ML11349A090

New York State (NYS) Pre-Filed Evidentiary Hearing Exhibit NYS000143, R. Ross, Ater Treeing Theories Current Status, Views and Aims,Proceedings of 1998 International Symposium on Electrical Insulating Materials, in Conjunction with 1998 Asi

ML11349A090
Person / Time
Site:	Indian Point
Issue date:	12/15/2011
From:	Ross R KEMA
To:	Atomic Safety and Licensing Board Panel
SECY RAS
References
RAS 21545, 50-247-LR, 50-286-LR, ASLBP 07-858-03-LR-BD01
Download: ML11349A090 (6)

Text

NYS000143 Submitted: December 15, 2011 Proceedings of 1998 International Symposium OIL Electrical Insulating Materials, in conjunction with 1998 Asian International Conference on El-l Dielectrics and Electrical Insulation and the 30th Symposium on Electrical Insulating Materials, Toyohashi, Japan, Sept. 27-30, 1998 WATER TREEING THEORIES - CURRENT STATUS, VIEWS AND AIMS R. Ross Transmission & Distribution Power KEMA, NL 6800 ET 9035 Arnhem, The Nethel;"lands INTRODUCTION CURRENT STATE OF UNDERSTANDING The matter of water treeing is both phenomenal in its Water in water trees economic impact and fascinating from scientific point of view. This is reflected in the vast number of papers The interactions of water with the polyethylene (PE that appeared since water trees were discovered in including both XLPE and LDPE) are described in Japan. Still, the subject requires attention to maintain [1,6]. Water is present mainly in four states: dissolved and increase insulatiun reliability. Recently, a review in the polymer; liquid in hydrophobic voids; liquid in was performed on the understanding of water treeing hydrophilic voids or channels; liquid at hydrophilic up to 1997 [ I]. The present paper is a condensed impurities or inclusions. The amount of water in each version of that review to which the reader be referred state is temperature dependent. On cooling down, for further details and little over 140 references. dissolved water starts to condense when its Subsequently, this paper discusses: the tc!rm 'water concentration equals th~ saturation level. Voids can treeing'; the most important properties of water trees; then be produced, probably mainly after extrusion.

the main water treeing theories; countermeasures.

PE in contact with extl~rnal (liquid) water, a "bath",

Definition of water treeing will be saturated. Watn in voids with hydrophobic walls will move oul, because the surface energy is The term 'water tree' was used in contrast to 'air always higher than at the PE-bath interface. Water trees' [2] refering to the ageing environment. Water moves into voids with hydrophilic walls. A typical trees have been identified with electrochemical trees pressure built up would be .1 MPa for a 1 11m void.

refering to a process [3]. They are also described by Dissolving species like salts trapped in voids gives a their image as diffuse structures with a bush or fan- pressure up to e.g. 25 MPa with a tangential pressure like appearance [4], being able tu disappear and of 12 MPa. Creep uccurs until the pressure equals the reappear on wetting. In polystyrene, they found non- polymer yield strength. Water trees are hydrophilic.

diffuse water trees [5]. The following definition is [1] Thus at least hydration should take place.

covers most findings over the past three decades:

Important characteristiics of water trees Water trees are degradation structures in a polymer that arc: There is a broad consensus on the most important 1 permanent properties uf water trees. Some of them appeared in 2 grown due to at least humidity and an electric field the proposed water tree definition.

3 electrically weaker than the original polymer when wet, but which are not a short circuit or local Permanent and hydrophilic Water trees can be dried, breakdown path but water will enter again in a sufficiently humid 4 substantially more hydrophilic than the original environment. Drying ar,d wetting can be carried out polymer. repeatedly. As water trees are permanent hydrophilic structures, the chemical' composition seems different Point 3 and 4 still have to be quantified. The process from the hydrophobic PE in addition to a likely of water treeing is a degradation phenomenon that change in physical structure.

produces water trees. The definition stales that waler trees are the degradation structures themselves. A Stainability Staining methods for water trees have string of water droplets would not be the water tree, been developed. A dydng mcthod using methylene but the inclusions and hydrophilic voids would. Also blue is recommended by Cigre [7]. Metbylene blue is the water tree does not disappear, but it just dries out. a redox indicator which is applied and buffered in its Though it is attempted to propose a clear definition, it oxidized state to maintain a bright color. The staining is not claimed that this definition will conclusively is not of chemical nature, but rather a diffusion and stake out the subject of water treeing. clustering of the dye in the water tree [IJ.

- 535 -

OAGI0001271_00001

Ouantitative water content The water content in Dielectric properties Wet water trees have different water trees is typically about 1%, i.e. a factor 100 dielectric properties than untreed PE. However, dry higher than in PE. Most of the water is in a liquid water trees have practically the same dielectric state in voids. In a needle test about 10% water was properties as unaged PE [16,9]. Wet water trees have found near the needle tip [8]. This high content is due an £ of 2.3 up to 3.6 [1]. With respect to tano, the to the very high ageing stress of 60 kV/mm. More results vary [10]. Nowadays, most agree that wet remote from the needle, the field is lower and a water trees have a higher tano, though still in thc content of 1-2% was found. In service-aged cables insulator range. Un treed PE has a tano=.0005. Wet lower water contents are found. In vented trees of samples having water trees from field-aged cables had field aged cables, infrared measurements showed tano"".002 at .01 kV/rnm and tano"".02 at .3 kV/mm

.8-.92% water [9]. The water content can be increased both at 50 Hz [12]. This is still in the insulator range .

by additional damage like boiling water trees or The charge carriers are probably ions. The different vacuum extraction plus rewetting. and non-linear tano opens the door to diagnose cables.

Voids The density of microvoids in the water tree tip The breakdown strength is reduced by sufficiently is similar to that in PE [10]. A higher void density is large and wet water trees [4,10]. As bow-tie trees stop often found for the main part of the tree. By TEM and after initial rapid growth whilst vented trees continue SEM also a greater void density is found in water to grow, the latter control the breakdown strength on trees than in PE [11,3]. On the other hand, some the long term [10]. However, in several cases, water evidence in the literature for voids is suspect: sample trees bridged the entire insulation without causing handling can easily produce voids in water trees by breakdown at operating voltage [17]. A relation boiling and/or drying wet samples by vacuum in an between the largest water tree length and breakdown exsiccator or an electron microscope. This is even strength has been reported [18,10]. In practice, used as a contrasting method [12]. Samples can be breakdowns often occur with impulse voltages.

dried with a drying agent like silica gel or phosphor pentoxide. Although many voids can have been caused CURRENT MAIN THEORIES or enlarged by researchers, the higher density of voids in water trees is generally accepted. The present technology and testing is based on more than one theory by lack of consensus. Though the Channels and tracks The prcsence of elongated theories tend to converge, still several theories exist:

structures seems to differ from case to case even in 1 electro-mechanical one investigation [12]. The various forms of water 2 diffusion trees in polystyrene in one experiment is a beautiful 3 electrochemical oxidation example [5]. Micro-channels as well as absence of 4 condition dependent model helium permeability are found for PE [13,14,15]. These theories will be addressed in the following.

The tenn 'channel' refers to an open duct allowing Electro-mechanical theories transport of water. There are however more subtle means like water hopping along hydratable sites. A Theories of water treeing by electro-mechanical forces, string of hydrophilic footholds (e.g. trapped ions) concern types of stress-cracking or electrically driven

'woven' through the matrix would provide tracks or diffusion of water.

paths for water, i.e. the water trees [3]. The presence of nanometre tracks are indicated by TEM studies e.g. Cracking An early consideration was that water

[11], dielectric measurements [9] and semI- treeing might be induced by supersaturation mainly.

permeability experiments [12]. As a result, crazing of the polymer would occur both in bow-tie and vented trees. Ildstad et al. stated that Methylene blue staining of water trees often seems to Maxwell stresses could cause a pulsating compressive reveal micro-branches. However, the same occurs if stress acting perpendicularly to the crack surface [19].

dye molecules cluster at specific cndgroups or ions in Particularly, internal stresses in the range of the PE nano-tracks. Such clusters can grow with other dye yield strengtll might be introduced during the molecules. Not only the water tree paths are stained, manufacturing process. Therefore, only small forces but also the (sub )micron environment. A macroscopic might be required to cause water treeing. Mcchanical analogon is a rope underwater looking thicker when damage on a smaller scaIc resulting in microcavities seawecd clusters at it. Similarly, water tree nano-tracks has been suggested as well [20]. The idea is similar.

can get a micro-channel appearance with dyeing. Electrical tields in combination with water lead to Though microchannels frequently occur, the main mechanical forces in the material. Chain scission elongated structures in water trees rather seem ionic occurs if the resulting stresses exceed the local nanometre tracks than hollow (sub)micron channels. mechanical strength of the polymer matrix .

.- 536 -

OAGI0001271_00002

The theories about electromechanical forc.!s describe Diffusion theories realistic phenomena. However, c1earcut evidence that it is the main ruling mechanism is certainly missing. The hydrophilic propenty of water trees indicates a Calculations by Zeller [6] and Steennis [LO] for PE change in chemical composition. One way to achieve indicate that the mechanical stresses are much weaker this, is by diffusion of hydrophilic species into the than in early estimations. The expected mechanical insulation. Unlike surfac:tants and apolar solvents like damage in the form of cracks is rarely found. pentane, hydrated ions will not penetrate PE easily.

Moreover, water-filled cracks and voids in PE would Diffusion can be driven by an electric field, which is still be hydrophobic. also called electro-osmosis [3]. Diffusion of ions and contaminations has been regarded as a side effect that On the other hand, electro-mechanical forces are contaminates the water tree. In the current theory supposed to take place to a certain cxtcnt. In needle solvable species are supposed to play an important tests, the electric field can be high enough for role and, moreover, the ions and other hydrophilic electromechanical deterioration [6]. In weaker species are regarded tu be essential parts of the water materials like oils, such forces can form channels [21]. tree backbone if not be the tree itself.

In cable insulation, high stresses can be expected from osmotic actions. Particularly for bow-tie trees starting There is overwhelming evidence that field driven from hydrophilic nuclei, the initial steps could be due diffusion of hydrophilic species occurs if such species to osmotic pressure. The role of pure osmotic pressure are available. In their discussion of electro-osmosis, in degradation is already played during production, Dissado et al. note that this phenomenon would curing and cooling down of the cable. The water particularly take place with AC fields [3], which is in content of the PE in buried cables is probably lower agreemcnt with practice. Experiments also show that than in freshly produced cables [1]. water trees rarely grow in the absence of such species.

Cables produced with cleaner materials suffer less On the level where thermodynamic quantities are water tree growth usually. Particularly semiconductive applicable, the mechanical fatigue model does not shields appeared sources for ions [23]. The impurtance seem sufficient. It is found though that stresses can of the field is alsu apparent. Hydrophilic species can exist locally, and that much lower stresses than the certainly provide the water tree with its hydrophilic yield strength can cause some local damage. The character. Therefore, the ingress of ions is regarded difference is that for water treeing a few local bonds rather instrumental to water treeing than a side effect.

may be broken, while for macroscopic fracture also the strongest bonds must be broken. When the However, it is also clear that ions do not unly act as discussion is taken at that level, the final verdict inert charge carriers. The chemical nature and physical cannot be expected at this stage. properties of ions have a great 'Impact on the growth rate of water trees. Some hydrated ions can Dielectrophoresis The theory of dielectrophoresis is corrosively attack polymer chains and cause a more more or less related to the electromechanical model. open structure, while smaller hydrated ions may have With dielectrophoresis the electric field pushes water a higher diffusion constant. Some ions can even act as into the polymer, but does not necessarily induce inhibitors. Other processes which help producing a mechanical damage. Patsch et a!. describe the theory more open structure or hydrophilic environment should of dielectrophoresis [22]. Once then: is an therefore be considered as well.

inhomogeneous field, dipoles will move towards sites with higher electric fields. Such sites can be due to an Electrochemical oxidation impurity or another disturbance. Such spots are widely regarded as initiation points for bow-tie trees. Another theory explaining the hydrophilic nature of water trees, states that the polymer itself is (electro-)

Dielectrophoresis certainly takes place. However, also chemically modified. A range of subtheories exists on from Zeller's calculations this seems not enough to the reaction details, which are still unclear. The cause permanent damage [6]. Like the case of the oxidation products can be discussed in greater detail.

mechanical fatigue models, dielectrophoresis cannot On the basis of those oxidation products the theory is account for the permanent hydrophilic behaviour of split into a pure oxidative theory and a combined watcr trees. Therefore, the theories on the basis of oxidation-diffusion theory.

pure electrophysical processes on water cannot account for water treeing. However, they can play a role in It is also considered that oxidation could take place combination with diffusion theories. before [24] or after [16] the actual water tree growth.

There seems to be no ckar evidence for separating the water treeing proce~s ,md oxidation~

- 537 -

OAGI0001271_00003

metal ions in addition to oxidation. Carboxylate groups o

II o o o can form dimers or larger structures by sharing their R-C-H metal ions. Thus nano-tracks can be formed which facilitate water transport. Carboxylate (salt) groups and aldehyde keton ester carbolWlic acid sulphates are considerably more hydrophilic than the dipolar groups including carboxylic acid.

Fig.I: polar oxidation products: carbonyls etc. O~

)C-R Pure oxidation The oldest oxidation theory concerns 10 if) OM 0'/

R-~G M pure oxidation. Water trees were regarded as electro- o R-q chemical trees quite early, mainly inspired by water OM etc.

carboxylate carbo?<ylate tree properties like permanent hydrophilicity and the group salt clustering need of other species. The polymer turns into a polar material containing oxidation products shown in Fig.I.

This provides a favourable environment for water. The Fig. 2: carboxylate groups oxidation can take place at void walls or develop a track through the amorphous phase with incidental In many cases carboxylate groups are found in water void formation. Suggestions for reaction details can be trees particularly in service aged cables. However, the found in various papers, but no specific intermediate concentration is sometimes very low and approaches reaction steps have been evidenced explicitly to the the detection limit of FTIR. Good spectra and a HCI-best of the author's knowledge. As oxidation gas test [12] are then required to evidence the mechanisms can be very complex, each guess seems carboxylate groups. In some other cases carboxylate equally valuable at this stage. groups are reported not to appear. This may be because the HCI-gas test had not been used, but the The required oxygen can have various sources, like carboxylate content varies with the ageing conditions oxygen dissolved in water [24] or electrolysis of water [26] and in certain cases carboxylate groups may not yielding various species among which radicals like be present at detectable levels if not absent. Far less oOH, which can oxidize PE [10]. Once radicals or oxidation products are found generally with needle ions are available, oxidation processes can take place tests. It is assumed that the importance of the leading to carbonyl groups in many cascs. oxidative formation of ionic groups is reduced at high stresses and when large quantities of other hydrophilic Pure oxidation introduces polar groups which can species are available. The theory is supported by the provide the permanent hydrophilic nature of water success of reducing water tree growth with the use of trees, though not as effectively as hydratable ions. In cleaner materials [27,23]. The reduced supply of metal contrast with practice, water treeing with distilled ions can become rate determining for this mechanism.

water at moderate stresses would seem possible owing to a pure oxidative mechanism. The need of species Condition dependent model like catalysts or cOlTosive contaminants might resolve this contradiction. Still, water treeing is enhanced by It is considered that water treeing is a complex of many species, that arc not oxidation catalysts [25]. processes [4,13,16,26]. Certain processes can be Though support exists for hydrophilic groups attached mutually substitutable to produce a very hydrophilic to polymer chains [12], the oxidation products shown environment: diffusion of ions or other hydrophilic in Fig.1 are frequently not found in water trees. This species into the polymer; oxidation of the polymer means that the required degree of oxidation is too low (preferably in combination with metal ion diffusion);

to be detected or that in many cases the pure oxidative and oxidation of antioxidants (idem). Similarly, voids theory is not applicable. An intermediate concept can be created by different processes involving chain could be that oxidation provides the paths while other scission like electro-mechanical stress and chemical processes also take place in parallel. reactions. Which process is dominant depends mainly on the set of ageing conditions in specific cases.

Combined oxidation and diffusion The second oxidative mechanism is a combination of oxidation This is the core of the condition dependent model, that and diffusion and produces ionic groups: carboxylate regards water treeing as a degradation phenomenon groups (Fig.2). Such a group can be obtained by e.g. that produces a hydrophilic network by one or more carboxylic acid trading its H+-ion for a metal ion, thus processes that are most favourable under the whole set becoming a carboxylate salt (group). Similarly sulphur of ageing conditions. The model tries to relate containing antioxidants can produce sulphates. The conditions and processes. The main driving force is production of ionic or salt groups requires diffusion of the electric field. The system seeks enhancement of

- 538 -

OAGI0001271_00004

the permittivity by intrusion of water and the Opinions During almost thirty years of water tree enhancement of the conductivity. Likely processes on awareness, many hypotheses were brought up. Clearly, the basis of degradation products in water trees are: even nowadays, still different opinions exist both on oxidation because of oxidation products; (mechanical propagation and inception mechanisms. Part of these and/or chemical) chain scission because of excess are undoubtedly due to different ageing conditions.

voids and of carboxylate groups; diffusion because of The main current theories are: electromechanical contaminants, electrolyte and counter ions. (cracking and dielectrophoresis), diffusion, electrochemical oxidation (pure and combined with As for evidence, various contradicting results in the diffusion) and condition dependent model.

literature can be traced back to different ageing conditions like the electrical parameters, the involved The latter model acknowledges the discrepancies of chemicals, etc. This supports the idea of processes that experimental findings where different processes partly can substitute each other to a large extent. Depending compete, partly concur. At least oxidation, chain on the exact conditions, such processes can concur or scission and diffusion are important during water tree compete. It is likely that changes in ageing eonditions growth. For propagation processes at service stresses will favour one process over another [1]. particularly electro-osmosis and oxidative processes are likely to occur on the basis of chemical analyses The model as such is not new as it combines the and dielectric properties. At higher stresses as in various processes that seem physically likely to occur needle tests, electro-me~chanical processes become within the range of practical ageing conditions. The more likely.

degradation is regarded both electrochemical and electrophysical in naturc. This means that accelerated Testing Though most modern cables are much less ageing studies should include some way of judging susceptible to water treeing, testing remains necessary which process(es) took place and evaluate whether this as a means of quality control. New diagnostic methods is relevant to practice. This might be clone with are emerging. The probable exchangeability of various chemical analysis afterwards or with dielectric water treeing processes urges to perform realistic tests.

measurements during ageing. The model should not For example, high electric fields selectively promote become a permit to neglect the critical analysis of physical processes both in cable and material tests.

mechanisms and experiments. The merit of the model Acceleration can be obtained also with higher is that it helps to find appropriate solutions when more frequencies. In material tests, problems occurred with than one mechanism can be active. unwanted electrolyte sources like steel or brass in contact with a well defined NaCI solution [1].

DISCUSSION AND CONCLUSIONS Research A problem in water tree research is of Though the subject of water treeing is still of great course, the slow growth of water trees in practice; i.e.

concern for the reliability of cables, great progress has typically <1 fU11Iday. Accelerating water treeing is been made both scientifically and technologically. The usually inevitable, but care should be taken with success of countermeasures may be the reason why the applying severe ageing conditions. The relevancy of subject draws less papers on recent confere~nces than various processes to service ageing should be in the eighties and early nineties. preserved and preferably evidenced. This can lead to systematic and effective countermeasures and to Technological progress Technology has made great correct testing.

progress after the identification of important key factors like humidity, cleanliness of materials and For studies of the water tree microstructure, the use of smoothness of surfaces. Water tight barriers like metal boiling or vacuum treatment is discouraged, because layers are used to keep the humidity low. This method the morphology is changed. Samples can be dried with is mainly applied in transmission cable for economic a drying agent. For staining, the standard methylene reasons. Swelling tape as drying agent with a PE blue method is suitable for length measurements. To sheath for low water permeation is an emerging evaluate the presence of sulphates or carboxylate method for distribution cables. The success of cleaner groups, the RCl-gas test is recommended. Particularly insulating and semiconductive materials probably due for low concentrations, the method is helpful.

to reducing the availibility of metal ions and hydrophilic species. Care must be taken that other ACKNOWLEDGEMENT materials in cables will not act as sources of contaminants. Smooth surfaces with triple extrusion The present paper is base~d on a review [1] carried out processes avoid electric field concentrations. Water by IWO (Institute for Science and Development, tree retardants have been developed based on various NL 6711GG 46 Ede, The Netherlands), in cooperation philosophies. with KEMA.

- 539-OAGI0001271_00005

REFERENCES [l5] J.D. Cross & J.Y. Koo, "Some Observations on the Structure of Water Trees", IEEE Trans. EI-19, pp.

[1] R. Ross, "Inception and propagation mechnisms 303-306, 1984.

of water trceing", Annual Report CEIDP, 1998 [16] J. Densley, "Ageing and Diagnostics in Extruded

[2] T. Miyashita & T. Inoue, "Treeing Phenomena Insulations for Power Cables", Proc. 5th ICSD, pp.

in Polyethylene-Coated Wire Immersed in Water", 1-15, 1995.

Electr. Eng. in Japan, VoL90, No.3, pp. 83-91, 1970. [17J G. Bahder, C Katz, J.H. Lawson & W.

[3J L.A. Dissado, S.V. Wolfe & J.C Fothergill, "A Vahlstrom, "Electrical and Electro-chemical Treeing Study of the Factors Influencing Water Tree Growth", Effect in Polyethylene and Crosslinked Polyethylene IEEE Trans. EI-18, No.6, pp. 565-585, 1983. Cables", IEEE Trans. PAS, VoL93, pp. 977-990, 1974.

[4] MT. Shaw & S.H. Shaw, "Water Treeing in [18] H. Franke, H. Heumann & D. Kaubisch, Solid Dielectrics", IEEE Trans. EI-19, No.5, pp. 419- "Testing Possibilities and Results Regarding Water 452, 1984. Aging of PEIXLPE Insulated Medium Voltage

[5] S.H. Shaw & M.T. Shaw, "Water Treeing Cables", Proc. Jicable 84, pp. 113-118, 1984.

Phenomena in Amorphous Dielectrics", Proc. 1990 [19] E. Ildstad, H. Bfudsen, H. Faremo & B. Knutsen, IEEE Int. Symp. on Electr. Ins., pp.187-190, 1990. "Influence of Mechanical Stress and Frequency on

[6] H.R. Zeller, "Thermodynamics of Water Water Treeing in XLPE Cable Insulation", Proc. ISEI, Treeing", IEEE Trans. EI-22, No.6, pp.677-681, 1987. pp. 165-168, 1990.

[7J P.B. Larsen, "Dyeing Methods Used For [20] Y. Poggi, V. Raharimalala, J.C. Filippini, .T.J. de Detection Of Water-Trees In Extruded Cable Bellet & G. Matey, "Water Treeing as Mechanical Insulation", Electra, pp. 53-59, 1983. Damage - Macroscopic and Microscopic Approach,

[8] CT. Meyer, "Water absorption during Water Influence of Test-specimen Process Parameters", IEEE Treeing", IEEE Trans. EI-18, No.1, pp. 28-31, 1983 Trans. EI-25, No.6, pp. 1056-1065, 1990.

[9J F. Stucki & J. Rhyner, "Physical Properties of [21] Z. Iwata, T. Fukuda & K. Kikuchi, Single Water Trees Extracted from Field Aged "Deterioration of Cross-linked Polyethylene due to Cables", Proc. 4th ICPADM, pp. 391-393, 1994. Water Treeing", 1972 Annual Report CEIDP, pp.

[lOJ E.P. Stccnnis & P.H. Kreuger, "Water Treeing 200-210, 1973.

in Polyethylene Cables", IEEE Trans. EI-25, No.5, pp. [22] R Patsch, A Paximadakis & P. Romero, The 989-1028, 1990. Role of Dielectrophoresis in the Water Treeing

[IIJ L.J. Rose, V. Rose & J.J. de Bellet, "The Effect Phenomenon", Proc. IEEE ISEI, pp. 160-164, 1990.

of Polymer Microstructure on Water-Tree Performance [23J S. Pelissou, "Impurities in XLPE Cable Resins",

in Crosslinked Polyethylene for Cable Insulation", Proc. 4th ICPADM, pp. 412-415, 1994.

Proc. 2nd ICSD, pp. 237-244, 1986. [24] B. Noirhomme, A.T. Bulinski, S.S. Bamji, J.P.

[121 R. Ross, "Composition, Structure and Growth of Crine & R.I. Densley, "Oxidation Phenomena in Water Trees in Polyethene", KEMA Scientific and Water Treeing", EPRI Report TR-102769, 1993.

Technical Report, pp. 209-271, 1990; also: KEMA, [25J R. Patsch, M. Ortholf & J. Tanaka, "Hydration Arnhem, ISBN 90-353-0080-7, 1990. of Ions - How does it influence Water Treeing?",

[13J W. Kalkner, U. Muller, E. Peschke, HJ. Henkel Proc. 5th ICPADM, pp. 410-413, 1997.

& R. von Olshausen, "Water Treeing in PE and XLPE [26] R. Ross, "Effect of ageing conditions on the type Insulated High Voltage Cables", Proceedings Cigre, of water treeing", IEEE ELInsuLMag., VoL9, No.5, paper 21-07, 1982. pp. 7-13, 1993.

[14J E. Moreau, C. Mayoux & C. Laurent, "The [27J K. Watanabe, H. Yagyu, K. Sasaki & M.

Structural Characteristics of Water Trees in Power Marumo, "Semiconducting Shield Materials for XLPE Cables and Laboratory Specimens", IEEE Trans. Cables", Hitachi Cable Review, No.8, pp. 15-20, 1989.

EI-28, No.1, pp. 54-64, 1993.

- 540 -

OAGI0001271_00006