Failure modes of outdoor polymeric insulators. Dave Thodey. .

Raychem L t d Outdoor insulators with a plastic external &as distinct from glazed porcelain or glass have now been in service for many years. They are generally constructed of three elements an lIlsulating mechanically strong structural core; metal end fi-s bonded in some way to the core which transfer themechanical stress betweenthe HV conductor and adjacent earthed structural metal work;and finally an exterior weather resistant plastic covering which provides sufficient creepage distance between HV and earth to prevent pollution flashover. All these elements either s a y or together can and do fail.

Pultmkd resin bonded glass fibre rod is the most popular strength bearing member. There is a wide range of and electrical properties and prices available in the market.The materials used are generally epoxy or vinyl ester resins in conjunction with an E or ECR glass fibre which itself is coated with a coupling agent to enhance the resin/glass bond. These rods may mechanically fhil by acid notchmg or electrically fail due to moisture ingress and interfacialt " g .

- E glass is particularly susceptible to acid natchtng but with a good design which avoids acid generadischarging voids adjacent to the glass, and a housing which is an effective acid acceptor, no problems should arise. In practice, I believe the problems have been restrictedto epoxy clad and PTFE clad insulators,neither of which had acid acceptor fillers in the housing. This type of design is now out of favour.
pultruded rod has relatively poor torsional stteogth and it is important to limit any torsional loads otherwise longmdinal craclaag occurs leadq eventuallyto internal electricaltraclang Mure. I n t e r n a l t " gfhilure is more commonly caused by poor bonding between resin and glass or s o " poorly cured resm. It takes several years ofmoisture ingress and slow degradation before the internaltracking failure t a k e splace. The wicking ,boiling and electric strength tests in IEC 1109 address this potensial fhilure mode.

M e t a l end fithngs are traditionally used to transfer the m e c h a n i c a l load form the RBGF rod to ad~acent earthed structures or HV cxmduars. They may be attached by adhesives; using a wedge shaped
mechanical lock; or crimping onto the rod using fiiction to transfer the mechanical force fiom core to end faemg. Crimping is t h e most popular choice as in most cases it gives the lowest cost system for a particular performance. The crimping process is n o t so simple under crimping allows the rod to shp out over crimping may break many strands. Then there is the question of the creep behaviour of the crimp and the resin glass matrix slowly relaxing the grip between them. IEC 1109 addresses the short time strength of the assembly (up to 4 days) but does n o t address the possibihty of t h e crimping pressures 'creeping' loose. End fittings 'letting go' is a reported Mure mode.

Epoxy adhesive systems have been widely used for applications at for instance British Rad with good success. They may however be susceptible to impact damage I have seen an insulator proof tested delivered, dropped and then the end Mbng fell o f f .

Wedge systems give excellent results and allow a very compact fittmg however their expense precludes their general adaptation. Corrosion of the end fithngs can cause severe problems they are dortunately right where all the aggressive ionic species present on the insulators migrate to. T h i s is a particular problem under the DC conditions which are frequently used on metro systems and in some applications, fittings made of gun metal have been specified to control the corrosion rate. Under power arc flashover conditions, aluminium end fittings are much more susceptible to mewablation damage than ferrous systems. It may be necessary to use arcing horns to obviate this possibility. At present there is no 'power arc' test included in IEC 1109. Finally, the polymeric weathersheds used to protect the glass fibre core and provide adequate external creepage have plenty of problems for both manufactures, users, and academics to Write about. These materials have to provide a good electrical interfacial seal to the glass fibre resin on one side and a

3/1

good, non traclang external surface on the other. This external surface: will be degraded by the weather,

SUrEace arcs and pollution- and the degraded surface must maintain good electrical properties.

PTFE is the only base polymer available which will work in the outdoor environment d o u t the extensive use of additives to try to address each polymers several weaknesses. The recipes whch result from these attempts are always a compromise - for example - increasing the 1 0 of an antitrackmg, anti erosion filler like alumina trihydrate d linevitably w o r s e n electrid puncture strength, mechanical properties, W resistance, and anfke electrid resistance. W screens like zinc oxide are electrical conductors - so addmg large quantrties may well give excellent W resistance but will do nothug for all your electrical properties. The requirements in IEC 1109 for the outer surface of the housing are limited to a 1000 hour salt fog test at a high salinity level and a recently introduced flamnabihty test of no relevance to outdoor insulators (it is a Bunsen burner test on a 3mm sheet). A 5000 hour test includmg W ageing equivalent to approximately 2 years in service is included as an appendix. The E E C inner interface spec test is a high rate of rise impulse test..Thisis not particularly easy to do and in my qerience does not give such useful guidance as to the performance and merit order of pamGular intafacial systems as the British Rail interfacial seal test devised by Bradwell and Wheeler which gives the user an indication of how fast an interface is deteriorating and how condllctive it has become.
The inner interface between rod and polymer may be electrically sealed with grease, adhesive, mastic, or bonded with coupling agents when the outer covering is overmoulded. Grease will performance on the IEC interfacetest but over a period of time the 03s in most greases w i l l the surroundmg polymers leaving a dry interface..or they may be pumped out systems can work well but are only practicable when using heat shrink housing system is the most popular but to do this very carefitl selection of the couphg agents used must be made...not maktng the rmstake of selecting for best mechanical bond...meChamcal and electrical performance do not have a lot to do with each other. There is no doubt that interfacial sealmg failure has caused a lot of problems in service .. nowadays the use of t e s t IElethods like the British Rail one has made it much easier to prove a new design.

If we look a t service experience of the outer surface of the housings, then it broad brush summary would be that with the exception of PTFE, a l l materials are deterior;i.ting as a result of W plus pollution damage.The attached table s u m the various candidate base polymers which have seen most service so far.
Silicone elastomers are a popular choice as a housing m a t e d and enjoy several misconceptions on why they have given generally good field performance. For instance .. 1) They cannot track because they contain no carbon.....nottrve ,they typically contain two carbon atoms for each silicon and will track on the IEC 587 test. 2) They cannot W degrade because the silicon -oxygen bond has a higher energy t h a n that available from a photon in the U.V spectrum. ... not relevant ..the other parts of the elastomer structure- the pendant methyl groups, any oxi&sed areas ,cham defects, the cross links necessary to hold the m a t e d together ..all these are weak links which may break under W light. 3) They bleed low molecular weight silicone oils to the surface and this hydrophobic layer suppresses leakage current..and this must be good.... well it probably is good for the surfice, but if oil is bleedmg out of the matrix ..water can go L a n d that is a bad thug. 4) The polymer thennally degrades to an i n s u l w ash and this must be.better than a cabon char. ....unfortumtely the insulahng ash strongly adheres to the polymer surface ,traps water and encourages canyon erosion. Other womes about the performance of silicones include disastrous degradation in the presence of NOx fiunes from arcing, h g h moisture uptake and h g h MVT, and susceptibility to bird damage. Despite this long list there is no doubt that the best silicone formulations have worked very w long periods of time - they have two major advantages from the point of view of the small insulator manufacturer.. 1)fdly formulated and proven materials are available from the polymer manufacturers ... 2) They are easy to process using widely available equipment.

Epoxy resins have not enjoyed a good reputation in the UK they have usually weathered badly in service and on test...resulting i n very high surface leakage currents and flashover problems. Elsewhere, in areas of low pollution..eg Finland they have worked adequately. Crosslinked polyethyl&e and its copolymers with eg vinyl acetate, ethyl acrylate, etc are very extensively and successively used throughout the world as MV cable terminations. This market is at present several times larger than the polymer insulator market. Nevertheless these materials have not been much used for insulator housings as few companies have managed to fonnulate them to be none trackrng both new and aged. Successful formulations often contain 10 or more components. They are relatively easy to stabilise against W plus pollution attack, and are easy to process. They have low MVTs and have much better physical properties than silicones or EPDMs EPDM rubbers are usually formulated with very high filler contents ( 65%) . This gives them outstandmg erosion and trackmg resistance but increases the difficulties inherent in the polymer of UV stabilisation. Water ingress problems at these high filler levels are also very difficult to avoid. The weathered surface of the insulatoryhas a large area of exposed filler, traps moisture and has a reputation like highly filled epoxies of conducmg very high leakage currents.These material are low cost and easy to process.

Unfortunately, the test requirements of IEC 1109 favour the selection of highly filled polymers to pass both the salt fog erosion test and the bunsen burner flammability test. This prejudices their W/pollution /wet electrical performance in favour of meeting either unrepresentative or irrelevant tests. These could be replaced by for instance weatherometer tests, power arc testing and/or IEC 587 track /erosion testing to determine susceptibihtyto the ignition sources likely in practice.

3/3

d
P 0 Y
L

!3

a .U
c)

cd

k
m

; I "
0
0

0 0

0 c

0 N
Y

e 4

77-

0 0

rr

m
CI

E:

e,
f, .Y
3 I

OJ

. I

E Y

3/4

0 1996 The Institution of Electrical Engineers. Printed and published by the IEE, Savoy Place, London WCSR OBL, U