; FRP Rods for Brittle Fracture Resistant Composite Insulators
Learning Center
Plans & pricing Sign in
Sign Out
Your Federal Quarterly Tax Payments are due April 15th Get Help Now >>

FRP Rods for Brittle Fracture Resistant Composite Insulators


  • pg 1
									                              FRP Rods for Brittle Fracture Resistant
                                     Composite Insulators
                                                            M. Kuhl
                                                CeramTec AG, Wunsiedel, Germany


                      Brittle fracture of FRP rods can lead to mechanical failures of composite
                      insulators even at low mechanical loads at operational service. Although this
                      fact has been known for 20 years, it may still be a problem of some designs of
                      composite insulators at present time. In order to find out counter measures
                      against brittle fractures, a study was carried out in the early eighties. It turned
                      out that brittle fracture is a problem of FRP material and that material
                      compositions exist resistant to brittle fracture. A brittle fracture resistant FRP
                      rod introduced 1983 in one particular design of composite insulators resulted in
                      15 years of excellent service performance.
                      This study deals with details of brittle fracture of FRP-rods. Test setups were
                      established to induce brittle fracture artificially. It was realised that brittle
                      fracture is some kind of stress corrosion related to the composition of the FRP
                      materials. A broad variety of FRP materials was evaluated showing the
                      influence of the components of FRP material on the brittle fracture behaviour
                      of FRP rods as well as the effects of different manufacturing processes. The
                      compositions of brittle fracture resistant FRP rods are disclosed. The results
                      from artificial testing are compared with brittle fractures of FRP rods that
                      occurred in composite insulators in operational service. Although no
                      quantitative correlation could be established, the trend concerning the material
                      behaviour of FRP rods is similar.

                  1 INTRODUCTION                                       better interfaces between the different materials of the
                                                                       composite components. New developed test standards [7, 8]
Composite insulators consist of a glass fiber reinforced plastic       are the tools to check the integrity of the composite structure.
rod (FRP rod), a shedded housing made of polymeric material            No broadly accepted test exists so far to check the composite
covering the FRP rod and metal end fittings attached to the            structure regarding brittle fracture of the FRP rod. This may be
ends of the FRP rod. The housing protects the FRP rod from             based on the fact that brittle fracture occured on a small
weathering and supplies the necessary creepage distance. This          number of insulators under operational service conditions in
composite structure consists of several interfaces which have          comparison to the big number of installed composite
to be designed and manufactured properly in order to avoid             insulators. In spite of this fact, a mechanical failure of an
ingress of moisture and pollutants from the surrounding                overhead transmission insulator may cause line dropping
environment into the interior of the composite structure.              which results in an outage of the transmission line. Aging of
Laboratory testing carried out more than two decades ago               composite insulators, and, in particular, aging of the interfaces
revealed some typical electrical and mechanical failures of            of the composite structure may be one of the reasons for the
composite insulators of early designs [1]. However, in the late        non-frequent occurrence of brittle fracture. Manufacturing
seventies a new kind of mechanical failure occurred on FRP             defects on a small number of insulators may contribute to the
rods of composite insulators installed in HV lines [2, 3, 4, 5],       failures as well as specific environments. These conditions
never seen before during laboratory testing. This kind of              cannot be simulated in a design test performed on a small
mechanical failure is called „brittle fracture“ of a FRP rod due       number of test specimens. The only possibility to eliminate
to the unusual fracture pattern of the fracture area of the            brittle fracture of composite insulators at all was considered by
concerned FRP rods. The failures occurred at very low                  using a brittle fracture resistant FRP rod as part of the
nominal operational mechanical service loads [3, 4]. The outer         composite structure. The result of this study has led to such a
features of the fracture areas are characterized by a razor cut        rod. It was introduced in the manufacturing of composite
fracture surface running perpendicular to the axis of FRP rods         insulators consisting of a housing made of HTV silicone
[6]. In those days it was thought brittle fracture was initiated       rubber in 1983. Such insulators have been manufactured in
by ingress of chemicals such as dilute acids into the composite        quite a big quantity during the last 15 years. The number of
structure [2, 4, 5]. The improvements carried out on composite         brittle fractures from those insulators is zero.
insulators in the last decades led to elastomeric housing
materials such as silicone rubber and EPDM rubber and to

It is interesting to note that for several designs of other
composite insulator types brittle fracture is still a problem
until today [9, 10, 11, 12]. Most of the concerned insulators
failed in USA because there is the biggest market for
composite insulators [11].
Brittle fractures on FRP rods can be traced to stress corrosion
of     E (electrical grade)-glass   filaments.  Stress   corrosion
involves an ion exchange mechanism. Sodium ions with a big
ion radius are replaced by hydrogen ions with a small ion
radius, resulting in an increase of stress in the glass surface of
the filaments [13]. E-glass filaments build up spiral flaws on
their surfaces after immersion in diluted acids [6, 14, 15]. The
flaws in the filaments initiate the failure of the composite
material which can be described by the mechanism of fracture
mechanics [16]. In order to improve the brittleness of FRP
material, several measures were proposed. Gel coats can be
applied on the surface. The correct choice of the matrix resin
as well as the use of an acid resistant glass composition for the
glass filaments may be successful [17, 18].
                                                                         Figure 1. Device for generating nitric acid at power frequency
                  2. EXPERIMENTAL                                        voltage 25 kV, 50 Hz.

All HV insulators under operational service conditions may be            1       steel bar
stressed frequently by electrical surface discharges when the            2       supporter
surfaces of the insulators are polluted and humidity penetrates          3       insulated copper wire 1,5 mm dia.
the pollution. Already Cavendish found in 1784 the generation            4       filter paper wrapped around
of nitric oxides and nitric acid by using electrical discharges in       5       holder (acrylic glass)
nitrogen and oxygen, while Birkeland and Eyde found an                   6       PVC tube OD 63 mm, ID 51 mm
industrial process proposed in 1905 for generating nitric acid           7       silver paint (ground electrode)
                                                                         8       water supply (deionized water)
by using electrical discharges in humid air [19]. It was very            9       acid collector
likely to assume that the electrical discharges on HV devices
may also generate nitric oxides and derivates such as nitric
acid. In order to show that nitric acid can be generated from            Mechanical failures of test specimens from stress corrosion
power frequency voltage in presence of air and humidity, a               implies the presence of two components, mechanical stress
                                                                         and simultaneous application of an environmental medium
device according to Figure 1 was built. The electrical
                                                                         such as nitric acid. To check FRP rods concerning their
discharge was carried out at 25 kVr.m.s. (50 Hz). Distilled
                                                                         resistance to stress corrosion some have suggested the use of
water was fed through the filter paper wrapped around the
                                                                         bending stress. Experiments carried out by the author under
energized electrode at a flow rate of 10 cm³/h. After a time
                                                                         bending stress showed that after crack initiation delamination
span of 1h the content of the water collector had a pH value of
                                                                         occurred lengthwise in FRP rods to different extents. The acid
2,9 and 2,52 mg nitrogen was measured. More evidence for
                                                                         can run out and the stress cannot be held constant due to the
the presence of nitric acid and nitric oxides and their derivates
                                                                         delamination in the composite material. More reproducible
were found on the surfaces of polluted insulators from
                                                                         results were obtained from test specimens loaded in tension.
experiments and tests. 110 kV silicone composite insulators
showed 4-5 % nitrogen in the pollution after 5 years of                  Test arrangement for testing FRP rods under tensile loads are
service. Salt was taken from the surface of a silicone                   shown in Figure 2. These arrangements include the advantage
                                                                         that assembled FRP rods can be evaluated. FRP rods clamped
composite insulator after passing the 1000 h salt fog test
                                                                         into the interior of the end fittings may undergo higher
described in [7]. The nitrogen content in the salt amounted to           mechanical stresses in the end fittings as can be expected for
2,5 %. A huge pollution content of nitrogen was reported from                                                                     h
                                                                         the free length of the rods between the end fittings. In t is way
DC insulators on the pacific coast of California [20]. The
authors of [20] considered that the nitrogen in the pollution            the influence of the stress in the end fittings can be estimated.
could be traced back to agriculture products and fertilizers.            Diluted nitric acid of 1 n HNO3 were chosen for simultaneous
                                                                         application of the environmental medium (1 n means 63 g
                                                                         conc. HNO3 added to 937 g water).

         Procedure I        Procedure II        Procedure III            Table 1 lists the applied epoxy resins and hardeners. The
                                                                         resins F, 0164 and 331 are aromatic diglycidylether (bisphenol
                                                                         A base) with an epoxy equivalent of about 190. X18 is a
                                                                         distilled version of the epoxy resin mentioned before (high
                                                                         purity). X100 is a cycloaliphatic diglycidylester (HHPA base)
                                                                         with an epoxy equivalent of 185. The hardeners MTHPA, PSA
                                                                         mixt, 1102 and MNA are liquid at room temperature. The
                                                                         hardeners HHPA, 4,4 MDA and DDS are solid at room
                                                                         temperature. Hardener EH 640 is a 4,4 MDA diluted in 30 %
                                                                         Table 1. Epoxy resins and hardeners used to manufacture the test
                                                                         Type                                 Trade name
                                                                         F                                    Araldite F
                                                                         0164                                 Ruetapox 0164
                                                                         X18                                  Lekutherm X18
                                                                         331                                  D.E.R. 331
                                                                         X100                                 Lekutherm X100
                                                                         MTHPA                                Ruetapox HX
                                                                         HHPA                                 Araldite 907
Figure 2 Test arrangement for the evaluation of brittleness of FRP
            .                                                            PSA mixt.                            Araldite 905
rods (tensile load and 1n HNO3, simultaneously).                         1102                                 Vers.Prod.1102
                                                                         4,4 MDA                              Araldite 972
Procedure I       End fittings under acid                                DDS                                  Araldite 976
Procedure II      Infusion of HNO3                                       MNA                                  Araldite 906
Procedure III     Acid on the free rod length                            EH 640                               Versamid EH640
                                                                         The glass fibers used consisted of either assembled or direct
Experiments with the three test arrangements shown in                    rovings in 2400 tex            or    4800 tex   (1 tex=1 gram/km,
Figure 2 resulted in rejecting procedure I because the acid              DIN60905). They were supplied from the companies Silenka,
attacked the metal of the end fittings in such a way that the            PPG (Pittsburgh Plain Glass), OCF (Owens Corning), Bayer
acidity suffered and led to unreliable test results. The best            AG, Ahlström, Norsk (Norsk Fiberglass), Vitrofil S.p.A.,
reproduction of test results could be obtained with the test             Gevetex (Stratifil) and were used as delivered.
procedure III, because there is no metal involved in the
chemical stability of the diluted acid. To manufacture FRP               All test specimens were tested under constant static load as
rods for this study, two manufacturing methods were used. For            shown in Figure 3. After loading diluted nitric acid of a
variation of different glass fibers a discontinuous manu-                concentration of 1 n was applied immediately and the time to
facturing procedure has been established. The glass fibers               break was recorded by an electrical clock connected via
were wound up on a rotating wooden sheet forming loops of                movement of the lever arm of the test set-up indicating hours,
glass rovings of a predetermined number. The loops were                  minutes and seconds. The tests were carried out indoors at
impregnated with heated resin mixture in a tub and then pulled           room temperature.
in steel tubes for crosslinking at elevated temperatures. For the
variation of different resin mixtures a continuous pultrusion
process was used. In this process it was simple to replace a
resin mixture by another mixture. The glass content of the
FRP rods from both manufacturing procedures amounted to
63-6 % by weight and the mechanical properties of the rods
from both procedures were most equal when the same
components were used. All FRP rods were manufactured from
an epoxy matrix resin because epoxy resin is the best r sin for
FRP rods used in HV application due to their excellent
mechanical and electrical properties, although remarkable
differences exist within epoxies. The resin mixtures were
prepared in ratios given by the manufacturer of the resin
mixtures. The curing state of the test specimens concerned a
curing state at 130°C for 10 hours, when no other treatment is
mentioned.                                                               Figure 3. Test set-up to cause artificial brittle fractures on assembled
                                                                         FRP rods.

  3. FEATURES OF BRITTLE FRACTURES                                             fracture patterns. Hence, they were obtained by much higher
                                                                               loads and required longer times to failure than necessary for
The first brittle fracture on a 420 kV silicone rubber insulator               stress corrosion failure.
occurred in 1978 [5] on a 24 mm FRP rod assembled with end                     The experience with brittle fractures of silicone insulators
fittings with a wide cleavage (Figure 4). The sealing of the end               showed that in every case failure of the sealing between the
fitting was opened, so that chemicals could enter the interior                 housing and at least one of the end fittings was involved.
of the end fitting.                                                            Brittle fractures on the free length of the FRP rod had never
The broken insulator was part of a double suspension insulator                 been found for this particular insulator design. Tests carried
string. The parallel insulator held the line and was brought                   out on test specimen with and without silicone sheath acc. to
down for evaluation purposes. Mechanical tests carried out on                  Figure 2 test procedure III, found that artificial brittle fractures
this insulator resulted in no reduction of the ultimate tensile                of the FRP rods can be obtained only from naked rods.
load. Brittle fracture can be simulated by means of test
procedure II (Figure 2). Both fractures, the natural brittle                        4. ARTIFICIAL BRITTLE FRACTURES
fracture (Figure 4) as well as the artificial brittle fracture
(Figure 5) show fracture surfaces arranged perpendicular to                       4.1 INFLUENCE OF NOMINAL TENSILE
the axis of the FRP rods and characteristic patterns of stress                                 STRESS
corrosion fractures which cannot be simulated any other way.
                                                                               The results shown in this study refer to nominal tensile
                                                                               stresses. They are defined as those tensile stresses calculated
                                                                               from the applied static tensile load divided by the unloaded
                                                                               cross section of the FRP rod. This is also applicable for the
                                                                               end fittings, however, it is a stress indication only and does not
                                                                               specify any real stresses quantitatively.
                                                                               The general characteristics of brittle fracture of FRP rods
                                                                               obtained from nominal tensile stress and simultaneous
                                                                               application of 1n nitric acid is shown in Figure 6.

Figure 4 Brittle fracture of a 420 kV silicone rubber suspension
insulator after 3 years of service.

                                                                               Figure 6. Load time characteristics of brittle fractures of FRP rods
                                                                               24 mm diameter.
                                                                               Curve 1 and 2 - Test procedure III
                                                                               Curve 3 - Test procedure II, wide cleavage

Figure 5. Brittle fracture of a test specimen acc. to test procedure II.       These results were obtained from 24 mm rods made of E-glass
                                                                               in postcured condition (180°C for 16 hours). In Figure 6
Several tests performed with cyclic loads on test specimens                    curve 1 envelops the most resistant E-glass composition (glass
without simultaneous application of acid resulted in different                 type OCF 859/resin type BAY X100/HHPA) and curve 2

envelops one of the most susceptible E-glass composition                  content < 0,15%) result in brittle fracture resistant FRP rods.
(glass type Silenka/resin type BAK/MTHPA). Both curves                    Apart from the boron free glass type Norsk ECR mentioned in
represent artificial brittle fractures obtained from the free             Table 2 there exist some more boronfree glass fibers as
length of FRP rods. Curve 3 envelops failures of FRP rods                 Stratifil (Gevetex), NT 712 (PPG) and S2 (OCF). 24 mm FRP
within end fittings designed as wide cleavage of the cone                 rods made of these glass fibers show the same or even better
wedge type. It can be seen that the stress caused by end                  brittle fracture resistance as found from Norsk ECR. Although
fittings can lead to a drastic reduction of the time to failure           boron free glass fibers used in FRP rods result in brittle
(FRP composition from curve 2). The section between curve 1                                                                     n
                                                                          fracture resistant FRP rods, the role of boron oxide i the glass
and curve 2 represents also the load time characteristics of all          composition is still unclear.
evaluated FRP rods from 16 mm up to 37 mm made of E-
glass.                                                                    The figures of the B203 content in Table 2 were determined by
It can be seen that considerable differences exist regarding the          means of flame photometry. Several measurements showed
brittleness of FRP rods made of E-glass caused by stress                  that the boron content of a particular E-glass type can vary
                                                                          from batch to batch. That may be one reason that the
corrosion. For all FRP rods made of E-glass, a load level
                                                                          correlation between boron content and time to failure is not as
exists indicating that at loads below this level brittle fracture
                                                                          good as expected.
does not occur. For the free length of the rods the level may
exist at 60 MPa and for the rod ends at 15 MPa under the
above described test parameters.                                          Another interesting aspect of brittleness of FRP rods is based
                                                                          on the pretreatment to which the FRP rods are subjected. In all
For composite insulators in HV transmission lines much                    cases the brittleness of FRP r   ods made of E- glass increase by
higher stress levels can be assumed at operational service.               application of postcuring (for Table 2 at 180°C during
From this point of view it is most likely that brittle fracture           16 hours). Postcuring of FRP rods for composite insulators
may occur on composite insulators under operational service               may occur during manufacturing of the elastomeric housing.
conditions if the FRP rod is manufactured from E-glass.                   Different manufacturing procedures for housings may result in
                                                                          different brittleness of the FRP rods. As shown in Table 2,
                                                                          different epoxy resins can also lead to different brittleness of
    4.2. BRITTLE FRACTURE RESISTANT                                       FRP rods with E-glass. The resin system X100/HHPA is
                 FRP RODS                                                 superior to the resin system 0164/MTHPA, however, system
                                                                          X100/HHPA is not resistant to hydrolysis. FRP rods made of
The brittleness of FRP rods caused by stress corrosion can be             this system do not pass the water diffusion test specified in [7].
                                                                          However, resistance to brittle fracture and resistance to
influenced by the glass composition as well as by the epoxy
                                                                          hydrolysis are not controversed in any case. There were
resin matrix and the interface between glass fibers and matrix.           various FRP rods made of E-glass and resins such as
Table 2. Variation of glass fibers: Breaking time in minutes of 24
                                                                          unsaturated polyesters found which showed high brittleness
mm FRP rods at nominal stress of 77 MPA, test procedure II, end           and bad resistance to hydrolysis as well.
fittings with wide cleavage
                                         Resin Type                       Table 3 shows some results concerning brittleness of FRP rods
                      Hydrolysis resistant     Hydrolysis                 regarding various impregnation resin mixtures. Between the
                      Bisphenol            A- susceptible                 epoxy resins based on Bisphenol A and an epoxy equivalent
                      Diglycidyl               HHPA – Diglycidyl          about 190 (F, 0164 and X18) the difference concerning
                      Ether / MTHPA            Ester / HHPA               brittleness is negligible. However, there is a considerable
Glass Type  B 03
              2       Cured       Postcured Cured         Postcured       effect of the type of hardener. The hardener affects the density
          content                                                         of crosslinking of the epoxy matrix. Dense crosslinking means
Silenka      5,7           79        73                                   a high distortion temperature and low impact strength.
Silenka      5,1         121         61          617          508
PPG          4,9           17
Bayer        4,8           61        34          401          180         It was found that a high impact strength of the pure
Vitrofil     4,6           79        51                                   impregnation resin improves the resistance to brittle fracture
OCF 859      4,3         155        114         5821        2982
Ahlström     4,0          498
                                                                          of FRP rods made of E-glass. It can be assumed that a matrix
OCF 424      3,3          103                                             of higher toughness results in lower crack propagation velocity
Norsk        2,6         893        551                                   during the stress corrosion process. On the other hand, a low
Norsk ECR   <0,01     > 18 720 > 18 720                                   interlaminar shear strength may act as barrier to stop crack
                                                                          propagation. For the concerned system F/EH 640, an
Table 2 shows several parameters influencing the brittleness of           unacceptable wet out characteristic was also found during
FRP rods. As can be seen, the boron content of the glass fibers           manufacturing.
is of great importance. Boron free glasses (B203

Table 3. Variation of resin mixtures: Breaking time in minutes of 24              corrosion can be expected earlier than in case of non-
mm FRP rods at nominal stress of 77 MPa, test procedure II, end                   postcured systems.
fittings with wide cleavage, E-glass Silenka
1)     inacceptable interlaminar shear strength

Resin       Hardener        Time to break (minutes)    Resin characeristic
                          cured         postcured
F          HHPA           333           253           Bisphenol A mixture
0164       HHPA           368           118           Bisphenol A regular
X18        HHPA           416           106           Bisphenol A distilled
F          PSA mixture    661           220
0164       PSA mixture    551           231

Resin       Hardener        Time to break (minutes)     Impact strength of
                          cured         postcured         pure resin in
F          DDS            67            17                      6
F          1102           217           79                      8
F          PSA mixture    661           220                     9
F          HHPA           333           263                    10
F          THPA           374           309                    15
F          4,4 MDA                      459                    30
F          EH640                        >13001)                50

The results presented in Table 2 and Table 3 emphasize the
effect of postcuring on the brittleness of FRP rods made of E  -
glass. It can be assumed that postcuring of FRP rods does not
affect the properties of the glass fibers itself. For the curing
state of the matrix and the interface some effects can be
assumed. Thus, some evaluations on the postcuring state of
epoxy matrix systems were carried out. Test specimens were
molded of various epoxy matrix systems and treated with
various curing states. The test specimens were tested for                         Figure 7. Properties of an epoxy matrix depending on the softening
bending strength and deflection [21], tensile strength and                        temperature T M. System DOW 331/MNA.
elongation [22], Youngs modulus [22] and density [23]. The
curing state was determined by measuring the softening
temperature of the resin acc. to Martens [24], called TM.                         Figure 8 shows the scattering of time to failure of the most
                                                                                  brittle FRP rod found during this study. It was also found that
Figure 7 shows the results obtained from the System
                                                                                  the more brittle a FRP rod system is the more s the scattering
331/MNA because this system exhibits the post curing effect
most impressively. As known from cast resin the mechanical                        of the test results. Apart from that the results from the end
                                                                                  fittings (test procedure II) scatter more than from free length
properties of epoxy resin systems improve with increasing
                                                                                  of FRP rods (test procedure III), test procedure III may be a
curing state. This is also the case for the system 331/MNA.
                                                                                  tool to check the brittleness of FRP rods in general. The test is
One can assume that the increasing curing state means an
                                                                                  easy to perform in a test arrangement shown in Figure 3 with
increasing density of crosslinking what results in a higher
                                                                                  test specimens shown in Figure 2 (Procedure III). The acid
stiffness of the material and in a more dense material.
                                                                                  container made of polyethylene should have such a size that
Figure 7 shows that these assumptions are not correct. It was                     the FRP rod is surrounded by liquid thickness not less than
found that all evaluated epoxy systems showed a decreasing                        1 cm and a liquid level of not less than 4 cm. The lower end of
Youngs Modulus and a decreasing density with increasing                           the acid container has to be attached and sealed to the surface
curing state to different extents. Curing and postcuring of                       of the FRP rod in order to prevent the acid from coming in
epoxy matrizes mean that the matrix during crosslinking is                        contact with the lower end fitting of the FRP rod. The acid
subjected to a swelling effect [25, 26]. In case of reinforced                    container should be covered to prevent evaporation of the
epoxy it can be assumed that a compression force acts on the                      liquid more than 5 % of its volume during the test period.
interface between the glass fibers and the epoxy matrix.
                                                                                  The test specimen prepared this way can be loaded in the test
Postcuring means also that the finish layer on the surface of
                                                                                  set-up shown in Figure 3 with a load high enough to cause a
the glass fibers is hardening. Both effects result in that
                                                                                  tensile stress of 340 MPa within the cross section of the free
propagation of cracks due to stress corrosion that is not
                                                                                  length of the FRP rod and maintain this stress for a time span
stopped at the interface, and that time to failure due to stress
                                                                                  of 96 hours. Immediately after applying the load, a nitric acid

of a concentration of 1n HNO3 (1 n means 63 g conc. HNO 3                                    brittle fracture patterns. Brittle fracture of FRP rods occurred
added to 937 g water) can be poured into the acid container.                                 in service at load levels far lower than the ultimate failing load
The acid must not come into contact with the end fittings of                                 of composite insulators. This is consistent with load levels
the test specimen. Brittle fracture resistant FRP rods can be                                found for artificial brittle fractures simulated with tensile
realised, if no fracture of the FRP rod occurs during the test                               stress and simultaneous application of diluted acids. These
period of 96 hours [14], see also Figure 6.                                                  facts have led to the conclusion that brittle fractures of
                                                                                             composite insulators in service are the result of stress
                      60                                                                     corrosion, initiated by diluted acid at the surface of E-glass
                                                                                             fibers under tensile stress.
                                                                                             Brittle fracture of composite insulators can be avoided by
                                                                                             using FRP rods made of boron free glass fibers. This class of
                                                                                             glass fibers is resistant to acids and stress corrosion at load
   Failure Time [h]

                                                                                             levels known from service experience [14].
                                                                                             In general brittle fracture failure of FRP rods made of E-glass
                                                                                             can vary in a wide range of time. The fractures are subjected
                                                                                             to a load time characteristic depending on the applied tensile
                                                                                             stress and the pH value of the corrosive medium. The nominal
                                                                                             stress can be enhanced by the design of the end fittings at the
                           0   1   2   3   4   5    6         7   8   9   10   11   12       ends of the FRP rods. Other general factors influencing the
                                               FRP Rod. Nr.
                                                                                             brittle fracture failures are the boron content of the glass
Figure 8: Scattering of time to failure. FRP rod 37 mm System Dow                            fibers, the toughness of the resin matrix and the curing state of
331/MNA/PPG E-glass, test procedure II (wide cleavage) nominal                               the resin matrix.
stress 77 MPa
                                                                                             These results obtained from artificial testing of FRP rods made
Brittle fracture of FRP rods made of E-glass can be initiated                                of E-glass showed that the choice of the material components
by any other diluted acid with pH values smaller than 4. Tests                               (glass, resin, hardener) can also lead to a non predictable
performed with strong organic acids (checked were formic                                     brittle fracture behaviour. In the frame of this study 62
acids, chlorinated acetic acid and oxalic acid) showed that                                  variations of FRP rods have been evaluated. Although some
these acids can also attack FRP rods. On the other hand, FRP                                 general trends resulting from the FRP material components
rods made of boron free glass fibers withstood all diluted                                   could be found, there were some results which could not be
acids. Water did not affect any of the evaluated FRP rod                                     explained, such as shown in Table 2 for the glass OCF 859 in
systems.                                                                                     combination with two different resins.
Glass monofilaments may contain tiny capillaries related to
gaseous bubbles in the molten glass during manufacturing of
the monofilaments. This was discovered in the 70ies for E-                                        5. NATURAL BRITTLE FRACTURES
glass. Some interesting effects concerning the electrical
performance of FRP rods containing capillaries were found by                                    5.1 FRP RODS EXPOSED OUTDOORS
the author. Test standards were developed in order to quantify
capillarity for the selection of FRP rods suitable for HV                                    It is common practise for manufacturers of composite
applications. A link between capillarity and brittle fracture of                             insulators to evaluate the load time characteristic of FRP rods
FRP rods could not be established. Today capillarity of FRP                                  under static loads as described in [7]. In most cases it takes
rods is not a problem anymore. Both E-glass as well as boron                                 several years to obtain a load time curve for a particular
free glass nearly free from capillaries is existing.                                         assembled FRP rod. Up to now nothing is specified about the
                                                                                             condition of test specimens for the realization of the load time
                                                                                             curves. Knowing about the impact of postcuring on brittle
           4.3. CONCLUSIONS FROM A RTIFICIAL                                                 fractures 24 mm FRP rods were exposed outdoors under static
                   BRITTLE FRACTURES                                                         loads in a test device similar to Figure 3. Rods with and
                                                                                             without a sheath of silicone elastomer in postcured condition
                                                                                             (180°C for 16 h) were tested.
Brittle fracture of FRP rods made of E-glass can be simulated
by applying tensile stress and simultaneous application of
                                                                                             During the first three years of outdoor testing no difference
diluted acids. The fracture patterns of artificial brittle fractures
are very close to the fracture patterns of brittle fractures of                              could be observed between test specimens with and without
broken composite insulators out of service. Tests conducted                                  silicone sheath. The obtained load time curves were similar.
with test specimens under static and dynamic loads without                                   However, after three years of loading visual inspection of the
                                                                                             naked FRP rods revealed some differences between FRP rods
application of acid led to fracture patterns not comparable to
                                                                                             of different compositions as well as differences between naked

and sheathed rods. While the sheathed rods showed minor                        by far the erosion seen on the 0164 system. It can be assumed
influence of weathering, all naked rods made of aromatic                       that crack propagation due to brittle fracture caused by stress
epoxy systems showed strong discolouration and strong                          corrosion of the glass fibers and the speed of matrix erosion
erosion on their surfaces. A layer of loose glass fibers covered               caused by weathering are in competition with each other. The
the surfaces of the rods due to erosion of the aromatic epoxy                  high speed of the matrix erosion of the F system inhibits crack
matrix the more the longer the time of exposure. The rods                      propagation due to stress corrosion.
composed from cycloaliphatic epoxy matrix showed by far
                                                                               The reason for brittle fractures observed on naked FRP rods
less surface erosion for the same exposure time.
                                                                               exposed outdoors (without voltage, without acid) are still
Table 4 Observations made on naked 24 mm FRP rods exposed
                                                                               unknown. It can be argued that diluted acid can exist in every
outdoors under static load.                                                    moderate climatic atmosphere as can be found in Germany.
                                                                               On the other hand, the evaluated epoxy systems are mostly
                        Artificial     Long-term Outdoor Exposure              crosslinked with acidic hardeners. Hydrolytic effects on the
                         Testing               without Acid                    matrix system may play a role. It can also be assumed that the
  Matrix      Glass      Time to     Nominal      Expo-     Obser-             UV spectrum of the solar radiation may cause nitric acid for
  Type        Type        break      Tensile       sure     vations            crack initation on the surface of FRP rods. The FRP rods
                       test proce-    Stress       time
                         dure II                                               covered with silicone elastomer did not show any signs of
                          (min.)     (MPa)       (years)                       brittle fracture. For the naked rods it can be concluded from
0164         Silenka        61        261          3,5     cracks and          Table 4 that the observations made on them follow the same
MTHPA                                                      delamina-           trend as found for the results from artificial testing.
                                                           tion on
                                                           FRP rod
                                                           surfaces,           5.2 FRACTURES ON SILICONE COMPOSITE
                                                           less surface
                                                           erosion                         INSULATORS
                                       102         6       brittle frac-
                                                           ture on free        Table 5 lists the brittle fractures of FRP rods made of E-glass
                                                           FRP rod             used in silicone composite insulators from operational service.
                                                           surface             As shown the majority of the broken insulators were installed
                                                           erosion             at harsh environmental conditions at high service voltages.
    F       OCF 859       220          252         9       FRP rod             In all cases the sealing between housing and the end fitting
 PSA mix-                                                  surface             opened and except in one case the live end fitting sides of the
   ture                                                    strongly
                                                                               insulators were involved. Brittle fracture on the free length of
X100        OCF 859      2982          217         14      One crack           the FRP rods did not occur in contrast to what is reported in
HHPA                                                       near end            [2] and [11]. A correlation between the expected service
                                                           fitting,            tensile stresses and the time to failure could not be found.
                                                           minor               However, 13 brittle fractures out of 19 occurred on tension
                                                                               insulators. The number of tension insulators in a HV
Table 4 shows the most interesting results from 24 mm rods                     transmission line is small compared to the number of
                                                                               suspension insulators. In spite of this fact brittle fracture
exposed outdoors under static loads in the plant location of
                                                                               occurred preferably on tension insulators which are usually
CeramTec AG. The environment of this plant can be
                                                                               more mechanically loaded than suspension insulators. This
characterized by a relatively clean atmosphere approx. 500 m                   may be an indication that tensile stress from service loading
above sea level. A lot of forests and agriculture dominate this                play a role concerning the statistical occurrence of brittle
area of moderate climate. The most interesting observation
made was the fact that one brittle fracture occurred on one
aromatic epoxy system and rods from other systems showed                       Table 5 presents the results of artificial testing of FRP rods
some cracks running perpendicular to the rod axis on its free                  from test procedure II. The comparison between the time to
length like a start of brittle fracture. In Table 4 are also listed            failure from artificial testing and from operational service
three epoxy systems and their time to break from artificial                    leads to a similar trend concerning the evaluated material
brittle fracture testing with nitric acids. There might be a                   composition. The resin matrix 331/MNA combined with the
correlation between artificial brittle fracture testing and those              E-glass type PPG 712 was found to be the most susceptible
observations made in outdoor exposure of FRP rods.                             system concerning brittle fracture during artificial testing, see
Questions arose why the aromatic F system from Table 4 did                     also Figure 8.
not break. Microscopic evaluation of the surface of those rods
revealed some differences to the 0164 system. Under the                        This system was also most susceptible under operational
microscope the same cracks could be seen on the F system as                    service conditions. The systems 0164/MTHPA/ Silenka and
seen macroscopically on the 0164 system; however, the degree                   F/PSA mixt/OCF 859 resulted in small differences from
                                                                               artificial testing.
of matrix erosion due to weathering on the F system exceeded

 Table 5. Brittle fractures experienced on silicone composite                                             the unusual fracture pattern of the FRP rods lead to the
 insulators.                                                                                              assumption that brittle fracture of FRP rods is initiated by
                                                                                                          stress corrosion. Nitric acid can be formed by electrical
                       Time                                             Brittle   Posi -
                FRP    to                                               Frac-     tion      Rea-          discharge in humid air and may have access to the surface of
                                                                                                          FRP rods.
tem     Type           II       Mfg.   (kV)                             Moun      Frac-     Brittle
                mm     Post -                                           ting      ture      Frac-         FRP rods have been tested artificially under tensile stress and
                       Cured                                                                ture
                       min.                                             years                             simultaneous application of 1n nitric acid.
                37     21       1979   420     triple
                                                           very se-
                                                                        0,08      life
                                                                                                          The results from this testing lead to the conclusions that brittle
                37              1979   420     dto.        dto.         0,08      ground    dto.          fracture of FRP rods
                37              1979   420     inv.V       dto.         0,4       life      dto.
                                                                                                          -    is a matter of stress corrosion
                                                                                  end                     -    follows the rules of fracture mechanics
                37              1979   420     suspens.    dto.         1,3       dto.      dto.
0164    Si-     24     121      1981   420     double      coastal      10        dto.      dto.          -    can be initiated by diluted acids and simultanous
                24              1981   420
                                                           dto.         10        dto.      dto
                                                                                                               application of tensile stress
                                                                                                          -    is a matter of stress corrosion of glass fibers containing
                24              1979   123     dto.        medium       6         dto.      dto.
                                                                                                               boron oxide
                24              1979   123     double      medium       15        dto.      dto.
                                                                                                          -    can be prevented by using boron free glass fibers.
                24              1979   420     triple      medium       12        dto.      dto.
                                               tension                                                    A broad variety of FRP rods made of different material
                24              1981   230     double                   2         dto.      dto.
                                               tension                                                    components were evaluated. It was shown that all E-glass
                24     220      1975   420     double
                                                           rural        3         life
                                                                                                          fibers used in FRP rods lead to rods susceptible to stress
mixt.   859                                                                                 cleav-        corrosion to different extents. Considering the parameters of
                24              1975   420     dto.        dto.         4         dto.      dto.          influence found, the occurrence of catastrophic mechanical
                                                                                                          failures of composite insulators installed outdoors on HV lines
                37              1975   420     invert. V   coastal      6         dto.      bad
                                                           severe                           sealing       can be related to the following reasons:
                37              1975   420     double      severe       7         dto.      dto.

                24              1977   123
                                               spacer      Alps         12        dto.      bad
                                                                                                          ?? Defect of the sealing between end fitting and FRP rod:
                                                                                            sealing          Moisture penetration into the inside of the end fitting with
                24              1975   420     suspens.    very         7         dto.      dto.
                                                           severe                                            generation of an acidic solution due to electrical activity
                24              1975   420     spacer      severe       20        dto.      bad
                                                                                                             (corona) in the area near the end fitting. Acid attack upon
                37              1976   420     double
                                                                        19        dto.      dto.             non-brittle fracture resistant FRP rods which are
X100    OCF     24     2982     1968   15      tension     very         29        dto.      dto.             permanently under mechanical load.
HHP     859                                    railway     severe                           wide
A                                                                                           cleav-
                                                                                                          ?? Glass fibres used for FRP rods contain B2O3.
                                                                                                          ?? Epoxy matrix is not suitable: The main parameters that
 Under operational service the performance of these two                                                      influence the brittle fracture resistance are the curing
 systems can be considered as equal. The average time to                                                     state, the toughness and the swelling characeristic of the
 failure of both systems under operational service conditions                                                matrix.
 results in 9.1 years (one can also assume that the time until
 opening of the sealing between housing and the end fitting is                                            ?? Interface between matrix and fibres is weak: Moisture in
 equal statistically). The system X100/HHPA/OCF859                                                           the interface, missing coupling agents and bad sizing can
 exhibited a remarkable resistance to the occurrence of brittle                                              lead to a weak fibre-matrix bonding. FRP rods with weak
 fracture in artificial testing as well as under operational                                                 fibre-matrix bondings are expected to be more sensitive to
 service. Although some thousand silicone composite insulators                                               brittle fracture.
 still exist utilizing this type of FRP rod in lines up to 245 kV
 with end fittings of wide cleavage, only one brittle fracture                                            It is most likely that more than one of these parameters
 under operational service conditions occurred within a service                                           determine the time to failure of a particular composite
 time of more than 25 years. Finally, a vast number of silicone                                           insulator installed on a HV line, as long as FRP rods are used
 rubber insulators have been installed since 1983 utilizing FRP                                           made of E-glass.
 rods containing boron free acid resistant glass fibers from                                              The brittleness of FRP rods can be checked simply by means
 which is known the rate of brittle fractures is zero. This fact is                                       of tensile stress in simultanous presence of diluted acids (test
 also consistent with the results of artificial testing.                                                  procedure III). The test results obtained from this testing are
                                                                                                          consistent qualitatively with the experience obtained from
                                6. CONCLUSIONS                                                            operational service of composite insulators installed in outdoor
 Composite insulators installed outdoors in HV lines can suffer                                           HV lines. The use of boron free glass fibers in FRP rods have
 from catastrophic mechanical failures at tensile loads far                                               led to a new generation of silicone composite insulators free of
 below their ultimate tensile strength. This fact together with                                           brittle fracture in 1983. This is proven by the vast number of

such insulators manufactured in the last 15 years. The number                [15] G.W. Ehrenstein and H. Spaude, “Rißbildung in
of brittle fractures from those insulators is zero.                               Elementarfasern durch Einwirkung korrosiver Medien und
                                                                                  erhöhter Temperatur“, Z. Werkstofftechn. 14 (1983) pp. 73-81
                                                                             [16] J. Aveston and J.M. Sillwood, “Long Term Strength of Glass
                      REFERENCES                                                  Reinforced Plastics in Dilute Sulphuric Acid“, Journal Mat.
                                                                                  Sci., 17 (1982) pp. 3491-3498
[1]    G. Karady, R. Vinet and N. Souchereau, “New Test Methods
      for Synthetic insulators“, CIGRE 1976 Session, paper 22-15             [17] P.J. Hogg, “Factors affecting the Stress Corrosion of GRP in
                                                                                  Acid Environments“, Composites, Vol. 14, No. 3, July 1983,
[2]   H. Weihe, R.E. Macey and J.P. Rynders: “Field Experience                    pp. 254-261
      and Testing of New Insulator Types in South Africa“, CIGRE
      1980 session, paper 22-03                                              [18] S. Torp, O. Syromsod and M. Onarheim, “Influence of Glass
                                                                                  Fiber Type on Laminate Corrosion Resistance“, 37th Annual
[3]   M. Cojan, J. Perret, C. Malaguti, P. Nicolini, J.S.T. Looms and             Reinforced Plastics / Composite Institute, The Society of
      A.W. Stannet, “Polymeric Transmission Insulators: Their                     Plastics Industry, INC. January 11-15, 1982, Session 9-E, pp.
      Application in France, Italy and the UK“, C IGRE 1980 session,              1-5
      paper 22-10
                                                                             [19] H. Römpp, “Chemie-Lexikon“, 6.               Aufl.,   Frankh’sche
[4]   A. Bradwell and J.C.G. Wheeler, “Evaluation of Plastics                     Verlagshandlung Stuttgart, 1966, pp. 6155
      Insulators for Use on British Railways 25 kV Overhead Line
      Electrification“, IEE Proc. Vol. 129, Pt. B. No. 3, May 1982,          [20] T.C. Cheng, C.T. Wu, “On the deletrious Effects of Nitrates on
      pp. 101-110                                                                 Insulator Surfaces under HVDC Conditions“, IEEE Trans.
                                                                                  Electr. Ins., Vol. E1 -14, No. 3, June 1979, pp. 171-174
[5]   E. Bauer, H. Kärner, K.H. Müller and P. Verma, “Service
      Experience with the German Composite Long Rod Insulator                [21] DIN 53 452, “Testing of Plastics; Bending test“, April 1977
      with Silicone-Rubber Sheds since 1967“; CIGRE 1980 session,            [22] DIN 53 445, “Testing of Plastics; Tensile Test“, August 1981
      Paper 22-11
                                                                             [23] DIN 53 479, “Testing of Plastics                and   Elastomers;
[6]   CIGRE WG 22-03, “Guide for the Identification of Brittle                    Determination of the Density, July 1976
      Fracture of Composite Insulators FRP rod“, ELECTRA No.
      143, 1992, pp. 61-69                                                   [24] DIN 53 458, “Testing of Plastics; Determination                of
                                                                                  Temperature of Deflection under Load (Martens Method)
[7]   IEC 1109 (1995), “Composite Insulators for a.c. Overhead
      Lines with a Nominal Voltage greater than 1000 V –                     [25] W. Fisch, W. Hofmann and R. Schmid, “Influence of Structure
      Definitions, Test Methods and Acceptance Criteria“.                         and Curing Conditions on the Density, Degree of Cure and
                                                                                  Glass Transition Temperature During the Curing of Epoxide
[8]   ANSI C29.11 (1989), “American National Standard for                         Resins“, Journal of appl. Polymer Science Vol. 13, pp. 295-308
      Composite Suspension Insulators for Overhead Transmission                   (1969)
      Lines – Tests“
                                                                             [26] H.-J. Booss and K.R. Hauschild, “Volumeneffekt bei der
[9]   “Interview with J.S.T. Looms“, Insulator News and Market                    Vernetzung von Epoxidharzen“, Kunststoffe 70 (1980) pp. 48-
      Report Vol. 2, No. 1, Jan./Feb. 1994, pp. 19-23                             50
[10] C.N. Ravera, P.J. Olivier, A.C. Britten and D.J. Swift,
     “Silicone Rubber Insulators on ESKOM’S AC Transmission
     Lines“, AC and DC Power Transmission, 29 April – 3 May
     1996, Conf. Publ. No. 423 IEE, 1996, pp. 70-75
[11] “Research of Brittle Fractures in Composite Insulators,
     Interview with Maciej Kumosa of the University of Denver“,
     Insulator News and Market Report, July/August 1997, pp. 46-
[12] M. Portillo Belinchon, “Technical Workshop on Composite
     Insulation for Power Lines, Final Report“; CIGRE SC 22 –
     Espana, Madrid, May 29, 1997
[13] A.G. Metcalf and G.K. Schmitz, “M echanism of Stress
     Corrosion in E-Glass Filaments“, Glass Technology, Vol. 13,
     No. 1, February 1972, pp. 5-16
[14] H. Dietz, H. Kärner, K.H. Müller, H. Patrunky, G. Schenk, P.
     Verma and H.J. Voss, “Latest Developments and Experience
     with Composite Longrod Insulators“, CIGRE 1986 session,
     paper 15-09


To top