Voltage dips in ship systems
J. Prousalidis, E. Styvaktakis, E. Sofras, I.K. Hatzilau, D. Muthumuni
Abstract: In this paper, voltage dips (an important aspect of
power quality) are presented in the context of ship power systems.
Voltage dips are analyzed with respect to their origin (fault, motor
starting, transformer saturation) considering the characteristics of
the on-board installations. Emphasis is given to voltage dips and
power quality issues of the All Electric Ship. Related issues are
presented and discussed.
Keywords: power quality, voltage dips(sags), ship power
systems.
I. INTRODUCTION
Ship Electric power plants similarly to continental grids
suffer from several power quality problems. The
particularly complicated structure of shipboard installations
comprising DC and AC subsystems of several operation
frequency and voltage levels is expected to worsen even
further with the advent of All Electric Ship (AES) buildings
(referring to full electric propulsion and extended
electrification of all shipboard installations).
Power quality phenomena in this context have already been
studied by the authors in previous papers [1-2]. The aim of
this paper is to identify and study the different types of
voltage dips (sags) that occur in ship systems considering
the particularities of these systems. In this way,
identification of the characteristics of different types of
voltage dips in ship system for planning and operating
purposes is done while situations that could cause problems
in ship operation are further investigated. Moreover, the
extraction of voltage dip features is highlighted for electric
system monitoring purposes. Protection issues relative to
voltage dips and shipboard standards for voltage quality
are discussed.
_________________________________________
The work of this paper was supported in part by the research project
"Pythagoras-II" within the "Operational Programme for Education and
Initial Vocational Training - EPEAEK-II"-frame.
J. Prousalidis is an Assistant Professor at the School of Naval and Marine
Engineering of National Technical University of Athens, 9 Heroon
Politechniou St, 15773 Athens, Greece (jprousal@naval.ntua.gr).
E. Styvaktakis, PhD, is with the Hellenic Transmission System Operator
HTSO SA, Amfitheas 11, N. Smirni, 17122, Athens, Greece
(estyvaktakis@desmie.gr).
E. Sofras is a PhD cantidate at the School of Naval and Marine
Engineering of National Technical University of Athens.
I. K. Hatzilau is professor with the Department of Electrical Engineering
and Computer Science in the Hellenic Navy Academy, Peiraeus, Greece.
D. Muthumuni, PhD is with the Manitoba HVDC Center, Manitoba,
Canada.
1-4244-0947-0/07 $25.00 © 2007 IEEE. 309
Simulations in PSCAD environment (the computer
program of Manitoba HVDC Research Center) are used for
analysis purposes.
II. BACKGROUND
A. Voltage dips
Voltage dips (a temporary reduction in rms voltage) are
responsible for the tripping of computers, electronic
equipment and process control equipment. The behavior of
these types of equipment can be a reset, incorrect operation
or shutdown depending on the design and the inherent
protection system of the load. For adjustable speed drives,
depending on the load of the drive, the reduction in speed or
torque might not be tolerated, by the process driven by the
drive. Problems are also caused to the drive controller or the
PWM inverter. Induction motors might not also manage to
reaccelerate after the reduction in speed that is caused by
the voltage dip [3]. Contactors are also sensitive to dips.
Voltage dips are characterized by their rms magnitude and
duration. Duration is the time that the rms voltage remains
below a certain threshold and dip magnitude is the
remaining rms voltage considering the phase with the
maximum voltage. A typical threshold is 10% of the
pre-event voltage.
B. Ship Power Systems - The All Electric Ship
Ship power systems require high levels of reliability as their
operation is closely linked to human safety [4]. The
increased use of electric propulsion leads to an increased
installed generating capacity, the use of different voltage
levels and, in general, to more complicated systems.
Considering that the load demand is typically comparable
to the generation capacity and that ship systems are isolated
it becomes obvious that, compared to the interconnected
terrestrial power systems, these systems present new
challenges.
The eventual complete electrification of all systems
onboard according to All Electric Ship (AES) concept is
expected to create a need for a closer look on the power
quality aspects of these systems. Thus, considering the
great variety of electric loads to be installed onboard in
conjunction with their huge energy demands leads up to the
conclusion that almost all AES- evolution trends would be
inherently correlated to several adverse power quality
phenomena. Considering voltage dips with respect to ship
power systems of interest is on:
1. Large motors and the voltage dips related to them that
may effect other loads in the system
2. System design and protection issues.
3. Motor drive converters applied to main and auxiliary
propulsion systems and their voltage dip ride through
capability.

III. FAULT-INDUCED VOLTAGE DIPS

Faults (due to the short circuit current) lead to voltage dips
of a magnitude that depends on the impedance of the source
(system’s strength), system configuration, fault impedance,
fault type and distance to the fault. Fault-induced events
present the most severe characteristics. Their duration
depends mainly on the protection system operation. That
varies from half-cycle (fuse operation) to several cycles Fig. 1. Simulated Ship Power System
(operation of circuit breakers).
Fig. 2 shows the voltage dip in the generators terminals
Advanced classification methods have been proposed that caused by a 3-phase fault in the 4 kV voltage level. It can be
consider all three phases motivated by the fact that the seen that voltage drops to approximately 0.4 pu (identical
relation between the phases is important for the equipment for all three phases) although voltage in the 4 kV system
performance [3] or by the fact that valuable information can becomes zero at the fault point. The influence of the motor
be extracted considering all three phases [5]. load can be seen both during the fault and after fault
clearing. Fault clearing causes a fast voltage increase to
The calculation of the dip magnitude for a fault somewhere 0.80 pu and then the voltage increases gradually towards
within a radial distribution system requires the point of the normal voltage due to the motor load influence.
common coupling (PCC) between the fault and the load to
1.2
be found. The dip magnitude (%) at the load position equals
the voltage at the PCC (neglecting all load currents): 1.1

1
ZF (1)
Vdip= E 0.9
Z F +Z S
Voltage rms (pu)

0.8
where ZS is the source impedance at the PCC, ZF is the 0.7
impedance between the PCC and the fault (including any
fault impedance), while E the corresponding pre-fault 0.6

voltage. 0.5

Motors that experience a voltage drop will temporary
operate as generators supporting voltage. This shows up in
the voltage recording as a slow decay in voltage magnitude,
during the fault. After fault clearing, motors re-accelerate
delaying the full recovery of voltage and creating a
post-fault dip as described in [3] that could cause motor
stalling.
A. Voltage dips and propagation
The electrical power system of Fig. 1 was simulated in
order to investigate the characteristics of fault-induced
voltage dips in naval systems. This system consists of four
generators and three different voltage levels (6 kV, 4 kV
and 0.38 kV). Induction motor load is a significant part of
the available generation.
Faults in the lower voltage level effect mainly loads fed by
the same busbar as the resulting voltage dips do not
propagate upwards in the system (the impedance of the
transformer is large compared to the source impedance in
medium voltage).
0.4
0.3
0.2
0 0.5 1 1.5 2 2.5 3 3.5 4
Time (sec)
Fig. 2. Voltage dip due to a 3-phase fault
Voltage dips propagate in the system and their
characteristics change as they propagate through
transformers. Fig. 3 shows the voltage dip in the generators
terminals caused by a 2-phase fault in the 4 kV voltage
level. The delta-star transformer that connects the two
voltage levels transforms the relationship of the voltage
drop in the three phases. A characterization method has
been proposed in [3] for voltage dips due to faults that takes
into account the different transformer connections. An
extensive analysis of voltage dips in naval systems, based
on this classification, has been presented by one of authors
in [6].
Considering Eq.1, it should be mentioned that the
magnitude of the drop depends on the available source
impedance.

310
 corresponding voltage dip depend on the induction motor
1.2
data (size, starting method, load, etc) and the strength of the
1.1 system at the point the motor is connected. The magnitude
1 of the dip depends strongly on the system parameters.
The
0.9
duration of the voltage dip due to motor starting depends on
a number of motor parameters with the most important
Voltage rms (pu)

0.8
being the motor inertia [3]. The duration of the dip is
0.7 prolonged if other motor loads are connected to the same
0.6 busbar, as they keep the voltage further down.
0.5

0.4
Fig. 5 shows a voltage dip caused by the starting of a
medium voltage auxiliary propulsion motor of 1.4 MW by
0.3
an 8 MVA generator. It can be seen that voltage drops to
0.2
0 0.5 1 1.5 2 2.5 3 3.5
0.85 pu and gradually recovers approximately 1 sec after.
Time (sec) The voltage dip is the same for all three phases.
Fig. 3. Voltage dip due to a 2-phase fault
Large power motors starting consecutively during ship
Therefore, the number of generators in operation effects maneuvering (e.g. driving thruster systems) can cause such
significantly the magnitude of the voltage dip experienced severe voltage dips.
by the loads.
Finally, single-phase faults consist a specific case of
voltage dips in most ship power systems due to their 1

ungrounded nature. In case of a faulted phase, the load is

still supplied via the healthy ones –which is the main reason
Voltage rms (pu)

0.95
of installing the ungrounded system onboard- but at a
significantly higher voltage level for both phases, namely
√3 or 1.73 p.u., see Fig. 4. Therefore, the system suffers 0.9
from a voltage swell rather than a dip, which stresses
equipment insulation.
0.85

1.8 0.8
0 0.5 1 1.5 2
1.6 Time (sec)
1.4 Fig. 5. Voltage dip due to motor starting
Voltage rms (pu)

1.2 Summarizing, voltage dips due to induction motor starting

1 are:
0.8 • non-rectangular: voltage recovers gradually.
0.6
• symmetrical: all phases present the same behavior.
0.4
V. TRANSFORMER SATURATION DIPS
0.2

0 When a transformer is energised under a no-load condition,

0 0.5 1 1.5 2 2.5 3 3.5 4
Time (sec)
the magnetizing current necessary to maintain the magnetic
flux in the core is in general only few percent of the
Fig. 4. Voltage dip (swell) due to a 1-phase fault nominal rated load current.
Summarizing, voltage dips due to faults are:
• rectangular: voltage recovers fast after fault clearing During transformer energising, a transient occurs to change
operation. the flux in the core to the new steady state condition. In
• symmetrical or asymmetrical: depending on the type of general this will cause the flux to go above the saturation
fault that caused them. value once each cycle until the average value of the flux
over a cycle has decayed to nearly zero. This temporary
over-fluxing of the transformer core causes high values of
IV. INDUCTION MOTOR DIPS the magnetizing current, which is highly asymmetrical and
During start-up of an induction motor takes current five to decays exponentially. This phenomenon is known as
six times larger than normal. This current remains high magnetising inrush current and its magnitude depends on
until the motor reaches its nominal speed. This lasts from the point on the wave where the energization switching
several seconds to one minute. The characteristics of the takes place and the core residual flux. As the core is forced

311
into saturation the transformer draws a large current from
50
the supplying network. When the voltage reverses its
polarity in the next half cycle, the maximum flux in the core 45

is less than the maximum flux density in the no-load

Harmonics (% of the fundamental)

40 second

situation. The transformer inrush current is therefore 35

asymmetrical and contains a DC component, which might 30
take seconds to disappear depending on the damping of the
25
system [5]. The voltage dip caused by the magnetising
third fourth
inrush current can be long in duration and drive other 20

near-by transformers into saturation (sympathetic 15

saturation, [7,8]). In general, any voltage change in the 10

transformer terminals (like a voltage dip) could lead it into 5 fifth

saturation due to the resulting transient in the core. 0

sixth

0 50 100 150 200 250 300

Time (msec)
Fig. 6 presents a transformer saturation voltage dip caused
by the energizing of a transformer in the network of Fig. 1. Fig. 7. Voltage harmonics in time due to transformer saturation
A sharp voltage drop (approximately 0.2 pu for the worst
phase) is followed by a gradual recovery. As can be seen in Considering that the use of harmonic filters is extended in
Fig. 7, the voltage presents temporary harmonic distortion. ship systems due to the multiple use of power converters
The Short Time Fourier Transform has been used for the this example shows that these filters must be disconnected
estimation of the harmonics (from 2nd to 5th) of the voltage during transformer energizing, a practice followed in many
of one of the phases of Fig. 6. The 2nd harmonic is industrial power systems.
contributing the most. This increased harmonic distortion
can cause undesired protection operation. 1
Voltage (pu)
0.5
Summarising, voltage dips due to transformer saturation 0 (a)
are: -0.5
• non-rectangular: voltage recovers gradually as the inrush -1
current decays. 0 50 100 150 200 250 300
Time (msec)
• non-symmetrical: each phase presents a different degree
Voltage magnitude (pu)

of saturation. 1
• rich in harmonics: due to the asymmetry of the inrush
current. 0.8
(b)

1
0.6
0 50 100 150 200 250 300
Voltage (pu)

0.5
Time (msec)
0
-0.5 Fig. 8. (up): Voltage waveforms during transformer energizing, (down):
corresponding voltage rms magnitude
-1
0 50 100 150 200 250 300
Time (msec)
VI. SHIPBOARD STANDARDS
Voltage magnitude (pu)

1.1

1 Table 1 shows the limits proposed by several classification

0.9
societies and international associations regarding the limits
of voltage transients. Voltage dips fall in this category.
0.8

0.7 TABLE I. Classification Society Rules & Standards ([8-11])

0 50 100 150 200 250 300
Time (msec)
REGISTER VOLTAGE TRANSIENT
ABS (2005) BV(2003)
Fig. 6. (up): Voltage waveforms during transformer energizing, (down): DNV(2001) GL(2004) ±20% (1,5s)
corresponding voltage rms magnitude PRS (2002) RINA(2005)
(LRS) (2001) +20%, -15% (1,5s)
The harmonic distortion caused by transformer saturation
IEEE Std 45-1998 ±16%(2s)
can lead to more severe dips if any system resonance is
STANAG 1008 (Ed.8, Ed.9) ±16%(2s) [ ±22%(2s) ]*
excited. Fig. 8 shows the voltage dip occurring when the USA MIL-Std-1399 ( like ST.1008 )
same transformer as above is energized through a busbar * under some circumstances
where a 3rd harmonic filter (single-tuned filter) is connected.
The 3rd harmonic filter causes a parallel resonance in a IEEE Std 45-1998 [9], additional to the values of Table 1,
frequency slightly lower than the 3rd harmonic and the provides guidelines regarding motor starting and limits for
harmonics of the transformer inrush current are amplified the resulting voltage dips. The lower allowed voltage dip is
leading to a more severe dip (0.1 pu lower than before). 30% (70% remaining voltage) in the case of a group of

312
motors that automatically restart upon closing a feeder automatic transfer switches, ferroresonance transformers,
breaker. Special attention is given to the generator sizing series connected voltage source converters, etc [3,14,15].
when large motors are connected to the system.
600
Load immunity is an issue that, at the moment, is not
S pik e T ra ns ien t covered extensively by standards and regulations. As the
500 use of sensitive equipment increases in naval systems, this
S p ike & T ra n sien t
Bo un d ary area a cc ord in g issue will become more important. For power converters
250 to S T AN AG 1 00 8 E d .9
that are used for propulsion purposes, the voltage ride
through capability is, up to a certain extend, cost related. By
adding capacitance (in low power electronics and AC
200
(%) change in nominal voltage

P ro h ib ited area
150
drives) and with appropriate control algorithms (for DC
drives) the sensitivity of the equipment to voltage dips is
100 reduced [3].

50
VII. VOLTAGE DIP RELATED ISSUES
+-- 10%

0
N OM INA L
VO LTAG E
As mentioned above, the duration of fault -induced dips is
mainly depended on the response of the protection system
-50

N o d am ag e a rea
(fault clearing) which, in the event of a fault, should isolate
-100
the part of the system where the fault is. Protection of ship
0.0001 0.001 0.01 0.1 1 10 100
Steady power systems is typically based on overcurrent elements
Tim e (sec) state

(relays or fuses). However, the protection coordination in

Fig. 9. ITIC compared to STANAG 1008 (dashed these systems is not an easy task. As the length of the cables
line) (applicable to 120 V equipment - 60 Hz) is typically short, the different parts of the system, in terms
of impedance, are separated mainly by the transformers.
STANAG 1008 [10] (a standard referring to the electrical
This makes selectivity difficult and leads to unnecessary
power plants in NATO naval vessels) states that:«The
protection operations and load interruptions.
definitions, user information and characteristics in this
agreement refer essentially to a healthy electrical power
The variations in source impedance is also a parameter that
supply system and include transient conditions that result
naval protection systems must take into account because
from normal system operations, such as motor starts and
these variations can lead to incorrect protection operations.
switching events. The STANAG does not describe system
At this point it should be also mentioned that different
behavior under abnormal conditions, e.g. short circuit
grounding practices are used in ship and this certainly
faults, loss of generator sets or malfunction of associated
influences protection operation. For ungrounded systems,
controls, as this behaviour depends heavily on the design
for example, detection of single phase faults can be a
characteristics of individual systems». Consequently,
difficult task.
STANAG 1008 does not include limits and guidelines with
respect to abnormal conditions for the design of naval
Faster protection operation and efficient protection
systems although these conditions produce transients and
coordination will reduce the duration of fault-induced
spikes are even more severe than the ones in normal system
voltage dips. Additionally, it is important to highlight that
operations.
by improving the protection of the ship systems the risk of
Comparing the voltage limit curve provided by STANAG
unwanted protection is minimized in the case of the other
1008 with the ITIC [13] (a curve often applied for assessing two types of events (motor and transformer related).
the performance of power systems in terms of power
quality) one can see (Fig. 9) that STANAG does not Sophisticated protection methods are available that can
accommodate dips of short duration and large magnitude accommodate the special characteristics of the naval
caused mainly by faults (up to 2 seconds the curve equals to system. For example, differential protection (unit
80% of nominal voltage). For shipboard standards, it is protection) improves significantly protection performance
important to take into account abnormal conditions: by discriminating faults by other types of disturbances. This
type of protection is used for transformers in terrestrial
(a) in the design stage in order to minimize their impact by systems. Furthermore, adaptive protection is able to take
making appropriate decisions from a system point of view into account the changes in the system (like the variations
(b) in coordinating load immunity and mitigation devices in source impedance) and adjust relay settings to the new
(c) in designing appropriate protection devices and conditions offering better performance [16-17]. In terms of
equipment with higher immunity to power quality design and operating practices, the increase in the size of
parameters. generating capacity of ship systems rises the issue of short
circuit current. As more generation is installed on-board the
Mitigation devices for voltage dips are ups, flywheel, equipment has to cope with higher currents in the case of a
fault. To avoid the cost of higher rating equipment it is

313
common practice to split the network into different parts.
Although this practice reduces the short circuit current, it
influences significantly the voltage quality of the network,
not only in terms of dips but also in terms of harmonics and
other power quality aspects. Stiffer systems (higher short
circuit current) are less sensitive to power quality
phenomena and in many cases the extra cost of higher rated
equipment is justified by the benefits in power quality.
Regarding motor dips, a number of technical solutions for
staring are available. They prevent severe voltage dips and
allow motor starting without problems, by limiting the
inrush current. These measures are [18]:
1. Starting-up via autotransformer
2. Starting-up via autotransformer and capacitor
3. Soft starting via power electronic devices
For transformer saturation events, as mentioned above,
special attention must be given to their interaction with
harmonic filters to avoid severe dips, increased harmonic
distortion, and protection maloperation.
VIII. CONCLUSIONS
In this paper an important aspect of power quality, voltage
dips, is presented with respect to on-board installations.
Voltage dips are categorized in terms of their origin into
three classes: fault-induced, motor starting and transformer
saturation. Voltage dips due to faults can be severe
depending on the voltage level where the fault takes place.
The influence of induction motor load on the voltage dip is
important as it can prolong its duration. The duration of
this type of dips depends on the protection system response
to the fault. Motor starting can be of considerable
magnitude depending on the motor, its load, other motors
near-by, etc. Transformer saturation dips are due to the
inrush current upon energizing. There are associated with
increased harmonic distortion that could lead to more
severe voltage dips in the presence of single-tuned
harmonic filters. Therefore, transformer energizing should
be avoided when harmonic filters are connected.
In terms of characterization, voltage dips due to faults
present an almost rectangular shape in their rms and can be
either symmetrical or asymmetrical depending on the type
of fault and the transformer connections between the fault
and the point where the dip is measured. Induction motor
dips are symmetrical and non-rectangular as the three
phases behave the same and voltage recovers back to
normal gradually. Transformer saturation dips are
non-rectangular and non-symmetrical as the voltage dip of
each phase depends on the point-on-the-wave where the
switching takes place.
A presentation of naval standards on voltage dips is also
given. Special attention is paid to STANAG 1008, while it
is highlighted that standards must take into account power
quality issues towards the complete electrification of ships.
Finally, considering the characteristics of ship power
314
systems, protection is probably the most critical issue as it
influences the dip duration, and up to a great extent, the
ability of loads to operate without problems during faults.
Faster protection schemes with efficient algorithms can
reduce the dip duration (by reducing the fault clearing time)
and minimize the risk of incorrect operation.
IX. ACKNOWLEDGEMENTS
The work of this paper is part of the research project "Pythagoras-II"
within the "Operational Programme for Education and Initial Vocational
Training - EPEAEK-II"-frame. The Project is co-funded by the European
Social Fund (75%) and Greek National Resources (25%).
X. REFERENCES
[1] I.K. Hatzilau, J. Prousalidis, E. Styvaktakis, F. Kanellos, S. Perros, E.
Sofras, ”Electric power supply quality concepts for the All Electric Ship
(AES)”, 2006 World Marine Transport Technology Conference, London
(UK), March 2006.
[2] I.K. Hatzilau, J. Prousalidis, E. Styvaktakis, E. Sofras “Voltage and
Current Spikes & Transients - Power Supply Quality aspects for the AES”,
Proceedings of 2005 All Electric Ship Symposium (AES2005), Paris
(France), October 2005.
[3] M.H.J. Bollen, “Understanding power quality problems: voltage sags
and interruptions”, IEEE Press, N.Y., 1999.
[4] J.V. Amy, “Considerations in the design of naval electric power
systems,” Proceedings of the Power Engineering Society Summer Meeting,
vol. 1, pp 331-335, 2002.
[5] E. Styvaktakis, M.H.J. Bollen, “Signatures of voltage dips:
transformer saturation and multistage dips”, IEEE Trans. on Power
Delivery, vol.18, no.1, pp. 265-270, January 2003.
[6] S. Spagkouros, J. Prousalidis, " Electric Power Quality problems in
ship systems: A classification method ", IASME TRANSACTIONS, Issue
2, Volume 1, April 2004, pp. 337-342.
[7] H.S. Bronzeado, P.B. Brogan, and R. Yacamini, “Harmonic analysis of
transient currents during sympathetic interaction,” IEEE Transactions on
PowerSystems, vol. 11, no. 4, pp. 2051–2056, November 1996.
[8] J.M. Prousalidis, A.G. Georgopoulos, N.D. Hatziargyriou, B.C. Papadias,
"Investigation of Transformer Sympathetic Inrush", Proceedings of 2nd
International Conference on Electromagnetic Transients, IPST ’97, Seattle
(USA), July 1997, pp. 197-202.
[9] IEEE Standard 45-1998, “IEEE Recommended Practice for Electrical
Installations on Shipboard”.
[10] STANAG 1008, ”Characteristics of Shipboard Electrical Power
Systems in Warships of the North Atlantic Treaty.
[11] USA MIL-STD-1399(NAVY), “Interface standard for Shipboard
systems – Section 300A – Electric Power, Alternating Current”.
[12] IEC-60092/101, “Electrical installation in Ships - Definitions and
General Requirements”, 1994.
[13] TC3 of the Information Technology Industry Council (ITIC) , “ITIC
(CBEMA) Curve”, 2000
[14] S. Samineni, B. Johnson, H. Hess, and J. Law, “Modeling and
Analysis of a Flywheel Energy Storage System with a Power Converter
Interface,” presented at the International Conference on Power Systems
Transients (IPST), New Orleans, USA, 2003.
[15] M. Moschakis, J. Prousalidis, N. Hatziargyriou," Performance
Assessment of STC used for Alternative Naval Power Supplying Units ",
IASME TRANSACTIONS, Issue 2, Volume 1, April 2004, pp.394-399.
[16] D. Whitehead and N. Fischer, “Advanced Commercial Power System
Protection Practices Applied to Naval Medium Voltage Power System,”
Electric Ship Technologies Symposium, Philadelpia, PA, July 2005.
[17] K.L. Butler, N.D.R. Sarma, C. Whitcomb, H. Do Carmo, and H.
Zhang, “Shipboard systems deploy automated protection,” IEEE
Computer Applications in Power, vol. 11, no. 2, pp. 31-36, April 1998.
[18] P. Vallianatos, J. Prousalidis, E. Styvaktakis, ”On starting-up large
power motors rotating high inertia loads in autonomous systems”,
Proceedings of International Conference on Electric Machines
(ICEM-2006), Chania (Crete), September 2006.