ML20140G548

Rev 0 to, Initial Review of Florida Power Crystal River Ampacity Derating Test Rept 95NK17030NC1973

ML20140G548
Person / Time
Site:	Crystal River
Issue date:	03/07/1997
From:	Nowlen S SANDIA NATIONAL LABORATORIES
To:	Ronaldo Jenkins NRC (Affiliation Not Assigned)
Shared Package
ML20140G552	List: ML20140G548 ML20141J045
References
CON-FIN-J-2503 95NK17030NC1973, 95NK17030NCI1973-R, 95NK17030NCI1973-R00, NUDOCS 9705120020
Download: ML20140G548 (18)

Text

_. _ _ _ _.. _. _ _._ _ __ - _._ _. _

e An Initial Review of the Florida Power Crystal River Ampacity Derating Test Report 95NK17030NC1,973 AIxtterReport to the USNRC March 7,1997 Revision 0 j

Prepared by:

Steven P. N9wlen Risk Assessment rnd Systens Modeling Dept.

Sandia NttionalLaboratories Albuquerque, New Mexico 87185-0747 Prepared for:

Ronaldo Jenkins ElectricalEngineering Branch Office of Nuclear Reactor Regulation U. S. Nuclear Regulatory Commission Washington,DC 20555 USNRC JCNJ2503 XA

@767/500zh map

TABLE OF CONTENTS:

3 Sscian

~

h 4

FORWARD.......................................

1.0 INTRODUCTION

1.1 Objective.................................................

1.2 Overview of the Licensee Ampacity Derating A j

Oripre. don ofReport.................. pproach.................

1.3

..........................I 2.0 LICENSEE RESPONSE OF RJLY 27,1995..........................

2.i Overview............................

2.2 Item 1 : Appendix B............................................ 3 2.3 Item 2: Aluminum Versus Steel Trays............................... 3

........... 3 3.0 A REVIEW OF THE CRYSTAL RIVER TEST PROGRAM...

3.1 Overview.......................................

3.1.1 Test Organization and Protocol..............

3.1.2 Test Articles Evaluated.................................

3.1.3 Instrumentation......................................

Fire Banier Systems Evaluated.............................. 5 3.1.4 Summary of Test Results.................................. 5 3.1.5 3.2 Anomalies Addressed in the Test Report................ l...........

.. 5 3.2.1 Problems with Base Line Test Ampacities for l' Conduits.......... 5 3.2.2 Failure to Achieve Temperature Balance....

3.3 Anomalies Noted by SNL........................................ 7

................... 7 3.3.1 Apparent Discrepancy in Clad Case for 4" Conduit with MTS-1 System....................................... 7 i

3.3.2 Discrepancy in Reporting of 1" Conduit Base Line i

Data for Test Article 2.............................

4.0

SUMMARY

OF FINDINGS AND RECOMMENDATIONS......i i

B

.~.

i a

FORWARD a1 The United States Nuclear Regulatory Commission (USNRC) has solicited the support 4

i National Laboratories (SNL) in the review of ficensee submittals associated with fire p I

and electrical engineering. This letter report represents the first report in an anticipated se review reports associated with ampacity derating submittals from the Florida Power (FPC) for the Crystal River Energy Complex (CREC). The submittal reviewed by SNL documents the results of a series of test sponsored by the licensees to assess the ofThermo-Lag 330-1 and Mecatin fire barriers when installed on cable trays and conduits. T 1

i report documents the results of an initial SNL review of the licensee submittal. De documents were submitted by the licensee in response to USNRC Generic Letter 92-08 and a su j

USNRC Request for Additional Information (RAI) ofJune 22,1995. This work was per as Task Order 1 ofUSNRC JCN J2503.

i l

4 i

I i

i l

1 i

~

j 1

1 i

1

}

i l

2Thermo-IAg 330-1 is a registered trademark ofnermal Sciences Inc iii

i

[

i 1.0 NrRODUCDON i.

j In response to USNRC Generic Letter 92-08 and a subsequent USNRC Reques Infonnation ofJune 22,1995, the Florida Power Corporation (FPC) Crystal Ri 3

Complex (CREC) provided documentation of the licensee position regar factors associated with its installed fire barrier systems. In particular, the lice i

two specific questions raised in the RAI and submitted to the USNRC a test i

the results of a series of test to assess the ampacity derating impact ofcerta and Mecatiss fire banier systems for the protection of cable trays and aa= Ale.

i SNL was asked to review these items under the terms of a general tech j

agreement JCN J2503, Task Order 1. The documents reviewed by SNL are:

{

i Letter, P. M. Beard, Jr., FPC, to the USNRC Document Control Desk, Item i

l 3F0795-05, July 27,1995.

I Letter, G. L. Boldt, FPC, to the USNRC Document Control Desk, Item 3F0696-14, June 26,1996, including an attached test report from Underwriters Labor

)

"Ampacity Test Investigation of Raceway Fire Barriers For Conduit and Ca i

I Systems," Report Nu;nber 95NK17030NC1973, May 7,1996.

i I

This letter report documents the initial results of SNL's review. It should be n' ote l

review is limited to a review of the licensee RAI responses and an asses 4

ofthe licensee submitted test report and the ampacity derating factors derived the not reviewed any documentation related to the full process by which the licensee has actual in-plant cable ampacity loads, and hence, this review does not include any as these practices.-

1.2 Overview of the Licensee Ampacity Derating Approach The licensee approach is based on an wys'msital determination of ampacity de certain fire barrier installations, presumably typical of those used in the plant. The tests performed consistent with the guidance provided in Draft 16 of the IEEE P848 test sta ampacity derating. The tests performed included three different types of test item conduits,4" conduits, and a 4"x24" cable tray, all consistent with the IEEE standard. Six fire banier systems were evaluated for each test item including a 1-hr TSI sy 330-1 at nominal 5/8" thickness, a 3-hr TSI system 'Ibermo-Lag 330-1 at nomin a 1-br Mecatiss upgrade system MPF.60 over the 1-hr TSI system, a 3-hr Mecat system MPF-180 over the 3-hr TSI system, a 1-hour Mecatis stand-alone system M 3-hr Mecatiss stand-alone systern hfrS 3.

1.3 Organiration ofReport Section 2 of this report provides a brief review of the specific licensee RAI resp 1995. This bulk of this review has focused on a technical review of the test rep l

I

1 M-:x-2. Section 3 provides a point by point review of the licensee test report includ*

i f'*"8'I0",0f otential areas ofconcern. Section 4 sunnarizes the SNL findings and l

P m.

..--_ _-..= -..

l O

d i

4 1

1' i

4 1

1 s

1 1

i i

i l

t a

G l

af e

2

~.

2.0 LICENSEE RESPONSE OF JULY 27,1995 a

2.1 Overview In an RAI ofJune 22,1995, the USNRC asked the ham to respond to two concems related t hs then planned ampacity dorating test program. The licensee response to these specific was provided by FPC in a letter ofJuly 27,1995. The subsections that follow provide SNL's assessment of these responses.

2.2 Item 1: Apia.h B The first of the USNRC RAIitems questions the adequacy of the licensee's intent to treat its ampacity derating tests as an " Appendix B QA program for $re protection." 'he licensee response appears to acknowledge that the tests will be performed consistent v% the USNRC request, ahhough the SNL reviewer is =~=Im d to assess the compliance of the licensee's test e

program with the Appendix B requirements. This is an area ofexpertise outside the scope of SNUUSNRC task ordering agreenwnt under which these efforts are being conducted. SNL ukes no recws.w.dations as to the sc~rability of this response.

2.3 Item 2: Aluminum Versus Steel Trays The second RAI hem raised by the USNRC in its letter of 6/22/95, questioned the appli tests wrducted using steel raceways to app'ications at the plant involving aluminum racew The licetsee response cites several points as justifying this approach:

The IEEE P848 standard calls for the use of steel raceways in testing.

The issue was discussed with IEEE P848 committee members, and the basis for this decision was cited.

The licensee and the P848 committee conclude that use ofsteel would conservatively bound the resuhs for aluminum. A technical argument to this effect is provided.

SNL finds the last of these three points to be the most telling That is, SNL agrees with the Ecensee assessment that the use ofsteel test samples would conservatively bound alami=>=

raceway applications. The supporting technical arBuments provided are reasonable. The reduced emissivity of an =1ami== raceway as compared to a steel raceway will primarily impar:M Ene ampacity measured in a test. With an aluminum raceway the base line ampacity s reduced. The clad ampacity would likely be impacted to a much lesser degree due to the fact that dad'Eg decreases the importsace ofradiation heat tmnsfer from the raceway surfaces in the overall heat bstance. Hence, the derating impact derived from a steel weway test should be more conservative than an equivalent aluminum raceway test.

In summary, SNL agrees wi@ the licensee assessment that the steel raceway test results wil conservatively bound the yptications m' volvirig aluminum raceways. No further actions on this hun are recommended.

3 n

{

l 3.0 A REVIEW OF THE CRYSTAL RIVER TEST PROGRAM l

3.1 Cmh i

j 3.1.1 Test Organization and Protocol De tests submitted by FPUCREC were actuaDy performed by Underwriters la "Special Services Investigation." The implications of this M-*Jon are simply that was not seeking a "UL listing" ofits barrier systems nor has h establishad the fo normdy associated with such a listing However, k should also be notad that the tests performed under the direct supervision and control of UL personnel. For example 1

barrier systems were installw by either "-:=c:: employees On the case o

=== A*-er employees (in the case of the Mecatiss products), aH installations were m and assessed by UL personnel to ensure compliance with the cited installations p Dese thetors are important in that they establishes a signi6 cant level ofindepende FPL and the banier manufacturers. In general, UL is eminently quah6ed to p derating tests in an appropriate manner, i

i The test protocol utilized in the test program was based on Draft 16 of the IEEE P8 This included the guidelines on the size, composition, and construction of the act of the provisions cf the standard were followed in testing. All of the test e y=== criteria were achieved with, in the end, only one notable exceptite (see discussion in Sect par:kular, data gathering and analysis all appear to have been aw.+1ished consistent 1

swhrd protocol. Also, the licensee tests did establish that steady state test conditions achieved consistent with the standard in all regards (linear regression, slope, a j

calculations were performed and achieved the standard objectives for steady state).

1 3.1.2 Test ArticlesEvaluated 1

The licensee test program involved four Test Articles. Each of these test articles amprised of three items; namely, a 1" conduit, a 4" conduit, and a 4"x24" ladd Each of the hems in each test article was evaluated in a base line condition, and s clad condition. Hence, there were a total offour 1" conduits, four 4" conduits, and four 4"x2 cable trays evaluated. Each of the 1" conduits had a single,4/C,10AWG, 600V lig control cable installed. For each cf the 4" conduits, a tightly wrapped bundle of 12 600V cables (a total of 36 individual conductors) was installed. De 4"x24" cable tra three layers of the 3/C 6AWG cable installed. All of these provisions are fully consisten test standard speci6 cations n

It should also be noted that the licensee test procedures included the testing of a full test a simultaneously. That is, in any given ampecity test, a 1" conduit, a 4" conduit and a 4"x2 tray were tested simultaneously in a comrnon test enclosure. The standard does sp40ca for such simultu.aous testing provided that the test specimens are not located one but rather, that a minimum clearance of 36" of horizontal space is provided bs;;s the test i Dese requirements were achieved in the licensee tests as demonstrated by the dimensio drawings providedin the test nport.

4 W, -

i 3.1.3 instrumentation i

AB of the instrumentation sped %11y required in the IEEE standard was i 3

In addition, the licensee included a cumber ofsupplemental thermocouple I

cable tray surface teir.

and inside for some). yer.tures, and fire barrier material surface temperatures (outs A!! test data was monitored using calibrated dsvices and UL is known to have an exceBent calibration process that is followed as routine practice. N j

regard were noted.

3.1.4 Fire Barrier Systems Evaluated cond the 4 x2 cab e tray r systems, each installed on the 1" conduit, the 4" i

A i-hour Thermo-Lag 330-1, single layer, fire barrier system alone A 3-hour Thermo-I.ag 330-1, single layer, fire barrier system alone A 1-hour Thermo-Lag 330-1 fire barrier system with a Mecatiss MPF-system j

A 3-hour Thermo-Lag 330-1 fire barrier system with a Macatiss MPF-

. system A 1-hour Mecatiss MTS-1 sptem A 3-hour Mecatiss MTS-3 system One important factor to note is that the fire barTier system for the cable tray 2 the installation of a sobd tray cover on the ladder back cable tray. It is implied instrumentation drawings that the base line tests did not include this tray cove appropriate. However, it is recommended that the licensee be asked to confirm th 3.1.5 9====7 oftest Results The test results obtained by FPC are summarized in Table 3.1. For each of the barrier con 6guration both the ampacity correction factor (ACF) and the am (ADF) is given. Recall that the relationship between these two values is as follows:

1 ADF = (1 - ACT) *100%

3.2 Anomalies Addressed in the Test Report 3.21 Problems with Base Line Test Ampacities for 1" CMa%

The licensee report includes a diacussion of an apparent discopancy involving for the 1" conduit specimens. In particular, the base line ampacity limits for the 1 Articles I and 2 differed signi6cantly from those for the. nominally identical 1" con Articles 3 and 4. This anomaly was ' vestigated at some lengths Although no firm m

was reached as to the cause of the discrepancy, it was ultimately concluded th Article 1 and 2 resultz were in error and had recorded a false-high ampacity limit overstated the actual base line current limits). These base line tests were repeated aA 5

,n i

j the barrier syster$ and the modified resuhs were in agreement with the Test A l'

The original base line results were discarded, and the newer values were used to calculated i

ACF/ADF values.

Table 3.1: Sur my of FPC/CREC test results.

i Test Item Barrier System ACF ADF(%)

l 1" C=#4 Tests 1-br L.+1ag No Result

• No Resuh*

3-hr h. + 1ag 1.04

-4.26 l

MPF-60 Upgrade 0.838 16.2 l

MPF-180 Upgrade 0.843 15.7 MTS-1 0.818 18.2 l

MTS-3 0.775 22.5

[

4" Conduit Test 1-hr TNu+Iag 1.03

-3.31 3-hr h. + 1ag 0.9 73 2.69 MPF-60 Upgrade 0.801 19.9

{

MPF-180 Upgrade 0.765 23.5 MTS-1 0.769 "

23.1" MTS-3 0.665 33.5 Cable Tray Tests 1-hr Thermo-Lag 0.590 41.0 l

3-hr Thermo-Lag 0.586 41.4

{

MPF-60 Upgrade 0.444 55.6 i

MPF-180 Upgrade 0.421 57.9 MTS-1 0.397 60.3 MTS-3 0.361 63.9

• No results reports due to problem with original base line iest and inability to repeat clad tes

" As discussed below, SNL will question these results.

As a part of the problem diagnosis pim, UL did change out the cable installed in the 1" j

conduits for Test Articles 1,2, and 3. This was apparently to ensure that the cables were not the problem, a fact verified by the results. However, this problem did compromise one of the 1,

item fire barrier system tests. That is, the Mecatiss MPF-60 upgrade to the 1" conduit in Test i

Article I had already been installed at the time the error was detected. This made it impo i

fully reproduce the test for the 1-hour Thermo-Lag only fire brrrier system without resortin construction of a whole new test article. This is because the original base line test had been det..l.e4 to be invalid, but the' clad test with only the%ermo-Lag fire barrier system could not be reproduced with the newly installed cable. It was considered inappropriate to compare base line test to a clad test that used a different set ofcable and th... occupies, and hence, n results for this test item arereported with only the 1-br Thermo-Lag fire barrier system.

The licensee treatment is considered adequate to res& % observed anomaly in the context of al of the base line tests. That is, the licensee docuraent mn - onsistency ber.s.four separate base line tests for each ofits 3 test items (l" conduit P m 'uits, and cable trays)is considered adequat: to ensure that the base line test results are r

~.. rue. Hence, in the context of the base 6

r-

,m--

rv y-w e--w

i i

line test results, SNL Snds that this anomaly was adequately resolved by 3

no actions in this regard are recommeaded.

[

However, SNL has concems that because the exact source of the anom

.a identi6ed, there is no clear assurance that the a similar anomaly did not impa j

and in particular, the clad test results. This issue will be taken up in Section 3.3.1 below.

i 3.2.2 Failure to Achieve Temperature Balance t

j The IEEE test standard speci6es that as a part of the test protocol the average cond temperature measured at the center of the test item should be compared to that meas

{

each side of the center. The center and side location values should agree to within *4*C l

certain of the licensee tests, this condition was not achieved. Most of these anomalies resolved through retesting of the test items. However, for one case the anom j

namely, the Test Article 2,1" conduit, base line test. This has the potential to impact j

Article 2 results for both the 3-hour Thenno-Lag barrier only, and the MPF-18 The test report cites that "Short ofinstalling a new instrumented cable, reinstalling t j

and TSI-3 hr systems and repeating the ampacity test series a third time, no remedy j

anomaly was available. A third series of ampacity tests to resolve the anomaly was not Wed." Hence, the reported results include crediting of the test results impacted b anomaly. No final resolution is pavided, and no Ammaionjustifying the use of these test j

is provided in the report.

}

The unresolved anomaly involves a tsr.yer.ture difference in one base line test betwe j

and 2 of 4.8'C as compared to the limit in the standard of 4*C. Hence, SNL notes that the differences is not grossly out ofcompliance with the standard. Further, SNL notes that i

line current measured for this conduit was very consistent with the other 1" test article

{

nominally identical to this item (30.2 versus a range ofvalues 8om 29.8 to 32.0A).

Given these observations, SNL Snds that this anomaly, while not fully resolved in the lice i

h=antation, is quite minor, and should not be considered to have significantly com j

test results. It is rww..r.sr.ded that these test results should be accepted despite this minor i

anomaly. No further actions on this anomaly are recommended.

1 3.3 AnomaliesNoted by SNL l-j 3.3.1

~

i Apparent Discrepancyin Clad Case for 4" Cond,uit with MTS-I System 4

As was noted in Section '3.2.1 above, the FPC/UL tests did experience some anomalous j

readings. In particular, false-high current values for two ofits base line test specimens we chtained. No clear and de6nitive source for this error was identi6ed, although it was that a loxe shunt connection had caused the problems. The failure to concisely iden for this problem leaves open the possibility that other tests may have been affected b Problem.

7

- - n-i.

As will be demonstrated in the foBowing discussion, SNL does consider that one te j,

did, in all likelihood, experience a similar problem that impacted the test results. The concern is associated with the clad ten for the MTS-1 system instcIled on the ' " conduit 4

i coMuit oftest Article 3). Based on SNL calculations, it is suspected that this test also suffer flrora a false-high current reading. Given that this is a clad test, a false-high curren ten wguld yield an overly optimistic ampacity derating impact.

1 In order to ilhistrate why SNL hss reached this conclusion, one must look at the availab j

somewhat unique way. One way to view the cable / conduit / barrier thermal system

{

electrical network analogy.1 hat is, one can think of heat flow as analogous to elec 4

temperature as analogous to voltage, and thermal resistance as analogous to electrical resist Using these analogies, the system can be viewed as a thermal resistance netw certain temperatures, heat Bows, and thermal resistance elements. Under this approach j

between two elements or nodes of the thermal system (Q) can be expressed as follows:

1 i

AT g.

R,3,,,,,

I i

where (AT) is the temperature difference beraw the thermal elements or nodes, an j

the thermal teristance tera the elements or nodes. Given this expression, if the heat load and temperature difference are known, then one can easily calculate the effective thermal resistance between the two elements. In the cas, of the ampacity derating tests, the heat load determined based on the cable current setting as follows:

i O= I*R.2

% ~.ua i

i where (RJ is the electrical resistance ofthe conductors, and (n

) is the number of conductors within the conduit.

Of most critical importance to the current discussion will be the thermal resistance that edsts between the cables and the conduit. This thermal resistance has been the investigation. The most concise treatment was that ofBuller/ Neher and Neher /McGrath'. In 2

these works this value was found to be a function of the conduit s;w and cable fill character only. It is especially important to recognize that this value will not be, inSuenced in a

'J external factors such as the ambient temperature, or the presence of a fire barrier system.

In fact, one ofthe fundamental pwpa ofa conduit ampacity derating test set is that the thermal behavior beraw the cables and the conduit should remain constant in the c

_ tests. That is one of the primary reasons why the IEEE standard requires tha 8F. R BuBer and J. R Neher, "The Thermal Resistance Between Cables and a Surrounding Pipe or Duet Wan," AIEE Transactions V69,1950 pgs 342-349.

8J. R Neher, ami M. R McGrath, "The Calculation of the Temperature Rise and Lead Capacity of Cable Systems," AIEE Transactions, Oct.1957, pgs 752-772.

8 I

j

test specimen (conduit and cables) be used to perfonn both the clad and base line tes j

unintended changes in the internal thermal behavior might easily bias the test results which are l

intended to resect only changes in the external thennal behavior.

l j

this em me simple check that can be performed to assess the consistency 1,eb;w a base une and a clad conduit ampacity test is to check the value of the internal cable to cor,dui thermal remstance m each test. This value should remain essentitlly eaadaat and significan devint ons would be indicative ofpotential problems. The only supplemental data, beyond required by the IEEE P848 standard, needed to perform this calculation is the temperature o e, he FPC/UL tests report these conduit tw,.isare values for most of th l

Using the FPC/UL data, and the two equations above, SNL has calculated the effective therma maistance ber;;een the cables and the conduits for all those tests which report conduit temperatures. A variety of temperature bases were tried (i.e., cable hot spot to individual conduit temperatures, cable hot spot to ave. age conduit temperature, and cable average temperature t conduit avange temperature). The results for each of these calculations were quite consistent.

2

{

For illustrative purposes, the dis,:ussions which follow will utilize the thermal resistance values j

based on the differerse bereceen the average temperature of the cable bundle at the center loc j

as reported by UL and the average temperature of the conduit surface Table 3.2 illustrates the d

results obtained by SNL for tha i" conduit test items.

As can be seen, the values derived are quite consistent with one notable exception. With l

given test article, the derived values are extremely consistent, varying in all cases by le and for most cases by less than lay less than *2.5%. Even compadng one test article to an the variation is no more than *10%. The one clear exception is the original base line test for Tes Article 1. In this one case, a value is derived that is far lower than any of the other cited values.

fact, the estimated base line thermal resistance is nearly 30% lower then the values obtained fo 3

two corr =paadag clad cases.

This calculation is fully consistent with the UL observation that the initial base line test for Test Article 1 suffered a false-high current reading. His false-high current would overstate th rate, and hence understate the thermal resistance factor given that the temperatures were measure correctly. As can be seen the thermal resistance calculation can provide a clear and accurate indication ofpotential problems in the test data. (It is fully expected that a sitnilar treatment for the Test article 2 initial base line run would reveal a similar effect, and that the same treatment f the repeated base line tests would reveal thermal resistance values far more consistent with the other cited values. Unfortunately, the report does not provide the necessary conduit tem data for these cases.)

Now consider the same process as applied to the 4" conduit tests. The results for these test '.tems are summarized in Table 3.3. The values are again calculated on the same basis; namely, av cable ts,.i. ore at the center location as reported by UL and the average conduit temperatu;

• The average conduit tw.gisare was calculated by SNL using the simple average of the ts.grEures for all conduit tLmoccupies installed on a given test item as reported by UL.

9

Table 3.2: Su== my cf tremal resistance calculation results generate i

FPCNL 1" conduit ampacity testa. All values are calculated based o j

the average temperature of the cable bundle and the average condu l

Test Article Barrier Configuration

• Cable to Conduit Thermal l

l Test Article 1 P=%3ce ('C*ft/W)

I Base fiaa 4.15 Clad: thr TSI i

5.91 Clad: MPF-60 i

Test Article 2**

5.84 a

j Clad: 3-hr TSI besier 6.16 i

i Test Article 3 _

Clad: MPF-180 6.43 i

Base IJa.

6.85 Clad: MTS-1 Test Article 4 6.80 i

Base Line 6.96 Clad: MTS-3

{

Range ofderived values:'" _

6.48 i

5.84-6.96 j

• Note that values can only be cal ~1*ed for the initial licensee runs.

identified configuration only. report conduit temperature data. Henc i

"The initial run of the base line case for Test Article 2 is unavailable d i

oversight in preparation of the UL re. port. See Section 3.3.2 for further discu

• " Range excludes Test Article I base line test.

I i

t Table 3.3: Summay ofTremal Resistance calculation results generated j

FPC/UL 4" conduit ampa% results. All values are calculated based on the average temperature of the cable bundle and the average conduit i

Test Article l

Barrier Configuration Cable to Conduit Thermal Test Article 1 Rosinance (*C*fi/W)

Base Line 1.84 Ciad: thr TSI 1.73 Clad: MPF-60 1.74 BaseIJne Test Article 2 1.86 Clad: 3-hr TSIbarrier 1.87 Clad: MPF-180 1.93 Test Article 3 Baseline

~

1.86 Clad:MTS 1 1.43 Test Aru4 Baseline 1.79 Clad: MTS-3 R* age ofPa=*s:'

1.77 1.73-1.93

• The range ofresults excludes the Test Article 3 clad test.

10

j In this case the overall values are much lower indicating a bette

}*

the cables and the conduit. More importantly, as with the 1" c i

consistent with one notable exception. Within any given test artic L

64% band. Even comparing Mr.;w test articles, the variations are all i

cae notable exception is the clad test for Test Article 3. In this one ca n.

e value is approximately 23% lower than the corresponding base line much larger than one should anticipate, rivals that of the. Test Article i

have been a problem, and is certainly an indication that a similar j

te.

be answered by " working backwards" through i

I i

measured temperatures are correct, and that the thermal resistance Wi.

i conduit during the clad test was in reality approximately the same as yielded these test conditions. The Erst step is to

" yer4ure difference and the base line case thermal resistance as follows:

AT g=

, _B 6.7 - 75. 7 = 5.914 W/ft R

1.86 w

Based on this heat flow rate we can now estimate the corresponding ofconductors and the electrical resistance values as follows:

y=

0 (5.914)

) R,,,,* n

= $ (5.15E-4) (36),g*,

om the ampacity derating calculation we must no conditions:

yr, y (90-4 0) (a+T,)

(90-40) (234.5+90.5)

( T,-T,) (a+90),17,9 ) (90.5-39.0) (234.5+90)

= 17

• 6A Finally, the ACF and ADF factors can be estimated using this modi 6ed e ampacity in comparison to the measured value of the base line ampacity as follo ACF=-

'J*d 17.6

=

= 0.667 I

26.4 neseusse ADF = (1.0 - ACF) e100g = 33,3g 11

.t j

4 j

Hence, this exercise has illustrated that the anticipated derating impact should have been on the

.i j

order of 33.3% versus the value of 23.1% cited in the test repon.

One might question the accuracy of this approcch. This can be demonstrated by once again i

returning to the 1" conduit case known to have been compromised by a false-high ampacity l

meaurement. If the same wocedure is repeated for the first run of the base line test, the predicted normali=d cable ampacity is found as follows':

i j

AT 88.2 - 57.1 l

Q=R

= 5.298 N/ft 5.87 i

i i

y=

0 (5.298)

$ R,,,,* n

,,, = )

(1.31E-3) (4) = 31. 8 A l

p,y (90-4 0) (a+T,)

1 (90-40) (234.5+90)

( T,-T,) (a+90) = 31.8 ) (90. 0-39. 9) (234.5+90)

= 31. 8A i

4 l

Hence, a " corrected" normalized base line ampacity of 31.8A is predicted as compared to the original value measured for this test of 37.8A. More importantly, the normali>d base line current i

for the repeat test was found to be 32.l A, a value very close to the predicted value 31.8A. The j

minor differences can easily be attributed to the fact that the cables were replaced ber;ca tests j

which would cause a minor change in the thermal conditions. This predicted value certainly fa well within the range of values cited in the test report for the normali*ad I" conduit base line ampacity values; namely, 30.2A to 32.1 A.

Given these observations, SNL finds that the clad test ampacity for the 4" conduit protect MTS-1 fire barrier system is suspect. The test was likely compromised by a false-high am i

limit similar to the false-high values obtained in the Test Article I and 2,1" conduit, original bas i

1 line tests. No other tests appear to have been impacted by a similar problem. This would have resulted in the calculation of a overly optimistic ADF value for this one barrier configuration, the 1

MTS-1 system on a 4" conduit. SNL recommends that this particular test result should not be j

accepted by the USNRC as representative of the ampacity derating impact of this fire barrier j

system.

3.3.2 Di%.y in Reporting c " ~ Conduit Base IJne Data for Test Article 2 In the actual data sets presented in Appendix B of the FPC/UL report, it appears that the data for the initial Test Article 1,1" conduit has been inadvertently substituted for the intended Test Article

• Thermal resistance was taim as the average of the other two "first run" cases for the Test Article 1,1" conduit as shown in Table 3.2; namely, the clad test for the 1-hr TSI barrier and the MFP-60 systems.

12

. =

2,1" conduit initial base line test. That is, the tabulated data given in Appendix B is identical in i

ev.iy tr. gard to that presented in Appendix A. This is a highly unlikely situation to occur by rinfom. Further, the values reported in the data summary tables for the initial base line test for the i

Test Article 2,1" conduit, do not match those given in Appendix B.

It appears that in preparing the test report, the data from the Test Article I test was inadvertently presented in Appendix B as well. This oversight is considered a minor discrepancy given that the Test Article 2 iddal base line test was not used in the fmal data analysis by the licensee. While it 1

would have been very interesting to supplement the discussions presented in Section 3.3.1 above I

with the additional Test Article 2 data, the data is clearly of no significant consequence to the final l

results of the licensee tests. This discrepancy in the hmentation should be noted simply "for the l

\\,

' I i

l l

i l

l

{

j i

l 1

l i

d 4

13

i 4 )*

4.0

SUMMARY

OF FINDINGS AND RECOMMENDATIONS

.. t I

With respect to the Licensee's specific responses to the two items raised in i

June 22,1995, SNL makes the following E=M=y and recommendations:

i i

AnnenAiv B OA tm:...x.: SNL make no recommendations regarding the j

{

acceptability of the licensee response to this RAI item. The nature of the R outside the scope ofexpertise of the SNL reviewer and the scope ofwork i

established for the general task ordering agreement under which these ac have been performed.

1 Ahiminum veren nea.1 Envr SNL finds the licensee arguments in this r i

a

table. SNL finds that the use ofsteel raceways in the licensee tests will conservatively bound the results when applied to aluminum cable racewa at the plant. No further actions on this item are recommended.

i i

With respect to the licensee submitted FPC/UL ampacity derating test repor

{

licensee tests have been performed and reported in a manner fully consiste j

Draft 16 test standard. Any raomalies observed by the testing laboratory h j

test report with the following exception:

i a

'Ihe base line test for the 1" cor.duit in Test Article 2 did not i

objective of uniform tempture in that the average cable temperature at th

}

location deviated from the average temperature at one of the two 4.8'C, as compared to the test standard which calls for deviations of no m 4*C. This anomaly is clearly noted in the test report, but no resolutio SNL finds that this is a minor anomaly that will not significantly impact t reliability and validity of the test results. SNL recommends that the test re this test item should be accepted by the USNRC, and that no further actio resolve this anomaly are warranted.

{

SNL did, however, note one apparent anomaly not observed by the testing not discussed in the test report. This anomaly is considered to have comprom one ofthelicensee tests:

SNL Gnds that the clad test ampacity for the 4" conduit protected by the fire banier system is suspect. The test was likely compromised by a ampacity limit indication similar to the false-high indications identifie laboratory in the Test Article 1 and 2, l'coreluit, original base line tests. N tests appear to have been impacted by a similar problem. This appare would have resuhed in the calculation ofa overly optimistic ADF value for banier configuradon, the MTS-1 system on a 4" conduit. SNL recommend this particular tes. resuh should not be accepted by the USNRC as rep the ampacity derating impact of this fire barrier system. The basis fo has been discussed in detailin Section 3.3.1 above.

g 14

..... - m ---

- ~' ~~ ~

~

~~

i i

• c',

Based on these Andags, SNL recom==ls that the USNRC accept the ampa results derived in the FPC/UL test report as representative of the derating imp i

banier systems with the exception of the MTS-1 system as installed on the.4 nc4ed above, it is recommended that this one test resuk not be accepted.

In addition, SNL cites the following item as a point for which clarification is is,Ti.,c

{

It is rww.. ended that the licensee be asked to mim that the cable tray ampacity tests were all performed in the absence ofany solid tray cover been installed on the test hems. This appears to have been the case base i

drawings pmvided, but is not explicitly stated in the report. A simple state confirming this observation would be sufficient to resolve any uncertainty j

regard.

i l

Finally, SNL makes the following observation:

i i

SNL notes that the data provided in Appendix B of the report for the orig line test of the Test Article 2, l' conduit test item, is an apparent duplicatio data set presented in Appendix A for the Test Article 1,1" conduit test item i

{

that the original base line test data for both test articles was not used in the fin data analysis, this apparent reporting oversight is considered ofno significa consequence. It is recommended that this observation be noted "for the record" only.

j 4

e I

e 15