ML20217H367

Requests That TU Electric Respond to Staff Comments Re Thermo-Lag Raceway Testing Proposal Concerning Thermo-Lag Fire Barriers at CP Steam Electric Station,Unit 1 to Resolve Issues Prior to Start of Any Fire Endurance Tests

ML20217H367
Person / Time
Site:	Comanche Peak
Issue date:	04/01/1998
From:	Polich T NRC (Affiliation Not Assigned)
To:	Terry C TEXAS UTILITIES ELECTRIC CO. (TU ELECTRIC)
References
TAC-M85536, NUDOCS 9804030287
Download: ML20217H367 (10)

Text

_ _ _ _ _

]

<

Mr. C. Lance Terry April 1, 1998 TV Electric Group Vice President, Nuclear Attn: Regulatory Affairs Department P. O. Box 1002 Glen Rose, TX 76043

SUBJECT:

REVIEW OF THERMO-LAG RACEWAY TESTING PROPOSAL RELATED TO THERMO-LAG FIRE BARRIERS AT COMANCHE PEAK STEAM ELECTRIC STATION - UNIT 1 (TAC NO. M85536)

Dear Mr. Terry:

By letter dated February 13,1998, Texas Utilities Electric Company (TU Electric) provided cdditionalinformation on its planned Thermo-Lag raceway testing. The NRC staff has reviewed the information and does not accept the test methodology proposed by TU Electric. The staff's comments on cable functionality and cable insulation testing issues are in Enclosure 1. Relevant input from Sandia National Laboratories is in Enclosure 2. Your contractor Omega Point Laboratory sent to the NRC in October 1992 an unsolicited proposal to perform cable insulation rasistance measurements during a fire test. This proposalis included as Enclosure 3. Other fire protection staff comments are in Enclosure 4. Enclosure 4 refers to your February 13,1998, 1stter and the enclosures to that letter.

It is requested that TU Electric respond to the staff's comments in Enclosures 1 and 4 to resolve the issues prior to the start of any % endurance tests. Please respond within 60 days from the date of this letter.

Sincerely, i

ORIGINAL SIGNED BY:

Timothy J. Polich, Project Manager Project Directorate IV-1 Division of Reactor Projects Ill/lV Office of Nuclear Reactor Regulation Docket Nos. 50-445 and 50-446

Enclosures:

As stated pg, my,y,s es O I OL L O M W BiE N IJ i h F cc w/encls: See next page DISTRIBUTION:

Docket File OGC PUBLIC GHill(4)

LMarsh ACRS TPolich (2)

TGwynn, RIV PDIV-1 r/f

\\ }

CHawes JHannon EAdensam (EGA1)

JCalvo I

Document Name: CP85536.LTR

• See p evious concurrence 0

gi i

OFC PM/PD4-1 LA/PD4-1 BC/EELB*

BC/SPLB*

PD/PDli/-1 JHanrhA NAME TPolich/vw CHa b JCalvo LMarsh DATE (///98[ 4/ l /98 3/26/98 3/30/98

'// /198 COPY YES/NO YES/NO YES/NO YES/NO YES/NO OFFICIAL RECORD COPY 9804030285 980401 PDR ADOCK 05000445 P

PDR

y -om (e

, p.'

t UNITED STATES

• I y

j NUCLEAR REGULATORY COMMISSION WASHINGTON, D.C. 2006 Hoot

'+4 * * * *

• April 1, 1998 Mr. C. Lance Terry TU Electric Group Vice President, Nuclear Attn: Regulatory Affairs Department P. O. Box 1002 Glen Rose, TX 76043

SUBJECT:

REVIEW OF THERMO-LAG RACEWAY TESTING PROPOSAL RELATED TO THERMO-LAG FIRE BARRIERS AT COMANCHE PEAK STEAM ELECTRIC STATION - UNIT 1 (TAC NO. M85536)

Dear Mr. Terry:

By letter dated February 13,1998, Texas Uw.ities Electric Company (TU Electric) provided additionalinformation on its planned Thermo-Lag raceway testing. The NRC staff has reviewed the information and does not accept the test methodology proposed by TU Electric. The staff's comments on cable functionality and cable insulation testing issues are in Enclosure 1. Relevant input from Sandia National Laboratories is in Enclosure 2. Your contractor Omega Point Laboratory sent to the NRC in October 1992 an unsolicited proposal to perform cable insulation resistance measurements during a fire test. This proposalis ir,cluded as Enclosure 3. Other fire protection staff comments are in Enclosure 4. Enclosure 4 refers to your February 13,1998, letter and the enclosures to that letter.

It is requested that TU Electric respond to the staff's comments in Enclosures 1 and 4 to resolve the issues prior to the start of any fire endurance tests. Please respond within 60 days from the j

date of this letter.

1 i

Sincerely, 0

Timothy J.

ofich, Project Manager Project Directorate IV-1 Division of Reactor Projects ill/IV Office of Nuclear Reactor Regulation Docket Nos. 50-445 and 50-446

Enclosures:

As stated cc w/encis: See next page 1

4

w e

o j

Mr. C. Lance Terry TU Electric Company Comanche Peak, Units 1 and 2 cc:

Senior Resident inspector Honorable Dale McPherson U.S. Nuclear Regulatory Commission County Judge P. O. Box 2159 P. O. Box 851 Glen Rose, TX 76403-2159 Glen Rose, TX 76043 Regional Administrator, Region IV Office of the Govemor U.S. Nuclear Regulatory Commission ATTN: John Howard, Director 611 Ryan Plaza Drive, Suite 400 Environmental and Natural '

Arlington,TX 76011 Resources Policy P. O. Box 12428 Mrs. Juanita Ellis, President Austin, TX 78711 Citizens Association for Sound Energy 1426 South Polk Arthur C. Tate, Director Dallas, TX 75224 Division of Compliance & Inspection Bureau of Radiation Control Mr. Roger D. Walker Texas Department of Health TU Electric 1100 West 49th Street Regulatory Affairs Manager Austin, TX 78756-3189 P. O. Box 1002 Glen Rose, TX 76043 Jim Calloway Public Utility Commission of Texas Texas Utilities Electric Company Electric Industry Analysis c/o Bethesda Licensing P. O. Box 13326 3 Metro Center, Suite 810 Austin, TX 78711-3326 Bethesda, MD 20814 George L. Edgar, Esq.

Morgan, Lewis & Bockius 1800 M Street, N.W.

. Washington, DC 20036-5869

M '

~

t.

g p rar

.+

UNITED STATES g

- g-3 NUCLEAR REGULATORY. COMMIS810N' WASHINGTON, D.C. 30006 0001

• • e,*

THERMO LAG RACEWAY TESTING ISSUES RELATED TO THERMO-LAG FIRE BARRIERS TEXAS UTILITIES ELECTRIC COMPANY-COMANCHE PEAK STEAM ELECTRIC STATION. UNIT 1 DOCKET NO. 50-445 1.

INTRODUCTION By letter TXX-98028 dated February 13,1998, Texas Utilities Electric (TUE) (the licensee) proposed to perform new fire barrier tests for Comanche Peak Steam Electric Station (CPSES),

Unit 1. The licantee stated that it intended to perform a " confirmatory / qualification" test of selected fire barrier configurations to resolve the two remaining Open Items pertaining to the staff's final acceptance of Thermo-Lag fire barriers for CPSES, Unit 1 as documented in the NRC -

letter from T. J. Polich to C. L. Terry dated May 22,1996. The subject licensee letter represents the latest follow up response after several meetings and telephone conference calls between the staff and licensee representatives on this matter.

The licensee is proposing to perform a fire endurance test in accordance with American Society for Testing and Materials (ASTM) standard E-119, to obtain temperature data foi specific Unit 1 3

Thermo-Lag fire barrier configurations which are installed on 2 inch conduits and 12 inch wide j

cable trays containing low cable filis. The primary objective of the subject testing is to confirm the results of a similar test performed in 1993 denoted as Test Scheme 13-2. The licensee's submittal also compared and commented on the differences between the test requirements contained in the NRC letter from S. C. Black to W. Cahill dated October 29,1992, and Generic Letter (GL) 86-10, Supplement 1.

Although the staff found that Test Scheme 13-2 was acceptable despite exceeding the fire endurance test temperature criterion, the enclosed cable fills for Test Scheme 13-2 did not bound all of the installed Unit 1 barriers on 12 inch cable trays and 2 inch conduits.

The staff review and recommendations on the subject licensee letter follows:

2.0 DISCUSSION After reviewing the licensee's submittal and related information, the staff makes the following observations and findings regarding the proposed raceway testing methods for Thermo-lag fire barriers installed at CPSES, Unit 1.

2.1 to TXX-98028. No. 4.c (Thermocouple Placement)

The licensee made the following statements in the REMARKS column regarding the thermocouple placement requirements:

c)

The primary purpose of this measurement is cable functionality. The NRC has approved a cable functionality methodology (see item NO. 9 below) W CPSES using thermocouples on the cable. The use of thermocouples on cables was found acceptable W CPSES bened on the types of cable used by CPSES and the fact that ENCLOSURE 1 J

~

only CPSES $ ant specrMc cables were being used%ualiRedin the Rn test. As a result, the NRC concluded that the CPSES testing was acceptable using this pant specrRc approach of pacing the thermocoupes on cables instead of the generic approach of placing the thermocouples on a bare copper wire.

}

Staff Feedback The licensee's assertion that the purpose for the placement of thermocouples on two AWG 8 stranded bare conductors is to demonstrate cable functionality is incorrect. As stated in GL 86-10, Supplement 1, "The bare copper wire is more responsive than cable Jackets to

{

temperature rise within the fire barrier enclosure. The temperature changes measured along the bare copper conductors provide indication of joint failure or material bum through conditions."

Therefore, thermocouple placement specifications are intended to obtain the temperature profile of the fire barrier configuration in order to assess the thermal performance of the fire barrier system consistent with the applicable fire test temperature and barrier condition criteria. It should be noted that the thermocouple placement specifications are identical regardless of the presence or absence of cables within the test specimen. Therefore, the issue of cable

functionality which applies only to the dispor.ition of deviations to the acceptance criteria is not

. directly related to the method of thermocouple placement. The staff position regarding the use of cable surface temperatures is indicated in the following paragraph from the NRC letter from S.C.

Black to W. Cahill dated October 29,1992 here forth denoted as the 10/29/92 criteria):

In the October 27,1992, meeting, wa discussed this concem [how you propose to evaluate the barrier's thermal performance) and your staff [TUE) indicated that the cable tray side rail and extemal conduit temperatures would be used to determine the temperature acceptance of the fire barrier system in addition, your staff (TUE) agreed, for cable trays, to also use the cable [ surface) thermocouple readings to supplement (emphasis added) the raceway thermocouples in assessing the thermal performance of the fire barrier system.

Consistent with the lessons leamed from the licensee's fire testing program the staff restated in GL 86-10, Supplement 1 that "The staff considers monitoring the cable temperature as the primary means of determining cable tray or raceway fire barrier performance to be nonconservative." Further clarification of the staff acceptance of the CPSES Unit 1 fire barrier configurations is detailed in Section 2.3 below.

2.2 to TXX-98028. No. 7 (Cable Insulation Tests)

. The licensee made the following statements in the REMARKS column regarding the cable insulation test requirements:

Even though TU Electric has a cable functionality methodology which has been accepted by the NRC (see NO. 9 below) and which relies on cable temperatures in lieu of meggerreadings, the use of meggertests during the fire test was evaluated.

2-k

e s

The Insulation Resistance (IR) testing required by the Ni10 has not been performedin conjunction with a Mm borderacceptance test. There are severalfactors which have contnbuted to why TU Electric does not feelit is prudent to introduce this test protocol into the CPSES testing at this time.

Of pdmary concem is TU Electric's desire to ensure the accuracy of the cable temperature data collected dunng the Rm test. The presence of a test voltage on the cable samples willpotentially eMiect the thermocouple readings of the monitored cables. The thermocouples used to monitorcable temperatures are extemely sensitive (1 millivolt of noise Induced onto the thermocouple will cause a 40* F errorin its temperature reading), the possibility that cable leakage currents orgroundloops could atFect the thermocouple readings is a signiMcant concem. Since a noise inducedinaccurate reading could cause a thermocouple to read high, orlow, the entire test methodology could become suspect.

While the possibility of noise Induced erroris TU Electnc's primary concem, it is not the only area of concem. Conducting the megger tests at the prescribed time intervals mayjeopardize the validity of the one-hour Rm test. In order to obtain IR values at 60 minutes, the test samples must be leM in the fumace while the megger tests are completed. This could result in a 70 to 60 minute fire test because of the

' baking"of the cables' samples while the 60 minute meggerreadings are being completed.

The testing laboratory performing the testing for TU Electric has stated that performing the megger test during the Rre test introduces additional safety hazards for personneland equipment. When these potentialsafety hazards are consideredin concert with the complicating technical factors discussed above, TU Elecinc has concluded it is not prudent to introduce this new aspect to the testing methodology for CPSES. TU Electric should continue to implement the cable functionality methodology allowed by the NRC letter of May 22,1996, to establish cable functionality, if required. As notedin Item No. 9 below, this CPSES methodology has been found acceptable for the Rre testing as performed by TU Electric when testing is performedin accordance with the 1992 criteria.

Staff Feedback The licensee's contention that insulation resistance (IR) measurements should not be performed by the proposed tests rests on two assertions: (1) lR test measurements pose a personnel safety hazard and (2) the IR test measurements themselves may adversely impact the thermocouple measurements taken during the fire test.

On the issue of personnel safety the staff agrees that it is the responsibility of the licensee and their testing laboratory's (i.e., Omega Point Laboratories (OPL)) management to ensure that the subject tests are conducted in a safe manner. However, we note that historically, several national testing laboratories have taken electrical measurements on cables while being subject to fire and other severe environmental conditions (i.e.,

Envirorimental Qualification tests). Attachment 2 provided citations of cable functionality tests which Sandia National Laboratories (SNL) has performed in the past. Further, OPL i

1 8

s provided the staff with an unsolicited proposal (See Attachment 3) to monitor insulation resistance values of cables enclosed within a protected raceway during an ASTM E-119 fire test. However, based upon the information provided by the licensee, it would appear that additional engineering resources would be needed to redesign the subject fire test to permit the measurement of IR values during the exposure fire.

The second issue that of the impact of IR measurements on thermocouple measurements, we note that the licensee successfully obtained thermocouple readings while simultaneously performing American Nuclear insurers (ANI) low voltage circuit continuity tests during the CPSES Unit 2 fire test program. Although the potential for noise induced error exist, we agree with the SNL assessment that the use of modest voltages can mitigate electrical noise problems in order to obtain the required IR measurements (see ).

Overall, the above objections raised by the licensee do not negate the responsibility of the fire barrier applicant to demonstrate cable functionality dg.ni.Qg the fire exposure as required by fire protection regulations. In order to address both the personnel safety and potential electrical noise concems the staff recommends that either the fire test design be modified to address the above concems or the licensee demonstrate cable functionality for undamaged cables by taking IR measurements during air ovens tests as described in GL 86-10, Supplement 1. Further clarification of the staff acceptance of the CPSES Unit 1 fire barrier configurations is detailed in Section 2.3 below.

2.3 to TXX-98028. No. 9 (Cable Functionalitv)

The licensee made the following statements in the October 29, Criteria Column regarding the cable functionality requirements:

[NRC letter of May 22,1996, from T. J. Polich to C. L Tony and enclosed safety evaluation] A test is acceptedif all of the cables within the test article meet the staff acceptance criterion of 1.0E6 0-1000ft when calculated using the SNL composite cable analysis method. This calculation is based on the actual measured cable temperatures. If the calculated cable IR values exceed the above criterion, then the effects of power cable operation are assumed to be bounded by the margin within the above criterion. Test articles that passed this initial criterion exhibited no other evidence of cable degradation indicating significant barrier failure (e.g., severe cable damage through chaning, bum through of barriermaterial).

[NRC letter of May 22,1996, from T. J. Polich to C. L Teny and enclosed safety evaluation]If the staff acceptance criterion is not met, SNL [Sandia National Laboratory] performed a single-point hot-spot analysis based on the measured hot-spot IR performance. The single hot-spot temperature for each cable in each test article was used as the worst-case peak temperature measured along the length of the cable specimen in the test article. This value corresponds to the single highest surface temperature measured along the length of a given cable during the entire test.

4

o

?

[NRC letter of May 22,1996, kom T. J. Polich to C. L Tony and enclosed safety evaluation] For those tests which failed to meet the acceptance cntena, the statt agreed to accept the SNL recommendation that cable hot-spot performance be used in the Knal determination Ibr acceptability of Mre benfer configurations. This recommendation is based on post cable-testing expedence in both EQ and Hre safety areas. This experience Indocates that the failure of an electrical cable due to high temperature exposures is likely to occur at a localized point rather than along the full length of an exposed cable... An evaluation of the cable performance at the hot-spot location during the Mre exposure tests would most accurately reRect this experience.

[NRC letter of May 22,1996, from T. J. Polich to C. L. Tony and enclosed sakty evaluation]In addition, a value of 1.0E3 (31000 M should be utilized as the sninimum acceptable IR limit for the predicted hot-spot behavior because: (1) the level of circuit degradation associated with Instrumentation, control, and power cable applications would be acceptable; and (2) a good correlation between fire exposure testing and sente accident steam exposures indicates that cable IR values in the range of the thresholdlimit provide a modest margin to thermal damage despite the uncertainty associated with the characterization of potential cable hot spots based on the test data.

Staff Feedback The licensee submittal of February 13,1998 reflects a basic misunderstanding of the staff's position regarding the use of cable functionality approaches to meet fire protection regulations and ti.a staff acceptance of specific CPSES Unit 1 fire barriers as documented in the NRC letter of May 22,1996, from T. J. Polich to C. L. Terry.

Section 50.48 of 10 CFR requires that each operating nuclear power plant have a fire protection plan that satisfies General Design Criterion (GDC) 3. GDC 3 requires that structures, systems and components important to safety be designed and located to minimize, in a manner consistent with other safety requirements, the probability and effects of fires. Fire protection features required to satisfy GDC 3 include features to ensure that one train of those systems necessary to achieve and maintain shutdown conditions be maintained free of fire damage. One means of complying with this requirement is to separate one safe shutdown train from its redundant train with a fire-rated barrier. The basic premise of the applicable staff Branch Technical Positions, GL 86-10 and the October 29,1992 criteria letter is to establish a fire rating based upon the thermal performance of the fire barrier system.

Cable functionality testing is intended to demonstrate that post-test deviations to the fire test j

acceptance criteria are acceptable for a specific fire barrior configuration. Cable functionality permits a marginali) scceptable fire barrier to be approved by the staff for use in a nuclear power plant. However, the licensee has characterized the qualitative and quantitative methods used by the staff to accept a specific fire barrier configuration on a marginal basis as a general criteria to replace the fire test acceptance criteria. In addition, the licensee's characterization of test criteria implies qualification of a fire barrier with specific cables whereas the fire protections requirements apply to the adequacy of the fire barrier alone based upon its thermal performance.

The fire barrier desig,1 should be sufficiently robust in order to account for any site specific variations with the tested fire barrier configuration.

t,.

l

~

6 4

~

On the issue of the staff review of the specific CPSES Unit 1 Fire barrier configurations as documented in the NRC letter of May 22,1996, from T. J. Polich to C. L. Terry and enclosed safety evaluation (here forth denoted as the NRC SER of May 22,1996) the following points were either overlooked or misunderstood by the licensee:

When questioned by the staff regarding the need to obtain IR measurements during the fire exposure test the licensee cited the need to perform ANI continuity tests to meet licensing commitments as the reason why IR measurements could not be taken during the CPSES Unit 2 fire tests. After the licensee discontinued the ANI continuity tests for Test Scheme 15-1 in the Unit 2 fire test program the licensee contended in its submittal dated August 8,1994, despite post-test deviations from the fire endurance acceptance criteria that "the matteris of no consequence for CPSES Unit i since such evaluations were not needed to demonstrate cable functionality" (Question 4, pp.18-20 of the NRC SER of May 22,1996). As reflected in the record, the staff has strived throughout the licensee's fire test program to obtain meaningful data in order to assess the ability of the subject fire barriers to keep cable systems free from damage during a fire.

Although the licensee recounted the quantitative evaluation steps which were taken by the staff to determine whether the test data for failed Unit 1 test configurations could demonstrate cable functionality what was overlooked was the qualitative criteria which was used by the staff to develop its findings of the NRC SER of May 22,1996.

One of the qualitative factors which supported the acceptable findings was the extensive staff review of the licensee's test program which included a vendor inspection of the test laboratory, continus1 inspection of barrier construction activities and observation of fire tests. For example, SNL recommend that their composite cable functionality analysis not be used when there is evidence that a cable was subjected to localized heating (see page 14 of the SNL Technical Evaluation Report included in the NRC SER of May 22,1996). Although Test Scheme 13-2 had a minor bum through indication, the staff determined that this indication would not invalidate the finding that Test Scheme 13 2 was acceptable despite exceeding the fire endurance test temperature criterion.

The exposed metal surface temperature was used during the staff evaluation as documented in the NRC SER of May 22,1996. Consistent with the 10/29/92 criteria and GL 86-10, Supplement 1 criteria, the staff expects to utilize the metal raceway temperatures in any future assessment of the thermal performance of a fire barrier system.

In summary, the staff agrees with the SNL position (see Attachment 2) that the analysis procedures used in the NRC SER of May 22,1996, should not be used to replace the guidance contained in GL 86-10, Supplement i nor as en attemative approach for general fire test evaluation in the future. The inferences arrived at through the use of temperature data on cable functionality provides only an indirect measure of the fire barrier's thermal performance. Further, the use of staff's quantitative and qualitative evaluation techniques as documented in the NRC SER of May 22,1996, may not provide reasonable assurance of cable functionality given the expected degraded thermal performance of the proposed test configuration. Therefore, the anticipated fire test results themselves support the need for the proposed test to obtain actual test measurements (i.e., IR measurements during the fire exposure) in order to resolve the subject Open item.

2.4 to TXX-98028. (Confirmatory Test vs. Qualification Test)

The licensee made the following statement regarding the expected results for the proposed test:

However, for the proposed test, TU Electric anticipates that although raceway surface temperatures in excess of those allowable by the acceptance methodology will be reached, acceptable temperatures should be maintained on cable surfaces. On this basis, TU Electric intends to use the same, previously accepted, inethodology to demonstrate functionality of the cables in the proposed test. Accordingly, TU Electric believes that changing the inethods used to 1) rnessure cable insulation resistance or

2) obtain temperature data within the barrier enclosures, willinvalidate the confirmatory nature of the proposed test.

Staff Feedback Based upon the licensee's submittal dated February 13,1998, the licensee proposes to conduct fire endurance testing on a test configuration 13-3 which is similar to CPSES Unit 1 Test Scheme 13-2. However, Test Scheme 13-3 is expected to demonstrate thermal performance characteristics inferior to Test Scheme 13-2. Given that Test Scheme 13-2 failed to meet the applicable fire endurance acceptance criteria it would be expected that the proposed Test Scheme 13-3 would also not meet the fire endurance acceptance criteria. Overall, the performance of fire testing on Test Scheme 13-3 differs significantly from the circumstances which existed during the CPSES Unit i fire test program by the following elements:

The absence of ANI continuity tests permits IR measurements to be taken on the specimen cables during the fire exposure. This action would readily establish whether cable functionality can be demonstrated for the subject fire barrier.

Cable functionality testing was not anticipated to be necessary prior to the

+

performance of fAe fire test for Test Scheme 13-2. The proposed fire test for Test Scheme 13 3 wG require cable functionality tests in order to resolve the outstanding

'{

Open item with the staff.

j Given the staff concems regarding the use of the SNL analysis in lieu of actual test data for the reasons stated in Section 2.3 above, the licensee's proposed test plan for Test Scheme 13-3 would still leave open questionable areas for further staff evaluation and comment.

Therefore, the rationale made in the licensee's submittal between Test Scheme 13-2 and the proposed test are insufficient to support the use of the previous test methodology and staff evaluation procedures which were used for Test Scheme 13-2.

3.0 CONCLUSION

S In summary, the staff does not endorse the test methodology proposed by the licensee for the performance of its " confirmatory test" as described in its submittal dated February 13,1998.

The staff recommends that either the fire test design be modified to take IR measurements during the performance of the proposed fire test or the licensee demonstrate cable functionality

)

for CPSES Unit 1 cables by taking IR measurements during air ovens tests as described in GL 86-10.

)

r-1 Sandia National Laboratories Letter Dated February 23,1998

SUBJECT:

Test Cable Performance Proposals

. ENCLOSURE 2

)

~-~

.,.- s2/as/ss ' EON es:40 FAI s0s.s443321 RIaK & RELIAB. CLUs'NR

-Q002 Secla N bratories PA. Sur M

r. r-um.m.m.eneser.OD February 23,1998 Ronaldo Jenkins Mail Stop 7 E 4, OWFN ElectricalEngineeringBranch Office ofNuclear Reactor Regula: ion U. S. NuclearRegulatory Commission Washington,DC 20555

Subject:

TUE Fire Test Cable Performance Proposais I have reviewed the information that you forwarded regarding the TUE arguments as to why they propose not to perform cable insulation resistance tests during the upcoming cs.T-atory fire barrier fire endurance test. The licensee arguments basically boil down to two assertions:

the insulation resistance tests pose a personnel safety hazard and the tests may adversely impact the thermocouple (TC) readings.

Frankly, these arguments are not compelling nor do they represent an adequate basis for not performing a circuit integrity measurement during the tests.

As to the issue ofpersonnel safety, SNL has routinely performed cable fbnctionality tests under a range of adverse -ysk.cr.:al conditions. This has included actual fire tests (see for example NUREG/CR-6173, Appendix A), fire environment aimat=*iaa tests (see for example NUREG/CR-5546), a wide range of equipment qualification (EQ) and severe accident tests that include immersion of energized cables in water and superheated steam exposures (see for example NUREG/CR-5772, volumes 1-3, NUREG/CR-5655, and NUREG/CR-6095), and hydrogen.

burn equipment survival tests (NUREG/CR-4763). We take the safety of our pemonnel very seriously and in all ofour experience we have never had a personnel safety incident of any kind deriving from our cable electrical integrity tests. Our experience shows that h is a relatively simple matter to set up and perform a reasonable cable. integrity test in a safe manner that will neither interrupt the normal test perforo.er.ce, nor lead to any siysf.cer.t personnel hazards. I have included an attachment that illustrates how such a system can be easily devised.

The second issue, that of the impact on TCs, can also be readily addressed. The liemaam is nominally correct that stray or noise electrical signals can impact a TC measuremanr. In reality, such problems are only rarely observed. Indeed, in our own fire and EQ tests we have never

suffered any adverse impact of a cable IR test on a TC measurement. As dimenmaad in the -

attachment, 'only very modest voltages are required to achieve the desired measurements. We

- commonly use 50 voks for instrument wires and 110-220 V for power or control cables. The currents involved are also quite small, and we typically ibaa our circuits at 0.1 1.0 A.

sospamws neemmemenwmanist 1

g we spuo a

- 02/.23/9s NON Os:43 FAI s0s $443321 RISE & RELIAB. CLUS'IER C003

.s t,.

~Jenkins 2

2'/23/98 Nwd ". if the licensee is concerned about this potential, then they can and should guard against it. This can be done using common laboratory pr.ciice. For example, if the licensee uses a 60 Hz AC power source for the cable integrity tests (by far the simplest approach given the available general facility power sources), then it is a trivial matter to include low-pass Stars in the measurement lines to eliminate any paaa+M 60 Hz noise problems. Careful grounding of the various elements would also eliminate any possible " ground loop" problems. SNL routinely uses

' such measures in many ofour test facilitias, especially those that use high energy power feeds.

One such fhcility of which I am per.c-dy fluniliar is the SNL Radiant Heat Facility. This facility has a number of 1000 MCM power cables routed in close proximity (in common 12" wide cable trays) to unshielded TC lead wires. The cables carry power loads'of 480V 34 AC at currents of up to several hundred amps. Use oflow pass filters is all that is required to mitigate any potential effects due to coupling between the cables; induced currents, or antenna effects. The licanaan proposed test presents a far less severe potential for TC error than is present in this SNL facility.

Instead of measuring cable performance, the licensee appears to be depending on the SNL temperature data analysis method used to evaluate the earlier TUE tests. I had originally proposed that method as a compromise solution given the " fait accompli" of a test set already performed without cable functionality measurements. I also recommended that it not be used in future testing to replace a direct cable performance measurement. In particular, I would draw your attention to the following passages firom SNL's letter report of 4/5/95 (see section 4.3 on pages 14 and 15 of that report):

"The failure of TUE to perform cable IR measurements during the test exposures has made the problem of determining the acc+sbility of the TUE tests much more difBcult and uncertain. SNL disagrees with TUE regarding the general ability of a test laboratory to make such measurements during the fire exposures."

"It is strongly recommended that cable IR measurements should be made during the fire exposure portion of the fire endurance testing protocol if cable ibactionality is to be used as the basis for test acceptance. While periodic IR measurements are preferred, at the very least a single-point measurement should be made at the height 'of the fire exposure hafats the termination of the flame exposure and before removal of the test article from the test cell. In the case ofTUE, temperatures have been used to estimate cable performance factors. However, SNL does not recommend that these analysis procedures should be considered an acceptable alternative for general test evaluation in the fbture. Reliance on calculations of the type performed here as the basis for test evaluation cannot bejustified

_ given the relative case with which cable IR measurements can be made during these fire tests."

I stand by this earlier recommervlatian That is, I recommend that (1) TUE can indeed make a cable fbnctionality performance measurement safely and without cce.p.emising any of the other test data, and (2) there is no compelling reason to rely on an uncertain method of analysis in lieu

? of a measurement that can be made directly. It is my recommendation that the USNRC insist that the licensee comply with the requirement to perfonn a cable A=-1---% measumment as a part of the fire test ifibactionality is to be used as an acc.pi.ees criteria.

m..i m

,,,02/p /ss mon 09:46 FAI 505 6443321 RISK & RELIAB. CLUSTER 0004

\\

Jes* ins 3

2/23/98 Yyou have any further questions, or need copies of any of the references I cited above, please let me know.

i i

i

' Sincerely, j

1. */

teven P.Nowien Accident and Consequence Analysis Department 6413 Copy to:

6413 S. P. Nowien j

l.

orns/es MON Os:4s FAI s0s s443321 RISE 6 REM AB. C W STER 000s Jenkins 4

2/23/98 Mah Common Practice in Cable Electrical Integrity Measurement 4

I IntrM~*iaa The objective of an electrical integrity test is very simply to measure the cable insulation resistance (IR). For a multi-conductor cable without ground the measurement is made from conductor-to-cWm. For either a multi-conductor cable with ground or a single conductor cable, a I

cWor-to ground test would be appropriate. In practice, even for a multi conductor cable the

.j test design will typicdy ground one of the conductors while energizing the others, and a conductor to system ground measurement of]R. is made. (If the cable includes a shield or " drain" conductor, this would be the grounded conductor).,

A relatively simple system using a modest voltage potential is d that is required. A single phase power source is typiedy sufficient although more sophisticated systems using three-phase power sources have also been used. Ultimately, the objectives can be readdy achieved using a very simple combination of a voltage source, ballast (or load) resistors, fuses, and a voltage measuring data logger. Ultimately, the tests can be performed using either a DC or AC power source. Each has certain advantages and disadvantages, but ultimately, either can provide the data needed.

2 The Voltage Source The energizing voltage need not be excessively high. Indeed, for an instrumentation cable, voltages as low as 50 V can be used effectively and are considered representative of the actual in-service voltage applied to such cables. For power or control cables, even a 110 V or 220 V

)

source taken from a wd outlet would be sufficient. The " trick" is to ensure that your l

measurement system is sensitive enough to measure the very low leakage currents one should l

expect when the cable insulation resistance approaches the threshold limit. This is dia-aad J

further below.

3 Load Current

]

It is ue a y to impose a load current on the cables in addition to a voltage potential provided

~ hat the chosen failure threshold has sufficient margin to allow for the cable self-heating effects t

I anticipated in a real application. As has been noted in previous SNL reviews, the criteria set forth by the USNRC in Suzanne Blacks letter of 10/29/92 would be appropriate failure thresholds to apply under test conditions with no ampacity load. -

4 Fusing One critical safety measure is to include fuses in each conductor power feed line. D-a adia: on.

I the system design, we typicdy use fuses rated at 0.1-1.0A in our systems. When designed properly, the maximum current flow will be limited to a readily calculated, and readily manipulated, value. Ihis value will typicdy be less than 1 A. The role of the fbse is to ensure that should a cable fault be realized, the power to the faulted conductors would be quickly

n.:.. n.- ~ _

., c2f w es.aoN os:4s FAI 50s s443321 RISK a RELIAB. CLUS'IER Ooos Jenkins

~

5-2/23/98 isolated. This also pigvides a second level ofprdW to the line power source to prevent a circuit breaker trip or other source overload.

5 The Ground Plane A second critical safety measure is to ensure that a proper common ground plane is established for the entire system. In the case of TUE and potential safety concerns, the conduit itselfis an ideal ultimate ground plane for the cable integrity tests. Hence, measures should be taken to ensure that the conduit is properly connected to the same ground plane as both the data logging system e d the power supply system (i.e., the utility service ground plane).

For the actual test cablAs, it is appropriate to consider the nature ofthe cable in establishing the need for a supplemental " local" ground plane. For example, instrument cables usually include a drain line or shield that would typically be grounded in practice. Hence, the drain or shield should also be grounded in the test using the same common' system ground. For multi-conductor cables without a shield or drain, one will typically ground one of the conductors and energize the others.

In this way, a conductor-to<onductor IR is obtained.

6 End Effects It is known that when a cable terminates within the fire exposure space, there are unique "and effects" that could impact the test results. In the proposed TUE tests this should not be an issue at all H9=a the cables pass entirely through the fire exposure segment of the test article with both ends isolated outside the exposure environment. Hence, the only measures that needed are to protect the un-connected end ofeach conductor to prevent inadvertent shorts or personnel exposures.

7 Measurement System Design The general design of a practical measuring system is relatively simple. Recall that the objective is to measure insulation resistance. In practice, this is generally accomplished by measuring leakage currents rather than by measuring resistance directly Leakage current can, in turn, be measured either directly, or more simply, by measuring the voltage drop across a ballast (or load) resistor placed in series _with the conductor. The voltage drop approach is generally easier because most common data loggers can easily and accurately measure a series ofvery small voltage drops whereas the ~ measurement ofvery small current flow's directly may require more advanced equipment. Hence, it is common to use a precision ballast resistor to convert the current flow to a measurable vohage signal based on the very simple V=IR relationship.

The key to the measurement system design is to ensure that the system has sufficient sensitivity to detect IR degradation well above the damage threshold. As will be demonstrated below, this is

. determined largely by the sensitivity of the available voltage measuring device. Given modern

.' data logging systems voltage measurements in the pV range can be readily achieved, and this is more than sufficient. To illustrate the design process wedder the following==ala.

v

=

,n

. y c /.23/esL mon os:s2 FAI 805 8443321 RISK o RELIAB. CLUSTER 0007

~ Jenkins 6-2/23/98 The first step is to determine the insulation resistaar= measured in the sample cable that would correspond to the normalized failure threnhald. To illustrate we will assume a normalized damage threshold of 10' O/1000'(normalized to a standard cable length of 1000 feet), and that the length

~ f cable being exposed in the Ere test is 20 feet (this would be the length of cable actually inside o

the test fbrnace and_ would exclude any additional cable lead located outside the test furnace, for example, above the test article top decking). Given this, the damage threshold to be inansured in

the actualcable sample would be:

Rm=10'x1000/20=5x10'O.

j If the energizing voltage is 50 V, then this implies that the corresponding leakage current at the damage threshold would be:

6,,=V/I6,,=50V/5 x 10'O=1 x10d A=1 pA.

Note that here we have ignored the role of the ballast resistor. Indeed, the next step is to size the ballast resistor. The objective is to choose a resistance value that is both very small in comparison to the anticipated cable IR (hence, we ignored it above), and that gives us a reasonable range for the vohage measurement. Ifwe use a 100 O ballast resistor (a common value), then the corresponding voltage drop for the ballast resistor at the failure threshold would be:

V

=I,,,16,,=(1x10-'A)x(1000)=1x10-* V=0.1 mV.

This is a readily measured voltage value. Indeed, one can readily measure voltages well below the pV range, and hence,2-3 orders ofmagnitude sensitivity in advance ofresching the damage threshold can be readily achieved.

Funher consideration must then be given to the fusing of the system. Ifwe actually see a cable dead short during the test, then the maximum current flow would be limited by the size of the ballast resistor,in this case,1000. Therefore the maximum current would be:

I =50V/1000-0.5 A.

Hence, the fbses can be set to a very low limit, such as 0.1 A for safety purposes, and yet the objectives of the test can be readily achieved. This last calculation also provides criteria for the selection of both the ballast resistor and any tiersfor...,. required to achieve the desired voltage.

That is, the maximum power dissipated in system at dead short conditions, even ignoring the limit imposed by the fbee,is:

P =1,*V=0.5Ax50V=25 W..

A precision resistor meeting these requirements can be readily obtained fiom any electronic parts supplier.

Note that other design factors should also be included. It is prudent to provide source cut-off switches to allow for inal**iaa of the power source once a ibutt has been realized or at the end of

02/,23/ss NON 09:54 FAI 50s 4443321 RISK a RELIAB.; CLUSTER

@ 00s

.j

- Jenkins 7

2/23/98 the test. Procedures should be established to ensure that at the er.d of the test the power source is ibily isolated, and disconnected from the test article as well.

Given this setup, there is really no reason not to run a continuos sample. However, procedures can also be established to periodically energize the system, allow a short settling time (on the order of a minute or so), and to make the measurements. The system could then be de energized until the next measurement point. The disadvantage of a periodic measurement is that if a severe fault occurs, there may be no advance warning of the fault. Rather, the system may simply blow a fuse as soon as poweris applied. This behavior should be anticipated With a continuos measurement, the system will provide warning ofan impending fault. Also note that even if continuously energized, the current flow rates will be so low that the cables will not add any signi6 cant heat to the system and hence will not alter the test results from a thermal perspective.

8 Data Analysis The data analysis at this point is quite trivial. Given the measured voltage drop across the ballast resistor, this is converted to an equivalent leakage current:

I%=VA This is in turn converted to the equivalent exposed cable composite IR as follows:

IR=(V

-V A.

In reality, Vballast Vsource, hence, this equation r' educes to:

IRw=VA.

The vrJue must then be "nor=U=i" to the equivalent resistance of a standard cable length (e.g.,

ohms per 1000 feet)as follows:

I eded-i This value can then be compared to the failure threshold directly to assess cable and barrier

)

performance.

9 A Sample Schematic Included in the Sgure below is a simple schematic of a system to achieve the design discussed above. The illustrated case assumes a 2-conductor instrument cable as the test object. Similar systems have been used in many different SNL test programs with great success. For a more detailed schematic refer, for example, to page 16 ofNUREG/CR-5655. All of the required u

s are readily available, and can likely be pulled "off the shelf" firom any major test e

laboratory. In our case, we used Hewlett-Packard Standard 3497 data loggers with 20 channel A-to-D voltage cards indalled.

~---~:--a--~----

- p,02,823/98 MON 09:57 FAI 505 4443321 RISK & RELIAB. CLUSTER C000 Jenkins 2/23/98 2/ C Ca bl e l

w 0

_k I F

• E x y o_ s__u_ r e_____

me a s u r e d vollage y

s baast (c i r'c u i i resisior fuse j l' V,ourc, A simple schematic for an insulation resistance test of a 2-conductor cable. Ih

Omega Point Laboratories Unsolicited Proposal No. P921020-02 Investigation of the High-Temperature, Short Time Behavior of Electrical Cables .g. ENCLOSURE 3 o

L Pl0i N Ni U d Q M1BiGl M Bq dGh@b 4 W%W l 9 m a.,a,sm+ w?m~ nem, w,y; e

gic

n.;&;gqm.<w

n wa . s,2 0%., se WeW:ma% YpgMg 3 4~ e .3:,9.r a,m 4 .Sm..w s. 4 ay m w m. gww,*p$ spegM 9MM INVESTIGATION OF THE HIGH-jv$indt} 'M'APN TEMPERATURE, SHORT TIME fif,MM GG m kycawr t Tpr%E ' BEHAVIOR OF ELECTRICAL

jrgg CABLES

.~5NP& WN$.p.? - . mz. Q.. ~w .w ,n. . en.- ~ -hhk,l e 5{, Unsolicited Proposal No. P921020 02 J we.... r.3s

s.. m..e,e.

w n ~? f) Y ?l',!{- 5.'a .,y-* $ f {. ~ j' A N A G F4 T N dr d.% y @ W, N 7 T t M @ 2 h [?@ S7 .Tk.M'hbD[IUh.h M W[9 & M?ttia&g@ P s m p @ s .M# mfMFSp S#mV&p!9%% fa F! $$M w%tM a99 A M]QxkWr ft hb ft4 X f5d5%r,9 u.y% %m& @ @f kev? f4kk%%??n& ww n -i h ??qhn's % ?N$ m:&-f i'

5 EN 9

B W.w.dyte@.J4.M}f,2 Ef mfe:Ja sq q v~!i V W $6 O $ @ hk.. 9SS%:d b Wen m% ~a e $$ $ M k %m t. K Jfg S:wr h. ;.s $c, v ag

j. 7 Wi'dL8EEIh[$y[:%m

%54i#N

fpf, d!7 M4

~ i Jh&nw.seng:rm$ep:w., wwww.k n w $$an,%;*q7sf& 1La Mk W Ma#nFikMQ Xh.,y&gt.xi,f5' fig 4%%s m MWf.. '%$g6% Mfdb67..oS3.%c,p%e& M 3E h Mf5h,p,c s.i.,8 w %am$dYfi,NN

1. kaM,dPA-

$@ M8f%we#m%e%ie25@W&taMn$aW add mp.W c ~ w% p &s p a q ' m a w a p u g Wgfn8&ffhM[EbiinV d M W $ l g ! 3 R i g ' f M e, e@#. h s & w# EMEWA1:ig sw#@p. Me M s tiMP?c wea wxmpe-An ng%h;En&w: ,+.:. -Mq&4 s = .%M.w* S vskW#GMMbiW. ~ * ~ ,~ ' SG s & g.. me M &@ W, 4 v ; g. ' ~.a .v % p m g&u",ay,,G. A*&., &w#4)sQwAvlr*i %1f

• 4*:

.t M r,.W yLAW g w ;,2 r,{s. y,,1 x,% e J., y,, ,R y'ry Qf geY.? w e 'NYY N h Q+ $ $ta=4.W ' kC,Lp,. y 2ft h $f d A n f h?'54 W .Odh% #6M N W 3' GRhEMPJ1R

o .3-(:. 1 pn a,. : O,...', , -. T. - i mw .i}.u :4 D ' g. - 4 *n, n4 A# 4 v 6. a,'s. '. ' ' * % / q~.,, 0 - s ".. y..[1 ~ c.- .g < y.,;: v. n.;..

n..

.. w e,. ::. - : .. a #, ~ .. ~ f yf4 f-', ,,.,,.a.. pg3 - .'.J., y,.;, q l' 3 .,. d, - '. 'g { C + 3 k, ; r.<- ~ -], . ' A %, *... .'4 . + ...) fr.,.. ..'.23t: j.,p,,. 4.,. t.,.' 3. -. t ;. .,-%, p,. p., ,s.,,.. g ..,..r. a .. +.. ., 7,. g a. ,..,,,,..w., og.f.~; ..3 4* ~ .' ' f */ '.,, L.'4.. l.[O, ;',--- Q i.'.a;.), l k. i ', 44 ( y w T.e,g., [Y. I *i p ] y j,,7./",]s 14 .'? .. ;A.,f,f ?, v *. t; i ;t,,.. - *..,~ ..,,.~-. A .J. + c4 -( ~ t g .s 3 .og ~.,%g.m. s -s -;

• s-

. h 4hp. QF. ;M.sQ,.,.M ha;sm V m Wp.~s.1;.Wh.wi. p. fv' t .s. s. .' { ..., M. y, 7.a. - ' ...e v .s,g,. .-. y.,-...,' e .,...,,~g...' . N fh ( ..o* .. m.* $ 4. A.. f,,n .'s n.. .x..: .. 7. w*... g. &s.-f, .c. ~ f .Y.,- %,.,., 4.,,,.%.. 4 5 IW !! F ' 6W D G : ' n 1 ? O q. W.,. g

• 1W.

.,.4. y .- g .~

6. %.7 2 TU!

ATLU,3MM ' M, wM' W.. d W,. -.. BE i AV DR OF ELECTR CAL gy; n.,,J W ' CABLES Y.< b W. 9: L aw. ' p'.. y,' ~ - d;- Unsolicited ProposalNo. P921020-M q 4; Q. M.. wwu.+.., \\e v, 4 k, \\m..A., t b -. _ _ _ a.m m: D,.,. E'h? h h~ w. m., a m u[, g u %. 0 f - Q N f't': Y ?r.c ?v. w;; eA.- B

• l -

mn 5 .r ~ __ m

• M c h --

e a- ' M.y? 3,,*:. p.M.;,.. r..it s rg.,,g.f L.V'.3 u ~ .. n, +3..., %, 3. " ~ -r.<p. -

.- n-uM ; m..,..., - + c., /

, J;..:. ... y %...,s;.> p A...s'h +,f.9 J. ' , c n,.. rf *1. m = .m lf.*L.,y, #, 7, 'c At a - n 1Y r,3 y - ; - 7 A,.r..y,rj ?. j .,3 + v. F.'. . 3 '.f;3.v.s t,; v-

J.

v u,? yn *y ) ,',.. :q n p). _ ' y . e *;> Lqp.a y+ s3-3(. \\ -,. v, v. y; >y c

.q _,

m.. ' sx. t .4, .;2 _g ;.p 2 .s., e.. ,^ : t- /\\ t ; e., -* r y; > n y..

. ia,r. y,. % < -

,.I '. e ' p. t ^ ?....,... 4 c : 4 >...;,v

n..m 4 v:

~ 1....;. .c ::., *. p '. >;y,. a

.. ^..

2 s. 'l ' $,. JT d-1 g. ,ws v Y f.. 9.,h.. c.1, f:. s, N l '. n? '. S".'! f '. ~, E J. N ~. 4. f W S.Y Of ,,.. m. .a 1 ' c,s~ '. *' o ; I^

• s

+ u.a y w' ' (spg' ^ LA%' y;' 'ff.^3.m. Ay,lf. %. y.. ..,',u.,;q,; :+ ~.'. 3....U.y :.g. -\\* p' '4 %p~ s' ' ' g' y f s. ~ ~ A n - 9 .. /. L 'u*_ . y :,' y 7 y n w % f y ~? w&. U 's H A +. f-M n .gn. w e ; F.q :, W '- u .-? %;w y 3 g m y,.y,ph b.maA~y a ; r.. 3.)2 4 w. qQ 23.3y J -.. i ;,. AIV . i - .e, +. .x .n l l

• l*> ~ f W :.!; j',; {k

[ ' l :. ';~ ^ 'M ' ' " {, - N ~ . t.,... A; h ' i N ~ 4<..... t-1, t .a.- j.. ;.- y, n ~,.. &,.,.,,, c g y ?. " *l.,**1 b $ "k-6 h. [...v .3 ~d \\ -,,r p,..- , f.g,,,,,,g9 p.pg c; s. , Q-Mgy) L' r.E .c l } l '"y ~ 3 :; ; ?g[' . ~. l l ~f4c .ve..g 1 g ,.. a I b $h A 4-. o. p.,,, s p r 1 ;, e,- x N. 3 .,gg. ( ,6 - ~. -.

i h, ,* ' ' Nuclear Basalatory th==ia asolicited Proposal Ns. P921020 02 October 20,1992 Page11 ~ l i u sru er I A method ofdetermining the electricalperformance of cables exposed to various exposures ofelevated temperatures over time periods of 60 and 180 minutes is presented. The cables are eve!!>eM for loss ofresistance i ' between each individual conductor and all others as well as electrical ground. The electrical resistance is measured at the rated AC voltage of 'the cable being evaluated. g This methodology has been developed to evaluate the ability of speciSc electrical cables to survive the conditions within an electrical raceway E during an externalBre exposure. This Unsolicited Technical Proposal has been presented as a possible l solution to problems and questions ofinterest to the U.S. nuclear 'I industry at this time. A Pricing Proposal will be presented upon request. I All inventions, designs and methodologies described in 'this document remain the property of Omega Point Laboratories, Inc. until acceptance \\ l of this proposal. E E /Of?//93. - ,x= ~ Deggary N. Priest Date E President E E E ,,? %, ,5 g L

3. ,..,,,',,. _ _ m. ~ ~ .f -, m- ~. - .f m i Unsolicited Proposal No. P921020 02 October 20,1992 Nuclear Regulatory Commission Page 1 i L INTRODUCTION l Electrical cables used to instrument and control critical equipment must continue to operate during an external fire exposure of a specified intensity and duration. To meet this requirement, various designs of" fire protective envelopes" are utilized. l The intent of these systems is to maintain the temperature within the raceway i below levels at which the cables lose ft.netionality. B The fire exposure tests are normally accomplished by constructing a specified En raceway configuration and installing cables instrumented with thermocouples within the raceway. Following the installation and curing of the fire protective envelope, the entire assembly is exposed to the time-temperature exposure i described in ASTM E119 Fire Testa of Building Construction and Materials for the amount of time for which rating is desired; normally 60 or 180 minutes. During the fire exposure period, the temperatures within the cable raceway are monitored at sufficient locations to allow an analysis of the temperature rise within j the system. While attempts have been made to limit these temperatures to i specified values, there always exists the possibility that whatever limita are chosen, cables may be used which will not survive those tempertures (for instance certain fiber optic cables will fog at relatively low temperatures). Alternatively, many i cables can continue to perform their function at temperatures signiScantly greater than the current E119 criteria of 250 F (139*C) increase above ambient for the average value and a 30% higher increase for any one individual point. I A more logical and technically velid method would be to evaluate a given fire protective envelope system so that the maximum internal temperature rise is known for a given external fire exposure period and a given thermal mass loading. That system then, would be rated for time and temperature: for instance; a system whose internal temperature rise at the end of a 60 minute external E119 time-temperature exposure was 204'C would be rated for 204*C rise at 60 minutes. While the development of the test requirements and criteria for fire protective envelopes falls outside the s:: ope of this proposal, information concerning the i maximum operating temperatures which individual cables can withstand is vital to a correct evaluation of any given system, and a method of determining those values follows. II. TECHNICAL APPROACH j A small scale test method is proposed, which allows electrical cables to be monitored 5 for high voltage resistance between all conductors within the cable and between all conductors and a steel raceway mock-up. The method involves instrumenting the J cable conductors so that the exitation voltage can be varied to meet the maximum R voltage rating of that specific cable and then exposing a given length of the cable to J

Se d M -. desdame and 1? a 's semesimed in abde desumussus ese nessuperop af oneses Pedas I : " _. - A Ime. (OPfJ med assy mas 6e neweed er espied audaeus de areasse, " " eropL d Unsolicited Proposal No. P921020 02 October 20,1992 Nuclear Regulatory Commission Page 2 a uniform, ramping temperature for a given period of time. The intent of this method is to enable the determination of the maximum final temperature which that specific cable can endure at the end of both one and three hour periods. In order to maintain a uniform exposure, the approach suggested is to install the d instrumented cable in a horizontal tube furnace and equilibrate the temperature to 90'C prior to the start of the test. The furnace will be controlled by a programmable temperature controller, so that the temperature can then be increased linearly to a predetermined maximum temperature at the end of the exposure period. A ramped exposure has been chosen as the most representative of the conditions within a protected cable raceway for the following reasons: During an external fire exposure, the temperatures within a protected raceway system tend to follow the lower section of a sigmoidal curve as the heat penetrates the envelope, often with a plateau, caused by any number of heat activation reactions (such as the endothermic driving off or chemically bound water), RACEWAY INTERIOR TEMPERATURE Classification Temperature Rise 9 i w E 5 3 1 2 4 4 e il e ma E-* v 7 Time l remaining below a straight line drawn between the beginning and ending temperatures on the plot. By equilibrating the cable temperature at it's maximum operating temperature (90'C) and then linearly ramping the furnace temperature to reach the desired level at the end of the preset time period, the researcher can, through a series of trials, determine the maximum temperature rise exposure for which the cable can retain its functionality over that time period. d n

Se " ds4pme asedassehedeansessestmedin anosdesmessed ase shepewpere eronsey Pedas s N Ann. deste amusassy uns to mammedeer asyded uds6ead she aardsson p__ '" aros Unsolidted ProposalNo. P92102002 October 20,1992 I Nuclear Regulatory Com=1uton Page 3 i CABLE TESTTEMPERATURE , ClassincatJon Temperatun Rise + 90*C ',,.e 6 o,.*' li e e4 8. 'E 8g n-w v Time The maximum temperature rise at which that particular cable would still function would then be it's classification temperature rise for that period of time. It is anticipated that the classification temperature rise for a one hour time period would be somewhat higher than that for a three hour time period. I Since protective envelopes are best classified in terms of the maximum temperature rise within the raceway, it is proposed that cables be rated by their maximum I temperature rise in this proposed test. For instance, a cable which will retain its functionality for a maximum 60 minute temperature rise classification of 200*C would be tested by ramping the temperature from 90 C to 290 C. Most raceways 8 involved in an external fire would be expected to be operating at a temperature somewhat below 90'C, but the selection of this initial temperature was made as " worst case." A square wave temperature exposure was not chosen, since it does not model the actual temperatures within the envelope. The ramped exposure was determined to be more representative, while still maintaining a degree of confidence in the results. 8 The starting temperature of 90 C was chosen since that is normally the marimum allowable temperature due to resistance heating and since some temperature above ambient should be chosen so that all tests begin at the same point. A relatively 8 small volume tube furnace was selected to ensure uniform temperatures across the entire cable specimen and to minimuze convective beating. h The end point criteria for the performance of the electrical cable will be determined as the failure of the resistance between any conductor and all other conductors and a small steel tray mock-up, and will be conducted by energizing the conductors with a 60 cycle high voltage set to the maximum voltage for which that cable is rated. An d ,8 o

. _. ~ ., ~... " -, - ~.~- - ~ -. - - m - e ay eens.dmed is.hde desen d are shepropersy.f omneys Pedas " -. d deme and ansehedess 2ne ' - pd.d.sd.h .he s.m -.. . fem Unsolicited Proposal No. P921020-02 October 20,~ 1992 I Nuclear Regulatory Commission Page 4 I allowable amount of current leakage has not yet been determined, and will depend on the acceptable level of resistance to be maintained. We will accept some l guidance on thatissue. III. PROCEDURE For heating the cable along a prescribed temperature ramp, a suitable horizontal tube furnace will be utilized, which can create demonstrably uniform temperatures along a 24 inch length of cable. The furnace will be of approximately 4" internal I diameter, and will be fitted with a small ladder-back tray mockup, consisting of two parallel lengths of 1/4" steel rod, with 2-1/2" long pieces of similar rod welded between them and spaced 2" o.c.: g , long enough to extend outside the heated area on both sides, 8 8 Simulated Ladderback Tray l This small tray, whose purpose will be to both support the cable during the test and to serve as a ground during resistance tests, will be laid down the longitudinal centerline of the tube furnace, and positioned at the bottom (with small pieces of I calcium silicate insulation [or equivalent) to insulate it thermally and electrically from the furnace). Prior to a test, the instrumented cable will be tied to every other tray rung, to keep it in position. The exact method of attachment will have to be i determined during the development of the apparatus and methodology. 1 i 8 i i fg d

t. me " vatpasrandumssendmassy somedened a sete desen(naat a4 dep,eparey of omegre 7Was 1 m u neu am ee..,an.,ewurans,ma eare em.s._ .cm. I Unsolicited Proposal No. P921020-02 October 20,1992 j I Nuclear Regulatory Commission Page 5 I e Furnace A Y k. m -. .i." E $M...

• g g{h h
• k Furnace Interior i;wn%

8 wn:n ~nr k [._, "M%y s. n <- p 4"s t C Hinge Mhd fh. h--

• ~

+? Test Cable j d)MMQ I %$$$W$%a ^ e M afad E n 5I?n... s*At" " "a w a[Q V gi: W y yp j 1%g%w M"W E%e l gT Small Tray EMGMb@h! $ MMS $@* NMI! l mawk CROSS-SECTION OF TEST FURNACE g The furnace will be electrically heated, with as many separate zones as is necessary 8 to ensure even heating along a 24" length of cable. The ends of the furnace where the cable exits shall be closed with a suitable material, probably ceramic fiber insulation. The determination of functionality will be accomplished by connecting each conductor in the test cable to a separate relay, which will have two switch d conditions: direct connection to the high voltage AC power supply or direct connection to the neutral (or chassis ground) of the system. The relays will be under computer control and only one of them will be in the High Voltage AC l position at any time. The computer will thus apply high voltage to one conductor and ground out all others and then watch for an appropriate time for current flow in the circuit (as indicated by a rectified voltage across the load resistor). The l computer will then return the first conductor to a ground state, energize the second j conductor and then once again watch for voltage, indicating current flow. In that manner the system will be continuously polled throughout the test. The relay g system will be constructed initially with 25 channels, and will be upgraded as

UnsoHdted Propeast No.P99103002 October 20,1992 Nuclear Regulasory Co==taata= Page 6 required to evaluate cables with higher numbers of conductors. A schematic of the proposed high voltage resistance monitoring circuit has been included in the appendix attached to this document. Once a minimum resistance to ground has been established as failure criteria, the load resistor and the fuse will be sized appropriately. The computerized data acquisition and control system to be utilized will be a Macintosh IIsi computer (Apple Computer Co.) =aneted to a John Fluke Company, Model HELIOS I Computer Front End. The computer will be programmed in g Microsoft QUICK BASIC so that the operator is given the option of selecting the number of channels to rotate through for exitation and monitoring. If, due to time 3 restraints in the sampling rate, the cycling period becomes to long, the computer B. will be utilized to control the switch positions only, with the voltage output from the circuit being connetml to a strip chart recorder for continuous monitoring. The furnace chosen for this project will be one that is programmable across the range of temperatures and rise rates required for these evaluations. IV. EVALUATION OF RESULTS As thermal energy enters a raceway (after passing through the protective envelope), it translates into temperature as a function of the thermal mass within the envelope. Within a moderate temperature range a given amount of thermal energy will raise the thermal mass a specific number of thermal degrees. For this reason, l most fire resistance test criteria are based on a limited increase in temperature j above the starting temperatures for both the average and individual high temperatures. For that reason, the important information to be gained by performing a fire resistance test on a protective envelope system is the amount of temperature increase within the system (of course, the amount of thermal mass in that system must be considered also). The temperature increase that the cables must be able to withstand to be qualified for use in a specific fire protective envelope system will dictate the minimum temperature rise which must be sustained in the cable qualification test. The cable g should retain functionality when the temperature is ramped from 90*C to 90'C plus the temperature increase in the protective envelope. EXAMPLE: Given that a specific protective envelope system reached a maximum internal temperature of 181'C above it's initial starting temperature over a 60 minute fire exposure test, then, in order to qualify for use in this system, a cable would have to be capable of functioning from 90'C ramping up to 271*C at the end of a j 60 minute cable qualification test.8

=. ne ! ":- :, deotene and '2 eentadaad in skie 2 " are sheproperty ofDeenre reint ! ' :--:_ ?.=,Ine. (OPEJ and nmes not be us%ed er enpied miehene she wronen permeieeten ofCPL Unsolicited Proposal No. P921020 02 October 20,1992 I Nudear Regulatory Commission Page 7 I l Once a large body of cable types have been classified for temperature tolerance for E 60 and 180 minute testa,it will be possible to determine the applicability of each for use in any previously tested envelope system. g V. CONCLUSIONS This proposal is to assemble and " debug" the basic apparatus presented for the determination of high temperature /short time tolerances of different electric cables, ) E and then to perform the evaluation on a variety of cable types from various I manufacturers. No efrort has been made to classify or identify the possible candidate cables, since that will be left to the sponsor. A Pricing Proposal will be I presented upon request, following a determination of the approximate number of cable types to be evaluated and the maximum number of conductors to be evaluated in a single cable. l E 1 Upon award of a contract to perform this project, the Laboratory will prepare a i detailed Test Plan, which will identify precisely all equipment and methods to be l E utilized, and will not proceed until this Plan has been approved by the sponsor. If I requested by the sponsor, Omega Point Laboratories will publish and maintain this test method as a public domain test standard for general use. Following the l completion of a test program, the assembled test apparatus will be maintained at this laboratory for use for classification of additional cables as requested by commercial clients. E Omega Point Laboratories is willing to make alterations to this proposal as requested by a potential sponsor and will gladly discuss any details pertinent to the g project. E E e 4 e E

3 ma sa sg, g,,,,,,e a s y ' _; 4 ,,g Q i" ** R donde en a n d ' ' ' *:h 10PfJ and may m # 6e memmed er p.g adm g g, m,g, #0PL g e. ~- 8' Unsolicited Proposal No. P921020-02 October 20,1992 Nuclear Regulatory Commission Page 8 8 8 8 8 8 8 8 APPENDIX ) 8 Q G G G G d ^ g

%.. s.,* _~* ' ~ ---- = m";

1. 12464W =- - - _ a446 -a.,i.. -

.. 3 a,, u g. W'seGiy, _a w .p~.n: mu -7 e-fairiM haa.s3pu an*d ..y utacedovocpkenehout ween poindon at oN not b .keenamn h.- 1 ~. control t....-.-.-.-.-.-.-....-....-.-..-...-.....-.-.a '1 '1 k_ conductor #1 g -e y NOTE: The circuit is shown with conductor #1 energized Relay #1 and all other conductors connected to neutral. 8-Following a short stabilization period (to be determined) Relays #1 & #2 will be switched, so 7.i.f' that conductor #2 is energized and all others 8 control connected to neutral. In this manner, all of the conductors in the test cable will be energized and watched for current leakage to neutral. v v l ^ s condu' - #2 ~ Relay #2 C' d computer relay control ...V.... V. g E s' i conductor #n 5 j,f g hM Relay #n simulated cable tray h .5 I 3 5 E' E O - 1000 VAC d. E 4 WW o Rt fuse > To Data c to N Acquisition Rectrier System ld o* t g e.,, g

n

d..,

NRC Staff Comments 'on Planned Thermo-Lao Raceway Testina

1..

.The proposed new test is a " qualification" test not a confirmatory test. The configurations that are being tested (cable fills & fire stops) have not been tested before. 2. The changes to the data collection specified by GL 86-10 Supplement 1 and the 10/29/92 will not change the licensing or design basis. The licensing and design basis is to provide 1 hour barriers. The licensing and design basis does not address instrumentation of test assemblies. 3. The' licensee cannot compare test results from the new tests with test results from old . tests as the important barrier parameters are different. 4. The licensee has provided no technical basis in Enclosure 1 that the placement of thermocouples every 12 inches on the tray ' side rails is equivalent to the 6 inch placWnont specified in GL 86-10 S1. 5. The licensee has stated in Enclosure 1 that the use of a #8 bare copper conductor is for cable functionality. This is not correct. The use of the #8 conductor is to provide the initial indication of a breech in the barrier during the fire exposure test. This was discussed at length with NEl (NUMARC) and the ACRS during their test program. NEl ' agreed to change their testing protocol "mid-stream" to include the #8 conductor. The staff's view is that the #8 conductor provides much better information~on barrier performance. In addition, the previous test (13-2) experienced "burnthrough" therefore, it is reasonable to expect that the ne'w test with less cable fill will also experience bumthrough. 6. The licensee states on page 1 of Enclosure 2 that they expect the raceway surface temperatures recorded during the proposed tests will exceed the acceptance criteria, specified in the 10/29/92 letter. Therefore, they are planning on not meeting their license. condition (1 hr barriers) and accepting the deviating barriers based on cable functionality (See #5 above). 7. The licensee states on page 2 of Enclosure 2 that the fire stop test will be conducted in accordance with IEEE 634. This is not acceptable. In the May 1996 SER the staff concluded that the IEEE 634 criteria was not applicable to fire stops installed in electrical raceway fire barrier systems. The IEEE 634 criteria is limited to penetration seals in walls, floors and ceilings. The appropriate fire stop test criteria is specified in GL 86-10 Si as the fire stops are an integral part of the electrical raceway fire barrier system. I ? ENCLOSURE 4 .}}