Professional Documents
Culture Documents
5CV.3.49 Version6
5CV.3.49 Version6
5CV.3.49 Version6
net/publication/344194402
CITATIONS READS
5 2,755
4 authors:
All content following this page was uploaded by Eduardo Abdon Sarquis Filho on 20 October 2020.
Eduardo A. Sarquis Filho, Andrés A. Zúñiga, João F. P. Fernandes, Paulo J. Costa Branco
IDMEC, Instituto Superior Técnico, Universidade de Lisboa,
Av. Rovisco Pais, 1049-001 Lisboa, Portugal
ABSTRACT: The present work aims to gather, analyze and organize the information available in the literature about
failure modes and failure rates in photovoltaic systems, mapping their origins and conditions under which they were
obtained to provide unambiguous and trustworthy information for reliability studies on general PV systems. A set of
guiding questions have been listed to clearly detail the original context of each publication. Filtering the references
with failure information based on field data, the most common failures registered in PV systems were identified. With
this information, a list has been created containing the failure rates for the major components in the PV system:
transformer, inverter, and PV array. In particular, the failures in the PV module are detailed further according to its
internal components and failure modes. This information can be used in reliability studies regarding the general
behavior of a PV system.
Keywords: PV System, Reliability, Failure Rate.
3 TYPICAL FAILURES OBSERVED IN THE FIELD Table III: Failure rate per unit per hour for the
components and sub-components in the PV system.
To understand the typical failures of a PV system and
its respective failure rates, references that present a large Component F.R. Ref.
number of systems under a diverse or moderate climate Sub-Component (failures/unit-h)
were selected and analyzed. In Table II, references [1], [2], PV Array
[3], [4], [7] and [9] fit in this profile. Mounting Struct. (per string) 0.845x10-6 [3]
Mounting Structure 0.101x10-6 [7]
Table II: Summary of references with field data and their PV Module 0.065x10-6 [3]
respective failure data collection characteristics. PV Module 0.0152x10-6 [7]
PV Module 0.025x10-6 [9]
Ref. #Systems Size Age PV Module 0.035x10-6 [9]
[1] 772 Diverse 2.7 average PV Module 0.04x10-6 [9]
[2] 350 Few kW to 70 MW Up to 7 years PV Module Connector 0.0056x10-6 [9]
[3] 63 0.2 MW to 10 MW No info PV String cabling 0.845x10-6 [3]
[4] 57 No info No info PV String cabling 0.002x10-6 [9]
[5] 11 10 kW to 600 kW 4 to 18 years Fuses 2.28x10-6 [3]
[6] 1 115 kW 5 years Fuses 0.063x10-6 [9]
[7] 1242 Avg. 26 kW Up to 14 years Breaker 6.075x10-6 [9]
[8] 1 4.6 MW No info Inverter
[9] No info 1.4 MW to 3.4 MW 2 to 7 years Generic – 3 kW 16.3x10-6 [3]
[10] 1 20 kW No info Generic – 30 kW 65.1x10-6 [3]
Generic – 100 kW 217x10-6 [3]
Table II (cont.): Summary of references with field data Generic – 26 kW 11.2x10-6 [7]
and their respective failure data collection characteristics. Generic – Central type 74.0x10-6 [9]
Generic – Central type 130x10-6 [9]
Ref. Location Climate Data Collection Generic – String type 15.1x10-6 [9]
Capacitor 17.8x10-6 [9]
[1] Europe Diverse Up to 2016
Capacitor 41.5x10-6 [9]
[2] Worldwide Diverse Jan/2010 to Mar/2012
[3] Italy and Spain Medit. 2012 to 2016 Contactor 8.31x10-6 [3]
[4] Italy Medit. No info Ctrl & Communication board 26.7x10-6 [9]
Ctrl & Communication board 63.7x10-6 [3]
[5] Arizona, US Hot dry 2009 to 2013
Cooling fan 26.7x10-6 [9]
[6] Serbia Moderate No info
[7] Japan No info 2006 to 2012 IGBT module 16.6x10-6 [3]
[8] Arizona, US Desert 2003 to 2007 IGBT module 8.9x10-6 [9]
Relays 2.77x10-6 [3]
[9] No info No info No info
Transformer 17.8x10-6 [9]
[10] Langley, Canada Moderate 2010
Table IV: Failure rate per unit per hour for the failure failure rates within the same order of magnitude, except
modes of the sub-components in the PV module. for the mounting structure, PV string cabling and fuses.
The difference between failure rates reported for the
Sub-Component P.O. F.R.* Ref. mounting structure can be correlated with the system size.
Failure Mode (%) (failures/unit-h) The rooftop PV systems reported in [7] present lower
PV cells 17.1% 11.1x10-9 [2] failure rate for the mounting structure than the PV power
Cell cracks 0.23% 0.147x10-9 [1] plants reported in [3]. In addition, maintenance inspections
Corrosion 1.21% 0.788x10-9 [1] are usually more detailed and frequent in PV power plants
Hot-Spot 3.28% 2.13x10-9 [1] than in rooftop PV systems. More inspections tend to
PID 8.28% 5.38x10-9 [1] increase the number of failure records.
Snail track 7.14% 4.64x10-9 [1] The differences in the values for string cabling and
Front Glass 58.5% 38.1x10-9 [2] fuses in references [3] and [9] are significantly large
Glass breakage 10.8% 7.02x10-9 [1] (respectively 400 and 36 times), but there is no clear
Encapsulation explanation for this discrepancy.
Delamination 3.95% 2.57x10-9 [1] Regarding the PV module subcomponents and their
EVA discoloration 19.4% 12.6x10-9 [1] failure rates presented in Table IV, the most frequent
Defective back sheet 4.88% 3.17x10-9 [1] failure modes observed in the PV module are EVA
Defective back sheet 11.4% 7.42x10-9 [2] discoloration, glass breakage, PID, Snail Track and
Junction-box 8.57% 5.57x10-9 [2] defective back sheet.
Corrosion 0.09% 0.057x10-9 [1] Although the modules have a lower failure rate than
Overheat 1.65% 1.07x10-9 [1] the inverters, the number of panels far exceeds the number
Bypass diode 4.28% 2.78x10-9 [2] of inverters in any photovoltaic installation. Therefore, the
Diode failure 1.61% 1.05x10-9 [1] comparison is fairer if made between the failure rates
normalized by the equipment nominal power. The failure
*Based on PV module generic failure rate of 0.065x10-6.
rates associated to the PV modules range from 0.08 to
In Table III, for better comprehension, the components 0.36x10-6 failures per kWp per hour. At its highest values,
this range overlaps with the inverter failure rate range
in the DC side are aggregated as subcomponents of the PV
which is 0.25 to 3.34x10-6 per kWp per hour.
array. Failure rates associated with major components are
considered a generic failure of the component.
Although references [1] and [2] do not contain explicit
6 FINAL CONSIDERATIONS
failure rate values, the percentage of occurrence provided
for the PV module sub-components in [2] and its failure
modes in [1] can be used to derive the rate of occurrence The collection of failure rates with a clear
for each failure mode based on the failure rate of the understanding of the context under which the PV system
has been operating is critical for reliability. In this work,
component.
the authors have endeavored to gather the most relevant
The conversion of the percentage of occurrence to
failure rate can be done through a simple multiplication. publications for the representation of typical failure
However, this can only be applied if the failure modes behavior in PV systems in general. The result is a list of
failure rate values derived from field data of hundreds of
considered in the definition of the PO are the same failure
PV systems.
modes considered in the calculation of the failure rate. In
the reviewed studies, there is no clear information about PV systems are subject to several types of failures, and
the failure modes related to these failure rates, but it is a general collection of field records of failures helps
understand what the most relevant in terms of occurrence
reasonable to assume that all permanent failures were
are. The failure rates filtered in this paper can be used in
considered. Subsequently, we have filtered only the
permanent failure modes in the PV module from [1] and reliability studies regarding the general behavior of a PV
defined the percentage of occurrence to be multiplied by system.
the PV module failure rate to find the failure mode rate of In the process, a set of guiding questions have been
listed to clearly detail the context under which the failure
occurrence. Table IV summarizes the values obtained.
data is collected. Some references do not have enough
information to answer all the questions presented. So,
5 DISCUSSION these questions also intend to highlight what are the
relevant information to be provided along with the failure
data to complete characterize its context.
According to the collection of failure rates listed in
Table III, the highest number of failures per unit-hour The work performed in this paper is the first part of a
among the PV system components is related to the inverter failure mode and effect analysis (FMEA) that will also
evaluates the impact of these failures in a PV system
and its internal components. The larger the inverter, the
performance as well as the ease of their detection using the
higher the failure rate, which can be explained by the
higher complexity of the circuit and the higher number of common inspection tools available.
internal components. The subcomponent in the inverter
that has the highest failure rate is the control and
7 REFERENCES
communication board, followed by the capacitors, the
cooling fan, the IGBT module, the contactor, and at last
the relays. [1] D. Moser, “Technical Risks in PV Projects, Report
on Technical Risks in PV Project Development and
The PV module has five failure rate values from three
PV Plant Operation,” p. 139, 2016.
studies, and all values are quite consistent with each other.
Same for the other components in Table III that have [2] A. Golnas, “PV system reliability: An operator’s
perspective,” in 2012 IEEE 38th Photovoltaic
Specialists Conference (PVSC) PART 2, Jun. 2012, Renewable Energy, vol. 36, no. 9, pp. 2334–2340,
pp. 1–6, doi: 10.1109/PVSC-Vol2.2012.6656744. Sep. 2011, doi: 10.1016/j.renene.2011.01.036.
[3] S. Gallardo-Saavedra, L. Hernández-Callejo, and O. [16] L. Cristaldi, M. Khalil, and P. Soulatintork, “A root
Duque-Pérez, “Quantitative failure rates and modes cause analysis and a risk evaluation of PV balance of
analysis in photovoltaic plants,” Energy, vol. 183, pp. system failures,” ACTA IMEKO, vol. 6, no. 4, p.
825–836, Sep. 2019, doi: 113, Dec. 2017, doi: 10.21014/acta_imeko.v6i4.425.
10.1016/j.energy.2019.06.185. [17] M. Perdue and R. Gottschalg, “Energy yields of
[4] M. Villarini, V. Cesarotti, L. Alfonsi, and V. Introna, small grid connected photovoltaic system: effects of
“Optimization of photovoltaic maintenance plan by component reliability and maintenance,” IET
means of a FMEA approach based on real data,” Renewable Power Generation, vol. 9, no. 5, pp. 432–
Energy Conversion and Management, vol. 152, pp. 437, 2015, doi: 10.1049/iet-rpg.2014.0389.
1–12, Nov. 2017, doi: [18] L. H. Stember, “Reliability considerations in the
10.1016/j.enconman.2017.08.090. design of solar photovoltaic power systems,” Solar
[5] J. M. Kuitche, R. Pan, and G. TamizhMani (Mani), Cells, vol. 3, no. 3, pp. 269–285, May 1981, doi:
“Investigation of Dominant Failure Mode(s) for 10.1016/0379-6787(81)90008-9.
Field-Aged Crystalline Silicon PV Modules Under [19] M. Mesić and T. Plavšić, “The contribution of
Desert Climatic Conditions,” IEEE Journal of failure analyses to transmission network maintenance
Photovoltaics, vol. 4, no. 3, pp. 814–826, May 2014, preferentials,” Engineering Failure Analysis, vol. 35,
doi: 10.1109/JPHOTOV.2014.2308720. pp. 262–271, Dec. 2013, doi:
[6] L. S. Cickaric, V. A. Katic, and S. Milic, “Failure 10.1016/j.engfailanal.2013.01.028.
Modes and Effects Analysis of Urban Rooftop PV [20] T. D. Han, M. R. M. Razif, and S. A. Sulaiman,
Systems – Case Study,” in 2018 International “Study on Premature Failure of PV Systems in
Symposium on Industrial Electronics (INDEL), Nov. Malaysia using FMEA and Integrated ISM
2018, pp. 1–7, doi: 10.1109/INDEL.2018.8637640. Approaches,” MATEC Web Conf., vol. 225, p.
[7] T. Oozeki, T. Yamada, K. Kato, and T. Yamamoto, 04004, 2018, doi: 10.1051/matecconf/201822504004.
“An Analysis of Reliability for Photovoltaic Systems [21] J. B. Basu, “Failure Modes and Effects Analysis
on the Field Test Project for Photovoltaic in Japan,” (FMEA) of a Rooftop PV System,” vol. 3, no. 9, p. 5,
in Proceedings of ISES World Congress 2007 (Vol. I 2014.
– Vol. V), Berlin, Heidelberg, 2009, pp. 1628–1632, [22] M. Catelani, L. Ciani, L. Cristaldi, M. Faifer, M.
doi: 10.1007/978-3-540-75997-3_334. Lazzaroni, and P. Rinaldi, “FMECA technique on
[8] E. Collins, M. Dvorack, J. Mahn, M. Mundt, and M. photovoltaic module,” in 2011 IEEE International
Quintana, “Reliability and availability analysis of a Instrumentation and Measurement Technology
fielded photovoltaic system,” in 2009 34th IEEE Conference, May 2011, pp. 1–6, doi:
Photovoltaic Specialists Conference (PVSC), Jun. 10.1109/IMTC.2011.5944245.
2009, pp. 002316–002321, doi: [23] R. Hu, J. Mi, T. Hu, M. Fu, and P. Yang,
10.1109/PVSC.2009.5411343. “Reliability research for PV system using BDD-based
[9] S. Baschel, E. Koubli, J. Roy, and R. Gottschalg, fault tree analysis,” in 2013 International Conference
“Impact of Component Reliability on Large Scale on Quality, Reliability, Risk, Maintenance, and
Photovoltaic Systems’ Performance,” Energies, vol. Safety Engineering (QR2MSE), Jul. 2013, pp. 359–
11, no. 6, p. 1579, Jun. 2018, doi: 363, doi: 10.1109/QR2MSE.2013.6625601.
10.3390/en11061579. [24] R. Laronde, A. Charki, and D. Bigaud, “Reliability
[10]P. Zhang, Y. Wang, W. Xiao, and W. Li, “Reliability of photovoltaic modules based on climatic
Evaluation of Grid-Connected Photovoltaic Power measurement data,” Int. J. Metrol. Qual. Eng., vol. 1,
Systems,” IEEE Transactions on Sustainable Energy, no. 1, Art. no. 1, 2010, doi: 10.1051/ijmqe/2010012.
vol. 3, no. 3, pp. 379–389, Jul. 2012, doi: [25] IRENA, “Solar energy data - installed capacity
10.1109/TSTE.2012.2186644. trends,” Infographic, Jan 2019. [Online].
[11]A. Charki and D. Bigaud, “Availability Estimation of Available:https://www.irena.org/en/solar
a Photovoltaic System,” in 2013 Proceedings Annual [26] D. C. Jordan and S. R. Kurtz, “Photovoltaic
Reliability and Maintainability Symposium (RAMS), Degradation Rates-an Analytical Review:
Jan. 2013, pp. 1–5, doi: Photovoltaic degradation rates,” Progress in
10.1109/RAMS.2013.6517744. Photovoltaics: Research and Applications, vol. 21,
[12]Z. J. Ma and S. Thomas, “Reliability and no. 1, pp. 12–29, Jan. 2013, doi: 10.1002/pip.1182.
maintainability in photovoltaic inverter design,” in
2011 Proceedings - Annual Reliability and
Maintainability Symposium, Jan. 2011, pp. 1–5, doi:
10.1109/RAMS.2011.5754523.
[13] E. Koutroulis and F. Blaabjerg, “Design
Optimization of Transformerless Grid-Connected PV
Inverters Including Reliability,” IEEE Transactions
on Power Electronics, vol. 28, no. 1, pp. 325–335,
Jan. 2013, doi: 10.1109/TPEL.2012.2198670.
[14] A. Colli, “Failure mode and effect analysis for
photovoltaic systems,” Renewable and Sustainable
Energy Reviews, vol. 50, pp. 804–809, Oct. 2015,
doi: 10.1016/j.rser.2015.05.056.
[15] G. Zini, C. Mangeant, and J. Merten, “Reliability of
large-scale grid-connected photovoltaic systems,”