-
Solar Energy Materials & Solar Cells 84 (2004) 255274
The applicability of accelerated life testingfor assessment of
service life of solar
thermal components
B. Carlssona,*, K. M .ollera, M. K .ohlb, M. Heckb, S.
Brunoldc,U. Freic, J.-C. Marechald, G. Jorgensene
aSP Swedish National Testing and Research Institute, P.O. Box
857, SE-501 15, Bor (as, SwedenbFraunhofer Institute for Solar
Energy Systems (ISE), Heidenhofstr. 2, D-79110 Freiburg, Germanyc
Institut f .ur Solartechnik (SPF), Hochschule Rapperswil HSR,
CH-8640 Rapperswil, Switzerland
dCSTB Centre Scientific et Technique du Batiment, F-38400
Saint-Martin DHeres, FranceeNational Renewable Energy Laboratory
(NREL), Golden, CO 80401-3393, USA
Received 31 October 2003; accepted 16 January 2004
Available online 2 June 2004
Abstract
To achieve successful commercialisation of new advanced windows
and solar fa@adecomponents for buildings, the durability of these
need to be demonstrated prior to installation
by use of reliable and well-accepted test methods. In Task 27
Performance of Solar Facade
Components of the IEA Solar Heating and Cooling Programme work
has therefore been
undertaken with the objective to develop a general methodology
for durability test procedures
and service lifetime prediction methods adaptable to the wide
variety of advanced optical
materials and components used in energy efcient solar thermal
and buildings applications. As
the result of this work a general methodology has been
developed. The proposed methodology
includes three steps: (a) initial risk analysis of potential
failure modes, (b) screening testing/
analysis for service life prediction and microclimate
characterisation, and (c) service life
prediction involving mathematical modelling and life
testing.
The applicability of the working scheme to be employed in the
development of durability
test procedures has been analysed for selective solar absorber
surfaces and polymeric glazing
materials in at plate solar collectors. The examples show the
great applicability of the general
methodology for accelerated life testing. This will allow much
shorter development cycle times
ARTICLE IN PRESS
*Corresponding author.
E-mail address: [email protected] (B. Carlsson).
0927-0248/$ - see front matter r 2004 Elsevier B.V. All rights
reserved.
doi:10.1016/j.solmat.2004.01.046
-
for new products and will allow improvements to be identied and
readily incorporated in new
products prior to market introduction.
r 2004 Elsevier B.V. All rights reserved.
Keywords: Solar thermal materials; Durability; Service life
prediction; Accelerated testing; Selective solar
absorber surface; Polymeric glazing material
1. Introduction
To achieve successful and sustainable commercialisation,
building products mustmeet three important criteria, namely minimum
cost, sufcient performance, anddemonstrated durability.Durability
assessment directly addresses all three segments of this triad.
First, it
permits analysis of life cycle costs by providing estimates of
service lifetime, O&Mcosts, and realistic warranties.
Understanding how performance parameters areaffected by
environmental stresses (for example by failure analysis) allows
improvedproducts to be devised. Finally, mitigation of known causes
of failure directly resultsin increased product longevity. Thus,
accurate assessment of durability is of para-mount importance to
assuring the success of solar thermal and building products.The IEA
Solar Heating and Cooling Programme, Task 27 on the Performance
of
Solar Facade Components started at the beginning of year 2000
with the objectivesof developing and applying appropriate methods
for assessment of durability,reliability and environmental impact
of advanced components for solar buildingfacades.For the work on
durability there are two main objectives. The rst was to
develop
a general framework for durability test procedures and service
lifetime predictionmethods that are applicable to a wide variety of
advanced optical materials andcomponents used in energy efcient
solar thermal and buildings applications. Thesecond is to apply the
appropriate durability test tools to specic materials/components to
allow prediction of service lifetime and to generate proposals
forinternational standards.This paper presents a general
methodology to meet the rst objective. A thorough
description of the methodology can be found in the Final Report
of IEA SolarHeating and Cooling Task 27 Project B1 on Durability
Assessment MethodologyDevelopment [1]. The methodology is also a
part of the forthcoming bookPerformance and Durability Assessment
of Optical Materials for Solar ThermalSystems [2].
2. General methodology
Many efforts have been made to develop systematic approaches to
service lifeprediction of components, parts of components and
materials so that all essentialaspects of the problem will be taken
into consideration, see e.g. Refs. [38].
ARTICLE IN PRESSB. Carlsson et al. / Solar Energy Materials
& Solar Cells 84 (2004) 255274256
-
In one such methodology, which was the focus in IEA Task 27, a
predictive failuremodes and effect analysis (FMEA) serves as the
starting point for service lifeprediction from accelerated life
test results as is illustrated schematically in Fig. 1.The analysis
is made on the component level.The diagram in Fig. 1 is based on a
similar scheme that was developed for the
purpose of accelerated life testing of selective solar absorber
surfaces in a joint casestudy of Task 10 of the IEA Solar Heating
and Cooling Program [8].
PENALTY is the level at which an assessment is made of the
economic effects of acomponent failure. Based on this assumption,
it is possible to set a reliability levelthat must be maintained
for a given number of years.
FAILURE is the level at which performance requirements are
determined. If therequirements are not fullled, the particular
component or part of component isregarded as having failed.
Performance requirements can be formulated on the basisof optical
properties, mechanical strength, aesthetic values or other criteria
related tothe performance of the component and its materials.
DAMAGE describes the stage of failure analysis at which various
types of damage,each capable of resulting in failure, can be
identied.
CHANGE is related to the change in the material composition or
structure thatcan give rise to the damage of the type previously
identied.
EFFECTIVE STRESS is the level at which various factors in the
microclimate,capable of being signicant for the durability of the
component and its materials can
ARTICLE IN PRESS
A. Initial risk analysis of
potential failuremodes
B. Qualification testing andscreening testing/analysis for
service life prediction
C. Service lifeprediction
PENALTY
Performance requirement
FAILURE
Qualification testing(envi- Mathematical model- ling (rate of
degrada-
DAMAGE Screening testing (acceler-ated aging at elevated
levels
tion process in terms of effective levels of deg-
of the degradation factors) Analysis of material change
radation factors) Life testing
CHANGE mechanisms)
Assessment of ex- pected service life
Reasonability of as- sessment/validation
EFFECTIVESTRESS
Microclimate characteriza- tion(influence of degradation
factors at materials level)
LOADS
(minimum material property)
ronmental resistance testing)
(identification of degradation
Fig. 1. Failure mode analysis for planning of accelerated tests
for service life prediction.
B. Carlsson et al. / Solar Energy Materials & Solar Cells 84
(2004) 255274 257
-
be identied. An important point here is that it is possible to
make quantitativecharacterisation.
LOADS is the level that describes the macro-environmental
conditions (climatic,chemical, mechanical), and which is therefore
a starting point for description of themicroclimate or effective
stress as above.Each step in the scheme on the left hand side of
Fig. 1 may be related to the
subsequent step by an appropriate deterministic or statistical
relationship. Therelation should dene the expected results of all
the various activities involved inaccelerated life testing, as
indicated on the right hand side of Fig. 1.
3. Initial risk analysis of potential failure modes
The rst step in the scheme illustrated in Fig. 1 is an analysis
of potential failuremodes with the aim of obtaining (a) a checklist
of potential failure modes of thecomponent and associated with
those risks and critical component and materialproperties,
degradation processes and stress factors, (b) a framework for the
selectionof test methods to verify performance and service life
requirements, (c) a frameworkfor describing previous test results
for a specic component and its materials or asimilar component and
materials used in the component and classifying their relevanceto
the actual application, and (d) a framework for compiling and
integrating all dataon available component and material properties
and material degradation technology.From a practical point of view,
but also from an economic viewpoint, an
assessment of durability or service life has to be limited in
its scope and focused onthe most critical failure modes. An
important part of the initial step in such anassessment is
therefore estimating the risk associated with each of the
potentialfailure modes of the component.The programme of work in
the initial step of service life assessment may be
structured into the following activities [9]: (a) Specify from
an end-user point of viewthe expected function of the component and
its materials, its performance and itsservice life requirement, and
the intended in-use environments; (b) Identifyimportant functional
properties dening the performance of the component andits
materials, relevant test methods and requirements for qualication
of thecomponent with respect to performance; (c) Identify potential
failure modes anddegradation mechanisms, relevant durability or
life tests and requirements forqualication of the component and its
materials as regards durability. Whenidentifying potential failure
modes, it is important to distinguish between (1) failuresinitiated
by the short-term inuence of environmental stress, the latter
representingevents of high environmental loads on the component and
its materials, (2) failuresinitiated by the long-term inuence of
environmental stress, the latter causingmaterial degradation so
that the performance and sometimes also the environmentalresistance
of the component and its materials gradually decrease, and (d)
estimationof the risk associated with each failure mode
identied.The result of the initial risk analysis of potential
failure modes may be documented
as shown in Tables 1ac using information from a case on
accelerated life testing
ARTICLE IN PRESSB. Carlsson et al. / Solar Energy Materials
& Solar Cells 84 (2004) 255274258
-
performed in Task 10 of the IEA Solar Heating and Cooling
Programme [8]; see alsoFig. 2.The rst activity species in general
terms the function of the component and
service life requirement from an end-user and product point of
view; see Table 1a,and from that identies the most important
functional properties of the componentand its materials; see Table
1b.How important the function of the component is from an end-user
and product
point of view needs to be taken into consideration when
formulating theperformance requirements in terms of those
functional properties. If theperformance requirements are not
fullled, the particular component is regardedas having failed.
Performance requirements can be formulated on the basis of
opticalproperties, mechanical strength, aesthetic values or other
criteria related to theperformance of the component and its
materials. Dening performance requirementshould be accompanied by
an assessment of the economic effects of a componentfailure. Based
on this, it is possible to dene a service life requirement or set
areliability level that must be maintained for a given number of
years.Potential failure modes and important degradation processes
should be identied
after failures have been dened in terms of minimum performance
levels. In general,there exist many kinds of failure modes for a
particular component and even thedifferent parts of the component
and the different damage mechanisms, which maylead to the same kind
of failure, may sometimes, be quite numerous.The objective of
analysis is to identify potential failure/ damage modes and
mechanisms that may lead to material degradation and the
development of damage, andassociated critical factors of
environmental stress or degradation factors; see Table 1c.The risk
or risk number associated with each potential failure/damage
mode
identied can be estimated by use of the well-known methodology
of FMEA; see e.g.Ref. [10]. The risk number is estimated by use of
the following factors: Severity (S),Probability of occurrence (PO),
and Probability of escaping detection (PD). The risknumber RPN is
the product of all these factors, i.e. RPN=S PO PD. In the
riskassessment of potential failure modes, the relevance of
durability and life data foundin the literature for the component
and its materials must be taken into account. Therisk assessment
should most advantageously be performed by a group of experts.The
estimated risk number is taken as the point of departure to judge
whether a
particular failure mode needs to be further evaluated or not.
The estimated risknumber may also be used to determine what kind of
testing is needed for quali-cation of a particular component and
its materials; see the example in Table 1c.
4. Screening testing and analysis for service life
prediction
4.1. Screening testing by accelerated ageing
Screening testing is thereafter conducted with the purpose of
qualitativelyassessing the importance of the different degradation
mechanisms and degradationfactors identied in the initial risk
analysis of potential life-limiting processes.
ARTICLE IN PRESSB. Carlsson et al. / Solar Energy Materials
& Solar Cells 84 (2004) 255274 259
-
ARTICLE IN PRESS
Table 1
Example of result from an initial risk analysis of potential
failure modes based on information taken from
the IEA Task 10 case study on selective solar absorber
surfaces
Function and general
requirements
General requirements
for long-term
performance during
design service time
In-use conditions
and severity of
environmental stress
(a) Specification of end-user and product requirements on
component
Efciently convert
solar radiation
into thermal energy
Suppress heat losses in
the form of thermal
radiation
Loss in optical
performance
should not result in
reduction of the
solar system energy
performance (solar
fraction) with more
than 5%, in
relative sense,
during a design
service time of
25 years
Behind glazing in
contact with air
Casing of collector
exchange air with the
ambient, meaning that
airborne pollutants
will enter collector
If the collector is not
rain tight the humidity
level of air in
the collector may
become high
Maximum
temperature 200C
Critical functional
properties
Test method
for determining
functional property
Requirement for
functional capability and
long-term performance
b) Specification of functional properties and requirements on
component and its materials
Solar absorptance (a)Thermal emittance (e)Adhesion (ad)
ISO CD 12592.2
ISO CD 12592.2
ISO 4624
Functional capability
a>0.92eo0.15ad>0.5MPa
Long-term
performance
Da+0.25De r0.05
Failure/damage mode/
degradation process
Degradation
indicator
Critical factors of
environmental stress/
degradation factors
Estimated risk of
failure/damage mode
from FMEA
(c) Potential failure modes, critical factors of environmental
stress/degradation factors, and risk assessment
by failure modes effect analysis (FMEA)
Unacceptable loss in
optical performance
PCa S PO PD Risk
RPN
High temperature
oxidation of
metallic nickel
Reection spectrum
Vis-IR
High temperature 7 2 8 112
Electrochemical corrosion
of metallic nickel
Reection spectrum
Vis-IR
High humidity, sulphur
dioxide (atmospheric
corrosivity)
7 5 5b 175
B. Carlsson et al. / Solar Energy Materials & Solar Cells 84
(2004) 255274260
-
ARTICLE IN PRESS
Table 1 (Continued)
Failure/damage mode/
degradation process
Degradation
indicator
Critical factors of
environmental stress/
degradation factors
Estimated risk of
failure/damage mode
from FMEA
Hydratization of
aluminium oxide
Reection
spectrum IR
Condensed water,
temperature
7 7 4b 196
FMEA involves estimation of Severity (S), Probability of
occurrence (PO), and Probability of escaping
detection (PD) of a particular failure/damage mode. The risk
number RPN is the product of all these
factors, i.e. RPN=S PO PD.aPC = Performance criteria =
Da+0.25De.bPD value resulted from possible failure of the
glazing.
Fig. 2. Principal components of the selective solar absorber
system studied in IEA Task 10 for use in single
glazed at plate solar collectors to be installed in domestic hot
water (DHW) systems [8]. (a) Component/
material: Selective solar absorber coating of anodised aluminim
pigmented with small metallic nickle
particles; (b) Cross section of at-plate collector; (c)
Application: Use in single-glazed at-plate solar
collector for Domestic Hot Water system.
B. Carlsson et al. / Solar Energy Materials & Solar Cells 84
(2004) 255274 261
-
When selecting the most suitable test methods for screening
testing, it is importantto select those with test conditions
representing the most critical combination ofdegradation factors;
see example in Table 2.
4.2. Analysis of material change during ageing
Using articially aged samples from the screening testing,
changes in the keyfunctional properties or the selected degradation
indicators are analysed with respectto associated material changes.
This is made in order to identify the predominantdegradation
mechanisms of the materials in the component. When the
predominantdegradation mechanisms have been identied also the
predominant degradationfactors and the critical service conditions
determining the service life will be known.Screening testing and
analysis of material change associated with deterioration in
performance during ageing should therefore be performed in
parallel. Suitabletechniques for analysis of material changes due
to ageing may vary considerably; seee.g. Ref. [11].In Table 3 an
example from the IEA absorber surface case study is shown that
demonstrates how different techniques for analysing material
changes resulting fromageing can be used to get information on what
material degradation mechanisms arecontributing to deterioration in
performance.
ARTICLE IN PRESS
Table 2
Programme for screening testing in the IEA case study on solar
absorber surfaces
Possible degradation
mechanism
Critical periods of high
environmental stress
Suitable accelerated test
methods and range of
degradation factors
High temperature oxidation of
metallic Ni particles
Stagnation conditions of solar
collector at high levels of solar
irradiation (no withdrawal of
heat from the collector)
Constant load high
temperature exposure tests in
the range of 200500C
Electrochemical corrosion of
metallic Ni particles at high
humidity levels and in the
presence of sulphur dioxide
Under starting-up and under
non-operating conditions of
the solar collector when the
outdoor humidity level is high
Exposure tests at constant high
air humidity (7595% RH),
constant temperature
(2050C), and in the presence
of sulphur dioxide (01 ppm)
Hydratization of aluminium
oxide and electrochemical
corrosion of metallic Ni
particles by the action of
condensed water
Under humidity conditions
involving condensation of
water on the absorber surface
Exposure tests under constant
condensation (sample surface
cooled 5C below surrounding
air which is kept at 95% RH)
and temperature conditions
ranging from 1090C
B. Carlsson et al. / Solar Energy Materials & Solar Cells 84
(2004) 255274262
-
4.3. Microclimate characterisation for service life
prediction
In order to be able to predict expected service life of the
component and itsmaterials from the results of accelerated ageing
tests, the degradation factors underservice conditions need to be
assessed by measurements. In Table 4 measurementtechniques that
were used in the IEA Task 10 absorber case study previouslyreviewed
are given as an example of what factors were needed to take
intoconsideration in this study. It is of extreme importance to
characterize the serviceconditions in terms relevant for the most
important degradation mechanismsidentied for the materials of the
component but also in terms relevant for andconvertible into the
test conditions for the environmental resistance tests to be
usedfor accelerated life testing.
ARTICLE IN PRESS
Table 3
Results from the IEA Task 10 solar absorber case study in the
analysis of material change upon durability
testing
Degradation
mechanism
Techniques for analysis
of material changes
Results
High temperature
oxidation of metallic Ni
particles
UV-VIS-NIR reectance
spectroscopy
AES depth proling
SEM-EDX
XRD
Reduction of absorption in solar range
corresponding to reduction in metal
concentration
Formation of Ni oxides
Small changes in surface morphology
Formation of NiO
Electrochemical
corrosion of metallic Ni
particles at high
humidity levels and in
the presence of sulphur
dioxide
UV-VIS-NIR reectance
spectroscopy
FTIR reectance
spectroscopy
AES depth proling
SEM-EDX
Reduction of absorption in solar range
corresponding to reduction in metal
concentration
Formation of sulphate
Increase in surface concentration of Ni
accompanied with sulphur at the
surface
Surface morphology affected and
detection of sulphur
Hydratization of
aluminium oxides and
electrochemical
corrosion of the
metallic particles by the
action of condensed
water
UV-VIS-NIR
reectance spectroscopy
FTIR reectance
spectroscopy
AES depth proling
SEM-EDX
Some changes hard to explain
Formation of hydrated forms of
aluminium oxide leading to increased
thermal emittance
Change in surface structure
Considerable change in surface
morphology
The following analytical techniques were used
ultraviolet-visible-near infrared-reectance spectroscopy
(UV-VIS-NIR reectance spectroscopy), Auger electron spectroscopy
depth proling (AES depth
proling), Fourier transform infrared reectance spectroscopy
(FTIR reectance spectroscopy), X-ray
diffraction (XRD), scanning electron microscopy with energy
dispersive X-ray spectroscopy (SEM-EDX).
B. Carlsson et al. / Solar Energy Materials & Solar Cells 84
(2004) 255274 263
-
If only the dose of a particular environmental stress is
important then thedistribution or frequency function of a
degradation factor is of interest. In the IEAabsorber case study,
only the distribution in the absorber temperature was neededfor
predicting the service life limited by high temperature
degradation; see Fig. 3a. Incase of service life prediction
considering degradation caused by the action of highhumidity and
condensed water, both air humidityabove a threshold of
99%andsurface temperature were taken into account; see Fig. 3b. In
case of service life
ARTICLE IN PRESS
Table 4
Techniques that were used in the IEA Task 10 absorber case study
for measurment of degradation factors
in solar collectors operating under service conditions
Degradation mechanism Degradation factors/
measurement variables
Sensors
High temperature oxidation
of metallic Ni particles
Temperature: Surface temperature
of absorber plate
Pt sensors in holders
screwed directly on the
absorber plate. To
accomplish a good thermal
contact heat sink compound
was used
Electrochemical corrosion
of metallic Ni particles at
high humidity levels and in
the presence of sulphur
dioxide
Atmospheric corrosivity:
Measurement of corrosion mass
loss rate of standard metal
specimens
Air pollutants: Measurement of
sulphur dioxide concentration
inside and outside of the solar
collector
Metal coupons of carbon
steel, zinc and copper and
evaluation of corrosion
mass loss according to ISO
9226
Exposed metal coupons
analyzed in respect of the
sulphate content of the
corrosion products by EDX
UVuorescence
instrument for direct
measurement of sulphur
dioxide concentration in the
air outside and inside of the
solar collector
Hydratization of aluminium
oxide and electrochemical
corrosion of metallic Ni
particles by the action of
condensed water
Humidity: Measurement of air
humidity in the air gap between the
absorber plate and glazing cover of
the collector
Time of condensation:
Measurement of specular
reectance of absorber surface
Surface humidity: Measured
relative air humidities converted to
relative humidity on surface by use
of measured surface temperatures
Capacitance humidity
sensors carefully shielded
from solar radiation and
thermal radiation of the
ambient
Special designed reectance
mode condensation sensor
B. Carlsson et al. / Solar Energy Materials & Solar Cells 84
(2004) 255274264
-
prediction considering electrochemical corrosion of the metallic
nickel particles theyearly corrosion of zinc in the air gap between
the absorber plate and the transparentglazing was determined.
ARTICLE IN PRESS
Fig. 3. Results from measurement of microclimate for the
absorber in the IEA absorber surface case study
(a) Absorber temperature frequency function for one year. For 1
month of the year the collector was under
stagnation conditions (upper diagram). (b) Absorber temperature
frequency function when RHX99% ofthat year (lower diagram).
B. Carlsson et al. / Solar Energy Materials & Solar Cells 84
(2004) 255274 265
-
5. Service life prediction from results of accelerated
testing
5.1. Mathematical modelling
To perform an accelerated test, D; means that the level of at
least one stress factor,X ; causing degradation is kept at a higher
level relative to the situation in service.Consequently this means
the time to failure in the accelerated test, tf ;D; will beshorter
relative to service life, ts: The ratio between the latter and the
former iscommonly referred to as the acceleration factor, A: If the
applied stress is constant intime also for service conditions, the
acceleration factor a can be expressed as
A tS=tf ;D aX gXD=gXS; 1
where the expression gXD=gXs is called the time transformation
function or theacceleration factor function. Examples of time
transformation functions can befound in e.g. reports by Martin [12]
and by Carlsson et al. [1,13]. The timetransformation functions, aX
; used in the absorber case study are shown in Table 5.
5.2. Accelerated life testing and assessment of expected service
life
Accelerated life testing means to quantitatively assess the
sensitivity to the variousdegradation factors on the overall
deterioration of the performance of thecomponent and its materials
in terms of the mathematical models set up tocharacterize the
different degradation mechanisms identied. Life testing
thereforerequires conducting a series of tests.From the accelerated
life test results the parameters of the assumed model for
degradation are determined and the service life then estimated
by extrapolation toservice conditions. If the service conditions
vary, effective mean values of stress needto be assessed from
measured service stress data [1,8]; see also Fig. 3 and Table 6.As
a result, it may be possible to express the importance of different
degradation
mechanisms in terms of expected service life values, see example
from the absorbersurface case study in Table 5.
5.3. Reasonability assessment and validation
By use of accelerated life testing, potential degradation
mechanisms limiting theservice life of a component may be identied.
However, it is important to point outthat it is only the service
life determined by the material degradation mechanismsobserved in
the accelerated tests at relative high levels of stress that can be
assessed.Life-limiting degradation mechanisms may exist that cannot
be identied by way ofaccelerated life testing because the knowledge
and experience in what may causedegradation of a particular
material in a component may be too limited.The best approach in
validating an estimated service life from accelerated testing,
therefore, is to use the results from the accelerated life tests
to predict expectedchange in material properties or component
performance versus service time andthen by long-term service tests
check whether the predicted change in performance
ARTICLE IN PRESSB. Carlsson et al. / Solar Energy Materials
& Solar Cells 84 (2004) 255274266
-
with time is actually observed or not. In 6.1 an example of such
a check forpolymeric glazing materials are presented.The results of
validation tests therefore can be used to revise a predicted
service life
and form the starting point also for improving the component
tested with respect toenvironmental resistance, if so required. It
should be remembered that the mainobjective of accelerated life
testing is to try to identify those failures which may lead
ARTICLE IN PRESS
Table 5
Estimated service life of the nickel-pigmented anodised
aluminium absorber surface in the IEA Task 10
absorber case study
Degradation mechanism Time transformation
function
Estimated service life with
PC=Da+De o 0.05a(years)
High temperature oxidation of
metallic Ni particles
aT=exp[(Ea/R) (1/TD1/Ts)]Ea=activation energy
R=general gas law constant
TD=temperature of test
Ts=effective mean temperature
at service
>105
Electrochemical corrosion of
metallic Ni particles at high
humidity levels and in the
presence of sulphur dioxide
aCo=tM,s/tM,DtM,s=time to reach a certainextent of corrosion of
reference
metal in service
tM,D=time needed to reach thesame extent of corrosion of
reference metal in test D
(zinc used as reference metal)
12
(The coating is assumed to
be installed in a non-
airtight highly ventilated
collector)
34
(The coating is assumed to
be installed in an airtight
collector with controlled
ventilation)
Hydratization of aluminium
oxide and electrochemical
corrosion of metallic Ni
particles by the action of
condensed water
a1T ;H tH expEH;T=R
T1H;eff T1D
9
(The coating is assumed to
be installed in an non-
airtight highly ventilated
collector)
TH,eff= effective mean
temperature of the absorber
surface when the relative
humidity in the air gap is equal
to or higher than 99%
tH=the time fraction of the year,time-of-wetness, during
which
the relative humidity in the air
gap is equal to or higher
than 99%
EH,T=Arrhenius activation
energy
Values are given for the different degradation mechanisms and
assuming that the different degradation
mechanisms are acting alone. How the effective mean values of
stress were calculated are shown in Table 6.aPC=0.05 corresponds to
a decrease in the solar system performance of 5%.
B. Carlsson et al. / Solar Energy Materials & Solar Cells 84
(2004) 255274 267
-
to an unacceptable short service life of a component. In terms
of service life, the mainquestion is most often, whether it is
likely or not, that the service life is above acertain critical
value.In order to validate the predicted service life data from
accelerated life testing in
the IEA absorber case study the actual service degradation in
optical performance ofthe nickel pigmented anodized aluminium
absorber coating was investigated [14].Samples from the coating
taken from collectors used in solar DHW systems for timeperiods of
10 years or more were analysed for that purpose. It could be
concluded
ARTICLE IN PRESS
Table 6
Denitions of effective mean values of different stresses used in
the IEA Task 10 absorber case study
Degradation mechanism Effective mean value of stress
High temperature oxidation of
metallic Ni particles
Effective mean temperature of stress at service TEFF,s was
dened as
expEa=R 1=TEFF;s
R Tmax
Tminexp Ea=R 1=T f T dT
Ea=activation energy
R=general gas law constant
f (T)=frequency function for absorber temperature
distribution during one year; see Fig. 3a
Tmin=minimum temperature during the year (K)
Tmax=maximum temperature during the year (K)
Electrochemical corrosion of metallic
Ni particles at high humidity levels and
in the presence of sulphur dioxide
Effective atmospheric corrosivity in the air-gap between
absorber plate and transparent glazing was determined by
use of metallic zinc as reference material. Zinc coupons
were exposed in the collector for one year and the metallic
mass loss was determined to 0.3 g/m2/year
Hydratization of aluminium oxide and
electrochemical corrosion of metallic Ni
particles by the action of condensed
water
Effective mean temperature of the absorber surface when
the relative humidity in the air gap is equal to or higher
than 99%, i.e. TH,eff, was dened as
expEH;T =R T1H;eff R TH;max
TH;minexpEH;T =R T1H
fHTH dTH
EH,T=Arrhenius activation energy
fH(TH)=the yearly based frequency function for the service
temperature of the absorber surface in a solar collector
when the relative humidity level exceeds 99%, being the
time fraction of a year when the service temperature is in
the interval T to T+dT and the relative humidity level
exceeds 99%; see Fig. 3b
TH,max=the maximum service temperature of the absorber
surface in a collector, when the relative humidity level
exceeds 99% (K)
TH,min=equal to 273K, as below this temperature ice is
formed on the surface of absorber
B. Carlsson et al. / Solar Energy Materials & Solar Cells 84
(2004) 255274268
-
that the agreement between degradation data determined for the
absorber samplesfrom the DHW systems and that from accelerated life
testing from the Task 10 studywas astonishingly good both from a
quantitative and a qualitative point of view; seeTable 7 and Fig.
4. For the absorber coating in a properly designed solar
collector,the service life seems good enough. For the absorber
coating in a non air tight solarcollector, probably because of
glazing failures, the humidity level is raised to suchhigh levels
that the service life is reduced to an unacceptable level.
6. Adoption of the general methodology for durability and
service life assessment topolymeric glazing materials
Polymeric glazing materials offer signicant potential for cost
savings both asdirect substitutes for glass cover plates in
traditional collector systems and as anintegral part of
all-polymeric systems. However, glazing materials should have
hightransmittance over the entire solar spectrum and must be able
to resist long term (25years) exposure to service conditions under
solar ultraviolet (UV) light and elevatedtemperature loads
sometimes exceeding 80C under shorter periods of
stagnationconditions of a solar collector. They must retain
mechanical integrity (for example,impact resistance and exural
rigidity) under these harsh environmental stresses.In the IEA
Working Group Materials in Solar Thermal Collectors, Polyvinyl
cloride (PVC) and UV stabilized Polycarbonate (PC) glazing
materials were studied.For validating the applicability of the
general methodology on durability assessmentdeveloped by IEA Task
27, it was therefore decided to make use of the results fromthis
previous IEA polymeric glazing study [1]. This study is also
reviewed in theforthcoming book Performance and Durability
Assessment of Optical Materials forSolar Thermal Systems [15].
ARTICLE IN PRESS
Table 7
Some general characteristics of six solar DHW systems from which
nickel-pigmented anodised aluminium
absorber samples were analysed after long period of service.
More details about the systems can be found
in Ref. [14]
DHW-system and
location
Age (years) Notes
DK 1 (Denmark) 12
DK 2 (Denmark) 11 One n replaced due to frost burst
DK 3 (Denmark) 10
DK 4 (Denmark) 10 One n replaced due to frost burst
DK 1 (Switzerland) 15 One collector leak has been repaired
DK 2 (Switzerland) 15 Plastic cover has been replaced once and
a
collector leak repaired
B. Carlsson et al. / Solar Energy Materials & Solar Cells 84
(2004) 255274 269
-
6.1. Initial risk analysis of potential failure modes of
polymeric glazing materials
and screening testing
One of the most critical material properties related to the use
of polymericmaterials as solar collector covers is the maximum
recommended service temperaturefor the material or in this case the
closely related deection temperature of the
ARTICLE IN PRESS
Fig. 4. Comparison between the estimated service life of an
absorber coating from actual service exposure
and from the results of accelerated life testing [14].
B. Carlsson et al. / Solar Energy Materials & Solar Cells 84
(2004) 255274270
-
material under mechanical load. This requirement, set by the
maximum covertemperature observed when the collector is under
stagnation conditions, narrowsdown the number of possible polymeric
materials considerably. The PC glazingmaterial (APEC 9353
UV-stabilized polycarbonate) meets this requirement but thePVC
material (Duroglas) does not, because its glass transition takes
place alreadyaround 80C. The PVC material also does not meet the
requirement on minimumservice temperature.The mechanical properties
of a polymeric glazing material may also be critical
when used as a collector cover. Especially as an effect of
ageing the mechanicalproperties of the polymeric materials may
deteriorate to a level at which failure mayoccur induced by e.g.
wind or snow loads and because of impacts from ying objectslike
hail.To summarize the analysis on the critical functional
properties of the two
polymeric glazing materials, it is evident that the PVC glazing
material does notqualify for use as a solar collector material.
However, the PC material was qualiedat this stage of
evaluation.With respect to durability, two kinds of failures were
considered, one related to a
gradual decrease in the optical performance of the cover
material and one related toinsufcient mechanical strength of the
material leading to either breakage or plasticdeformation of the
cover.Reduction in the optical performance of a PC glazing may be
due to yellowing.
This kind of chemical degradation of a PC material, due to
photooxidation orthermal oxidation, may also result in
deterioration in the mechanical properties ofthe material, which
may result in a breakage of the cover when subjected to a
highmechanical load. It was believed that photooxidation and
thermal oxidation resultingin an unacceptable optical performance
were the most important and needed to befurther studied by ageing
testing. Failures resulting from an insufcient
mechanicalperformance caused by ageing are believed to be of less
probability of occurrence.The optical performance of a PVC glazing
material may be deteriorated by
dehydrochlorination causing yellowing. As the lowest as well as
the highest expectedcover temperature for the considered
application are outside of the recommendedservice temperature range
of PVC, mechanical failures due to that most likely willoccur. But,
it was also believed that dehydrochlorination and
photooxidationresulting in an unacceptable optical performance are
important and need to befurther studied by ageing testing. Failures
resulting from an insufcient mechanicalperformance caused by ageing
were believed to be of second order interest.From the screening
tests it could be concluded that the dominating mechanism of
degradation of the PC glazing material was photooxidation and in
the case of thePVC material photochemical dehydrochlorination. The
two most dominatingdegradation factors were thus temperature and
UV-radiation.
6.2. Service life prediction from accelerated ageing
Because only the degradation mechanism of photooxidation seemed
to contributesignicantly to the service life of the PC glazing, the
time-transformation function as
ARTICLE IN PRESSB. Carlsson et al. / Solar Energy Materials
& Solar Cells 84 (2004) 255274 271
-
shown in Eq. (2) was used for modeling of accelerated test data.
The same timetransformation was used to model the degradation of
the PVC glazing.
aPO IP expEa=RTEFF;D=IP expEa=RTEFF;s; 2
where aPO is the acceleration factor for photooxidation, I the
intensity ofphotoactive light, p the material dependent constant
which value has to bedetermined by accelerated testing, Ea the
activation energy, R the general gas lawconstant, T the temperature
[K], s the index for service; D the index for test, EFF
theeffective mean value.By conducting a series of accelerated tests
at elevated levels of surface temperature
and UV-light intensity, the parameters p and Ea in Eq. (2) could
be determined forthe two glazing materials by Jorgensen et.al.
[16]. By measuring glazing surfacetemperartures and UV-light
intensities during outdoor exposure, it was then possiblefor them
to predict expected degradation in optical performance by making
use ofthe results from accelerated ageing. As shown in Fig. 5 the
agreement betweenpredicted and actually measured degradation in
optical performance was quite good.
7. Conclusions
Durability assessment, accelerated life testing, and service
lifetime prediction areall very important elements in the
development of solar products for achieving
ARTICLE IN PRESS
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0 50 100 150 200 250 300 350 400Time of Exposure (days)
Chan
ge in
Tra
nsm
ittan
ceMeasured; PC; COPredicted; PC; COMeasured; PVC; COPredicted;
PVC; COMeasured; PC; AZPredicted; PC; AZMeasured; PVC; AZPredicted;
PVC; AZ
Predicted; PC; CO
Predicted; PC; AZ
Predicted; PVC; CO
Predicted; PVC; AZ
Fig. 5. Measured vs. predicted change in hemispherical
transmittance between 400500nm for the PC and
PVC glazing materials during outdoor exposure at test sites in
Colorado (CO) and Arizona (AZ) [16]. The
service life with PC=change in global transmittance o0.05 for
the PVC glazing material (Duroglas) wasestimated to be around 1
2of a year at the test site in Arizona and 3
4of a year for the test site in Colorado.
The service life with PC=change in global transmittance o0.05
for the PC glazing material (APEC 9353UV-stabilized polycarbonate)
was estimated to 5 years at the test site in Arizona and to 6 years
at the test
site in Colorado.
B. Carlsson et al. / Solar Energy Materials & Solar Cells 84
(2004) 255274272
-
successful and sustainable commercialization. It has been shown
that the systematicapproach to service lifetime prediction
described in this paper may be a valuable toolwhen applied to
specic products in their special applications and stress
conditions.The relation between materials performance and system
performance constitute
the starting point in dening performance requirements and
criteria and requireknowledge on measurement of material
performance properties and on systemperformance testing and
simulation. The environmental stress conditions for theproducts
need to be properly characterized and measured to serve as a base
forservice life prediction. To identify and characterize lifetime
limiting degradationprocesses with respect to changes in materials
performance and properties,accelerated aging tests are performed
for screening the relevant stress parameterlevels and for life
testing. Mathematical models are employed for analyzing
theaccelerated test data and the environmental stress conditions
for estimation ofservice life.Service life prediction at an early
stage of product development will enable a fast
improvement of the product quality. Materials or designs can be
chosen withsufciently durability or protection against
environmental stresses. This will reduceleading times for new
products.
References
[1] B. Carlsson, K. M .oller, M. K .ohl, M. Heck, S. Brunold,
J-M. Marechal, G. Jorgensen, General
methodology of accelerated testing for assessment of service
life of solar thermal components, Final
Report of IEA Solar Heating and Cooling Task 27: Solar Facade
Components - Performance,
durability and sustainability of advanced windows and solar
components for building envelopes,
Project B1: Durability assessment methodology development,
December 2004, Report to be ordered
from SP Swedish National Testing and Research Institute, P.O.Box
857, SE-50115 Bor(as, Sweden.
[2] B. Carlsson, General methodology, in: A. W. Czanderna (Ed.),
Performance and Durability
Assessment of Optical Materials for Solar Thermal Systems,
Elsevier Science, Amsterdam, 2004,
in press.
[3] G.B. Gaines, R.E. Thomas, G.C. Derringer, C.W. Kistler, D.M.
Brigg, D.C. Carmichael,
Methodology for designing accelerated aging tests for predicting
life of photovoltaic arrays, Batelle
Colombus Laboratories, Final Report ERDA/JPL-954328-77/1,
1977.
[4] C. Sj .ostr .om, Overview of methodologies for prediction of
service life, in: L. Masters (Ed.), Problems
in Service Life Prediction, NATO ASI Series E No. 95, Martin
Nijhoff Publishers, Dorderchk, 1985.
[5] CIB W 80/RILEM 71-PSL, Final Report Materials and
Structures, Vol. 19, No. 114, 1986.
[6] ISO 15686 Building and constructed assetsService life
planningPart 1: General principles, ISO
International Standardization Organization Geneva, Switzerland,
2000.
[7] ISO 15686 Building and constructed assetsService life
planningPart 2: Service life prediction
procedures, ISO International Standardization Organization
Geneva, Switzerland, 2001.
[8] B. Carlsson, U. Frei, M. K .ohl, K. M .oller, Accelerated
life testing of solar energy materialscase
study of some selective solar absorber coatings for DHW systems,
International Energy Agency,
Solar Heating and Cooling Programme Task X: Solar Materials
Research and Development,
Technical Report, SP- Report 1994:13, 1994.
[9] B. Carlsson, Methods for service life prediction, Handbook
in Life Time Technology, Swedish
Defence Material Administration , 1993-06-21 (in Swedish)
(Chapter 4,5).
[10] R.E. Mc Dermott, R.J. Mikulak, M.R. Beauregard, R. Mikylak,
The Basics of FMEA, Productivity
Inc., 1996, ISBN: 0527763209.
ARTICLE IN PRESSB. Carlsson et al. / Solar Energy Materials
& Solar Cells 84 (2004) 255274 273
-
[11] K. M .oller, Analytical techniques for studying solar
materials degradation processes, in: A.W.
Czanderna (Ed.), Performance and Durability Assessment of
Optical Materials for Solar Thermal
Systems, Elsevier Science, Amsterdam, 2004, in press.
[12] J.W. Martin, Time transformation functions commonly used in
life testing analysis, Durability of
Building Materials 1 (1982) 175.
[13] B. Carlsson, Mathematical models for service life
prediction, Survey of service life prediction methods
for materials in solar heating and cooling, International Energy
Agency, Solar Heating and Cooling
Programme Task X: Solar Energy Materials Research and
Development, Technical Report, Swedish
Council for Building Research Document D16, 1988.
[14] B. Carlsson, K. M .oller, U. Frei, S. Brunold, M. K .ohl,
Comparison between predicted and actually
observed in-service degradation of a nickel pigmented anodized
aluminium absorber coating for solar
DHW systems, Sol Energy Mater. Sol. Cells 6 (2000) 223.
[15] G. Jorgensen, S. Brunold, B. Carlsson, M. Heck, M. K .ohl,
K. M .oller, Case study on polymeric
glazing, in: A. W. Czanderna (Ed.), Performance and Durability
Assessment of Optical Materials for
Solar Thermal Systems, Elsevier Science, Amsterdam, 2004, in
press.
[16] G. Jorgensen, C. Bingham, J. Netter, R. Goggin, A.
Lewandowski, in: D. R. Bauer, J. W. Martin
(Eds.), Service Life Prediction of Organic Coatings, A Systems
Approach; ACS Symposium Series
722, American Chemical Society, Oxford University Press,
Washington DC, 1999, pp. 170185.
ARTICLE IN PRESSB. Carlsson et al. / Solar Energy Materials
& Solar Cells 84 (2004) 255274274
The applicability of accelerated life testing for assessment of
service life of solar thermal componentsIntroductionGeneral
methodologyInitial risk analysis of potential failure
modesScreening testing and analysis for service life
predictionScreening testing by accelerated ageingAnalysis of
material change during ageingMicroclimate characterisation for
service life prediction
Service life prediction from results of accelerated
testingMathematical modellingAccelerated life testing and
assessment of expected service lifeReasonability assessment and
validation
Adoption of the general methodology for durability and service
life assessment to polymeric glazing materialsInitial risk analysis
of potential failure modes of polymeric glazing materials and
screening testingService life prediction from accelerated
ageing
ConclusionsReferences