FMECA Technique on Photovoltaic Module

Marcantonio Catelani1, Lorenzo Ciani1, Loredana Cristaldi2, Marco Faifer2, Massimo Lazzaroni3, Paola Rinaldi4

1
Dipartimento di Elettronica e Telecomunicazioni - Università degli Studi di Firenze - Firenze, Italy
marcantonio.catelani@unifi.it - lorenzo.ciani@unifi.it
2
Dipartimento di Elettrotecnica - Politecnico di Milano - Milano, Italy,
loredana.cristaldi@polimi.it - marco.faifer@polimi.it
3
Dipartimento di Tecnologie dell’Informazione - Università degli Studi di Milano - Crema (CR), Italy
massimo.lazzaroni@unimi.it
4
Dipartimento di Elettronica, Informatica, Sistemistica - Università degli Studi di Bologna- Bologna, Italy
prinaldi@deis.unibo.it

Abstract — The solar photovoltaic industry has seen rapid use conditions. Maintainability is defined, instead, as the
expansion in the past decade with an ever-increasing share of the ability of the item under given condition of use, to be retained
electricity-generating capacity for the world. For the emerging in, or restored to, a state in which it can perform a required
photovoltaic (PV) industry the assessment of the quality and function, when maintenance is performed under given
reliability of its products is becoming more and more important. conditions and using stated procedures and resources. It’s easy
To this aim and to ensure an optimal design and material choice, to understand that the evaluation of reliability and
a failure modes, effects and criticality analysis methodology maintainability performances can be made if are known the
(FMECA) to classify the occurrence, the severity and the impact failure conditions for the equipment under test.
of all possible failure mechanisms on the PV system has been
introduced. This helps to eliminate or reduce the impact of In order to maximize the aforementioned characteristics,
potential failure modes before the completion of the design and failure event would be taken into account, studied and
before failures occur in the field. By means of this analysis it can predicted. Many techniques can be used in order to
be noticed that a crucial aspect in PV systems is the cleaning characterize failure events. In particular, we refer to Failure
status of the panel surface. In this paper this problem has been Modes and Effects Analysis (FMEA) and the Failure Modes
analyzed by means of an experimental activity. Finally a method Effects and Criticality Analysis (FMECA), because before
for the assessment of the PV panel condition has been proposed. trying to eliminate or even reduce the PV degradation and
failure modes, a thorough understanding of their origin and
Keywords – FMEA, FMECA, Dependability, Reliability,
behavior is required [4]-[6]. On the basis of the knowledge of
Failure mode, Fault, Diagnostic, Perceived Solar Radiation Level.
the failure modes and failure mechanisms both techniques are
I. INTRODUCTION able to give improvements of Maintainability and Safety
performances of the equipment under test.
Nowadays an increasing worldwide interest in sustainable
energy production and use is indisputable. To this aim The paper is organized as follows. In Section II a brief
different energy sources that will have a role in distributed overview of the FMECA analysis is given. Successively, in
generation and sustainable energy systems can be now Section III the FMECA technique applied on a PV panel will
considered such as photovoltaic (PV) panels, fuel cells stack, be described. Based on the results of this analysis, in Section
and batteries of various chemistries. In this paper PV panels IV and Section V the measurement set-up and the test phase
are taken into account as promising application in alternative are discussed and, finally in the conclusions are presented. A
energy source. In fact in more recent years the number of plant brief review of the main concepts on FMECA will be also
based on the use of PV panels are increased and, reported.
consequently, the number of installed PV panels increased too.
Although photovoltaic modules are a very reliable source of
electrical energy, field results [1], [2] indicate that the modules II. BASIC CONCEPTS ON FMECA
can fail or degrade in a number of ways. In this new scenario This section presents a brief review of some concepts
it’s fundamental to guarantee the functionality of both the concerning FMECA analysis technique [7]-[9]. Such method,
plant and the PV panels but also, in the same time, the treated in the technical standards edited by the International
reliability and maintainability performances. According to the Electrotechnical Commission (IEC) and in particular by the
Standards [3], the reliability can be considered as the ability of Technical Committee IEC TC 56 – Dependability, can
an item – in our case the PV panel – to maintain its adequately give both qualitative and quantitative evaluations
functionality in the time, under specified environmental and about failure modes. The methods of analysis that will be

978-1-4244-7935-1/11/$26.00
Authorized ©2011 IEEE FEDERAL DO PARA. Downloaded on October 17,2023 at 22:21:18 UTC from IEEE Xplore. Restrictions apply.
licensed use limited to: UNIVERSIDADE
discussed in this section allow evaluating the failure modes to Risk Priority Number (RPN), given by the following:
which a system is or could be subjected to. Furthermore, the
results obtained allow the designer to identify any
RPN = S ⋅ O ⋅ D
modification which must be made in order to improve the where S (Severity) represents the estimate of how strongly
RAMS (Reliability, Availability, Maintainability and Safety) the effects of a failure impacts on the system or user
requirements of the product, item or system. The potentialities (personnel or customer for example) and it’s related to the
of the FMECA technique in the electrical and electronics defined boundaries of the analyzed system; O (Occurrence)
application field are shown also in other several cases [10-12]. represents the probability that a failure mode will be
FMECA is an inductive analysis method that starts from manifested in a determined time that usually coincides with
the lowest level, for example the failure of a single component the useful life of the component under examination. At this
(mechanical, electrical, etc.) that has an effect on the entire aim, it is necessary to have access to detailed information
system. In such a case, a profound and detailed knowledge of regarding the reliability of components/devices utilized, for
the system and its structure is required to study failures. example the failure rate or field data feedback. The parameter
Inductive methods are generally rather stringent and well D (Detection) is the estimate of the possibility of
designed to identify all the individual failure modes. An identifying/diagnosing and eliminating/preventing the onset of
analysis of this type gives its best results when conducted in a breakdown before its effects are manifested on the system or
the final planning or design phase although it can be profitably personnel. This number is usually ranked in reverse order
used even in different phases of the design process. from the severity or occurrence numbers: the higher the
detection number D, the less probable is the possibility of
FMECA technique permits to analyze a system in order to identifying the failure and vice versa. Starting from these
identify potential failure modes (the manner in which an item considerations the lower probability of detection leads to a
fails), their cause and effect (the consequence of a failure higher RPN; this indicates the necessity to resolve the failure
mode) on performance and, when applicable, their effect on mode with maximum priority and speed.
the safety of people, on the environment and on the system.
Finally the concept of criticality is introduced. Criticality is a For each of these indices (S, O, D) it is necessary a
way to quantify the attention it is opportune to give to a detailed analysis in order to identify their appropriate values,
determined failure, event or non-conformity and depends both related to the type of application and the operating
on the probability of its occurrence and the gravity of the environment. The indices assessment for the PV module under
consequences it may have. The attention to dedicate to an examination can be summarized as shown in Table I.
event depends first of all on the effect it may have on the
TABLE I. CRITERIA FOR THE OCCURENCE, SEVERITY AND
safety of personnel, on the damage it can cause with DETECTABILITY OF POTENTIAL HAZARDS
subsequent losses and its effect on the availability of service.
It is rather difficult to define a generally valid criterion to Occurrence Severity Detection Score
evaluate criticality because the concept of the seriousness of (O) (S) (D)
the consequences and their probability of their happening Improbable: Automatic
come into play. The seriousness level can vary and can be Unlikely to occur detection
evaluated differently if the objective, for example, regards the Insignificant 1-2
in the life of the with
safety of personnel, damage and relative losses, or the system warning
availability of service. Criticality is defined by means of a Remote: Automatic
scale of values that allows evaluating the seriousness of Marginal:
Occurs 1 time in detection
consequences in function of the criteria taken into Low produced 3-4
consideration. In [8] a classification with four principal levels a maintenance without
power decreasing
of the gravity of consequences is presented. However, it would cycle warning
be noted that different levels may also be used in function of Occasional:
Critical: Detected
the specific application under investigation. Occurs 2 times in
High produced by the 5-6
a maintenance
Different types of critical failure modes can be identified. power decreasing operator
cycle
A scale of criticality based on the following categories is Probable: Impossible
generally valid (as described in [8]): a) death or injury to the Very critical:
Occurs 3 times in to detect
public or company personnel; b) damage to this or other temporary block of 7-8
a maintenance by the
equipment; c) economic damage deriving from loss of output the PV system
or loss of system functions and finally, d) inability to perform cycle operator
a function due to inability of equipment to properly perform Frequent:
its principal function. The choice of criticality categories Likely to occur at
Catastrophic:
requires careful study and prudence. It is necessary to take into least once a Impossible
possible injuries of 9-10
account all the factors that have an impact on the evaluation of month in the to detect
the user or operator
the system, its performance, costs, programs, safety and risks. operating life of
the system
A criticality analysis is further performed to enable a
priority ranking among the identified failure modes and
effects. This ranking is done using a quantitative index, called

Authorized licensed use limited to: UNIVERSIDADE FEDERAL DO PARA. Downloaded on October 17,2023 at 22:21:18 UTC from IEEE Xplore. Restrictions apply.
 III. DISCUSSION ABOUT FMECA ON PV SYSTEMS type of maintenance to be applied. It would be noted that a
Renewable energy systems are fairly reliable, but like any condition based maintenance (CBM) program can be very
complex system, they may fail, and the effects of failures interesting in this situation. In fact, in order to take into
should be analyzed and accounted for. Based on this topic, this account the necessity to remove the powder on the panel, for
section presents an assessment of the failure modes, effects example, different kinds of approaches are possible:
and criticality of a PV system. Table II shown an abstract of • Maintenance based on the time: once or twice a year the
FMECA analysis results and the major failure modes are panels are cleaned; this way to proceed is not able to get
analyzed and classified in terms of criticality. This analysis the best performance due to the fact that the cleaning
provides critical input for future improvements in module operation is performed without take into account the
reliability and it provides important data on the long term wear actual situation of the panel surface. For example the
out or failure of PV modules. As already said, the RPN was operation is performed after rain time when a natural way
used to prioritize all potential failure modes to decide upon to remove the powder has operated. This way to operate is
actions that would lower the risk by reducing the likelihood of similar to an open loop control system.
occurrence and improving controls for detecting the failure. • Maintenance based on a well-defined condition. In this
Furthermore, considering the most critical hazard, based on case a monitoring activity is necessary. Monitoring of the
RPN value, it is possible to plan a series of tests on the PV plant or panel efficiency can give information concerning
module under examination in order to optimize its when it is opportune to clean the panel surface. It would
performances during the installation on the field. be highlighted that the maintenance is, sometimes, a very
In particular, the obtained results suggest a good expansive operation both form economical point of view
maintenance activity of the panels. In fact the higher RPN and pollution point of view. This way to operate is similar
value is related to the presence of powder on the panel and its to a closed loop control system.
effects on the whole PV system. But also other high RPN On the base of these considerations in the following the
values can be decrease by means of an optimize planning of problem of powder on panel surface will be analyzed. In
maintenance: for example those failures due to clamp/terminal particular the severity of the problem and a possible method of
damaged or PV cell fracture. A further suggestion regards the identification will be discussed.

TABLE II. AN EXTRACT OF THE PV MODULE FMECA ANALYSIS

Possible Compensating
Failure Detected Local Final
It Failure provision S O D RPN Recommendation
mode symptom Effect effect
cause against failure

Non Sandwich closing Aging of Produced Support 4 3 5 60

conformity of in not correct way elements inside power structure
front glass the sandwich and decreasing maintenance
(scratch , reduction of the
bubble, long term
intrusion) reliability

Open circuit PV cell fracture Produced power Produced Annual 5 4 5 100 Laying care
lowering power inspection of
decreasing the PV module
Aging of contact Absorbed power Produced Spear parts 4 4 5 80 Periodic
between PV panel lowering power change maintenance of
decreasing the contacts

Short circuit Misalignment Out of service of Produced 6 4 6 144

between PV cell a part of the PV power
Module

string module decreasing

Heart failure Sandwich Protection Plant out of Produced Periodic 5 3 6 90
insulation fault systems service power inspection of
nulling the plant
Hot-spot Welding not Junction Produced PV panel 5 3 6 90
correct – perforation with power installation in
Shadowing and localized decreasing area without
by-pass diode overheating and continuous
failure PV module shadowing
damage
Powder Atmospheric Produced power Produced Structure 6 7 6 252 Monthly cleaning
event decreasing power maintenance
decreasing
Connection Clamp/Terminal Produced power Produced Spear parts 5 5 6 150
box defective damaged decreasing power change
decreasing

Authorized licensed use limited to: UNIVERSIDADE FEDERAL DO PARA. Downloaded on October 17,2023 at 22:21:18 UTC from IEEE Xplore. Restrictions apply.
 IV. EXPERIMENTAL SETUP panels performances we are interested in. In fact since an
As stated before, the FMECA shows that the maximum analysis based on a comparison will be done, as
value of RPN is related to the presence of powder on the panel aforementioned, the crucial characteristic of the light source is
surface. By increasing the powder a decrease of the generated its repeatability and stability.
power can be registered. Starting from the results obtained by Another important characteristic of the developed setup is
the FMECA, an experimental activity has been carried out. In its capability to guarantee a constant temperature on the
particular the aim of the experimental activity was to directly surface of the analyzed panel. This is key because, as well
quantify the effect of the powder on the PV panels known, the PV panel performance heavily depends on its
performances. In Fig. 1 the utilized experimental setup is temperature [13].
shown.
250
240
A
PV V Load
230
220

Radiation (W/m2)
210
200
190
Fig. 1: Experimental setup for measurement on PV.
180
The core of the experimental setup is the light source that 170
has been properly designed. In fact, in order to analyze the PV 160
panel performances degradations a repeatable and stable light 150
source must be used. For this reason a luminous source made -10.5 0 10.5
up by a matrix of 3 × 3 LED array has been designed and Position (cm)
assembled. In particular 9 arrays of 4 LED Oslon Star White PV Panel active area
by OSRAM have been used. This source allows achieving, in
a quite easy way, a repeatable output; in fact it can be shown
that by properly controlling the supply current and the cooling Center of PV panel
system these components feature a good emitting stability. It Fig. 3: Radiation distribution.
must be noted that the light spectrum obtained by this source
is quite different by the one of the sun (Fig. 2).
1.0
V. EXPERIMENTAL RESULTS
0.9
0.8 With the experimental setup depicted in Fig. 1, two PV
0.7 panels in different cleaning conditions have been compared. In
Power ratio

0.6 particular panels built with polycrystalline technology have

0.5 been analyzed (TABLE III). The panel dimensions are
0.4 24.5×23.2 cm while its active area is 21 × 18.5 cm.
0.3
TABLE III. : PV PANEL CHARACTERISTICS
0.2
0.1 Pmax 5W
0.0 Vpm 17.5 V
350 450 550 650 750 Ipm 0.285 A
Wavelenght (nm) Voc 21.3 V
Fig. 2: Power spectrum of the LED.
Isc 0.31 A

However, for the analysis we want to perform it is more A first analysis has been done by evaluating the effect of
important to guarantee the repeatability and the reproducibility the ambient pollution on the PV panel performances. For this
of the measurements. The source has been characterized analysis the panel PV1 has been used. In Fig. 4 and Fig. 5 the
mapping its emission by means of a radiometer (NR01 - I-V curve and the P-V curve obtained with the clean PV panel
Hukseflux). The source light radiation at a distance of 28 cm and after a 15 days long exposition to the outside pollution are
is shown in Fig. 3. By analyzing the light source emission, it reported (average values of the air pollution are: PM10 equal
can be noticed that the radiation it is not perfectly uniform. In
to 46 μg/m3 and PM2.5 equal to 38 μg/m3 [14]). The little
fact, if an area of 21×18.5 cm is considered, corresponding to knee on I-V characteristic is due to the not perfect uniformity
the active area of the PV panels that have been employed for of the light source. However, it can be noticed that this non
this analysis, a maximum variation of 30% has been measured ideality of the source does not affect the comparison: the PV
(Fig. 3). The maximum variation is measured at the right side panel behavior in the two conditions is compatible. By means
of the aforementioned figure. Also this not ideality, with of this comparison it can be shown that even a short exposition
respect to the sun, is not crucial for the evaluation of the PV of the PV to pollution affects its performances. In fact a

Authorized licensed use limited to: UNIVERSIDADE FEDERAL DO PARA. Downloaded on October 17,2023 at 22:21:18 UTC from IEEE Xplore. Restrictions apply.
reduction of 1.6% of the generated power, at the MPP, has 45
Increasing Powder quantity
Clean
been measured. 40

A further investigation has been performed by using the 35 Powder 4 g/m2

second PV panel (PV2). In this case, measurements have been 30
performed after having applied different concentration of

Current (mA)
Powder 20 g/m2
powder on the PV panel surface. The experimental results are 25

plotted in Fig. 6 and Fig. 7: as expected the higher the 20

concentration of the powder the lower the generated power. In Powder 40 g/m2
15
order to highlight the effects of the powder on the MPP, its
reduction with respect to the powder concentration has been 10
plotted in Fig. 8. 5
45 0
0 5 10 15 20
40
Voltage (V)
35
Fig. 6: PV2 panel: experimental results for cleaned and not cleaned panel.
30
Current (mA)

25
20
600 Powder 4 g/m2
15 Clean

10 15 days after cleaning 500

5 Clean
0 400
Power (mW)

0 5 10 15 20
Voltage (V) 300 Powder 20 g/m2
Fig. 4: I-V characteristic of PV1 panel.

600 200
Powder 40 g/m2
15 days after cleaning
500 Clean 100

400
Power (mW)

0
0 5 10 15 20
300 Voltage (V)

200 Fig. 7: PV2 panel: experimental results for clean and not clean panel.

100 0
-5
-10
MPP Reduction %

0
0 5 10 15 20 -20
-27
Voltage (V)
-30
(b) -40
Fig. 5: P-V characteristic of PV1 panel. -51
-50
An interesting analysis can be performed by analyzing the
-60
plot reported in Fig. 9. In this figure the characteristic solar 0 5 10 15 20 25 30 35 40 45
radiation with respect to MPP and concentration of powder Powder concentration (g/m2) [0 = clean PV]
with respect to MPP, with a constant solar radiation of
240 W/m2, are reported. This plot allows to evaluate the
cleaning status of the PV panel, taking into account both MPP Fig. 8: PV2 panel: plot of MPP Reduction Vs powder concentration.
and solar radiation values.

Authorized licensed use limited to: UNIVERSIDADE FEDERAL DO PARA. Downloaded on October 17,2023 at 22:21:18 UTC from IEEE Xplore. Restrictions apply.
 0 VI. CONCLUSIONS
230
5
210 In this paper an analysis of the PV system fault modes is
10
190
presented and discussed. The analysis has shown that the
Radiation (W/m2)

15 higher RPN value is related to the presence of powder on the

Powder (g/m2)
Radiation
170
150
20 panel surface. For this reason an experimental activity oriented
25 to a further evaluation of this problem has been performed.
130
30 The obtained results confirm the severity of this problem.
110
35
Moreover a proposal of techniques for the continuous
90
Powder evaluation of the PV panel efficiency has been also presented.
70 40
50 45 REFERENCES
250 300 350 400 450 500 550
[1] S. E. Forman, “Performance of experimental terrestrial photovoltaic
MPP (mW)
modules,” IEEE Transactions on Reliability, vol. R-31, pp. 235–245,
1982.
Fig. 9: Radiation vs MPP and powder concentration vs MPP (computed with a [2] L. N. Dumas and A. Shumka, “Photovoltaic module reliability
radiation of 240 W/m2) (PV2). improvement through application testing and failure analysis,” IEEE
Trans. Reliability, vol. R-31, pp. 228–234, 1982.
The MPP reduction due to the powder leads to the [3] IEC 60050-191 - International Electrotechnical Vocabulary. Chapter
definition of the Perceived Solar Radiation Level (PSRL). 191: Dependability and quality of service.
[4] Meyer, E.L.; van Dyk, E.E.; "Assessing the reliability and degradation
When the panel is clean the PSRL is equal to the actual solar of photovoltaic module performance parameters," IEEE Transactions on
radiation while the increase of pollution on the panel surface Reliability , vol.53, no.1, pp. 83- 92, March 2004.
causes a decrease of the PSRL. This reduction can be [5] Kuznetsova, V.A.; Gaston, R.S.; Bury, S.J.; Strand, S.R.; “Photovoltaic
evaluated by analyzing the MPP as shown in Fig. 9. On this reliability model development and validation” Proc of 34th IEEE
Photovoltaic Specialists Conference (PVSC), 2009, Page(s): 432 – 436.
plot it can be seen that for a given value of MPP two [6] Collins, E.; Dvorack, M.; Mahn, J.; Mundt, M.; Quintana, M.;
conditions can be defined: clean panel with a given value of “Reliability and availability analysis of a fielded photovoltaic system”,
solar radiation or dirty panel with a higher value of solar Proc of 34th IEEE Photovoltaic Specialists Conference (PVSC), 2009,
radiation. For example, for the analyzed panel PV2, a value of Page(s): 2316 – 2321.
MPP of 350 mW can be obtained with a solar radiation of [7] U.S.A. Department of Defence, MIL-STD-1629 Procedures for
Performing a Failure Mode, Effects and Criticality Analysis, 1980.
170 W/m2 and a clean panel or a solar radiation of 240 W/m2 [8] IEC 60812, Analysis techniques for system reliability – Procedure for
and a panel on which surface a powder concentration of about failure mode and effects analysis (FMEA), 2006.
27 g/m2 is present. Actually the effective solar radiation is [9] SAE J1739, Potential Failure Mode and Effects Analysis in Design,
70 W/m2 lower than the actual solar radiation. 2009.
[10] Seung J. Rhee, Kosuke Ishii, “Using cost based FMEA to enhance
Starting from these considerations, in order to evaluate the reliability and serviceability”, Advanced Engineering Informatics,
efficiency of the panel, for each working temperature it is Volume 17, Issues 3-4, July-October 2003, Pages 179-188
[11] Sachin Kumar, Eli Dolev, Michael Pecht, “Parameter selection for health
possible to define the relation solar radiation – MPP. These monitoring of electronic products”, Microelectronics Reliability,
curves can be used for estimating the value of PSRL and Volume 50, Issue 2, Feb ruary 2010, Pages 161-168.
therefore the loss of performances of the panel. In fact by [12] P. Chaparala, Li Erhong, S. Bhola, “Reliability qualification of
knowing the MPP of the panel, the PSRL can be estimated. photovoltaic smart panel electronics”, Proc. 17th IEEE International
Now by considering the actual value of radiation, obtained by Symposium on the of Physical and Failure Analysis of Integrated
Circuits (IPFA), 2010 , pp. 1-4, 5-9 July 2010.
means of a radiometer, the value of MPP of panel perfectly [13] T. Markvart, L. Castañer, Practical handbook of photovoltaics.
clean can be defined. The difference between the two values Fundamental and applications, Elsevier, Oxford, 2003.
of MPP is an index of the reduction of performances. [14] http://ita.arpalombardia.it/ITA/qaria/doc_RichiestaDati.asp
Therefore on the base of this index a CBM approach can be
implemented for the PV module.

Authorized licensed use limited to: UNIVERSIDADE FEDERAL DO PARA. Downloaded on October 17,2023 at 22:21:18 UTC from IEEE Xplore. Restrictions apply.