ML12089A404
| ML12089A404 | |
| Person / Time | |
|---|---|
| Site: | Indian Point  |
| Issue date: | 03/29/2012 |
| From: | - No Known Affiliation |
| To: | Atomic Safety and Licensing Board Panel |
| SECY RAS | |
| Shared Package | |
| ML12089A391 | List: |
| References | |
| RAS 22115, 50-247-LR, 50-286-LR, ASLBP 07-858-03-LR-BD01 | |
| Download: ML12089A404 (2) | |
Text
ENT000245 Submitted: March 29, 2012 194 Chapter 3 Design and Manufacture of Extruded Solid-Dielectric Power Distribution Cables lead to dielectric failure. Reports of cable failures suspected to be caused by water trees about the time of the original study by Vahlstrom [39] also originated in Japan [42].
These unexpected discoveries of possible unsatisfactory service life of polyethylene and crosslinked polyethylene power cables were disturbing to the power utilities and the cable industry and led to urgent studies of the problem in man y countries including Canada [43- 46]. Most of the work ha s been carried o ut on HMWPE and unfilled XLPE. The quantities of EPR and filled XLPE cables in underground distribution are relative ly small ; furthermore the opaque nature of these materials makes the detec-tion of trees difficult. Trees can be observed easily in XLPE a nd HMWPE cables because these insulations are translucent. Water treeing of dielectrics is discussed and illustrated in Chapter 2. Water trees are apparent by their characteristic dark appea r-ance when contrasted to the translucent, thin wafers cut from the cable insulation.
Because of the lack of field evidence, it is not possibl e to conclude that EPR cables have failure rates comparable to HMWPE and XLPE. It has been reported that the treeing susceptibility of EPR cables is about equal to that in XLPE and less than that in HMWPE [47].
Water trees, sometimes called electrochemical trees, have basic characteristics dif-ferent than electrical trees. E lectrica l trees are characterized by the occurrence of partial discharge, require high electric stress to initiate and rapidly lead to catastrophic dielec-tric failure. Water trees can be initiated at much lower dielectric stress, grow very slowly, are associated with no measurable partial discharge, and may completely bridge the insu lation from conductor to shield without dielec tric breakdown , although the dielectric strength is much reduced , in partic!ar the direct current (dc) breakdown value [48]. Electric treeing, a well-known phenomenon since the early days of electrical engineering. occurs in poorly designed or overstressed insulation systems. The mechan-ism of failure is known and understood. Water treeing is a process studied intensively since circa 1970. but the incepti on and growth of the treelike st ructure has no univer-sally agreed theoretical baSIS. The literature on water treeing is large because the inves-tigations, although only encompassing a short time period. are intensive. Bernstein [49]
in his review or wate r treeing theory. gi\'Cs the major requirements and factors influen-cing the growth but suggests that the mechanism of inception is not known . It is accepted that two fundamental conditions arc required: (i) a polar liquid , usuall y water, must be present and (ii) \'oltage stress. Electrical trees require only voltage stress.
Other factors, listed in no particular order of importance, have been enumerated by Bernstein [49]. These arc aging time. material nature, contaminants/impurities , tem-perature, temperature gradient. cable design. magnitude of operating voltage stress, test frequency, antioxidant, voltage stab ili ze rs. water nature , and semiconducting layer type.
It has been shown that water in the interstices of the stranded conductor greatly enhances the tree growth even when the cable is immersed in water, particularly when a temperature gradient exists in the insulation [50]. Badher et al. proposed a physical model of aging in polymeric cables [51] and later proposed short- and long-term elec-trical tests based on the model [52]. Lyle and Kirkland reported on an accelerated aging test procedure for the gro wth of water trees and determination of cable life when subjected to various test conditions [53]. The importance of water in the strand and increased temperatures were again demonstrated.
Although the problem of water treeing has not been elimina ted, manufacturers are improving processing technology to improve the service life. Cables manufactured
,uti on Cables Section 3.5 Solid-D ielectric Insulati on Techniq ues 195
. wa ter trees today are improv ed over those that exhibit ed early failures circa Japan [42]. 1970. Tape strand shieldin g is now unacce ptable. The numbe r of voids, protrusions thylene and , and contam inants have been signific antl y reduced . Tree-re tardant XLPE insulati on
.les and the compo unds are now widely used. Some experts believe that voltage stress should be reduced
- s includi ng even though this action would increas e the first cost of the cable. Hermet ically sealed ll1d unfilled sheaths have been develop ed [54], althoug h these designs might be econom ically attracti jistribu tion ve only for trans-mission voltage s. Experie nce indicat es that such sheaths provide s the detec- good service lives for cables rated at 138 kV and higher. The great importa nce of keeping VPE cables water out of the strand is now accepte d , and several manufa cturers now offer a
- cussed and strand filler to preven t water ingress. Moder n cable design has improv ed marked ly over uk appear- the last decade , but a final so lution to water treeing and premat ure failure in wet environ insulati on. ments has not yet been assured . Figure 3.12 illustra tes the rising trend of fai lure rates.
EPR cables Recent field inves-tigation s [5 5] led to the conclus ions that failures in XLPE begin ed that the to intensif y after la- IS years of service and that water treeing and interna l defects (inclusi
- han that in ons) we re the most often iden tified problem s . Unfort unately , there is an unavoi dable time lag betwee n possibl e effectiv e correct ive measur es and corrobo rative evidenc eristics dif- e from service failure rates.
- e of partial phic dielec-grow very 3.5 SOLID-DIELECTRIC INSUL ATION TECHNIQUES
- tely bridge though the Since the 1950s there ha ve been great advanc es in the techniq ues of insulat ing medium -
breakd own voltage solid-d ielect ri c cab les. Former ly the insulati ons were the thermo setting* com-
)f electrical pounds based on butyl rubber or the thermo plastic compo und ,
polyeth ylene. The butyl le mechan - rubber insulan t wa s applied by an extrude r, called a tuber. The vu lcan ization of the in tensi vel y insulat ion was a separat e operation in an open steam vessel or autocla ve.
no UI1lver- Semico nductin g fabr ic tapes were utilized for conduc tor and insulat ion shieldin g. In
- the in ves-
~nstein [49]
~ 20.0
)rs influen -
own. It is 0)
- 10.0 1 ~/ 1
'~3 0.
id, usually ~ 5.0 tage stress .
lerated by o
o E
' '1:'
- ities, tem- ~ 1.0 7-1/~ 17'
%~;-:
stress, test 'J>
~ 0.5
- ting layer -/-
~
o - t>-~r-;t
-"+-0 a ,~_
L/- f -
~/~1 ,-+-=-\~!
~
- or greatly "
r1 y when a Figure 3.12 Failul"C r" tcs or po lymeric power E 0.1 a physica l cable rated 5- 35 kV: ' Irter Thue [4 6]: (a) and Z 0.05 (b). Average failure r;t te I*o r ,til c"blcs o perat-1964 1966 1968 1970 1972 1974 1976 1978 term elec- ing in a particula r ye' lr (eI) XLPE 'Illd (b) PE. Curves a and b ated aging (c) Failure I*ate vel*s u, number o i" years in ser- 2 4 6 8 10 12 14 16 life when vice: XLPE + PE. Year in service (curve c) trand and
- A thcrlll osC I i, d ef ined as a m a teri a l, cured under hea t.
that docs no t soften when rehea ted. Thi s definition is not stric*tl l" truc fo r cro sslin ked pol ye th ylene because
- turers are it docs soften and ha s lo wer ph ysica l strength tl1<ln Ill o, t \ ulclI1i/ed rubbers a t high temperat ures tUfactured . H owever. for practical cable applicati on s. th e definiti o n is , uit ahk .