an electrical power system used in a MEA is presented. The MEA electrical power system includes six generators, two per turbine and two per auxiliary power unit (APU). The main generators (the ones connected to the turbines) are either of permanent magnet synchronous machine (PMSM) or switched reluctance machine (SRM) type. The aggregated rating of them is reaching up to 1 MVA.

The main generators outputs (to the main AC bus), a three-phase AC voltage with a variable frequency (ACVF) ranging between 360-800 Hz and a RMS L-N voltage of 115 V for the A380 and 230 V for the B787. The APU is also connected to the main AC bus and is activated in case of an emergency and for power supply on ground while the engines are not running. The variable frequency of the main AC bus is a result of the elimination of the mechanical gearbox between the turbine and the generators in accordance to the MEA concept. This results in a more reliable generation with less maintenance and lower costs compared to the complex constant speed drive generation used by the IDG in conventional aircraft [3]. The direct connection of the genera- tors to the turbine also allows the generator to act as a starter motor for the turbine, reducing the overall weight of the aircraft [2].

FIGURE 1.3: Electrical power distribution of the conventional B767 (left) and the MEA B787 (right) [3].

Chapter 1. Introduction 4

FIGURE1.4: Electrical power distribution system in MEA B787 and A380 [1].

The APU and the redundancy of the system makes sure that the most impor- tant components of the aircraft have power even under emergencies and fail- ures. Still, the reliability and the performance of the elements in the system is of great importance. The generation and the conversion of power needs to be robust and deliver a high power quality. There are several aircraft standards including standards created by aircraft companies which electrical power systems in aircraft need to comply with. The two general industrial stan- dards are the DO-160G [4] used for civil aircraft and the MIL-STD-704F [5]

used for military aircraft (also applicable on older/conventional civil air- craft). As seen in figure 1.4, the modern MEA electrical power distribution system consist of several voltage buses supplying several different loads.

To obtain the voltage for each bus different kinds of converters are needed.

Commonly, the 28 VDC bus is obtained by a transformer rectifier unit (TRU), the secondary AC bus at 115 VAC (not necessary in A380) is obtained by an auto-transformer unit (ATU) and finally the 270 VDC bus (or in the B787 a 540 VDC bus) is obtained by a auto-transformer rectifier unit (ATRU). The loads of each bus differ but common loads are actuators used for both pri- mary and secondary flight-control [2]. The actuators load demand varies with the mission profile and are often in need of a low continuous power supply during flight and a high peak power supply during starting and land- ing of the aircraft.

The last mentioned conversion, i.e. the rectification of the 115/230 VAC sup- ply to 270/540 VDC, is a part of a product that Saab [6] offers to their cus- tomers. In this thesis an investigation of this type of rectification on aircraft is carried out and the focus will be on evaluating (in a simulation environment) the fulfillment of the power quality requirements according to the mentioned standards. Both passive and active topologies are considered, discussed and compared. Simulation models are designed and presented. Lastly, a modu- lar concept is discussed where different power supplies and loads are con- sidered.

Chapter 1. Introduction 5

1.2 Objective and Goal

The different types of power supplies used in aircraft leads to new design so- lutions for each converter case which in turn drives cost. Therefore, it would be advantageous to keep certain design elements (e.g. inductors, capacitors and ratings) common independently of input power and obtain a modular concept. Further on, as the development of power systems and electronics moves forward in accordance to the MEA concept the topology used for rec- tification today has to be evaluated considering power quality, weight and reliability. Hence, the goal of the master thesis is to investigate topologies for rectification on-board aircraft and perform simulations in order to evalu- ate power quality and establish to what extent design elements can be kept common for a rectifier used for aircraft utility equipment, i.e if a modular concept can be obtained. Both short duration (peak power) drive, continu- ous drive and different power inputs are simulated. Figures of merit such as weight, reliability and cost should also be considered and finally future rectifier topologies and aspects should be discussed.

1.3 Methodology

First, a literature study and a market research was done to get an insight in existing airborne power system and more specific in the topologies used to rectify AC to DC in aircraft. The information was obtained from white papers, articles, journals and informative web pages but also from earlier projects and internal documents at Saab. Three topologies for rectification were investigated further and compared regarding price, size, weight, relia- bility and performance.

From analysis and evaluation of the literature study and the market research two of the three investigated topologies were chosen to be modeled for sim- ulation in MATLAB/Simulink. Simulations with different input voltage and frequency were done to establish to what extent components could be kept common, independently of the input. To validate the performance of the models, power quality requirements and characteristics defined in two stan- dards (one for civil and one for military aircraft) were used.

Finally, the simulation result was evaluated and summarized to conclude if a common modular design concept could be obtained and how good the power quality was. Future aspect and topologies were discussed and an overall evaluation of the whole procedure of the master thesis was done.

1.4 Delimitations

The thesis only investigates the rectification for following input power sources

• Variable frequency 360-800 Hz, 230 VAC (L-N, RMS)

Chapter 1. Introduction 6

• Constant frequency 400 Hz, 115 VAC (L-N, RMS) The output voltage and power range are limited to

• 270 VDC and 540 VDC

• Continuous power ≤ 2 kW

• Peak power ≤ 10 kW

Due to secrecy of the aircraft companies, the models are only evaluated using the following general standards

• DO-160G

• MIL-STD-704F

1.5 Disposition

The power quality requirements obtained from the aircraft standards are summarized and presented in chapter 2. The outcome of the literature study, the market research and a summary of rectifier topologies in aircraft are pre- sented in chapter 3. The most promising rectification topologies from chap- ter 3 are designed for simulation and each step of the modeling are presented in chapter 4. The result of the simulations, both during validation and with different inputs, are discussed and presented in chapter 5. Finally, a conclu- sion is drawn considering the simulation result and the overall thesis out- come. Also future aspects and work are discussed in chapter 6.

7

Chapter 2

Power Quality Requirements of Aircraft

2.1 Standard MIL-STD-704F and DO-160G

The power quality requirements and characteristics used in today’s aircraft are described in MIL-STD-704F [5] and DO-160G [4] and are both widely used for verification of electrical applications. The MIL-STD-704F standard is for military and older civil aircraft and is more strict when it comes to voltage characteristics compared to the DO-160G standard but does not cover EMC and voltage spikes. The DO-160G standard is for newer civil aircraft and is more comprehensive and covers test procedures as well. An extract from the scopes of the standards are presented as follows.

MIL-STD-704F

This standard establishes the requirements and characteristics of aircraft electric power provided at the input terminals of electric utilization equipment. The pur- pose of this interface standard is to ensure compatibility between the aircraft electric system, external power, and airborne utilization equipment.

DO-160G

This document defines a series of minimum standard environmental test conditions (categories) and applicable test procedures for airborne equipment. The purpose of these tests is to provide a laboratory means of determining the performance charac- teristics of airborne equipment in environmental conditions representative of those which may be encountered in airborne operation of the equipment.

In the next three sections, requirements stated in the standards concerning voltage, frequency, power, current and quality in VAC (ACVF and ACCF) and VDC are summarized and the parameters and limits used for the devel- opment and validation of the simulation models are presented.

Chapter 2. Power Quality Requirements of Aircraft 8

2.2 Voltage requirements

The voltage supply should be a three-phase grounded neutral system with a phase sequence of A-B-C as illustrated in figure 2.1 and the supply voltage should be a sine wave fulfilling the specified voltage quality requirements in the standards. To get an overview of these requirements a summary and a comparison between the two standards for VAC and VDC are presented in table 2.1. To obtain an estimation of the limits for 230 VAC the values in table 2.1 could be multiplied by two [4]. It is assumed that a multiplication by two of the 270 VDC limits in table 2.1 is a valid estimation for the 540 VDC limits as well.

FIGURE2.1: Phase sequence and phase shift of supply voltage with a counterclockwise phase rotation [5].

Some of the main voltage quality requirements for VAC is the distortion fac- tor and crest factor. The distortion factor of VAC is equal to the total harmonic distortion of the voltage (T HDV). The THD is defined as the ratio of the RMS value of the harmonics (i.e exclusive of fundamental) to the RMS value of the fundamental [4], [5]. The T HDV is formulated in equation 2.1. For a pure sine wave the THD is equal to zero since no harmonics are present.

T HD_V =

pP∞ n=2V_n²

V₁ · 100% (2.1)

V₁ =fundamental RMS voltage V_n=n^th harmonic RMS voltage

The crest factor of the voltage (CFV) is defined as the ratio of the absolute peak value to the RMS value. A pure voltage sine wave has a CF equal to

VRM S

√ 2 V_{RM S} =√

2and a pure DC voltage (no peaks) has CF equal to 1. The CFV

is formulated in equation 2.2.

Chapter 2. Power Quality Requirements of Aircraft 9 TABLE2.1: Steady state voltage requirements for ACVF, ACCF

and DC [4], [5].

ACVF (115 V, 360-800 Hz) MIL-STD-704F DO-160G

RMS L-N Supply voltage 108-118 V 100-122 V

RMS voltage unbalance ≤3.0 V ≤8.0 V

RMS voltage modulation ≤2.5 V ≤5.0 V, P-V

Voltage phase difference 120^◦±4^◦ 120^◦±6^◦

Distortion factor ≤5% ≤10%

Crest factor 1.41±0.10 1.41±0.15

DC component ±0.10 V ±0.20 V

ACCF (115 V, 400 Hz) MIL-STD-704F DO-160G

RMS L-N Supply voltage 108-118 V 100-122 V

RMS voltage unbalance ≤3.0 V ≤6.0 V

RMS voltage modulation ≤2.5 V ≤5.0 V, P-V

Phase difference 120^◦±4^◦ 120^◦±4^◦

Distortion factor ≤5% ≤8%

Crest factor 1.41±0.10 1.41±0.15

DC component ±0.10 V ±0.20 V

DC (270 V) MIL-STD-704F DO-160G

DC Voltage 250-280 V 220-320 V

Distortion factor 0.015 N/A

Ripple amplitude ≤6.0 V ≤8.0 V

CF = | ˆV |

V_{RM S} (2.2)

V =ˆ peak voltage V_{RM S} =RMS voltage

For VDC, ripple is the main topic when concerning voltage quality (in ad- dition to correct voltage amplitude). Ripple is defined as the variation of voltage about the steady state DC voltage during steady state operation [5].

The requirements stated in the two standards covers distortion factor, dis- tortion spectrum and ripple amplitude. The distortion factor is equal to the ripple factor of the output voltage of the rectifier (RFV). The RF is defined as the ratio of the RMS value of the alternating voltage component on the DC voltage to the DC steady state voltage [4], [5]. The RFV is formulated in equation 2.3. The ripple amplitude Vripple is the maximum absolute value of the difference between the steady state and the alternating voltage com- ponent on the DC voltage [5]. The Vripple is formulated in equation 2.4. The allowed ripple amplitude of a certain frequency component is defined in the distortion spectrum illustrated in 2.2. With use of the spectrum the allowed

Chapter 2. Power Quality Requirements of Aircraft 10

peak-to-peak voltage ripple (vr) at a switching frequency fsw can be calcu- lated as in equation 2.5. This is used later for a first approximation of the DC-link capacitance in chapter 4, section 4.3.

RF_V = VRM S,o,AC

V_o (2.3)

V_ripple = | ˆV_o,AC| (2.4)

V_o,AC = V_o− V_{AV G,o}

V_o,AC =output AC voltage V_o=output voltage V_{AV G,o} =AVG output voltage V_{RM S,o,AC} =RMS output AC voltage

FIGURE 2.2: Maximum distortion spectrum for 270 VDC sys- tems [5].

vr= 2√

2 · 10^Adist(fsw)20 (2.5) A_dist(f_sw) =distortion RMS amplitude

Chapter 2. Power Quality Requirements of Aircraft 11

2.3 Power and Frequency requirements

The power in an electrical system can be described using different quanti- ties. For an AC-system, the complex power is often considered. The complex power (S) is the complex sum of the active and reactive power and is defined as in equation 2.6 [7].

S = P + jQ (2.6)

P =active power Q =reactive power

j =imaginary unit

The apparent power (|S|) is for an AC system the magnitude of the complex power but can also be expressed for systems with distortion via the RMS voltage and the RMS current as in equation 2.7 [4].

|S| = V_{RM S}I_{RM S} (2.7)

V_{RM S} =RMS voltage I_{RM S} =RMS current

For AC-systems, the active power (P ) and the reactive power (Q) is the real and imaginary part of the complex power, respectively, as seen in equa- tion 2.6. The formulas defining these two are expressed in equation 2.8 and equation 2.9, respectively [7].

P = |S| cos(φ) (2.8)

Q = |S| sin(φ) (2.9)

|S| = apparent power

φ =phase shift between sinusoidal voltage and current

The active power is generally defined as the average of the product of the instantaneous voltage and current, as expressed in equation 2.10

P = 1 T

Z T 0

v(t)i(t)dt (2.10)

T =period (reciprocal value of f ) v(t) =instantaneous voltage

i(t) =instantaneous current t =time

The power factor (PF) is defined as the ratio of the active power (P ) flowing to the load to the apparent power (|S|) in the system, as expressed in equa- tion 2.11. It describes how efficiently the system transfer the active power and since aircraft power system does not allow bidirectional power flow the PF can be a value between 0 and 1 (if bidirectional power flow is allowed the value is between -1 and 1) [4], [5]. How the PF affects the system is easier to

Chapter 2. Power Quality Requirements of Aircraft 12

see if equation 2.7 is inserted in equation 2.11 and is rewritten with regards to the current, as in equation 2.12.

P F = P

|S| (2.11)

I_{RM S} = P

VRM SP F (2.12)

P =active power

|S| = apparent power VRM S =RMS voltage

I_{RM S} =RMS current

Equation 2.12 shows that a PF close to unity (1) will lead to a lower RMS- value of the current which in turn implies lower ohmic losses and that is why unity power factor is desired in electrical systems [8]. Unity PF is achieved when there is no distortion and the phase shift between voltage and current is equal to zero and this occurs when the system behaves essentially like a resistor. If the phase shift is negative the PF is said to be lagging and this oc- curs when the system is inductive (current is lagging the voltage). Similarly, the PF is said to be leading if the phase shift is positive and occurs when the system is capacitive (current is leading the voltage). A PF equal to unity, a lagging PF and a leading PF are illustrated in figure 2.3.

A summary and a comparison between the two standards concerning power and frequency requirements for VAC are presented in table 2.2. The PF value from MIL-STD-704F is when operating at 50 percent or more of its rated load current and for a load greater than 500 VA. When operating at more than 100 VA the system shall not have a leading PF [5]. The PF value from DO- 160G is during full load and for a load greater than 150 VA [4].

TABLE2.2: Steady state power and frequency requirements for ACVF and ACCF [5][4].

ACVF (115 V, 360-800 Hz) MIL-STD-704F DO-160G

Frequency 360-800 Hz 360-800 Hz

Power Factor (Lag) ≥0.85 ≥0.8

Power Factor (Lead) 1 ≥0.968

ACCF (115 V, 400 Hz) MIL-STD-704F DO-160G

Frequency ±7 Hz ±10 Hz

Power Factor (Lag) ≥0.85 ≥0.8

Power Factor (Lead) 1 ≥0.968

Chapter 2. Power Quality Requirements of Aircraft 13

FIGURE2.3: Unity, lagging and leading power factor.

For a discretized system there is a sufficient condition to make sure that the discrete sequence of samples capture all of the information from a continuous (time) signal of defined bandwidth. This condition is defined by the Nyquist- Shannon sampling theorem [9]. For a certain switching frequency the sample time should be according to equation 2.13 to avoid aliasing.

T_s≤ 1

2f_sw (2.13)

T_s=sampling time f_sw =switching frequency

2.4 Current and EMC requirements

There are different types of emission that have to be considered when deal- ing with power electronics on aircraft. The requirements concerning these emissions are divided in three frequency ranges as presented in table 2.3.

The requirements regarding frequencies up to 150 kHz are covered by the current harmonics limitations, presented in table 2.4. The limits are defined up to 40th harmonic which correspond to a frequency of 32 kHz considering

Chapter 2. Power Quality Requirements of Aircraft 14 TABLE2.3: Definition of frequency range and requirement con-

cerning EMC [4].

Frequency range [MHz] Requirement

0 - 0.150 Current harmonics

0.150 - 100 Conducted emission

100 - 6000 Radiated emission

the maximum supply frequency of 800 Hz, hence limits of current harmon- ics above 32 kHz are not defined. The second frequency range deals with the conducted emission (CE) and covers the common mode (CM) and differ- ential mode (DM) currents and the last range is the radiated emission (RE).

Limits of the CE and RE are presented in appendix A.2, figure A.1 and A.2, respectively.

TABLE 2.4: Limits of harmonics in phase currents [4]. ˆI1 is the amplitude of the fundamental peak current.

Harmonic Order Limit

3^rd, 5^th, 7^th 0.02 ˆI₁

Odd Triplen Harmonics (n = 9, 15, 21, ..., 39) 0.1 ˆI1/n

11^th 0.1 ˆI₁

13^th 0.08 ˆI1

Odd Non Triplen Harmonics 17, 19 0.04 ˆI₁ Odd Non Triplen Harmonics 23, 25 0.03 ˆI1

Odd Non Triplen Harmonics 29, 31, 35, 37 0.3 ˆI₁/n Even Harmonics 2 and 4 0.01 ˆI1/n Even Harmonics > 4 (n = 6, 8, 10, ..., 40) 0.0025 ˆI₁

Harmonics > 40 N/A

The occurrence of emission disturbances in power equipment is inevitable due to the fact of non-ideal components. To suppress and minimize the dis- turbance the design and placement of components and modules has to be considered. This includes the choice of using surface mounting or through hole mounting, minimizing cross-section areas of sensitive nodes and loca- tion of components such as X and Y capacitors (described below), just to mention a few [10]. Apart from the design and placement, current harmon- ics are suppressed by harmonic injection and active/passive filter, conducted emission with the help of EMI filter and radiated emission by shielding.

The EMI filter often suppress both CM and DM currents. The part of the EMI filter suppressing CM currents (CM filter) often consists of a choke coil and a line bypass capacitor (Y) connected to the chassis ground and sup- presses disturbances which are conducted on all lines in the same direction, sometimes referred to as asymmetrical interference. The part of the EMI filter

Chapter 2. Power Quality Requirements of Aircraft 15

suppressing DM currents (DM filter) often consists of a across-the-line capac- itor (X) and suppresses disturbances which are conducted on the signal line and ground line in the opposite direction to each other, often referred to as symmetrical interference [11][10]. A graphical representation of a CM and DM filter are illustrated in appendix A.2, figure A.3.

16

Chapter 3

Rectifier Topologies in Aircraft

To meet the requirements of today’s aircraft standards basic rectifier topolo- gies such as the uncontrolled (passive) three-phase diode bridge rectifier (B6U) and the controlled (active) three-phase thyristor/transistor bridge rec- tifier (B6C) are not sufficient without any additional circuits [12]. The power quality of the B6U is too low due to the high input current THD and low PF and the B6C is a bidirectional rectifier meaning it allows power flow in both directions which is prohibited in aircraft. Additionally, the reliability of the B6C is too low due to the risk of a shoot-through resulting in a short circuit of the DC-link [12]. These disadvantages and prohibitions can be avoided by implementing additional circuits and complex rectification topologies. For aircraft, the conventional topology to rectify AC to DC today is the passive transformer rectifier unit (TRU) and in the latest aircraft such as B787 and A380 an auto-transformer rectifier unit (ATRU) is used [1]. In the future, ac- tive solutions such as the active power factor correction unit (APFC) might be an alternative as well.

The ATRU utilize the B6U together with an auto-transformer (AT), interphase transformers (IPT) and a DC-link capacitor resulting in a robust system with improved power quality compared to the basic B6U [12].

The APFC can be implemented using different design topologies: ∆-switch, Y-switch, Vienna, Buck, Boost and SWISS just to mention a few. The typically active rectifiers are comprised of a B6C and a DC-link capacitor where the B6C is controlled to obtain a unity power factor at the supply. The APFC topologies are often instead based on B6Us together with a switch-configura- tion. The slightly more complex design and control can still assure that no shoot-through occurs and that a constant DC-link voltage together with a unity power factor are obtained.

In the following sections the ATRU and the most promising APFC for air- craft rectification according to [2], [12]–[15] are investigated. Also, a sum- mary where the discussed topologies are compared and evaluated consider- ing power quality, weight, reliability and cost will be presented.

Chapter 3. Rectifier Topologies in Aircraft 17

3.1 Auto-transformer rectifier unit

The ATRU is a passive rectifier comprised of an AT, one or several B6Us de- pending on topology and two IPTs. According to [12] the employment of an AT instead of an isolated transformer reduces the weight and volume signif- icantly. As mentioned previously, there exist several design topologies of the AT, ranging from 6 to 30-pulses using winding configurations such as zig- zag, ∆ and hexagon just to mention a few. In this thesis the focus will be on the 12-pulse topology and for that topology, the conventional winding con- figuration is the ∆-configuration. An overview of a full 12-pulse ATRU sys- tem based on this winding configuration is illustrated in figure 3.1. In addi- tion, three examples of 18-pulse AT topologies based on the ∆-configuration and a 12-pulse AT topology based on zig-zag configuration are presented in appendix A.2.1, figure A.4 and A.5, respectively. Further on, only the 12- pulse topology will be discussed.

FIGURE3.1: 12-pulse ATRU based on a ∆-configured AT [16].

The AT divides the three-phase input A, B, C into two separate three-phase systems 1,3,5 and 2,4,6, with a 30^◦ phase shift. The two systems are rectified by the B6Us and then switched in parallel using IPT#1 and IPT#2 resulting in a DC-voltage with a 12-pulse ripple on the output [12]. The 5th and 7th harmonics of the input current can be significantly reduced in an ATRU. How effectively the reduction is depends on the design topology and how close to rated power it is operating at [17]. Some of the challenges in the construction and design procedure are

• Choosing winding material

• Selecting modular shape

• Defining and optimizing core geometry and aspect ratio

• Optimizing interactions between windings (leakage inductance, prox- imity effects)

To maximize the cancellation of the harmonics, fractional turns are required which in reality are approximated by full turns. Further, the ∆-configuration

Chapter 3. Rectifier Topologies in Aircraft 18

can be used to circulate the triple harmonics produced in the rectification process. Design and topology decision is explained and evaluated in [16]

and [12] and therefore shall be omitted here for the sake of brevity.

The main advantages of the ATRU are the reliability and robustness due to the lack of active electronic components. It also shows high overload capacity which is advantageous when it comes to applications requiring high peak-to- continuous power demand [12], [14], [18]. According to [16], industrial and aerospace power quality requirements, such as DO-160, can be satisfied with a 12-pulses ATRU. Additionally the ATRU does not need a common-mode filter at the input, due to the high common mode impedance [12], [17].

The main disadvantages of the ATRU are the weight and the unregulated output voltage. The high weight is due to the auto-transformer which ac- counts for approximately 50% of the total weight [12]. The unregulated out- put voltage is due to the passive attribute, which in turn results in voltage droops as the ATRU gets loaded. The droop is a function of the input source impedance and the leakage reactance of the transformer and can result in substantial voltage droops during full load [17]. The ATRU does neither of- fer soft start nor over current protection and an input differential-mode EMI- filter needs to be added [17], [18]. Additionally, the power factor of a passive solution, such as the ATRU, cannot be controlled. Without EMI-filter, the power factor is a function of loading and is typically lagging for an ATRU.

With the EMI-filter the power factor can be made to range from leading to lagging depending on load, though leading power factor is not desired in power systems [17].

3.2 Active power factor correction

3.2.1 ∆-switch and Y-switch rectifier

The ∆-switch and the Y-switch rectifier are both active three-phase two-level unidirectional boost rectifiers. An circuit overview of the ∆-switch and the Y-switch is presented in figure 3.2a and 3.2b, respectively. The topologies are comprised of one inductor LN i per phase, a B6U Dip, Din with a bidi- rectional (current) bipolar (voltage) switch-configuration Sij (i, j ∈ {1, 2, 3}) connected between the two diodes of the B6U on each phase-leg and finally a DC-link capacitor Co. The switch-configuration can be realized by two tran- sistors (a switch-pair) in series on each phase-leg, as illustrated in figure 3.3.

The switch-pair on each phase-leg are connected in a ∆-coupling between the phases for the ∆-switch and in a Y-coupling between the phases for the Y-switch, hence the names. The individual switches are denominated as (top to bottom, left to right) S12, S21, S23, S32, S31and S13. In [12] it is stated that the conduction losses of the ∆-switch topology are lower compared to the Y-switch topology and therefore only the ∆-switch topology is considered from now on.

Chapter 3. Rectifier Topologies in Aircraft 19

(A)

(B)

FIGURE 3.2: APFC topologies (A) ∆-switch and (B) Y-switch both with metal oxide semiconductor field effect transistors

(MOSFETs) as switching devices [2].

FIGURE 3.3: Switch-pair setup where i, j ∈ {1, 2, 3} denotes each individual switch [2].

The three-phase input VN i is rectified by controlling the three switch-pairs Sij and Sji in a manner that keeps the DC-link voltage constant and the in- put power factor as close to unity as possible. The diodes in the rectifier Dip

and Din make sure a short-circuit of the DC-link does not occur but on the other hand they have to commutate at switching frequency [2]. In order to minimize switching losses due to this, diodes with a low reverse-recovery current are essential. In addition, choosing control strategies for the switches has to be done carefully. A hysteresis control may increase the effort of EMI- filtering and a constant switching frequency control may result in too high supply current THD if the switching frequency is too low [15]. A survey cov- ering different control strategies is presented in [19] and therefore shall be omitted here for the sake of brevity.

Chapter 3. Rectifier Topologies in Aircraft 20

The presence of the inductors on the AC side has an impact on several con- verter properties. It gives a continuous input current and the attribute of a boost rectifier which implies an output voltage Vo with a relation to the sup- ply voltage VN i as expressed in equation 3.1 [2]. The right hand side is the peak value of the line-to-line supply voltage. Having an output voltage of at least this peak-line-to-line supply voltage will guarantee that the system does not get over-modulated. If over-modulation occur, it could lead to distortion in the supply current [2].

V_o ≥√ 3√

2V_{N i} (3.1)

Hence, the inductors are often referred to as boost inductors. The total weight of the ∆-switch rectifier is mainly determined by them, since they are approx- imately 50% of the total weight [12]. However, the capacitor needed in the EMI-filter is small due to the presence of the boost inductors [15].

The main advantages of the ∆-switch rectifier are the active attribute en- abling a power factor at unity and a constant DC-link voltage enhancing the overall power quality of the rectifier [12]. Additionally, the ∆-switch rectifier has a high reliability and low complexity compared to other APFC topolo- gies and can handle a phase loss without changing controller structure [15].

The main disadvantage of the ∆-switch rectifier, or APFC topologies in gen- eral, is the EMI topic [12]. They need both DM and CM-filtering, which is a disadvantage concerning weight and volume. However, it is thought that the development and improvement of power electronics and computational power of digital signal processors will enable a reduction of filter compo- nents [17]. In addition, boost solutions require additional circuitry for soft start-up to minimize the inrush current [13].

3.2.2 Vienna rectifier

The Vienna rectifier is an active three-phase three-level unidirectional boost rectifier [20]. A circuit overview is presented in figure 3.4a. The topology is comprised of one inductor LN iper phase with the same purpose as the ones used in the ∆-switch rectifier, a B6U DF +, DF −with an advantageously inte- grated bidirectional switch-configuration Si (i ∈ {1, 2, 3}) on each phase-leg and finally two DC-link capacitors Co at the output creating the three-level property.

The switch-configuration is comprised of a transistor as switch and a diode bridge. The diode bridge is comprised of two diodes DN +, DN −connected to the supply and two diodes DM +, DM −connected to the midpoint M between the output capacitors Co. During operation, the current flows through two diodes for every switching state which implies conduction losses. In order to reduce the losses a topology called six-switch Vienna rectifier can be used [2].

A circuit overview of this topology is presented in figure 3.4b. Compared to the original Vienna rectifier the diodes DM +, DM −in the switch-configuration connected to the midpoint M and the switch Siare replaced by a switch-pair

Chapter 3. Rectifier Topologies in Aircraft 21

S_i+, Si−. Considering all three phases, it sums up to a total of six switches, hence the name six-switch Vienna rectifier. Considering the reduced conduc- tion losses the focus will be on the six-switch topology from now on.

(A)

(B)

FIGURE3.4: APFC topologies (A) original Vienna and (B) six- switch Vienna with MOSFET as switching device [2].

Chapter 3. Rectifier Topologies in Aircraft 22

The operation of the six-switch Vienna rectifier uses the same approach as the

∆-switch rectifier, i.e the rectification, the constant DC-link voltage and the unity power factor are achieved by controlling the switch-pairs. The diodes connected to the supply DN + and DN − are commutated with supply fre- quency meanwhile the diodes in the B6U DF + and DF − commutates with switching frequency. Similarly to the ∆-switch rectifier, diodes with a low reverse-recovery current are needed for the B6U, whereas for all other diodes standard rectifier diodes can be applied [2]. Due to the boost attribute, the minimal output voltage Voin relation to the supply voltage VN iis also for this topology as stated in equation 3.1, according to [2]. Finally, the main advan- tages and disadvantages of the six-switch Vienna rectifier are also similarly to the ones for the ∆-switch topology. Though, according to [15] the efficiency of the six-switch Vienna rectifier is slightly lower compared to the ∆-switch rectifier for a certain power supply and load as presented in table 3.1. The ef- ficiency here and further on in the report refers to the efficiency of a electrical system as defined in appendix A.1.

3.3 Summary and Comparison

In [15] a comparison between a constructed ∆-switch and six-switch Vienna rectifier is carried out and the outcome is presented in table 3.1. In the same table, values obtained for a 12-pulse ATRU from [18] are also presented. A slightly higher efficiency for the ∆-switch topology is achieved compared to the Vienna topology given the specified supply and load. Though, [15]

also state that the six-switch Vienna rectifier might be more efficient for a higher supply voltage (e.g. 230 VAC). The ATRU shows the highest efficiency, though the given value is for different supply voltage and output voltage than for the active topologies. The biggest different can be seen on the sup- ply current THD between the active and passive topology. This is also veri- fied in a comparison carried out in simulation environment of a conventional B6U, a 12-pulse ATRU and a ∆-switch rectifier done in [12]. The results are presented in table 3.2. Apart from the power quality which supply current THD is a major part of, there is a close race between the passive topology (ATRU) and the active ∆-switch topology. Considering the development in power electronics which will improve the reliability and lower the cost, the active solutions will get even better. Meanwhile, a possibility to improve the ATRU by development of the magnetic material technology is unlikely in the near future [17]. But one possible way to improve it is to increase the weight and implement an 18-pulse ATRU instead, which would especially improve current THD [18]. One can also see that a B6U is the overall best solution of the three when disregarding the power quality. However, power quality is of big importance. As mentioned in chapter 2, a low THD and a power factor close to unity results in higher efficiency. The power quality is especially essential when considering the MEA concept which will lead to an increase in electric power systems which in turn implies that a good power quality is important for the overall system performance. Further on, com- plexity of a topology generally determines the cost and failure rate. That is

Chapter 3. Rectifier Topologies in Aircraft 23

why the passive solutions are a better choice cost-wise and reliability-wise since active solutions is in need of semiconductor devices such as transistors, additional control and gate drive circuitry. That active solutions at the mo- ment are more expensive than passive ones is also verified in [17]. Finally, in [13] a comparison between a six-switch Vienna rectifier and a 12-pulse ATRU is carried out and the result shows a higher efficiency for the ATRU but with a higher weight for a supply voltage of 115 V and an output power of 5000 W.

One of the reasons for the ATRU being the commonly used rectifier topology today is (in addition to the high reliability and efficiency) that it has, as de- scribed earlier, a capability to provide current in excess of its nominal value.

This is used in applications where the peak-to-continuous demand is high which is the case of actuators on aircraft. The active solution is in this case limited by the durability of the switching devices and the power rating of the rectifier has to increase for application where overloading capability is needed [17].

TABLE 3.1: Comparison of ∆-switch rectifier, six-switch Vi- enna rectifier and 12-pulse ATRU. Values are obtained from [2][15][18], and values that were not available in the reports are

marked with "-".

∆-switch Vienna 12-pulse ATRU

Supply L-N RMS voltage [V] 115 115 230

Supply frequency [Hz] 400 400 400

Switching frequency [kHz] 72 72 passive

Output voltage [V] 400 400 270

Output power [kW] 5.0 5.0 4.5

Supply current THD 2-3% - 11%

Power factor 0.999 - 0.986

Efficiency 94.9% 94.4% 95.0%

Input EMI-filter Yes Yes No

Power density [kW/dm³] 1.91 1.91 0.995 Power weight ratio [kW/kg] 1.32 1.32 1.82

TABLE 3.2: Comparison of B6U, 12-pulse ATRU and ∆-switch rectifier [12]. The topologies are evaluated from very good (++)

to very bad (- -).

B6U 12-pulse ATRU ∆-switch

Power quality - - + ++

Weight ++ - - -

Reliability ++ + - -

Cost ++ o -

Efficiency ++ + -

Chapter 3. Rectifier Topologies in Aircraft 24

In summary, the active topologies ∆-switch and Vienna offer today better power quality but a slightly lower efficiency compared to the ATRU. The cost, the reliability and the ability to vary the output power with ease is in favor of the ATRU, but in the future the active topologies might replace the ATRU. Considering these comparisons, the models to simulate is chosen and the design of them is presented in the next section.

25

Chapter 4

Design of Simulation Models

Due to time limitations not all the topologies presented in chapter 2 were modeled and simulated. The 12-pulse ATRU and the ∆-switch rectifier were chosen to study in more detail. The reason for this choice was that the ATRU was the topology that Saab uses at the moment and that the ∆-switch recti- fier has a slightly higher efficiency for an input voltage of 115 V according to [15]. In all simulations EMC, net inductance and resistance, transient be- havior, saturation of the AT and start-up circuitry to minimize inrush current are neglected due to time limitation and secrecy. The modeling and all simu- lations are carried out in MATLAB/Simulink using the Simscape Power System - Specialized technology toolbox.

4.1 Power supply

The power supply was in all simulations modeled as a balanced three-phase Yg-connected sinusoidal voltage source using the Three-Phase Source block.

The phase voltages (Va(t), Vb(t), Vc(t)) generated from the block can mathe- matically be expressed as in equation 4.1.

V_a(t) = ˆV_ssin(ωt + δ_a) V_b(t) = ˆV_ssin(ωt + δ_b) V_c(t) = ˆV_ssin(ωt + δ_c)

(4.1)

Vˆ_s =√ 2V_s

Vˆ_s =peak L-N supply voltage Vs =RMS L-N supply voltage ω =angular frequency = 2πf f =supply frequency

δ_a =phase of voltage in phase a δ_b =phase of voltage in phase b δ_c =phase of voltage in phase c

For each simulation the supply voltage (amplitude and phase) and the fre- quency were set according to the requirements specified in chapter 2, sec- tion 2.2 and 2.3 respectively.