ML20206K149
| ML20206K149 | |
| Person / Time | |
|---|---|
| Site: | Sequoyah  |
| Issue date: | 04/08/1987 |
| From: | Gridley R TENNESSEE VALLEY AUTHORITY |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| NUDOCS 8704160281 | |
| Download: ML20206K149 (20) | |
Text
'
e TENNESSEE VALLEY AUTHORITY CH ATTANOOGA. TENNESSEE 37401 5N 1578 Lookout Place APR 081987 U.S. Nuclear Regulatory Conunission ATTN: Document Control Desk Washington, D.C.
20555 Centlemen:
l In the Matter of
)
Docket Nos. 50-327 Tennessee Valley Authority
)
50-328 SEQUOYAH NUCLEAR plt.NT (SQN) - CABLE TEST PROGRAM
References:
1.
Letter from B. J. Youngblood to S. A. White dated March 9, 1987, " Evaluation of Sequoyah Units 1 and 2 Cable Pulling and Bend Radii Concerns" 2.
Memorandtun from Thomas W. Alexion to B. J. Youngblood dated February 20, 1987, " Status of Cable Pulling Review Efforts" 3.
Letter from B. J. Youngblood to S. A. White dated January 15, 1987, "Reconunended Actions for Resolution of Specific Cable Pulling Concerns, Sequoyah Units 1 and 2" The purpose of this letter is to formally document TVA's cable test program and to resolve NRC's concerns regarding cable pullbys, januning, and lack of support of vertical cable runs with 90-degree condulets at the top of the run. These concerns were provided in reference 1 above and were discussed at meetings between TVA and NRC in Bathesda, Maryland, on January 29, 1987; March 3, 1987; and March 11, 1987. A TVA draf t response to NRC's reconunended actions and NRC's draf t conunents on this response were provided in reference 2.
NRC's evaluation and concerns on cable pulling were further.
detailed in reference 3, the Technical Evaluation Report (TER), which was provided to TVA at the March 11, 1987 meeting.
The purpose of the program is to conduct in situ nondestructive tests on representative worst-case cables. This program will support the adaquacy of SQN's installation practices and will verify the performance capabilities of the cable system in view of the concerns.. The proposed criteria for cable selection, testing, inspection, and acceptance for cable pullbys, januming, and cables supported by 90-degree condulets are provided in enclosures 2, 3, and 4, respectively. TVA will replace any damaged conductors discovered during sampling before restart of the affected unit. TVA will also support 10 CFR 50.49 silicon rubber insulated cables inside containment that are under significant strain inside a conduit with 90-degree condulet at the-top'of a long vertical run before restart of the affected unit.
i l
i 8704160281 870408 g6[
i PDR ADOCK 05000327 P
- l. l An Equal Opportunity Employer I
.---,.---,,...,._.,,,..,,,,._,_,.m
.,,_,_,w,_.__
m.
,,_~ ~,m m.,,,,,
,,,v.
r q
,s U.S. Nuclear Regulatory Connission 1091987 In evaluating this cable test program, TVA requests NRC to consider the merits and basis of this program as listed in enclosure 1. lists the commitments made in this letter.
Very truly yours.
TENNESSEE VALLEY AUTHORITY t/-
R.
- ridley, irector Nuclear Safe y and Licensing Enclosures cc (Enclosures):
Mr. G. G. Zech, Assistant Director for Inspection Programs Division of TVA Projects Office of Special Projects U.S. Nuclear Regulatory Commission 101 Marietta Street, NW, Suite 2900 Atlanta, Georgia 30323 Sequoyah Resident Inspector Sequoyah Nuclear Plant 2600 Igou Ferry Road Soddy Daisy, Tennessee 37379
I ENCLOSURE 1 SEQUOYAH NUCLEAR PLANT i
BASIS OF TVA CABLE TEST PROGRAM 1.
The SQN installation configuration does not significantly differ from other plants of its vintage. In many respects.the installations reflect conservatism such.as the installation of numerous pull points and short conduit runs. The walkdowns performed by TVA recognized these SQN installation practices.. This should be considered when determining the extent of testing required for resolution of issues on SQN.
2.
Recent interviews of electricians who had actually performed the cable installation, conducted by the NRC/ Consultants, indicated a thorough understanding of the cable installation process and the relevant concerns.
These interviews substantiated conformance to proper construction practices used throughout the industry and utilization of the engineering approved installation procedures and specification, and confirmed the presence of a Quality Control Inspector on all Class 1E cable pulls.
In addition, the close working relationship between construction and engineering was outlined including the evaluation made before each Class IE pull to determine the best method of installation, which included determining the direction of the pull and the need for manual assistance at pull points to reduce tensions.
The testimony in these interviews was contrary to employee concerns in this area.
3.
The test voltage by TVA is in accordance with the recognized industry standard test (IEEE 383-1974, Section 2.3.3.4) for determining that cables are adequate to perform their intended Class 1E function. This test accounts for the known and accepted reduction in cable dielectric strength because of the normal rigors of shipping, handling, installation, and operation.
In addition, this test voltage is consistent with testing previously performed, and accepted by NRC, for the identical concern of cable pullbys at the Grand Gulf station. This test voltage was selected over the ICEA factory voltage since testing of installed cables at the factory voltage exceeds manufacturers' recommendations for installation and maintenance-proof testing. As such, use of the factory test voltage at this stage in the life of SQN cables may damage cables which are otherwise acceptable and inhibit a root cause determination of any test failures.
4.
The in situ de high potential tests by TVA, which will be performed in the presence of moisture, correspond to the recognized industry test for determining the operational condition of the cable (IEE 383-1974) while simulating the condition expected during a DBE.
This test method does not degrade or damage the cables in question and therefore allows for determination of the true cause of any test failures.
IEEE 400-1980 recommends this type of testing not only for determining the present-operating condition of the cable but also for detecting cables which 'are approaching failure. This test method was previously utilized at Grand Gulf, and accepted by the NRC, as a method to determine the existence and extent of cable pullby damage.
f
, 5.
The acceptance criteria by TVA for the de high-voltage testing is in accordance with that by which the cables final acceptability is determined in the factory. To assist in the establishment of the acceptance criteria, TVA requested an evaluation and recommendation from an independent consultant in the nuclear cable industry.
The results of the evaluation and the recommendations of the consultant, Reinhold Luther, as well as a summary of his qualifications are included in attachment 1.
6.
SQN has completed tests on its cable system consistent with current day industry practices that establish adequacy of the installation.
n
ATTACH!ENT 1 l
REINHOLD LUTHER, P.E.& LS.
735 IMUTTLE MEAOCW AVI.
.... c.....
neou<n esens sas.nes March 15, 1997
.W.S. Raughley, Chief Electrical Engineer Tennessee Valley Authority - V 8 C 126 C-K 400 West Su=it hill Drive hoxville, TN 37902 SU3 JECT: In-Situ D.C. Testing of Low Voltage Power and Control Cables.
Dear Mr. Raughley,
As requested, I have investigated the various aspects of high voltage D C testing of non shielded lov voltage power and control : ables. Although it is not customary to test these cables in the field, it is possible to do so and to prove that they are as good - allowing a slight =argin for insulation degradation during shipping and installation -as they left the cable anufactu-rer's fac111 tes.
Insulation characteriscies, ratings, :esting and other aspects of these cables are covered by applicable Insulated Cable Engineer's Association ( ICEA )
standards. These standards require that each length of co=pleted cable be high voltage tested at specified voltages based on insulation thickness, utilizing either alternating or direct-current voltages, even though these test voltages are below the intrinsic.e.apabilities of the cables, they are deened to have enou2h cargin over the operating voltages to be suitable for whatever appli-cation they are intended, including harsh environ =ent and LCCA. I have con-tacted six major manufacturers of this category of cables ( Boston Insulated Wire, Brand-Rex, General Electric, Ierite. Okonite and Rockbestos ) and found that while so=e use alternating and so=a direct - current voltages for this final test, all of ther. reported that the testa are' effective and sone cables fail the teet and are rejected. It follows therefore, that cables containing spots that vould have failed at slightly higher test voltages leave the fac-tory and are installed without any questions regarding their integrity.
As pointed out earlier, no further testing of these cables is required until the'y are placed in service. It is logical to assume, however, that some insulation degradation takes place during the transit from the factory, hand-ling and installation, no matter how stringent and exact the specifications.
Since in-situ testing of these cables is not normally done, industry guide-lines are. lacking, buc all of the manufacturers contacted recommended that field test voltages should be approximately 80 percent of those utilized in the factory. This value seems to be acceptable. In my opinion to test at a higher voltage could not only damage the cable, but also lead to endless debates as to whether the cable failed because of damage during installation or an incipient fault created in the factory.
-1
r The ICEA standards state that ".. the ra$e of increase from the initially applied voltage to the specified test voltage shall be approxi-mately uniform and shall be not more than 100 percent in 10 seconds nor less than 100 percent in 60 seconds. I see no reason to deviate from this pro-cedure.
As far as acceptance criteria is considered this, by definition, is a withstand test (so-no go). The test voltage is applied and held for the specified ti=e If the test set does not trip, the cable passes the test. It was confir=ed by all canufacturers contacted that this is the pass / fail cri-teria they use, although it should not be made a part of the pass / fail cri-teria, I do reco==end that the leakage currents at the final test voltage be recorded for engineering infor=ation only.
In conclusion, I reco= send that for the proposed in-situ D C test on
'non shielded low voltage control and power cables, TVA utilize a 240 volts per mil'of nominal insulation thickness test voltage. The application of the test voltage shall be at a uniform rate that reaches the final test value in between 10 to 60 seconds. The cables pass the test if the leakage currents. stabilize at the end of the test duration ($sinutes). I also reco=cend that the leakage currents be recorded for engineer'ing infor:atirn only.
Please es11 en at (203) 223-3491 if you have questions regarding the above.
Very Truly Yours lR.LutherP.E..bY Consultant Generating Station Cables e..
- i..
l' SIGNIFICANT ACHIEVEMENTS IN CABLE ENGINEERING R. LUTHER GENRRAL Member IEEE/ Power Engineering Society.
Secretary of the Insulated Conductors Committee - Past Chairman of the Cable Accessories Subcommittee.
Member Association of Edison Illuminating Companies (AEIC) Cable Engineering Section..This is a utility-sponsored organization that prepares specifications for distribution and transmission cables.
Member EPRI's Underground Transmission Task Force.
(Term expires December 31, 1985.) Presently Chairman of Extruded Cable Subcommittee and member of Cas Cable Subcommittee (past chairman).
Member NU's Distribution Research Subcommittee.
Member NU's Transmission Research Subcommittee.
Advisor - University of Connecticut Institute of Materials Science / Electric Materials Research center.
Registered Professional Engineer and Land Surveyor, State of Connecticut.
CENERATING STATION CABLES Actively participated in Work Group 12-32.
This work resulted in IEEE Standard G83, " Standard for Type Test of Class 1E Electric Cables. Field Splices, and Connections for Nuclear Power Generating Stations." This standard was endorsed by the NRC and forms the basis for all nuclear cable purchases.
Conducted technical evaluation and made purchase recommendations for cables utilized in Millstone unit 2.
Although IEEE Standard 383 was not yet published, all cables met its requirements, which, in turn, was an asset during ilcensing procedures.
Assisted Generation Engineering in the resolution of numerous cable problems in units 1 and 2 at the Millstone complex, Connecticut Yankee, and during construction of Middictown unit 4.
Led NU's efforts in upgrading cables for fire resistanco at Connecticut Yankee and Millstone 1, as mandated by NRC after the Browns Ferry fire.
Prepared NU's specifications for nuclear grade instrumentation, control,
- Low-voltage power, and high-voltage power cables.
Was instrumental in establishing NU's nuclear cable stock program.
Established approved supplier's list for nuclear station cables.
Conducted technical evaluation and made purchase recommendations for cables utilized in Millstone unit 3.
Was instrumental in forming and actively participated in Insulated Conductors Committee's Task Force 12-40, which resulted in IEEE Standard 634, " Standard Cable Penetration Fire Stop Qualification Test."
This standard is now being utilized for Millstone unit 3.
Presented paper to the American Chemical Society, October 1978, in Boston,
" Performance Requirements for Rubber-Insulated Cables in Nuclear Generating Stations." A condensed version of the paper was published in DuPont Company's " Elastomer News" in seven languages.
It resulted in requests for copies of the paper from all over the world, including China, Japan, India, South Africa, etc.
Presently providing support to Generation Engineering and project personnel in resolving several "Non-Conformance and Disposition Reports" regarding cable installation at Millstone #3.
Critical for NRC operating license.
DISTRIBUTION CABLES Presented paper at the ECNE T&D Committee Meeting, September 21-22, 1978,
" Evaluation of Cross-Linked Polyethylene and High Molecular Weight Polyethylene Insulated Cable Performance."
Industry Advisor for EPRI's Project 194, " Laser Detection of Voids and Contaminants in Polyethylene Insulated Cables." Results of this project formed the basis for a technical paper presented at the IEEE 1982 Summer Power Mocting.
Coauthored with Prof. Tanaka, of the University of Connecticut, " Analysis of Cables with Visible Halos " presented at the 1982 International Symposium on Electric Insulation, June 9, 1982.
Discovered a manufacturing defect in cables delivered to NU during 1976-78 period. Corroborated with NU's purchasing departmcnt to reach settlement with the manufacturer with a possible payback to NU of $100,000. Prepared and issued OAP 214, " Failed Cables Subject to Manufacturer's Rebate."
Corroborated with Distribution Standards to establish an approved supplier's list for Ethylene Propylene Rubber (EPR) insulated cables.
(Derite, Okonite, and Anaconda.
Anaconda subsequently stopped manufacturing these cables.)
Prepared sections on Undernround Systems and Materials Research for NU's Distribution Research Plan.
Prepared and issued OAP 215, "DB Electrical Cable - Failure Tracking and Replacement."
Prepared DER 82-15 "EPR vs KLP Insulated Cables."
(Nov. yet published because of Ed Cardner's objections.)
r Presented a paper at the ECNE T&D Committee Meeting, September 22-23, 1983, "Polyethyl ene Insulated Cable Failure Tracking and Replacement Program at Northaast Utilities."
Prepared DER "DB Cable-Replacement Projects - Priority Ratings." (Being reviewed by W. L. Wagner.)
Scheduled to lecture at the T&D Expo (sponsored by T&D magazine) May 14, 1985, on "The Failure Syndrome in UD Cables."
UB TRANSMISSION CABLES Project Engineer during construction of the 138-kV submarine cable across Long Island Sound.
Coproduced 16mm color / sound movie, " Power Across the Sound," which was widely shown in the U.S. and abroad.
Together with Peter Judd, visited utilities and cable manufacturers in England, Norway, and France, and published report, " Transmission and Distribution Policies and Practices in England, Norway and France, with Particular Reference to Environmental Consideration and Underground Installations." This report formed the basis for NU's recommendations to the Power Facilities Evaluation Council (PFEC).
Authored NU's publication, " Underground Transmission, State of the Art."
At that time, this was the only basic treatise on U. G. Transmission in the U.S.
Organized and chaired a panel discussion, " Underground Transmission -
Search for a Policy," at the U.C. Transmission and Distribution Conference, Apell 1974, Dallas, Texas.
Led NU's participation in the many repales of the Long Island Sound cable. Was instrumental in persuading LILCO to discontinue Pirelli's role in the repair process (rigid vs flexible splice).
Alerted management to possible thermomechanical bending (TMD) problems at Northfield Mountain's 345-kV pipe type cable.
Subsequent X-rays identified problem, and repairs were made before electrical failure could occur.
Convinced NU's management that the Scovill Rock River Crossing design, as proposed in the 1975 application PFEC, was prone to TMB and downhill migration problems and should be redesigned.
Led NU's delegation to Norway to redesign the Scovill Rock River Crossing and to obtain cost estimate for it.
This design was utilized for the 1980 application to PFEC.
Testified before PFEC during the Scovill Rock hearings as NU's cable expert. This testimony reaffirmed the validity of the technical aspects and the material costs for the project as proposed.
End result was that existing ilnes remained overhead.
(Approximate savings to NU $20,000,000.)
As member of NU's Transmission Research Subcommittee, prepared section on AC Underground Systems for the Transmission Research Plan.
Gave a deposition in Cynthia Carlson vs HELCO regarding technical aspects of the case.
(Case subsequently settled out of court.)
2/28/85 l
l 5
4 H
i 1
y j
l ENCLOSURE 2 SEQUOYAH NUCLEAR PLANT CABLE PULLBYS Cable Selection criteria The total population of conduits containing safety-related circuits at SQN through which one or more cable pulls subsequent to the initial installation have occurred will be identified and screened using the criteria outlined below. The purpose of the screening process is to identify a population of conduits which have some credibic chance of having sustained conductor insulation damage because of pullbys. Within this bounded population, a further screening will be performed to create additional subpopulations consisting of 20 conduits each. The first group of 20 conduits will represent-the " worst" (i.e., highest pullby damage potential) conduits in the screened population; the second group of 20 conduits will represent the second " worst" group of cenduits and so forth. For the purposes of this evaluation, the representative worst-case conduits and cables will be considered to be any that meet, or come closest to meeting,'the following:
1.
The conduit contains a minimum of seven cables.
2.
There were at least two separate pullbys in the conduit.
3.
At least three PVC jacketed cables were present in the conduit before the final pullby.
4.
The initial pull and at least two pullbys were made before Polywater J was used as a lubricant (August 1984).
S.
The total length and number of bends between a set of adjacent pull points exceeds or comes closest to exceeding the requirements of G-38 Appendix F.
For the purposes of this evaluation, straight through condulets ("C" type) will not be considered as pull points.
6.
The conduit contains at least two condulets.
7.
Preference shall be given to those conduits which meet all of the above and any or all of the following:
a.
At 1 cast two of the pullbys made before August 1984 contained a minimum of two cables with hypalon or chlorinated polyethylene (CPE)
- jackets, b.
If it is determined through analysis that any of the cable pulls required mechanical assistance.
c.
The length of the conduit between potential lubrication points exceeds the requirements of G-38 Appendix F.
e
, d.
If the direction of pull of any of the pullbys is determined and the conduit bends were near the pull point (i.e., end of pull).
e.
If the direction of pull of any of the pullbys is determined and the first portion of the conduit was an upward or horizontal section such that lubricant could not be poured into the conduit.
Cable Test and Inspection Criteria As stated by NRC in reference 1,.the predominate concerns during cable pullbys are for saw-through and mashing of the jacket and insulation.
In all cases any potential damage would occur to the jacket first and to the insulation only if the protective jacket had been totally compromised in the area of concern. The postulated failure mode is the degradation of dielectric properties because of a reduction in insulation wall thickness, perhaps influenced by the presence of moisture. The following test program is specifically designed to identify if such a potential condition exists.
Sampling will begin by selecting a random sample of 15 conduits from the
" worst" of the subpopulations of 20 conduits. This sample size is sufficiently large to provide 95-percent confidence that no greater than five percent of the conduits in the first group contain cable damage discernable by the test method being employed, provided that no damaged circuits are detected in the sample.
1.
After having randomly selected the 15 conduits from the " worst" of the subpopulations of 20 conduits, the conduits' installed configuration shall be reviewed to determine the feasibility of introducing water into the conduit system without compromising equipment integrity or personnel safety.
It is preferred that the entire conduit be subjected to the water inj ection. However, as a minimum, the conduit segment which contains the field or factory bend which would have seen the highest pull tension during the cable pullbys must be configured to allow for such testing.
If the conduit configuration does not allow for wet testing of a segment of conduit of sufficient length or number of bends to be justified as the segment with the greatest potential for damage because of pullbys, a new conduit from the subpopulation shall be selected in its place and evaluated as above.
2.
With the cables in the dry conduit, perform a 5-minute de high-voltage test at 240 volts / mil of insulation (per IEEE 383-1974, Section 2.3.3.4).
The test will be performed between all conductors and shicids, if applicable, bundled together and the conduit. When cables of different insulation thicknesses are installed in the same conduit, those of the same nominal thickness will be bundled together and tested'with respect to all other conductors and the conduit tied together.
If the termination points of all cables in the conduit are not at the same location, several tests may be performed, at various locations, between a group of bundled conductors and all other conductors and the conduit tied together.
Based on the failure mode of concern and considering the presence of moisture along the cables in the conduit, the path of least resistance from each conductor is through the moisture to the conduit.
Testing from conductor to conductor is not jequired nor appropriate.
The voltage stress across the insulation of two conductors, even if adjacent to each otker, is approximately half of that between one of the conductors and the ground plano provided by the motsture.
. 3.
The entire length of each of the conduits or, if the entire length does not allow, the segments of the conduits justified as having the greatest potential for damage will be ' subjected to the introduction of ordinary tap water. The purpose is to ensure, as a minimum..that sufficient water is introduced so that moisture is present along the entire surface of the ~
installed cables in the conduit segment of concern.
It is not intended or expected that the cables will remain submerged, but rather that a tracking -
path will be established which would identify, through the de high potential testing, any reduction in the cable's integrity or ability to perform its intended function.
4.
While in the presence of moisture, each conductor shall be subjected to a 5-minute de high-voltage test as specified in item 2 above.
i Cable Acceptance Criteria 1.
The cables must pass the in-conduit de high-voltage tests, both dry and in t
the presence of moisture. This test is a go-no-go test, and the cables pass the test if the leakage currents stabilize at the end of the test duration (5 minutes for the wet conduit test). Although not part of the pass / fail criteria, the leakage currents at the final test voltage will be recorded for engineering information only.
2.
If the cables. fail either of the high-voltage tests, the location of the failure will be determined, and'the cables will be removed from that conduit segment and inspected to determine the cause.
If the cause is determined to be isolated and the result of other than cable pullbys, the test on that conduit shall be considered nonconclusive and a new conduit segment from the subpopulatid.1 will be selected and the tests repeated.
If damage to the conductor insulation which is determined to be due to pullbys is inadvertently discovered during the test program by means other than the in situ high-voltage testing, it also will-be. considered as a failure in this test program. The possible findings from the initial sample and the subsequent actions to be taken as a result of failures which are due to cable pullbys are as follows:
A.
If two or more conduits containing damaged conductors are observed in the initial sample, the five remaining conduits in the group will also be tested. Corrective action will be taken to replace all observed damaged conductors. The second worst group of 20 conduits will then i
be sampled using the same procedure as for the first group of conduits.
A finding of two or more conduits with damaged conductors in the second group would also lead to a 100-percent examination of the remaining circuits in that group, implementation of appropriate corrective action, and sampling examination of the third group of 20 conduits.
- m
. m
.-..m
-m
S
~
_4_
r B.
If, in the worst group of 20~ conduits, only one conduit in the random
/
sample of 15 conduits is founo to have damage to the conductors, the remaining five conduits in this group will also be examined. A finding of one or more additional conduits in this group of five will result in corrective action and require the sampling inspection of the y
second worst group of 20 conduits, etc., as described for item A.
C.
If, in the first (worst) group of 20 conduits, either no conduits with damaged conductors are found in the random sample of 15 conduits, or if one conduit with damaged conductors is found in the random 'aimple of 15 conduits but no similar conditions oc' cur in the remaining five conduits of this group, it may be concluded with 95-percent confidence (100-percent confidence if all 20 conduits in the group are examined) that no greater than five percent of the conduits in that group have n
damage to conductors' exceeding the test criterion.
e D.
Because the groups of 20 conduits each are to be sampled in the order 3
of highest to lowest potential for conductor damage due to pullbys, w<
and if for any group it can be ' demonstrated with at least' 95-percent confidence that at least 95 percent of the conduits are free of
~
conductor damage exceeding the test criteria, it can be inferred with
~'
greater than 95-percent confidence that any conduit groups ranked o
lower than the group also meet the 95-percent reliability criterion.
on that basis the sampling will be terminated at the point where the first 95/95 demonstration can be made; lower ranked groups are
. inferred also to meet the acceptance criterion.
Damaged conductors discovered during sampling will be replaced.
Those cables in each group which pass the test will be considered acceptable.
TECHNICAL BASIS: The sample size chosen for the conduit pullby sampling plan was determined using the hypergeometric likelihood density function method described by Goodman (1984). The applicable equation is:
(x' (N-x) wo p = pr(xf-m j k.n,N,)
(k, ik-x.
=
- o (n+11 N
Equation 1
+1 x=k j
i Where
.x 100 = the target confidence level (as a percent) k = the nuniber of conduits with damaged i
conductors observed in the sample o = the maximum number of conduits containing m
damaged conductors that occur in the population from which the sample is drawn and still meet the target reliability criterion n = the size of the random sample N = the size of the population from which the sample is drawn, and iI (b[
a) =
at b!(a-b)!
' For random sampling from a population of size N = 20 conduits, with k = 0 conduits in the sample exhibiting conductor damage because of the potential concern (pullbys or jamming), and mo = 1 the maximum number of conduits in the total population that can have conductor damage and still meet the target reliability criterion (95 percent = 19 of 20 conduits having no conductors with damage exceeding the measurement criterion), solution of Equation 1 for incrementally increasing n yicids first exceeding 0.95 (i.e., 95-percent confidence) for n = 15 conduits as the minimum required sample size.
REFERENCES:
- Goodman, J., 1984, " Sampling Inspection of Nuclear Power Plants." Proceedings of<the 1984 Statistical Symposium on Nation Energy Issues, Seattle, Washington, October 16-18, 1984.
NUREC/CP-0063.
/
/
l g
=
M f
b
ENCLOSURE 3 SEQUOYAH NUCLEAR PLANT CABLE JAMMING Cable Selection Criteria The total population of conduits containing safety-related circuits at SQN will be identified and screened using the criteria outlined below. The purpose of the screening process is to identify a population of conduits which have some credible chance of having sustained conductor insulation damage because of cable jamming. Within this bounded population a further screening will be performed to create additional subpopulations consisting of 20 conduits each. The first group of 20 conduits will represent the " worst" (i.e., highest potential for cable jamming damage) conduits in the screened population; the second group of 20 conduits will represent the second " worst" group of conduits and so forth.
For the purposes of this evaluation, the representative worst-case conduits and cables will be considered to be any that meet, or come closest to meeting, the following:
1.
The conduit contains three single-conductor cables (larger than No. 10 AWG) of the same conductor size.
2.
The cables are routed in a conduit such that the ratio of'the inside diameter of the conduit divided by the average diameter of one of the cables is in the range of 2.8 to 3.1.
This is in accordance with IEEE 690-1984 Section A9.2.4.4 and as stated, accounts "for tolerances in cable and conduit sizes, and for ovality in the conduit at a bend.
(An initial review indicates that SQN units 1 and 2 have approximately 46 conduits which meet the above criteria all containing low-voltage unshielded power cable with cross-linked polyethylene insulation.)
3.
The conduit contains field-or factory-fabricated bends.
4.
The total length and number of bends between a set of adjacent pull points exceed or come closest to exceeding the requirements of G-38 Appendix F.
For the purposes of this evaluation, straight through condulets ("C" type) will not be considered as pull points.
5.
There were at least two potential pulling points before the segment of conduit of highest interest (i.e., the segment which would have experienced the highest pull tension).
For the purposes of this evaluation, straight through condulets ("C" type) will not be considered as pull points.
6.
Preference shall be given to those conduits which meet all'of the above and any or all of the following:
a.
If it is determined that the conduit bends were made in the field as opposed to the factory; b.
If it is determined through analysis that the cable pull required mechanical assistance.
e
, Cable Test and Inspection Criteria Jamming is a consideration when three single conductor cables of like diameter are being pulled into a conduit. As described in IEEE Standard 690-1984, Section A9.2.4.4, ".
. Up to a ratio of 2.5 the cables are constrained into a triangular configuration. However, as the value approaches 3.0, jamming of the cables could occur and the cables would freeze in the duct causing serious cable damage.
Once jamming occurs, it can only be overcome by " brute force." Either the pulling rope or conductors break or the insulation and jacket are deformed by crushing to relieve the wedging action.
In either case the damage extends over a considerable length of the cable and is severe.
Based on the fact that the damage caused by jamming is expected to be severe and to be consistent with the test program for pullbys, TVA proposes to dry and wet test the cables in the conduit.
The following test program is specifically designed to determine whether or not such damage has occurred.
Sampling will begin by selecting a random sample of 15 conduits from the
" worst" of the subpopulations of 20 conduits. This sample size is sufficiently large to provide 95-percent confidence that no greater than five percent of the conduits in the first group contain cable damage discernable by the test metb<:4 being employed, provided that no damaged circuits are detected in the sample.
1.
After having randomly selected the 15 conduits from the " worst" of the subpopulations of 20 conduits, a field measurement of the actual cable diameter will be made to verify that the selected cable does, in fact, fall within the range specified in Cable Selection Criteria No. 2.
If they do not, another conduit from the subpopulation will be selected.
2.
After having determined that the selected conduits and cables do fall within the specified range, the conduits' installed configuration will be reviewed to determine the feasibility of introducing water into the conduit system without compromising equipment integrity or personnel safety. It is preferred that the entire conduit be subjected to the water injection. However, as a minimum, the conduit segment which contains the field or factory bend which would have seen the highest pull tension during cable pulling must be configured to allow for such testing.
If the conduit configuration does not allow for wet testing of a segment of conduit of sufficient length or number of bends to be justified as the segment with the greatest potential for damage, a new conduit will be selected in its place and evaluated as above.
3.
With the cables in the dry conduit, perform a 5-minuto de high-voltage test at 240 volts / mil of insulation (per IEEE 383-1974, Section 2.3.3.4).
The test will be performed between all conductors bundled together and the conduit. Based on the failure mode of concern and considering the presence of moisture along the cables in the conduit, the path of least resistance from each conductor is through the moisture to the conduit.
Testing from conductor to conductor is not required nor appropriate. The voltage stress across the insulation of two conductors, even if adjacent toeaghotherg sappoxggggglghagfgfgggtbetweenoneoftheconductors s
. 4.
The entire length of each of the conduits or, if the entire length does not allow, the segments of the conduits justified as having the greatest potential for damage will be subjected to the introduction of otdinary tap.
water. The purpose is to ensure, as a minimum, that sufficient water is introduced so that moisture is present along the entire surface of the installed cables in the conduit segment of concern.
It is not intended or expected that the cables will remain submerged but rather that a tracking path will be established which would identify, through the de high potential testing, any reduction in the cable's integrity or ability to perform its intended function.
5.
While in the presence of moisture, each conductor will be subjected to a 5-minute de high-voltage test as specified in item 2 above.
Cable Acceptance Criteria 1.
The cable must pass the in-conduit de high-voltage tests, both dry and in the presence of moisture. This test is a go-no-go test, and the cables pass the test if the leakage currents rLabilize at the end of the test duration (five minutes for the wet conduit test).
Although not part of the pass / fail criteria, the leakage currents at the final test voltage will be recorded for engineering information only.
2.
If the cables fail either of the high voltage tests, the l'ocation of the failure shall be determined, and the cables will be removed from that conduit segment and inspected to determine the cause.
If the cause is determined to be isolated and the result of other than cable jamming, the test on that conduit will be considered nonconclusive and a new conduit segment from the subpopulation shall be selected and the tests repeated.
If damage to the conductor insulation which is determined to be due to cable jamming is inadvertently discovered during the test program by means other than the in situ high-voltage testing, it also will be considered as a failure in this test program.
The possible findings from the initial sample and the subsequent actions to be taken as a result of failures which are due to cable jamming are as follows:
A.
If two or more conduits containing damaged conductors are observed in the initial sample, the five remaining conduits in the group will also be tested. Corrective action will be taken to replace all observed damaged conductors.
The second worst group of 20 conduits will then be sampled using the same procedure as for the first group of conduits. A finding of two or more conduits with damaged conductors in the second group would also lead to a 100-percent examination of the remaining circuits in that group, implementation of appropriate corrective action, and sampling examination of the third group of 20 conduits.
_4 B.
If, in the worst group of 20 conduits, only one conduit in the random sample of 15 conduits is found to have damage to the conductors, the remaining five conduits in this group will also be examined. A finding of one or more additional conduits in this group of five will result in corrective action and require the sampling inspection of the second worst group of 20 conduits, etc.,
as described for item A.
C.
If, in the first (worst) group of 20 conduits, either no conduits with damaged conductors are found in the random sample of 15 conduits, or if one conduit with damaged conductors is found in the random sample of 15 conduits but no similar conditions occur in the remaining five conduits of this group, it may be concluded with 95-percent confidence (100-percent confidence if all 20 conduits in the group are examined) that no greater than five percent of the conduits in that group have damage to conductors exceeding the test criterion.
D.
Because the groups of 20 conduits each are to be sampled in the order of highest to lowest potential for conductor damage because of cable jamming and if for any group it can be demonstrated with at least 95-percent confidence that at least 95 percent of the conduits are free of conductor damage exceeding the test criteria, it can be inferred with greater than 95-percent confidence that any conduit groups ranked lower than the group also meet the 95-percent reliability criterion. On that basis the sampling will be terminated at the point where the first 95/95 demonstration can be made; lower ranked groups are inferred also to meet the acceptance criterion.
Damaged conductors discovered during sampling will be replaced. Those cables in each group which pass the test will be considered acceptable.
TEJ1fNICAL BASIS:
See enclosure 2, " Cable Pullbys" l
1
r EZCLOSURE 4 SEQUOYAH NUCLEAR PLANT CABLES SUPPORTED BY 90-DEGREE CONDULETS Cable selection criteria The following criteria will be utilized to select a worst-case conduit and cables for evaluation of potential damage which could have resulted from supporting cables by 90-degree condulets. The test quantity is in accordance with NRC recommendations in references 1 and 3 identified in the letter.
During the selection process, if a single " worst case" is not obviously apparent, several representative " worst-case" conduits may be examined.
Selection of the " worst" case will then be based upon examination of the cables in the 90-degree condulets to determine the cable with the highest force being exerted by the condulet corner. For the purposes of this evaluation, a representative worst-case conduit and cables will be considered to be any that meet the following:
1.
The conduit will contain only cables with silicon rubber insulation.
(An initial review indicates that SQN unit 2 has approximately 340 conduits
'which contain only 10 CFR 50.49 cables with silicon rubber insulation.)
2.
The conduit shall have a minimum of five cables and a minimum fill of 20 percent.
(An initial review indicates that approximately 180 of the above 340 conduits contain five or more cables.
Af ter applying the minimum fill requirement to these 180, it is expected that a sufficient population will exist to select a representative worst case. The intent of the minimum fill and number of cables requirement is to obtain a conduit in which the cables lie on top of one another, thereby exerting more force on the lower cables.)
3.
The cables will be supported by a 90-degree condulet at the top of the run.
4.
The cables will have a vertical drop immediately below the 90-degree condulet which exceeds the requirements of NEC Article 300-19.
preference should be given to a conduit that exceeds the NEC Article 300-19 requirements by the greatest amount when compared with other similar installations.
Cable Test and Inspection Criteria As stated by NRC in reference 1 above, the predominate concern from supporting cables by 90-degree condulets is the potential for cutting the insulation by the corner of the condulet. The following test program is specifically designed to identify if such a potential condition exists:
With the cables in the conduit, pct cm a 5-minuto de high-voltage test at
^
240 volts / mil of insulation (per IEEE 383-1974 Section 2.3.3.4).
The test I
will be performed between each conductor and the remaining conductors in the conduit and the conduit tied togethor.
1 1
2-Cable Acceptance Criteria The cables must pass the in-conduit de high-voltage test.
This test is a go-no-go test, and the cables pass the test if the leakage currents stabilize at the end of the test duration (five minutes).
Although not part of the pass / fall criteria, the leakage currents at the final test voltage will be recorded for engineering information only.
If the cables fail the high-voltage test, the location of the failure will be determined, and the cables will be removed from that conduit segment and inspected to determine the cause.
If the cause is determined to be isolated and the result of other than supporting the cable by a 90-degree condulet, the above test will be considered nonconclusive and a new " worst-case" conduit segment will be selected and the test repeated.
If the failure is due to supporting the cable by 90-degree condulets, the extent of the problem will be evaluated and the need for further corrective action determined.
If the representative worst-case cable passes the test, it will demonstrate that damage because of supporting cables by a 90-degree condulet has not occurred. Furthermore, before restart and in accordance with NRC recommendations in references 1 and 3 above TVA will evaluate all conduits inside containment which contain 10 CFR 50.49 silicon rubber insulated cables. The cables'in those conduits, which contain a 90-degree condulet at the top and which have a vertical drop immediately below the condulet which exceeds the requirements of NEC Article 300-19, will be examined. Only those cables which are exposed to a significant strain will be resupported before restart. Those cables which could not easily be lifted off the condulet corners by hand or were found to be severely indented will be considered to be under significant strain. The support method which will be utilized is expected to be the insertion of a rubber-like pad between the cables and the inner radius of the condulet. The radius of curvature of those cables not in contact with the inner radius of the condulet la much greater, resulting in a much lower bearing pressure. This method, therefore, will eliminate the failure mode of concern identified by URC in reference 1, that is, conductor creep through the insulation to the metal condulet corner, e
e
.. ~
J
r ENCLOSURE 5 SEQUOYAH NUCLEAR PIANT CABLE TEST PROGRAM LIST OF COMMITMENTS 1.
TVA will perform a 5-minute de high-voltage test in situ dry and with moisture to ascertain the presence of damage caused by pullbys on a statistically significant sample.
Cables that fall the tests will be replaced before restart of the affected unit.
2.
TVA will perform a 5-minute de high-voltage test in situ dry and with moisture to ascertain the presence of damage caused by jamming on a statistically significant sample.
Cables that fall the tests will be replaced before restart of the affected unit.
3.
TVA will perform a 5-minute de high-voltage in situ dry test for a worst-case conduit to ascertain the presence of damage by cutting for vertical cables with long runs supported by a 90-degree condulet at the top of the run. Cables that fail the tests will be replaced before restart of the affected unit.
4.
TVA will support 10 CFR 50.49 silicon rubber insulated cables inside containment in conduits which contain a 90-degree condulet at the top and which have a vertical drop immediately below the condulet which exceeds the requirements of NEC Article 300-19.
This will be completed before restart of the affected unit.
l 1
.