This article has been accepted for publication in a future issue of this journal, but has not been

fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3044884, IEEE
Transactions on Transportation Electrification
Ni et. al.: Input-Output Small-Signal Stability Analysis of A PLL-Free Direct Power Controlled Partially Power Decoupled 1
More-Electric Shipboard Propulsion System

Input-Output Small-Signal Stability Analysis of

A PLL-Free Direct Power Controlled Partially
Power Decoupled More-Electric Shipboard
Propulsion System
Kai Ni, Member, IEEE, Chun Gan, Member, IEEE, Yihua Hu, Senior Member, IEEE, and Ronghai Qu,
Fellow, IEEE

V, I, E Voltage, current and internal voltage magnitudes

Abstract — With the partially power decoupled system R, L, C Resistance, inductance, capacitance
architecture, a doubly-fed induction motor-based shipboard Z, Y Impedance and admittance
propulsion system (DFIM-SPS) has the potential of increasing the Lls, Llr Stator and rotor leakage inductances
system safety level by decreasing the penetration of power Lm Mutual inductance
electronics. In this paper, a phase-locked-loop-free (PLL-free) Converter-side, stator, and rotor inductances (Ls = Lm +
direct power control (DPC) strategy is proposed for a DFIM-SPS Ls, Lr
Lls; Lr = Lm + Llr)
to increase the system dynamic performance. The system
small-signal stability analysis is carried out based on the P, Q Active and reactive powers
source-load structure, where a salient-pole synchronous generator
S Switching function
supplies power to the DFIM electric drivetrain. A detailed dq
impedance model of DFIM-SPS is established with the power θm, θr Mechanical rotor angle, electrical rotor angle
controller gains of back-to-back power converter (BTBPC) Synchronous angular frequency, mechanical rotor
ω e , ω m, ω r
angular speed, electrical rotor angular speed
considered. The proposed source-load dq impedance model of
Tem, Tm Electromagnetic and mechanical torques
PLL-free DPC-based DFIM-SPS can easily reflect the equivalent
impedance produced in the stator and rotor circuits. The effects n p, H Number of pole pairs, inertia constant
on system stability caused by the equivalent impedance variations kp, ki Proportional gain, integral gain
due to power controller gain changes are evaluated in detail. The H(s) Transfer function
frequency response of the elements in the DFIM impedance Hss(s), Hr(s)
PI transfer functions for SSC and LSC power
matrix is investigated first, and then the minor loop gain matrix is controllers
obtained. The system input-output small-signal stability is ICE Internal combustion engine
analyzed based on Nyquist stability criterion. Time-domain SSC; LSC Source-side converter; load-side converter
simulations are carried out in Matlab/Simulink to validate the SG Synchronous generator
effectiveness of the proposed PLL-free DPC strategy for Gm; Pm Gain margin and phase margin
DFIM-SPS. Re; Im Real and imaginary axes
AVR Active voltage regulator
Index Terms — Doubly-fed induction motor, shipboard
PLL Phase-locked loop
propulsion system, impedance model, direct power control,
PWM Pulse width modulation
Nyquist stability criterion, input-output small-signal stability
Subscripts
s, r, t Stator-side, rotor-side, total variables
NOMENCLATURE
ss, ls SSC and LSC related variables
v, i, ψ Instantaneous voltage, current and flux
l, m Leakage and mutual variables
f Frequency
k, f, g Damper, field, and generator variables
u Control signal
dc dc-bus related variables
p, i Proportional and integral controller variables
Manuscript received Jul. 10, 2020; revised Oct. 09, 2020; accepted Dec. 09,
nom Nominal values
2020. This work was supported by the National Natural Science Foundation of
China under Grant 52007071 and 51907073, the China Postdoctoral Science a, b, c Phases A, B, C
Foundation under Grant 3004131154 and 2020M672355, and the Applied Direct and quadrature components referred to the
d, q
Basic Frontier Program of Wuhan under Grant 2020010601012207. synchrnous reference frame
(Corresponding Author: Chun Gan) Superscripts
K. Ni, C. Gan and R. Qu are with the State Key Laboratory of Advanced * Reference value
Electromagnetic Engineering and Technology, School of Electrical and
Electronic Engineering, Huazhong University of Science and Technology, Note: All the bold variables represent vectors and matrices.
Wuhan 430074, China. (e-mail: [email protected]; [email protected];
[email protected])
Y. Hu is with the Department of Electronics Engineering, Heslington,
University of York, York, YO10 5DD, U.K. (E-mail: [email protected])

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 15,2021 at 06:13:06 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3044884, IEEE
Transactions on Transportation Electrification
Ni et. al.: Input-Output Small-Signal Stability Analysis of A PLL-Free Direct Power Controlled Partially Power Decoupled 2
More-Electric Shipboard Propulsion System
I. INTRODUCTION current proportional-integral (PI) controllers can be eliminated.
In addition, the synchronous reference frame is usually
W ith the fast development of power electronics technology,
the concept of “more-electric ship” (MES) emerges in
recent years, and it has become the production standard for the
established for decoupled control by tracking the voltage/flux
vector position through the use of phase-locked loop (PLL). For
large-scale shipyards all over the world [1]. Although the a DFIM-SPS, the input power to the DFIM electric drivetrain is
advantages of flexible on-board space and weight allocation, supplied by synchronous generators, whose speed changing
more degrees of freedom for shipboard propulsion system (SPS) process can directly influence the system frequency, thus
layout design, increased power efficiency, noise and vibration deteriorating the PLL performance. In this case, the DPC
attenuation, etc. are achieved by extensive power electronics strategy for a DFIM-SPS without PLL is attractive, where the
integration, the easy-to-breakdown feature and nonlinear virtual synchronous reference frame is applied [18].
dynamics of a power electronic device result in potential threats This paper aims to implement small-signal stability analysis
to the safe operation of an MES [2-4]. On top of that, the dc for a PLL-free DPC-based DFIM-SPS by establishing
power distribution architecture is gaining popularity due to the source-load impedance models. The impedance-based stability
flexibility of power regulation brought about by the wide use of criterion was proposed for grid-connected inverters in [19] and
power electronics [2, 5, 6]. For such a system architecture, is widely applied for power electronics-based systems, which is
extremely fast and precise fault detection and isolation are simple and concise to use without the need of knowing every
required, bringing about high technical difficulty and cost in detail of the targeted system [20]. Measured impedances are
fault protection [7, 8]. applied to evaluate the system stability at each interface, and
A doubly-fed induction machine based SPS (DFIM-SPS) is they intrinsically model not only the physical components but
a promising solution to the aforementioned problems [9]. also the controller effects.
Firstly, dual power flow paths based on both ac transmission Through the use of dq transformation for impedance models,
and back-to-back power converter (BTBPC) are presented, the stability of three-phase power converter interfaces can be
instead of totally relying on power electronics for on-board assessed by regarding it as two single-input single-output
power delivery. Therefore, the breakdown of BTBPC does not systems. Alternatively, a three-phase system can be treated as a
halt the continuous operation of MES. Secondly, cost-effective single-phase one by dividing it into positive and negative
fault protection can be realized as the dc power proportion is sequences in the phase domain [21]. Since the small-signal
greatly reduced, and cheap and mature ac circuit breakers are stability issues are widely presented for grid-connected voltage
used to replace the expensive dc ones. Therefore, more system source converters (VSCs), a number of research works on
safety is achieved with less power electronics integrated. For impedance-based stability analysis have been carried out by
the ship types with low demand for dc power and service loads, applying either the dq [22, 23] or sequence impedance model
such as cargo ships and platform supply vessels, the advantages [24-26] for grid-connected VSCs.
of DFIM-SPS can be utilized to the maximum extent. The The authors in [27, 28] applied the Nyquist stability criterion
system structure of DFIM-SPS is displayed in Fig. 1. In the to a DFIM with a series-compensated network. However, as the
proposed work, the targeted application of DFIM-SPS is a effects of PLL and dc-bus voltage control were not included,
cargo ship. inaccurate analysis can be presented for a VOC-based DFIM.
Vector-oriented control (VOC) and direct power control With the full model characteristics considered, a modular
(DPC) strategies are generally applied for an induction motor modelling technique was developed for a DFIM in [29].
[10-17]. For the purpose of improving the dynamic Moreover, a modified sequence impedance modelling method
performance, the DPC strategy is preferred as the redundant [30] was adopted so that a three-port BTBPC small-signal
model can be constructed to consider both the converter and
dc-bus control effects. However, it is not convenient for
revealing the effects of stator and rotor circuits on the
impedance model. In [31], the frequency coupling
characteristic modeling and stability analysis of DFIM were
carried out with a single-in-single-out stability criterion
proposed. Furthermore, the analysis and reshaping on the
impedance characteristic of DFIM by considering a
symmetrical PLL in an improved control strategy were
conducted in [32]. Although the impedance model of a DFIM
has been investigated in these papers, there is no study on the
impedance modelling and stability analysis of DFIM-SPS,
especially for the case where a PLL-free DPC-based
DFIM-SPS is presented.
The internal dynamic characteristics of the electric drivetrain
of a DFIM-SPS was investigated based on an interface friendly
BTBPC average model in the electromechanical control
Fig. 1. The system structure of DFIM-SPS timescale in [33]. Furthermore, a synthetic internal voltage

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 15,2021 at 06:13:06 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3044884, IEEE
Transactions on Transportation Electrification
Ni et. al.: Input-Output Small-Signal Stability Analysis of A PLL-Free Direct Power Controlled Partially Power Decoupled 3
More-Electric Shipboard Propulsion System
vector was established for the electric drivetrain of DFIM-SPS VS
in [34], where the phase-amplitude dynamics of it was VL = ZLI = (2)
1 + ZS / ZL
investigated with respect to the rotor speed control, reactive
control and PLL control effects. Nevertheless, only the electric The small-signal stability of the whole system is to be judged
drivetrain was included in the abovementioned works, and the based on the following transfer function,
overall system small-signal stability was not analyzed.
1
In this paper, a detailed dq impedance model of DFIM-SPS H ( s) = (3)
is established by incorporating the proposed PLL-free DPC 1 + ZS ( s ) / Z L ( s )
strategy. The source-load system structure is formed with the The Nyquist stability criterion is illustrated as follows. All
salient-pole synchronous generator (SPSG) and DFIM acting as zeros of the denominator of H(s) need to be in the open
the system input and output ports, respectively. This paper will left-half-plane. The number of right-half-plane (RHP) poles of
investigate how effective the Nyquist stability criterion is for the impedance ratio [ZS(s)/ZL(s)] should be equal to that of
input-output small-signal stability analysis of a DFIM-SPS counter-clockwise encirclement for the Nyquist map of
controlled by the proposed PLL-free DPC strategy. The [ZS(s)/ZL(s)] around the critical point (-1, 0) in the complex
small-signal stability of the derived model is analyzed with plane. Especially, when there is no RHP pole, the Nyquist map
respect to the DFIM power controller gain setup. The impacts of [ZS(s)/ZL(s)] should not encircle (-1, 0). The stability margin
on the system stability are to be reflected in the impedance can be generally observed according to the distance from the
matrix elements. Moreover, the input-output system Nyquist locus to the critical point.
small-signal stability is further analyzed based on the minor
loop gain matrix, which is obtained as the impedance ratio of III. PLL-FREE DPC OF DFIM-SPS
source to load.
The paper is arranged in the following structure: In Section II, In the DFIM-SPS, an SPSG is the input terminal of the
the Nyquist stability criterion is briefly introduced for the sake system and it is responsible for providing a stable voltage
of analyzing impedance-based models. In Section III, the supply to the propulsion load. In contrast, the DFIM electric
proposed PLL-free DPC strategy is illustrated for both the drivetrain is the output terminal, and its behavior significantly
source-side converter (SSC) and load-side converter (LSC). affects the system stability.
Then, the dq impedance models of DFIM and SPSG are For the DFIM-SPS with PLL-free DPC strategy, stator flux
established in Sections IV and V, respectively. The orientation is implemented to align the d-axis in the same
input-output system small-signal stability analysis is carried out direction as that of the stator flux. The stator flux angle θe is
in Section VI. Followed by that, case studies are presented in directly estimated in the two-phase stationary reference frame
Section VII to validate the proposed PLL-free DPC strategy for (αβ) according to the following algorithm,

 tan( s  /  s ),  s  0
a DFIM-SPS. Finally, the conclusion is presented in Section
VIII. e =  (4)
 tan( s  /  s ) +  ,  s  0
II. IMPEDANCE-BASED NYQUIST STABILITY CRITERION
If ψsα = 0, then
Compared with the state-space-based system stability
analysis method in the time domain, the frequency-domain  / 2,  s  0
method based on impedance modelling is more convenient to 
be applied when the loads are critical in system stability  e = − / 2,  s  0 (5)
0,  s = 0
analysis. The system can be regarded as source and load 
subsystems that are represeneted by input and output
impedances. Based on the linearization of system at a specific After deriving the stator flux angle θe, the variables in the
operation point, the small-signal stability can be analyzed by two-phase arbitrary rotating reference frame (dq) can be
applying the Nyquist stability criterion to the ratio of the source obtained through Park Transformation [35].
output impedance to the load input one. The dq voltage and flux equations of a DFIM are expressed
For the simplest small-signal system model with only one as [36]
source and one load established by the impedance-based
modelling method, the source output and load input vsdq = Rsisdq + d sdq / dt + j e sdq
 (6)
impedances are denoted as ZS and ZL, respectively. The current
vrdq = Rrirdq + d rdq / dt + j slip rdq
flowing through the equivalent circuit is
VS VS  sdq = Lsisdq + Lmirdq
I= = (1)  (7)
ZS + ZL ZL(1 + ZS / ZL)  rdq = Lrirdq + Lmisdq
Therefore, the load input voltage is obtained as where ωslip = ωe – ωr.
In the steady state, the variation in the stator flux dψsdq/dt is
neglected, therefore ψsdq can be estimated by

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 15,2021 at 06:13:06 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3044884, IEEE
Transactions on Transportation Electrification
Ni et. al.: Input-Output Small-Signal Stability Analysis of A PLL-Free Direct Power Controlled Partially Power Decoupled 4
More-Electric Shipboard Propulsion System
 sdq = (vsdq − Rsisdq) / je (8)  2 Lss
 vssd = Qss −  e ssq
A. Power Calculation  3 e sdTs
 (16)
The stator-side and SSC-side active and reactive power Ps, vssq = − 2 Lss Pss +  e ssd
Qs, Pss and Qss are calculated as [37] 
 3 e sdTs
 Ps = 1.5(vsdisd + vsqisq ) The control signals for SSC are then produced through the PI
Qs = 1.5(vsdisq − vsqisd )
 controllers for the power control loop.
 (9)
 Pss = 1.5(vsdissd + vsqissq ) ussd = Hss ( s )(Qss − Qss ) +  eLssissq
*

Qss = 1.5(vsdissq − vsqissd )  (17)

ussq = − Hss ( s )( Pss − Pss ) +  e( sd − Lssissd )
*

B. PLL-Free DPC for SSC where

The stator flux of DFIM ψs can be calculated according to the
kiss
integration of the stator-side and SSC-side voltages [38].
Hss ( s) = kpss + (18)
dissdq s
 sdq =  Vsdqdt =  (Vssdq + Lss )dt =  ssdq + Lssissdq (10)
dt C. PLL-Free DPC for SSC
Therefore, Based on (7), the stator currents can be obtained,

 sdq − ssdq  sdq Lm rdq

issdq = isdq = − (19)
Lss
(11)
 Ls  LsLr
According to (6), (9), (10) and (11), by neglecting the stator where the leakage coefficient σ = 1 – Lm2/LsLr.
resistance and the dynamic process of stator flux ψs, and with According to (6), (9) and (19), by neglecting the stator
the assumption of ψsq = 0, the SSC-side active and reactive resistance and the dynamic process of stator flux ψs, and with
power can be further expressed as the assumption of ψsq = 0, the stator active and reactive power
can be further expressed as
 Pss = 1.5(vsdissd + vsqissq )
  Ps = 1.5(vsdisd + vsqisq)
= 1.5(− e sq  sd − ssd +  e sd  sq − ssq ) 
 Lss Lss =1.5[− e sq (  sd − Lm rd ) +  e sd (  sq − Lm rq )]
= −1.5 e sd ssq / Lss   Ls  LsLr  Ls  LsLr
(12) = −k e sd rq
  (20)
Qss = 1.5(vsdissq − vsqissd ) Qs = 1.5(vsdisq − vsqisd )
  sq − ssq  sd − ssd 
= 1.5(− e sq −  e sd ) =1.5[ e sq ( Lm rq −  sq ) +  e sd ( Lm rd −  sd )]
 Lss Lss   LsLr  Ls  LsLr  Ls
= 1.5 e sd ( ssd − sd ) / Lss  Lr
= k  e sd ( rd −  sd )
As the d-axis stator flux ψsd is constant, the deviations of Pss  Lm
and Qss are only related to the changes in ψss, therefore where kσ = 1.5Lm/(σLsLr).
As the d-axis stator flux ψsd is constant, the deviations of Ps
Pss = −1.5( e sd / Lss ) ssq and Qs are only related to the changes in ψrdq, in which case the
 (13)
Qss = 1.5( e sd / Lss ) ssd following equations can be obtained,

The voltage equation for SSC is Ps = − k  e sd  rq

 (21)
vssdq = d ssdq / dt + j
e ssdq (14) Qs = k  e sd  rd
By discretizing (14) within a sampling period Ts, the With the rotor resistance neglected, the rotor voltage can be
deviations in the dq SSC fluxes ψssd and ψssq are expressed as derived as

 ssd = (vssd +  e ssq )Ts vrdq = d rdq / dt + jslip rdq (22)
 (15)
 ssq = (vssq −  e ssd )Ts By discretizing (22) within a sampling period Ts, the
deviations in the dq rotor fluxes ψrd and ψrq are
By substituting (13) into (15), the expressions for SSC dq
voltages are derived as

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 15,2021 at 06:13:06 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3044884, IEEE
Transactions on Transportation Electrification
Ni et. al.: Input-Output Small-Signal Stability Analysis of A PLL-Free Direct Power Controlled Partially Power Decoupled 5
More-Electric Shipboard Propulsion System

 rd = (vrd +  slip rq)Ts updated as

 (23)
 rq = (vrq −  slip rd )Ts vsdq = ( Rs + sLls )isdq + sLm(isdq + irdq ) + j e sdq
 (28)
By substituting (21) into (23), the expressions for rotor dq vrdq = ( Rr + sLlr )irdq + sLm(isdq + irdq ) + j slip rdq
voltages are derived as
The SSC and LSC can be respectively regarded as the
 Qs voltage sources behind impedances.
vrd = k e sdTs −  slip rq According to Fig. 2, the stator and rotor voltages can be
 (24) obtained based on the active and reactive power control effects.
vrq = − Ps +  slip rd
 k  e sdTs vsd = Hss ( s )(Qss* − Qss ) + ( Rss + sLss )issd

 +  eLssissq
The control signals for LSC are then produced through the PI
vsq = − Hss ( s )( Pss* − Pss ) + ( Rss + sLss )issq
controller for the power control loop. 
 +  e( sd − Lssissd ) (29)
 Ps
vrd = Hr ( s )(Qs* − Qs ) +  slipPs / (k  e sd )
urd = Hr ( s )(Qs * − Qs ) +  slip
 k  e sd 
 (25)
vrq = − Hr ( s )( Ps * − Ps ) +  slip[(Qs / k  e sd )
urq = − Hr ( s )( Ps * − Ps ) +  slip ( Qs + Lr  sd ) 

 k  e sd Lm  + ( Lr sd / Lm)]
where According to (29), SSC and LSC are equivalent to two
voltage sources behind the corresponding impedances. The two
kir
Hr ( s) = kpr + (26) voltage sources are represented as
s
 Essd = Hss ( s )Qss*
The control block diagram of a PLL-free DPC-based  (30)
 Essq = − Hss ( s ) Pss +  e sd
*
DFIM-SPS is shown in Fig. 2.

IV. IMPEDANCE-BASED MODELLING OF DFIM  Elsd = Hr ( s)Qs*


In this part, the DFIM impedance model is built with an ideal  Lr (31)
 Elsq = − Hr ( s ) Ps +  slip sd
*
voltage source vs. The impedances in the dq reference frame
will be revealed with the power control loop effects considered.
 Lm
Combine (4) and (5) and transform the equivalent circuit into The equivalent impedances at the SSC and LSC sides of the
the s-domain, DFIM electric drivetrain are presented as Zss and Zls,
respectively. When observing the circuit from the stator side,
vsdq = Rsisdq + s ( Lsisdq + Lmirdq ) + j e sdq the slip, which equals ωslip/ωe, is to be considered for the
 (27)
vrdq = Rrirdq + s ( Lrirdq + Lmisdq ) + j slip rdq steady-state rotor-side parameters. According to (27) – (31), the
equivalent circuit of DFIM impedance model in the dq frame is
For the convenience of expressing the equivalent circuit by displayed in Fig. 3.
the voltage equations at both the stator and rotor sides, (27) is

Fig. 3. Equivalent circuit of DFIM dq impedance model

Fig. 4. Updated equivalent circuit of DFIM dq impedance model under

Fig. 2. DFIM-SPS control block diagram PLL-free DPC strategy

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 15,2021 at 06:13:06 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3044884, IEEE
Transactions on Transportation Electrification
Ni et. al.: Input-Output Small-Signal Stability Analysis of A PLL-Free Direct Power Controlled Partially Power Decoupled 6
More-Electric Shipboard Propulsion System
The dq equivalent circuit of DFIM impedance model in Fig. vsd = Essd + [ Rss + sLss + Hss ( s)]issd +  eLssissq
3 is derived based on the ac coupling between the stator and vsq = Essq + [ Rss + sLss + Hss ( s)]issq −  eLssissd
rotor circuits. For the sake of deriving the impedance matrix for 
 (35)
vrd = Elsd + Hr ( s)isd + ( slip / k  e)isq
the DFIM propulsion load, the equivalent circuit is updated by
only presenting the relationship between the dq stator voltages
and currents, which is shown in Fig. 4.
vrq = Elsq + Hr ( s)isq − ( slip / k e)isd
It can be seen that there are coupling terms in Zss and Zls,
Then, by substituting (32) into (35), the relation between the
which are expressed as the corresponding voltage drops in the
stator dq voltages and currents, and that between the rotor
updated equivalent circuit of DFIM dq impedance model. In
voltages and currents are derived as follows.
order to further modify the equivalent circuit, the mutual
impedance (Zm = sLm) and the rotor-side impedance should be vsd = Essd + [ Rss + sLss + Hss ( s)]issd +  eLssissq
paralleled and then series connected to the stator-side one. As vsq = Essq + [ Rss + sLss + Hss ( s)]issq −  eLssissd
Zm is much larger than either the impedance at the stator or rotor 
 (36)
vrd = Elsd − Hr ( s)ird − ( slip / k  e)irq
side, it can be neglected in the further updated DFIM dq
impedance equivalent circuit. Therefore,
vrq = Elsq − Hr ( s)irq + ( slip / k  e)ird
irdq = −isdq (32)
According to (36), the expressions for Zss and Zls can be
Based on (32), the further updated equivalent circuit of obtained as follows.
DFIM dq impedance model is illustrated in Fig. 5. As the
direction of isdq is opposite to that of irdq, the equivalent  Zss = Rss + sLss + Hss ( s ) − j eLss
 (37)
resistance Zls becomes negative when observing the circuit
 Zls = − Hr ( s ) + j slip / (k  e)
from the stator side.
In addition, since the proposed model is established in the According to Fig. 5, the impedances for the stator and SSC
per unit system, (9) is updated as circuits are parallel connected to obtain the ultimate impedance
model for the DFIM electric drivetrain.
 Ps = vsdisd + vsqisq
Qs = vsdisq − vsqisd  Zsdd Zsdq 
 Zsdq = 
Zsqq 
(38)
 (33)
 Pss = vsdissd + vsqissq  Zsqd
Qss = vsdissq − vsqissd where

For the steady-state operation, with the stator resistance  Zsdd = Rs + Rr e /  slip + sLls
neglected, the values of vsd and vsq are 0 and 1 in the per unit  + sLlr + Hr ( s ) e /  slip
system, respectively. Therefore, (33) is further updated as 
 Zsdq = − eLls −  eLlr + 1/ k 
 Ps = isq  (39)
Qs = −isd  Zsqd =  eLls +  eLlr − 1/ k 
  Zsqq = Rs + Rr e /  slip + sLls
 (34)
 Pss = issq 
 + sLlr + Hr ( s ) e /  slip
Qss = −issd
 Zssdd Zssdq 
For the steady-state operation of a DFIM in the per unit Zssdq =  
system, ψsd = 1 and ψsq = 0. Substitute (30), (31) and (34) into  Zssqd Zssqq  (40)
(29), the following expressions are obtained.
 Rss + sLss + Hss ( s )  eLss 
=
 − eLss Rss + sLss + Hss ( s ) 
The subscript dd/qq represents the voltage response on the
d/q axis with respect to the current perturbations in the
corresponding axis, and the subscript dq/qd indicates the
coupling effect between d and q axes. The corresponding
admittance matrices are obtained as

Ysdd Ysdq   Zsqq / det(Zsdq) − Zsdq / det(Zsdq) 

Ysdq =  = 
Ysqd Ysqq   − Zsqd / det(Zsdq) Zsdd / det(Zsdq) 
Fig. 5. Further updated equivalent circuit of DFIM dq impedance model under
PLL-free DPC strategy (41)

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 15,2021 at 06:13:06 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3044884, IEEE
Transactions on Transportation Electrification
Ni et. al.: Input-Output Small-Signal Stability Analysis of A PLL-Free Direct Power Controlled Partially Power Decoupled 7
More-Electric Shipboard Propulsion System
Yssdd Yssdq   Zssqq / det(Zssdq) − Zssdq / det(Zssdq)  paper. The d-axis is aligned in the direction of rotor flux, and
Yssdq =  =  the field winding exists on the d-axis. Additionally, there is a
Yssqd Yssqq   − Zssqd / det(Zssdq) Zssdd / det(Zssdq)  damper winding on the d-axis. While for the q-axis, there are
(42) only parallel-connected damper windings, and one q-axis
Therefore, the DFIM admittance matrix is obtained as damper winding is applied in this paper for simplicity. The field
voltage of SG is derived through active voltage regulator (AVR)
YDFIMdd YDFIMdq  control. Since the input and output variables of the AVR are
YDFIMdq = Ysdq + Yssdq =   (43)
both voltages, there is no impedance shaping effect caused by
YDFIMqd YDFIMqq 
SPSG control in this case.
The corresponding DFIM impedance matrix is To distinguish the variables for SPSG from those for DFIM,
the subscript “g” is added. The dq voltage and flux equations
 ZDFIMdd ZDFIMdq  for SPSG are illustrated as shown below.
ZDFIMdq =  
 ZDFIMqd ZDFIMqq  d gsd
(44) 
 YDFIMqq / det(YDFIMdq) −YDFIMdq / det(YDFIMdq)  vgsd = − Rgsigsd + dt −  g gsq
=  
 −YDFIMqd / det(YDFIMdq) YDFIMdd / det(YDFIMdq)  vgsq = − Rgsigsq + d gsq +  g gsd
The details of det(Zsdq), det(Zssdq), and YDFIMdq are displayed  dt

as follows.  d f
vf = Rfif + (48)
det(Zsdq) = ( Lls + Llr ) s + 2[ Rs + ( Rr + kpr ) e /  slip]
2 2
 dt
 d kd
( Lls + Llr ) s + [ Rs + ( Rr + kpr ) e /  slip]2 + [( eLls + vkd = Rkdikd + dt
(45)
 eLlr − 1/ k )2 + 2( Lls + Llr )kir e /  slip ] + 2[ Rs + 
vkq = Rkqikq + d kq
( Rr + kpr ) e /  slip](kir e /  slip) s −1 + (kir e /  slip) 2 s −2  dt
det(Zssdq) = [ Rss + sLss + Hss ( s)]2 +  e 2 Lss 2 (46)  gsd = − Lgsdigsd + Lgmd (if + ikd )
 gsq = − Lgsqigsq + Lgmqikq
YDFIMdd = Ysdd + Yssdd 
YDFIMdq = Ysdq + Yssdq  f = Lfif + Lgmd (−igsd + ikd ) (49)

 (47)  kd = Lkdikd + Lgmd (−igsd + if )
YDFIMqd = Ysqd + Yssqd 
YDFIMqq = Ysqq + Yssqq  kq = Lkqikq − Lgmqigsq
To express the impedance model of SPSG, (48) and (49) are
V. IMPEDANCE-BASED MODELLING OF SPSG combined and transformed into the s-domain as follows,
An SPSG acts as the power source of DFIM-SPS in this
vgsd = −( Rgs + sLgls )igsd + sLgmd (−igsd + if + ikd ) −  g gsq
vgsq = −( Rgs + sLgls )igsq + sLgmq (−igsq + ikq ) +  g gsd

vf = ( Rf + sLlf )if + sLgmd (−igsd + if + ikd ) (50)
vkd = ( Rkd + sLlkd )ikd + sLgmd (−igsd + if + ikd )

vkq = ( Rkq + sLlkq )ikq + sLgmq (−igsq + ikq )

According to (50), the equivalent circuit of SPSG dq

impedance model is presented in Fig. 6.
In an SPSG, the d-axis mutual inductance Lgmd is much larger
than the q-axis one Lgmq, in which case Lgmd is ignored when
paralleling the rotor-side and mutual impedances, while Lgmq
cannot be omitted.
On the d-axis, the field and damper winding impedances are
paralleled to derive the rotor-side d-axis impedance, which is

Zfkd = (Rkd + sLlkd ) || (Rf + sLlf ) (51)

On the other hand, the rotor-side q-axis impedance is derived

Fig. 6. Equivalent circuit of SPSG dq impedance model for (a) d-axis (b) q-axis
by paralleling q-axis damper winding and mutual impedances,

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 15,2021 at 06:13:06 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3044884, IEEE
Transactions on Transportation Electrification
Ni et. al.: Input-Output Small-Signal Stability Analysis of A PLL-Free Direct Power Controlled Partially Power Decoupled 8
More-Electric Shipboard Propulsion System

Fig. 7. Updated equivalent circuit of SPSG dq impedance model by ignoring Fig. 8. Further updated equivalent circuit of SPSG dq impedance model for (a)
Lgmd for (a) d-axis (b) q-axis d-axis (b) q-axis

Zkq = (Rkq + sLlkq) || sLgmq (52) YDFIMdq = ZDFIMdq−1 (56)

For the purpose of expressing the SPSG impedance model Under the proposed PLL-free DPC strategy, only the setup of
more clearly, only the stator dq voltages and currents are SSC and stator power controller gains shapes the impedance.
employed as the variables in the updated equivalent circuit, The primitive power controller parameters used in system
which is displayed in Fig. 7. According to Kirchhoff’s Current stability analysis are given in Appendix.
Law and the Law of Shunt, the following conditions are
A. SSC Power Control Effects on System Stability
satisfied.
The small-signal stability of the DFIM electric drivetrain is
if + ikd = igsd to be investigated first as it functions as the unique load in the
 (53) case considered in this paper. The effects on the small-signal
ikq = igsqsLgmq / ( Rkq + sLkq ) stability of DFIM electric drivetrain caused by the SSC power
controller parameter setup are analyzed from two aspects.
Based on (53), the equivalent circuit of SPSG impedance Firstly, set the SSC proportional power controller gain kpss to
model can be further updated to reveal the dq impedance 0.1, 1 and 10 respectively with the SSC integral power
derived according to stator voltages and currents, which is controller gain kiss remaining at 5. Secondly, the SSC integral
shown in Fig. 8. Finally, the dq impedance matrix for SPSG is power controller gain kiss is set to 0.5, 5 and 50 respectively
obtained as with the SSC proportional power controller gain kpss remaining
 Zgdd Zgdq  at 1. The corresponding Bode plots are shown in Figs. 9 and 10.
Zgdq =  
 Zgqd Zgqq 
 − Rgs − sLgls − Zfkd  g[ sLgmq 2 / ( Rkq + sLkq) − Lgsq]
= 
  g (− Lgsd + Lgmd ) − Rgs − sLgls − Zkq 
(54)

VI. INPUT-OUTPUT SMALL-SIGNAL STABILITY ANALYSIS FOR

DFIM-SPS UNDER PLL-FREE DPC STRATEGY
Based on the input-output DFIM-SPS source-load structure,
the impedance-based stability analysis is carried out by
constructing the minor loop gain matrix, which is calculated as
the impedance ratio of the SPSG impedance matrix to the
DFIM one. The minor loop gain matrix LGdq is derived as
LGdq = ZgdqYDFIMdq (55)
where
Fig. 9. Frequency response of ZDFIMdq with the SSC proportional power
controller gain kpss set to 0.1, 1 and 10 respectively

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 15,2021 at 06:13:06 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3044884, IEEE
Transactions on Transportation Electrification
Ni et. al.: Input-Output Small-Signal Stability Analysis of A PLL-Free Direct Power Controlled Partially Power Decoupled 9
More-Electric Shipboard Propulsion System
From Fig. 9, it can be seen the elements in all channels keep On the other hand, the phase margins for the dd- and
stable for all the three different values of kpss. The gain margins qq-channel elements ZDFIMdd and ZDFIMqq are both negative
for the dd- and qq-channel elements are infinite for all the three when kiss = 0.5, indicating that unstable cases can happen for the
SSC proportional controller parameters, while the highest DFIM electric drivetrain during the transient process, as can be
phase margin is derived when kpss = 1, and the lowest one is seen from Fig. 10. In addition, positive gain and phase margins
presented for the case when kpss = 0.1. For the dq- and are available for the other cases. For the qd-channel element
qd-channel elements ZDFIMdq and ZDFIMqd, the phase margins are ZDFIMqd, the gain margin keeps increasing with the increment of
all infinite for the three controller parameters, and they all have kiss, and the phase margins are all infinite.
positive gain margins. Therefore, small-signal stability of The input-output small-signal stability for DFIM-SPS with
DFIM electric drivetrain is ensured for the cases with kpss respect to the changes in kpss and kiss is presented with the
values of 0.1, 1 and 10, when the other controller parameters Nyquist plots for LGdq in Figs. 11 and 12. It can be seen from
remain at the primitive values as shown in Appendix. Fig. 11 that the critical point (-1, 0) is not encircled by any
Nyquist locus in all the channels except that in the qd-channel.
From the enlarged view of LGqd, the critical point is encircled
by the Nyquist plot when kpss = 1 is applied. In this case,
unstable operation of DFIM-SPS can occur if coupling between
the variables in the d- and q-axes exists. On the other hand,
according to Fig. 12, when the value of kiss is set to 0.5, the
critical point (-1, 0) is encircled by the Nyquist locus of the dq-
and qq-channel elements LGdq and LGqq of the loop gain matrix,
which indicates instability.
However, for the case of kiss = 0.5, negative phase margins
are obtained according to the frequency response shown in Fig.
10, in which case the Nyquist plots are not reliable for
analysing the input-output small-signal stability for the overall
system [19]. Since the case of kiss = 5 has the same Nyquist
plots as shown by the red ones in Fig. 11, they are not displayed
again in Fig. 12.
B. LSC Power Control Effects on System Stability
Fig. 10. Frequency response of ZDFIMdq with the SSC integral power controller As the LSC is responsible for controlling the operation of
gain kiss set to 0.5, 5 and 50 respectively
DFIM, the corresponding effects of power controller gain setup
needs to be investigated in detail. Firstly, set the LSC
proportional power controller gain kpr to 0.1, 1 and 1000
respectively with the LSC integral power controller gain kir
remaining at 8. Secondly, the LSC integral power controller
gain kir is set to 0.008, 8 and 80 respectively with the LSC
proportional power controller gain kpr remaining at 1. The

Fig. 11. Nyquist plots of LGdq with the SSC proportional power controller gain Fig. 12. Nyquist plots of LGdq with the SSC integral power controller gain kiss
kpss set to 0.1, 1 and 10 respectively set to 0.5 and 50 respectively

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 15,2021 at 06:13:06 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3044884, IEEE
Transactions on Transportation Electrification
Ni et. al.: Input-Output Small-Signal Stability Analysis of A PLL-Free Direct Power Controlled Partially Power Decoupled 10
More-Electric Shipboard Propulsion System
corresponding Bode diagrams are shown in Figs. 13 and 14. = 0.1 as shown in Fig. 15. However, since the phase margin of
It can be seen from Fig. 13 that for the given values of kpr ZDFIMdq is negative as can be seen from Fig. 13, the Nyquist
when kir remains at 8, the dd- and qq-channel elements ZDFIMdd plots in Fig. 15 cannot fully reveal the small-signal stability of
and ZDFIMqq can achieve the stable state for the cases of kpr = 0.1 DFIM-SPS when kpr = 0.1. Similarly, although the Nyquist
and kpr = 1. However, negative phase margins are presented plots in Fig. 15 show stable performance for the DFIM-SPS
when kpr = 1000, which results in unstable operation. Besides, when kpr = 1000 as there is no encirclement around the critical
for the dq-channel element ZDFIMdq, negative phase margins are point, the system can still encounter unstable situations because
achieved when kpr = 0.1 or 1000. For the case of ZDFIMqd, the the Bode plots show negative phase margins in the dd-, dq- and
gain margin decreases with the increase of kpr. From Fig. 14, qq-channel elements. Moreover, according to Fig. 16, the
stable operation can be achieved for ZDFIMdd and ZDFIMqq when Nyquist plot for the qq-channel element LGqq encircles the
kir = 0.008 or 8, while the gain margin for the case of kir = 80 is
negative. In addition, a large kir value threatens the system
stability in terms of the frequency response of ZDFIMdq.
The input-output small-signal stability for DFIM-SPS with
respect to the changes in kpr and kir is presented with the
Nyquist plots for LGdq in Figs. 15 and 16. The critical point (-1,
0) is encircled by the Nyquist locus for LGdq and LGqd when kpr

Fig. 15. Nyquist plots of LGdq with the LSC proportional power controller gain
kpr set to 0.1 and 1000 respectively

Fig. 13. Frequency response of ZDFIMdq with the LSC proportional power
controller gain kpr set to 0.1, 1 and 1000 respectively

Fig. 14. Frequency response of ZDFIMdq with the LSC integral power controller Fig. 16. Nyquist plots of LGdq with the LSC integral power controller gain kir
gain kir set to 0.008, 8 and 80 respectively set to 0.008 and 80 respectively

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 15,2021 at 06:13:06 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3044884, IEEE
Transactions on Transportation Electrification
Ni et. al.: Input-Output Small-Signal Stability Analysis of A PLL-Free Direct Power Controlled Partially Power Decoupled 11
More-Electric Shipboard Propulsion System
critical point when kir = 0.008, and those for the dd- and
dq-channel elements encircle the critical point when kir = 80. As
the negative phase margin is presented for the dq-channel
DFIM impedance element ZDFIMdq and negative gain margins
are developed for the dd- and qq-channel DFIM impedance
elements ZDFIMdd and ZDFIMqq when kir = 80, the purple Nyquist
plots in Fig. 16 are not reliable for overall system stability
analysis.
According to the analysis above, the small-signal stability of
a DFIM-SPS controlled by the proposed PLL-free DPC
strategy with respect to the variations in power controller gains Fig. 17. Case 1 simulation results of (a) total input active power Pt in MW and
(b) electromagnetic & mechanical torques Tem & Tm in pu for the PLL-free
is displayed in TABLE I. DPC-based DFIM-SPS for the SSC power controller gain setup in Figs. 9 - 12
TABLE I (left: kiss = 5 with kpss changes; right: kpss = 1 with kiss changes; all the cases
SMALL-SIGNAL STABILITY OF A DFIM-SPS CONTROLLED BY THE PROPOSED should be stable)
PLL-FREE DPC STRATEGY WITH RESPECT TO THE VARIATIONS IN POWER
CONTROLLER GAINS
The value of kpss with
0.1 1 10
kiss = 5, kpr = 1, kir = 8
Stable if the
coupling between dq
Small-signal stability Stable Stable
variables is
eliminated
The value of kiss with
0.5 5 50
kpss = 1, kpr = 1, kir = 8
Stable if the
Not coupling between dq
Small-signal stability Stable
applicable variables is
eliminated
The value of kpr with Fig. 18. Case 1 simulation results of (a) total input active power Pt in MW and
0.1 1 1000
kpss = 1, kiss = 5, kir = 8 (b) electromagnetic & mechanical torques Tem & Tm in pu for the PLL-free
Stable if the DPC-based DFIM-SPS for the LSC power controller gain setup in Figs. 13 - 16
Not coupling between dq (left: kir = 8 with kpr changes; right: kpr = 1 with kir changes; the cases of kpr =
Small-signal stability Unstable
applicable variables is 1000, kir = 8 and kpr = 1, kir = 80 should be unstable)
eliminated
The value of kir with
0.008 8 80
kpss = 1, kiss = 5, kpr = 1
Stable if the
coupling between dq Not
Small-signal stability Unstable
variables is applicable
eliminated

It can be summarized from TABLE I that stable operation of

the DFIM-SPS controlled by the proposed PLL-free DPC
strategy can be derived for most of the cases on condition that
the coupling between dq variables is fully eliminated. However,
when a large kpr or a small kir is applied, unstable performance Fig. 19. Case 2 simulation results of (a) total input active power Pt in MW and
will be derived. Moreover, the small-signal stability of the (b) electromagnetic & mechanical torques Tem & Tm in pu for the PLL-free
whole DFIM-SPS is not identifiable for certain scenarios, DPC-based DFIM-SPS for the SSC power controller gain setup in Figs. 9 - 12
(left: kiss = 5 with kpss changes; right: kpss = 1 with kiss changes; all the cases
where negative phase margins are presented for the Bode plots
should be stable)
of the DFIM electric drivetrain.

VII. CASE STUDY

A. Case 1
A case study is carried out in Matlab/Simulink with the
sampling time of 5μs by applying different power controller
gains to investigate the time-domain performance of
DFIM-SPS. The load torque first increases from 0.1pu to 0.2pu
at 3s, and then it further increases from 0.2pu to 0.7pu at 7s.
The simulation results of total input power active Pt and the
electromagnetic & mechanical torques Tem & Tm are displayed Fig. 20. Case 2 simulation results of (a) total input active power Pt in MW and
in Figs. 17 & 18 to investigate the control performance for the (b) electromagnetic & mechanical torques Tem & Tm in pu for the PLL-free
SSC and LSC power controller gain setup in the stability DPC-based DFIM-SPS for the LSC power controller gain setup in Figs. 13 - 16
analysis section for Case 1, and also verify the small-signal (left: kir = 8 with kpr changes; right: kpr = 1 with kir changes; the cases of kpr =
1000, kir = 8 and kpr = 1, kir = 80 should be unstable)

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 15,2021 at 06:13:06 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3044884, IEEE
Transactions on Transportation Electrification
Ni et. al.: Input-Output Small-Signal Stability Analysis of A PLL-Free Direct Power Controlled Partially Power Decoupled 12
More-Electric Shipboard Propulsion System
stability analysis results. operation shown by the Nyquist plots for the case of kir = 0.008
It can be seen that by applying the proposed PLL-free DPC in Fig. 16 is trustworthy as there is neither negative gain nor
strategy for DFIM-SPS, fast and accurate tracking performance phase margins for the elements in the dq impedance matrix of
of the total input active power and electromagnetic torque are DFIM electric drivetrain when kir = 0.008 as displayed in Fig.
achieved, and all the three lines with different colors coincide 14, and unstable operation is observed for this case.
completely. However, when the LSC integral controller gain kir Furthermore, the case of applying kir = 80 shows stable
changes to 0.008, the input active power and electromagnetic operation for DFIM-SPS, as the purple Nyquist plots in Fig. 16
torque can no longer track the changes in the load, indicating cannot reveal the real small-signal stability due to the negative
unstable operation. According to the small-signal stability gain and phase margins shown in Fig. 14.
analysis in the previous section, the DFIM-SPS should also be
C. Case 3
unstable when kpr = 1000 with other 3 power controller gains at
the primitive values, while it did not happen. Therefore, other In addition, for the sake of investigating the steady-state
operation conditions need to be further investigated. performance of a DFIM-SPS for the cargo ship propulsion
application, a case study that emulates the operation of a cargo
B. Case 2 ship with the loading and unloading processes of cargos during
Since the above-mentioned case is not able to reveal some the cruise taken into account is implemented. The primitive
unstable operation scenarios as illustrated in Figs. 11, 12, 15 controller parameters listed in Appendix are applied in this case.
and 16, another case with an abrupt load torque change from In the first hour of operation, there is no cargo on the ship,
0pu to 1pu is implemented at 2s, and it decreases from 1pu to therefore the load torque Tl equals 0. Then, at the beginning of
0.2pu at 5s, and then it increases to 0.7pu at 9s. The simulation the second hour, the load torque increases to 0.1pu to represent
results of total input active power Pt and the electromagnetic & the light loading condition, and it further increases to 0.5pu at
mechanical torques Tem & Tm are displayed in Figs. 19 & 20 to the start of the fourth hour of operation. From the sixth hour, the
investigate the control performance for the SSC and LSC power cargo ship operates at full load. Afterwards, the load shedding
controller gain setup in the stability analysis section for Case 2, process begins from the eighth hour, and Tl decreases from 1pu
and also verify the small-signal stability analysis results. to 0.6pu. For the last hour of operation, the ship turns back to
It can be observed from Fig. 19 that stable operation can be the light loading condition with Tl = 0.2pu.
obtained for different power controller parameters used for The input power to the DFIM electric powertrain and the
SSC in the stability analysis section, and all the three lines with load & electromagnetic torques during this operation process
different colors coincide completely. Since the DFIM electric are displayed in Fig. 21. It can be seen that the electromagnetic
drivetrain is controlled by decoupling the dq variables, the torque precisely tracks the load torque, and the input power
unstable situation shown in the dq-channel loop gain for the changes with respect to the loading and unloading processes.
case of kpss = 1 in Fig. 11 is avoided. In addition, due to the
presence of negative phase margins in LGdd and LGqq for the
case of kiss = 0.5, the instability shown by the Nyquist plots in
Fig. 12 is not reliable. For Fig. 20, the unstable operation is
presented when kpr = 1000 or when kir = 0.008, while the other
two lines still coincide with each other. According to Fig. 13,
negative phase margins are shown for the case of kpr = 1000,
therefore the stable operation displayed in Fig. 15 is not
guaranteed, which is why the tracking processes of the input
power and electromagnetic torque fail. Moreover, the unstable

Fig. 22. HIL results of total input active power Pt in MW and reactive power Qt
in MVar with the primitive controller parameters

Fig. 21. Simulation results of (a) total input active power Pt in MW, and (b) load
& electromagnetic torques Tl & Tem in pu for the PLL-free DPC-based
Fig. 23. HIL results of total input active power Pt in MW and reactive power Qt
DFIM-SPS for the cargo ship application with loading and unloading processes
in MVar with the primitive controller parameters except that kpr changes to
considered
1000

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 15,2021 at 06:13:06 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3044884, IEEE
Transactions on Transportation Electrification
Ni et. al.: Input-Output Small-Signal Stability Analysis of A PLL-Free Direct Power Controlled Partially Power Decoupled 13
More-Electric Shipboard Propulsion System
D. Hardware-in-the-loop Validation 2) SG parameters used in simulation
For further validation of the proposed control strategy in a Pgnom = 10MW; Vgnom = 4.16kV; fgnom = 50Hz; ngp = 20; Rgs =
more practical operation condition, hardware-in-the-loop (HIL) 2.85*10-3pu; Lgls = 0.114pu; Lgmd = 1.19pu; Lgmq = 0.36pu; Rf =
operation is implemented by establishing the power circuits of 5.79*10-4pu; Llf = 0.114pu; Rkd = 0.0117pu; Llkd = 0.182pu; Rkq
BTBPC with the connection to an SPSG in the software of = 0.0197pu; Llkq = 0.384pu.
dSPACE DS1007, and it outputs the control variables of 3) Primitive power controller parameters
BTBPC to the target PC. Besides, the power switches in the
kpss = 1; kiss = 5; kpr = 1; kir = 8.
BTBPC are controlled according to the control signals from the
target PC. The power controllers are enabled at 2s. The
REFERENCES
waveforms of the total input active and reactive power by
[1] E. Skjong, R. Volden, E. Rodskar, M. Molinas, T. A. Johansen, and J.
applying the primitive controller parameters and by changing Cunningham, "Past, Present, and Future Challenges of the Marine
kpr to 1000 are respectively shown in Figs. 22 and 23. It can be Vessel’s Electrical Power System," IEEE Trans. Transport. Electrific., vol.
seen that when the primitive power controller parameters are 2, no. 4, pp. 522-537, Dec. 2016.
[2] Z. Jin, L. Meng, J. M. Guerrero, and R. Han, "Hierarchical Control Design
used, the total input active and reactive power can achieve for a Shipboard Power System With DC Distribution and Energy Storage
constant values and maintain stable operation, while unstable Aboard Future More-Electric Ships," IEEE Trans. Ind. Informat., vol. 14,
operation is presented when kpr is changed to 1000. no. 2, pp. 703-714, Feb. 2018.
[3] S. Yang, D. Xiang, A. Bryant, P. Mawby, L. Ran, and P. Tavner,
"Condition Monitoring for Device Reliability in Power Electronic
VIII. CONCLUSION Converters: A Review," IEEE Trans. Power Electron., vol. 25, no. 11, pp.
In this paper, a detailed dq impedance model is established 2734-2752, Nov. 2010.
[4] J. Fang, H. Li, Y. Tang, and F. Blaabjerg, "On the Inertia of Future
for a PLL-free DPC-based DFIM-SPS in a cargo ship, whose More-Electronics Power Systems," IEEE J. Emerg. Sel. Topics Power
input-output small-signal stability is analyzed with respect to Electron., vol. 7, no. 4, pp. 2130-2146, Dec. 2019.
the power controller gain setup. The stator and rotor circuit [5] V. Nougain, S. Mishra, and A. K. Pradhan, "MVDC Microgrid Protection
using a Centralized Communication with a Localized Backup Scheme of
impedances can be clearly revealed in the presented impedance Adaptive Parameters," IEEE Trans. Power Del., vol. 34, no. 3, pp.
model. An SPSG is chosen as the power source for the 869-878, Jun. 2019.
DFIM-SPS. The frequency response of the elements in the [6] U. Javaid, F. D. Freijedo, W. van der Merwe, and D. Dujic, "Stability
Analysis of Multi-Port MVDC Distribution Networks for All-Electric
DFIM dq impedance matrix with respect to the BTBPC power Ships," IEEE J. Emerg. Sel. Topics Power Electron., vol. 8, no. 2, pp.
controller gain variations is investigated in detail. Moreover, 1164-1177, Jun. 2020.
the Nyquist stability criterion is applied to the derived minor [7] P. Cairoli and R. A. Dougal, "Fault Detection and Isolation in
Medium-Voltage DC Microgrids: Coordination Between Supply Power
loop gain matrix for system small-signal stability analysis. Converters and Bus Contactors," IEEE Trans. Power Electron., vol. 33, no.
From the results of stability analysis and case study, the 5, pp. 4535-4546, May 2018.
following points are obtained. [8] K. Satpathi, A. Ukil, and J. Pou, "Short-Circuit Fault Management in DC
Electric Ship Propulsion System: Protection Requirements, Review of
1) The DFIM electric drivetrain is able to achieve stable Existing Technologies and Future Research Trends," IEEE Trans.
operation with SSC proportional power controller gain changes, Transport. Electrific., vol. 4, no. 1, pp. 272-291, Mar. 2018.
while unstable operation of DFIM-SPS can still occur if the [9] K. Ni, Y. Hu, Z. Wang, H. Wen, and C. Gan, "Asynchronized
controls of dq variables are not well decoupled. Synchronous Motor-Based More Electric Ship—Less Power Electronics
for More System Reliability," IEEE/ASME Trans. on Mechatron., vol. 24,
2) When unstable situations are encountered for the DFIM no. 5, pp. 2353-2364, Oct. 2019.
impedance elements, the results derived from the Nyquist plots [10] M. E. Zarei, C. Veganzones Nicolas, and J. Rodriguez Arribas, "Improved
for the overall DFIM-SPS small-signal stability analysis are not Predictive Direct Power Control of Doubly Fed Induction Generator
During Unbalanced Grid Voltage Based on Four Vectors," IEEE J. Emerg.
reliable. Sel. Topics Power Electron., vol. 5, no. 2, pp. 695-707, Jun. 2017.
3) Stable operation of DFIM electric drivetrain cannot ensure [11] H. Nian and L. Li, "Direct Power Control of Doubly Fed Induction
small-signal stability of the whole DFIM-SPS when a small Generator Without Phase-Locked Loop Under Harmonically Distorted
Voltage Conditions," IEEE Trans. Power Electron., vol. 33, no. 7, pp.
LSC proportional power controller gain is applied. 5836-5846, Jul. 2018.
4) The unstable performance can only be reflected in certain [12] D. Sun, X. Wang, H. Nian, and Z. Q. Zhu, "A Sliding-Mode Direct Power
operation conditions. Control Strategy for DFIG Under Both Balanced and Unbalanced Grid
Conditions Using Extended Active Power," IEEE Trans. Power Electron.,
5) For the stable operation scenarios, load torque variations can vol. 33, no. 2, pp. 1313-1322, Feb. 2018.
be swiftly and accurately tracked by the electromagnetic torque [13] Y. Zhang, J. Jiao, and D. Xu, "Direct Power Control of Doubly Fed
by applying the proposed PLL-free DPC strategy for a Induction Generator Using Extended Power Theory Under Unbalanced
Network," IEEE Trans. Power Electron., vol. 34, no. 12, pp. 12024-12037,
DFIM-SPS. Dec. 2019.
[14] N. E. Ouanjli et al., "Modern improvement techniques of direct torque
APPENDIX control for induction motor drives-a review," Protection and Control of
Modern Power Systems, vol. 4, no. 2, pp. 136-147, 2019.
1) DFIM parameters used in simulation [15] L. Xiong, J. Wang, X. Mi, and M. W. Khan, "Fractional Order Sliding
Mode Based Direct Power Control of Grid-Connected DFIG," IEEE Trans.
Pnom = 5MW; Vnom = 4.16kV; fnom = 50Hz; Cdc = 10mF; Rs = Power Syst., vol. 33, no. 3, pp. 3087-3096, May 2018.
0.023pu; Lls = 0.18pu; Rr = 0.016pu; Llr = 0.16pu; Lm = 2.9pu; H [16] J. Hu, Y. Li, and J. Zhu, "Multi-objective model predictive control of
= 3.5s; np = 3; Rss = 0.003pu; Lss = 0.3pu. doubly-fed induction generators for wind energy conversion," IET
Generation, Transmission & Distribution, vol. 13, no. 1, pp. 21-29, 2019.

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 15,2021 at 06:13:06 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3044884, IEEE
Transactions on Transportation Electrification
Ni et. al.: Input-Output Small-Signal Stability Analysis of A PLL-Free Direct Power Controlled Partially Power Decoupled 14
More-Electric Shipboard Propulsion System
[17] N. E. Ouanjli et al., "Direct torque control of doubly fed induction motor [38] D. Zhi, L. Xu, and B. W. Williams, "Improved Direct Power Control of
using three-level NPC inverter," Protection and Control of Modern Power Grid-Connected DC/AC Converters," IEEE Trans. Power Electron., vol.
Systems, vol. 4, no. 3, pp. 196-204, 2019. 24, no. 5, pp. 1280-1292, May 2009.
[18] L. Li, H. Nian, L. Ding, and B. Zhou, "Direct Power Control of DFIG
System Without Phase-Locked Loop Under Unbalanced and Kai Ni (S’17-M’20) was born in Jiangsu, China. He
Harmonically Distorted Voltage," IEEE Trans. Energy Convers., vol. 33, received the B.Eng. (Hons) degree and Ph.D. degree
no. 1, pp. 395-405, Mar. 2018. in Electrical Engineering from the University of
[19] J. Sun, "Impedance-Based Stability Criterion for Grid-Connected Liverpool, Liverpool, UK, in 2016 and 2019,
Inverters," IEEE Trans. Power Electron., vol. 26, no. 11, pp. 3075-3078, respectively. After obtaining the support from the
Nov. 2011. Postdoctoral International Exchange Program, he
[20] X. Wang, L. Harnefors, and F. Blaabjerg, "Unified Impedance Model of joined the School of Electrical and Electronic
Grid-Connected Voltage-Source Converters," IEEE Trans. Power Engineering at Huazhong University of Science and
Electron., vol. 33, no. 2, pp. 1775-1787, Feb. 2018. Technology from Dec. 2019 as a postdoctoral
[21] A. Rygg, M. Molinas, C. Zhang, and X. Cai, "On the Equivalence and researcher. His research interests include modelling,
Impact on Stability of Impedance Modeling of Power Electronic control and stability analysis of doubly-fed induction
Converters in Different Domains," IEEE J. Emerg. Sel. Topics Power machines, power electronic converters, and aircraft/shipboard power systems.
Electron., vol. 5, no. 4, pp. 1444-1454, Dec. 2017. Dr. Ni was the recipient of the 2019 Chinese Government Award for
[22] B. Wen, R. Burgos, D. Boroyevich, P. Mattavelli, and Z. Shen, "AC Outstanding Self-financed Students Abroad.
Stability Analysis and dq Frame Impedance Specifications in
Power-Electronics-Based Distributed Power Systems," IEEE J. Emerg. Chun Gan (M’14) received the B.S. and M.S.
Sel. Topics Power Electron., vol. 5, no. 4, pp. 1455-1465, Dec. 2017. degrees in electrical engineering from China
[23] B. Wen, D. Boroyevich, R. Burgos, P. Mattavelli, and Z. Shen, "Inverse University of Mining and Technology, Jiangsu, China,
Nyquist Stability Criterion for Grid-Tied Inverters," IEEE Trans. Power in 2009 and 2012, respectively, and the Ph.D. degree
Electron., vol. 32, no. 2, pp. 1548-1556, Feb. 2017. in electrical engineering and motor drives from
[24] M. K. Bakhshizadeh et al., "Couplings in Phase Domain Impedance Zhejiang University, Hangzhou, China, in 2016.
Modelling of Grid-Connected Converters," IEEE Trans. Power Electron., He is currently a Professor with the School of
vol. 31, no. 10, pp. 6792-6796, Oct. 2016. Electrical and Electronic Engineering, Huazhong
[25] W. Wu et al., "Sequence-Impedance-Based Stability Comparison University of Science and Technology, Wuhan, China.
Between VSGs and Traditional Grid-Connected Inverters," IEEE Trans. From 2016 to 2018, he was a Research Associate with
Power Electron., vol. 34, no. 1, pp. 46-52, Jan. 2019. the Department of Electrical Engineering and
[26] W. Wu et al., "Sequence Impedance Modeling and Stability Comparative Computer Science, University of Tennessee, Knoxville, TN, USA. He has
Analysis of Voltage-Controlled VSGs and Current-Controlled VSGs," authored/coauthored more than 80 peer-reviewed technical papers, including
IEEE Trans. Ind. Electron., vol. 66, no. 8, pp. 6460-6472, Aug. 2019. more than 50 IEEE Transaction papers. He has over 20 issued/published
[27] L. Fan and Z. Miao, "Nyquist-Stability-Criterion-Based SSR Explanation invention patents. His research interests include electrical motor drives,
for Type-3 Wind Generators," IEEE Trans. Energy Convers., vol. 27, no. electrical motor converters, motor design, electric vehicles, electrified
3, pp. 807-809, Sep. 2012. transportation, and high-efficiency power converters.
[28] Z. Miao, "Impedance-Model-Based SSR Analysis for Type 3 Wind Dr. Gan was the recipient of the 2018 Highlighted Paper Award from IEEE
Generator and Series-Compensated Network," IEEE Trans. Energy TPEL, the 2018 Marie Sklodowska-Curie Actions Seal of Excellence Award
Convers., vol. 27, no. 4, pp. 984-991, Dec. 2012. from European Commission, and the 2019 Best Paper Award from ICEMS.
[29] C. Zhang, X. Cai, M. Molinas, and A. Rygg, "Frequency-domain
modelling and stability analysis of a DFIG-based wind energy conversion Yihua Hu (M’13-SM’15) received the B.S. degree in
system under non-compensated AC grids: impedance modelling effects electrical engineering in 2003, and the Ph.D. degree
and consequences on stability," IET Power Electron., vol. 12, no. 4, pp. in power electronics and drives in 2011, both at China
907-914, Jan. 2019. University of Mining and Technology. Between 2011
[30] A. Rygg, M. Molinas, C. Zhang, and X. Cai, "A Modified and 2013, he was with the College of Electrical
Sequence-Domain Impedance Definition and Its Equivalence to the Engineering, Zhejiang University as a Postdoctoral
dq-Domain Impedance Definition for the Stability Analysis of AC Power Fellow. Between 2013 and 2015, he worked as a
Electronic Systems," IEEE J. Emerg. Sel. Topics Power Electron., vol. 4, Research Associate at the power electronics and
no. 4, pp. 1383-1396, Dec. 2016. motor drive group, the University of Strathclyde.
[31] Y. Xu, H. Nian, T. Wang, L. Chen, and T. Zheng, "Frequency Coupling Between 2016 and 2019, he was a Lecturer at the
Characteristic Modeling and Stability Analysis of Doubly Fed Induction Department of Electrical Engineering and Electronics,
Generator," IEEE Trans. Energy Convers., vol. 33, no. 3, pp. 1475-1486, University of Liverpool (UoL). Currently, he is a Reader at Electronics
Sep. 2018. Engineering Department at The University of York (UoY). His research
[32] H. Nian, B. Hu, C. Wu, L. Chen, Y. Xu, and F. Blaabjerg, "Analysis and interests include renewable generation, power electronics converters & control,
Reshaping on Impedance Characteristic of DFIG System based on electric vehicle, more electric ship/aircraft, smart energy system and
Symmetrical PLL," IEEE Trans. Power Electron., pp. 1-1, 2020. non-destructive test technology. He is the associate editor of IEEE Transactions
[33] K. Ni, Y. Hu, R. Liang, H. Wen, and M. Alkahtani, "Internal Voltage on Industrial Electronics, IET Renewable Power Generation, IET Intelligent
Phase-Amplitude Dynamic Analysis With Interface Friendly Transport Systems and Power Electronics and Drives.
Back-To-Back Power Converter Average Model for Less Power
Electronics-Based More-Electric Ship," IEEE Access, vol. 7, pp. Ronghai Qu (S’01–M’02–SM’05–F’18) was born in
93339-93351, Jul. 2019. China. He received the B.E.E. and M.S.E.E. degrees
[34] K. Ni, Y. Hu, C. Gan, C. Gong, and H. Wen, "Synthetic Internal Voltage in electrical engineering from Tsinghua University,
Phase–Amplitude Dynamics Investigation for Electric Drivetrain Beijing, China, in 1993 and 1996, respectively, and
Small-Signal Model in Electromechanical Control Timescale for a Wound the Ph.D. degree in electrical engineering from the
Rotor Induction Machine-Based Shipboard Power System," IEEE Trans. University of WisconsinMadison, in 2002. In 1998,
Transport. Electrific., vol. 6, no. 2, pp. 844-855, June 2020. he joined the Wisconsin Electric Machines and Power
[35] P. C. Krause, O. Wasynczuk, T. C. Oconnell, and M. Hasan, "Tesla's Electronics Consortiums as Research Assistant. He
Contribution to Electric Machine Analysis," IEEE Trans. Energy Convers., became a Senior Electrical Engineer with Northland,
vol. 32, no. 2, pp. 591-598, Jun. 2017. a Scott Fetzer Company in Watertown, NY, USA, in
[36] G. Abad, J. s. Lo´pez, M. A. Rodrı´guez, L. Marroyo, and G. Iwanski, 2002. Since 2003, he had been with the General
Doubly Fed Induction Machine: Modeling and Control for Wind Energy Electric (GE) Global Research Center, Niskayuna, NY, as a Senior Electrical
Generation. Hoboken, NJ, USA: Wiley, 2011. Engineer with the Electrical Machines and Drives Laboratory. He has authored
[37] D. Zhi and L. Xu, "Direct Power Control of DFIG With Constant over 50 published technical papers and is the holder of over 40 patents/patent
Switching Frequency and Improved Transient Performance," IEEE Trans. applications. From 2010, he has been a professor with Huazhong University of
Energy Convers., vol. 22, no. 1, pp. 110-118, Mar. 2007. Science &Technology, Wuhan, China. Prof. Qu is a full member of Sigma Xi.

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 15,2021 at 06:13:06 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TTE.2020.3044884, IEEE
Transactions on Transportation Electrification
Ni et. al.: Input-Output Small-Signal Stability Analysis of A PLL-Free Direct Power Controlled Partially Power Decoupled 15
More-Electric Shipboard Propulsion System
He has been the recipient of several awards from GE Global Research Center
since 2003, including the Technical Achievement and Management Awards.
He also is the recipient of the 2003 and 2005 Best Paper Awards, third prize,
from the Electric Machines Committee of the IEEE Industry Applications
Society at the 2002 and2004 IAS Annual Meeting, respectively.

2332-7782 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on May 15,2021 at 06:13:06 UTC from IEEE Xplore. Restrictions apply.