Over Current Protection in PV Array Installations
Over Current Protection in PV Array Installations
Martina Calais1, Nigel Wilmot2, Andrew Ruscoe2, Oscar Arteaga2, Hari Sharma2
1
School of Engineering and Energy
Murdoch University
Murdoch WA 6150
[email protected]
2
Research Institute for Sustainable Energy (RISE)
Murdoch University
Murdoch WA 6150
[email protected]
ABSTRACT
In photovoltaic (PV) arrays with several strings in parallel certain fault conditions may
lead to potentially damaging reverse currents for PV modules. The Australian Standard
AS/NZ5033 Installation of Photovoltaic (PV) Arrays recommends fuses to protect both
cabling and PV modules in case of the occurrence of these fault conditions. The above
standard forms the basis of the development of the international IEC draft standard on
Installation and Safety Requirements for Photovoltaic (PV) Generators. This
international standard development work highlighted the need for improved
understanding of the application of fuses in PV array and this paper consequently
examines, in detail, requirements for fusing in PV arrays in view of various fault and
environmental conditions with respect to protecting both cabling and PV modules.
It reviews and discusses fusing requirements and ratings of the current PV array
installation standard. Test results of typical fuse characteristics of fuses used in PV
arrays are presented. This testing utilised both a laboratory power supply and a PV
array. The paper also examines the relationships between acceptable reverse current
levels and exposure durations of reverse currents on PV modules and compares them
with trip current of fuses and typical time delays experienced with fuse tripping.
INTRODUCTION
Over-current protection in PV arrays is made difficult by the inherent current limited
nature of PV technology. That is, PV module current only increases marginally when
going from operating to short circuit condition and hence there may not be sufficient
over-current flowing through a fuse or circuit breaker to allow it to operate. Throughout
this paper we will refer to this as the limited fault-current problem.
However, as the number of parallel interconnected modules or strings of modules
increases, potential fault currents and potential reverse module currents do as well. In
this paper, we aim to address the question of how many strings in parallel can be
connected in a PV array without over current protection and why. Furthermore we
discuss the case when we exceed a safe number of PV strings in parallel, what over
current protection is required and how do we select the appropriate string protection
device.
ISES-AP - 3rd International Solar Energy Society Conference – Asia Pacific Region (ISES-AP-08)
Incorporating the 46th ANZSES Conference
25-28 November 2008 Sydney Convention & Exhibition Centre
M.Calais, N. Wilmot, A. Ruscoe, O. Arteaga, H. Sharma
below its ISC. However, as mismatch increases, the operating point may move beyond
ISC. On the other hand, connecting a load reduces the reverse current through the shaded
module. In conclusion, the operation of a parallel connection of PV modules under
partial shading is uncritical in terms of over-current conditions, if modules of the same
type are used.
1 reflected
in v-axis
0
-1
-2
P3
-3
-4
G 2 = 0 W/m2
-5 T 2 = 25°C
-6
0 5 10 15 20 25
Voltage (V)
ISES-AP - 3rd International Solar Energy Society Conference – Asia Pacific Region (ISES-AP-08)
Incorporating the 46th ANZSES Conference
25-28 November 2008 Sydney Convention & Exhibition Centre
3
M.Calais, N. Wilmot, A. Ruscoe, O. Arteaga, H. Sharma
reverse current has now increased above the maximum operating current, it is then
possible to protect for this occurrence.
Fig. 2: Earthed PV array with two strings in parallel with earth fault. Due to the fault
and a limited DC voltage operating range of the inverter it may not continue to convert
power from the array and the top right module is then exposed to a reverse current.
It should be noted that the fault case described above investigates reverse currents in a
PV array with n strings in parallel. In large PV arrays with several sub-arrays in parallel
the analysis of fault currents and its fault current protection increases in complexity,
which is addressed by Haeberlin (2007). In this paper we limit our discussion to a PV
array with n strings in parallel.
the operating temperature may then reach up to 150°C, a value where extended
operation is considered undesirable.
Ideally for individual modules the PV module manufacturer should know the
permissible reverse current. However, this information may not be provided directly.
IEC 61730.2 (International Electrotechnical Commission 2004) has a module reverse
current test which is related to the size of the manufacturer’s recommended series fuse
size. This test passes 135% of the series fuse rating in reverse through the module. The
amount of module reverse current that can be safely carried (for hopefully limited time)
without long term module damage is then specified by the module manufacturer
through the specification of a maximum series fuse rating.
The Australian Standard AS/NZS5033 (Standards Australia 2005a) relates the number
of parallel strings without over-current protection to the module reverse current rating
(Ireverse) and the module short circuit current (ISC): For example, if
2 ⋅ I SC ≤ I reverse < 3 ⋅ I SC then the number of strings without over-current protection would
be 3. For panels not tested to IEC 61730.2 or where no manufacturer’s series fuse rating
is given, AS5033 requires over-current string protection in every string.
AS/NZS 3000:2007 (Standards Australia 2007) clause 2.4.3 sets out types of fuses
which are suitable. Essentially the selected fuses are required to meet the AS60269
series (Standards Australia 2005b). The AS/NZS60269 series provides the requirements
and defines the specification of fuses. The terminology and parameters describing fuses
used in this standard series are listed in Table 1.
To choose a suitable fuse to provide over-current fault protection, let us review what
needs to be protected and the potential fault conditions. Essentially, all the faults relate
to situations where a string becomes a “load” to the rest of the array. The easiest
situation to understand is the ground fault in an earthed array (see Fig. 2). Other
possibilities would include a short circuit in a PV string wiring or a double earth fault in
an unearthed array.
ISES-AP - 3rd International Solar Energy Society Conference – Asia Pacific Region (ISES-AP-08)
Incorporating the 46th ANZSES Conference
25-28 November 2008 Sydney Convention & Exhibition Centre
5
M.Calais, N. Wilmot, A. Ruscoe, O. Arteaga, H. Sharma
We now consider a 5 string PV array (i.e. exceeding the number of parallel strings
without protection by one string, n=5) and we then require string over-current
protection according to AS/NZS5033. We need to choose a fuse so that it will operate
in a suitable time in a fault situation. The maximum current available in a fault situation
in one string of the array is equal to (n − 1) ⋅ I SC = (5 − 1) ⋅ 5.5 A = 22 A . In Figure 3 the
“4xIsc” line represents this level.
ISES-AP - 3rd International Solar Energy Society Conference – Asia Pacific Region (ISES-AP-08)
Incorporating the 46th ANZSES Conference
25-28 November 2008 Sydney Convention & Exhibition Centre
6
M.Calais, N. Wilmot, A. Ruscoe, O. Arteaga, H. Sharma
30
Where:
I series is the manufacturers maximum
25 recommended string fuse
I reverse is the test current used in the
4xI sc module as per the reverse current test of
20 I reverse IEC61730.2
I trip_min , I trip_max is the AS/NZS5033
3xI sc recommended fuse rating range
15 I (# A) is a typically available fuse value
I (10 A)
I (12 A)
Itrip_max
I (8 A)
Iseries
The rating for the over-current protection device according to AS/NZS5033 is as per PV
module manufacturer, in our case: 15 A (see Tables 2 and 3). However, if a fuse is
based on this rating, it may not operate until optimal solar irradiance conditions are
present for a significant time. This can be seen in Figure 3 as the 4 × Isc line passes
through the If band of Iseries. As defined in Table 1, the If band in Figure 3 represents
current levels that may take 1 or more hours to operate a fuse. Further to this, PV will
only supply this fault current during the middle of the day and a low impedance fault
ISES-AP - 3rd International Solar Energy Society Conference – Asia Pacific Region (ISES-AP-08)
Incorporating the 46th ANZSES Conference
25-28 November 2008 Sydney Convention & Exhibition Centre
7
M.Calais, N. Wilmot, A. Ruscoe, O. Arteaga, H. Sharma
must exist. As the fusing current If in this example will never be exceeded, a fast
operation will not occur. Depending on the solar conditions, a fault current greater than
16.5 A is required for at least 1 hour before the fuse may operate. Obviously, if
additional strings are connected, the potential fault current increases and the levels
required to operate the fuse arise quicker.
Choosing a fuse below the maximum (e.g. 10 A as in the Table 3) can be necessary,
particularly when np is exceeded by only one or two strings. In this situation, fault
currents exceeding 11 A will be present sooner and the likelihood of the fuse operating
is higher. It can be observed in Figure 3 that the If band for I (10 A) is entirely below the
4 × Isc line and largely below the 3 × Isc line. This means that on a good solar day, If will
be exceeded during the middle of the day and will cause the fuse to operate.
FUSE TESTING
The issues described in the example above, the limited fault-current problem and the
problem that fault current levels vary and may be reduced during conventional times
were demonstrated with testing of three types of fuses at RISE.
In selecting typical fuses, we assumed ISC for a common PV panel to be around 5A. If
the manufacturer does not give any recommendation with regards to the rated trip
current of the fault current protection device (i.e. Iseries in Table 2), the maximum size
string-fuse allowed by AS/NZS5033 would be 10A (see Case 2 above) and this would
present the worst case scenario for the limited fault-current problem. Three readily
available DC fuse types were selected to approximate this scenario (see Table 4).
Part A of the fuse testing involved a programmable power supply feeding into a short
circuit via the fuse under test. Current was stepped up from 5A to several different
currents levels and the length of time taken for the fuse to operate (pre arc time) was
measured. It should be highlighted that each data point represents a different fuse and
only a single trial was used for each data point. The experimental results are compared
with the manufacturers’ specification in Figure 4.
ISES-AP - 3rd International Solar Energy Society Conference – Asia Pacific Region (ISES-AP-08)
Incorporating the 46th ANZSES Conference
25-28 November 2008 Sydney Convention & Exhibition Centre
8
M.Calais, N. Wilmot, A. Ruscoe, O. Arteaga, H. Sharma
1000
0
DII spec
100 DII exp
0 DMM spec
DMM exp
PV
10 spec
PV
0
exp
Pre Arc Time (sec)
1
0
0.
1
0.0
1
0.00
1 1 20 30 40 50 60 10
0 0
Current (Amps)
Fig. 4: Fuse testing part A - Pre arcing time for three types of fuses: Experimental
results and manufacturers’ specification.
It can be seen in Figure 4 that the manufacturers specified pre arc time was consistently
exceeded for the PVfuse but the other two fuses generally met specification (with the
exception of the DMM fuse at 20 Amps). It should be noted that all three fuses were
tested at 15A operation but only the ultra rapid fuse became open circuit at this level (in
only a few minutes). Although this complies with the fuse specifications, it highlights
the importance of fuse type selection with regard to the limited fault-current problem.
That is, 15A is equivalent to three PV strings being short circuited through a failed
fourth string in our example. Hence a 4 parallel string array with this fuse sizing is
essentially unprotected for 2 out of 3 of these fuse types.
Part B of the fuse testing was undertaken utilising the PV array at RISE on several fine
days during March 2008. The array was arranged into 4 parallel strings of 15 BP275 PV
modules in series. The same fuse selection was used as in the previous testing. It should
be noted that ISC = 4.75 A for BP275 modules and hence the selected rated fuse currents
were slightly higher than maximum allowed by AS/NZS5033 (i.e., In = ITRIP_MAX = 2 ISC
= 9.5A). One string was connected through the fuse under test into short circuit. The
ISES-AP - 3rd International Solar Energy Society Conference – Asia Pacific Region (ISES-AP-08)
Incorporating the 46th ANZSES Conference
25-28 November 2008 Sydney Convention & Exhibition Centre
9
M.Calais, N. Wilmot, A. Ruscoe, O. Arteaga, H. Sharma
three additional strings were then connected in parallel and pre arcing times were
measured with results shown in Table 5.
Tab. 5: Fuse testing part B: Fuse pre arcing time when connected to real array.
Case Fuse type Transient Pre arc time
1 SIBA PVFuse (PV) connected to 4 strings at 2:30pm No break
with I = 18.9 A and falling
2 SIBA PVFuse (PV) connected to 4 strings at Break at 12:40pm 19.8A
11:30am with I = 17.4 A and
rising
3 Bussman DMM (DMM) connected to 4 strings at 20A Break in 514s
This testing highlighted a further difficulty that a real PV array has with respect to the
limited fault-current problem. That is, real PV strings are rarely capable of supplying ISC
for a sustained period of time and hence there is a further difficulty in getting fuse
protection to operate (e.g., consider Case 1 in Table 5). The critical nature of fuse type
selection is once again highlighted comparing the ultra rapid fuse to the others.
Several important issues that have not been considered in this testing require
investigation. These include the possibility of fuses blowing at lower than expected
current because they are in a hot environment or deterioration and the possibility of
fuses blowing due to current exceeding ISC because of cloud or snow reflection
Haeberlin (2007) recommends that rated fuse currents should generally be between 1.4-
2.4 of ISC and 1.6-2.4 of ISC in regions where higher power generation due to snow and
associated light reflection can be expected. In addition, it is probable that the type of
termination and cabling that connects to a fuse will influence fuse temperature and
hence pre arc times.
In a three string array – any fault current protection used may never trip due to the
characteristics of the fuses, described previously. The cable is then sized appropriately
to safely handle the maximum current that could flow which is equal to two times the
maximum expected short circuit current (i.e. (3-1) x (Isc x 1.25) (as per requirements of
AS/NZS5033). The maximum expected short circuit current is considered within the
AS/NZS5033 standard to be 1.25 times expected short circuit current which accounts
for high incident radiation events greater than 1000 W/m2.
In a situation where np is not exceeded and there are 4 or more strings then the cable can
be sized similarly according to (n-1) x Isc x 1.25 or string fault current protection can be
ISES-AP - 3rd International Solar Energy Society Conference – Asia Pacific Region (ISES-AP-08)
Incorporating the 46th ANZSES Conference
25-28 November 2008 Sydney Convention & Exhibition Centre
10
M.Calais, N. Wilmot, A. Ruscoe, O. Arteaga, H. Sharma
used and cable sized according to that. The string over-current fault protection device
should be selected such that its rated nominal current is between 1.25 x ISC and 2 x ISC.
As the number of strings increases beyond this, the potential available fault current
increases. In these situations, with more strings, it is recommended not to exceed the
module manufacturer’s recommendation for series fuse rating. In addition, ensure that
the maximum fault current is a level that is greater than the conventional fusing (If)
current for the selected fuse. In such a fault situation it will follow that potentially
damaging currents in the PV array string are stopped by the fast operation of the
selected fuse.
When selecting fuses, careful consideration should be given to the manufacturer’s pre
arc time data. Experiments on this have shown that this fuse characteristic can vary
considerably between fuse types and this has important ramifications for the limited
fault-current problem. It should also be noted that manufacturer data seemed overly
optimistic for one type of fuse tested. It is hoped that this is an unusual occurrence.
In selection of fuses other factors (not detailed here) should also be considered to ensure
selection of correctly rated devices for the PV array installation:
ISES-AP - 3rd International Solar Energy Society Conference – Asia Pacific Region (ISES-AP-08)
Incorporating the 46th ANZSES Conference
25-28 November 2008 Sydney Convention & Exhibition Centre
11
M.Calais, N. Wilmot, A. Ruscoe, O. Arteaga, H. Sharma
REFERENCES
Haeberlin, H. (2007): Photovoltaik - Strom aus Sonnenlicht fuer Verbundnetz und
Inselanlagen, AZ Verlag, CH-5001 Aarau, ISBN 978-3-905214-53-6, (in German:
Photovoltaics: Solar energy for grid connected and stand alone systems). Chapter 4:
Solarmodule und Solargeneratoren (Photovoltaic Modules and Photovoltaic Arrays).
p. 121-196.
ISES-AP - 3rd International Solar Energy Society Conference – Asia Pacific Region (ISES-AP-08)
Incorporating the 46th ANZSES Conference
25-28 November 2008 Sydney Convention & Exhibition Centre
12