Design, analysis, and comparison of automatic flux

This document is downloaded from DR‑NTU (https://dr.ntu.edu.sg)Nanyang Technological University, Singapore.

Design, analysis, and comparison of automaticflux regulator with automatic voltageregulator‑based generation system for dc marinevessels

Satpathi, Kuntal; Ukil, Abhisek; Pou, Josep; Zagrodnik, Michael Adam

2018

Satpathi, K., Ukil, A., Pou, J., & Zagrodnik, M. A. (2018). Design, analysis, and comparison ofautomatic flux regulator with automatic voltage regulator‑based generation system for dcmarine vessels. IEEE Transactions on Transportation Electrification, 4(3), 694‑706.doi:10.1109/TTE.2018.2826439

https://hdl.handle.net/10356/96796

https://doi.org/10.1109/TTE.2018.2826439

© 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must beobtained for all other uses, in any current or future media, includingreprinting/republishing this material for advertising or promotional purposes, creating newcollective works, for resale or redistribution to servers or lists, or reuse of any copyrightedcomponent of this work in other works. The published version is available at:https://doi.org/10.1109/TTE.2018.2826439

Downloaded on 09 Nov 2021 19:50:41 SGT

1

Design, Analysis and Comparison of AutomaticFlux Regulator with Automatic Voltage RegulatorBased Generation System for DC Marine Vessels

Kuntal Satpathi, Student Member, IEEE, Abhisek Ukil, Senior Member, IEEE,Josep Pou, Fellow, IEEE and Michael Adam Zagrodnik

Abstract—The generation system in the dc marine vessel isexpected to be comprised of wound rotor synchronous generator(WRSG) which is interfaced with the active front-end rectifier.For such system, WRSG flux and dc voltage at the activefront-end rectifier output are controlled independently. WRSGflux is maintained by the field excitation circuit of WRSG forwhich automatic voltage regulator (AVR) has been traditionallyemployed. AVR operates by regulating the WRSG terminalvoltage, thus demands additional tuned filter for the applicationin dc marine vessel. This filter increases the size of the generationsystem along with additional cooling requirements at higher load.With this regard, this paper introduces automatic flux regulator(AFR) based field excitation system where the WRSG flux ismaintained directly and is independent of the measurement ofthe terminal voltage. Small-signal analysis is done to design thecontrol loop of the AFR based field excitation circuit. To verify itsefficacy, comparison with the AVR based dc generation system isconducted. The comparison has been done on the basis on controlresponse and marine operation.

Index Terms—DC Bus Voltage Control, DC Shipboard Opera-tion, Excitation System, Flux Estimation, Synchronous Generator.

NOMENCLATURE

C DC-link capacitorCf Filter Capacitance of the LC filterio Output dc current of AFE rectifierθp Reference angle generated by PLL (Phase locked loop)ipds, i

pqs d-q current of WRSG in PLL reference frame

vpds, vpqs d-q voltage of WRSG in PLL reference frame

Manuscript received November 20, 2017; revised March 05, 2018; acceptedMarch 30, 2018. The work was conducted within Rolls-Royce@NTU Corplab under Corp Lab@ University Scheme. (Corresponding authors: KuntalSatpathi; Abhisek Ukil.)K. Satpathi and J. Pou are with the School of Electrical and ElectronicEngineering, Nanyang Technological University, Singapore 639798 (e-mail:[email protected]; [email protected]).A. Ukil is with the Department of Electrical and Computer Engineer-ing, The University of Auckland, Auckland 1010, New Zealand (e-mail:[email protected]).M. A. Zagrodnik is with the Advanced Technology Centre, Rolls-RoyceSingapore Pte. Ltd., 6 Seletar Aerospace Rise, Singapore 797575 (e-mail:[email protected]).

irds, irqs d-q current of WRSG in RRF (rotor reference frame)

vrds, vrqs d-q voltage of WRSG in RRF

θs Reference angle for stator flux reference frame (SFRF)iMs, iTs M-T current of WRSG in SFRFvMs, vTs M-T voltage of WRSG in SFRFKexc Gain of Type-ST ExciterLs WRSG stator inductanceLf Filter inductance of the LC filterp Number of poles of WRSGRs WRSG stator resistanceR′fd Field winding resistance of WRSG (referred to stator)Rd Damping resistance of LC filtervdc Output dc-link voltageXs/Ls Synchronous Impedance/InductanceXmd/Xmq WRSG d-q axis reactanceXlf WRSG field circuit reactanceλrds, λ

rqs WRSG Flux linkage in RRF

λMs, λTs WRSG Flux linkage in M-T axisψMs, ψTs Flux linkage per second of WRSG in M-T axisτ ′do WRSG open circuit time constantτexc Time constant of Type-ST Exciterωb, ωr Base speed and rotor speed of WRSG

I. INTRODUCTION

In recent years, the paradigm shift in the energy conversiontechnologies of the emerging marine vessels have witnesseda transition from the ac to dc power systems. This transitionto the dc marine vessel is primarily strengthened by the fuelefficient operation of the interfaced diesel generator (DG)running at the optimized speed [1]–[5]. With the requirementof stringent emission control as per the guidelines issuedby the International Maritime Organization, such dc powerconversion systems in marine vessels are expected to beadopted in the near future [2], [6], [7]. It is further supportedby the availability of efficient ac/dc [8] and dc/dc [9] powerelectronic conversion devices. Thus, the modeling, control andthe operation of the dc generation system becomes an important

NTULSC

Rectangle

NTULSC

Text Box

2

aspect for such emerging marine vessels with some of thefunctional aspects being inherited from the conventional acgeneration systems.

The ac generation system in the conventional ac marinevessel comprises of the fixed frequency DG to comply with thestrict frequency and phase synchronization requirements [10],[11]. On the other hand, DG of the dc generation system indc marine vessel is expected to be interfaced with the activefront-end (AFE) rectifier, enabling it to run asynchronouslyand operate in variable speed according to the load demand.This particular operation leads to the lower specific fuel oilconsumption (SFOC) resulting in increased fuel efficiency [2],[7]. It is to be noted that the common control requirementsfor ac and dc generation system is to maintain the terminalbus voltage while operating the interfaced ac generator (of theDG) within the prescribed flux limits. For the marine vesselsrequiring higher power, wound rotor synchronous generator(WRSG) might be preferred choice for the interfaced acgenerator on account of lower cost, robust design and superiorfault current limiting capability which is partially achievedthrough field de-excitation system [12], [13]. As a result, theexternal field excitation system of the interfaced WRSG is ofsubstantial importance while designing control framework ofthe generation system.

A. AC Generation System

In the ac marine vessels, WRSG is generally interfaced withthe ac bus [10]. Such cases emulate the control requirements forthe isolated ac grid conditions where WRSG terminal voltagei.e. the ac bus voltage is maintained by the voltage regulatorbased field excitation system and is conventionally knownas ‘automatic voltage regulator (AVR)’. The availability andmeasurement of the sinusoidal voltage at the WRSG terminalshave supported the widespread utilization of AVR in the acpower systems, and in traditional ac marine vessels [14], [15].The functional block diagram of AVR as depicted in Fig. 1comprises of three basic blocks (a, b & c) for voltage regulationand three advanced blocks (d, e & f) for load compensation andstability of the multi-machine system [15]. AVR compares theterminal voltage of the WRSG against a defined set-point andsubsequently varies the field current/field voltage to maintainthe desired terminal voltage for different loading conditions. Inother words, AVR indirectly controls the flux of the WRSG bydirectly maintaining the terminal voltage to a speed dependentset-point.

B. DC Generation System: Problem Statement

The control objective of the dc generation system is toregulate the dc bus voltage at the AFE rectifier output whileindependently operating the WRSG within the prescribed fluxlimits. Since the WRSG is interfaced directly with the AFEconverter, its terminal voltage is influenced by the AFE con-verter switching voltages and dc-link voltage. Although the

d.Terminal Voltage Transducer and Load Compensator

b. Exciterc. Generator

and Power Systema. VoltageRegulator

e. Excitation SystemStabilizer

f. Power SystemStabilizer

+

V ref

V cV err V T

I T

I FD-

EFDV R

V F

V SV S1

Fig. 1. Functional block diagram of AVR based field excitation control systemof WRSG.

(a) (b)

Fig. 2. Generator terminal voltage when (a) interfaced directly with theconverter and (b) interfaced with the filter inductor.

internal voltage of the generator (eT ) is sinusoidal, it is theterminal voltage (vT ) after the synchronous impedance (Xs)which is of switching voltage type as illustrated in Fig 2(a).Hence, the measurement of the voltage directly at the WRSGterminal would not reflect its performance thus limiting theapplicability of AVR for WRSG terminal voltage and fluxcontrol. However, AVR can be suitably integrated by filteringout the switching voltage across the WRSG terminal thusenabling sinusoidal measurement of vT . To regulate the dc-link voltage, vector control of the AFE rectifier would bedone considering the WRSG (with AVR control) as weakgrid network [16]. Hence the type of interfaced filter plays asignificant role in devising control framework for AVR baseddc generation system. For instance, vT while using the L filteris shown in Figure 2(b). The inclusion of the high inductance Lwill limit the maximum dc voltage at the AFE converter output,slowing down the dynamic response of the system [17]. Thus,LCL filters might be an alternative choice to the L filter [17]–[19]. LC filter can be installed at the generator terminal whereXs along with external LC would thus act as LCL filter.

Such interfaced filters for the applicability in AVR baseddc generation system increases the space and weight alongwith increased cooling requirements at higher loads. Moreover,during the variable speed operation, the voltage reference set-point to the AVR must be changed continuously to regulateterminal voltage/frequency (vT /f ) ratio hence maintaining the

NTULSC

Text Box

3

flux of the WRSG.

C. Automatic Flux Regulation based DC Generation System

To avoid interfacing of the filter for the integration ofAVR, this paper proposes flux regulation based field excitationsystem of WRSG and is termed as Automatic Flux Regulator(AFR). Instead of regulating the terminal voltage, AFR directlyregulates the machine flux of the WRSG thus indirectly regu-lating the terminal voltage at the operating speed. Unlike AVR,AFR is independent of the measurement of the WRSG terminalvoltage and is an integral part of the flux estimation controlroutine of the vector control of WRSG [20]. Further, it is in-dependent of the dc bus voltage control at AFE rectifier outputwhich is maintained by controlling torque producing currentof the WRSG. This control might be perceived to be derivedfrom the field oriented control of the synchronous machinedrive which has received limited attention [20]–[25] due tothe dominance of permanent magnet synchronous generator(PMSG) drives [26]–[29] and induction generator drives [30]–[33]. The vector control should be done at the reference framerotating at synchronous speed to make the control parametersdc quantity [34]. However, it has been reported that the controlin the rotor reference frame (RRF) would lead to installationproblems such as the requirement of higher voltage rating ofWRSG and interfaced AFE rectifier at full load [35]. Thishas been explained in Subsection III-A. Hence, vector controloperation at the stator flux reference frame (SFRF) is selectedin this paper which proves to be advantageous as the torqueand flux producing current components are decoupled fromeach other. Further, there is no need of detailed flux equationsin M-T axis of SFRF as the aim is to maintain the total fluxof the WRSG using the AFR.

D. Research Gaps and Contributions

As detailed before, AVR based field excitation system di-rectly regulates the WRSG terminal voltage while indirectlycontrolling the WRSG flux. On the contrary, AFR basedfield excitation system directly regulates the WRSG flux thusindirectly controlling the WRSG terminal voltage. In spite ofthe advances in the control paradigm of the AFE rectifier;there has been no study to best of the authors’ knowledge onthe design, control and subsequent comparison of these twofield excitation control methods intended for the dc generationsystem for the application in the emerging marine vessels. Thusthe following contributions have been reported in this paper:

1) Vector control of AFE rectifier has been devised consid-ering AVR controlled WRSG as weak grid. The AVRoperation is supported by filter designed at the desiredcut-off frequency. Type-ST based AVR [15] has beenchosen for its lower footprint and is expected to besuitable for the marine applications where space andweight reduction is prime concern [6]. This has beenincluded in Section II.

2) Vector control of WRSG drive has been devised with theAFR based field excitation system. The control systemof AFR based field excitation system is developed bydetailed small signal analysis of the variation of theWRSG flux. The derivation of AFR is done analogouslywith the AVR based field excitation system for easier andfair comparison. This has been reported in Section III.

3) Control loop design of AVR and AFR based field excita-tion system of WRSG for dc generation has been madecompliant with IEEE Std. 421.2 [15]. Subsequently,the comparative analysis based on time and frequencydomain analysis has been performed and described inSection IV.

4) Operation of AVR and AFR based dc generation sys-tem for various marine operating conditions have beencompared and reported in Section V. The operation suchas load change, step change, fault conditions have beenvalidated with hardware-in-loop (HIL) environment.

Discussions on the findings are reported in Section VI andthe paper is concluded in Section VII.

II. AVR BASED DC GENERATION SYSTEM

A. Selection of AVR

In comparison with the Fig. 1, the basic blocks (a, b and c)have been considered for the AVR control. The control loopschematic of the Type-ST based AVR for regulation of vT isshown in Fig. 3. The parameters of the WRSG is indicated inAppendix A. Type-I generator transfer function [13], [36] isconsidered as shown in Fig. 3 where τ ′do is calculated from (1)which comes to be 3.75 s [13].

τ ′do =1

R′f

Xmd +Xlf

ωb. (1)

First order transfer function has been considered for theexciter with τexc = 0.01 s and Kexc = 1. The AVR isdesigned for the bandwidth of 5 Hz as per IEEE Std 421.2 [15].GAV R(s) is the controller of the AVR which is used to regulatethe terminal voltage vT of WRSG to the desired setpoint vSPT .The detailed analysis to determine the GAV R(s) is done inSection IV.

Fig. 3. Control system diagram of the Type-ST based AVR for WRSG terminalvoltage control.

B. Filter Design for AVR based Field Excitation System

Passive damping with damping resistor Rd has been consid-ered in this paper to mitigate the potential resonant problems.The general schematic of vT control using AVR with the LC

NTULSC

Text Box

4

(a)

Lf , Rf

Cf

Ls

Vconv(s)E

T(s)

Machine Side Converter Side

LC Filter

Rd

VT(s)

(b)

Fig. 4. (a) Schematic of LC filter interfaced with machine and (b) singlephase representation of LC filter.

filter is shown in Fig. 4(a). The single phase diagram of the LCfilter is shown in Fig. 4(b), where ET (s) is the internal voltageof the WRSG, VT (s) is the terminal voltage and Vconv(s) isthe converter input voltage which is the source of the switchingvoltages. The transfer function Fr(s) to determine VT (s) isillustrated in (2) where filter elements are consistent with theannotations of Fig. 4 and the filter inductor winding resistance(Rf ) is neglected.

Fr(s) =VT (s)

Vconv(s)=

1/LfCf + sRd/Lfs2 + sRd/Lf + 1/LfCf

. (2)

The filter parameters are computed by considering (3) [17]–[19] and is shown in Table I:

Lf =Vdc D (1−D)

fsw δIp, Cf =

1

4π2f2cLf

, Rd = 2ζ

√LfCf

,

(3)where, D is the duty cycle, Lf is chosen as per the allowableripple content of the generator output current (δIp), Cf ischosen based on the desired cut-off frequency (fc) and Rdvalue is chosen for the given damping ratio (ζ). The cut-offfrequency (fc) should be in the range depicted by the equationstated in (4) [18]:

10 fL < fc <fsw2, (4)

where, fL is the line frequency (60 Hz) and fsw is theswitching frequency of the converter (5000 Hz). The bode plotof Fr(s) with various damping ratio is shown in Fig. 5(a).Waveform of vT corresponding to parameters of Table I isshown in Fig. 5(b).

C. System Description and Controller Design

The schematic of the AVR based dc generation system isshown in Fig. 6. The sinusoidal generator terminal voltage

TABLE IPARAMETERS OF THE LC FILTER

f c (Hz) Lf (H) Cf (F) Rd (Ω) ζ D

1000 3.1 x 10−4 1.58 x 10−4 1.9526 0.7 0.5

(a)

0.5 0.505 0.51 0.515 0.52 0.525 0.53 0.535 0.54

−1000

−500

0

500

1000

Time (s)

Vol

tage

(V

)

(b)

Fig. 5. (a) Bode plot of LC filter for various damping ratios and (b) terminalvoltage of WRSG after interfacing LC filter (vTa: Red, vTb: Blue, vTc:Green).

(vTa, vTb, vTc) at the WRSG terminal is obtained by em-ploying LC filter and is maintained at desired set-point by theAVR. The two-level voltage source converter (2L-VSC) is theAFE converter which perceives this source as weak grid andcontrols the dc bus voltage and line current by vector controloperation [16]. The control is carried out at the reference framedictated by the PLL. The reference angle for control (θp) iscomputed by the PLL installed at the generator terminals.At this PLL reference frame the vpds and vpqs are decoupledfrom each other. The currents at PLL reference frame ipdscontrols the active power while the ipqs controls the reactivepower output provided the ipds is aligned with the vpds. Theipqs is set to zero to enable unity power factor operation. Thephasor representation of such operation is shown in Fig. 7.By this method, operation of the AVR to maintain the WRSGflux/voltage and AFE rectifier operation to maintain dc-linkvoltage becomes independent of each other. The switchingfrequency of the converter is set at 5000 Hz, the bandwidthof current control loop is set at 1000 Hz, voltage control loopis set at 100 Hz and the bandwidth of AVR control is set at5 Hz [15]. The voltage and current control loop is describedbelow.

1) Voltage Control Loop: By neglecting the power lossesacross the filter in Fig. 6, it is assumed that the power generatedat Point-A equals to the power output of the converter (Point-B). This can be illustrated by:

3

2(vpdsi

pds + vpqsi

pqs) = C

dvdcdt

vdc + vdcio, (5)

NTULSC

Text Box

5

Generator

Diesel Engine

Diesel Generator (DG)

DCDC

Fuel Input (uF )

PI Control

Fuel Injection System Coupling Shaft

-

+

ω Ls

K pc+K ics

ω Ls

Current Control Loop

Voltage Control Loop

PI- Control Voltage Control Transfer Function

Current Control Transfer Function

SP

WM

Mo

du

lati

on

2L- VSC

1/ωro

Js+2k loss

k pmτ pm s+1

e−τd

C vdc

-

+ K pc+K ics +

PI-Control

Generator Field Excitation System

- +

+

- +

- +

- + K pv+K iv

s - +

ω Ls

ω Ls

+- -

+- +

PI-Control

Gcap (s)

Gcap (s)

Gvap (s)

iqsp−ref=0

vdcref

ia ,ib ,ic

idsp−ref

iqsp

idsp

v tref

v tmeas

L f R f

Rs

eT LsPmech P loadω

r

ωrref ω

r

P load i fd e fdv ta ,v tb ,v tc

C fRd

Point-B

vdc3-phase

PLL

v tav tbv tc

θp

v tav tbv tc

iaibic

vdsp

vqsp √vds2+vqs2 v t

meas

abc

d q0

abc

d q0

idsp

iqsp

vdsp

vqsp

GAVR(s)

vdsp

vqsp

Point-A

θp

θp

io

Fig. 6. Overview of the functional block diagram of AVR based dc Generation System.

Fig. 7. Phasor diagram for vector control of AFE rectifier interfaced with theWRSG regulated by the AVR at PLL reference frame.

where, vpd , vpq and ipd, ipq are calculated at point-A of Fig. 6.As per the proposed control technique at PLL reference frame,vpq = 0 and ipd = 0 (for unity power factor). Thus, by applyingsmall signal analysis in (5), the plant transfer function for thevoltage control loop (Gpva(s)) is derived as shown in:

Gpva(s) =3

2

V pdssCVdc + Io

. (6)

2) Current Control Loop: The line currents ia, ib, ic flowsfrom point A to point B through the filter inductor withinductance Lf and inductor winding resistance Rf . The voltageat the converter input terminal is maintained to get the requiredcurrent flow. The plant transfer function for current loop(Gpca(s)) is shown:

Gpca(s) =1

Rf + sLf. (7)

III. AFR BASED DC GENERATION SYSTEM

A. Selection of Reference Frame

Selection of the reference frame is important to implementthe AFR based dc generation system. The reference frameshould be synchronously rotating to make the control param-eters dc quantities thus enabling easier controllability. As aresult RRF or SFRF might be the feasible reference frames.At RRF, the torque (T) is given by:

T = 3pirqsλ

rds − irdsλrqs

4. (8)

The flux equation along d- and q- axis of RRF is given by:

λrds = −Llsirds + Lmd(−irds + i′rfd + i

′rkd), (9)

λrqs = −Llsirqs + Lmq(−irqs + i′rkq1 + i

′rkq2). (10)

By neglecting the damper winding currents (i′rkd, i

′rkq1, i

′rkq2) and

setting irds = 0, the d-axis flux can be regulated by the fieldcurrent, ir

′

fd while the torque can be controlled by varying irqs.In such a case, the torque equation becomes:

T = 3pirqsλ

rds

4= 3p

irqsLmdi′rfd

4. (11)

Although the d-axis flux and q- axis current could be controlledindependently, the flux along d and q axes are not decoupledcompletely. This is because, the d-axis flux is controlled by thefield current while the q-axis flux is dependent on the q-axiscurrent, which is illustrated below:

λrds = Lmdi′rfd, (12)

λrqs = Lmqi′rqs. (13)

The control strategy at RRF, irfd will try to keep the λrdsconstant. However, an increase in load will increase irqs whichwill further increase λrqs as per (13). Thus, the total machineflux, λs =

√λrds + λrqs, also increases with increase in load.

This will result in increased operating flux and terminal voltageat higher loads thus requiring higher ratings of the interfacedAFE rectifier [35].

It is thus inferred that selection of reference frame is crucialto implement the AFR based dc generation system. The flux isrequired to be decoupled at all operating scenarios and shouldpreferably be controlled by the field exciter circuit. On theother hand, the torque output of WRSG could be controlled bythe vector control operation of the AFE rectifier. This operationis possible if the control is done at SFRF. The phasor diagramrepresenting the various reference frames of WRSG is shown inFig. 8. The torque equation of WRSG when the vector control

NTULSC

Text Box

6

is carried out at SFRF is shown as [37]:

T = 3piTsλMs − iMsλTs

4. (14)

B. System Description of the Vector Control of WRSG

The proposed AFR based dc generation system is illustratedin Fig. 9 where by theory, λTs = 0 and λMs = |λs| (WRSGflux) when the control is done at SFRF (M −T axis). Further,the WRSG is operated at unity power factor by setting iMs =0. Thus, with the mentioned objectives, the torque equationreduces to:

T = 3piTsλMs

4. (15)

The output power or equivalently the output dc bus voltage ismaintained by varying iTs according to varying load demand.λMs or equivalently λs of WRSG is independently controlledby AFR based field excitation. The whole control frameworkcan be divided into five major control blocks, as shown inFig. 9, and are discussed below:B1. Estimation of the phase voltages (vabc) of WRSG is done

by the dc-link voltage and the switching combination(Sabc) of the AFE rectifier [8].

B2. After vabc is estimated, the angle to convert the controlparameters into SFRF is computed. This angle (θs) iscalculated by tan−1(λβ/

√λ2α + λ2

β) where λα and λβare flux components in the α and β axes which are instator reference frame (Fig. 8).

B3. The flux linkages per second in M-T axis (SFRF) i.e.ψMs and ψTs are estimated utilizing θs computed inStep B2. By theory, ψTs = 0, and the ψMs(= ψs,WRSG flux linkages per second) is fed to the AFR basedexcitation control.

B4. The AFR based excitation control maintains ψMs or(equivalently λs) to specified set-point defined by theuser. This is maintained by varying the field current andfield voltage in accordance to changing load angle (δ).

B5. The power control loop consists of voltage and currentcontrol loop. The design of these loops and the selected

Fig. 8. Phasor diagram showing the relative orientation of d− q and M − Taxes for WRSG.

bandwidth are similar to the design presented in Sec-tion II-C respectively with the only difference is thatthe present control is done in M-T axis. This loop isindependent of the flux control by AFR in Step B4. Thevoltage loop (Gvf ) is computed from the small signalanalysis of the power balance equation in (16) and isillustrated in (17). The current loop transfer function isalso shown in (17).

3piTsλMs

4ωr = C

dvdcdt

vdc + vdcio. (16)

Gvf (s) =3

4

pλMsωrsCVdc + Io

; Gcf (s) =1

Rs + sLs. (17)

C. Control Loop Derivation of AFR Based WRSG Field Exci-tation System

One of the objective of AFR is to control the stator flux (λs)by field excitation circuit which can equivalently be achievedby regulating the stator flux linkage per second, ψs( = ψMs)to the desired value. The voltage equation of WRSG equationsin terms of flux linkages/second in M-T frame is illustrated in(18a) and (18b) [37].

vMs = −rsiMs −ωrωbψTs +

1

ωb

d

dtψMs, (18a)

vTs = −rsiTs +ωrωbψMs +

1

ωb

d

dtψTs, (18b)

and the WRSG terminal voltage vt is given by (18c),

v2t = v2

Ms + v2Ts, (18c)

`

Generator

2L-VSC

PI Controller

ΨTs

= 0

MT0 current controller

MT0

abcSPWM

Switching pulses for VSC

(2 −1 −1

−1 2 −1−1 −1 2 )

13

λα=∫(vα−iα r s)dt

λ β=∫(vβ−iβ r s)dt

λ s=√(λα2+λ β

2)

ΨMs

MT0abc

MT0abc

αβ0abc

αβ0abc

λs

θs

θs

Ψs

ref

B4. Automatic Flux Regulator (AFR)

DCDC

5(a) Voltage Control Loop

5(b) Current Control Loop

B3. Computing Flux in SFRF Frame

B2. Computing Angle for SFRF conversion

B1. Estimation of Phase Voltagesψ Ms=∫ω b(vMs+r s iMs+

ω rω b

ψ Ts)dt

ψ Ts=∫ω b(vTs+r s iTs−ω rω b

ψ Ms) dt

Converter Switching

Combination

B5. Output Power Control

PROPOSED CONTROL FRAMEWORK

+-K AFR( s)

iαβ0

vαβ0

vdcref

vdc

iTsref

i0ref

=0

iMsref

=0 vMsref

vTsref

v0ref

θs

vdc

iMT0

vMT0

+ -

vdc

vabc S abc

ωr

i fd e fd

iabc

X

C

io

Fig. 9. Overview of the vector control of WRSG drive for dc bus voltagecontrol and AFR for WRSG control.

NTULSC

Text Box

7

where, vMs & vTs are phase voltages, iMs & iTs are linecurrents, ψMs & ψTs are flux linkages per second in M-Taxis. rs is WRSG stator resistance, ωb and ωr are base speedand generator speed respectively. As ψMs = constant andψTs = 0 in M-T axis; by neglecting rs, the equation ofterminal voltage becomes:

v2t = v2

Ms + v2Ts, (19a)

v2t =

( 1

ωb

d

dtψMs

)2

+(ωrωbψMs

)2

(19b)

where, ψMs = ψs. Thus by applying small signal analysis on(19b), the relation between ψMs and vt can be found out.

(Vt + vt)2 =

( 1

ωb

d

dt(ΨMs + ψMs)

)2

+(ωrωb

(ΨMs + ψMs))2

,

(20a)

d

dt(ΨMs + ψMs

)= ωb(Vt + vt)

√1−

(ωr(ΨMs + ψMs)

ωb(Vt + vt)

)2

.

(20b)

The RHS of the (20b) is of the form Ak(1− x)1/2 which canbe expanded binomially and linearised by neglecting higherorder terms to get (20c),

d

dt(ΨMs + ψMs

)≈ ωb(Vt + vt)

(1− 1

2

(ωr(ΨMs + ψMs)

ωb(Vt + vt)

)2),

(20c)

d

dtψMs ≈ 2ωbvt −

ω2r

ωb

ΨMs

VtψMs. (21a)

Applying Laplace transform

sψMs(s) ≈ 2ωbvt(s)−ω2r

ωb

ΨMs

VtψMs(s). (21b)

Re-arranging the terms we will get the relation betweenψMs(s) (= ψs(s)) and Vt(s) as shown in (21c).

ψs(s)

vt(s)≈ ωb

12s+ 1

2ω2

r

ωb

Ψs

Vt

. (21c)

This relation between ψs(s) and vt(s) will be used to deter-mine the plant transfer function and to design the controllerof the AFR based field excitation system. The AFR basedcontroller is implemented as shown in B4 block of Fig. 9.The schematic for the AFR based WRSG control is depictedin Fig. 10 where the Ψs is maintained at the desired setpointΨSPs dependent on the GAFR(s). GAFR(s) is determined by

the plant transfer function of generator, exciter and D(s) asanalyzed in Section IV.

IV. CONTROLLER RESPONSE ANALYSIS

In the dc generation system, the aim of both AVR andAFR based field excitation system is to regulate the operatingflux of the WRSG within the prescribed operating limits.AVR does this by maintaining the terminal voltage of WRSG

Fig. 10. Functional block diagram of the field excitation circuit in the proposedAFR based control.

against speed dependent setpoint whereas the AFR does thisby regulating the WRSG flux to its rated value which isdescribed in Section II and Section III respectively. In thissection, comparative analysis of the AFR with AVR basedfield excitation is carried out to support its effectiveness andhighlight its potential benefits.

A. Controller Design and Frequency Response Analysis

One of the prime difference between the control structureof the AVR and AFR based excitation system is the need ofadditional transfer function D(s) (in (22)) in the AFR basedsystem as marked in Fig. 10.

D(s) =ωb

12s+ 1

2ω2

r

ωb

Ψs

Vt

(22)

GAV R(s) and GAFR(s) shown in Fig. 3 and 10 are thecontrollers of the AVR and AFR which are used to maintainthe terminal voltage and machine flux respectively at definedsetpoints and within the required bandwidth. As per, IEEEStd 421.2, bandwidth of 5 Hz is chosen for the field regulat-ing applications. Conventionally, GAV R(s) and GAFR(s) aremodeled by proportional controllers (P-controllers) to maintainthe desired bandwidth, which will have persistent steady-stateerrors. Thus, in addition to the P-controller, this paper alsoconsiders K-factor based controller which is based on thepole-zero placement technique [38]. In this way, steady-stateand transient responses could be studied when the type ofcontrollers are altered. The general structure of the P-controlleris:

GAV RP(s) =

KP

1 + sτP, (23)

where, KP is the gain and τP is the time-constant of theamplifier. Generally, τP is very small constant value while KP

is adjusted to get the desired bandwidth. On the other hand,the general structure for the K-factor based controller is:

GAV RK(s) =

KK

s

(1 + s/ωz1 + s/ωp

). (24)

In such controllers the gain KK , pole ωp and zero ωz couldbe adjusted to obtain the desired bandwidth and phase margin.For a fair comparison, bandwidth of both AVR and AFR havebeen fixed at 5 Hz. The controller transfer functions for AVRand AFR are shown in Table II. Bode plot of both systems arecompared in Fig. 11. It can be observed that the responses ofboth AVR and AFR are similar in the low frequency region.

NTULSC

Text Box

8

TABLE IICONTROLLER TRANSFER FUNCTIONS FOR AVR AND AFR SYSTEM

Control AVR(GAV R(s)

)AFR

(GAFR(s)

)Proportional Control

124

1 + 0.001s

63

1 + 0.001s

K-factor Control77.18

s

(1 + s/31.42

1 + s/161.97

)28.38

s

(1 + s/0.45

1 + s/217.96

)

Due to the presence of D(s), the gain and phase margin ofAFR are lesser than AVR but within permissible limits. Thefrequency-domain metrics for such normal operating conditionsare presented in Table III.

A F R B a s e d E x c i t a t i o n S y s t e m

1 0 - 1 1 0 0 1 0 1 1 0 2 1 0 3- 3 6 0- 2 7 0- 1 8 0

- 9 00

A V R B a s e d E x c i t a t i o n S y s t e mPhas

e (de

g)

F r e q u e n c y ( H z )

1 0 - 1 1 0 0 1 0 1 1 0 2 1 0 3- 1 5 0- 1 0 0

- 5 00

5 0

A V R B a s e d E x c i t a t i o n S y s t e m

Magn

itude

(dB)


(a)

1 0 - 1 1 0 0 1 0 1 1 0 2 1 0 3- 3 6 0- 2 7 0- 1 8 0

- 9 00


F r e q u e n c y ( H z )

1 0 - 1 1 0 0 1 0 1 1 0 2 1 0 3- 1 5 0- 1 0 0

- 5 00

5 0


Magn

itude

(dB)

A F R B a s e d E x c i t a t i o n S y s t e mPhas

e (de

g)


(b)

Fig. 11. Comparison of the bode plot of AVR and AFR based excitationsystem for (a) P-control and (b) K-factor based control.

B. Step Response Analysis

Fig. 12 compares the step response of the closed-loopvoltage control for AVR and closed-loop flux control forAFR based field excitation system. As the bandwidth of bothsystems is kept same, the rise times are reasonably comparable.Although other parameters such as peak amplitude of AFRis different from AVR but are within permissible limits asprescribed by IEEE Std 421.2 [15]. The time-domain metricsfor such normal operation are depicted in Table III.

0 . 0 0 0 . 0 5 0 . 1 0 0 . 1 5 0 . 2 0 0 . 2 5 0 . 3 00 . 0

0 . 2

0 . 4

0 . 6

0 . 8

1 . 0

1 . 2

Volta

ge (p

u)

T i m e ( s )

P - C o n t r o l K - F a c t o r

(a)

0 . 0 0 0 . 0 5 0 . 1 0 0 . 1 5 0 . 2 0 0 . 2 5 0 . 3 0

0 . 0

0 . 2

0 . 4

0 . 6

0 . 8

1 . 0

1 . 2

Flux (

pu)

T i m e ( s )

P - C o n t r o l K - F a c t o r

(b)

Fig. 12. Comparison of the step response of (a) AVR and (b) AFR based fieldexcitation system using both P-control and K-factor based control.

C. Response to Change in Parameters and Operating Limits

The controllers GAV R(s) and GAFR(s) are designed con-sidering the plant transfer functions of exciter and WRSG.

Inaccurate determination of the plant transfer functions mightadversely effect the field regulating operation. To test the ro-bustness and operating capabilities of the AVR and AFR basedfield excitation system, the system parameters such as Kexc andτexc of the exciter and τ ′do of WRSG are varied [39]. Time-domain, frequency-domain and stability performance metricsare monitored which are shown in Table III. It is observed thatthe AFR based system offers similar characteristics than theAVR based system.

V. OPERATIONAL ANALYSIS

With reference to the modeling and control frameworkdeveloped in Section II, III and IV; this section presents theoperation of both AVR and AFR based dc generation system.This operation has been validated with the Opal-RT OP5600based HIL real-time simulation platform [2], [40], [41]. Themarine loads and contingencies have been tested for both typesof generation systems for detailed one-to-one comparison.

A. Output with Step Change in Set-point of Control Parameter

Step change in the terminal voltage (for AVR) and fluxset-points (for AFR) are changed from 1.0 pu to 1.1 pu at4 s and to 0.9 pu at 8 s to study its impact on the fieldcurrent requirements and output dc bus voltage. The resultsare shown in Fig. 13 and 14 respectively. Both P-controland K-factor based controller are implemented and compared.As expected, it is seen that there is a persistent steady-stateerror for proportional control while the control variable closelyfollows the set-point for the K-factor based method. It canfurther be seen that the dc bus voltage is more sensitive tothe step change in AVR control. The overshoot in the dc busvoltage for AVR is 1.17% while with AFR is 0.73%. In bothcases the overshoot is within permissible limits according tothe IEEE Std 421.2 [15].

B. Decoupling of the Control Parameters

For the AFR based control, ψMs (or equivalently λMs)should be decoupled from ψTs (or λTs). Similarly, decouplingof vpds from vpqs is required for the AVR based control. For unitypower factor iM (in AFR) and ipq (in AVR) is maintained atzero. The control parameters must be de-coupled at all loadcondition and also during the load change transients. This isshown by the load change operation as shown in Fig. 15. Thegenerator output is varied from 1300 kW to 1850 kW at 5 sand again restored to 1300 kW at 8 s. Fig. 15(a) and 15(b)shows the decoupled control parameters during this operation.The K-factor based control is used for both AVR and AFRbased dc generation system.

C. Verifying the Stability Limits

The stability limits of both the AVR and AFR operating withP-control are primarily dominated by the selection of gains

NTULSC

Text Box

9

TABLE IIIPERFORMANCE METRICS OF AVR AND AFR BASED FIELD EXCITATION SYSTEM TO VARYING SYSTEM PARAMETERS

Conditions Parameters AVR Based System AFR Based SystemP-Control K-Factor P-Control K-Factor

Normal OperatingConditions

Time DomainParameters

Rise Time (s) 0.0455 0.0385 0.0403 0.0372Settling Time (s) 0.0718 0.1271 0.1054 0.1233Overshoot (%) 0.7973 7.5146 2.99 8.45Undershoot (%) 0 0 0 0

Frequency-DomainParameters

Gain Margin (dB) 33.3637 7.7131 9.91 5.58Phase Margin (deg) 71.1844 60.7919 66.17 59.26Bandwidth (Hz) 5.01 5.03 5.06 5.00

Stability Limits Gain (K) 4137 595 624 158

50% Increase inboth Exciter TimeConstant and Gain


Rise Time (s) 0.0443 0.0395 0.0281 0.0284Settling Time (s) 0.1284 0.1403 0.1492 0.1524Overshoot (%) 5.174 13.936 18.99 18.48Undershoot (%) 0 0.77 3.63 3.61




50% Reduction ofboth Exciter TimeConstant and Gainn


Rise Time (s) 0.0527 0.0407 0.1093 0.1152Settling Time (s) 0.0962 0.0652 0.2018 0.2018Overshoot (%) 0 1.6129 0 0Undershoot (%) 0 0 0 0


Gain Margin (dB) 36.35 10.6746 22.91 23Phase Margin (deg) 79.34 68.71 82.65 82.35Bandwidth (Hz) 5.19 5.2 2.65 2.64


50% Increasein τ ′do


Rise Time (s) 0.0754 0.0611 0.0676 0.0691Settling Time (s) 0.1339 0.1892 0.1178 0.1232Overshoot (%) 0 2.2518 0 0Undershoot (%) 0 0 0 0




50% Reductionin τ ′do


Rise Time (s) 0.0221 0.021 0.0203 0.0205Settling Time (s) 0.0734 0.1552 0.109 0.1128Overshoot (%) 11.05 27.74 19.83 18.92Undershoot (%) 1.2 8.1 4.72 4.7




3 4 5 6 7 8 90 . 8 00 . 8 50 . 9 00 . 9 51 . 0 01 . 0 51 . 1 01 . 1 51 . 2 0

Volta

ge (p

u)

T i m e ( s )

W R S G T e r m i n a l V o l t a g e S e t p o i n t W R S G T e r m i n a l V o l t a g e ( K - F a c t o r ) W R S G T e r m i n a l V o l t a g e ( P - C o n t r o l )

(a)

3 4 5 6 7 8 91 . 6 5

1 . 7 0

1 . 7 5

1 . 8 0

1 . 8 5

1 . 9 0

Field

Curre

nt (pu

)

T i m e ( s )

F i e l d C u r r e n t ( K - F a c t o r ) F i e l d C u r r e n t ( P - C o n t r o l )

(b)

3 4 5 6 7 8 91 4 7 01 4 7 51 4 8 01 4 8 51 4 9 01 4 9 51 5 0 01 5 0 51 5 1 01 5 1 51 5 2 01 5 2 5

Volta

ge (V

)

T i m e ( s )

D C B u s V o l t a g e ( K - F a c t o r ) D C B u s V o l t a g e ( P - C o n t r o l )

(c)

Fig. 13. (a) WRSG terminal voltage, (b) WRSG field current and (c) dc-link voltage for P-control and K-factor based GAV R(s) with step change in WRSGterminal voltage setpoint.

NTULSC

Text Box

10

3 4 5 6 7 8 91 . 3 01 . 3 51 . 4 01 . 4 51 . 5 01 . 5 51 . 6 01 . 6 51 . 7 0

F l u x S e t p o i n t W R S G F l u x ( K - F a c t o r ) W R S G F l u x ( P - C o n t r o l )

T i m e ( s )

Flux (

Wb)

(a)

3 4 5 6 7 8 91 . 1 61 . 1 81 . 2 01 . 2 21 . 2 41 . 2 61 . 2 81 . 3 0

Curre

nt (pu

)T i m e ( s )

F i e l d C u r r e n t ( K - F a c t o r ) F i e l d C u r r e n t ( P - C o n t r o l )

(b)

3 4 5 6 7 8 91 4 8 01 4 8 51 4 9 01 4 9 51 5 0 01 5 0 51 5 1 01 5 1 51 5 2 0

Volta

ge (V

)

T i m e ( s )

D C B u s V o l t a g e ( K - F a c t o r ) D C B u s V o l t a g e ( P - C o n t r o l )

(c)

Fig. 14. (a) WRSG flux, (b) WRSG field current and (c) dc-link voltage for P-control and K-factor based GAV R(s) with step change in WRSG flux setpoint.

1 0 0 01 2 5 01 5 0 01 7 5 02 0 0 0

0 . 0 00 . 2 50 . 5 00 . 7 51 . 0 0

3 4 5 6 7 8 90

5 0 01 0 0 01 5 0 02 0 0 02 5 0 0

Powe

r (kW)

P o w e r

Volta

ge (p

u) v p d s v p q s

Curre

nt (A

)

T i m e ( s )

i p d s i p q s

(a)

1 0 0 01 2 5 01 5 0 01 7 5 02 0 0 0

0 . 0 0 0

0 . 7 4 5

1 . 4 9 0

3 4 5 6 7 8 90

5 0 01 0 0 01 5 0 02 0 0 02 5 0 0

Powe

r (kW)

P o w e r

M s T s

Flux (

Wb)

i T s i M s

Curre

nt (A

)

T i m e ( s )

(b)

Fig. 15. (a) Decoupling of vpds and vpqs; ipds and ipqs for AVR based systemand (b) decoupling of λMs and λTs; iMs and iTs for AFR based systemwhen generator output is altered.

Kpa (for AVR) and Kpf (for AFR). Analysis of the root locusdiagram in Figure 16(a) shows that the gain Kpa increasingbeyond 4137 leads to instability in the AVR control. This isevident from the study of the corresponding change in fieldvoltage Efd is shown in Figure 16(c). The Efd fluctuatesfrom 0 pu to 6 pu when Kpa > 4137 making the AVRbased control inoperable. Similarly, root locus diagram inFigure 16(b) infers that the gain Kpf increasing beyond 624would lead to instability in the AFR system. This is supportedby the corresponding change in field voltage Efd is shownin Figure 16(d). It can be observed that for higher gains(Kpf> 624), Efd fluctuates from 0.5 pu to 3.5 pu makingthe AFR based excitation control inoperable.

D. WRSG Loading and Field Current Requirements

Loading of the propulsion system is emulated in Fig. 17(a)where the WRSG is gradually loaded from the no-load tonearly full-load (1850 kW) condition. In this process, it isobserved that the field current requirement for the AFR based

−1400 −1200 −1000 −800 −600 −400 −200 0 200−600

−400

−200

0

200

400

600Root Locus of AVR Based Field Excitation System with Varying Gain "K

pa"

Real Axis

Imag

inar

y A

xis

Kpa

= 4137

Kpa

= 4137

Kpa

= 4137

(a)

−1000 −800 −600 −400 −200 0 200−200

−100

0

100

200Root Locus of AFR Based Field Excitation System with Varying Gain "K

pf"

Real Axis

Imag

inar

y A

xis

Kpf

= 624

Kpf

= 624Kpf

= 624

Kpf

= 624

(b)

2 4 6 8 10 12 14 16 18 20Time (s)

0

1

2

3

4

5

6

7

Gai

n (K

pa)

2 4 6 8 10 12 14 16 18 20Time (s)

0

1

2

3

4

5

6

7

Gai

n (K

pa)

x10E

3

2 4 6 8 10 12 14 16 18 20Time (s)

−1

0

1

2

3

4

5

6

7

Fie

ld V

olta

ge, E

fd (

pu)

2 4 6 8 10 12 14 16 18 20Time (s)

−1

0

1

2

3

4

5

6

7

Fie

ld V

olta

ge, E

fd (

pu)

Printed for Opal−RT 1/1

(c)

2 4 6 8 10 12 14 16 18 20Time (s)

0

100

200

300

400

500

600

700

800

Gai

n (

Kp

f)

2 4 6 8 10 12 14 16 18 20Time (s)

0

100

200

300

400

500

600

700

800

Gai

n (

Kp

f)2 4 6 8 10 12 14 16 18 20

Time (s)

0

1

2

3

4

5

Fie

ld V

olt

age,

Efd

(p

u)

2 4 6 8 10 12 14 16 18 20Time (s)

0

1

2

3

4

5

Fie

ld V

olt

age,

Efd

(p

u)

Printed for Opal−RT 1/1

(d)

Fig. 16. The stability limits of the AVR and AFR based control with thevariation of (a) Kpa and (b) Kpf ; the variation of output field voltage whilechanging (c) Kpa for AVR and (d) Kpf for AFR.

dc generation system is less than that of the AVR baseddc generation. For AVR based system, the active power andreactive powers are regulated by ipds and ipqs, respectively. Onthe other hand, active and reactive powers in AFR based systemare regulated by iTs and iMs, respectively. For similar loadingscenarios the variation of ipds and ipqs (for AVR) and iTs andiMs (for AFR) are shown in Fig. 17(b). In this paper, ipqs (forAVR) and iMs (for AFR) are maintained at zero for unitypower factor operation. Conversion of these currents from theirrespective reference frames (PLL reference frame for AVRand SFRF for AFR) to RRF is done, which is indicated inFig. 17(b). It is seen that the d-axis current at RRF for AVRbased dc generation has higher magnitude than that of the AFRbased dc generation system. It is known that the field currentrequirement depends on the WRSG flux and to counter thed-axis current at RRF [37]. Thus, this requirement is higherfor the AVR based dc generation system which has lower field

NTULSC

Text Box

11

current as compared to the AFR based dc generation system.This becomes one of the major advantages of using the AFRbased dc generation for dc marine vessels.

2 4 6 8 1 00 . 0

0 . 5

1 . 0

1 . 5

2 . 0

2 . 5

3 . 0 G e n e r a t o r L o a d i n g F i e l d C u r r e n t : A V R F i e l d C u r r e n t : A F R

T i m e ( s )

Field

Volta

ge (p

u)

0

5 0 0

1 0 0 0

1 5 0 0

2 0 0 0

Powe

r (kW)

(a)

2 4 6 8 1 00 . 00 . 20 . 40 . 60 . 81 . 01 . 2

Curre

nt (pu

)

T i m e ( s )

A F R i T s i M s d - a x i s c u r r e n t i n R R F q - a x i s c u r r e n t i n R R F

i p d s i p q s d - a x i s c u r r e n t i n R R F q - a x i s c u r r e n t i n R R F

A V R

D i r d s

(b)

Fig. 17. (a) Variation of field current for AVR and AFR based dc generationsystem when WRSG is loaded from no-load to nearly full-load (1850 kW). (b)Current components in SFRF (for AFR based system) and in PLL referenceframe (for AVR based system) along with the decomposed components in theRRF.

E. Operation during Variable Speed

The dc marine vessels are expected to integrate the variablespeed DGs. During such operation, the operating parametersuch as the machine flux of the interfaced WRSG should bemaintained at nominal values. For AVR based system, theterminal voltage set-point of WRSG should be altered con-tinuously with the operating speed so that constant vt/f ratiois maintained and machine flux does not reach the saturationregion. On the other hand, in the AFR based WRSG, machineflux is directly regulated at the desired setpoint without theneed of any set-point variation as compared to the AVR basedsystem. This is a major advantage of the AFR based generationsystem. This operation is shown in Fig. 18 when the speed ischanged from 1 pu to 0.85 pu at 5 s.

For AFR based system in Fig. 18(a), WRSG flux is main-tained at desired setpoint (1.49 Wb) while the field currentincreases to regulate the flux at lower speed. For the AVRbased system the WRSG terminal voltage setpoint is changedaccording to the WRSG operating speed. With the changingspeed, the WRSG terminal voltage setpoint is changed from1 pu to 0.85 pu as shown in Fig. 18(b) and 18(c). When thesetpoint is changed abruptly from 1 pu to 0.85 pu, there isa transient effect in the field current and the machine flux.However, if the voltage setpoint is altered for every operatingspeed, the transient effect in WRSG field current and flux isminimized.

F. Operation during the Fault Conditions

The fault resistance of 0.01 Ω is introduced at the terminalof AFE rectifier for both AFR and AVR based dc generationsystem when it was delivering 2000 kW power to study theresponse of the field excitation circuit and the output faultcurrent. For both cases, the field voltage Efd rises to themaximum upper limit (Fig. 19) which set as per IEEE Std421.2 [4], [15]. The initial fault current due to capacitor

discharge is equal at 12 kA for both cases. The ac fault currentdelivered by the WRSG in case of AVR based system is limitedby the filter inductor, hence it is lower than the AFR basedsystem which can be seen from Fig. 19(b, d).

VI. DISCUSSION

Based on the operation and control of the AVR and AFRbased dc generation system, the following discussions are cited:

1) Conventionally, the generation system in traditional acmarine vessels is based on AVR based field excitationsystem, which is dependent on the measurement of theterminal voltage of WRSG. However, in the 2L-VSCbased dc generation system, the terminal voltage canonly be measured by employing a tuned LC filter, whichincreases the space and weight requirements and furtherincreases the cost. Contrarily, the proposed AFR basedcontrol does not require the measurement of terminalvoltage thus not requiring an additional LC filter. This isthe principal advantage of the AFR based dc generationsystem as compared to the AVR one. This reducesthe complexity of the control, helps in restricting thesize and space requirements and avoids the need ofadditional cooling requirements for the filter during fullload conditions.

2) AVR regulates the terminal voltage while AFR maintainsthe WRSG flux. Thus, the modeling and derivation of therequired ‘terminal voltage to flux’ transfer function andcontrol loop design for the field excitation circuit havebeen derived for the AFR based dc generation system.This was the prime research gap in the existing literature.This design has been done analogously to the AVR basedfield excitation system.

3) DC-link voltage control for the AVR based dc generationsystem is carried out by the vector control operationof the interfaced AFE rectifier at PLL reference frame.This is done by considering the WRSG and interfacedLC filter as weak grid network. It serves as the basemodel and is used for comparison with the AFR baseddc generation system. The dc-link voltage control forAFR system is done by vector control operation of theWRSG fed AFE rectifier at SFRF.

4) Controller response analysis with the emphasis on thevariation of system parameters in both AVR and AFRyields similar results. This makes AFR a suitable alter-native to the AVR based dc generation system. The AFRbased dc generation system is evidently promising whichis supported by the stability analysis and the detailedcomparative analysis with AVR based dc generation inboth time- and frequencydomain.

5) AFR directly regulates the WRSG flux, thus changingthe speed of WRSG would not effect the operatingparameters. However, AVR requires the vSPT to be variedin according to the changing speed. It is seen that theabrupt change of vSPT results in transient variation of the

NTULSC

Text Box

12

4 5 6 7 8 9 1 00 . 8 0

0 . 8 4

0 . 8 8

0 . 9 2

0 . 9 6

1 . 0 0

1 . 0 4

Spee

d (pu

)

S p e e d ( p u ) F i e l d C u r r e n t ( p u ) F l u x ( W b )

T i m e ( s )

1 . 1 0

1 . 1 5

1 . 2 0

1 . 2 5

1 . 3 0

1 . 3 5

1 . 4 0

1 . 4 6 01 . 4 6 51 . 4 7 01 . 4 7 51 . 4 8 01 . 4 8 51 . 4 9 01 . 4 9 51 . 5 0 0

(a)

4 5 6 7 8 91 . 11 . 21 . 31 . 41 . 51 . 61 . 71 . 81 . 9

Flux (

Wb)

T i m e ( s )

V o l t a g e S e t p o i n t C h a n g e d A c c o r d i n g l y V o l t a g e S e t p o i n t C h a n g e d A b r u p t l y V o l t a g e S e t p o i n t N o t C h a n g e d

(b)

4 5 6 7 8 91 . 6

1 . 7

1 . 8

1 . 9

2 . 0

Curre

nt (pu

)

T i m e ( s )

V o l t a g e S e t p o i n t C h a n g e d A c c o r d i n g l y V o l t a g e S e t p o i n t C h a n g e d A b r u p t l y V o l t a g e S e t p o i n t N o t C h a n g e d

(c)

Fig. 18. (a) WRSG flux and field current when the operating speed of DG is reduced from 1 pu to 0.85 pu at 5 s for AFR based dc generation system.Variation of (b) WRSG flux and (c) field current requirement when operating speed is reduced from 1 pu to 0.85 pu at 5 s for AVR based dc generation systemand when terminal voltage setpoint is unchanged, abruptly changed and varied in according to the WRSG speed from 1 pu to 0.85 pu at 5 s.

TABLE IVAVR VS AFR BASED DC GENERATION SYSTEM

Parameter AVR based dc Generation System AFR based dc Generation System

WRSG Excitation Control Parameter Terminal Voltage (vT ) Flux Linkage per second (ψs) / Flux (λs)Reference Frame for WRSG & AFE con-verter control As per PLL Stator Flux Reference Frame

De-coupling Parameters ipds and ipqs; vpds and vpqs iMs and iTs ; ψMs and ψTs

Additional Requirements Need of LC filter Need of converter switching combinationsFootprint More due to LC filter Expected to be lessLosses More for passive damping based LC filter Expected to be lessVariable Speed Operation Manual change of voltage set-point No change of set-pointVoltage Sensors 3 (ac) + 1 (dc) 1 (dc)

Cost More due to additional voltage sensors, LCfilters and cooling requirements

Lesser than AVR due to direct connection of WRSG withVSC resulting in less number of voltage sensors and LCfilters

9.9 10 10.1 10.2 10.3 10.4 10.5Time (s)

−101234567

Fie

ld V

olta

ge, E

fd (

pu)

(a)

9.9 10 10.1 10.2 10.3 10.4 10.5Time (s)

−101234567

Fie

ld V

olta

ge, E

fd (

pu)

9.9 10 10.1 10.2 10.3 10.4 10.5Time (s)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

Cur

rent

(A

)

(b)

9.9 10 10.1 10.2 10.3 10.4 10.5Time (s)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

Cur

rent

(A

) x1

0E4

9.9 10 10.1 10.2 10.3 10.4 10.5Time (s)

−101234567

Fie

ld V

olta

ge, E

fd (

pu)

(c)

9.9 10 10.1 10.2 10.3 10.4 10.5Time (s)

−101234567

Fie

ld V

olta

ge, E

fd (

pu)

9.9 10 10.1 10.2 10.3 10.4 10.5Time (s)

−0.20

0.20.40.60.8

11.21.4

Cur

rent

(A

)

(d)

9.9 10 10.1 10.2 10.3 10.4 10.5Time (s)

−0.20

0.20.40.60.8

11.21.4

Cur

rent

(A

) x1

0E4

Printed for Opal−RT 1/1Fig. 19. (a) Field voltage & (b) dc fault current for the AFR based systemand (c) field voltage & (d) dc fault current for AVR based system.

WRSG flux and field current. As a result, it is requiredthat vSPT has to be altered at every operating speed forsmoother operation. Thus, additional arrangements areneeded to provide the speed dependent vSPT .

6) It is observed that for similar loading condition fieldcurrent requirement for AFR based dc generation systemis more than that of AVR based dc generation system.This is because the AVR based control is done at PLL

reference frame and AFR based control is done at SFRF.When converted into RRF, the d-axis current (in RRF) ishigher in the case of the AVR based control. Thus higherfield current is needed to counter this d-axis current andalso to establish the WRSG flux. Therefore, the AFRbased system would require lower rated field circuit ascompared to the AVR based system. This becomes anadditional advantage of using AFR based dc generationsystem.With regards to the discussion, the comparison of bothtypes of dc generation system is illustrated in Table IV.

VII. CONCLUSION

This paper investigates the suitability of automatic fluxregulation (AFR) based excitation control of the WRSG for theapplication in the dc generation system for emerging marinevessels. The efficacy of the proposed method is supplementedby the comparative analysis with the conventional automaticvoltage regulation (AVR) based field excitation control. Thisstudy would be helpful in the development and selection ofthe type of field excitation system to control the WRSG forsuch applications. Detailed modeling for both systems aredone which substantiates the suitability of the AFR based dcgeneration system in the emerging dc marine vessels.

NTULSC

Text Box

13

APPENDIX

WRSG Rating : 2048 kVA, 690 V, 60 Hz, 8 pole. ArmatureCircuit : Rs = 0.002591 pu, Xls = 0.007 pu , Xd = 1.87 pu,X ′d = 0.219 pu X ′′d = 0.116 pu Xq = 1.05 pu X ′′q = 0.121 pu.Field Circuit : R′f = 0.002 pu X ′f= 0.5 pu.

REFERENCES

[1] Z. Jin, G. Sulligoi, R. Cuzner, L. Meng, J. C. Vasquez, and J. M.Guerrero, “Next-generation shipboard DC power system: Introductionsmart grid and DC microgrid technologies into maritime electricalnetworks,” IEEE Electrific. Mag., vol. 4, no. 2, pp. 45-57, Jun. 2016.

[2] K. Satpathi, V. M. Balijepalli, and A. Ukil, “Modeling and real-timescheduling of DC platform supply vessel for fuel efficient operation,”IEEE Trans. Transport. Electrific., vol. 3, no. 3, pp. 762-778, Sep. 2017.

[3] K. Satpathi, A. Ukil, N. Thukral, and M. A. Zagrodnik, “Modelling ofdc shipboard power system,” in Proc. Int. Conf. Power Electron. DrivesEnergy Syst. (PEDES), pp. 1–6, Dec. 2016.

[4] K. Satpathi, A. Ukil and J. Pou, “Short-circuit fault management inDC electric ship propulsion system: Protection requirements, review ofexisting technologies and future research trends,” IEEE Trans. Transport.Electrific., vol. 4, no. 1, pp. 272-291, Mar. 2018.

[5] B. Zahedi and L. E. Norum, “Modeling and simulation of all-electricships with low-voltage DC hybrid power systems,” IEEE Trans. PowerElectron., vol. 28, pp. 4525–4537, Oct. 2013.

[6] T. McCoy, “Electric ships past, present, and future,” IEEE Electrific.Mag., vol. 3, pp. 4–11, Jun. 2015.

[7] J. F. Hansen, J. O. Lindtjorn, K. Vanska, and O. Abb, “Onboard DCgrid for enhanced DP operation in ships.” in Proc. Dynamic PositioningConference, Houston. 2011.

[8] J. D. Irwin, Power electronics and motor drives, CRC Press, 2011.[9] D. Jovcic and L. Zhang, “LCL DC/DC converter for DC grids,” IEEE

Trans. Power Del., vol. 28, pp. 2071–2079, Oct. 2013.[10] G. Sulligoi, A. Vicenzutti and R. Menis, “All-electric ship design:

From electrical propulsion to integrated electrical and electronic powersystems,” IEEE Trans. Transport. Electrific., vol. 2, no. 4, pp. 507-521,Dec. 2016.

[11] A. Boveri, F. D’Agostino, A. Fidigatti, E. Ragaini, and F. Silvestro,“Dynamic modeling of a supply vessel power system for DP3 protectionsystem,” IEEE Trans. Transport. Electrific., vol. 2, pp. 570–579, Dec.2016.

[12] Ahmed Abdel-Hafez, “Power generation and distribution system for amore electric aircraft: A review”, INTECH Open Access Publisher, 2012.

[13] H. Yazdanpanahi, W. Xu and Y. W. Li, “A novel fault current controlscheme to reduce synchronous DG’s impact on protection coordination,”IEEE Trans. Power Del., vol. 29, no. 2, pp. 542-551, Apr. 2014.

[14] G. Fusco and M. Russo, “Adaptive voltage regulator design for syn-chronous generator,” IEEE Trans. Energy Convers., vol. 23, no. 3, pp.946-956, Sep. 2008.

[15] “IEEE guide for identification, testing, and evaluation of the dynamicperformance of excitation control systems,” IEEE Std 421.2-2014, pp.163, Jun. 2014.

[16] A. Yazdani and R. Iravan, Voltage-sourced converters in power systems:Modeling, control, and applications, John Wiley & Sons, 2010.

[17] Y. W. Li, “Control and resonance damping of voltage-source and current-source converters with LC filters,” IEEE Trans. Ind. Electron., vol. 56,no. 5, pp. 1511-1521, May 2009.

[18] N. Geddada, M. K. Mishra and MV Manoj Kumar. “SRF based currentcontroller using PI and HC regulators for DSTATCOM with SPWMswitching,” Int. J. Elect. Power Energy Syst., vol. 67, pp. 87-100, May2015.

[19] C. C. Gomes, A. F. Cupertino, and H. A. Pereira. “Damping techniquesfor grid-connected voltage source converters based on LCL filter: Anoverview”, Renew. Sustain. Energy Rev., vol. 81, pp. 116-135, Jan. 2018.

[20] A. K. Jain and V. T. Ranganathan, “Modeling and field oriented control ofsalient pole wound field synchronous machine in stator flux coordinates,”IEEE Trans. Ind. Electron., vol. 58, no. 3, pp. 960-970, Mar. 2011.

[21] S. D. Sudhoff, E. L. Zivi, and T. D. Collins, “Start-up performance ofload-commutated inverter-fed synchronous machine drives,” IEEE Trans.Energy Convers., vol. 10, no. 2, pp. 268-274, Jun. 1995.

[22] D. Fodorean, A. Djerdir, I. A. Viorel, and A. Miraoui, “A double excitedsynchronous machine for direct drive application design and prototypetests,” IEEE Trans. Energy Convers., vol. 22, no. 3, pp. 656-665, Sep.2007.

[23] P. C. Kjaer, T. Kjellqvist, and C. Delaloye, “Estimation of field current invector-controlled synchronous machine variable-speed drives employingbrushless asynchronous exciters,” IEEE Trans. Ind. Appl., vol. 41, no. 3,pp. 834-840, May-Jun. 2005.

[24] F. Nozari, P. A. Mezs, A. L. Julian, Chiping Sun, and T. A. Lipo,“Sensorless synchronous motor drive for use on commercial transportairplanes,” IEEE Trans. Ind. Appl., vol. 31, no. 4, pp. 850-859, Aug.1995.

[25] S. Shinnaka and T. Sagawa, “New optimal current control methodsfor energy-efficient and wide speed-range operation of hybrid-fieldsynchronous motor,” IEEE Trans. Ind. Electron., vol. 54, no. 5, pp. 2443-2450, Oct. 2007.

[26] M. van der Geest, H. Polinder, J. A. Ferreira, and M. Christmann,“Power density limits and design trends of high-speed permanent magnetsynchronous machines,” IEEE Trans. Transport. Electrific., vol. 1, no. 3,pp. 266-276, Oct. 2015.

[27] M. F. M Arani and Y. A. R. I. Mohamed, “Assessment and enhancementof a full-scale PMSG-based wind power generator performance underfaults,” IEEE Trans. Energy Convers., vol. 31, pp. 728-739, Jun. 2016.

[28] F. Gao, X. Zheng, S. Bozhko, C. I. Hill and G. Asher, “Modal analysis ofa PMSG-based DC electrical power system in the more electric aircraftusing eigenvalues sensitivity,” IEEE Trans. Transport. Electrific., vol. 1,no. 1, pp. 65-76, Jun. 2015.

[29] M. M. Chinchilla, S. Arnaltes, and J. C. Burgos, “Control of permanent-magnet generators applied to variable-speed wind-energy systems con-nected to the grid,” IEEE Trans. Energy Convers., vol. 21, no. 1, pp.130-135, Mar. 2006.

[30] U. K. Madawala, T. Geyer, J. B. Bradshaw, and D. M. Vilathgamuwa,“Modeling and analysis of a novel variable-speed cage induction gener-ator,” IEEE Trans. Ind. Electron., vol. 59, pp. 1020–1028, Feb. 2012.

[31] B. Lequesne, “Automotive electrification: The nonhybrid story,” IEEETrans. Transport. Electrific., vol. 1, no. 1, pp. 40-53, Jun. 2015.

[32] S. L. Prakash, M. Arutchelvi, and A. S. Jesudaiyan, ”Autonomous PV-array excited wind-driven induction generator for off-grid application inIndia,” IEEE Trans. Emerg. Sel. Topics Power Electron., vol. 4, no. 4,pp. 1259-1269, Dec. 2016.

[33] V. Nayanar, N. Kumaresan, and N. Ammasai Gounden, “A single-sensor-based MPPT controller for wind-driven induction generators supplyingDC microgrid,” IEEE Trans. Power Electron., vol. 31, no. 2, pp. 1161-1172, Feb. 2016.

[34] R. W. De Doncker, F. Profumo, M. Pastorelli and P. Ferraris, “Compar-ison of universal field oriented (UFO) controllers in different referenceframes,” IEEE Trans. Power Electron., vol. 10, no. 2, pp. 205-213, Mar1995.

[35] K. Satpathi, N. Thukral, A. Ukil, and M. Zagrodnik, “Flux estimationbased DC bus voltage control in marine DC power system,” in Proc.Annu. Conf. IEEE Ind. Electron. Soc. (IECON), Florence, Italy, pp. 1815-1820 Oct. 2016.

[36] P. Kundur, N. J. Balu, and M. G. Lauby, Power system stability andcontrol, vol. 7. McGraw-hill New York, 1994.

[37] P. C. Krause, O. Wasynczuk, S. D. Sudhoff, and S. Pekarek, Analysis ofelectric machinery and drive systems, John Wiley & Sons, 2013.

[38] R. W. Erickson and D. Maksimovic, Fundamentals of power electronics,Springer Science & Business Media, 2007.

[39] H. Shayeghi, A. Younesi, and Y. Hashemi, “Optimal design of arobust discrete parallel FP+ FI+ FD controller for the automatic voltageregulator system”, Int. J. Elect. Power Energy Syst., vol. 67, pp. 66-75,May 2015.

[40] OPAL-RT. OPAL-RT Technologies Official Website. Accessed: Feb. 10,2016. [Online]. Available: http://www.opalrt.com/

[41] S. Abourida, C. Dufour, J. Belanger, G. Murere, N. Lechevin, and B. Yu,“Real-time PC-based simulator of electric systems and drives,” in Proc.Annu. IEEE Appl. Power Electron. Conf. Expo., vol.1. Mar. 2002, pp.433–438.

NTULSC

Text Box

Documents

Design, analysis, and comparison of automatic flux