Maintenance Management For Photovoltaic Power System Survey

Renewable and Sustainable Energy Reviews 134 (2020) 110347
Contents lists available at ScienceDirect
Renewable and Sustainable Energy Reviews

journal homepage: http://www.elsevier.com/locate/rser
Survey of maintenance management for photovoltaic power systems

Alfredo Peinado Gonzalo a, Alberto Pliego Marugán b, Fausto Pedro García Márquez, Prof. c, *
a
University of Birmingham, Birmingham, United Kingdom
b
ESIC Business & Marketing School, Madrid, Spain
c
Ingenium Research Group, Universidad Castilla-La Mancha, 13071, Ciudad Real, Spain
A R T I C L E I N F O A B S T R A C T
Keywords: The sustainability of the global energy production systems involves new renewable energies and the improve
Photovoltaics ment of the existing ones. Photovoltaic industry is growing thanks to the development of new technologies that
Renewable energy increase the performance of photovoltaic systems. These systems are commonly subject to harsh environmental
Maintenance
conditions that decrease their energy production and efficiency. In addition, current photovoltaic technologies
Fault detection
Fault prevention
are more sophisticated, and the size of photovoltaics solar plants is growing. Under this framework, research on
Solar power energy failures and degradation mechanisms, together with the improvement of maintenance management, becomes
essential to increase the performance, efficiency, reliability, availability, safety, and profitability of these sys
tems. To assess maintenance needs, this paper presents a double contribution: an exhaustive literature review
and updated survey on maintenance of photovoltaic plants, and a novel analysis of the current state and a
discussion of the future trends and challenges in this field. An analysis of the main faults and degradation
mechanisms is done, including the causes, effects, and the main techniques to detect, prevent and mitigate them.
1. Introduction technologies.
PV solar energy generated 385 GWp in 2017, compared to 2008,
The use of renewable energies has been rising in the last decade due when it was only 14 GWp [9]. Fig. 1 shows the solar PV generation
to national and supranational government policies [1,2], together with capacity over the last 10 years [10]. There is an exponential growth on
the location optimization of renewable power plants [3], where solar the last years, being China, US and Germany the main investors. CSP has
energy is one of the main renewable energies. The main solar power also increased its capacity to more than 3000 MW [11], but it still is less
types are concentrated solar power (CSP) and photovoltaic (PV). CSP than 1% of the energy generated by PV [12].
concentrates the solar energy to heat a transfer fluid to generate energy The PV industry needs of novel and advanced technologies, together
like a thermal plant [4,5]. This technology presents four main configu with robust approaches for advanced analytics to support the decision
rations: Parabolic trough collector (PTC), solar power tower (SPT), makers [13,14]. New technologies will lead to reach the competitiveness
linear Fresnel reflector (LFR) and parabolic dish systems (PDS). Other in this sector, together with better efficiency, sustainability and main
technologies, such as concentrated solar thermoelectric, are in devel tenance policies [13,15,16]. Valuable research has been conducted to
opment stage and are not economically feasible yet [6]. PV solar energy review aspects of PV technology, but it is limited to aspects such as the
is based on the PV effect on semiconductor materials. Electricity is operation and maintenance issues, the effects of degradation mecha
generated when sunlight incises on solar cells. This electricity can be nisms, or the development in different countries. Most of these reviews
directly used, stored or sent to the grid [7,8]. Solar cells are mainly only discuss the advantages and disadvantages of the approaches
separated depending on their crystalline structure and materials, without enough detailing of the procedures and structures of the
whereas PV systems are separated into stand-alone and grid-connected. methods. To expand the research, this paper first focuses on the main
The combination of solar cells, modules and arrays allows this tech degradation mechanisms affecting PV technology, considering their
nology to be used in small scale devices, such as calculators, or in large causes and consequences; and then complements it with the most recent
power plants, making PV one of the most flexible energy generation research dedicated to their detection, prevention, and mitigation. The
* Corresponding author.
E-mail addresses: axp864@student.bham.ac.uk (A. Peinado Gonzalo), alberto.pliego@esic.edu (A. Pliego Marugán), faustopedro.garcia@uclm.es (F.P. García
Márquez).
https://doi.org/10.1016/j.rser.2020.110347
Received 23 March 2020; Received in revised form 29 July 2020; Accepted 3 September 2020
Available online 12 September 2020
1364-0321/© 2020 Elsevier Ltd. All rights reserved.
A. Peinado Gonzalo et al. Renewable and Sustainable Energy Reviews 134 (2020) 110347
topics and field of study covered by each publication are summarized in manifested the benefits derived from a correct maintenance manage
the corresponding tables. The main contributions of this paper to the ment and the necessity for the identification of the best practises within
current research are: the photovoltaic maintenance field. The methodology followed in this
work is aligned with the procedures followed during the mentioned
- This paper explains the principal faults in PV technology, and the projects, which is based on three fundamental stages:
main causes and consequences on the affected parts of PV modules.
Similar publications on this field are [17], on the challenges faced for - First, seeking information. This paper is the result of an exhaustive and
the operation and maintenance of PV systems, considering their in deep searching of literature, mainly scientific papers. This literature
spection and different failure modes; or [18,19] on the development has been found through the most important scientific databases
of PV energy in India and England, respectively. Other reviews focus (SCOPUS, Web of Science, IEEE Xplore, ScienceDirect, etc.). The
on one type of defects, such as Costa et al. [20] on soiling research research works are carefully summarized. The relevance of these
and development; Reese et al. [21], on protocols to measure corro works is also analysed by considering some indicators such as the
sion and degradation caused by outdoor exposure; Pawluk et al. [22], number of cites of the paper or the prestige of the journal according
on the influence factors of snow power losses; Hamdi et al. [23], on to the rankings.
the effects of humidity; or Kumar and Kumar [24], who made general - Second, the classification of literature. The selected papers are ordered
review of PV technology and its failure modes. This paper analyses and classified following a coherent structure. A contribution of this
the most important defects affecting PV energy and their influence study is that it does not focus only on maintenance activities, but also
on the affected components. describes the possible faults (causes and effects) that may occur.
- The most important issues related to failure prevention and mitiga Therefore, this paper presents an operative perspective when
tion techniques on PV are presented. Other reviews are focused on describing the main maintenance techniques, but also a purely
monitoring and fault detection systems [25]; the application of experimental perspective when describing the faults and degradation
thermography for solar fields monitoring [26]; or the use of mechanisms.
self-cleaning coatings on PV panels to prevent dust deposition [27]. - Third, the analysis and discussion of the state of the art. This stage is one
Most literature in this field is focused on a specific technique, but this of the most important contributions of this paper since new knowl
paper covers the most relevant literature about failure prevention edge is generated. By using metadata (such as topic, date of publi
and mitigation from a general perspective. cation, cites, etc.) of the research works, a quantitative analysis is
- This paper summarizes the most relevant publications analysed done to obtain statistical information of the current state of main
through two informative tables (Tables 2 and 4), where all the ref tenance in PV solar plants. Moreover, due to the numerous papers
erences can be consulted at a glance. These tables are organised studied in the previous stages, a qualitative analysis of the state of
attending to the main topic, degradation mechanism, type of study, the art is carried out. Both qualitative and quantitative analysis allow
experimental location, and the corresponding reference. These tables to extract important conclusions, to discover trends and to identify
can be a useful index for readers to look for specific information. weaknesses in this topic. This new information is essential for the
- Each section contains not only a collection of relevant papers but also identification of challenges and future research lines.
a brief discussion of the current state of the art. Moreover, after
presenting the state of the art, quantitative and qualitative analyses As the result of the described methodology, the paper is structured as
of the main requirements and improvement in this field are done. follows: Section 2 explains the fundaments of the PV technology, sys
The main trends and future challenges are gathered from these tems and materials; Section 3 shows the main degradation mechanisms
analyses. affecting PV technology, the publications about them and the main
conclusions extracted from them. Section 4 summarizes the main
This paper is the product of several European research projects research studies about prevention and mitigation of failures in PV
(NIMO (FP7-ENERGY-2008-TREN1:239462) [28], OPTIMUS (FP7-E technology. Section 5 presents the discussion of the state of the art
NERGY-2012-2-322430) [29] and INTERSOLAR (Fp7-sme-2013- analysis and summarizes the main references. Finally, Section 6 shows
605028) [30]) in the context of maintenance of renewable energy sys the conclusions and future challenges.
tems. In this framework, companies of the photovoltaic sector have
Fig. 1. Evolution of PV energy generation in the last years. Figure adapted from Ref. [10].
2
2. Photovoltaic technology highly defective, having the lowest efficiency of all silicon-based
materials, 13.4%.
The basic unit in a PV solar power plant is the solar or PV cell. It is a - Gallium arsenide: It has a high level of absorptivity and higher effi
device that generates electricity from the PV effect (movement of elec ciency than silicon (28.8%), but it is considerably more expensive. It
trons in certain materials when irradiated by electromagnetic radiation), is used in concentrator systems and space applications.
produced in a PN junction [31]. In this component, see Fig. 2, electrons - Cadmium telluride: It is one of the most promising thin film materials
are able to move to the p-type part of the junction when the n-type part due to its lowest production cost among them. It has a high ab
of the junction is irradiated with more energy than the band-gap energy sorption coefficient and good efficiency (21.5%).
of the material, creating a voltage and an electric current [31,32]. - Copper indium diselenide: It can achieve high energy conversion effi
Solar cells are usually connected in series, parallel or combined ciency (21.5%) without suffering degradation. Some drawbacks are
forming a PV module to produce enough voltage and power. The PV its difficult manufacturing and toxic gases created in the process.
array is a structure consisting of PV modules. Its power generation - Organic PV materials: There is a high variety of materials where
ranges from hundreds of watts to megawatts. The possibilities for flexibility and colour are important, separated into sublimed small-
connection are the same as for PV cells: series, parallel or combined. The molecule and solution processed polymer/fullerene solar cells.
following level are the PV power systems, qualitatively classified into However, their efficiency is relatively low (11.5%), and are highly
stand-alone and grid connected (Fig. 3). They can be also quantitatively prone to degradation and corrosion.
classified according to their power generation capacity [31]. - Quantum dot: These cells are made with synthesized, low-
Stand-alone systems are the ones in which energy generation cor temperature solution processing, and their band gap is adjustable
responds to the demand. They have been historically useful in in by composition and size. They have large voltage and current losses,
stallations where it is not common to connect demand with the main and a distribution of quantum dot sizes resulting in distribution of
grid, e.g. rural houses. It can be used in small electronic devices, like band gap energies. Their efficiency is the lowest of all materials
calculators or watches. These systems can also be divided into with or (9.9%).
without storage, and hybrid systems. The last ones combine generation
of the PV solar panels with other system(s) when the power coming from To summarize, PVs is a well understood and largely studied tech
the modules is not enough. These systems are usually supported, for nology, based on the PV effect that appears on a wide variety of mate
example, by a wind turbine, a diesel generator or a cogeneration engine. rials. It has a large range of applications as a consequence of its
Grid-connected are large systems which provide energy and may be scalability, from calculators to power plants. However, it is subjected to
connected to the network via public grid or house grid. They are divided harsh environmental conditions in industrial applications that are likely
depending on their power range (Fig. 3). There is some discussion on the to cause failures on the modules.
boundaries of this classification, such as the consideration of hybrid as a
type of PV systems on their own. 3. Faults and degradation mechanisms in photovoltaics
The PV materials must present several characteristics for its suit
ability, based on semiconductor properties [33], such as: band gap be The performance of PV modules is conditioned mainly by two fac
tween 1.1 and 1.7 eV, readily available, non-toxic components, good PV tors: the light received by the modules and the condition of cells and
conversion efficiency, manufacturing technique suitable for large pro connections. Regarding the light received, the most important faults are
ductions and long-term stability. There is not complete agreement on the soiling, shading and snow deposition. The condition of cells and con
type of PV materials, but the two most common are crystalline and thin nections are due principally by degradation, delamination hot spots, or
film, among other categories as organic or quantum dot materials [34, cracks.
35]. Some of the most relevant materials are:
3.1. Soiling
- Single-crystal silicon: It has good stability and properties. The previous
existence of silicon in electronics contributed to reduce its price. Its Soiling is the deposition of dust, dirt, pollen and/or other elements
efficiency is the higher among silicon-based materials, 24.2%. on the surface of the PV modules. It is one of the most relevant aspects to
- Poly-crystalline silicon: Being cheaper than single-crystal, it is the be considered in solar energy, especially in desertic areas close to the
most popular material for PV energy although its efficiency, 19.2%, equator. In consequence, research works focus more on dust than other
is lower. This is compensated by its ease of assembly. depositions. Its effects are generally short term and temporary, since
- Amorphous silicon: Its absorptivity is about 40 times greater than soiling is usually a reversible fault. Ilse et al. [36] explain the basis of
crystalline silicon. Its production process requires less energy, soiling and condition the key parameters of dust: concentration,
reducing its cost. However, it degrades under sun exposure and is composition, size distribution, accumulation and removal, and envi
ronmental parameters. The causes of dust accumulation are divided into
three groups [37]: environment factors (wind, temperature, humidity or
pollution); dust nature (sand, clay, bacteria or carbon), and installation
characteristics (material, orientation, location, type of area …).
Dust characterization and its effects depend on the location of the
installation. John et al. [38] developed artificial soiling tests in labora
tory to compare the composition and the effects of dust from locations in
Mumbai and Jodhpur (India). Samples from those sites showed strong
variations both in size and composition. For instance, the authors
observed greater losses in samples from Mumbai (18% of losses with 5
g/m2 of dust) than from Jodhpur (10% of losses with 5 g/m2). Rao et al.
[39] compared a clean panel with a dusty panel for laboratory and
outdoor locations. They observed that, under similar conditions, the
dusty panel had not remarkable losses in open circuit voltage compared
to the clean one, but they found power losses ranging between 45 and
Fig. 2. Schematics of a PN junction and its energy generation. Figure adapted 55% in the indoor tests and 6–8% in the outdoor tests. Javed et al. [40]
from Ref. [31]. made a comparison between dusts from different desertic places from all
3
Fig. 3. Classification of PV systems. Figure adapted from Ref. [31].
over the world. Other works about dust characterization have been losses in crystalline than in amorphous materials, ranging from 3% with
developed by Javed et al. [41], Zitouni et al. [42] and Burton et al. [43]. an angle of 10◦ to 0.5% with 60◦ .
Power losses due to soiling have been measured in laboratory The effects of snow deposition have been modelled according to
(controlled irradiation and soiling rate conditions) and outdoor condi different parameters, e.g. air temperature, tilt angle, irradiance or ma
tions, whose results depend on the materials and dust deposition. For terial to improve operation and maintenance aspects. Marion et al. [56]
indoor conditions, Burton and King [44] observed linear changes on used different sensors to measure snow losses and modelled a system to
losses up to 20% with 2 g/m2 of dust on glass covers. For outdoor assess surveillance and maintenance of the modules. Powers et al. [57]
conditions, Saidan et al. [45] obtained power losses greater than 18% for considered temperatures, humidity and snowfall, among other param
one month in Iraq; García et al. [46] observed yearly losses averaging eters, to model losses and compare them on cleaned and covered mod
11.9% on horizontal surfaces; Mejia et al. [47] determined losses up to ules, developing a model to predict snow deposition and losses for
7.2% in dry periods in California; Caron and Litmann [48] measured several years.
peaks of more than 11% of losses in California. Finally, Kimber et al. Image processing, heat transmission and fluid dynamics are used for
[49] observed a decreasing trend in the daily efficiency of 0.2% in dry modelling the snow sliding and melting, see for example the research
climates, meaning between 1.5 and 6.2% yearly, depending on plant works of Bogenrieder et al. [58] and Rahmatand et al. [59].
location; Gostein et al. [50] observed soiling rates ranging between 0.5% Finally, ice is produced in the same conditions as snow, but it has not
and 5% per month; and Pedersen et al. [51] defined a transmittance loss been studied enough [60,61]. Carriveau et al. [62] explain its effects on
rate of 0.1% per 10 mg/m2 approximately. renewable energies in general, and commented the decrease of energy
In summary, soiling is a phenomenon that affects mostly PV in production because of high ice adhesion on PV panels. Pawluk et al. [63]
stallations mainly in desert locations. Most of the publications focus on explained the effects of ice adhesion and sliding on solar modules related
dust characterization than on power losses. Different studies have to snow deposition and light irradiation.
determined large variations between indoor and outdoor tests. Its effects In conclusion, snow deposition is a seasonal phenomenon happening
are usually short term and easily mitigated by proper cleaning of the on places with cold climates, that causes severe light obstruction, ice
surface. This fault affects the installation intensity, but not the voltage. deposition and possible moisture ingress. It is modelled according to
environmental and panel parameters and causes remarkable power
3.2. Snow and ice deposition losses on affected modules. Sliding and melting are modelled consid
ering the influence of fluid dynamics and heat transmission, but they
Snow and ice are relevant in cold countries because the climate fa have not been studied enough. Studies related to ice are few and usually
vours their apparition. Some of the main problems generated are total or focus on sliding and melting rather than the losses produced.
partial obstruction of radiation, reduction of efficiency, cracking and
delamination or mechanical loads. One advantage is the decrease of 3.3. Outdoor exposure: corrosion, delamination and degradation
temperature of the modules [52], but there are drawbacks, e.g. light
obstruction and moisture ingress on the modules. There is a large Outdoor conditions and exposure to sunlight generate short and
amount of research works related to its effects, or to prevent and miti long-term consequences on PV modules, deeply related within them
gate this phenomenon. Usually, snow deposition is quantified through selves. Therefore, research usually explains their correlation and effects
the power losses, image processing or temperature indicators [53]. together. The main faults related to long term exposure are corrosion
Regarding power losses, Pawluk et al. [22] made a literature review and delamination, due to moisture ingress and general degradation of
on different climates and the influence factors, such as climate, adhesion the modules.
and angle. They also review the most remarkable prediction and Corrosion affects more in organic PVs, but also modifies the crys
modelling techniques for short and long term; and mitigation methods, talline and thin film modules characteristics. There are protocols to
such as coatings and heating. Power losses are clearly seasonal, measure the influence of stability, influence of weather, temperature,
appearing only on cold months, as remarked by Heidari et al. [54]. They etc. on this fault. Reese et al. [21] explained some of the main protocols,
showed that snow deposition is only remarkable from October to April in tests and requirements to measure corrosion and degradation on PV
USA, producing yearly energy losses of 30–35% on obstructed panels modules under different light, temperature and irradiation conditions.
compared to 5–15% on unobstructed panels. Andrews et al. [55] Regarding to the outdoor variables, Gevorgyan et al. [64] studied the
designed a structure of solar panels with different tilt angles, and stability of different modules after 10000 h of testing in several coun
measured yearly the energy losses due to snow by using image pro tries. The modules were prepared and encapsulated to ensure their
cessing. They observed a low influence of angle on losses, and few more long-term stability.
4
Many studies on corrosion focus on the moisture effects. Park et al. moisture ingress in a PV module in South Africa. It is observed that
[65] developed a method to calculate the accumulated degradation and physical degradation is clearly visible and varies with the distance to the
performed accelerated tests to obtain the results. Osterwald et al. [66] frame, and this delamination allows moisture to penetrate. There was a
studied electrochemical corrosion in thin-film modules using acceler decreasing of performance (more than 16%) and temperature variations
ated tests, varying the moisture of the testing chamber. Karl et al. [67] within the different cells. The paper concludes that moisture regresses in
investigated lifetime of organic PV cells through comparison with cal hot and dry months, improving the performance.
cium corrosion tests. Related to moisture ingress is the phenomenon named ‘snail trails’. It
Delamination occurs due to the loss of adhesion between the occurs when local areas of a PV cell react with moisture causing a
encapsulant and the photosensible parts of the cells, see Fig. 4. It facil chemical reaction with the silver, sulphur and phosphorus contained,
itates the moisture and corrosion, leading to a performance reduction. It study done by Meyer et al. [75]. This defect causes losses of electrolu
usually appears on metalized regions, as gridlines or bus bars. These minescence, making it easy to be detected. According to Yang et al. [76],
defects are irreversible and contribute to the long-term degradation of this phenomenon also causes performance losses, around 9%, but tem
the modules. There is a variety of methods to measure delamination and perature does not increase in the affected cells.
interface adhesion. Tracy et al. [68] developed a method consisting in Regarding general module degradation, research works are divided
the preparation of a standardized PV module and the subsequent mea into cause definition and long term effects. Respect to the cause defi
surement of the adhesion energy and chemical analysis. Jansen and nition: Kounouhéwa et al. [77] summarized the different degradation
Delahoy [69] evaluated the delamination in glass substrates applying an causes, components affected and their effects; van Dyk et al. [78]
electrical current for up to 15 min and measuring the effects after the investigated the electrical and physical degradation of the module under
cooling of the sample. Tracy et al. [70] quantified the energy per surface outdoor exposure,; and Bouraiou et al. [79], tested degradation, soiling,
required to separate two adhesive sheets, and proposed two different delamination or corrosion of exposed PV modules in Saharan environ
tests, for PV modules and for adhesive materials. ment. Regarding long term exposure, the most remarkable effects are
Terwilliger et al. [71] developed an extensive study of PV module permanent damage and power losses. For instance, the long term
packaging materials and their response to adhesion tests under different exposure was studied by Silvestre et al. [80] after 5 years in Madrid,
conditions. They measured the evolution of adhesion and moisture Limmanee et al. [81], 4 years in Thailand and Chandel et al. [82], after
transport. The tests were made on a wide variety of samples ranked 28 years in western India.
according to their performance on several aspects, such as adhesion, To conclude, the effects of outdoor exposure are irreversible and
glass cleaning or shear strength. They concluded that: the approach usually affect the long-term behaviour of the PV modules. Faults
provided better results than commercial products; combined packaging explained in this section are closely related, as corrosion and delami
may allow moisture ingress but also egress of harmful substances; and nation are produced by moisture ingress and cause long-term degrada
the conditions of 85 ◦ C and 85% moisture are too aggressive for testing, tion of the modules. Their effects are appreciable through the power
but 60 ◦ C and 70% moisture are too mild. They also set that the ideal profile and output of the modules.
conditions for testing must be still determined.
The long term effects of delamination and discoloration are studied
by Park et al. [72] on 25 year old modules in coastal environment, see 3.4. Cracks
Fig. 4. The initial observations showed a decreasing of almost 18% of
power generation, but no dielectric breakdowns on the modules. A vi Cracks are considered breakages in the structure of PV modules.
sual test to detect the main defects was developed, discovering a clear They are unpredictable, and the quantification of losses is complicated
deterioration in the modules, together with a decreasing of electrical since their effects are unknown. Usually, the cracks are microscopic and,
properties in the defective cells and corrosion in the zones with metal along the time, they grow and cause the apparition of larger ones [83].
contact of the cells. Discoloration is also studied by Sinha et al. [73], The main crack types in a PV module are shown in Fig. 5.
who classified its causes and used several non-destructive tests to According to Köntges et al. [84], cracks appear both in the transport
identify its effects on long term exposed modules. of the PV modules and in mechanical tests, due to vibrations and reso
Van Dyk et al. [74] analysed the effects of the delamination and the nance modes of the modules. They compared the statistical distribution
of cracks and observed a predominance of parallel and 45◦ oriented
Fig. 4. Degradation of a 25 year old PV module [72].
5
Fig. 5. Crack classification for crack sorting [84].
cracks, responsible of an approximated 70% of them. The results show a 3.5. Hot spots
common distribution among modules from different manufacturers,
concluding that mechanical loads during transport are the main cause of This fault consists in the increasing of temperature of PV solar cells
crack appearance and growth. due to, for example, partial shading of the solar module or deficient
Kajari-Schröder [85] studied crack distribution after mechanical connections between cells, causing the energy dissipation in the form of
load tests. They focused on the crack formation under homogeneous heat instead of electricity [93,94].
static mechanical loads, e.g. snow or wind. The electroluminescence was Moretón et al. [95] applied several methods to identify possible hot
measured after the tests to determine crack distribution and orientation spots in two solar plants in the centre of Spain. They discovered cracks
of the modules. Authors observed that 50% of cracks are parallel to the and bubbles in some cells by visual inspection, whilst thermography
busbar, whereas diagonal, dendritic and several direction cracks are shown the actual location of hot spots (Fig. 6). This study sets that the
responsible of around 15% of the cracks. Kilikevičius et al. [86] studied power losses of the whole module are a reliable source of information,
the effects of dynamic loads, such as vibrations or wind, on the perfor but not the relation between temperature difference and PV cell power
mance and crack apparition of PV modules. They achieved that the loss. Unmanned aerial vehicles (UAV) equipped with a thermal camera
maximum fracture occurred at 16 Hz, with a maximum amplitude of 7 have been demonstrated to be a well-functioning technique [83], see, for
mm at 20 Hz. example, the researchers done by de Oliveira et al. [96] on a solar farm,
Käsewieter [87] studied a PV cell to measure the variation of elec and Bharadwaj et al. [97] in a test site.
trical resistance under cyclic bending loads. On the first bending cycle, Simon and Meyer [98] made tests in darkness to identify the effect of
the creation of the crack was visible and audible, with a force drop and
increase of the system resistance. The posterior cycles show an incre
ment of the resistance, with less value at each time. Sander et al. [88]
evaluated the electroluminescence of the modules with different loads to
observe its evolution. They observe cracks before testing due to sol
dering or lamination in the manufacturing process. Both tests identified
an exponential increment of the size and number of cracks during the
experiments.
Image processing is also applied for crack detection by using
different methods considering the electroluminescence of the PV cells.
Dhimish et al. [89] developed a 3D model to examine micro cracks
enhanced with logical gates. Song et al. [90] used an algorithm based on
the named Gaussian pyramid and wavelet modulus. Other related
studies were done by Dhimish et al. [91], analysing related size and
direction of cracks to performance, and by van Mölken et al. [92], using
electric modelling and electroluminescence measurements to study
micro cracks and their effects.
In conclusion, cracks appear due mainly to mechanical loads, and
their principal effects are changes in electroluminescence and power
decreases. Most of literature focuses on crack distribution and properties
evolution, usually taking advantage of the changes in the
electroluminescence.
Fig. 6. Hot spots in a PV module [95].
6
impurities on the apparition of hot spots. Once the hot spots were technologies and techniques are discussed, as well as the relevance of
identified, three areas were defined: hot spot centre, outside hot spot the I–V curves.
and non-hot spot. They were analysed by thermography and micro Triki-Lahiani et al. [25] made a review on monitoring and fault
scopy. The observations shown a direct correlation between areas of detection systems. They explained the different failure modes of PV
high impurity concentration and hot spot heating, as well as the com modules and their effects and causes, showing some of the main pa
ponents of these impurities: oxygen, carbon, iron and platinum. Ac rameters used to monitor and detect failures (irradiance, temperature,
cording to that, Ramspeck et al. [99] developed a setup oriented to voltage and current, power, etc.). They also explained some of the
process control of PV cells. The method consists in applying voltage to different data analysis and processing techniques used, as circuit simu
the cell and force it to work on reverse conditions, and then create a lations, statistical analysis, artificial neural networks or predictive
thermography of the cell, all in less than half a second. The process models. Another review is done by Packard et al. [109], who analysed
control compares reverse power and temperature, and it accepts or re the inspection of PV components and installations, and developed a
jects the cells depending on their relation. This process allows the standardized form to write down all the necessary information.
quality of the final product to be improved with a relative low CMS can also be done remotely using online configurations [110].
investment. Tina and Grasso [111] studied the effectivity of a stand-alone PV
There are many publications related to simulation and modelling of appliance and a remote CMS software capable of storing and plotting
hot spots using different software. For instance, Deng et al. [100] data from sensors (power, temperature, voltage, etc.), showing real time
designed a model to define heat dissipation under different shading data and sending e-mail alerts about the appliance condition. Spataru
conditions. Solheim et al. [101] modelled the thermal properties of a et al. [112] used temperature and irradiance to show power output and
solar cell considering its several layers to calculate more precisely the made short-term predictions under different conditions, obtaining good
thermal distribution. The simulation confirmed that the size of the hot sensitivity to deviations and capability to detect power losses above 5%.
spot is inversely proportional to the temperature on it. Rajput et al. Watson et al. [113] presented two techniques to continuous condi
[102] modelled hot spot behaviour and validated the results in outdoor tion monitoring. These are a spectrophotometric analysis to record light
conditions using a high speed thermocamera, PV modules with verified absorption of the cell and corrosion progression for 24 h, and digital
hot spots, and a meteorological station to measure environmental con image acquisition during 100 h. The results differentiated between iron,
ditions. They discovered that wind speed reduces significantly the zinc and aluminium, and molybdenum, tungsten and stainless steel,
temperature of hot spots, similar behaviour of opaque and used for the coating. These results showed the accuracy of the methods
semi-transparent modules, and a negative correlation between the hot to measure corrosion, and the different resistance to corrosion of the
spot temperature and its quantity. used metals. Pavan et al. [114] utilised two functioning PV plants to
Other publications focus on electrical simulation and testing to correlate soiling and power losses with a CMS consisting of: an acqui
analyse hot spots. Kim and Krein [103] developed an electrical model sition board; a radiation sensor; two sensors per module for ambient
dependent on temperature, and Kim et al. [104] detect hot spots using temperature acquisition, and; a data logger and the server for the ac
several measurements based on electrical properties. Hot spots under quired dataset. The system analyses the predicted and measured power
shading conditions have been studied by Geisemeyer et al. [105] and produced before and after cleaning the PV modules, showing an almost
Rossi et al. [106]. perfect correlation for their model.
To summarize, hot spots are defects with different causes that pro Application of different approaches have been proven also effective.
duce a temperature increase and power losses on the affected cells and Touati et al. [115] used LabVIEW and sensors measuring ambient con
modules. The technique most used to detect them is thermography. This ditions and power output, sending them to an online platform and
fault has been modelled electrically and by computer simulations. showing the evolution of power output and pattern recognitions before
It has been stated that the most relevant degradation mechanisms in and after cleaning the modules. Other publications regarding PV
PV energy depend on the light received and on the condition of cells and monitoring using LabVIEW are done by Chouder et al. [116] and Rezk
connectors. Regarding light received, short-term and reversible effects et al. [117], Fuentes et al. [118] using the open platform Arduino, and
are caused by soiling and snow deposition, both producing light Shariff et al. [119], who made a brief literature review and used a Zigbee
obstruction and power losses but with different causes. Long term effects processing board for wireless online data acquisition.
appear under outdoor exposure and are mainly corrosion, delamination Infrared testing and image processing are common inspection tech
and degradation, which are irreversible and interrelated. Cracks and hot niques in PV, there, many studies use these two techniques. Tsanakas
spots are the main fault mechanisms according to cells and connections. et al. [26] reviewed the main aspects of online thermography of PV
Preventive/Predictive maintenance tasks should be done to avoid and modules and the defects identified through this technology, together
reduce the appearance and consequences of these faults. with the characterization of the intensity-voltage (I–V) curve of the
modules. They explained publications of image processing to automat
4. Failure detection, prevention and mitigation techniques in ically identify and classify failures such as degradation, cracks, defective
photovoltaics connections or shadowing. Other works related to the use of thermog
raphy and image processing have been presented by Gupta et al. [120],
The correct PV module condition may be compromised by the faults who compared the performance of high speed thermographic cameras to
or failures, where preventive/predictive maintenance should be done detect temperature variations on solar cells, Jaffery et al. [121], that use
correctly to avoid and minimize them. Nowadays, these maintenance Fuzzy Logic systems to monitor and predict possible failures, and
tasks are mainly based on condition monitoring systems (CMS) [107, Buerhop et al. [122], studying the reliability of IR in an operating PV
108]. This section analysed CMS into three groups: failure detection, the plant to identify different defects.
act of identifying anomalies, defects or failures in systems; prevention, There are other techniques to detect failures: the six-layer detection
application of measurements and design changes of components and algorithm, explained by Dhimish et al. [123], capable of detecting
systems to avoid failures; and mitigation, application of measurements failures through the power output of the PV module; the parallel algo
and design changes to minimize the effects. rithm by Dhimis et al. [124], detecting different failures and shading
effects using the power output of grid-connected modules; statistical
4.1. Failure detection techniques approaches, such as the ones by Harrou et al. [125] and Garoudja et al.
[126].
This section is focused on the CMS and failure detection in PV in UAVs are also used for embedding CMS for PV panels, enabling to
stallations, both individual PV panels and complete farms. New inspect large surfaces in short amounts of time [127,128]. Herraiz et al.
7
[129] presented a review on CMS for PV solar plants based on ther focused on surface modifications and there are many reviews and clas
mography. Quarter et al. [130] embedded digital and thermographic sifications. Andersson et al. [143] classified four types of hydrophobic
cameras to a UAV. They inspect panels and detect faults as cracks, hot surfaces: structured, liquid infused, smooth and hybrid, and comment
spots or snail trails remotely. Grimmacia et al. [131] used a UAV to the research opportunities of them, as the effect of albedo or minimum
inspect both a rooftop installation and a PV plant in the north of Italy angle for snow sliding. Jelle et al. [144] divided them into four groups:
using also an embedded camera and a thermographic camera. Márquez repellent, micro and nanostructured, superhydrophobic and icephobic.
and Ramirez [132] embedded a CMS for solar power plants inspections Fillion et al. [145] focus on icing prevention, reviewing the most rele
with radiometric and thermographic sensors. vant aspects to be considered when designing an icephobic coating, as
The intensity-voltage (I–V) and power-voltage (P–V) curves are used ice nucleation, adhesion, condensation and durability. Other studies are:
to PV module condition analysis. I–V and P–V show the correlation Jelle [146], explaining research opportunities for the future on snow
between intensity or power output, respectively, and voltage output of a and ice prevention; and Andrews et al. [147], studying the effectiveness
PV module, and are significantly affected by soiling, hot spots, partial of hydrodynamic surface coatings considering their optical and snow
shading or shunts. These curves are modelled by González-Longatt shedding effects.
[133] considering parameters as temperature, power, voltage or current Fatigue is also studied in order to prevent mechanical failure, e.g.
intensity. Some fault effects are explained by Schill et al. [134] taking how many mechanical cycles a module or connector can withstand
into account different levels of soiling on the I–V curve. Moretón et al. before cracking or degrading. Paggi et al. [148] tested the effects of
[95], previously mentioned on section 3.5, employed them on hot spots cyclic loads on crack apparition and electrical recovery, making mea
detection, and Meyer and Van Dyk [135] used them to shunt resistance surements in the modules after a certain number of cycles. They
analysis. observed that recovery decreases with the cycles and when measuring
To summarize, CMS is necessary to ensure proper functioning of PV cracks. For ribbons, connector between cells, fatigue leading to mis
modules and solar farms. Many tools and dedicated software are used connections may cause hot spots and temperature increases on the
with this purpose, aiming at a continuous inspection of the installations. modules. Pander et al. [149] used a finite element model (FEM) simu
Therefore, when faults are detected, actions can be done quickly. New lation, a small PV module and copper ribbon to estimate its fatigue life.
technologies and automation, as the UAVs used for inspection, largely They studied the effects of displacement on connections between cells
reduce the time needed and increase the efficiency of inspection tasks. with a mechanical testing, and analysed them regarding the simulation
I–V and P–V curves are powerful tools that can help to identify different results. Their results shown a good estimation for the simulation and the
faults and act in consequence. Nowadays, research works are required to effectivity of fatigue testing for lifetime estimation. Other studies related
efficiently inspect the massive surfaces of PV solar farms with the aim to to fatigue in ribbon connectors can be found in: Bosco et al. [150],
increase their efficiency and life cycle. comparing simulations and real tests on cell connectors; Wiese et al.
[151], on microscopical effects of copper ribbons; and Meier et al. [152],
4.2. Prevention techniques considering the combined effects of temperature and mechanical
loading.
This section collects the main techniques used to prevent failures, In conclusion, prevention in PV energy is not as studied as CMS and
such as coatings and surface modifications for dust and snow deposition, fault detection. It focuses mainly on dust and snow prevention. Since
or fatigue analysis to study the statistical apparition of faults. The most of the defects on PV solar modules are irreversible, like degrada
research is more focused on prevention than mitigation because most of tion, corrosion or cracks, research efforts focus on avoiding the appari
the faults are irreversible. tion of these defects and modifying module surface before installing
Bahattab et al. [136] compared the performance of anti-soiling sur them to minimize power losses and faults on them. However, research is
faces in laboratory and outdoor conditions for surface modifications, required on degradation prevention, finding new materials and
considered as the changes in the geometry of the module surface. They improving isolation on modules to avoid moisture ingress and the
concluded that their laboratory tests can be reproduced in outdoor apparition of cracks on the cells.
conditions, but the different dust natures make unlikely a direct transfer
of results. Walwil et al. [137] modified several PV modules using glass
texturing and coatings to compare their effects through their power
output and temperature. They concluded that any modification supposes
an increase of power output between 1 and 8% for texturing and 5–12%
for coating. Other articles focused on surface modification are: Bhushan
et al. [138], studying the cleaning efficiency of hydrophobic surfaces
based on the lotus surface leaf; Nayshevsky and Lyons [139], analysing
the behaviour of hydrophobic-hydrophilic surfaces under different
moisture and configuration conditions; and Tucci et al. [140], focused
on modelling electrostatic treatments on surfaces.
Regarding coatings and products applied to a surface not previously
modified, Fathi et al. [141] compared the effectivity of a hydrophobic
coating on solar panels with and without surface treatment. In the test,
dust is added to a glass surface to compare how water is distributed over
it. They measured lower temperatures and power losses on the treated
modules, proving the beneficial effects of this treatment. Glaubitt and
Löbmann [142] cover a sample surface using a coating with nanopores,
and later subject to a dust exposure test. The results show a clear dif
ference between coated and uncoated surface under different moisture
and temperature experiments, where the transmittance losses in coated
surface are less than 2%, while in uncoated glass surpass 7%. Fig. 7
shows the differences between coated and uncoated surfaces after
testing.
The research regarding prevention of snow and ice deposition is Fig. 7. Coated and uncoated glass tubes after soiling testing [142].
8
4.3. Mitigation techniques main degradation mechanisms and, on the other hand, the principal
detection, prevention and mitigation techniques applied. Table 1 sum
As most of faults are irreversible, research is more focused on pre marizes the main causes, effects and detection, together with the im
vention and detection than mitigation. The only reversible defects provements, prevention or mitigation approaches.
whose effects can be mitigated are dust and snow deposition. This sec Table 2 summarizes the main degradation mechanisms, the type of
tion explains the approaches to minimize their effects, mostly cleaning study, the location and the main issue covered.
the soiled modules or removing the snow and ice on them. Table 3 collects data about the ten countries with the largest PV
Regarding soiling, most of studies on mitigation are focused on energy production. The amount of publications has been approximated
cleaning methods or their classification. Alshehri et al. [153] distin according to ScienceDirect. The variables that may be related to the PV
guished and analysed manual and mechanized tools by hydraulic sys energy production have been considered, such as population, the size of
tems, robotic systems and deployable robotics. Jamil et al. [154] divided the country, the GDP or the average solar irradiation. The correlation
cleaning systems into labour, natural, water, mechanical and surface between the PV energy production and each variable has been estimated
modifications. through the Pearson’s correlation coefficient.
The methods employed to mitigate soiling are also divided into water Some conclusions can be extracted from this analysis. First, the
and dry based. Regarding water based, Moharram et al. [155] studied correlation between the R&D expenditure and the PV energy is the
the influence of water and detergents on the performance of PV panels highest correlation, even higher that the correlation with GPD or the size
for a programmed routine, obtaining that the use of surfactants in a of the country. It leads to conclude that although the PV is a well-known
programmed schedule optimizes cleaning with a minimum use of water. technology, the investment in R&D is still necessary to improve the ef
Mazumder et al. [156] focused on high relative humidity locations and ficiency and the profitability of these systems. The amount of publica
introduce techniques as coating modifications or active and automated tions is also highly correlated with the PV production, being a direct
water-cleaning methods. Dry cleaning was studied by Aly et al. [157], consequence of the R&D expenditure.
who propose a four-staged cleaning process: air spraying to remove A null correlation has been found between the average solar irradi
coarse dust and particles; a polyurethane foam roller passing over the ation and the PV production. It could mislead the reader since the solar
surface to clean dust and sand stuck; other spraying to eliminate possible irradiation is a crucial factor for the PV production. This result reveals
dust rubbed, and; a electrostatically charged duster passes to remove that the differences between the production of these countries is not due
very fine dust particles. to their differences in the solar resource. In other words, regarding the
Other works on soiling cleaning and mitigation have been done by Li ten most important PV producers, there are other influential factors,
et al. [158], using the Kinect software to model the topology of the such as economic or political issues.
system to be cleaned, and Jones et al. [159], optimizing cleaning rou
tines for PV plants in Saudi Arabia. 5.1. Discussion on degradation mechanisms
Regarding snow and ice mitigation, research is focused on heating
panels to melt snow and ice. Weiss and Weiss [160] proposed a heating There are many different degradation mechanisms affecting PV
system on the bottom of the panels that causes avalanches and removes technology, both cells and modules, in different aspects. Even though
snow successfully after 15 min. Rahmatmand et al. developed two ap different materials and manufacturing processes are required to create
proaches on this topic: one focused on electrical heating [161], PV cells and modules, all of them create a similar energy production
considering different inclinations with reverse current and a heater, this system.
last one the more efficient, cleaning almost completely a panel after 20 Soiling and snow deposition are important issues to be considered
min, and; other approach on the effects of circulating hot water under and, therefore, they are widely studied because PV modules require a
the panels [162], removing the snow after 16 min heating. large surface to harness energy. Moreover, snow deposition causes me
To summarize, research done on mitigation is not as extended as on chanical loads on PV panels that can generate cracks and moisture
prevention, since most of the defects are irreversible. The studies are ingress, leading to corrosion and degradation, making it one of the most
focused on soiling and snow removal. Investigation on soiling is mainly aggressive degradation mechanisms affecting PV energy, especially in
done on cleaning under different methods, while research on snow and cold countries. On the other hand, cracks, caused by mechanical loads,
ice removal is based on heating the panels to melt and remove them. generate connection issues in the PV cells that may lead to hot spots and
However, this supposes a challenge and opportunity for future research important power losses on the modules.
on this topic, both in the faults explained and the ones not mentioned. Corrosion seems to have a higher impact over organic PV materials.
However, there are not many publications over this subject, therefore, it
5. Discussion cannot be considered definite. The effect of the PV material type on
potential failures is currently a gap in the literature. Moreover, the
On one hand, this paper analyses the main causes and effects of the behaviour of degradation mechanisms concerning the type of PV
Table 1
Causes, effects and detection of main degradation mechanisms.
DEGRADATION MECHANISM CAUSES EFFECTS DETECTION IMPROVEMENTS, PREVENTION
OR MITIGATION
Soiling Dust and particle Light obstruction, power Visual inspection, thermography, Surface modifications, coatings
deposition losses power output
Snow and ice Snow and ice deposition, Light obstruction, power Visual inspection, thermography, Coatings, surface modifications
low temperatures losses, moisture ingress power output
Outdoor exposure: corrosion, Outdoor exposure Power losses, cell Power output, visual inspection Material engineering, module
delamination and degradation degradation, insulation
Cracks Mechanical stress, bad Power losses, hot spots, Electroluminescence, thermography, Material engineering, handling.
handling breakages visual inspection
Hot spots Mismatches, defective Temperature increase, power Power output, electroluminescence Connection handling,
connections losses
Fatigue Use, outdoor exposure, Cell degradation, power losses Power output. Connection handling,
mechanical stress
9
Table 2
Degradation mechanisms in PV technology and studies related to them.
DEGRADATION MECHANISM TYPE OF STUDY STUDIES LOCATION TOPICS/KEYWORDS
RELATED
Soiling Causes and [20] Global Literature review.

effects [36] Germany Literature review, dust characterization.
[37] Global Cause analysis, literature review.
[38] India Dust characterization, gravimetric analysis, light
absorption.
[39] India Laboratory and outdoor, I–V curve.
[40] Qatar Dust characterization, seasonal variations.
[41] Qatar Dust characterization, cleanliness index
[42] Morocco Dust characterization, monocrystalline silicon,
[43] USA Artificial modelling, VT.
[44] USA Artificial soiling, dust characterization, light absorption
[45] Jordan Artificial soiling, I–V curve, dust characterization
[46] Spain Optical losses, modelling
[47] USA Conversion efficiency, effects of cleaning
[48] USA Optical losses, different ambients.
[49] USA Several ambients, outdoor exposure.
[50] USA, Arabia, Several stations, performance analysis.
Australia
[51] Norway, South Africa VT, electrical measurements.
Detection [114] Southern Italy Economic impact of soiling, several stations, regression
methods.
[134] Germany I–V curve, partial shading.
Prevention [136] Saudi Arabia Different materials, dust characterization,
[137] Saudi Arabia Texturing, coating, desert conditions.
[138] USA, Germany Cleaning, hydrophobic surfaces
[139] USA Hydrophobic surfaces, different configurations
[140] Italy Modelling, Surface treatment, electrostatic treatment
[141] Germany Hydrophobic coating, experimental testing.
[142] Germany Sol-gel processing, porous coating, several materials.
Mitigation [140] Italy Dry cleaning, system modelling, influence of temperature.
[153] Saudi Arabia Review, system description.
[154] Malaysia Literature review, mitigation methods.
[155] Egypt Desert ambient, several products, behaviour modelling.
[156] USA Literature review, mitigation methods, high humidity.
[157] Saudi Arabia, Qatar Dry cleaning, prototype, automated system.
[158] China, Italy Kinect, 3D mapping.
[159] Saudi Arabia Schedule optimization, several materials, cost analysis.
Snow and ice deposition Causes and [52] Norway Snow characterization, light obstruction.
effects [53] USA Instrumentation, image processing, temperature
[22] Canada Literature review, influence factors, mitigation.
[54] USA Power losses, effect of angle.
[55] Canada, USA Image recognition, power losses, different materials.
[56] USA Image processing, modelling,
[57] Canada, USA Power performance, snow characterization, image
processing.
[58] Germany Different materials, image processing.
[59] Canada Snow sliding, modelling.
[62] International Ice, renewable energy, mitigation, literature review.
[63] Canada Ice adhesion, different materials.
Prevention [143] Norway, Canada, USA Snow repulsion, coating comparison.
[144] Norway Surface modificaion, classification.
[145] Canada Literature review, surface modification.
[146] Norway Snow and ice, research opportunities, urban solar cells.
[147] Canada, USA Surface coatings, snow.
[163] Norway, Canada, USA Snow repulsion, coating comparison.
Mitigation [160] Austria Heating system, electrical modelling, thermography
[161] Canada Heating system, different scenarios,
[162] Canada Hot wáter heating system
Outdoor exposure: corrosion, delamination and Causes and [21] International Protocol and tool establishment, organic, corrosion.
degradation effects [64] International Organic PV, laboratory, corrosion.
[65] South Korea Accelerated tests, corrosion evolution.
[66] USA Thin film modules, temperature dependence, water vapour,
corrosion.
[67] Germany Organic, EL, corrosion.
[68] USA Adhesion, modelling, metrology.
[69] USA Laboratory technique, procedure proposal.
[70] USA Adhesion, delamination, modelling.
[71] USA Packaging materials, water ingress.
[72] South Korea Electrical characterization, outdoor conditions.
[73] India Different NDT, cause identification, different modules.
[74] South Africa Outdoor conditions, moisture ingress.
[75] Germany Cause analysis, EL, laboratory testing, snail trails.
[76] China Urban ambient, EL, thermography, snail trails
(continued on next page)
10
Table 2 (continued )
DEGRADATION MECHANISM TYPE OF STUDY STUDIES LOCATION TOPICS/KEYWORDS
RELATED
[77] Benin Organic PV, literature review.

[78] South Africa Thin-film module, outdoor exposure.
[79] Algeria Degradation, soiling, delamination.
[80] Spain Long term exposure, thin film, modelling.
[81] Thailand Long term exposure, tropical ambient, comparison with
other ambients.
[82] India Long term exposure, tropical ambient, IR, VI.
Detection [113] United Kingdom Metal substrates, light measurement.
[133] Venezuela I–V curve, Matlab modelling.
Cracks Causes and [84] Germany Classification, EL electrical characterization,
effects [85] Germany Mechanical testing, FEM, EL.
[86] Lithuania Dynamical loads, FEM.
[87] Germany, Spain Electrical resistance measurement, mechanical testing.
[88] Germany FEM, mechanical testing, several materials.
[89] United Kingdom, 3D modelling, microcracks.
Algeria
[90] China Wavelet, Gaussian pyramid.
[91] United Kingdom EL, outdoor conditions, power output.
[92] Germany Electrical modelling, thermography, EL.
Detection [26] France, Germany Literature review, image processing, classification.
Hot spots Causes and [95] Spain I–V curve, thermography, EL.
effects [96] Brazil UAV, I–V curve, thermography.
[97] Sweden I–V curve, polycrystalline silicon.
[98] South Africa Inspection in darkness, laboratory, induced current.
[99] Germany Process control, thermography.
[100] China Simulation, impact of defects.
[101] Norway FEM, different types.
[102] India Modelling and validation, different materials.
[103] USA Induced current, electrical and thermal modelling.
[104] USA Detection, electrical modelling.
[105] Germany Partial shading, induced current, simulation.
[106] Italy Shading, thermal modelling,
Detection [130] Italy IR, UAV, remote monitoring.
[131] Italy IR, UAV, remote monitoring.
Fatigue Prevention [148] Italy Mechanical and recovery testing, FEM analysis.
[149] Germany Fatigue, numerical testing, connectors.
[150] USA, Japan Mechanical testing, FEM simulations, several conditions.
[151] Germany Different temperatures, microscopic analysis.
[152] Germany Thermo-mechanical testing, system modelling.
material is also a field with little information. modifications and coatings, forcing a compromise between efficiency
Short-term appearing degradation mechanisms, as instantaneous and failure avoidance. Fatigue, both thermal and mechanical, is one of
cracks or monthly soiling and snow deposition, usually lead to long term the main factors leading to disconnections and electrical failures over
ones, e.g. degradation and corrosion. This implies that research and time, that can be minimized but not avoided.
prevention on the former ones is likely to reduce the effects and Regarding mitigation, research aims to solve problems associated
appearance of the latter ones. Therefore, research on environmental with soiling and snow deposition, as these are the only reversible
issues affecting PV modules, i.e. production and operation, methods that degradation mechanisms. Generally, automated or dry-cleaned routines
prevent mechanical loads on the PV panels must be addressed to in are applied on the modules to minimize their effects.
crease their life service and profitability. In summary, detection, prevention and mitigation techniques are
employed to ensure a proper work of PV modules. The references on
5.2. Discussion on detection, prevention and mitigation techniques prevention and mitigation have been usually focused on soiling and
snow avoidance, but rarely considers other defects, leading to a
There are a large number of different technologies applied to detect, knowledge gap on this field. The development of artificial intelligence
prevent and mitigate degradation mechanisms on PV energy. Most of and big data techniques can generate useful tools to maintain solar farms
them can be detected through CMS and non-destructive testing [127, due the amount and variety of data. Therefore, research on this field is
167]. The main references about CMS systems is collected in Table 4, increasing.
considering also the techniques, the localisation and the content of the
study. 5.3. Future challenges and trends
Thermography is one of the most relevant techniques, as many de
fects cause heat variations on the affected component. Its combination A quantitative analysis of the state of the art is presented in this
with UAVs and image processing made it a good approach for quick and section. This analysis shows the approximate amount of publications
automatic inspection of solar farms. I–V and P–V curves also provide regarding degradation mechanisms in PV technology in Fig. 8, and the
information on the type and location of faults, and the application of proportion of publications on each defect in Fig. 9. Data has been ob
data processing techniques allow to minimize maintenance time and tained by searching specific keywords in the main scientific databases
costs. (SCOPUS, Web of Science, IEEE Xplore, ScienceDirect), for instance
Given the irreversible nature of most of failure mechanisms in PV, “photovoltaic” plus the corresponding defect for each year. These data
the research is more focused on prevention than mitigation, mainly on should be considered as an approximation of the values providing sta
soiling and snow deposition. Most of the techniques are based on surface tistical information. A positive trend is observed for references focused
11
on every defect, where references on the publications on corrosion and locations, and that it has a short-term effect that can be mitigated
snow deposition are double. Corrosion is being, therefore, the most through proper maintenance actions. It has been detected that the
studied fault. Finally, publications on soiling have multiplied by 7 in the number of studies published about soiling has doubled in the last ten
last ten years. years. Regarding the snow or ice deposition, a lack of studies has been
There have not been significant variations in the latest years ac detected in this field, being most of them oriented to prevent the ice, and
cording to the proportions of publications related to defects in PV very few to evaluate the losses produced by this phenomenon. Among
technology. Corrosion by itself focuses almost 40% every year and, the damaging agents due to weathering, corrosion is the most studied
together with cracks and snow deposition, sum up to 70% of publica one (around 40% of all the publications about faults of photovoltaics).
tions. Fatigue, delamination and hot spots are evenly distributed around This fault type causes irreversible damages that affect severely to the
8%, and soiling has duplicated the proportion of publications since 2009 power output of the PV modules.
from 2.5% to 5.8%. Cracks are mentioned in more than 15% of the publications about PV
The following limitations and future challenges are extracted defects. They can be caused by multiple agents, and most of the litera
considering the quantitative analysis done: ture focuses on their detection, usually through electroluminescence
variations. Other fault that can be generated by multiple causes are
- Most of research on PV degradation mechanisms focuses on corro hotspots, which produce temperature gradients and important power
sion, snow deposition and crack apparition. Other faults should be losses. Thermography is a usually employed technique for its detection.
considered in the researches because there are not many publications In order to ensure a correct state of the PV modules and solar farms,
focused on them. There is also a lack of information on how degra
dation mechanisms affect the different PV materials and technolo
Table 4
gies, considering both the likeliness and magnitude in which they
Failure detection, prevention and mitigation studies in PV technology.
may be affected.
- Though publications in soiling have increased considerably, their Technique Studies Location Content
related
number is still much lower in respect to the other degradation
mechanisms, and usually focused on dust characterization or pre Condition [25] Tunisia Literature review,
Monitoring and modelling, several defects.
vention and mitigation. Research on short and long-term effects
failure detection [109] USA Inspection form, several
could be carried out to increase comprehension on this defect. locations.
- Few publications have been found to study the consequences of the [111] Italy Online monitoring, data
appearance of different degradation mechanisms on the same mod processing and acquisition.
ules. A proper combination of these defects could be addressed to [112] Denmark Online monitoring, failure
prediction.
study the effects of this combination. [115] Qatar LabVIEW, online
- Since development of PV technology has occurred in a relatively monitoring, literature
small period of time, there are few studies on the effects of long-term review.
degradation of the modules or their treatment after service. Research [116] Spain LabVIEW, data comparison
and simulation.
on this topic is likely to increase as more solar modules reach the end
[117] Saudi Arabia, LabVIEW, data processing.
of their life cycle, offering research opportunities on this field. Egypt
[118] Spain Arduino, autonomous
6. Conclusions system, live monitoring.
[119] Malaysia Literature review, Zigbee,
online monitoring.
The paper has presented an overview of the main faults and degra [120] Germany High speed camera, image
dation mechanisms associated to photovoltaic (PV) installations. processing, IR.
Moreover, the principal techniques procedures employed to mitigate [121] India Fuzzy logic, machine
learning, image processing.
them have been analysed.
[122] Germany Image processing, I–V
Although PV Technology can be employed in a wide range of systems curve, thermography.
due to its scalability, this paper focuses on the industrial field where the [123, UK Fault detection, P–V curve
solar modules must enface environmental factors that cause degrada 124] modelling.
tions and defects. Soiling, snow deposition, corrosion or delamination, [125, Algeria, Saudi Statistical detection, P–V
126] Arabia, Spain curve modelling, different
are the main phenomena caused by the weather conditions exposure of models.
the PV modules. In addition, cracks or hot spots can also appear causing [135] South Africa I–V curve, shading, shunt
losses in the efficiency of the PV installation. resistance.
The references considered state that soiling usually affects to desert
Table 3
Degradation mechanisms in PV technology and studies related to them.
PV energy Publications Population Area [Km^2] GDP [billions of Normal direct irradiation per R&D expenditure [billions
production [GWh] [164] [164] USD] [165] year [kWh/m2] [166] of USD] [164]
China 176,1 32353 1,40E+09 9,60E+06 14216,503 1491 553,4

USA 62,6 29421 3,29E+08 9,16E+06 21344,667 1892 511,1
Japan 56 23366 1,26E+08 378000 5176,205 1112 165,7
Germany 45,4 26819 8,31E+07 357000 3963,88 900 118,8
India 32,9 17481 1,38E+09 3,29E+06 2971,996 1573 66,5
Italy 20,1 13177 6,05E+07 3,01E+05 2025,866 1582 27,4
UK 13 14987 6,66E+07 2,44E+05 2829,163 710 44,8
Australia 11,3 12024 2,52E+07 7,69E+06 1417,003 2759 23,3
France 9 16762 6,52E+07 6,44E+05 2761,633 1157 60
S. Korea 7,9 9484 5,18E+07 1,21E+05 1656,674 1301 91,6
Pearson’s correlation coefficient 0,82 0,68 0,69 0,66 0,00 0,85
with PV energy production
12
Fig. 8. Publications related to defects in PV technology.
Fig. 9. Proportion of publications related to defects in PV technology.
condition monitoring and detection techniques are employed. Novel of the named snail trails on organic materials; A survey on the economic
technologies and approaches, such as unmanned aerial vehicles or and energy production losses associated to the apparition of the degra
artificial intelligence, lead to more efficient maintenance tasks and, dation mechanisms, and; The economic and energy production benefits
subsequently, to a better performance of the installations. of the detection, prevention and mitigation techniques applied to PV
Through the literature analysis, it has been detected that preventive power production.
maintenance in photovoltaic energy is less studied than condition
monitoring or fault detection. The prevention activities are usually
Declaration of competing interest
oriented to avoid the apparition of irreversible faults, e.g. corrosion or
cracks.
The authors declare that they have no known competing financial
Research on mitigation is not enough covered, since most of the
interests or personal relationships that could have appeared to influence
faults of PV modules are irreversible, and their effects often cannot be
the work reported in this paper.
attenuated. Methods and techniques for cleaning and dry the modules
are usually the most studied topic. Therefore, the analysis of mitigation
Acknowledgements
techniques on different faults can suppose a research opportunity.
This paper also presents a discussion of the actual context and trends
The work reported herewith has been financially supported by the
of the research on PV energy. Regarding the main PV energy producers,
Spanish Ministerio de Ciencia e Innovación, under the Research Grant
a strong correlation between the amount of publications and the PV
RTC2019-007364-3.
energy production has been detected, which is closely linked to the
research and development expenditure. Moreover, it has been detected
that the solar irradiation is not the factor that produces the main pro References
duction inequalities between these countries.
[1] Márquez FPG, Karyotakis A, Papaelias M. Renewable energies: business outlook
Finally, some limitations and future challenges are identified, for 2050. Springer; 2018.
example: a further explanation of the effects of degradation mechanisms [2] Hosseinabad ER, Moraga RJ. The evaluation of renewable energy predictive
over the different technologies and PV materials, such as the apparition modelling in energy dependency reduction: a system dynamics approach. Int J
Appl Manag Sci 2020;12(1):1–22.
13
[3] Feili H, et al. Ranking of suitable renewable energy location using AHP method [37] Maghami MR, et al. Power loss due to soiling on solar panel: a review. Renew
and scoring systems with sustainable development perspective. In: 6th Sustain Energy Rev 2016;59:1307–16.
international conference on economics, management, EngineeringScience and [38] John JJ, et al. Study of soiling loss on photovoltaic modules with artificially
art; 2018. Brussels, Belgium. deposited dust of different gravimetric densities and compositions collected from
[4] Muñoz CQG, et al. New pipe notch detection and location method for short different locations in India. IEEE journal of photovoltaics 2016;6(1):236–43.
distances employing ultrasonic guided waves. Acta Acustica united Acustica [39] Rao A, et al. Influence of dust deposition on photovoltaic panel performance.
2017;103(5):772–81. Energy Procedia 2014;54:690–700.
[5] Jiménez AA, et al. Artificial intelligence for concentrated solar plant maintenance [40] Javed W, et al. Characterization of dust accumulated on photovoltaic panels in
management. In: Proceedings of the tenth international conference on Doha. Qatar. Solar Energy 2017;142:123–35.
management science and engineering management. Springer; 2017. [41] Javed W, et al. Photovoltaic performance degradation due to soiling and
[6] Zhang HL, et al. Concentrated solar power plants: review and design characterization of the accumulated dust. In: 2016 IEEE international conference
methodology. Renew Sustain Energy Rev 2013;22(Supplement C):466–81. on power and renewable energy (ICPRE). IEEE; 2016.
[7] Fahrenbruch A, Bube R. Fundamentals of solar cells: photovoltaic solar energy [42] Zitouni H, et al. Experimental investigation of the soiling effect on the
conversion. Elsevier Science; 2012. performance of monocrystalline photovoltaic systems. Energy Procedia 2019;157:
[8] Goetzberger A, Hoffmann VU. Photovoltaic solar energy generation. Springer 1011–21.
Berlin Heidelberg; 2005. [43] Burton PD, King BH. Application and characterization of an artificial grime for
[9] Agency, I.R.E., Renewable power generation Costs in 2017. 2018, IRENA: IRENA photovoltaic soiling studies. IEEE Journal of Photovoltaics 2014;4(1):299–303.
Publications. p. 160. [44] Burton PD, King BH. Artificial soiling of photovoltaic module surfaces using
[10] Bp BP. Energy outlook [cited 2018 October]; Available from: https://www.bp. traceable soil components. In: 2013 IEEE 39th photovoltaic specialists conference
com/en/global/corporate/energy-economics/energy-outlook.html; 2018. (PVSC). IEEE; 2013.
[11] Council WE. World energy resources. Solar 2016; 2017 [cited 2019 January]; [45] Saidan M, et al. Experimental study on the effect of dust deposition on solar
Available from: https://www.worldenergy.org/wp-content/uploads/2017/03/ photovoltaic panels in desert environment. Renew Energy 2016;92:499–505.
WEResources_Solar_2016.pdf. [46] García M, et al. Soiling and other optical losses in solar-tracking PV plants in
[12] Papaelias M, Márquez FPG, Ramirez IS. Concentrated solar power: present and navarra. Progress in Photovoltaics: Research 2011;19(2):211–7.
future. In: Renewable energies. Springer; 2018. p. 51–61. [47] Mejia F, Kleissl J, Bosch J. The effect of dust on solar photovoltaic systems.
[13] Pliego Marugán A, García Márquez FP, Lev B. Optimal decision-making via binary Energy Procedia 2014;49:2370–6.
decision diagrams for investments under a risky environment. Int J Prod Res [48] Caron JR, Littmann B. Direct monitoring of energy lost due to soiling on first solar
2017;55(18):5271–86. modules in California. IEEE Jounal of Photovoltaics 2013;3(1):336–40.
[14] Pliego Marugán A, García Márquez FP, Lorente J. Decision making process via [49] Kimber A, et al. The effect of soiling on large grid-connected photovoltaic systems
binary decision diagram. Int J Manag Sci Eng Manag 2015;10(1):3–8. in California and the southwest region of the United States. In: 2006 IEEE 4th
[15] García FP, Pliego A, Trapero JR. A new ranking method approach for decision world conference on photovoltaic energy conference; 2006.
making in maintenance management. In: Proceedings of the seventh international [50] Gostein M, Caron JR, Littmann B. Measuring soiling losses at utility-scale PV
conference on management science and engineering management. Springer; power plants. In: 2014 IEEE 40th photovoltaic specialist conference. PVSC);
2014. 2014.
[16] Moraga R, Rabiei Hosseinabad E. A system dynamics approach in air pollution [51] Pedersen H, Strauss J, Selj J. Effect of soiling on photovoltaic modules in Norway.
mitigation of metropolitan areas with sustainable development perspective: a Energy Procedia 2016;92:585–9.
case study of Mexico City. Journal of Applied Environmental Biological Sciences [52] Andenæs E, et al. The influence of snow and ice coverage on the energy
2017;7(12):164–74. generation from photovoltaic solar cells. Sol Energy 2018;159:318–28.
[17] Hernández-Callejo L, Gallardo-Saavedra S, Alonso-Gómez V. A review of [53] Marion B, Rodriguez J, Pruett J. Instrumentation for evaluating PV system
photovoltaic systems: design, operation and maintenance. Sol Energy 2019;188: performance losses from snow. Golden, CO (United States): National Renewable
426–40. Energy Lab.(NREL); 2009.
[18] Sahoo SK. Renewable and sustainable energy reviews solar photovoltaic energy [54] Heidari N, et al. Impact of snow and ground interference on photovoltaic electric
progress in India: a review. Renew Sustain Energy Rev 2016;59:927–39. system performance. IEEE Journal of Photovoltaics 2015;5(6):1680–5.
[19] Nolden C, Barnes J, Nicholls J. Community energy business model evolution: a [55] Andrews RW, Pollard A, Pearce JM. The effects of snowfall on solar photovoltaic
review of solar photovoltaic developments in England. Renew Sustain Energy Rev performance. Sol Energy 2013;92:84–97.
2020;122:109722. [56] Marion B, et al. Measured and modeled photovoltaic system energy losses from
[20] Costa SCS, Diniz ASAC, Kazmerski LL. Solar energy dust and soiling R&D snow for Colorado and Wisconsin locations. Sol Energy 2013;97:112–21.
progress: literature review update for 2016. Renew Sustain Energy Rev 2018;82: [57] Powers L, Newmiller J, Townsend T. Measuring and modeling the effect of snow
2504–36. on photovoltaic system performance. In: Photovoltaic specialists conference
[21] Reese MO, et al. Consensus stability testing protocols for organic photovoltaic (PVSC), 2010 35th IEEE. IEEE; 2010.
materials and devices. Sol Energy Mater Sol Cell 2011;95(5):1253–67. [58] Bogenrieder J, et al. Technology-dependent analysis of the snow melting and
[22] Pawluk RE, Chen Y, She Y. Photovoltaic electricity generation loss due to snow–A sliding behavior on photovoltaic modules. J Renew Sustain Energy 2018;10(2):
literature review on influence factors, estimation, and mitigation. Renew Sustain 021005.
Energy Rev 2019;107:171–82. [59] Rahmatmand A, Harrison SJ, Oosthuizen PH. Numerical and experimental study
[23] Hamdi RT, et al. Humidity impact on photovoltaic cells performance: a review. of an improved method for prediction of snow melting and snow sliding from
International Journal of Recent Engineering Research Development 2018;3(11): photovoltaic panels. Appl Therm Eng 2019:113773.
27–37. [60] Muñoz CQG, Márquez FPG. Future maintenance management in renewable
[24] Kumar M, Kumar A. Performance assessment and degradation analysis of solar energies. In: Renewable energies. Springer; 2018. p. 149–59.
photovoltaic technologies: a review. Renew Sustain Energy Rev 2017;78:554–87. [61] Pinar-Pérez JM, Márquez FPG. Managing costs and review for icing problems. In:
[25] Triki-Lahiani A, Bennani-Ben Abdelghani A, Slama-Belkhodja I. Fault detection Renewable energies. Springer; 2018. p. 97–109.
and monitoring systems for photovoltaic installations: a review. Renew Sustain [62] Carriveau R, et al. Ice adhesion issues in renewable energy infrastructure. J Adhes
Energy Rev 2018;82:2680–92. Sci Technol 2012;26(4–5):447–61.
[26] Tsanakas JA, Ha L, Buerhop C. Faults and infrared thermographic diagnosis in [63] Pawluk R, Chen Y, She Y. Observations of ice at the interface between snow
operating c-Si photovoltaic modules: a review of research and future challenges. accumulations and photovoltaic panel surfaces. In: 2018 6th international
Renew Sustain Energy Rev 2016;62:695–709. renewable and sustainable energy conference (IRSEC). IEEE; 2019.
[27] Sarkın AS, Ekren N, Sağlam Ş. A review of anti-reflection and self-cleaning [64] Gevorgyan SA, et al. Interlaboratory outdoor stability studies of flexible roll-to-
coatings on photovoltaic panels. Sol Energy 2020;199:63–73. roll coated organic photovoltaic modules: stability over 10,000h. Sol Energy
[28] ( IP), F.P.G.M.. Development and Demonstration of a novel integrated condition Mater Sol Cell 2013;116:187–96.
monitoring System for wind turbines (NIMO) (ref.: FP7-ENERGY-2008-TREN1: [65] Park N, Han C, Kim D. Effect of moisture condensation on long-term reliability of
239462). FP7 program. EUROPEAN - PROJECTS; 2009. crystalline silicon photovoltaic modules. Microelectron Reliab 2013;53(12):
[29] ( IP), F.P.G.M.. Demonstration of methods and tools for the optimisation of 1922–6.
operational reliability of large-scale industrial wind turbines (OPTIMUS) (Ref.: [66] Osterwald CR, McMahon TJ, del Cueto JA. Electrochemical corrosion of SnO2:F
FP7-ENERGY-2012-2-322430). FP7 Program, EUROPEAN - PROJECTS; 2013. transparent conducting layers in thin-film photovoltaic modules. Sol Energy
[30] ( IP), F.P.G.M.. Structural health Monitoring of concentrated solar power plants Mater Sol Cell 2003;79(1):21–33.
(INTERSOLAR) (ref.: FP7-SME-2013-605028). FP7 program. EUROPEAN - [67] Klumbies H, et al. Water ingress into and climate dependent lifetime of organic
PROJECTS; 2013. photovoltaic cells investigated by calcium corrosion tests. Sol Energy Mater Sol
[31] El-Ghoemy. Photovoltaic solar energy: review. Int J Sci Eng Res 2012;3(1). Cell 2014;120:685–90.
[32] Dzimano GJ. Modeling of photovoltaic systems. The Ohio State University; 2008. [68] Tracy J, Bosco N, Dauskardt R. Encapsulant adhesion to surface metallization on
[33] Goetzberger A, Hebling C. Photovoltaic materials, past, present, future. Sol photovoltaic cells. IEEE Journal of Photovoltaics 2017;7(6):1635–9.
Energy Mater Sol Cells 2000;62(1–2):1–19. [69] Jansen KW, Delahoy AE. A laboratory technique for the evaluation of
[34] Polman A, et al. Photovoltaic materials: present efficiencies and future electrochemical transparent conductive oxide delamination from glass substrates.
challenges. Science 2016;352(6283):aad4424. Thin Solid Films 2003;423(2):153–60.
[35] Aghaei M, et al. An overview of solar cells materials classification. In: [70] Tracy J, et al. Encapsulation and backsheet adhesion metrology for photovoltaic
NatGrad2013. Proceedings of the 2nd National Graduate Conference; 2014. modules. Prog Photovoltaics Res Appl 2016;25.
[36] Ilse KK, et al. Fundamentals of soiling processes on photovoltaic modules. Renew [71] Jorgensen GJ, et al. Moisture transport, adhesion, and corrosion protection of PV
Sustain Energy Rev 2018;98:239–54. module packaging materials. Sol Energy Mater Sol Cell 2006;90(16):2739–75.
14
[72] Park NC, et al. The effect of encapsulant discoloration and delamination on the [105] Geisemeyer I, et al. Prediction of silicon PV module temperature for hot spots and
electrical characteristics of photovoltaic module. Microelectron Reliab 2013;53 worst case partial shading situations using spatially resolved lock-in
(9):1818–22. thermography. Sol Energy Mater Sol Cell 2014;120:259–69.
[73] Sinha A, Sastry OS, Gupta R. Nondestructive characterization of encapsulant [106] Rossi D, et al. Modeling and detection of hotspot in shaded photovoltaic cells.
discoloration effects in crystalline-silicon PV modules. Sol Energy Mater Sol Cell IEEE Trans Very Large Scale Integr Syst 2015;23(6):1031–9.
2016;155:234–42. [107] Marquez FG. An approach to remote condition monitoring systems management.
[74] Van Dyk E, Chamel JB, Gxasheka AR. Investigation of delamination in an edge- Railway Condition Monitoring; 2006. p. 156–60.
defined film-fed growth photovoltaic module. Sol Energy Mater Sol Cells 2005; [108] Márquez FPG, Pedregal DJ. Applied RCM 2 algorithms based on statistical
88:403–11. methods. Int J Autom Comput 2007;4(2):109–16.
[75] Meyer S, et al. Snail trails: root cause analysis and test procedures. Energy [109] Packard CE, Wohlgemuth JH, Kurtz SR. Development of a visual inspection data
Procedia 2013;38:498–505. collection tool for evaluation of fielded PV module condition. Golden, CO (United
[76] Yang H, et al. Power degradation caused by snail trails in urban photovoltaic States): National Renewable Energy Lab.(NREL); 2012.
energy systems. Energy Procedia 2016;88:422–8. [110] Muñoz CQG, et al. A new condition monitoring approach for maintenance
[77] Kounouhéwa B, et al. Comparison of degradation mechanisms in organic management in concentrate solar plants. In: Proceedings of the ninth
photovoltaic devices upon exposure to a temperate and a subequatorial climate. international conference on management science and engineering management.
Chem Phys Lett 2015;640:201–14. Springer; 2015.
[78] van Dyk EE, et al. Investigation of the degradation of a thin-film hydrogenated [111] Tina GM, Grasso AD. Remote monitoring system for stand-alone photovoltaic
amorphous silicon photovoltaic module. Sol Energy Mater Sol Cell 2007;91(2): power plants: the case study of a PV-powered outdoor refrigerator. Energy
167–73. Convers Manag 2014;78:862–71.
[79] Bouraiou A, et al. Experimental evaluation of the performance and degradation of [112] Spataru S, et al. Photovoltaic array condition monitoring based on online
single crystalline silicon photovoltaic modules in the Saharan environment. regression of performance model. In: Photovoltaic specialists conference (PVSC),
Energy 2017;132:22–30. 2013 IEEE 39th. IEEE; 2013.
[80] Silvestre S, et al. Degradation analysis of thin film photovoltaic modules under [113] Watson T, et al. Corrosion monitoring of flexible metallic substrates for dye-
outdoor long term exposure in Spanish continental climate conditions. Sol Energy sensitized solar cells. Int J Photoenergy 2013;2013:8.
2016;139:599–607. [114] Massi Pavan A, Mellit A, De Pieri D. The effect of soiling on energy production for
[81] Limmanee A, et al. Degradation analysis of photovoltaic modules under tropical large-scale photovoltaic plants. Sol Energy 2011;85(5):1128–36.
climatic conditions and its impacts on LCOE. Renew Energy 2017;102:199–204. [115] Touati F, et al. Investigation of solar PV performance under Doha weather using a
[82] Chandel SS, et al. Degradation analysis of 28 year field exposed mono-c-Si customized measurement and monitoring system. Renew Energy 2016;89:
photovoltaic modules of a direct coupled solar water pumping system in western 564–77.
Himalayan region of India. Renew Energy 2015;78:193–202. [116] Chouder A, et al. Monitoring, modelling and simulation of PV systems using
[83] Arcos Jiménez A, Gómez Muñoz C, García Márquez F. Machine learning for wind LabVIEW. Sol Energy 2013;91:337–49.
turbine blades maintenance management. Energies 2018;11(1):13. [117] Rezk H, et al. Performance of data acquisition system for monitoring PV system
[84] Köntges M, et al. Crack statistic of crystalline silicon photovoltaic modules. In: parameters. Measurement 2017;104:204–11.
26th European photovoltaic solar energy conference and exhibition; 2011. [118] Fuentes M, et al. Design of an accurate, low-cost autonomous data logger for PV
[85] Kajari-Schröder S, et al. Spatial and orientational distribution of cracks in system monitoring using Arduino™ that complies with IEC standards. Sol Energy
crystalline photovoltaic modules generated by mechanical load tests. Sol Energy Mater Sol Cell 2014;130:529–43.
Mater Sol Cell 2011;95(11):3054–9. [119] Shariff F, Rahim NA, Hew WP. Zigbee-based data acquisition system for online
[86] Kilikevičius A, Čereška A, Kilikevičienė K. Analysis of external dynamic loads monitoring of grid-connected photovoltaic system. Expert Syst Appl 2015;42(3):
influence to photovoltaic module structural performance. Eng Fail Anal 2016;66: 1730–42.
445–54. [120] Gupta R, et al. In-line shunt detection in solar cells by fast lock-in infrared
[87] Käsewieter J, et al. Cracks in solar cell metallization leading to module power loss thermography. In: Proc. 22nd European photovoltaic solar energy conference and
under mechanical loads. Energy Procedia 2014;55:469–77. exhibition 2007; 2007.
[88] Sander M, et al. Systematic investigation of cracks in encapsulated solar cells after [121] Jaffery ZA, et al. Scheme for predictive fault diagnosis in photo-voltaic modules
mechanical loading. Sol Energy Mater Sol Cell 2013;111:82–9. using thermal imaging. Infrared Phys Technol 2017;83:182–7.
[89] Dhimish M, et al. Development of 3D graph-based model to examine photovoltaic [122] Buerhop C, et al. Reliability of IR-imaging of PV-plants under operating
micro cracks. J Sci: Advanced Materials and Devices 2018;3(3):380–8. conditions. Sol Energy Mater Sol Cell 2012;107:154–64.
[90] Song M, et al. Crack detection algorithm for photovoltaic image based on multi- [123] Dhimish M, et al. Diagnostic method for photovoltaic systems based on six layer
scale pyramid and improved region growing. In: 2018 IEEE 3rd international detection algorithm. Elec Power Syst Res 2017;151:26–39.
conference on image, vision and computing (ICIVC). IEEE; 2018. [124] Dhimish M, Holmes V, Dales M. Parallel fault detection algorithm for grid-
[91] Dhimish M, et al. The impact of cracks on photovoltaic power performance. J Sci: connected photovoltaic plants. Renew Energy 2017;113:94–111.
Advanced Materials and Devices 2017;2(2):199–209. [125] Harrou F, et al. Reliable fault detection and diagnosis of photovoltaic systems
[92] van Mölken JI, et al. Impact of micro-cracks on the degradation of solar cell based on statistical monitoring approaches. Renew Energy 2018;116:22–37.
performance based on two-diode model parameters. Energy Procedia 2012;27: [126] Garoudja E, et al. Statistical fault detection in photovoltaic systems. Sol Energy
167–72. 2017;150:485–99.
[93] Herraiz ÁH, Marugán AP, Márquez FPG. Photovoltaic plant condition monitoring [127] Márquez FPG. A new method for maintenance management employing principal
using thermal images analysis by convolutional neural network-based structure. component analysis. Struct Durab Health Monit 2010;6(2):89–99.
Renew Energy 2020;153:334–48. [128] Herraiz ÁH, et al. Remotely operated vehicle applications. In: Non-destructive
[94] Ramirez IS, Marquez FPG. fault detection and identification for maintenance testing and condition monitoring techniques for renewable energy industrial
management. In: International conference on management science and assets. Elsevier; 2020. p. 119–32.
engineering management. Springer; 2020. [129] Herraiz ÁH, Marugán AP, Márquez FPG. A review on condition monitoring system
[95] Moretón R, Lorenzo E, Narvarte L. Experimental observations on hot-spots and for solar plants based on thermography. In: Non-destructive testing and condition
derived acceptance/rejection criteria. Sol Energy 2015;118:28–40. monitoring techniques for renewable energy industrial assets. Elsevier; 2020.
[96] Oliveira AKVd, et al. Aerial infrared thermography of a utility-scale PV plant after p. 103–18.
a meteorological tsunami in Brazil. In: 2018 IEEE 7th world conference on [130] Quater PB, et al. Light unmanned aerial vehicles (UAVs) for cooperative
photovoltaic energy conversion (WCPEC) (A joint conference of 45th IEEE PVSC, inspection of PV plants. IEEE Journal of Photovoltaics 2014;4(4):1107–13.
28th PVSEC & 34th EU PVSEC); 2018. [131] Grimaccia F, et al. Planning for PV plant performance monitoring by means of
[97] Bharadwaj P, Karnataki K, John V. formation of hotspots on healthy PV modules unmanned aerial systems (UAS). International Journal of Energy and
and their effect on output performance. In: 2018 IEEE 7th world conference on Environmental Engineering 2015;6(1):47–54.
photovoltaic energy conversion (WCPEC) (A joint conference of 45th IEEE PVSC, [132] Márquez FPG, Ramírez IS. Condition monitoring system for solar power plants
28th PVSEC & 34th EU PVSEC); 2018. with radiometric and thermographic sensors embedded in unmanned aerial
[98] Simon M, Meyer E. Detection and analysis of hot-spot formation in solar cells. Sol vehicles. Measurement 2019;139:152–62.
Energy Mater Sol Cells 2010;94:106–13. [133] González-Longatt FM. Model of photovoltaic module in Matlab. Ii Cibelec; 2005.
[99] Ramspeck K, et al. In-line thermography for reliable hot spot detection and p. 1–5. 2005.
process control. Energy Procedia 2014;55:133–40. [134] Schill C, Brachmann S, Koehl M. Impact of soiling on IV-curves and efficiency of
[100] Deng S, et al. Research on hot spot risk for high-efficiency solar module. Energy PV-modules. Sol Energy 2015;112:259–62.
Procedia 2017;130:77–86. [135] Meyer EL, Van Dyk EE. The effect of reduced shunt resistance and shading on
[101] Solheim HJ, et al. Measurement and simulation of hot spots in solar cells. Energy photovoltaic module performance. In: Photovoltaic specialists conference, 2005.
Procedia 2013;38:183–9. Conference record of the thirty-first IEEE. IEEE; 2005.
[102] Rajput P, Tiwari GN, Sastry OS. Thermal modelling and experimental validation [136] Bahattab MA, et al. Anti-soiling surfaces for PV applications prepared by sol-gel
of hot spot in crystalline silicon photovoltaic modules for real outdoor condition. processing: comparison of laboratory testing and outdoor exposure. Sol Energy
Sol Energy 2016;139:569–80. Mater Sol Cell 2016;157:422–8.
[103] Kim KA, Krein PT. Photovoltaic hot spot analysis for cells with various reverse- [137] Walwil HM, et al. Comparative studies of encapsulation and glass surface
bias characteristics through electrical and thermal simulation. In: 2013 IEEE 14th modification impacts on PV performance in a desert climate. Sol Energy 2017;
workshop on control and modeling for power electronics (COMPEL); 2013. 142:288–98.
[104] Kim KA, et al. Photovoltaic hot-spot detection for solar panel substrings using AC [138] Bhushan B, Jung YC, Koch K. Self-cleaning efficiency of artificial
parameter characterization. IEEE Trans Power Electron 2016;31(2):1121–30. superhydrophobic surfaces. Langmuir 2009;25(5):3240–8.
15
[139] Nayshevsky I, Xu Q, Lyons AM. Hydrophobic–hydrophilic surfaces exhibiting [154] Jamil WJ, et al. Performance degradation of photovoltaic power system: review
dropwise condensation for anti-soiling applications. IEEE Journal of on mitigation methods. Renew Sustain Energy Rev 2017;67:876–91.
Photovoltaics 2019;9(1):302–7. [155] Moharram KA, et al. Influence of cleaning using water and surfactants on the
[140] Tucci M, et al. Dry cleaning process of crystalline silicon surface in ⃞
a Si: H/c
⃞ Si performance of photovoltaic panels. Energy Convers Manag 2013;68:266–72.
heterojunction for photovoltaic applications. Thin Solid Films 2002;403:307–11. [156] Mazumder MK, et al. Environmental degradation of the optical surface of PV
[141] Fathi M, Abderrezek M, Friedrich M. Reducing dust effects on photovoltaic panels modules and solar mirrors by soiling and high RH and mitigation methods for
by hydrophobic coating. Clean Technol Environ Policy 2017;19(2):577–85. minimizing energy yield losses. In: Photovoltaic specialist conference (PVSC),
[142] Glaubitt W, Löbmann P. Anti-soiling effect of porous SiO 2 coatings prepared by 2015 IEEE 42nd. IEEE; 2015.
sol–gel processing. J Sol Gel Sci Technol 2011;59(2):239. [157] Aly SP, et al. Novel dry cleaning machine for photovoltaic and solar panels. In:
[143] Andersson P-O, Jelle BP, Zhang Z. Passive snow repulsion: a state-of-the-art Renewable and sustainable energy conference (IRSEC), 2015 3rd international.
review illuminating research gaps and possibilities. Energy Procedia 2017;132: IEEE; 2015.
423–8. [158] Li M, et al. Fast and robust mapping with low-cost Kinect V2 for photovoltaic
[144] Jelle BP, et al. Avoiding snow and ice formation on exterior solar cell surfaces–a panel cleaning robot. In: Advanced robotics and mechatronics (ICARM),
review of research pathways and opportunities. Procedia Engineering 2016;145: international conference on. IEEE; 2016.
699–706. [159] Jones RK, et al. Optimized cleaning cost and schedule based on observed soiling
[145] Fillion RM, Riahi AR, Edrisy A. A review of icing prevention in photovoltaic conditions for photovoltaic plants in central Saudi Arabia. IEEE Journal of
devices by surface engineering. Renew Sustain Energy Rev 2014;32:797–809. Photovoltaics 2016;6(3):730–8.
[146] Jelle BP. The challenge of removing snow downfall on photovoltaic solar cell [160] Weiss A, Weiss H. Photovoltaic cell electrical heating system for removing snow
roofs in order to maximize solar energy efficiency—research opportunities for the on panel including verification. In: 2016 IEEE international conference on
future. Energy Build 2013;67:334–51. renewable energy research and applications (ICRERA). IEEE; 2016.
[147] Andrews RW, Pollard A, Pearce JM. A new method to determine the effects of [161] Rahmatmand A, Harrison SJ, Oosthuizen PH. An experimental investigation of
hydrodynamic surface coatings on the snow shedding effectiveness of solar snow removal from photovoltaic solar panels by electrical heating. Sol Energy
photovoltaic modules. Sol Energy Mater Sol Cell 2013;113:71–8. 2018;171:811–26.
[148] Paggi M, et al. Fatigue degradation and electric recovery in Silicon solar cells [162] Rahmatmand A, Harrison SJ, Oosthuizen PH. Evaluation of removing snow and
embedded in photovoltaic modules. Sci Rep 2014;4:4506. ice from photovoltaic-thermal (PV/T) panels by circulating hot water. Sol Energy
[149] Pander M, et al. Lifetime estimation for solar cell interconnectors. In: 28th 2019;179:226–35.
European photovoltaic solar energy conference; 2013. [163] Andrews RW, et al. A new method to determine the effects of hydrodynamic
[150] Bosco N, et al. Evaluation of dynamic mechanical loading as an accelerated test surface coatings on the snow shedding effectiveness of solar photovoltaic
method for ribbon fatigue. In: Photovoltaic specialists conference (PVSC), 2013 modules. Sol Energy Mater 2013;113:71–8.
IEEE 39th. IEEE; 2013. [164] De Paiva DF, et al. Decision making on generator for wind turbines using the AHP
[151] Wiese S, Meier R, Kraemer F. Mechanical behaviour and fatigue of copper ribbons methodology. In: 2019 IEEE milan PowerTech, PowerTech 2019; 2019.
used as solar cell interconnectors. In: Thermal, mechanical & multi-physics [165] Zhong S, et al. A reliability-and-cost-based fuzzy approach to optimize preventive
simulation, and experiments in microelectronics and microsystems (EuroSimE), maintenance scheduling for offshore wind farms. Mech Syst Signal Process 2019;
2010 11th international conference on. IEEE; 2010. 124:643–63.
[152] Meier R, et al. Reliability of copper-ribbons in photovoltaic modules under [166] Energydata.info. Global solar atlas. In: [cited 2019 15th december 2019];
thermo-mechanical loading. In: Photovoltaic specialists conference (PVSC), 2010 available from: Global Wind Atlas; 2019. https://globalsolaratlas.info/.
35th IEEE. IEEE; 2010. [167] Márquez FPG, Papaelias M, Karyotakis A. Introduction to non-destructive testing
[153] Alshehri A, et al. Dust mitigation in the desert: cleaning mechanisms for solar and condition monitoring techniques for renewable energy industrial assets.
panels in arid regions. In: Smart grid conference (SASG), 2014 Saudi Arabia. IEEE; Elsevier; 2020.
2014.
16

Maintenance Management For Photovoltaic Power System Survey

Uploaded by

Copyright:

Available Formats

Maintenance Management For Photovoltaic Power System Survey

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Maintenance Management For Photovoltaic Power System Survey

Uploaded by

Copyright:

Available Formats

Renewable and Sustainable Energy Reviews 134 (2020) 110347

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews

Survey of maintenance management for photovoltaic power systems

Fig. 3. Classification of PV systems. Figure adapted from Ref. [31].

Fig. 4. Degradation of a 25 year old PV module [72].

Fig. 5. Crack classification for crack sorting [84].

Soiling Causes and [20] Global Literature review.

[77] Benin Organic PV, literature review.

China 176,1 32353 1,40E+09 9,60E+06 14216,503 1491 553,4

Fig. 8. Publications related to defects in PV technology.

Fig. 9. Proportion of publications related to defects in PV technology.

You might also like