PROGRESS IN PHOTOVOLTAICS: RESEARCH AND APPLICATIONS
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
Applications
Prevention of Islanding
in Grid-connected
Photovoltaic Systems
M. E. Ropp*, M. Begovic and A. Rohatgi
School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0250, USA
Recently there has been a resurgence of concern about islanding of grid-connected
photovoltaic (PV) systems. This condition occurs when the PV system continues to
energize a section of the grid after that section has been isolated from the main utility
voltage source. Generally, islanding is undesirable because it poses a safety hazard to
utility service personnel, and also because it can lead to asynchronous reclosure which
can damage equipment. It is therefore important that PV systems incorporate
methods to prevent islanding. The purpose of this paper is threefold: (1) to critically
review the literature on islanding prevention methods for PV systems and discuss
their strengths and shortcomings; (2) to review and analyze the islanding behavior of
four converters which are prominent in the literature in order to demonstrate the
implementation and e€ectiveness of some islanding prevention methods; and (3) to
introduce a new islanding prevention scheme, active frequency drift with positive
feedback, which overcomes many of the shortcomings of existing schemes. It is
concluded that no `perfect' islanding prevention method yet exists, but also that
many existing methods or combinations thereof work very well in practical situations.
Finally, it is noted than an investigation of what constitutes `sucient' islanding
prevention is needed. Copyright # 1999 John Wiley & Sons, Ltd.
INTRODUCTION
I
n many countries throughout the world, programs are being implemented to encourage the
installation of grid-connected photovoltaic (PV) systems. The proliferation of such systems has led to
a resurgence of concern about the problem of islanding. Islanding occurs when a PV system feeds
power into a section of the utility system which has been isolated from the utility voltage source. Consider
the con®guration shown in Figure 1, a PV system connected to a feeder line which is in turn connected to
the utility grid through a transformer and some sort of switch (a recloser, breaker, fuse etc.). The PV
system consists of a PV array and a power conditioning unit (PCU). A local load is also connected to the
feeder line. If the switch were opened, under certain conditions it is possible for the PV PCU to continue
to energize the isolated section of the grid and supply power to the local load. This is islanding, and the
isolated section of the utility being powered by the PV system is referred to as an island of supply or simply
an island. Utilities frequently use the term `renewable energy island' to di€erentiate this situation from
other types of islands. Although this distinction is frequently important, for brevity we will use the term
`island' throughout this paper without ambiguity.
* Correspondence to: M. E. Ropp, School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta,
GA 30332-0250, USA
CCC 1062±7995/99/010039±21$17.50
Copyright # 1999 John Wiley & Sons, Ltd.
Received 3 March 1998
Revised 31 July 1998
40
M. E. ROPP, M. BEGOVIC AND A. ROHATGI
Figure 1. Schematic of a PV system connected to a utility feeder which can be isolated from the utility by the switch
at the right
The amount of time between the disconnection of the utility and the shutdown of the PCU is referred to
as the run-on time. Islanding events are typically subdivided into two categories: long-term, with run-on
times of 1 sec or more, and short-term, with run-on times of less than 1 sec.1,2
The primary concern with long-term islanding is one of safety.1±8 Maintenance or repair personnel
arriving to service the isolated feeder may be unaware that it is still energized, which could lead to
personal injury. This is of particularly great concern in the case of scheduled maintenance, when the
switch would be manually operated by service personnel who will immediately commence work on the
isolated system. In this case, islanding of even a few tens of seconds could be dangerous.
Another problem associated with both long- and short-term islanding is that the PV system, which
relies on the utility voltage to provide a phase and frequency reference for its output current, may lose
synchronism with the utility while the switch is open. The utility could then reclose on a PV system which
is out of phase.2,9 Most PV PCUs are two quadrant devices, designed for unidirectional power ¯ow from
the DC to AC side only. This is done for economic reasons; two-quadrant converters are less expensive
than four-quadrant converters. However, during an out-of-phase reclosure, there are intervals in which
the polarities of the voltage and current are opposite. During these intervals, the converter is absorbing
power from both the PV and utility sides, which can lead to destructive component failures in the PCU.
It has been postulated that another possible problem with short-term islanding is that it can interfere
with the arc-clearing function of protective relays.9 However, there is much debate over whether this is a
signi®cant issue. This subject will be discussed in more detail shortly.
One ®nal problem which is increasing in relevance is that some islanding prevention methods interfere
with each other, leading to longer run-on times and possibly failure to detect islanding if several PV
systems are present in the island. In some cases, this can happen even if all the PV systems in the island are
using the same islanding prevention scheme. This situation has been termed the `multi-inverter case', and
it could become increasingly common with the proliferation of small, roof-mounted PV systems and the
development of PCUs for AC PV arrays, in which case there could be tens or even hundreds of PCUs in an
island.
At this time, no universal standard for a maximum PV PCU run-on time has been adopted. Many
utilities in the United States have selected 1 sec as an acceptable time. Current IEEE standards
recommend 2 sec,8 but the new IEEE-929 standard, which speci®cally addresses this and many other
issues related to PV-utility interconnection, recommends di€erent run-on times depending on the nature
of the islanding situation and which protection mechanisms are likely to operate.10 In Japan, the target
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
PREVENTION OF ISLANDING
41
time is 0.5±1 sec;11 in Germany, a maximum run-on time of 5 sec has been proposed.12 In general, an
islanding prevention method should accomplish these goals:
(I) Detection of islanding and disconnection of the PV system from the utility, regardless of the initial
state of the system, perturbations, composition of the load, or presence of other equipment such as
other PV systems (i.e. the multiple inverter case).
(II) Detection of islanding which is suciently fast to guarantee safety and safeguard the reliability and
integrity of the utility and PV systems.
(III) Disconnection of the PV system only when islanding is actually occurring (no nuisance trips).
A considerable body of work exists on islanding and islanding prevention methods. The purpose of
this paper is threefold: ®rst, to collect and review the existing methods, discuss their strengths and
weaknesses, and compare them; second, to review and analyze the behavior of four existing converters to
demonstrate the e€ectiveness and implementation of these methods; and third, to propose a novel
islanding prevention scheme, active frequency drift with positive feedback (AFDPF), which largely
overcomes the shortcomings of the existing methods.
ISLANDING PREVENTION BY STANDARD PROTECTION SYSTEMS
Grid-connected PV systems are required to have an overvoltage relay (OVR), an undervoltage relay
(UVR), an overfrequency relay (OFR), and an underfrequency relay (UFR) which disconnect the
PV system from the utility in the event that the magnitude or frequency of the PCU's terminal voltage
goes beyond certain limits.10 Under most circumstances, these relays will prevent islanding. To understand this, consider the con®guration shown in Figure 2. When the recloser is closed, real and reactive
power PPV ‡ jQPV ¯ows from the PV system to node a, and power Pload ‡ jQload ¯ows from a to the load.
Summing at node a, we see that
DP ˆ Pload ÿ PPV
1†
DQ ˆ Qload ÿ QPV
Figure 2. PV system/utility feeder con®guration showing de®nitions of power ¯ows and terms
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
42
M. E. ROPP, M. BEGOVIC AND A. ROHATGI
are the real and reactive power ¯owing into the feeder from the utility. It should be pointed out that PV
PCUs typically operate with a unity power factor, so under most conditions QPV ˆ 0 and DQ ˆ Qload .
The real and reactive power being consumed by the load are given by
h
i
*
Pload ˆ Re V~a *I~ load ˆ 2Va Iload cos f
h
i
*
Qload ˆ Im V~a *I~ load ˆ 2Va Iload sin f
2†
where cos(f) is the load displacement power factor (dpf) and Va and Iload are the RMS values of va ,
the instantaneous voltage at a, and iload , the load current. The superscript asterisk denotes complex
conjugation. Assuming that the load can be represented as a parallel RLC circuit, these expressions may
be written in terms of Va as follows:
Va
Rload
2
3
Va
Va
ÿ
ˆ Va 4
1 5
oL
oC
Pload ˆ Va
3†
Qload
4†
When the recloser opens, DP and DQ will both go to zero. The behavior of the isolated system will depend
on DP and DQ at the instant before the recloser opens to form the island, which we denote DP ÿ and DQ ÿ .
There are four cases in which the OVR/UVR or OFR/UFR will prevent islanding.11
(1) DP ÿ 4 0. In this case, the PV system is producing less real power than is required by the local load
(Pload 4 PPV). From equation (3) we see that when the switch opens and DP becomes zero, Pload will
decrease, meaning that Va must also decrease since Rload can be assumed to be constant over this time
interval. This decrease can be detected by the UVR, and islanding is prevented.
(2) DP ÿ 5 0. In this case, Pload 5 PPV , and power is ¯owing into the utility system. Now, when DP
becomes zero, Pload must increase and Va will also increase. This condition can be detected by the
OVR, and again islanding is prevented.
(3) DQ ÿ 4 0. This case corresponds to a lagging power factor load, or a load whose reactive component
is inductive. After the recloser opens, DQ ˆ 0. However, as previously mentioned, QPV is usually
zero, and therefore Qload ˆ 0. This requires the term in square brackets in equation (4) to become
zero, meaning that the inductive part must drop and the capacitive part must increase. Equation (4)
shows us that in order for this to occur the frequency o of va must increase. This increase in o can be
detected by the OFR.
(4) DQ ÿ 5 0. This case corresponds to a leading power factor load, or one which is primarily capacitive.
As in case (3), when DQ becomes zero, the inductive and capacitive parts of equation (4) must balance
so that Qload ˆ 0, and this requires o to decrease. This can be detected by the UFR.
Note that cases (3) and (4) can be expressed in terms of a phase condition. The PV system will cause the
frequency to change until the following condition is satis®ed:
argfR
ÿ1
ÿ1
‡ joC ÿ j…oL† g ˆ 0
5†
This occurs at the load's resonant frequency, ores ˆ (LC) ÿ0.5. At ores , steady state is reached and no
further change in o occurs. If ores lies outside the trip limits of the OFR/UFR, islanding will not occur. It
bears repeating at this point that all PV PCUs for utility interface applications are required to have OVR/
UVR and OFR/UFR protection.10 Therefore, if either the real or reactive power of the load and PV
system are not matched, and the thresholds for the OVR/UVR and OFR/UFR are properly chosen,
islanding will not occur.
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
PREVENTION OF ISLANDING
43
With this understanding of the operation of the four standard relays, it is now appropriate to revisit the
issue of interference of PV systems with the arc-clearing function of utility reclosers. Reclosers are circuit
breakers which open when an overcurrent condition caused by an arc or other fault is detected, and then
automatically reclose after a short time interval. During the open interval, the ionized air which forms the
conducting channel of the arc should dissipate, thereby allowing the arc to clear itself without an extended
interruption of electrical service. If the arc does not clear during the ®rst open interval of the recloser, it
will open again for a longer period of time. Some reclosers repeat this procedure a third time. If the
overcurrent condition persists after all the open intervals, the recloser opens and `locks out', meaning that
it will not attempt to reclose until it is manually reset by repair personnel. For a con®guration like that
shown in Figure 2, it has been postulated9 that, if the impedance of the arc is suciently high, the PV
system could continue to maintain the ionized air channel, preventing the arc from clearing during the
recloser's ®rst open interval. This could lead to unnecessarily long interruptions of power and a decrease
in the reliability of electric power service to the load. However, let us consider what can be expected to
happen if, for example, a ground fault is connected in Figure 2 between the utility feeder and ground
anywhere inside the isolated system. In this simple model, the fault appears as a resistance connected in
parallel with the load. Arcs and faults which draw enough current to trip utility reclosers have very low
impedances.12,13 If such a fault were present inside the recloser in Figure 1, it would short out the load,
leading to a decrease in the voltage at node a which would be detected by the PV system, since the UVR is
required to deactivate the PCU very quickly (within a few line cycles10) if the voltage at a drops to a very
low level. A fault with suciently high impedance that the drop in Va would not be large enough to trip
the UVR would not draw enough current to cause a standard recloser to open.13,14 Therefore, with present
equipment, it appears that PV systems should not interfere with the arc-clearing function of reclosers.
However, high-impedance fault detection devices are under development,13,14 and widespread use of such
detectors could change the islanding protection requirements for PV systems. This complex issue is one
reason for the aforementioned lack of consensus on the maximum allowable run-on time for PV PCUs,
and therefore it is an area in which further research is needed.
SHORTCOMINGS OF THE STANDARD PROTECTION SYSTEMS:
THE NONDETECTION ZONE
We have thus far examined four cases in which the OVR/UVR and OFR/UFR of a PV system will prevent
islanding. Unfortunately, there is another possible case: DP ÿ ˆ DQ ÿ ˆ 0. This corresponds to a case in
which the PV power production is matched to the load power requirement, and the load dpf is unity. In
this case, when the switch is opened no change occurs in the isolated system, and the OVR/UVR and
OFR/UFR do not operate. In reality, DP ÿ and DQ ÿ do not have to be exactly zero for this to occur
because the magnitude and frequency of the utility voltage can be expected to deviate slightly from
nominal values, and therefore the thresholds for the four relays cannot be set arbitrarily small or else the
PV system will be subject to nuisance trips. This limitation leads to the formation of a nondetection zone
(NDZ), as shown in Figure 3. Studies have shown that the probability of DP ÿ and DQ ÿ falling into the
NDZ of the OVR/UVR and OFR/UFR can be signi®cant.15,16 It is therefore important that PV systems
incorporate methods to prevent islanding in the case in which DP ÿ ˆ DQ ÿ 0.
METHODS OF ELIMINATING THE NDZ
Passive methods
Passive methods for islanding prevention involve monitoring the PCU's terminal voltage (va) for some
condition which indicates islanding. These methods are discussed below.
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
44
M. E. ROPP, M. BEGOVIC AND A. ROHATGI
Figure 3. Demonstration of the nondetection zone in which the standard four relays cannot detect PV
system islanding
Voltage harmonic monitoring
The voltage harmonic monitoring method does not rely on a real or reactive power mismatch for islanding
detection. Instead, the PV PCU monitors the total harmonic distortion (THD) of va and shuts down if this
THD exceeds some threshold. There are two mechanisms which can cause the harmonics in va to increase
when islanding begins. One of these is the PV PCU itself. A PV PCU will produce some current
harmonics in its AC output current, as all switching power converters do. A typical requirement for a gridconnected PV PCU is that it produce no more than 5% THD of its full rated current.17 Under normal
operation, the utility, being a `sti€' voltage source, forces an `undistorted' sinusoidal voltage (THD 0)
across the load terminals, causing the (linear) load to draw an undistorted sinusoidal current. Summing at
node a, we see that under this condition the harmonic currents produced by the PCU will ¯ow out into the
grid. When the recloser opens, the harmonic currents produced by the PCU will ¯ow into the load, and by
Ohm's law and superposition, these will produce harmonics in va .18 These voltage harmonics can be
detected by the PCU, which can then assume that the PV system is islanding and discontinue operation.
The second mechanism is the voltage response of the transformer shown in Figure 2. If the switch which
disconnects the utility voltage source from the island is on the primary side of the transformer, the
secondary of the transformer will be excited by the output current of the PV system. However, because of
the magnetic hysteresis of the transformer, the voltage response of the transformer to this (approximately)
sinusoidal excitation is highly distorted. In particular, it contains a large third harmonic component.18
This phenomenon has been veri®ed experimentally; investigators have found that the third harmonic
voltage at node a can increase by a factor of ®ve or more when islanding begins because of the distortion
introduced by standard pole-mounted transformers.18
In theory, the voltage harmonic monitoring method promises to be highly successful in detecting
islanding under a wide range of conditions,18 and its e€ectiveness should not change signi®cantly in the
multiple-inverter case. However, it su€ers from a serious implementation diculty: it is not always
possible to select a trip threshold which provides reliable islanding protection but does not lead to
nuisance tripping of the PV system. It is clear that a threshold must be selected which is: (a) higher than
the THD which can be expected in the grid voltage; but (b) lower than the THD which will be produced
during islanding by the two mechanisms described above. Let us assume that the PV PCU produces 5%
THD in its output current, the maximum allowable limit. For a resistive load fed by this current, the THD
of va will also be 5%, but for a parallel RLC the load impedance decreases with increasing frequency
above ores , and so there can be less distortion in the voltage response than in the exciting current. It is
therefore clear that the THD threshold will have to be set lower than 5%. In reality, the utility voltage
distortion which we assumed to be 0 in the foregoing discussion can actually be expected to be 1±2%
under normal conditions, but there are many conditions, such as the presence of power electronic
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
PREVENTION OF ISLANDING
45
converters which produce current harmonics at frequencies at which the utility system has resonances,
which can cause this value to increase signi®cantly.19 Also, transient voltage disturbances, particularly
large ones such as those which accompany the switching of capacitor banks,20 could be interpreted by
PV PCUs as a momentary increase in THD, depending on the measurement technique used. It is clear
that in some cases it is not possible to select a threshold that meets criteria (a) and (b). It may be possible
to overcome this problem using digital signal processing and harmonic signature recognition, but these
techniques cannot be implemented cost-e€ectively in small PV PCUs. For these reasons, the harmonic
monitoring technique has not been used commercially.
Transient phase change or phase jump detection
Another method of islanding prevention, phase jump detection (PJD), involves monitoring the phase
between the inverter's terminal voltage and its output current for a sudden `jump'.6,18 Consider the case
in which the load in Figure 2 has a nonunity dpf (nonzero voltage-current phase angle). Under normal
operation, the PCU's output current waveform will be synchronized to the utility voltage by detecting
the rising (or falling) zero crossings of va . This is done through the use of a phase-locked loop (PLL). A
schematic of the PLL circuit is shown in Figure 4.21,22 The input line ®lter removes noise and higherorder harmonics from the input signal, in this case va . The phase comparator generates a signal whose
magnitude is proportional to the phase error between its input signals, which are the ®ltered va and the
output of the voltage-controlled oscillator (VCO). In most modern PLLs, the phase comparator uses
some sort of a measurement of the time between rising zero crossings of its two input waveforms to
determine the phase error. The loop ®lter removes AC components from the output of the phase
comparator. The DC output signal of the phase comparator and loop ®lter is used to adjust the output
of the VCO in such a way as to reduce the phase error between its output and va . In a PV inverter, the
VCO output provides the waveform template for the PCU output current iPV and thus has the same
phase and frequency iPV . Therefore, the function of the PLL is to synchronize va and iPV . To implement
PJD, we simply need to add a device which measures the DC output signal of the phase comparator and
loop ®lter and generates a signal to deactivate the PCU when this signal reaches or exceeds some
threshold.
PJD works as shown in Figure 5. When the utility is deactivated, suddenly it is the PV current iPV which
becomes the ®xed phase reference, since iPV is still following the waveform template provided by the PLL.
However, since the frequency has not changed, the current-voltage phase angle of the load must be the
same as before the utility switch opened, and therefore va must `jump' to this new phase as shown in
Figure 5. At the next rising zero crossing of va , the resulting phase error between the `new' voltage and the
PCU's output current, if greater than the threshold, will cause the monitoring device in the PLL to
generate its `stop' signal, and islanding is prevented. If we make the simplifying assumption that the
Figure 4. Basic con®guration of a phase-locked loop
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
46
M. E. ROPP, M. BEGOVIC AND A. ROHATGI
Figure 5. Figure explaining the operation of the phase jump detection method
system response is instantaneous, we can write an approximate phase criterion for PJD similar to that for
the OFR/UFR:
ÿ1
ÿ1 argfR ‡ joC ÿ j…oL† g 5 fth
6†
where o is the frequency of va and fth is the phase threshold at which a stop signal is generated. If equation
(6) is satis®ed at the utility frequency o0, then islanding will not occur.
The advantages of PJD are its sensitivity and ease of implementation. Unfortunately, its disadvantage
is that it has an NDZ for unity dpf loads, within the existing NDZ of the four standard relays. This NDZ
can be moved away from the unity-dpf load region by operating the PCU at a nonunity power factor, but
as was previously mentioned it cannot be moved very far without making the PCU more expensive.
Phase-jump detection therefore can shrink but does not eliminate the NDZ of the four standard relays.
Slide-mode frequency shift
In the slide-mode frequency shift (SMS) method, the current-voltage phase angle of the PCU, instead of
always being controlled to be zero, is made to be a function of the frequency of va as shown in Figure 6.11
The S-shaped phase response curve of the PCU is designed such that the phase of the inverter increases
faster than the phase of most unity-dpf loads in the region near the utility frequency o0 . This makes o0 an
unstable operating point for the PCU. While the utility is connected, it stabilizes the operating point o0 by
providing the phase and frequency reference. However, after the switch opens, the phase-frequency
operating point of the load and PV system can be at an intersection of the load line and PCU phase
response curve. Consider the load line of the unity-dpf load shown in Figure 6. The load line and PCU
curve intersect at (o0 , 0), but if there is any small perturbation of the frequency of va away from o0 , the
instability of the PCU at o0 causes the PCU to reinforce the perturbation and drive the system to a new
operating point, either at o1 or o2 depending on the direction of the perturbation. The PCU phase curve
can be designed in such a way that o1 and o2 lie outside the NDZ of the OFR/UFR.
SMS is implemented through the design of the input line ®lter to the PLL. This ®lter controls the phase
characteristic of the PCU because it controls the reference signal for the PLL. Therefore, all that is
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
PREVENTION OF ISLANDING
47
Figure 6. Narrow frequency range graph showing the phase vs frequency characteristic of a PCU utilizing the SMS
islanding prevention method. The S-shaped curve is the PCU phase-frequency characteristic; the dashed lines are the
load `lines' [inverted because of the minus sign in equation (7)]. The second load line passes through (o0, 0),
indicating a unity-dpf load
required is to design the input ®lter to have the desired phase characteristic. We can write an expression for
the steady-state frequency attained by this method in terms of another phase criterion
argfR
ÿ1
ÿ1
‡ joC ÿ j…oL† g ˆ ÿ arg fG… jo†g
7†
where G( jo) is the transfer function of the input line ®lter. Consider the case in which the phase shift of
the line ®lter is zero at the utility frequency and the load has a near-unity dpf. If equation (7) has more
than one solution (see Figure 6), the solution at the utility frequency is unstable and islanding will not
occur.
This scheme has been shown to be highly e€ective, both theoretically and experimentally.6 It works for
purely resistive loads, whose phase response curves lie in the frequency axis in Figure 6. It also works for a
wide range of RLC loads. However, some RLC loads have phase response curves such that the phase of
the load increases faster than the phase of the PV PCU, meaning that equation (7) has only one solution at
o0 . This makes the nominal line frequency a stable operating point and renders SMS ine€ective. We have
performed computer modeling to demonstrate this fact, and the results are shown in Figures 7 and 8. In
these simulations, an SMS phase response curve from a commercial inverter6 is plotted against the phase
responses of several parallel RLC loads. These examples show that SMS has an NDZ for RLC loads with
relatively small values of L but large values of C (Figure 7), or low-power loads in which R is large
(Figure 8). An additional problem with SMS is that it relies on an uncontrollable, externally-supplied
perturbation, which makes predictions of the run-on time of an SMS-equipped PV system dicult.
Active methods
In order to eliminate the shortcomings of the passive NDZ elimination methods, several active methods
have been developed. Active methods involve changing the system con®guration or control of iPV in such a
way as to cause a change in va when islanding. The active methods include:
Output variation or `impedance measurement'
Since the PV system appears to the utility as a current source supplying current given by
iPV ˆ IPV sin …oPV t ‡ fPV †
8†
there are three parameters of the PV system output which can be varied: the amplitude IPV , the
frequency oPV , and the phase fPV . In the output variation method, a perturbation is periodically
applied to one of these parameters.18 If the utility is disconnected, this variation will force a detectable
change in va which can then be used to prevent islanding. In so doing, the PCU is in e€ect measuring
dva/diPV , and therefore this method is also called impedance measurement. Its primary advantage is that
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
48
M. E. ROPP, M. BEGOVIC AND A. ROHATGI
in theory it has no NDZ; for a single PV system with any local load, if the load and PV powers are
balanced upon disconnection of the utility, the output variation of the PCU will upset this balance and
cause the UVR to trip. However, output variation has signi®cant disadvantages. One of the most serious
of these is that its e€ectiveness decreases in the multi-inverter case. This happens even if all PCUs in the
Figure 7. Plot of an actual SMS phase-frequency characteristic and the phase responses of several RLC loads. The
dark S-shaped curve is the SMS phase characteristic; the other curves are the RLC phase responses [inverted because
of the minus sign in equation (7)]. R is held ®xed; L and C are varied, maintaining unity power factor at 60 Hz. For
the bottom three loads in the legend, SMS cannot detect islanding
Figure 8. Plot of an actual SMS phase-frequency characteristic and the phase responses of several RLC loads. The
dark S-shaped curve is the SMS phase characteristic; the other curves are the RLC phase responses [inverted because
of the minus sign in equation (7)]. L and C are ®xed and resonant at 60 Hz; R is varied. For the top three loads in the
legend, SMS cannot detect islanding
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
PREVENTION OF ISLANDING
49
island are using output variation, unless the variation is somehow synchronized. The reason is that as
more PCUs are added to the island, the amount of variation introduced by each PCU into the total iPV
being generated by all PV systems is reduced, and eventually the variation becomes so small that the
change in va becomes undetectable. Also, for high-power PV systems, output variation can lead to grid
instability, voltage ¯icker, and several other problems. Therefore, output variation is not suitable for
multiple small systems or for single large systems. These disadvantages have led many to conclude that
this method is of little practical value.23
Active frequency drift
In the active frequency drift (AFD) method, the PV PCU uses a slightly distorted output current to cause
the frequency of va to drift up or down when the utility is disconnected. One example of a waveform that
implements upward frequency drift is shown in Figure 9, along with an undistorted sine wave for
comparison. TVutil is the period of the utility voltage; TIpv is the period of the sinusoidal portion of the
current output of the PV system; and tz is a dead or zero time. The ratio of tz to half the utility voltage
period, TVutil/2, is referred to as the `chopping fraction' (cf). During the ®rst portion of the ®rst half-cycle,
the PV system's current output is a sinusoid with a frequency slightly higher than that of the utility
voltage. When the PV output current reaches zero, it remains at zero for time tz before beginning the
second half-cycle. For the ®rst part of the second half-cycle, the PV output current is the negative half of
the sine wave from the ®rst half-cycle. When the PV current again reaches zero, it remains at zero until the
rising zero crossing of the utility voltage. When we apply such a current waveform to a resistive load in the
absence of a utility voltage, its voltage response will follow the distorted iPV , and therefore va will reach a
rising zero crossing tz seconds before it would have had the utility still been connected. This is interpreted
by the PV system as an increase in the frequency of va . The PV system then increases its frequency to
attempt to maintain the relationships shown in Figure 9. The resistive load again responds by advancing
the negative to positive zero crossing of va by tz , which is again interpreted by the PV system as an increase
in frequency, and this process continues until the frequency has drifted far enough from nominal to be
detected by the OFR.
Figure 9. Example of a waveform used to implement the AFD method of islanding prevention. A pure sine wave is
also shown for comparison
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
50
M. E. ROPP, M. BEGOVIC AND A. ROHATGI
AFD is also highly e€ective in detecting a wide range of islanding conditions. However, it too has an
NDZ. Recall from the OFR/UFR discussion that, if the local load is capacitive, the voltage frequency in
the island will exhibit a tendency to decrease, partially counteracting the upward frequency drift of the
PCU. If the percentage of zero time of the inverter output current is ®xed, it has been shown experimentally that there will always be a particular value of capacitance which can be added to a resistive load
that will result in a downward frequency drift that exactly cancels the upward frequency drift of the PCU,
and under this condition islanding can continue inde®nitely.23 Therefore, AFD is known to have an NDZ
for such RC loads. However, it has recently been shown that AFD also has an NDZ for a range of parallel
RLC loads. For such loads, a phase criterion for a stable steady state frequency (a limit cycle) for an
AFD-equipped PCU can be written as follows:24
argfR
ÿ1
‡ … joL†
ÿ1
‡ joCg
ÿ1
ˆ ÿ05otz ˆ ÿ05pcf
9†
where cf is the chopping fraction de®ned previously. Once the system reaches this steady-state frequency,
which will be slightly higher than the load's resonant frequency, no further frequency increase occurs. If
the steady-state frequency lies within the trip thresholds of the OFR/UFR, and the magnitude response of
the RLC load is such that the voltage does not go beyond the OVR/UVR thresholds, islanding can
continue inde®nitely.24 Using this equation it has been shown that AFD simply moves the NDZ of the
OFR/UFR to leading-dpf loads for RLC loads with small C and large L, and also narrows the NDZ for
large C and small L.24 However, a large cf can be required to obtain a signi®cant bene®t, which leads to
signi®cant distortion of the PCU's output current.
A new approach: active frequency drift with positive feedback
We have been able to partially overcome the disadvantages of AFD by adding positive feedback, creating
a new method called active frequency drift with positive feedback AFDPF.24 In the AFDPF method, cf is
varied during each current cycle of the PV PCU according to the expression
cfk ˆ cfkÿ1 ‡ F…Dok †
10†
where cfk is the chopping fraction in the kth cycle, and F(F:R ! R) is a mapping of the sampled
frequency error in the kth cycle, Dok ˆ ok ÿ o0 , onto the correction of cf. Using a linear F, it has been
demonstrated that AFDPF has a signi®cantly smaller NDZ than does AFD; in some cases a reduction of
NDZ size of several orders of magnitude was obtained.24 There are two primary reasons for this reduction
in NDZ size. One is that an AFDPF-equipped PV PCU is capable of reinforcing a negative frequency
deviation, if F has odd symmetry. If Dok is negative, cf will decrease, and can even become negative so that
the converter has a downward-drifting tendency. In this way, the converter `cooperates' with loads which
tend to drive the frequency of va down, instead of trying to counteract them as AFD does. The second
reason for the NDZ size reduction is that the `leveling o€' of the frequency does not occur until the
frequency deviations are much larger than is the case with AFD. This is most easily seen by examining a
simpli®ed situation in which the system response is assumed to be instantaneous and F is assumed to be
linear (F ˆ KDok). Then, substituting equation (10) into (9) yields
ÿ1
argfR
‡ … jok L†
ÿ1
‡ jok Cg
ÿ1
ˆ ÿ05p…cfkÿ1 ‡ KDok †
11†
This approximate phase condition shows that as the frequency error increases, the required phase of the
load increases, and so does the deviation of the frequency away from the load's resonant frequency which
is required to attain steady state. This e€ect impedes the attainment of a steady state frequency until much
larger frequencies are reached than would be the case with AFD, giving a greater chance for detection by
the OFR/UFR.
Another major advantage of AFDPF is that it does not lose e€ectiveness in the multiple inverter case.
Both AFD and AFDPF should maintain their e€ectiveness in the case in which all inverters in an island
use the same scheme, but AFDPF can also work in conjunction with several other schemes. For example,
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
PREVENTION OF ISLANDING
51
consider a case in which there are two PV systems, one equipped with AFDPF and one with harmonic
detection. When the utility is disconnected and AFDPF begins to increase cf, the THD in the AFDPF
PCU's output current will increase.24 This increase in distortion can help to trip the PCU using harmonic
detection, increasing the probability that one of the converters will trip and islanding will stop. AFDPF
can also provide the perturbation required to make SMS work. AFDPF is therefore well-suited for
application to multiple small PV systems or to AC arrays.
AFDPF is clearly a very promising islanding prevention scheme. However, it does have some drawbacks. One of the most serious is that, if the measurement of the frequency of va is inaccurate, AFDPF will
cause a PV PCU to have a higher steady-state THD in its output current than would AFD. This problem
can be minimized by appropriately choosing cf0 and F, and by good ®lter and zero-crossing-detector
design. Also possible but less likely is the potential for instability in areas of the utility system which
experience high levels of PV penetration if the utility voltage source is not suciently sti€. The suitability
of this method for large PV systems is therefore unclear. Finally, although AFDPF signi®cantly reduces
the size of the NDZ, it does not eliminate it. AFDPF has an NDZ for RLC loads with large C and small
L, similar to (but smaller than) that found for SMS. Research on AFDPF is currently ongoing.
Reactance insertion
The reactance insertion method1,2 is unique among islanding prevention methods in that it does not rely
on the PCU to detect the islanding condition. Instead, a large reactance, usually a capacitor bank, is
installed on the utility system inside the potential island at point b as shown in Figure 10. The switch is
normally open. When the recloser opens, the capacitor bank switch closes after a short delay. In the case
where DP ÿ and DQ ÿ were nearly zero, the sudden addition of a large reactive impedance will imbalance
the reactive power requirement, causing a frequency decrease which the UFR can detect. This method
o€ers several advantages. It is highly e€ective in preventing islanding1,2 as long as the small delay is
allowed between the time of recloser opening and the time of capacitor insertion. This delay is necessary
to ensure that insertion of the capacitor will not actually create a balanced situation between the PV
system and a lagging load. Capacitors of this type are readily available, and utilities have a great deal of
experience with them. The same capacitor bank could also be used for reactive power support, with only
minor changes in the switching logic to allow it to maintain its unit-islanding function. However, two
readily-apparent drawbacks to the reactance insertion method are: (1) there may be multiple switches in
series leading into the potential island, meaning that each series switch might need to be equipped with a
switchable capacitor bank (depending on the load con®guration); (2) this method cannot prevent shortterm islanding, partially because of the speed of action of the capacitor switches and partially because of
the necessary delay in switching. This method actually has a third problem which is more political than
technical: it requires the installation of equipment on the utility side of the point of common coupling,
which is usually taken to be the utility's electric meter. Utilities generally look unfavourably on such an
arrangement.
Figure 10. The reactance insertion method. This is the con®guration of Figure 1, now equipped with a switchable
capacitor bank at point b
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
52
M. E. ROPP, M. BEGOVIC AND A. ROHATGI
Summary of existing islanding prevention methods
Table I summarizes the islanding prevention methods discussed in this paper. Each islanding prevention
scheme is categorized according to the types of load conditions under which it can prevent islanding, and
the `Remarks' column addresses issues not easily quanti®ed elsewhere in the table. It is clear from the
table that no `perfect' islanding prevention scheme has yet been devised; all existing methods involve a
tradeo€ between e€ectiveness in islanding prevention, power quality, simplicity and cost-e€ectiveness.
None of the methods simultaneously meets all three of the goals de®ned in the Introduction of this paper.
RESULTS OF TESTING OF FOUR PV PCUs
Although many PV PCUs have been designed and built, there are three units which appear frequently in
the literature and whose islanding characteristics have been carefully scrutinized, and a fourth whose
islanding behavior has been described recently. These are the Gemini, manufactured by WindWorks in
Mukwonego, Wisc; the Teslaco, manufactured by the Teslaco Corporation in Irvine, Calif.; the APCC,
made by the American Power Conversion Corporation in Billerica, Mass; and an unnamed converter
made by Sanyo Corporation which uses the slide-mode frequency shift method and is referred to herein as
the SMS.
The Gemini
The Gemini PV PCU was an early 6 kW line-commutated inverter made by WindWorks. It utilized SCRs
in a bridge con®guration with a line-frequency transformer. It has been found3,6,25 that this PCU is prone
to inde®nite run-on times under certain conditions. However, this behavior was related to the fact that the
Gemini was line-commutated and was not equipped with all four of the standard relays. Furthermore,
its current distortion was very high (a characteristic of all line-commutated inverters, which produce
essentially a square-wave output current), and for this reason inverters of this type are no longer used for
PV-utility applications. Therefore, the Gemini is not considered further here.
The Teslaco
Figure 11 shows a circuit diagram of the Teslaco, a 4 kW, self-commutated PCU designed speci®cally for
PV applications.6,26 It is transformer isolated, using a high-frequency (20 kHz) push-pull waveshaping
stage on the transformer primary. The amplitude of the push-pull stage output is modulated by a 120 Hz
fully-recti®ed sine wave. The transformer secondary is connected to a recti®er, a low-pass ®lter, and a
120 Hz `unfolding stage' consisting of a transistor bridge which inverts every other cycle of the fullyrecti®ed sine. This results in a 60 Hz sine wave at the output of the PCU. The Teslaco's PLL, along with
an OVR/UVR, implement the converter's islanding prevention scheme, which is based on the phase-jump
detection method. The PLL generates the 120 Hz fully recti®ed sine wave reference used to generate the
switching commands for the power stage. The phase di€erence between this reference and the inverter's
terminal voltage is monitored, and if it exceeds 68 the PCU shuts down. However, the con®guration of the
Teslaco's PLL includes one additional element: a one-cycle time delay, drawn in Figure 4 using dotted
lines. Because of this delay, the PLL actually compares the phase of the input signal (va) during the present
cycle to the phase of the VCO wave form from the previous cycle. Using standard control theory, the delay
causes the PLL to have a pole in the right half of the phase plane,6 indicating that it is unstable. In the
absence of a stable voltage reference from the utility, the instability of the PLL would cause any slight
phase error to grow exponentially, eventually becoming greater than 68 and shutting down the converter.
Two studies involved simulating the behavior of the Teslaco using the ElectroMagnetic Transients
Program (EMTP)4±6 or specially-designed models and software.7 The modeling studies all indicated that
the Teslaco's islanding prevention method was e€ective; the simulated converter was found not to island
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
Islanding
protection
method
Works in multiinverter case?a
OVR/UVR
Islanding protectionb:
Under which of these conditions does the given method protect against islanding?
DQ ÿ 6ˆ 0
DP ÿ ˆ DQ ÿ 0,
purely resistive load
DP ÿ ˆ DQ ÿ 0,
resonant RLC load
Remarks
Yes
Yes
No
No
No
Small NDZ around DP ÿ ˆ 0 because thresholds
cannot be set arbitrarily small
OFR/UFR
Yes
No
Yes
No
No
Small NDZ around DQ ÿ ˆ 0 because thresholds
cannot be set arbitrarily small
PJD
Yes
No
Yes
No
No
Can be highly e€ective, but NDZ lies within that of
OFR/UFR, OVR/UVR
Harmonic
detection
Yes?
See remarks
See remarks
See remarks
See remarks
Detection not dependent on power matching, but
rather on THD of va . Threshold setting is a
problem for this method
Output variation No
(impedance
measurement)
SMS
Yes?
Yes
Yes
Yes
Yes
Noc
Yes
Yes
Yes
Fails in multi-inverter case, can cause stability/
¯icker problems for large PV systems, reduces
system eciency
Experimentally shown to be highly e€ective, but has
NDZs for low-power loads and low L, high C loads
AFD
Yes?
Noc
Yes
Yes
No
Ine€ective for low L, and high C loads that are near
unity dpf at utility frequency; requires output
current distortion
AFDPF
Yes?
Noc
Yes
Yes
Yes
Shown by simulation to be highly e€ective except
for very low L, high C loads that are near unity dpf
at the utility frequency; may produce more output
distortion than AFD
Reactance
insertion
Yes
Yes
Yes
Yes
Yes
Requires installation of equipment on utility side of
point of common coupling; cannot prevent shortterm islanding
a
Assuming all inverters to be using the same islanding prevention scheme. If multiple schemes are used, the particular combination in question would have to be analyzed. A
question mark denotes that in theory the method should be e€ective in this case, but no modeling or simulation data is available.
bThese columns describe the load circumstances under which each scheme is designed to prevent islanding.
c
Success of these methods depends strictly on the R±L±C makeup of the load; they will work for any DP ÿ if phase conditions are met.
53
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
DP ÿ 6ˆ 0
PREVENTION OF ISLANDING
Copyright # 1999 John Wiley & Sons, Ltd.
Table 1. Comparison of the islanding prevention methods discussed in this paper
54
M. E. ROPP, M. BEGOVIC AND A. ROHATGI
Figure 11. Circuit diagram of the Teslaco PCU. From (6)
inde®nitely under any condition, including the case in which DP ÿ and DQ ÿ are both zero. The longest
run-on time with a linear local load was found to be 1.948 sec (116.9 line cycles) and occurred when DP ÿ
was equal to 30% of the real power required by the load and DQ ÿ was slightly greater than zero. The
presence of a motor load in the island lengthened the run-on time, with run-on times tending to increase
as the rotational inertia of the motor increased, but in no case was the run-on time of the PCU longer than
4 sec. The present of multiple PCUs (as many as four) in the simulated island seemed to have no e€ect on
run-on times.
However, laboratory tests of the Teslaco have led to con¯icting results. Experiments done in conjunction with the EMTP simulations found that run-on times of the Teslaco in all tested cases were within
0.2±0.3 sec of those predicted by EMTP.3,6 However, other experiments showed that longer run-on times
of up to 6 sec, and even inde®nite islanding,25 were possible if the isolated system contained rotating
generators or other power conditioners with lagging dpf and capacitive compensation (like the Gemini).
This observation is important for two reasons: one, the Teslaco is susceptible to run-on if there is a large
capacitor in the island, just as the SMS and AFD-based schemes are; and two, if there are generation units
in the island which can act as voltage sources, such generators seriously impede the ability of nearly any
PV islanding protection system to safely prevent islanding. (It should be noted that in the case in which
voltage sources in the island caused the Teslaco to run on, the voltage sources were themselves islanding.)
Also, one of the Teslacos tested ran on inde®nitely due to a defective capacitor in the PLL,6 punctuating
the need for stringent quality control in PV PCUs. The ®nal word on the Teslaco method of islanding
prevention seems to be that it is e€ective under many conditions, but (a) it can be made to run on
inde®nitely with an RLC load with large C and small L, and (b) even in cases in which it works, it allows
relatively long run-on times and therefore may have diculty meeting applicable standards.10
The APCC
The APCC PCU (also called the `SunSine') was named after the American Power Conversion Corporation which produced it. A simpli®ed circuit diagram of the APCC is shown in Figure 12. This converter
utilizes a buck con®guration waveshaping stage which produces a fully-recti®ed sine wave. This output is
fed into an unfolding stage (labeled `unwrapper' in Figure 12) which unfolds the recti®ed sine into a full
sine wave across the inputs of a line-frequency transformer.
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
PREVENTION OF ISLANDING
55
Figure 12. The APCC PV PCU. This is a simpli®ed diagram, as more exact diagrams were considered proprietary.
From (6)
The APCC also uses a PLL to synchronize its output to the utility and to enable the islanding
protection function. The islanding prevention scheme used in the APCC augments phase jump detection
(PJD) with the slide-mode frequency shift (SMS) method. The Bode plot of the phase and magnitude
response of the AC line ®lter on the input to the APCC PLL are shown in Figure 13. At steady state, this
line ®lter has a slightly leading phase, meaning that the APCC operates normally with a small (28) leading
phase angle. Let fL be the phase angle of the load impedance, fi be the phase angle between the PCU
terminal voltage and output current (the phase of the input line ®lter), and fe be the phase error between
the PCU's terminal voltage and its internal voltage reference produced by the PLL. In steady state, fe is
zero and fi is about 28 leading. At the instant when the utility switch opens to form the island, fi will
become equal to fL , and fe will jump to 28 ÿ fL . The APCC controller shuts the PCU down if the
absolute value of fe becomes greater than 28. If the load were inductive, the PCU would shut down almost
instantly because fe in that case would always become greater than the 28 threshold and the PJD portion
of the APCC's islanding prevention scheme would operate. The same would be true if the load were
capacitive with a leading phase angle of greater than 48. If the load had a leading phase angle in the range
08 5 fL 5 48, the SMS part of the APCC scheme would operate, driving the phase error to a value larger
than 28 and again resulting in converter shutdown.
EMTP simulations of the APCC6 indicate that it shuts down almost immediately under all conditions,
with the longest run-on times of 927.9 msec (55.7 line cycles) occurring for the case when the local load is
slightly capacitive. The only laboratory results found for the APCC during this literature search indicate
that the APCC's anti-islanding scheme allows shorter run-on times than the Teslaco's, but can still island
for almost four and a half seconds when other power conditioners are present and the reactive
requirements of the load and all interconnected power conditioners are very closely matched, and for up
to 17 sec (1020 line cycles) if more than one APCC is present in the island.25 (It is important to note at this
point that it is unclear whether all of the papers referenced here discuss the same APCC converter.
Atmaram et al.25 state that their APCC operated at a slightly lagging power factor, whereas the one
studied by Jones et al.6 operated at a slightly leading pf, and also the converter studied by Atmaram et al.25
is described as a `transistor bridge' con®guration, not the buck con®guration shown in Figure 12.6)
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
56
M. E. ROPP, M. BEGOVIC AND A. ROHATGI
Figure 13. A plot of the phase response of the ac line ®lter used in the APCC converter's line-locking PLL, generated
using the line ®lter transfer functions given in (6)
A considerable amount of ®eld data is available for the APCC.27±29 Thirty APCC units were installed
on residences in Gardner, Mass in 1985±1986 and were carefully monitored for 10 years. During that
time, no islanding events were noted. This data suggests that, in practice, the islanding prevention method
used in the APCC works very well; that is, the conditions under which this method fails are not observed
in practice. Therefore, this study has found no experiment or ®eld case in which the APCC converter
islanded inde®nitely.
The SMS: another implementation of the slide-mode frequency shift method
The slide-mode frequency shift method as described above has been implemented in the SMS converter,
made by Sanyo Corporation and ocially certi®ed for utility-PV interface applications in Japan.30 Like
the APCC, this PCU receives its sinusoidal reference from the utility through the ®lter which has a
frequency response like that shown in Figure 13. The manufacturers refer to it as a `twin-peak bandpass
®lter' due to the shape of its magnitude response. In fact, it appears that the only di€erence between the
SMS and APCC is that the SMS actually depends on a UFR/OFR combination for PCU shutdown30 and
not directly on a phase error in its PLL. The topology of this converter was not given.
Ishida et al.30 present test results which show that the SMS shuts down in 724 msec (about 43.5 line
cycles) in the case where DP ÿ and DQ ÿ are both zero under full-load conditions. However, the described
condition is the only one under which this converter was tested, although in reality longer run-on times
could occur for other conditions,3,6 such as a lighter load as shown in Figure 8. In all likelihood, the
performance of this converter should be very similar to that of the APCC, and the previous description of
the SMS method should accurately describe this converter's behavior.
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
PREVENTION OF ISLANDING
57
CONCLUSIONS
Critical evaluation of existing islanding prevention schemes indicates that the problem of islanding in
grid-connected PV systems has not been completely solved. The summary of these methods presented in
Table I reveals that no existing islanding prevention method simultaneously meets all of the three goals
given in the Introduction of this paper. We have shown that two types of parallel RLC loads appear to
cause particular diculty: (a) RLC loads which have a near-unity dpf at the utility frequency and have a
high value of C and a low value of L; (b) low-power loads in which R is large.
However, this work also indicates that evidence exists to suggest that, in practical situations,
some existing methods actually work very well, with minuscule possibility of inde®nite run-on because
the conditions which lead to long run-on times are never practically observed or are of no consequence.4,6,25,27±30 In particular, the schemes used in both the Teslaco and the APCC appear to be quite
e€ective in practice.
The e€ect of multiple inverters on some islanding protection schemes is not completely clear.
In theory, as shown in Table I, many methods should perform equally well regardless of the number of
PCUs in the island. However, experience has sometimes shown otherwise. For example, research suggests
that the APCC allows longer run-on times in the multi-inverter case,25 indicating that SMS
may lose e€ectiveness in this case. However, simulations showed that the Teslaco did not appear to
behave di€erently when interconnected with other Teslacos.6 Table I also indicates that some of the
methods have not been tested experimentally. The multiple inverter case de®nitely warrants further
investigation.
This paper also shows that the conditions for `sucient' islanding protection are not clearly de®ned. In
order to make this de®nition clear, two problems must be resolved. First, there is a lack of consensus on
what the maximum allowable run-on time should be, which arises at least in part from a lack of
understanding of the interaction between PV systems and utility protective equipment, particularly in the
presence of high-impedance faults. This question must be investigated. Second, it is generally thought that
an islanding prevention scheme should prevent islanding under all conditions, when in fact all conditions
are not observed in the ®eld. For example, this study shows that parallel RLC loads which are near
resonance at the utility frequency and have large C and small L cause diculty for most islanding
prevention methods; however, there are very few such loads (if any) in a practical utility system which are
described by such a model. Similarly, we have presented evidence that the presence of generators which
can act as independent voltage sources renders ine€ective the islanding prevention methods presented
herein, but again in practice utilities will not allow voltage sources which themselves have no islanding
prevention mechanism to be connected to the grid. A consensus must be reached as to what loads, i.e.
range of R, L and C and load composition, are `practical' so that islanding preventing methods may be
designed to meet these targets.
Finally, in addition to the methods reviewed in this work, we have introduced a new option for
islanding protection, active frequency drift with positive feedback (AFDPF). This method has the
potential to be one of the most e€ective islanding prevention schemes available. The advantages of
AFDPF include its very small NDZ and that it remains e€ective in the multi-inverter case. Its drawbacks
are its complexity and that it can decrease the power quality of a PV system by increasing the THD in iPV .
Further research on this promising islanding prevention method is in progress.
Acknowledgements
This paper would not have been possible without the generous and carefully-considered input of the
following people: Richard Bass (Georgia Tech), Ward Bower (Sandia National Laboratories), Bill Brooks
(North Carolina Solar Energy Center), John Kennedy (Georgia power), Greg Kern (Ascension Technology), John Stevens (Sandia National Laboratories), and Robert Wills (Advanced Energy Systems).
Their contributions are gratefully acknowledged by the authors.
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
58
M. E. ROPP, M. BEGOVIC AND A. ROHATGI
REFERENCES
1. A. Kitamura, M. Okamoto, F. Yamamoto, K. Nakaji, H. Matsuda and K. Hotta, `Islanding phenomenon
elimination study at Rokko test center', Proceedings of the 1st IEEE World Conference on Photovoltaic Energy
Conversion, pt. 1, 1994, pp. 759±762.
2. A. Kitamura, M. Okamoto, K. Hotta, K. Takigawa, H. Kobayashi and Y. Ariga, `Islanding prevention
measures: demonstration testing at Rokko test center for advanced energy systems', Proceedings of 23rd IEEE
Photovoltaic Specialists Conference, 1993, pp. 1063±1067.
3. T. R. Sims, R. A. Jones and A. F. Imece, `Investigation of potential islanding problems of a line-commutated
static power converter in photovoltaic systems', IEEE Transactions on Energy Conversion, 5(3), 429±435 (1990).
4. R. A. Jones, T. R. Sims and A. F. Imece, `Investigation of potential islanding of a self-commutated static power
converter in photovoltaic systems', IEEE Transactions on Energy Conservation, 5(4), 624±631 (1990).
5. A. F. Imece, R. A. Jones, T. R. Sims and C. A. Gross, `An approach for modeling self-commutated static
power converters for photovoltaic islanding studies', IEEE Transactions on Energy Conversion, 4(3), 397±401
(1989).
6. R. A. Jones, T. R. Sims and A. F. Imece, Investigation of Potential Islanding of Dispersed Photovoltaic Systems,
Sandia National Laboratories report SAND87±7027, Sandia National Laboratories, Albuquerque, 1988.
7. G. Vachtsevanos and H. Kang, `Simulation studies of islanded behavior of grid-connected photovoltaic systems',
IEEE Transactions on Energy Conversion, 4(2), 177±183 (1989).
8. J. Stevens, `Utility intertied photovoltaic system islanding experiments', Proceedings of the 17th IEEE
Photovoltaic Specialists Conference, 1987, pp. 1134±1138.
9. P. Longrigg, `E€ects on electrical distribution networks of dispersed power generation at high levels of
connection penetration', International Journal of Ambient Energy, 3(4), 199±214 (1992).
10. IEEE-P929: Recommended Practice for utility Interface of Photovoltaic (PV) Systems, 2nd draft of IEEE
Standard P929, September (1997).
11. S. Yuyama, T. Ichinose, K. Kimoto, T. Itami, T. Ambo, C. Okado, K. Nakajima, S. Hojo, H. Shinohara,
S. Ioka and M. Kuniyoshi, `A high-speed frequency shift method as a protection for islanding phenomena of
utility interactive PV systems', Solar Energy Materials and Solar Cells, 35, 477±486 (1994).
12. German Prestandard DIN VDE 0126, `Automatic disconnecting facility for photovoltaic installations with a
rated output 44±6 kVA and a single-phase parallel feed by means of an inverter into the public low-voltage
mains', October 1997.
13. C. L. Benner and B. D. Russell, `Practical high-impedance fault detection on distribution feeders', IEEE Transactions on Industry Applications, 33(3), 635±640 (1997).
14. B. M. Aucoin and R. H. Jones, `High-impedance fault detection issues', IEEE Transactions on Power Delivery,
11(1), 139±148 (1996).
15. H. Kobayashi and K. Takigawa, `Statistical evaluation of optimum islanding preventing method for utility
interactive small scale dispersed PV systems', Proc. 1st IEEE World Conference on Photovoltaic Energy
Conversion, 1994, pp. 1085±1088.
16. M. Begovic, M. E. Ropp and A. Rohatgi, `Determining the suciency of standard protective relaying for
islanding prevention in grid-connected PV systems', Proceedings of the 2nd World Conference on Photovoltaic
Energy Conversion, Vienna, Austria, May 1998.
17. S. J. Ranade, N. R. Prasad, S. Omick and L. F. Kazda, `A study of islanding in utility-connected residential
photovoltaic systems, Part I: models and analytical methods', IEEE Transactions on Energy Conversion, 4(3),
436±445 (1989).
18. H. Kobayashi, K. Takigawa and E. Hashimoto, `Method of preventing islanding phenomenon utility grid with
a number of small scale PV systems', Proc. 21st IEEE Photovoltaic Specialists Conference, 1991, pp. 695±700.
19. D. Handran, R. Bass, F. Lambert and J. Kennedy, `Simulation of distribution feeders and charger installation
for the Georgia Tech Olympic Electric Tram System', Proc. 5th IEEE Workshop on Computers in Power
Electronics, August 11±14 1996, pp. 168±175.
20. T. E. Grebe, `Application of distribution system capacitor banks and their impact on power quality', IEEE
Transactions on Industry Applications, 32(3), 714±719 (1996).
21. R. Best, Phase-Locked Loops: Theory, Design, and Applications, 2nd ed., McGraw-Hill, New York (1993).
22. P. V. Brennan, Phase-Locked Loops: Principles and Practice, Macmillan Press, New York (1996).
23. G. Kern, `SunSine300: utility interactive AC module anti-islanding test results', Proc. 26th IEEE Photovoltaic
Specialists Conference, 1997.
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)
PREVENTION OF ISLANDING
59
24. M. E. Ropp, M. Begovic and A. Rohatgi, `Analysis and performance assessment of the active frequency drift
method of islanding prevention', accepted for publication in the IEEE Transactions on Energy Conversion, 1999.
25. G. H. Atmaram, A. H. Ayoub, J. Benton and W. Bower, `Test results of islanding experiments on gridinteractive residential power conditioners', Proc. 15th IEEE Photovoltaic Specialists Conference, 1985, pp. 1326±
1335.
26. A. Cocconi, S. Cuk, R. D. Middlebrook and T. S. Key, Design and Development of a 4 kW/20 kHz dc-Isolated
Power Conditioner for Utility-Interactive Photovoltaic Applications, Sandia National Laboratories report
SAND87-0353 UC-83, Sandia National Laboratories, Albuquerque, September 1987.
27. J. J. Bzura, `The New England electric photovoltaic systems research and demonstration project', IEEE Transactions on Energy Conservation, 5(2), 284±289 (1990).
28. J. J. Bzura, `Performance of grid-connected photovoltaic systems on residences and commercial buildings in
New England', IEEE Transactions on Energy Conversion, 7(1), 79±82 (1992).
29. J. J. Bzura, `Photovoltaic research and demonstration activities at New England Electric', IEEE Transactions on
Energy Conversion, 10(1), 169±174 (1995).
30. T. Ishida, R. Hagihara, M. Yugo, Y. Makino, M. Maekawa, A. Takeoka, R. Suziki and S. Nakano, `Antiislanding protection using a twin-peak band-pass ®lter in interconnected PV systems, and substantiating
evaluations', Proc. 1st IEEE World Conference on Photovoltaic Energy Conversion, 1994, pp. 1077±1080.
Copyright # 1999 John Wiley & Sons, Ltd.
Prog. Photovolt. Res. Appl. 7, 39±59 (1999)