Electric quality concepts for the All Electric Ship (AES)

I.K. Hatzilau, Prof. Dr.-Ing., Hellenic Naval Academy (HNA), GREECE J. Prousalidis, Ass. Prof., National Technical University of Athens (NTUA), GREECE E. Styvaktakis, PhD, Hellenic Transmission System Operator (HTSO), GREECE F. Kanellos, PhD , GREECE S. Perros, MSc, Cdr Hellenic Navy (HN), GREECE E. Sofras, PhD Cand., National Technical University of Athens (NTUA), GREECE

SYNOPSIS

Power Supply Quality (PSQ) has become an important concern for ship electrical systems. Deviations of voltage from nominal values and wave-shapes could cause several malfunctions, interruption or even damages of onboard equipment and systems. In contrast to the All Electric Ship perspective domination, limited research work has been reported (even for more conventional ship types) on the investigation and classification of PSQ phenomena. The paper aims : To make a short but coherent description and categorisation of the PSQ phenomena (causes, consequences, characteristic parameters etc). To investigate, discuss and compare the relevant issues covered by different standards (international and national, military and commercial) applied in shipboard applications, taking also into account the standard status of similar continental power systems. To lead to conclusions, where possible, on which specific standardized issues appear closer to the current and forthcoming status of ship electric networks and the relevant PSQ problems. Key-Words : Power Supply Quality, Standards, All Electric Ship

INTRODUCTION AND BACKGROUND

Power Supply Quality (PSQ) is a term referring to a wide variety of disturbances in electric networks either ship or continental ones [1], [2], [3], [4]. Thus, PSQ problems refer to deviations of electric quantities from nominal values and wave-shapes, which can cause several malfunctions, interruption or even damages of equipment and systems. Concerning ship systems and in contrast to the All Electric Ship (AES) perspective domination with extensive electrification and significant technological progress in the field of sophisticated power subsystems installed onboard, limited research work has been reported (even for more conventional ship types) on the investigation and classification of PSQ phenomena. On the contrary, most ship-standards have been based on measurements and analyses/approaches performed almost twenty to thirty years ago, while no actual amendments on this topic have been made at least in comparison with the corresponding major changes introduced in standards dedicated exclusively to continental grids.

------Author’s Biography Prof. Dr.-Ing. I.K. Hatzilau (Electrical & Mechanical Engineer from NTUA/1965, Dr.-Ing. from University of Stuttgart/1969). After few years in the industry, he joined the Academic Staff of Dept of and Computer Science in HNA. He is representative of HN in NATO AC/141(NG/6)SG/4 dealing with warship-electric systems. Dr. J. Prousalidis (Electrical Engineer from NTUA/1991, PhD from NTUA/1997). Assistant Professor at the School of Naval and Marine Engineering of National Technical University of Athens, dealing with electric energy systems and electric propulsion schemes on shipboard. Dr. E. Styvaktakis (Electrical Engineer from NTUA/1995, MSc in Electrical Engineering from UMIST/1996, PhD from Chalmers/2002). He is currently with the Hellenic Transmission System Operator dealing with power system studies. Dr. F. Kanellos (Electrical Engineer from NTUA/1998, PhD from NTUA/2003). He currently works as consultant engineer for General Secretary of Research and Technology of Greece. Cdr. S. Perros, (Engineer Officer from the HNA/1984, M.Sc in Electrical Engineering from Naval Postgraduate School in Monterey California, USA/ 1992). After many years of service in HN warships, he is assigned now in HN General Staff. He is also an associate of the Dept of Electrical Engineering and Computer Science in HNA. E. Sofras (Naval Architect and Marine Engineer from NTUA/2004). Currently, he is pursuing postgraduate studies as a PhD candidate at NTUA, dealing with issues of ship electric networks. This paper aims : - To make a short but coherent description and categorisation of the PSQ phenomena (causes, consequences, characteristic parameters etc). - To investigate, discuss and compare the relevant issues covered by different standards (international and national, military and commercial) applied in shipboard applications, taking also into account the standard status of similar continental power systems. - To lead to conclusions, where possible.

The standards mainly studied for the purposes of this paper are : - IEEE 1159 [5] and IEEE 519 [6] on general and harmonic distortion PSQ issues respectively. These standards refer to continental grids, however, as they have recently issued, they have introduced several new aspects on PSQ terminology and problem resolution. - IEEE 45 [7] and IEC-60092/101 [8] mainly dealing with general ship network installations. Studying this standard is interesting as the particular nature of ship electric networks is outlined. - Classification Society Rules [9] mainly referring to normal operation of electric networks of commercial ships. These rules, in general, have not taken into account the significant upcoming changes in the network field and are rather poor in stipulating standardized terminology and limitations. An exception worth mentioning is that of LRS, where in case of naval ships it makes reference to STANAG-1008. - STANAG-1008 [10] referring to the electrical power plants in NATO naval vessels. This is a standard that has come out of USA MIL-STD-1399(NAVY)–Section 300A [11] under the responsibility of the NATO AC/141(NG/6)SubGroup/4. This standard refers actually to normal operation of warships and has introduced a series of PSQ interesting issues. However, despite the strong interest of most navies in AES concept, no significant attention has been so far paid to standardizing critical related aspects such as High Voltage application, operation in fault conditions and electric propulsion system.

From a certain point of view, PSQ phenomena can be classified into two main categories, i.e.: • Steady-state phenomena which have inherent periodicity and they are repeated on almost constant basis without significant changes (e.g. there is no decaying mechanism), such as harmonics and unbalance, while there is also a stochastic parameter included in certain cases like notching, flickering etc. • Transient-state phenomena, which are of limited duration and they are characterized by significant changes (there is definitely a decaying mechanism), such as spikes and transients like dips, swells. It is worth noting that although all these phenomena are treated as Power Quality aspects, they actually refer to the fundamental electric supply quantities, i.e. voltage, frequency and possibly current. The first two consist the main system characteristic quantities that have to be kept intact and well within the limits set by the standards. The current on the contrary, is a quantity affecting power quality in an indirect manner and via voltage drop on system and equipment impedances, in particular. Nevertheless, should the current does not comply with the limitations set by certain standards, it is highly possible that voltage and /or frequency will be influenced, too.

The PSQ problems dealt with in this paper are Harmonic distortion, Spikes / Transients / Dips / Swells, Voltage and Frequency Modulation (due to pulsed loads), Voltage Unbalance, Leakage Capacitive Currents. Additionally, some other PSQ issues (Voltage interruptions, Frequency ) are discussed.

HARMONICS

Harmonic power quality refers to the existence of distorted periodic voltage or current waveforms, which, can be expressed via mathematical Fourier analysis, as the superposition of an infinite series of frequencies, the fundamental one (the so-called power frequency) and its multiples, the high-order harmonics Vn, In :

∞ ∞ v(t) = ∑Vn 2 sin(n.ωt +θ v,n ) i(t) = ∑ I n 2 sin(n.ωt +θ i,n ) (1) n=1 n=1

Harmonic distortion is mainly due to power electronic devices used to couple and control different operating voltage and frequency levels e.g. in shaft generator systems or in electric motor drives. In contradiction to other power quality problems, harmonic distortion is a steady-state phenomenon existing on a constant time basis, which means that the stresses occurred, though not severe, can eventually provoke an accumulative result often not easily explained. The impacts of harmonic voltage and/or current distortion include a wide range of phenomena [1], [2], [6], [12] : • Extra heating losses in electric machinery and cable wiring (leading to either premature aging and de-rating of the equipment due to overheating or to extra cooling requirements). • Decrements in accuracy of measuring equipment, which are not designed for non-sinusoidal electric quantity measurements. • Excitation of resonance phenomena resulting in significant and/or overcurrents. • False tripping of protective switchgear (e.g. fuse blowing, or incorrect thermal relays actuation). • Failure of equipment sensitive to harmonics. • Electromagnetic Interference (EMI) problems with sensitive electronic equipment (navigation, communication, control and automation) • Erection of mechanical oscillations, vibrations, mechanical stresses and noise due to harmonic torque ripples produced.

Harmonic power quality is faced by many standards as voltage quality, as voltage can be directly controlled and regulated by the power system. On the contrary, current is determined by the various loads supplied, therefore, current quality can not be easily controlled or even monitored, despite that current distortion is also reflected as a voltage distortion, via the voltage drop on circuit impedances. Hence, much attention must be paid to current quality defined by the entity of loads installed onboard or at least the major ones, i.e. those of significant power demand with respect to electric system capacity.

Regarding appropriate indices to demonstrate the level of distortion, besides the individual voltage and current harmonics Vn and In, [ see Eq. (1) ], Total Harmonic Distortion (THD) seems to be the most coherent one, expressed as the portion of high order harmonics with reference to the fundamental component :

2 2 ∑Vn ∑ I n (2) n=2 n=2 THDV = THDI = V1 I1

It is worth noting that according to THD definition, both the fundamental component and the high order harmonics should be measured simultaneously. However, should in Eq. (2), the fundamental equals an average quantity over a specific time interval, then the index is called Total Demand Distortion (TDD).

Concerning further harmonic power quality problems, certain harmonics, which are supposed to be non-existent e.g. the “triplens” (integer multiples of 3) on the 6-pulse converters can still have a fairly significant measurable value, while the so-called “interharmonics” and “sub-harmonics”, i.e. non-integer multiples and partial multiples οf the fundamental frequency respectively comprise a sort of distortion not well defined or predicted.

As shown in Fig 1 and 2 and Table I, while most standards provide voltage harmonic distortion limitations only few of them define figurative restrictions on current one.

6,0% 8,0% GL 7,0% 5,0% (odd) STANAG 1008/Ed. 8 (Power <1 kVA) IEC-60092 IEEE-519 6,0% STANAG IEEE-519 4,0% 5,0% 1008 STANAG 1008/Ed. 8 (Power >1 kVA) 1L

/I 4,0% 3,0% n I Vn/V1 LRS 3,0% 2,0% 2,0% IEEE-519 (even) 1,0% 1,0% 0,0% 0,0% 0 102030405060708090100 0,00 20,00 40,00 60,00 80,00 100,00 120,00 Harmonic Order n Harmonic Order n

Fig 1. Graphic representation of individual voltage Fig 2. Graphic representation of current harmonic limits harmonic limits set by various standards set by various standards (only STANAG 1008/Ed. 8 and IEEE-519 have current harmonic limitations. The graph is valid provided the base reference values of the two standards coincide). Table I. Qualitative comparison among standards

STANAG 1008 GL LRS IEEE-519 IEC-60092 Edition 8 Edition 9 THDV Yes (5%) Yes (5%) Yes (8%) Yes (8%) Yes (5%) Yes (5%) Vn Yes (3%) Yes (3%) Yes Yes (1.5%)* Yes (3%) Yes (3%) (see Figure 1)

THDI No ?*** No No Yes (5%)** ? ***

In Yes ? *** No No Yes ? *** (see Figure 2) (see Figure 2) * limitation valid for n≥25 ** TDDI *** non-numerical limitation , but action or no-action depended on Pdist,k and Ssc is recommended.

Regarding mitigation measures towards amelioration of harmonic distortion problem, these comprise either intervention of appropriate filters or the introduction of improved switching techniques of the power electronic converters or even the synthesis of more complicated converter topologies.

Concerning Ed.8 of STANAG 1008 [10], it has been reported in numerous technical essays [13], [14], [15], that its distortion limits of current harmonic distortion set (Fig 3a) are rather strict, as, according to field measurements, they are often violated by several equipment onboard. However, this maynot result, in exceeding the corresponding voltage harmonic distortion levels, which is often kept well below the limits set by almost all standards STANAG 1008, Edition 9 (2004), on the other hand sets no tangible numerical limits but only defines courses of actions to be taken depending on the rated power of harmonic pollutant devices with respect to system short circuit capacity, see Fig 3b. Nevertheless, when a course of action is to be taken, this comprises studies via simulation of the entire ship electric energy system operation, provided of course that all necessary data are available by the equipment manufacturers [16].

∑ Ρ dist,k log(I /I ) n 1,fL Small Loads S sc

< (1kVA/60Hz or Analysis required log(100) 0.2kVA/400Hz and 2% 2A/115V/400Hz)

No action log(3) Big Loads > (1kVA/60Hz or 0.2kVA/400Hz and 1% 2A/115V/400Hz) No action N o acti o n

0.1% 0. 5 log(33) log(100) log(n) Maxk(Pdist,k) S sc

Fig 3a. Ed.8 STANAG 1008 - Guidelines on current Fig 3b. Ed. 9 of STANAG 1008 – Guidelines on harmonics current harmonics In : Individual Harmonic of each separate load Pdist,k : power of devices distorting current I1,fL : Full load fundamental current waveforms Ssc : short circuit power level of the supply system

Moreover, two strongly related harmonic distortion PS-issues with inherent stochastic nature are the “notching” and ”noise” phenomena, to which not much attention is paid by any ship Standard.

More specifically, notching is a periodic voltage disturbance caused by the normal operation of power devices due to commutation [IEEE 1159]. Three-phase converters are the most important source of voltage notching [IEEE 519]. Notching is a steady state phenomenon but unlike the harmonic pollution issues described previously, notching is associated with higher frequencies. This phenomenon can cause problems to power electronic equipment that use zero crossings for frequency or time calculation. Furthermore, noise comprises signals with broadband spectral content lower than 200 kHz superimposed upon the power system voltage or current in phase conductors, or found on neutral conductors or signal lines [IEEE 1159] and it is caused by power electronic devices, control circuits, arcing equipment and switching power supplies. Noise disturbs electronic devices such as microcomputer and programmable controllers. Solutions to noise problems include filters, isolation etc.

Harmonic and notching problems are expected to significantly increase in the AES perspective as their main causes i.e. power electronic devices are to be extensively applied onboard either for AC/DC conversion or for electric motor variable speed drives.

SPIKES, TRANSIENTS, DIPS, SWELLS

Electric networks suffer from “short time duration non-periodical disturbances” of voltage and current caused by switching operations (making or breaking), short-circuits, fuse blowing-up or even strikes [17], [18], [19]. The name of these disturbances varies depending on certain characteristics of them (e.g. transients, spikes, dips, swells etc).

A transient and a spike are a sudden, non-power frequency change in voltage or current, or both. They can be unidirectional or oscillatory in polarity. Lightning and switching actions are typical causes of spikes and transients. For example a “spike” can occurs in the initial phase of the “transient” phenomenon following the energizing of a load ( see Fig 4, [20] ). The energizing drives the network to oscillations at the resonant frequency of the whole system. The evolution of the disturbance (waveshape, peak-value, etc) will be influenced by many parameters such as : load impedance, the elements of the Thevenin equivalent source of the supply network at the position of the load, the time instant inside the 2π voltage period the energisation take place and the characteristics of the switch and the non-linear and partly statistic behavior of the electrical characteristics of the arc.

Fig 4.

Spikes introduced in AC power system during energisation of : (4) : a PC-screen

(8) : a PC-Terminal

Spikes and Transients are characterized considering the actual values of the voltage waveform.

Voltage dips are events that present a temporary decrease in the rms voltage. They are caused by a increase in current somewhere in the system e.g., during motor starting and energizing [21]. A voltage swell is a temporary increase in rms voltage. Swells are usually caused by fault conditions or capacitor inrush. Voltage Dips and Swells are characterized by their rms magnitude and duration.

Voltage and Current spikes and transients have several adverse consequences such as equipment failure and improper operation of the entire system, like : - insulation breakdown, - misfiring of semiconductor switches by the rate of voltage change, leading to device failures or even errors in data processing devices. - False tripping of adjustable speed drives. - Regarding surge protective devices transients in voltage as well as in current are crucial. Failures have been identified that are linked to current and not to energy.

Voltage dips are responsible for the tripping of computers, electronic equipment and process control equipment. Like transients, swells can be harmful to electrical insulation [3].

In IEEE1159 [5], power electromagnetic phenomena are classified into 7 groups. Groups 1 and 2 are transients and short duration variations (impulsive transient, oscillatory transient, voltage dips and voltage swell). Each group is subdivided further into classes considering duration, rise time, spectral content, amplitude etc.

According to to Editions 7, 8 and 9 of STANAG 1008 [10], the definitions and the limits for voltage transient and spikes can be summarized as follows (in parentheses the Ed.9 changes), also seen in Fig 5 : Voltage transient is a sudden change in voltage (Ed.9: in the peak amplitude of the voltage), which exceeds the user voltage tolerance limits for a time longer than 1 msec and less than 2 sec (Ed.9: the typical time duration is between a fraction of a cycle and 2 sec). Limits : AC systems +/-16% (under circumstances +/-22%). 24V DC systems 18-35V. Voltage spike is a voltage transient of short (less than 1 msec) duration (Ed.9: A voltage spike is usually less than a quarter cycle in duration). Limits : AC systems 2.5kV (440V supply) or 1kV (115V supply). 24V DC systems 600V. The terminology and the “philosophy” of STANAG 1008 in not always consistent with the standards that are widely used in continental systems [22], [23]. These differences impose difficulties in directly applying IEEE documents into naval electrical systems which would be very favourable considering that nowadays onboard conventional COTS technology used for continental grids is mainly installed for economical reasons. This COTS exploitation is one of the key perspective aspects of AES technology.

Nevertheless, some of the most problematic issues of STANAG 1008 are the following :

- Spike peak limits have not been measured since at least 25 years ago, i.e. when neither equipment electrification nor protective surge suppressors have been extensively introduced [24], [25].

- Spike/Transient amplitude(DV)-duration(Dt) diagram. STANAG 1008 Ed.8 distinguishes abruptly – and in contradiction to the real natural phenomena - between “spike” and “transient” in the amplitude(DV)-duration(Dt) diagram. Other standards ( e.g. ITIC [26] ) use an intermediate “sloped” section of the curve, which separates the permissible and not-permissible areas in the amplitude-duration diagram. Additionally, in Ed.9 of STANAG 1008 the decisive characteristics for distinguishing transients from spikes are not clear and this may lead to confusion Fig 6.

- Spike waveshape, Test-impulse Measurements and theoretical analysis show, that most voltage spikes in small-size low-voltage systems (as ship systems) have oscillatory waveshapes, unlike the unidirectional waves usually specified in high-voltage insulation standards. In STANAG 1008 only the "Lightning-Impulse"-IEC 60060-1 is recommended for test purposes. However, other standards include also the "Ring Wave"-0.5µsec-100kHz (oscillatory waveshape). Additionally the lightning impulse spike test has characteristics, that are based on the concept of energy (100 joules on 2 ohm load). But it has been shown ( [27], [28] ) that this is a misleading oversimplification. E.g. the energy, that a metal-oxide must handle, depends also on amplitude, waveform, source impedance and varistor characteristics and NOT only on the energy calculated by the voltage waveform. Additionally there are cases, where the failure mechanism is totally independent from energy. Current surges have also a critical effect on the failure mode and they are also system depended [29].

The AES concept is expected to derive complicated transient phenomena and spikes to a significant extent. For instance, dip problem will be magnified especially in the case of starting of big propulsion motors or Pulsed loads etc. Motor drive converters can provoke (besides harmonics) switching transients, mainly due to problems during commutation of power electronic switches. Concerning the initial peak value of spikes, as widely known from experience on continental grid, the stray capacitances of all components onboard would play an important role. The introduction of High Voltage (levels > 1000 V) increases the requirements for insulation in either steady- or transient- state. The probable integration of HVDC power supply could introduce several switching transients overvoltages due to the circuit breaking of a DC current [30]. Zonal concept and ring type networks, designed to increase power supply reliability, could magnify Spikes & Transients problems, if the system protection scheme is not well set ( e.g. selectivity etc ). Economic operation of the AES high powered electric system requires dynamic reconfiguration on a real-time basis, which actually means a significant number of switch change-overs with the subsequent switching transients.

600 Spike Transient

500

STANAG 1008 250

e 200 Problematic Spike & ag

lt Transient Overlap area in

vo Ed. 9 of STANAG 1008 l 150 na i m

o ITIC

n 100 i curve Prohibited area ge

an 50 h

c +-- 10%

(%) 0

-50 No damage area -100 0.000 0.001 0.01 0.1 1 10 10 Steady Time (sec) state

Fig 5. Transients conditions Fig 6. ITIC compared to STANAG 1008 (dashed according STANAG 1008, Ed.8 . line) (applicable to 120 V equipment – 60 Hz) (a) acceptable spike, (b) acceptable transients, (c) unacceptable spike, (d) unacceptable transient. t(1,2) and t(5,6) < 1msec. t(3,4) and t(7,8) > 1msec.

MODULATION / PULSED LOADS

Voltage and frequency periodic or quasi-periodic variations such as might be caused by regularly or randomly repeated loading with frequency less than nominal are refered in shipboard Standards (STANAG 1008, IEEE 45) as “modulation”. In Fig 7, voltage and frequency modulation is shown. For quantifying voltage or frequency modulation the difference between maximum and minimum value is used as a percentage of the double of the nominal value. 2% and 0,5% limits are proposed for voltage and frequency, respectively. In continental systems this low value systematic fluctuations of voltage are measured by the flickermeter and the phenomenon is referred to as flicker due to the effect that it has on lighting. Flicker is quantified by the use of long-term and short-term indexes, Plt and Pst, respectively (IEEE 1453 [31]). However, modulation/flicker is not considered for frequency.

Fig 7 Voltage and Frequency modulation

The future electric power system of naval warships in the context of the AES will supply energy to sophisticated systems for propulsion, electric guns, electric launchers, high power sensors, navigation etc. [32], [33], [34], [35]. Some of new naval weapon systems, referred as “pulsed loads”, require high power (in order of several GWs) for a very short interval in order of a few seconds and even milliseconds are the main cause of voltage/frequency modulation. The operating characteristics of the aforementioned emerging electrical loads will pose many new challenges for these power systems as these lead to voltage and frequency modulation in an Integrated Power System (IPS). Examples of pulsed loads include : • Electromagnetic Aircraft Launch Systems (EMALS) • Electromagnetic Guns ( Rail Guns, Coil Guns, Lasers, High Energy Microwaves ) • Radars, sonars, communication systems etc.

Experience from power systems indicates that the primary effects of pulse loading are manifested in the areas of [36] : • Voltage flicker • Transient stability (the ability of synchronous machines to remain in synchronism after a large disturbance) • Dynamic stability or long-term stability (the ability of a system to remain stable when generator swings are primarily dictated by the response of their control systems) • Excitation of torsional frequencies in generators

The aforementioned may affect the operation of several subsystems of a ship such as radarscopes, communication equipment, missile guidance systems, weapon systems, gear systems etc [33].

According to STANAG 1008 pulsed loads should not exceed the limits specified by the following equations :

Qpulse < 0.065*Ssupply and Ppulse < 0.25*Ssupply [3]

(P, Qpulse = active, reactive power of the pulsed load, Ssupply = full rated apparent power of the supply at the occurrence of the pulse).

Probably, the most effective way to assess all power quality problems related to pulsed loads is time-domain simulations of such systems taking advantage of the available powerful simulation tools (PSCAD, EMTDC, Matlab etc.) [37].

Pulsed loads integration requires an adequate and well-designed energy storage system. The discharge characteristics required will determine the storage system to be used. Energy storage options include : • Electrochemical : Batteries and Capacitors • Mechanical : Flywheels • Electromagnetic : Superconducting Magnetic Electric Storage Power sources for electromagnetic guns require technologies that are very different from the other applications due to very high pulse powers and very short durations (50GW and 6-10 ms). Compensated pulsed alternators, homopolar generators, capacitor banks and flux pumps are more suitable for millisecond-long pulses. Flywheels, concerning energy storage systems, stand much closer to the commercial applications, and currently are extensively researched.

In conclusion, future integration of electrical pulsed loads into warships will pose new challenges in the design of their power systems in order to ensure operational characteristics imposed by standards. Energy storage systems will play significant role in the operation of an IPS [34], [35], [38]. One energy storage technology will not probably be the optimum solution for every application provided that power and energy densities must be adequate to achieve the power and energy levels required for navy applications within the allowable economic, size and weight constraints.

VOLTAGE UNBALANCE

In a symmetrical three phase system, phase voltages are equal in magnitude and their phase angle differs by 120°. Deviation from this symmetry (unbalance) is caused by several reasons: • Asymmetrical distribution of single phase or two phase loads. • Single phase generation (small photovoltaic panels, fuel cell units etc). • Problems with joints and switching equipment (open circuits, loose connections, blown fuses etc). Unbalance causes problems to: • Induction motors: the rotating magnetic field of the negative sequence component (as a result of the unbalance) is opposite to the field of the positive sequence component. Therefore the machine does not produce its full torque. Additionally, the bearings suffer mechanical damage because of induced torque components at double system frequency and the stator and, especially, the rotor are heated excessively (leading to faster thermal ageing). Heat results from currents induced by the fast rotating (in the relative sense) opposite/negative magnetic field, as seen by the rotor. The extensive electrification of all systems onboard according to AES concept will include the installation of increased number of electric machines increasing, thus, the system problematic operation due to voltage unbalance. • Synchronous generators: phenomena similar to those described for induction machines, but mainly excess heating. • Transformers: parasitic losses in parts such as the tank are caused due to circulating zero sequence/homopolar currents in the delta winding of transformers. • Electronic power converters which are faced with additional, uncharacteristic, harmonics. Excessive currents associated with this could lead to tripping of drive systems and thermal stress. Harmonic distortion imposed to the rest of the system is also a concern.

Unbalance characterization – Continental Standards :

EN50160 [39], employs the ratio between negative and positive sequence component for voltage unbalance characterization. This should not exceed 2% for more than 95% of the time for rms values of negative and positive sequence components (average values obtained in 10 min intervals).

IEEE Standards and NEMA : Two definitions are given in IEEE 1159. One using the ratio of negative and positive components of voltage (as in EN50160) and also: «the maximum deviation from the average phase voltage, referred to the average of the phase voltage». This definition is also given in IEEE 112 [40]. The IEEE 100 [41] gives a slightly different definition: «the difference between the highest and the lowest rms voltage referred to the average of the three voltage». NEMA (National Electrical Manufacturers Association, USA) proposes the same definition with IEEE 112 but using phase-to-phase voltage [42]. It is shown in [43] that the 2 IEEE definitions, give significantly different results. If the three phase voltages are: va=0.97, vb=α , vc=α (α=1∠120°) then the ratio of negative to positive sequence component is 1.01% equal to the NEMA definition. However, the definition given by IEEE 112 gives 3.03% and the definition given by IEEE 100, 2.02%. In an example given in [42] for an unbalance where the negative to positive sequence component ratio is 23.8%, the NEMA definition gives 20%, which is an acceptable difference. For more realistic values of unbalance this difference becomes even smaller.

Unbalance characterization – Naval Standards :

STANAG 1008 Ed.8 uses the negative to positive sequence component ratio for voltage unbalance characterization as in IEEE 1159 and EN50160. The following formula, see also in [44], [45], is provided for a simpler calculation of this ratio using rms phase-to-phase voltages (vab, vbc, vca):

2 4 4 4 β - 3β - 6()vab + vbc + vca k 2 = where β = v 2 + v 2 + v 2 (4) 1 2 4 4 4 ab bc ca β + 3β - 6()vab + vbc + vca

A 2% limit is set for the above index. The NEMA definition [Pillay01] is given also as a quick approximation to the above formula. As mentioned above these two definitions give similar results.

In USA MIL-STD-1399 unbalance is measured «as the difference between the highest and the lowest line-to-line voltage». This definition is expressed with the following formula:

max(v , v , v ) - min(v , v , v ) VU (%) = 100 ab bc ca ab bc ca (5) nominal voltage

A 3% limit is given for 60 Hz/440 V power supplies.

2 An example is given to show that Eq. 4 and Eq. 5 produce different results. If va=0.97, vb=α , vc=α (α=1∠120°) then the ratio of negative to positive sequence given by Eq. 4 is equal to 1.01%. However, the ratio of Eq. 5 is equal to 1.5%. If va=0.95 then Eq. 4 gives 1.69% and Eq. 5 gives 2.53%.

LEAKAGE CAPACITIVE CURRENTS

Leakage capacitance phenomenon refers to capacitive leakage currents circulating between the main ship electric network and the ship's hull via stray equipment capacitances and/or their EMC filter input capacitances [46], [47], [48], [49], [50].

It is worth noting that stray capacitances actually exist between conductive elements at differential potential levels and at a distance from one another, e.g. two phases or one phase and the zero potential point (by default the ideal ground, which in case of a shipboard installation is the ship’s hull). In practice, however, only the capacitive elements between one phase and earth are those of significant value and, hence, corresponding importance for the phenomena discussed. Some examples of unintentional capacitances are spacings on printed circuit wiring boards, insulations between semiconductors and grounded heatsinks, and the primary-to-secondary capacitance of isolating transformers within the power supply.

The leakage capacitive current is mainly due to the following three mutually interrelated factors:

- the most common unearthed nature of electric system which is actually earthed via the high impedance capacitances of the neutral points of wye-connected transformer/generator windings supplying energy to downstream loads ( this earthing type is particularly selected in shipboard installations as it provides continuous power supply even in case of single-phase earth faults; this can not be done by low resistance in neutral earthing point). Nowadays, that High Voltage equipment is installed aboard following-up the AES realization, alternative grounding schemes are considered - especially for limiting short circuit currents - some of which have already been successfully integrated onboard, e.g. “Petersen’s coil earthing”. - stray capacitances (distributed or lumped type) of equipment such as machine windings and power distribution cables. In AES applications where High Voltage (in the range of 1kV up to 11 kV) equipment is extensively integrated, these stray capacitances are expected to increase by an order of magnitude. - capacitances of equipment input filters mainly used for EMC improvement. It is worth noting that in most cases, the filter (and hence its capacitance, too) is always connected to the electric grid disregarding if the main equipment is ON or OFF. As there is an augmenting interest in resolving any EMI problems in modern warships, especially in the advent of AES concept where all pieces of equipment are to be electrified and EMI-free, these filter capacitances are expected to contribute in an incremental manner to the leakage capacitive current circulation phenomenon.

In all cases, the resultant cumulative capacitance is of rather low impedance at 50/60 or 400 Hz operating frequency enabling, thus, the circulation of a fairly large capacitive current, in either steady-state (including harmonic distortion problems, see also above/below) or in transient-state e.g. in single-phase fault cases. The maximum leakage current is governed by standards and regulations and depends upon the type of equipment. The leakage current is given by :

IL = 2.π.U.f.C (6) where "IL" is the leakage current; "U" the voltage across the capacitor; "f" the frequency of the mains voltage across the capacitor; and "C" the capacitance.

The possible consequences, most of which are of adverse nature are related to an unexpectedly large capacitive current due to the high cumulative capacitance are enumerated –not necessarily in priority order- in the following:

- personnel safety : as even the steady-state circulating leakage current can be dangerous or lethal for the human factor - fuse failures or tripping due to the high capacitive current developed, leading either to system unbalanced operation or to complete power supply failure. - problems with insulation monitoring system installed onboard due to unearthed system. More specifically, the commonly used solution for unearthed electric system comprising the monitoring of continuous flow of an injected AC pulse might encounter several problems as the capacitive leakage current can be mistaken as “false monitoring signal”. - provocation of possible resonances especially in conjunction with harmonic distortion problem [51] : Thus, the high valued capacitance besides offering a low impedance circulation path for high order harmonics, it also lowers system natural frequencies ( natural frequencies are proportional to the square root of the ratio of system total over the corresponding capacitance, i.e. √(L/C) ) being responsible for resonance raisings in case of incidents with matching frequency. - deterioration of harmonic power quality problems as the total harmonic impedances decrease due to the capacitive susceptances offering a more conductive circulation path for harmonic currents; via the voltage drop caused by the harmonic current flow (see also harmonic distortion section) the harmonic voltage quality is affected, too.

However, there are also some positive consequences due to the high valued capacitance, namely the comparatively low transient overvoltages developed during switching operations, as these overvoltages are reciprocally proportional to the square root of the capacitance.

Not all regulations/rules consider the problem as of primary importance, therefore not all standards set any limits. For instance, STANAG-1008 (Ed. 8 and 9) and USA-MIL-STD-1399 set the following limitations – mitigation measures :

The total leakage current should not exceed 20 A, while referring to filter capacitances they should be less than 0.1 µF at the 60 Hz and 0.02 µF at the 400 Hz power system, respectively. Moreover, it is recommended that all single-phase equipment are connected between two phases rather than one phase and the ground. Further mitigation measurements include resonant earthing system via “Petersen coils”, which can counter-balance the leakage capacitive currents. Mains filters should be mounted as close as possible to power entry so that high frequency interference does not bypass the filter. To achieve higher attenuation or an increase in the effective working frequency range more complex filters can be made using more common mode or differential mode or capacitors.

It is worth noting, that in practice significant larger capacitances (followed by high valued respective capacitive currents) have been measured aboard several ships, raising major questions like among others for human factor safety.

OTHER PSQ ISSUES

Besides the aforementioned PSQ features, some additional issues should be considered : power supply interruptions and power frequency variations.

Power supply interruptions are an important aspect of power quality and it must be included in a monitoring survey with all the relevant details (cause and duration) in order to evaluate the performance of the electrical system, identify the weak components, apply enforcements and select appropriate power quality mitigation equipment. Although USA MIL-STD-1399(NAVY) (a source document for STANAG 1008) considers supply interruptions as a power quality parameter, STANAG 1008 does not, due to the fact that it considers only normal operation events.

Referring to power frequency variations, one must note that power system frequency is related to the rotational speed of the generators on the system which depends on the balance between the load and the capacity of the available generation. When this dynamic balance changes, small changes in frequency occur. In isolated power systems, as the ones in ships, problems occur when the governor response to abrupt load changes is not adequate to regulate within the bandwidth required by frequency sensitive equipment (example: processes that use power frequency for timing).

CONCLUSIONS

In this paper, power quality is seen from the naval power system point of view towards the complete electrification of all systems onboard and AES in particular :

The different aspects (origin, characteristics, effect on loads, characterisation methods) of several power quality phenomena are presented. These phenomena are: harmonics (an important issue due to the increasing use of power electronics), short duration voltage events (spikes, transients, dips and swells) and voltage unbalance (a threat for the operation of induction motors). Several issues considering the relevant continental and naval standards were discussed and differences between them were highlighted. Additionally, two topics with power quality implications are identified : voltage quality problems due to pulsed loads (which impose significant stress on the power system) and leakage capacitive currents (a distinctive characteristic of ships with particular effect on personnel safety, short-circuit faults, harmonic distortion, transients and EMC).

It becomes evident that power quality is a factor of increasing importance on the design and operation of naval electrical power system. The increasing needs of modern ships (advanced propulsion systems, modern onboard weapons etc) require studies with respect to the distinctive characteristics of ships and measurements surveys to depict power quality on existing systems. This kind of work (initiated by the relevant work on continental systems) will be reflected on future naval standards. Especially, STANAG 1008, which covers warships, although more detailed that standards provided by classification societies, has some vague points that have to be enriched.

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%).

REFERENCES

[1] R. Dugan, M. McGranaghan, H.W. Beaty, “Electrical Power Systems Quality”, Mc-Graw Hill, New York, 1996. [2] G. Vokas, D. Gazithellis, C. Katsikaris, G. Peponis, “Electric Power Quality”, Technical Chamber of Greece, Athens, May 1993 (In Greek). [3] M.H.J. Bollen, “Understanding power quality problems: voltage sags and interruptions”, IEEE Press, N.Y., 1999. [4] J.K. Mindykowski, “Assessment of Electric Power Quality in Ship Systems Fitted with Converter Subsystems“, Shipbuilding and Shipping Ltd, Gdansk , 2003, ISBN 83-88621-07-6 [5] IEEE Std 1159-1995, “IEEE recommended practice for monitoring electric power quality” [6] IEEE Std 519-1992, ‘IEEE Recommended Practice and Requirements for Harmonics Control in Electrical Power Systems’, IEEE, New York, 12 April 1993. [7] IEEE Standard 45-1998, “IEEE Recommended Practice for Electrical Installations on Shipboard” [8] IEC-60092/101, ‘Electrical installation in Ships - Definitions and General Requirements’, 1994. [9] Rules and Regulations of Classification Societies [10] STANAG 1008, ”Characteristics of Shipboard Electrical Power Systems in Warships of the North Atlantic Treaty Navies”, NATO , Edition 8, 21 Febr. 1994. / Edition 9, 24 Aug. 2004. [11] USA MIL-STD-1399(NAVY), “Interface standard for Shipboard systems – Section 300A – Electric Power, Alternating Current”. [12] M McGranahan, ‘Overview of the Guide for Applying Harmonic Limits on Power Systems-IEEE 519A’’, Proceedings of 8th International Conference of Harmonics and Quality Power, Athens (Greece), 16-18 October 1998, pp. 462-469. [13] NATO AC/141(SG/6)SG/4 GERMANY paper (02)96/10-03.04.96 : ‘STANAG 1008, Harmonic Distortion, Compilation of Replies to Questionnaire’ [14] NATO AC/141(SG/6)SG/4 GREECE paper (03)00/28-03.04.00, ‘Interpretation of Current Distortion Limits of STANAG 1008 Ed. 8’. [15] I K Hatzilau, A Magoulas, S Perros, D Kavoulakos, E Sakiotis, E Christofis, F Martinos, J Prousalidis, “Harmonic Power Quality On Naval Ships Electric Systems (Standards Overview - Field Measurements)”, Proceedings of 3rd Mediterranean Conference and Exhibition on Power Generation Transmission Distribution and Energy Conversion (MEDPOWER 2002), 4-6 November 2002, Athens (Greece). [16] J. Prousalidis, I.K. Hatzilau, S. Perros, “Harmonic electric power quality concepts for the electrified ships (AES)”, Proceedings AES2003 broadening the horizons, 13, 14 February 2003, Edinburgh, pp. 279-290. [17] IEEE “Surge-Trilogy” about “surges” : IEEE Std C62.41.1TM-2002, IEEE Std C62.41.2TM-2002 and IEEE Std C62.45.1TM-2002. [18] http://www.eeel.nist.gov/817/817g/spd-anthology/ [19] NATO AC/141(SG/6)SG/4, U.K., ADLPS/7/9/16-Sept.87 and 91/15-03.06.91, “Discussion paper on Voltage Spikes on shipboard power supply systems” [20] I.K. Hatzilau, M.E. Fafalios, “Phenomenological investigation of voltage spikes introduced by energizing low power electronic equipment”, Proceedings Int. Conf. on “Circuits, Systems and Computers ‘96”, HNA, Piraeus, 15-17 July 1996, Vol.2 pp. 513-520. [21] 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. [22] NATO AC/141(SG/6)SG/4, GREECE, 05/99(09) – 12.01.99. “Voltage Spikes and Transients (STANAG 1008) – Summary of Answers” [23] I.K. Hatzilau, J. Prousalidis, E. Styvaktakis, E. Sofras, ”Voltage and Current Spikes & Transients - Power Supply Quality aspects for the AES”, Proceedings of 5th Int. Symposium AES 2005 ‘Civil or military AES’, 13, 14 Oct. 2005, Paris. [24] S.F. Cannova, "Short-Time Voltage Transients in shipboard electrical systems", IEEE IAS 1972 Annual Meeting Conf. Rec. and IEEE Transactions IA-9, No.5, Sept/Oct 1973. [25] L. Bachman, M. Gullberg, F. Strickler, H. Sachs, “An Assessment of Shipboard Power Line Transients.”, Proceedings, International IEEE EMC Symposium, Aug. 1981. [26] TC3 of the Information Technology Industry Council (ITIC) , “ITIC (CBEMA) Curve”, 2000 [27] T. Key, A. Mansoor, F.D. Martzloff, “No joules for surges: relevant and realistic assessment of surge stree threats”, 7th International Conference on Harmonics and Quality of Power, Las Vegas, October, 16-18, 1996, (p.p. 67-72) [28] G. Lindes, A. Mansor, F.D. Martzloff, D. Vannoy, “Surge Recordings That Make Sense: Joules Deposition: Yes! - ‘Joule Content’: Never!”, Proceedings, PQA’97 USA, 1997 [29] F.D. Martzloff, “Surge recordings that make sense: Shifting focus from voltage to current measurements”, Proceedings, EMC’96 ROMA Symposium, September 1996 [30] J.O. Flower, C.G. Hodge, “Stability and transient-behavioural assessment of power-electronics based dc-distribution systems”, 1st WSEAS Int. Sub-Conference on Electroscience and Technology for Naval Engineering and All Electric Ship, Athens, July 2004. [31] IEEE Std 1453-2004, “Recommended Practice for Measurement and Limits of Voltage Fluctuations and Associated Light Flicker on AC Power Systems” (Adoption of CEI/IEC 61000-4-15:1997+A1:2003) [32] http://www.onr.navy.mil, US Office Naval Research. [33] NATO AC/141(SG/6)SG/4, GREECE, 17(GR)03.01 – 14.03.2003. “AES-Chapter 8, PULSED LOADS, Summary of Answers” [34] J.V. Amy, D.H. Clayton, T.J. Mc Coy, “The evolving nature of loads in an electric warship”, AES 2003, 13-14 February, UK. [35] Dr Janet, Dr Wegner, “Long Term Scientific Studies/47 on Electric Power Systems”, NATO Research and Technology Organization, 17 April 1998. [36] Howard A. Smolleck Satish J. Ranade Nadipuram R. Prasad Rudolph O. Velasco, “Effects of Pulsed – Power Loads upon an Electric Power Grid”, IEEE/PSS 1990 Sumner meeting, Minneapolis, Minnesota, July 15-19, 1990. [37] J. Prousalidis, S. Perros, I.K. Hatzilau, P. Buchanan, D. Muthumuni, “Introducing a COTS simulation tool for ship electric power quality studies”, INEC 2004, 16-18 March, The Netherlands. [38] A J Donaldson, BSc, MSc, MINucE, Ceng, “Energy storage – new technologies and new roles”, INEC 2002, 23-25 April, SECC, Glasgow, UK. [39] European Standard EN-50160, “Voltage characteristics of electricity supplied by public distribution systems”, CENELEC, Brussels, Belgium, 1994. [40] IEEE Std. 112-1991, IEEE Standard test procedure for polyphase induction motors and generators [41] IEEE Std. 100-1996, IEEE Standard Dictionary of Electrical and Electronic Terms. [42] P.Pillay and M. Manyage, “Definitions of Voltage Unbalance”, IEEE Power Engineering Review, pp. 50-51, May 2001. [43] M.H.J. Bollen, “Definitions of Voltage Unbalance”, IEEE Power Engineering Review, pp. 49-50, November 2002. [44] NATO AC/141(SG/6)SG/4, FRANCE, 91/09-28.10.91, “Stanag 1008 – Taux de desequilibre d’ un reseau triphase” ( Document Electricite de France d’aout 1984, J.P. Meyer, Contribution a la determination du taux de desequilibre d’ un reseau triphase ). [45] A. Robert and J. Marquet, ‘Assessing Voltage Quality with relation to Harmonics, Flicker and Unbalance, WG 36.05, Paper 36-203, CIGRE 92. [46] R Bal, G W Stephens, ‘Neutral Earthing of Marine Electrical Power Systems’, Trans I Mar E (TM), Vol. 95, Paper No 32, 1982. [47] NATO AC/141(SG/6)SG/4, UK, 92/19-01.10.92, “Revised summary of replies received to UK questionnaire on the use of gapacitive EMC filters” [48] NATO AC/141(SG/6)SG/4, GERMANY, (45)93/22-11.11.93, “Leakage capacitance” [49] NATO AC/141(SG/6)SG/4, GERMANY, (06)97/14, “Leakage capacitance. Effects of capacitance between line and earth” [50] T.C. Dalton, B. Desai, «All electrical systems are grounded – the difference is in the type and degree of isolating impedance», INEC2004, Marine technology in transition, 16-18 March 2004, Amsterdam. [51] NATO AC/141(SG/6)SG/4, THE NETHERLANDS, 09(NL)99/03, 1-3-1999 , “Harmonic Currents and Leakage Capacitance”