ML20149E088

Forwards Request for Addl Info Re Thermo-Lag Related Ampacity Derating Issues for Byron & Braidwood Stations Re GL 92-08

ML20149E088
Person / Time
Site:	Byron, Braidwood
Issue date:	07/09/1997
From:	Dick G NRC (Affiliation Not Assigned)
To:	Johnson I COMMONWEALTH EDISON CO.
References
GL-92-08, GL-92-8, TAC-M82809, NUDOCS 9707180095
Download: ML20149E088 (5)

Text

- . .. .

July 9, 1997 Ms. Irene Johnson, Acting Manager Nuclear Regulatory Services Commonwealth Edison Company Executive Towers West III 1400 Opus Place, Suite 500 Downers Grove, IL 60515 i i

l

SUBJECT:

REQUEST FOR ADDITIONAL INFORMATION REGARDING THERM 0-LAG RELATED AMPACITY DERATING ISSUES FOR BYRON AND BRAIDWOOD STATIONS (TAC NO. M82809)

Dear Ms. Johnson:

By letter dated March 21, 1996, as supplemented July 12, 1996, Commonwealth Edison Company (Comed) responded to the staff's Request for Additional i Information (RAI) dated December 4, 1995, related to Generic Letter  ;

(GL) 92-08, "Thermo-Lag 330-1 Fire Barriers," for the Braidwood Station. The  !

supplemental response included a new set of analytical calculations and l ampacity methodology for a range of fire barrier installations. The staff, in I conjunction with its contractor, Sandia National Laboratories (SNL), has reviewed the submittals and has identified a need for further information, as discussed in the enclosed RAI.

Please provide your response to the RAI so that we may continue to review your submittal s . Due to contractor scheduling restrictions, please respond to the RAI within 60 days of receipt of this letter.

Sincerely, ORIGINAL SIGNED BY:

George F. Dick, Jr., Project Manager Project Directorate III-2 Division of Reactor Projects - III/IV Office of Nuclear Reactor Regulation Docket Nos. STN 50-454, STN 50-455, /

${g n. STN 50-456, and STN 50-457

//

Eo gg

Enclosure:

RAI f cc w/ encl: See next page Distribution:

g " Docket? File PUBLIC PDIII-2 r/f og J. Roe, JWR E. Adensam, EGAl R. Capra So.a. C. Moore (2) G. Dick (3) S. Bailey OGC, 015B18 ACRS, T2E26 R. Lanksbury, RIII R. Jenkins, RVJ u...oENT NAME: BRAID \BR82809.RAI 8

To racilve a copy of this document, Indicate in the box: "C" = Copy without enclosures *E" = Copy with enclosures "N' == No copy 0FFICE PM:PDIII-2 ( La:'1MII-2 Os PM:FDIIIf2# l /; D:PDIII-2 lE NAME SBAILEY </G Cil0CRE\ GDICK U/M RCAPRA >

DATE 07/7/97 ~ 07/T/97 07/Of/97' " ' 07/1 /97 0FFICIAL RECORD COPY b!!!hl0fIfll$b!!

I. Johnson Byron /Braidwood Power Stations

. Commonwealth Edison Company cc: i l

Mr. William P. Poirier, Director George L. Edgar i Westinghouse Electric Corporation Horgan, Lewis and Bochius Energy Systems Business Unit 1800 M Street, N.W.

Post Office Box 355, Bay 236 West Washington, DC 20036 Pittsburgh, Pennsylvania 15230 i Attorney General ,

Joseph Gallo 500 South Second Street l Gallo & Ross Springfield, Illinois 62701 1 1250 Eye St., N.W.

Suite 302 EIS Review Coordinator Washington, DC 20005 U.S. Environmental Protection Agency 77 W. Jackson Blvd.

Michael I. Miller, Esquire Chicago, Illinois 60604-3590 l Sidley and Austin

! One First National Plaza Illinois Department of f

Chicago, Illinois 60603 Nuclear Safety i Office of Nuclear Facility Safety l

Howard A. Learner 1035 Outer Park Drive l Environmental law and Policy Springfield, Illinois 62704 l Center of the Midwest l

203 North LaSalle Street Commonwealth Edison Company Suite 1390 Byron Station Manager Chicago, Illinois 60601 4450 North German Church Road Byron, Illinois 61010 i U.S. Nuclear Regulatory Commission .

l Byron Resident Inspectors Office Kenneth Graesser, Site Vice President )

4448 North German Church Road -

Byron Station '

Byron, Illinois 61010-9750 Commonwealth Edison Station l 4450 N. German Church Road l Regione,i Administrator, Region III Byron, Illinois 61010 U.S. Nuclear Regulatory Commission 801 Warrenville Road U.S. Nuclear Regulatory Commission Lisle, Illinois 60532-4351 Braidwood Resident Inspectors Office Rural Route #1, Box 79 Ms. Lorraine Creek Braceville, Illinois 60407 Rt. 1, Box 182 Manteno, Illinois 60950 Mr. Ron Stephens Illinois Emergency Services Chairman, Ogle County Board and Disaster Agency l Post Office Box 357 110 East Adams Street l Oregon, Illinois 61061 Springfield, Illinois 62706 Mrs. Phillip B. Johnson Ci', airman 1907 Stratford Lane Will' County Board of Supervisors  !

j Rockford, Illinois 61107 Will County Board Courthouse l l Joliet, Illinois 60434 l

i i

2-Commonwealth Edison Company Braidwood Station Manager Rt. 1, Box 84 Braceville, Illinois 60407 Ms. Bridget Little Rorem Appleseed Coordinator 117 North Linden Street Essex, Illinois 60935 Document Control Desk-Licensing Commonwealth Edison Company 1400 Opus Place, Suite 400 Downers Grove, Illinois 60515 Mr. H. G. Stanley

-Site Vice President  !

Braidwood Station l Commonwealth Edison Company RR 1, Box 84 Braceville, IL 60407

.y l

RE00EST FOR ADDITIONAL INFORMATION THERMO-LAG AMPACITY DERATING ISSUES COMMONWEALTH EDISON COMPANY BYRON STATION. UNITS 1 AND 2 AND BRAIDWOOD STATION. UNITS 1 AND 2 l

DOCKET NOS. STN 50-454. STN 50-455. STN 50-456 AND STN 50-457 l

1.0 BACKGROUND

j By letter dated February 15, 1995, Commonwealth Edison Company (Comed, the licensee) submitted documents that were requested during phone conversations

between Comed and NRC staff, related to Generic Letter (GL) 92-08, "Thermo-Lag l 330-1 Fire Barriers," for the Braidwood Station. In their response of March 28, 1995, the licensee indicated that their analytical approach has been shown conservative to actual ampacity derating test results, and ampacity testing is not planned for abandoned in-place Thermo-Lag fire barriers. The staff, in conjunction with its contractor, Sandia National Laboratories (SNL),

completed a preliminary review of the licensee's analytical approach based on

.the February 15, 1995, submittal and identified a number of open issues and

! concerns requiring clarification by the licensee. The staff transmitted a Request for Additional Information (RAI) to the licensee on. December 4,1995, (erroneously referred to in the attachment as the November 2, 1995, RAI)

L requesting a response on the outstanding issues and concerns.

By letter dated March 21, 1996, as supplemented July 12, 1996, Comed provided a response to the staff's RAI. The supplemental response included a new set of analytical calculations that was identified in the July 12, 1996, submittal as Calculation BYR-96-082/BRW-96-194, Revision 0, covering a range of fire barrier installations for both cable trays and conduits. Calculation t BYR-96-082/BRW-96-194, that was also applicable to Byron Station, introduced a l new methodology and addressed ampacity analyses for single cable trays in "special" barrier configurations, multiple horizontal trays in a single fire barrier enclosure, single. vertical cable' tray risers, and multiple cable tray j risers in a single fire barrier enclosure. i The staff has completed its review of the licensee's supplemental submittal and requests that the following questions be addressed by the licensee. .)

2.0 ADDITIONAL INFORMATION REQUIRED i SNL has identified several points of concern regarding Calculation BYR l 082/BRW-96-194, which was included in Comed's submittal dated July 12, 1996.

(See Section 3.0 of the attached SNL letter report.) Please address the following questions:

2.1 It appears that Comed' base case comparisons are not applied on a consistent basis. In particular, the licensee is comparing a calculated l ENCLOSURE l

I '

l I______-_._______,__.____. y g. ,ne%

2-clad case ampacity limit to a base case ampacity derived on a different basis. The estimates of fire barrier Ampacity Derating Factor (ADF) should be based on self-consistent treatment of the clad and base line

cases. In this case, it is considered critical to assess both the i clad and base line ampacity limits using a self-consistent thermal l model. If the thermal model is used to predict the clad ampacity I

limits, then a thermal model fully consistent with the clad case analyses should also be used to assess the base line ampacity limits as well. The licensee is requested to implement a thermal model for the analysis of the base line case ampacity that is fully consistent with I

its clad case analyses, and to then base its final ampacity derating assessments on a comparison of the clad and base line thermal analysis results.

! 2.2 The licensee has presented a table of heat intensity versus depth of fill values (Item 13 of page 13 of BYR-96-082/BRW-96-194). This table is in apparent conflict with the heat intensity values cited by Stolpe

-and in the ICEA standard P-54-440. The cited values appear to modestly over-state allowable heat intensity limits, and hence might lead to optimistic estimates of the cable ampacity limits.

2.2.1 The licensee is requested to establish the basis for how this heat intensity table was developed and how it is applied in practice, and to reassess the ampacity limit calculations in light of this apparent discrepancy.

2.2.2 The licensee is also requested to provide the supporting calculation cited in BYR-96-082/BRW-96-194 as the basis for this table (i.e., Calculation ESI150-1, Revision 0).

2.3 The licensee cites in Item 2 on page 12 of BYR-96-082/BRW-96-194 that the base line ampacity for a 3/C, #6 AWG, 600 V cable with a 2.5" depth of fill is 27.5 A. The basis for this value is not clear. SNL was unable to reproduce this limit using standard approaches to ampacity analysis given that the licensee thermal model has cited the ICEA definition as the basis for fill depth calculations. The licensee is requested to describe, in detail, how this value was obtained, or alternately the subject calculation should delete references to and reliance upon this value as the " base line ampacity" for the cases examined.

2.4 Several references are made in BYR-96-082/BRW-96-194 to a "SilTemp l Sheet,".but the fire barrier descriptions do not include a discussion of any 'such sheet used in the installation process. The licensee is requested to clarify if such a material is used in its fire barrier constructions.

l l

Attachment:

Report to U.S. NRC, Revision 0, dated May 2, 1997, prepared by Sandia National Laboratories I

l w ,,m-m,<. --m.--or +w- - . , .,,en- n w -

A Review cf the Braidwood Station Calculation BYR96 082/BRW-96-195 on Fire Barrier Ampacity Derating Factors for Special Configurations 3'*

4 AIAtterReport totheUSNRC Revision 0 May2,1997 Prepamiby:

Steve Nowlen .  ;

SandiaNatiorallaboratodes Albuquerque,New Mexico 87185-0747 (505)S45-9850 i

7. reced for:

. Ronaldo Jenkins Electrical Fag =-:--iss Branch ,

Office ofNuclear Reactor Regulation ,

U.S.NuclearRegSn %nmisman r Washington,DC 20555 l USNRC JCN J-2503, Task Order 3 e

. ATTAC K NT I

b -

g

i TABLE OF CONTENTS:

Sostian i

a

. Eass  !

FORWARD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii .

l ..

i . .

~ '

1.0 INTRODUCTION

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 '

1.1 Backsround . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Objectives of the Current Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 4

l 1.3 Overview of the U~aw Approacb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2  :

i 1.4 C.ciL. ion of This Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

< 2.0 THE LICENSEE APPROACH TO ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . 4 i

!. 2.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 I i 2.2 Methods of Calculation for the Base Une Ampacity . . . . . . . . . . . . . . . . . . 4 i 2.2.1 The 1CEA Ampacity Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 l- 2.2.2 1CEA Heat Intensity Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2.2.3 Use of a Thermal Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.3 The Licensee 'Ihermal Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

23.1 Overview of Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 1-

, 32 Critical Modeling Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 i

3.0 POTENTIAL POINTS OF CONCERN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

[ 3.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

3.2 Points of Concern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

, 3.2.1 Consistency . . . . . . . . . . . . . . . . ..........................13

, 3.2.2 Possible Error in ICEA versus Stolpe Application . . . . . . . . . . . . . 14 1

3.2.3 Licensee Heat Intensity Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 1

3.2.4 References to "SilTemp Sheet" . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

) 33 SNL Analysis of the Base Line Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 l 3.3.1 Overview and Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 l l 3.3.2 SNL Base Une Analysis for the Sensitivity Study Cases . . . . . . . . 19 l i 3.3.3 SNL Base Line Analysis Resuhs for the Licensee Case Studies . . . 19 l l 3.3.4 9==-y of Findings and Recommendations . . . . . . . . . . . . . . . . 20

3.4 Implications for the G-63 Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 4

4.0

SUMMARY

OF REVIEW FINDINGS AND RECOMMENDATIONS . . . . . . 22 i j 4.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 I

j 4.2 Remaining Points of Concern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

} 4.3 Impli:ations of the Current Finding for the G.63 Calculations . . . . . . . . . . 23 APPENDIX A: ...................................................24 J

i 4

i

, l f

1 u -

i

FORWARD The United States Nuclear Regulatory Commission (USNRC) has solicited the support of Sandia National Laboratories (SNL) in the review of utility submittals associated with fire protection and electrical engineering. His letter report documents the results of an initial *

, s..

SNL review of a set of supplemental calculations included in the licensee response to an

• 3 USNRC RAI sent to the Braidwood Station on November 2,1995 (Calculation BYR96- .,.

I 082/BRW-96-194). Dese rww calculations deal with the issue of fire banier ampacity '

derating factors for cable tray fire barrier systems involving certain special cable tray fire barrier configurations. His report is the fourth in a series of review reports, of which the first three dealt with calculations for more standard barrier configurations and were I prepared under USNRC JCN J2017, Task Order 6. The current efforts were performed ,

under USNRC JCN J2503, Task Order 3, Subtask 5. ...

. h.'

.l*

.g

.4

$*:$E

~

• $5-0I.

t. .

t t'.*.*

.:t;.

• .+

7.!

t *

r.i.

. l,-

3 .

I .

h.:

1. -

e:.

m .t 2,)

.:i

1.0 INTRODUCTION

1.1 Backsround In response to USNRC Generic Imter 92-08, Braidwood Station provided documentation ,

of the licensee position regarding ampacity derating factors associated with its installed ~

fire barrier systems. In particular, the originallicensee submittals included documentation '

of analyses performed to assess the adequacy ofin-plant cable ampacity factors for two types of Appendix R cable tray and conduit fire barrier systems; namely, Thermo-Lag and Darmatt. On August 25,1995 SNL prepared a letter report which do-nanted the resuhs of an extensive review of the licensee approach and calculations'. Based largely on the  !

findings of this review, on Now 2,1995 the USNRC sent a Request for Additional l Information (RAI) to the licensee requesting clarification of several po*mts of concern identified in the review.

In a letter dated Maich 21,1996 the licensee provided an initial response to this RAI.

SNL provided a reviesv of this initial respouse to the USNRC in a letter report dated August 16,1996. Submuently, the USNRC forwarded an additional set ofresponse documents to SNL for review. This supplemental set oflicensee response documents included documentation of an entirely new set of analytical calculations covering a range fire barrier installations for both cable trays and conduits. SNL submitted a third letter report on December 20,1996 that focused on the updated calculations that wnwponded to those originally submitted by the licensee and reviewed by SNL in 1995. The licensee documents provided for SNL review were:

- I.stter, Denise Saccomando to USNRC Document Control Desk, March 21,

, 1996 with one =Mment. (This is the initial response document reviewed in the August 16,1996 letter report.)

- 14ter, John B. Hosmer to the USNRC Document Control Desk, July 12,1996 with three attached licensee calculations (BYR-96-059/G-70-96 092 Rev. O, BYR-96-082/BRW-96-194 Rev. O, and G-63 Rev. 4).

1.2 Objectives of the CurrentReview The current report documents an initial reWew by SNL of the licensee's new calculations presented in BYR-96-082/BRW-96-194, PAev. O. Initially, it had been intended that this would be a final review of the licensee RAI response. However, as the review progressed potential areas of concern not previously noted were identified which might sig46=1y compromise the reliability of the calculation resuhs Hence, the objective of this review then became to identify potential areas of concern that might warrant a supplemental RA1.

See letter S. Nowlen, SNL, to R. Jenkins, USNRC, Dated 8/25/95 and the attached letter report "A Review of the Braidwood Station Analysis ofFire Barrier Ampacity Derating Factors," DraA Revision 0, 8/25/95. .

l _

1

This effort represents a contimation ef work initially undertaken as a part cfgeneral technical support contract JCN J-2017, Task Order 6. The current efforts have been .

performed as a part of the follow on efforts under JCN J2503, Task Order 3, Subtask 5.

1.3 Overview oftheIJoensee Approach In general the limne** approach to amd'y derating is based on analysis with limited ,

, ::M validation of the analysis process. This review has focused on the licensee Calculation i, BYR-96-082/BRW-96-194 Rev. 0 which covers single cable trays in "special" banier '

configurations, multiple horizontal trays in a single fire barrier enclosure, single vertical cable tray risers, and multiple cable tray risers in a single fire barrier enclosure. .:

^

The approach taken by the licensee in its calculations is rather unique. In most ampacity thermal models the objective is to calculate a limiting ampacity for a given set of .'

predefined cable installations, typically for both the base line and clad conditions. -

However, in the Braldwood submittal the licensee sets an arbitrary value of the actual I.]

current, and then calculates a limiting depth offill for that current level. This is then M, compared to standard values of ampacity for'the base line case to determine the derating ,

impact. Such an approach to analysis can be risky because it is important to establish that L-the thermal model is consistent with the base line ampacity tables. This is diamened ..

further beloW.

• For each case a similar approach to analysis is undertaken: 2 A nominal " base line" ainpacity limit of 27.5A is assumed for all cases. This .

value is cited as being the base line ampacity for a 3/C 6AWG 600V light power cable installed in a tray to a depth of fill of 2.5". (As will be noted gj below, SNL takes exception to this characterization, but this concern :n; uhimately has no r'al impact on the results of the analysis.) W:

IE Using this value as an assumed input, the licensee implements a thermal model 2:E to calculate the maximum depth of fill under clad conditions for which this  :-

current value will result in a cable hot-spot temperature of 90*C. ., j, Y:4 t- The clad ampacity limit is converted to an equivalent " Heat Intensity" factor fl{.;:- .

for the clad case in a manner similar to that taken by Stolpe and by the ICEA 'M

===d*y standard for open top cable trays (P-54-440).  ;

s This clad case heat intensity value is then compared to a " base line" heat  :.[?j.

intensity value derived for the same depth of fill from a table developed by the :p Sz ^ ^ . 3, .

The am=d+y derating factor (ADF) is then calculated based on a ratio of the  :,}

two heat intensity values in a manner similar to that taken when actual cable .:~.y current limits are compared.

Q.}:

..e 9.

. E .'

W.:

W

While somewhat unusual, this process could, in principle, work. As will be discussed further below, SNL has identified four points of concern regarding these calculations, one ofwhich is considered p=11y significant. SNL's most serios concerns are primarily associated with issues ofmodeling consistency. These concerns are diameW at length in Section 3.2 below. . .

~

1.4 Ci-- -= ton ofDis Report Chapter 2 provides a more detailed discussion of the licensee modeling approach, and certain fundamental %ts critical to the understanding of that approach. Chapter 3 .

provides for a direct review of the licensee BYR-96-082/BRW-96-194 Rev. O calculation, and in particular, focuses on potential shoitcomings in that method as currently .

documented. Chapter 4 summadzes the SNL recomr-ada+1ons regarding the adequacy of r ae responses and calm 1=daan these w,n e

1

. 3 t

t; 2.0 THE LICENSEE APPROACH TO ANALYSIS 1 2.1. Overview '

license Calculation BYR96-082/BRW-96-194 documents the results of a set of case S

> analyses performed to assess the derating impact for a number of special barder and cable ,

, 5 9 tray configurations. In particular, the assessments include 3 hour3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> barriers wroyiM of ;T i

. both single and double layers ofmatedal, *Ihermo-l.ag and Darmatt materials, horizontal-and vertical cable trays, and cases with either a single tray or multiple trays in the same fire Jfj'

'g barrier erdosure. 'Ihese cal =1=+1aan include certain very unique analysis approaches, and ' .y

! include a number of assumptions that should result in a net conservative result ifvie r dy . '.i ;

i -

applied.

J3

.:a3 The objective of this chapter is to describe the details of the licensee analvsis Mad, and i: 1 to identify the critical assumptions that wiB impact the reliability and conservatism of the N{3 =

resuhs. These hems are taken up primarily in Section 2.3 below. As a pi E-imry to '

these deme = tons, Section 2.2 provides a briefreview ofcertain concepts and approaches

f.

c.i .

that are critical to an understanding of the licensee analyses. The specific items of concern /g4 regarding the licensee implementation ofits analysis model will be taken up separately in JQ Chapter 3.  :.

,3 i 2.2 Methods of Calculation for the Base line Ampacity j.} 4

44 -

'Ibere are at least three methods by which one can estimate the base line ampacity of a

. ll]

cable in an open cable tray; namely, (1) use of the standard tables of ampacity such as the .. . ,

, ICEA P-54-440 tables, (2) use of the Stolpe/ICEA concept of heat intensity limits, and (3) 7.,g implementation of a thermal model. The licensee analysis does implement a thermal . rg model, but is also heavily dependent on the second of these approaches, the heat intensity Mg.

approach, although this approach is not Wally common in practice. Hence, an $$:

~ understanding of the heat intensity approach is critical to understanding the licensee

calculations. S.j E.

y l

.3 The dem== tons in this section serve two purposes. First, each of these three methods of ;jp.3 t analysis will be briefly summarized in order to establish a firm understanding of each. In addition, as a part ofeach discussion, a single case example taken directly from the  :/' .f i

licensee analyses will be considered to iDustrate each approach. These case analysis results will also support *=~== tons in Section 3 below.

.i?

d, 4

9l.b.

• The specific case example considered by SNL wiisponds to the base line case equivalent f.eg to the clad tray cases -vaminad in the first six calculations presented in the licensee  :.j submittal.; namely, those cases used'by the licensee to assess the sen:itivity of the resuhs ')$:j

^

to the assumed width of the side rail-to-barrier air gap and to the cable mass-to-bottom :d banier panel air gap. 'Ihe base line case considers a 4"x24" cable tray filled to a depth of .M 2.5" based on the ICEA definition of fill depth (see discussion in 3.2.2 below) with 3/C, #6 W-

, AWG,600 V cable with an outside diameter of 0.953". SNL will evaluate this same tray (.f.'k,

for the base line condition using each of the three methods of analysis. Note that the  ;

Y(

licensee assumed a " base line" ampacity for this case of 27.5 A.

!lg 4

k5 s H Y

a

~

2.2.1 DelCEA AmpacityTables

- Use of the ICEA P54-440 standard tables of ampacity for open top cable trays is by far the most common approach taken to estimate cable ampacity limits for general cable tray applications. This includes those cases requidng the application of a separately derived estimate of a fire banier ==p=% derating factors. The process can be illustrated for the specific case example cited in 2.2 above quite en:Hy. In & example it is assumed that the depth offill,2.5", is consistemt with the lCEA definit on of e parameter. SNL has made this assumption br== the thermal model clearly uses the ICEA dermition ofdepth of fJl '

(see further discussion of depth of fin issues in 2.2.2 below).

The ===4ty iimit of a 3/C, 6 AWG,600V cable can be calculated using either Table 3-3 l or 34; the final resuh will be the same. Using Table 34, the base ==p=W for a 6AWG cable with a 2.5" depth of fill is given as 30A. His value must be corrected for the actual l cable diameter as compared to that assumed in the standard. This is donc using the j equation in Section 2.3 of the erandard as foBows:

• • • 2 I= y****'* = 0. 953" (3 0A) = 31.77A dm 0.900" \

where 0.953" is the cable diameter cited by the " =m. and 0.900" is the cable diameter cited in the ICEA tables; Hence, the tabulated base line an==@ would be this modified value,31.77 A. -

2.2.2 ICEAHeatIntensity Approach A second and less commonly applied method of ebmining cable ampacity limits set forth in the ICEA standard is to use the " heat intensity" limits in Appendix A of P54440. His method is generally applied when mii.po!ation of the tabulated limits in needed (for smaller cables, or depth of fill values outside the tabulated range of 1"-3"). The f.mdamental basis of Stolpe's work, and of the ICEA tables, is that the allowable maximum value of heat intensity, that is the rate ofheat generation per unit volume of cables, is a function of cable tray depth of fill only. It is therefore assumed that for any combination of cables loaded into any open top cable tray, the maximum beating rate of each conductor in the tray can be calculated bened only on knowledge of the total tray depth offill and the cori yonding heat intensity. It is also important to note that the licensee has cited Stolpe as the basis for their tre, and hence, the results which follow should be viewed as an appropdate basis for assessment of the licensee results.

In order to iDustrate e approach let us first briefly review the coispi of heat intensity as defined by Stolpe. For e pr.e.sation we will use the notation of Stolpe in which (Q) is the heat intensity, (q) is the actual heat generation rate for a single conductor, and (a . ) is the number of er=Mors in the cable. Heat intensity (Q) is basicaDy the allowable total heat generation rate per foot of cable tray per square inch of cable cross- ,

section. (Note that e is distinguished from the " Watts per foot" awhui in that the value heat intensity is volume based whereas " Watt, per foot" is simply length based, 4

5

allswable heat per foot cf tray.) MathematicaDy, this can be expressed for a single cable as (see St:1pe, eq.1): .

g , 4 * % e,ecor, a

An.eu.a .

i _. ._

t

. N, .

, Note that (Q) is typically given in units W/ftlin so that (q) is given in units of W/ft, and J-2

- (W is the cross-sectional area of the nMm cable (s) in units ofin . If one is '.

given the allowable maximum heat intensity and the characteristics of the cable, then the .

i allowable rate of heat generation per conductor, (q), and the cables maWmum current (I) -

i can be calculated. Given the conductor heating rate, the corresponding maximum cable ..

current is easily calculated using (see Stolpe, eq. 2):

4 " I'Renue i '::

where (RJ is the AC electrical recietance per foot of cable. 'lil i

t In this process, there is one additional issue which must be understood. This has to do with how the cross-sectional area of the cable and hence the depth of fill of a cable mass is "/;:.

} calculated. The most obvious approach for a single cable is to simply assume the cable is '.

circular, and to calculate the cable cross-section as: . .-

• nd' {'

A e2. " 4 ..

.O 3.s 4.'. :

However, under the definition provided in the ICEA standard, the area to be used is the 'e;t.t.

equivalent area of a square section fully surrounding the cable, rather than the actual cross ,j.y section of the round cable itself(see Section 2.2 of the standard): g.

A, ,f = d 8 '

v. ~

i:4

.n:

This issue also has implications for the calculation of depth of fill as well. In general, the depth of fill is related to the number of cables in the tray (Q, the width of the tray (w,), and the cable cross-sectional area as foHows:

= ^ .

s,,,, . . gcur->. p

.m

n However, the answer obtained is, quite obviously, dependent of which method is used to i:i.

calculate the cable cross section. Ifone uses the actual circular cross section, then depth offillis given as: ?l:

p r .-

g.

6 *:

!s Dr

n.oa.,n d' f222 4y wr 1

h effect, one is samuning that the cables are very tightly packed with e-*t.ny no air ..

gaps. It is this defmition that was used by Stolpe in his original paper. h contrast, the ICEA standard very clearly assumes the equivalent square cable area because depth of fill .

is sp+=ny cited as being calculated as: .

D.aa., d' d,u, = ,

mr De ultimate lesson to be taken firom this discussion is that in order to properly apply the heat intensity approach, the cross section of a single cable, the depth of fill of the cable mass, and the heat intensity must all be based on the same analysis approach. In particular, a 2.5" fill as defined by Stolpe is not the same as a 2.5" fill as defined by the ICEA. In fact, a 2.5" "lCEA fill" is really equal to a (2.5*x/4) or 1.%" "Stolpe fill".

However, for a given tray, either method should yield essentially the same result.

I As a final note to this discussion, observe the fact that the ICEA cited heat intensity limits are actually identical to those cited by Stolpe at anv eiven death of fill. Heat intensity is a function of depth of fill only, and W*nu Stolpe and the ICEA have used different definitions for this parameter, one might not expect a one to.one correspond =* between )

the Stolpe and the ICEA heat intensities. However, is it simple to verify that, for example,

the ICEA heat intensity at 2.5" "ICEA fill" is identical to the Stolpe heat intensity for a 83.3% or 2.5" "Stolpe fill." As noted above, for a given cable anangement the Stolpe method will yield a smaller fill depth than will the ICEA approach. Hence, the Stolpe fill i method would yield a higher heat intensity limit due to the lower fill. The ampacity one i would calculate ends up e=*t.ny the same because the higher best intensity is offset by l the correspondingly lower cable cross-section area given by the Stolpe versus ICEA  !

l method. Hence, in the end a given cable is allowed to generate approximately the same level of heat in either case, and hence, ' ends up with a *r*intly the same ampacity limit.

Returning to the our specific case example, the ama W limit of a 6AWG wire in a cable i- tray with a 2.5" depth of fill can be easily determined. However, the result will depend on l whether we consider 2.5" depth of fill to be a "Stolpe fill" or an "ICEA fill." Based on an examination of the licensee thermal model, it is clear that in that model the licensee has implemented a definition of depth of fill based on the ICEA approach (see further j l discussion in Section 2.3.1 below). Hence, a comparison based on the ICEA definitions  !

I appears most appropriate. However, for illustrative purposes, SNL will consider both i

possible definitions.

L 7

For a 2.5" ICEA fdl,1CEA speci6es a heat intensity limit cf 1.784W/Alin8. In accordance with the ICEA approach using the square equivalent, the crou section cf the cable is givenby:

A , = dr = (0.953fn)2 e 0.9084nz

.. a..  :

Rearranging Stolpe eq. I as given above, we can now calculate the eBowable heating rate l for each of the three conductors in the licensees cable based on the heat intensity as  !

specified in theICEAtables:

• l Orcza-2.s Aean2. , 1.784 0.908 Gm, 1 = 0.540 W/ft -

n meson 3 '

.' \

yi Reanunging Stolpe's eg. 2, the conesponding cunent can be calculated as: '.f

~

y#C** .

I czA J 0.540 R ,, .) = 32.45 A l

$ 0.000513 As one expects, this value is quite similar to the actual value extrapolated directly from the ICEA tables (31.77 A as hsW above). The minor difference, about 2%,is easily ,.

attributed to round-offin the standard tables. -

For illustration let us assume that the 2.5* were in fact based on Stolpe's definition of fill d depth. The corresponding heat intensity must be obtained by e.polation he== Stolpe kj! I only plots his results for up to a 2.4* depth of fdl (80% fill of a 3' cable tray). The value ijj estimated by SNL for a 2.5" fill based on extrapolation of Stolpe's plot is 1.75 W/fVm2. :l;:

As noted above, this is roughly identical to the ICEA heat intensity values cited above. B. l However, if this is a *Stolpe fil!" then the one must also use the circular cable cross-5.

section in the calculation of aBowable cunent. Hence, the ama*% limit for this 2.5"  : I "Stolpe fiH"is given by: .

f.,9.

1 J=D 0n = 0.953 (1.75) (3.14 ) = 28.5 A -e.-

2) )

n ,R,, 2 (3) (5.13E-4 ) i,;.r; '

5;

'Ihus, if the 2.5" is assumed to be based on Stolpe's definition of fill depth then an 'l?

ampacity of 28.5 A which is much more consistent with the 27.5 A value cited by the W. i.

M, As a final example, l'et us assume that the 2.5" fib was based on the ICEA definition. In -;.

. this case as cited above the equivalent "Stolpe fil!" would in fact be about 1.96*. Using this value, the heat intensity limit given by Stolpe would be about 2.4 W/fVm2, and the l4 corresponding ampacity limit would be:  ;,;

S i::

.s:

4:

l 1 ,

y,D 0n , 0.953 (2. 4 ) (3.14 ) = 33.3 A

2) n o R, 2 $ (3) (5.13E-4) 1 i .

4 As noted above, when the depth of El and cable cross-section are treated consistently, the '

i Stolpe method yield ===aan.ny the same mma ahy,33.3 A, as did the ICEA ==*hd 32.45 A. The key is to ensure that depth offill and cable cross-section are used l

consistently.

2.2.3 Use of aThermalMo' del l

l A third approach to the meimadon of base line ama=W values is to apply a thermal model j and to simply calculate the allowable =p+y limit for a given case. This approach is

! rarely, if ever, used *m practice for cable trays, but is relatively common for the calculation -

! of conduit mmpeakies (the National Electric Code, NEC, specifically allows this approach for example). With a direct calculation of actual ampacity limits for a given case there is more potential for error. This is because modeling assumptions and any mienkes made in l

the development or execution of the thermal model will directly impact the results.

However, in the context of relative analysis of ampacity derating factors e is, by far, the preferable option.

In particular, in the context of the ampacity derating issue, the objective is to assess the r,1=dve hnpact of the fire barrier system on cable ==reehy limits. It is p=Uy important in this process to ensure that both the base line and clad case maaeahy values i have been =====ad on a consistent basis. 'Ihis holds for both testing and analysis based l spproaches. In the specific case of testing, it is important that the sample be tested in both  !

the base line and clad conditions, and that these two test results be compared. It is not  !

considered appropriate to, for example, compare a tested clad ampacity to a base line l ampacity from the standard tables. The tepic of the licensee analysis in this regard will be l taken up in Chapter 3. l In the case of our specific example problem, the licensee has not provided any thermal

= adding resuhs to assess the base line amp =ahy. However, given the effort that has been applied to the clad case analyses, the implementation of a fully consistent base line case analysis is quite easily accomplished. SNL, as a part of this review, has implemented such a model. The MATHCAD analysis Se for this case is presented in Appendix A, and the model will be taken up in detail in Chapter 3.

The final resuh of the SNL base line thermal model analysis estimated the base line amp =ahy to be 32.58 A. This value is in remarkably good agreement with the ICEA table-based value of 31.77 A, and with the ICEA heat intensity based value of 32.45 A.

However, the value is, once again, signiAa**1y higher than the 27.5 A value cited by the licensee as the base line ampacity for this case.

9

2.3 DeIlcensee HermalModel s

2.3.1 OverviewofApproach De thermal model devcbped by the licensee employs as set ofwell known and well ,;

characterized heat transfer correlations incorporated into a single analysis package which is " tuned" to suit a specific case analysis as needed. There is, however, one aspect of the < ,?

model which is unique in comparison to other ==a d+y dorating analyses reviewed by

.SNL. ~

In a typical cable tray fire barrier .gcity derating analysis a given cable tray '97:

l configuration is dermed in terms of the physical characteristics of the cable tray and cables.

This hypothetical tray is then analyzed with a thermal model to determine the current load l that wiu yield a predicted conductor temperature of 90*C while the ambient is at 40*C. Ti l De calculation is r=aaa'ad for the base line and for the clad conditions, and the ra=*ing ' 's.

=ma A+y values compared to meimate the derating impact. In this approach, the physical 1,',2 characteristics of the system remain fixed, and the cable current is used as the " floating" ,

parameter that is adjusted to match the desired thermal conditions. In effect, this mimics  ?

the process of an ampacity derating test in which currents are adjusted for a given specimen to obtain the desired thermal conditions. .S In contrast, the licensee calculations in BYR96-082/BRW-96-194 are based on a

,f substantially different aoproach. The one aspect of the licensee submittal that is most  :

unique is that the licensee analysis has used the cable depth of fill as the " floating" -

l parameter in the analysis, rather than the cable current. That is, the licensee modeling approach is to fix the cable ampacity at the outset of the calculation to a preset value, and 9

Y;.Q to then adjust the depth offill until the hot spot temperature condition is satisfied. This ':;.,

approach can be made to work because reducing the depth of fill is the same as reducing :d the numhar of cables assumed to be present in the tray. This, in turn, reduces the total  ;:.i

. .~f.

heat load on the system and hence reduces the estimated temperature drops through each ':q element of the thermal model. The depth of fill, i.e. the total heat load, can be adjusted .9 until a match is obtained. *>-

The results of the initial thermal analysis are an eBowable depth of fill for the chosen

" nominal (rated) ampacity of the cable" for the cable tray in a clad condition. In order to y

.l.M establish the amaadty derating factor, the corr ==aaading base line current must be M.i estimated. This is accomplished by using the heat intensity approach and a set ofspeci.il :y.g licensee derived heat intensity versus depth of fill values. That is, the depth of fin .: . ..

determined in clad case analysis is used to establish an allowable base line heat intensity I..:

based on a pre-set "look-up table." His heat intensity value is used in turn to establish the '., {'

allowable ampacity limit. His value is then compared to the clad case ampacity in the We;-

mandard manner to determine the derating impact. n

.::i In principle, the concept of adjusting the depth of fill to match the thermal conditions is an }I:?

.i acceptable approach to the problem. However, as will be noted in Chapter 3 SNL has },:

certain concerns related to the licensee implementation of this thermal model that could ( i:

significantly impact the results of the licensee analysis.

M l

N w

10 :w:.

ys$

2.3.2 CriticalM~Wg Assumptions The licensee's analysis has ir.wryo,ned a raunber of mada11ag assumptions, both

. conservative and non-conservative, that will significantly impact the analysis results. h some regards, the licenses thermal model has followed standard and accepted engineering ,

practices. Theseinclude: - - -- ---

- Standard, modem correlations for the convection rates both inside and outside of the fire barrier have been employed in an apparently proper manner. For the outside surfaces, the orientation of the surface has been included. For the, inside surfaces both orientation and gap width have been considered as appivprise. (Note that this had been a point of concern raised in the original 1995 review.)

- The analysis as presented by the licensee is clearly daemnantad and relatively easy to follow. The model has been implemented using w = Wally available computer software, and includes proper hadling of all units in the analysis.

Sources for all correlations have been cited.

- The analysis has included consideration of the heat transfer effects within the cable mass itself, albeit in a rather simpliSed inanner.

Those assumptions considered by SNL to be of a non-conservative nature include:

-- Full credit is given for heat transfer through the sides of the cable tray system in addition to the heat transfer from the upper and lower surfaces. This i practice is we..ry to common practice in which heat transfer from the sides is j neglected. h general, so long as the base line and clad cases are treated

! consistently, this would not be a significant point of concern. It is the potential

! incontietencies in this treatment that are of concern to SNL. .

- h treating the sides of the cable tray, the licensee has assumed that the entire height of the cable tray side rail will be at the same tw,yernure as the surface of the cable mass (also assumed to be at a uniform temperature). This assumption is non-conservative for two reasons. First, there will be some temperature drop from the cable to the side rail due to contact thermal resistance bn c.;,r, the two items. Second, the cables only come into contact

with a fraction of the side rail (typical depth of fills for the licenwe calculations are in the range of l' or less). Hence, the rest of the side rail will act somewhat like a fin to lower the average temperature of the side rail below that of the cable mass. The licensee treatment assumes a side rail temperature higher than that to be expected in reality which could over-estimate the role of the side rails in the heat transfer process. Here again, the primary concern of SNL is related to the consistency of the treatment.

- h treating the bottom of the solid bottom cable trays used by the licensee, h has been e-mad hat t the temperature of the bottom plate will be identical to 11 r - - - - - - - - ,-. - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ . _ _ _ _ _ _ _ _

l

~

that cf the cable mass avface. This b pa+;.ny non-conservative W="- it .

Ignores the thermal contact resistance between the two aufaces. However, ' l given the configuration of the licensee trays, solid bottom, this will likely be a minor affeet in this case. .

. However, oh**ing these potential sources ornon-conservatism are several other  ;. -

l j

a:runption of a conservative nature. These include: '

l r -

i For all analyses involving muhiple trays in a single enclosure, the licensee has l assumed that any un-powered cable trays present will act as perfect insulators -

I and fully block the heat transfer through the " shadowed" portion of the fire .,

,l barrier. .::, .

An amianivity of 0.4 has been assumed for the unpainted Thermo-Lag surface. .c ..

This value is much lower than the 0.9 value typically mammwi in such sisy. .es.  :- }.,

This willbe conservativewan- it will reduce the heat rejection capacity of ,.

l the fire barrier system. .

,]

Painted barriers are assumed to have an amianivity of 0.9 at the outside (cited . Et

{ as typical of a painted surface) but will retain the 0.4 emissivity assumption i

noted imn=ti=+ely above for the inner anfaces. Hence, a lesser level of j conservatism is realized when a painted barrier is considered. ..

A sensitivity analysis was performed to assess the impact of the assumed tray '?

2 rail height, side rail-to-banier gap width, and bottom tray-to-barrier gap height it.:

on the estimated derating knpact. The most conservative values (the minimum Wi; values) have been used in subsequent analyses. .?-l! ,

. Overall, the not impact of the licensee modeling assumptions abould resuh in a reasonable assurance that the ampacity derating impact has been conservatively estimated in this h

.g regards. As will be noted in Chapter 3, there are aspects of the thermal model that still .T.

render the final resuhs of questionable reliability. , M;

)::::'-

<*,:q 4

li:;.I

9 5 ".l .

? *

<*t i-w:p 65 12  ?:

y%.-

- 1 3.0 POTENTIAL POINTS OF CONCERN 3.1 Overview As a part of this review, SNL has identified four points of concern in the licensee's -

)

analyses. These items are taken up in Section 3.2 below. The most significant of these '

concerns relate to the manner in which the licensee has performed the base line case assessments. 'Ihe concerns in this regard relate to both the lack of consistency betw the base line and clad case assessments, and to certain errors and inconsistencies in how the actual values were obtained. As will be noted below, SNL finds that the licensee  ;

calculations have been seriously compromised by these problems. As an ahernative, SNL J considers that the most appropriate basis for an analysis of this type is for the clad case  ;

1 and base line case to both be calculated using a consistent thermal model. SNL has implemented such a model for the base line case analyses as a part ofits review efforts.

'Ihe approach to modeling taken by SNL and the resuhs of these analyses are &m-4 in Section 3.3 below. Finally, SNL has considered whether on not the observed concerns will impact the earlier Braldwood analyses for standard configurations, Calculation G-63, j as discussed in Section 3.4 below.

l 3.2 Points ofConcern I 3.2.1 Consistency

! Of the concerns identified by SNL the one th'at is of most significance is the observation that the licensee is compadng a calculated clad case ampacity limit to a base line ampacity derived on an entirely different basis. In any ampacity derating n====>* either vihd or analytical, it is critical that both the base line and clad ampacity values

have been determined on a consistent basis. By comparing a given analysis resuh to a standard table of ampacity limits, the licensee is potentially comparing " apples to oranges" and this renders the results of the analysis highly suspect.

The concern can be easily illustrated through an analogy to the testing approach. In particular, it had once been proposed that ampacity derating values be based on a comparison of a tested value of clad case ==pehy to the ICEA tabulated value of the open tray base line ampacity. This practice was considered inappropriate, and is no longer allowed in the testing standard. Ampacity derating values in testing must now be based on the comparison of a clad case test resuh to a coirspording base line case test result. This ensures self-consistency in the test results.

The intent of ampacity derating is to establish the actual relative impact of the file barrier system on the ampacity limits. In testing, it has been concluded that this can only be accomplished by compadng a base line test to a clad case test where each has been performed under consistent test conditions. By the same reasoning, the only .ypiepriate basis for the analysis of ampacity derating factors is the comparison of clad and base line analyses in which each analysis has been performed using the same =~Wg ermphan, correlations, and parameters.  ;

~

13 l

The primary basis for this constraint is that in previous efforts' SNL has hustrated that calculations cf the absolute ampaity limits cf a cable tray system in comparison to experimental or tabulated values can be very dif5 cult, and that these calculations are i subject to signi5 cant variation based on nominally innocuous modeling changes. It was,  ! ,

however, found that it was much easier to reliably reproduce the relative ampacity ... ,

derating impact from an experiment using a pair of self-consistent thennal models. That ..  :. )

is, the relative cal =1&n of ADF was relatively simple provided that the base line and clad cases were analyzed on a consistent basis. This is because the relative ADF impact is Fi. I

. f.

actuaHy presented as the ratio of two calculated ====% values, and hence, minor errors . .

in the thermal model tend to be self-canceling provided that consistency is maintained . .

.('t'.

In the case of the liennw submittals this coanietaacy ht;;w the base line and clad case .

analyses has not been ==in+=iaai In particular, the licensee has only exercised its thermal .',-

model in the analysis of the clad cases under consideration. De base line case ampacity limit is not assessed using a corresponding thermal model, but rather, is based on tabulated *:1 values for the base line ampacity (as derived from the licensee's own table of best intensity . :f.5 .

values versus cable depth of fill). No assurance whatsoever has been provided that the '.i. I licensee thermal model is at au consistent with the cited heat intensity limits. This is :iT:

. considered to represent a serious flaw in the licensee approach that renders the results * ' .:

• • P**'-

?:

Y g Findinn and Reenmmendatinne SNL finds that the licensee's comps'ison of a clad case .; ,.

ampacity limit to a base line ampacity derived from tabulated ampacity values (or beat ,: ;

intensity limits) is an inappropriate basis for the derivation of ampacity derating factors. It .

is recommended that the licensee be asked to reassess its ampacity derating estimates .

using a thermal model for the base line case analyses that is fully consistent with the ':;,ij ,

thermal model used to estimate the clad case ====% limits. (See further discussion of a  :? .

SNL implementation of such a model in Section 3.3 below.)

.9.{.E

q.:

3.2.2 Possible Error in ICEA versus Stolpe Application g::p;-

. . w.

In calculating ampacity limits, it would appear that the licensee has made an error of ,f analysis. This ima m both the individual sensitivity analysis cases cited by the licensee G s, and the speci6c special configuration cases. De error is related to the manner in which  : :::-

the licensee has calculated used depth of fill to calculate base line =mamel'y limits. *-

/. .. .:.

As was noted in 2.2.2 above, Stolpe and the 1CEA provide different definitions for how .,

depth of fill is to be calculated. These differences also apply directly to how the cross- y:1'

. /.

sectional area of an individual cable is calculated as well. So long as one is consistent in -$

.. the calculation of both values, then essentially the same result will be obtained for either '.@;il case. In the licensee analyses this consistency appears to have been violated. fj.,

.:!i:1 8

See " Fire Barrier System Cable Ampacity Derating: A Review ofExperimental -).

and Analytical Studies," A Letter Report to the USNRC, August 15,1995, Fm' al, preyered (?g.  !

by SNL underUSNRC JCN J2018. l.;.f .

14 .,eg:j id5-

-l:5: ,

i .

l.

In the licensee's clad case analyses an examination cf the thermal model quickly and

clearly reveals that the ICEA defmition of depth of fill has been used. For example, see j ,

page 16 of the licensee calculation. Near the middle of the page the value %" is i calculated from the given depth offdl. This equation also defmes depth offill as it is j dermed by the ICEA standard. The depth of fdl defmition is also repeated at the top of page 20, and again, it is quite clear that the ICEA dermition has been applied. This same J  ;

definition has been used thr=ghaut the hainare of the licensee clad case analyses as well. ,.

i 1 j In the analysis, the clad case ampachy is used directly to set the total heat load on the 1

system using the electricat ranicance of the conductors and the total number of conductors ,

i present (qql.2*n.*n - ). Thus, the total heat load and the depth of fdl are l

! intimately linked through the conductor count p.i.g.rter (n.*n - 1. Giventhisitis a l quite apparent that the clad ampacity limit is based on the ICEA definition of fdl depth. 1 j This applies to both the sensitivity and special configuration analyses.

[

Consider now the sensitivity cases. As was illustrated in Section 2.2 above, a direct application of the 1CEA P-54-440 standard yields a base line ampacity limit signih*1y

]

l j higher than the 27.5 A values cited by the licensee. SNL has explored two potential i explanations for this discrepancy:

i i -

The only way that SNL can come even close to reproducing this value based on a direct application of the ampacity tables is to assume that the 2.5" is based I

i on Stolpe's definition of fill depth so that the Stolpe definition ofcable cross-j section applies. If this interpretation is correct, then the result has been to j artificially reduce the base line ampacity in comparison to the clad ampacity,

and hence, to artificially reduce the calculated ADF.

l

} -

One additional explanation considered by SNL is that the value of 27.5 may l

! have been derived for a cable tray with a solid cover. Recall that in the L Braidwood G-63 calculations, the licensee cited that solid bottom cable trays with solid steel covers were used predominantly at the plant. Waaaa testing '

, was cited as indicating a 15% derate because of the solid tray covers. Indeed, j if the values of 27.5 is divided by 0.85 (in order to remove" a 15% derating of l j the ICEA value), a modi 6ed ampacity of 32.35A is obtained. This is, in fact, j . very dose to the value derived from the ICEA heat intensity tables (see

discussion in Section 2.2.2 above). This is one alternative explanation for how I this values was derived. -

j Either explanation is considered a significant po*mt of concern. If the value derives from j using a mixed definition of depth of fill versus cable cross-section, then this is a clear .

! error. Ifit derives from an assumption of a 15% derate for a covered tray, then it is also j an inappropriate basis for the base line case which should resect an open top tray. (Recall

that many of the cases e=iaad in BYR96 092/BRW-96-194 did not include a solid j cover on the cable tray. This included all of the sensitivity cases. It would be

! inappropriate for the licensee to apply a base line dy for a covered tray to these j analyses. The base line case should be the solid bottom, open top tray case.)

,1 - .

, 15

ne impact of this apparent disw.percy on the sensitivity cases in very dgniScant. For example, if the base line ampacity assumed for the Srst cf these sensitivity studies (see -

pages 16-28 of 297) is raised to the ICEA ampacity (31.77A) then the calculated ADF increases to 47A% as compared to 39.3% calculated by the licensee. ,{

, .i It is unclear if the ADF values from any of these case studies will be used by theTicensee _ Jg l in actual applications. The summary ofresults presented at the end of the calculation does not include these sensitivity cases, and hence, it is generally assumed that the values wiil

") .

not be used. It should also be noted that this change would not impact the fmal sensitivity  ;

maaeions in that the relative ranking of the case study ADFs would remain the same.

. .y .

3:.

The same observation also applies to the speciSc case studies included by the lir== to $

assess the special conSgurations. As was noted above, all of the clad case analyses have  :

,;if '

1 been based on the ICEA definition offill depth. De cerig-:=%g base line case analyses 1.)

for the various applications are all presented in the very last set of calculations which begin ' ";f on page 292. For the base line case analysis, the licensee simply sets the depth of 611  : ?.,j

. derived in an individual clad case analysis, uses a "look-up table" to get the r~rr pending base line beat intensity for that depth of fill, and then calculates the corresponding 3

.sj

==W. The error is clearly seen in this last step. The fourth equation on page 293 '~. .

clearly defines the ampacity relationship used (see equation:"Ampacity(DOF)=..."). This  !.0 equation is clearly based on the Stolpe deSnitic;n of fill depth (note the presence of the 5f.

factor z/4 in this equation which is only used when the Stolpe definitions are invoked). ~d 1

.g l Findings and Recommendations: SNL finds that the licensees has failed to establish an .i adequate basis for its assumption that the base line ampacity limit of a 6 AWG,3/C,600 V J.5 cable in a open cable tray with a 2.5" depth ofcable fill is 27.5 A. It is r=+= == 4ed that g.

the licensee be asked to remove references to this value, and to assess the base line .t::y; ampacity values using a compatible thermal model for an open tray case without solid F.l.} l covers (such as that Eerme=1 in Section 3.2.3 below).

Further, SNL finds that in performing its assessments, the licensee has apparently

?d mixed ICEA and Stolpe deSnitions of depth of fill in an apparently inappropriate manner. I:,.

De licensee results have significantly understated the ampacity limits for the base line '- l case. As was noted in Section 3.2.1 above, it is ultimately SNL's iE+= Ha+1aa that .y;i '

the licensee should be asked to base it base line assessments on a consistent thermal 3YG model. If this broader recomrnanda+ ion is implemented, then SNL's concern in this $$

i rding inconsistencies in the depth of fill treatment would be rendered moot.

Tl 3.2.3 IJcensee HeatIntensityTables

.  :?.

.r sl

.::c ;

As a part ofits submittal, the licensee has presented its own table of heat intensity values 7$ '

similar to that provided in the ICEA tables (see licensee item 13 on page 13 of 297). His  ;: ' . ' ,

licensee table allegedly was derived from the work of Stolpe and the ICEA tables, ..-

although the values do not appear to be compatible with either of those two works. $,. .,

W.: ,

As was noted in Section 2.2.2 above, the heat intensity values of the ICEA and Stolpe are  ?:

essentially identical for any given value of the depth of fill, even though the two sources d.@

use different definitions of fill depth. In comparison to the ICEA/Stolpe tables, the .W.J.

Q:;:

w::

16 . 9 3,

&n -

i .. a.

_ ____.________.____e.__. _ _ . _ _ . _ _ _ _ .. .

l . .

]

l . Ecensee table appears to modestly over-state the limiting best *mtensity values for all depth i

of fills cited. For example, at a depth of fill of 2.5*, the ICEA/Stolpe best intensity limit is i

1.784 W/filin'. In comparison, the licensee cites a value somewhere bet uw 1.962 and 2.095 W/ft/in 2(the licensee values at 2.594" and 2.473" respectively, SNL has not l attempted to interpolate between the two values). ,

l The basis for this disupf s unclear. i 'Ibe Econsee has not diemsead how its own heat ., -

i intensity table was derived in the submittal. The submittal does cite a separate licensee  !

! document (Calculation ES11501 Rev. O, Enaneae ref 16). However, this document is not available for review.

l I 1 i SNL does note that at the end of each calculation the licensee has presented a conversion . l' 1 for heat intensity for " square cables" to best intensity for "round cables." This correction

does indicate that the licensee has some myyridstics of the differences between the two j methods. However, the basis for and pmpose of this " correction"is unclear. As noted d

above, despite the differences in depth of fill between " square cables" as per the ICEA and "round cables" as per Stolpe, both present essentially identical heat intensity limits for a given fill depth. Hence, the objective of the licensee is entirely unclear in this regard.

Findines and Reconwar.datione SNL finds that the licensee cited heat intensity table j (item 13 on page 13 of 297 of the licensee =_le' e nt) is in apparent conflict with the

ICEA/Stolpe tables of heat intensity. It is recommended that the licensee be asked to resolve the apparent discrepancy. It is also recommended that the licensee be asked to i review its actual cable ampacity calculations to ensure that this discrepancy has not

adversely impacted its assessment of cable ampacity limits.

l 3.2.4 References to"SilTemp Sheet" i

j In several of the licensee calculation, there is a reference made to the %%== of the i SilTemp sheet" (see for example page 24 of 297,4* line from the bottom). No references i to the use of a SilTemp sheet, a glass fiber blanketing material, in the construction of the licensee fire barriers is provided elsewhere in the = Amin =1. It is pra=>=ad by SNL that I this is a " spurious" reference carried over from some previous application of the model, 1 but this should be verified.

Findings and Recommendations: SNL has identified references in the Hecaeae cal ~1=+1aan to a "SilTemp sheet" that appear inconsistent with the licensee descriptions of their fire barrier inst 11=+1aa_= lt is reconunandad that the licensee should be asked to verify if such a Si! Ten. > sheet is used in any ofits fire barrier systems, and if so, how the material is used. The ' ensee should verify that each ofits case analyses have either considered the sheet ii used, or not considered the sheet ifit is not used.

3.3 SNL Analysis oftheBaseLine Case 3.3.1 Overview and Approach 17 e . + . . 9 e ... 8 e

-e ,r - - - __ . _ _ _ _ _ _ _ _ _ _ . _ _ _ . _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _

\

While the licensee has not provided any thermal analysis of the base line cable tray cases in f

its =hD=L it b a relatively trivia! == nan to implement a base line case model given the work already performed by the licensee. In previous efforts SNL had already implemented a MATHCAD Se to reproduce the licensees single tray ampacity calmlations from Calm 1=dna G-63. As a part of the current review, SNL modified this file in order to e develop a base line cable tray thermal analysis model fully consistent with the licensee's

..g own thermal models for its analysis of the clad case ampacities as presented in the current "','

calculation. In implaaaadag this model, SNL has retained all of the critical modeling

===>=adaae used in the baaaa' analyses. This includes: -

SNL has retained an of the physical and thermal pivgdes specified by the .

licensee (for the cables, the cable tray, and the ambient). .:1' .

j$,'

i FuH credit has been given for radiation and convection from the top of the cable mass, from the sides of the cable tray, and from the bottom of the cable tray. 'l???

s.i The cable tray sides and bottom have been assumed to be at the same

.'!?

temperature as the surface of the cable mass

- .02 '

The surface of the cable mass has been assumed to be characterized by a single -

tswr.ture (i.e., an isothermal surface, as per the licensee's model). .

14.:,.

All of the exact same convection correlations have been applied.

,fii The same simplified model for heat transfer within the cable mass has been I

. .'.[:

applied.

j.l[

- "d SNL has retained the ICEA definition of fiU depth as per the licensee clad case

.i:::

thermal model.

@h

cM

+

SNL has calculated the base line case ampacity limit in essentia!!y the same manner as that '

applied by the licensee to the clad case analysis. The major difference in the SNL .y. .

implementation is that the depth offill is set to the value derived by the licensee for the clad case, and the ampacity is adjusted until the desired thermal conditions (90*C hot

$(li.i::

.M.

spot) are schleved. However, as in the 5censee clad case analyses, the depth ciffill is .',)

3 uniformly based on the ICEA definition, and the total heat load on the tray is calculated  ;!-

i directly from the cable current using the %*or count and electrical resistance factors. .}y?.

i.

This completely eliminates the calculation's dependence on any pre-set table or==a=A+y .Y:!.

or heat intensitylimits. '#

I rd The result of this exercise is a thermal model for the base line case that is fully consistent 5.),'

i

with the licensee's treatment for clad cable tray cases. Appendix A provides a listing of '-}. ,

the SNL implemented base line cable tray thermal model.  ?!.d I M'

.N:

?.v

!  :.:6.s

.W* i f'". * -

1

]g .*.

u:q.=

e 1

__ l

33.2 SNL Base Line Analysis for the Sensitivity Study Cases As an initif case study, SNL assessed the base line ampacity limit for a single, open top, 4'x24" cable tray with a 2.5" depth of fill of the #6 AWG cables considered in all of the Ecensee analyses. This case wei-yonds directly to the first six analyses presented by the ,

Ecensee in its submittal; namely, those used to perform the air gap sensitivity assessments

~

for both the side rails and the tray bottom. All si:: of these case studies can be considered .

to have the exact same base line condition.

The resuhs of this thermal model indicate a base line ampacity Emit for this case of .

32.58 A. This value is remarkably similar to the values derived from the ICEA tables (31.77 A) and from the ICEA heat intensity approach (32.45 A). However, the value derived is sigrihily higher than the nominal base line value cited in the Econsee study (27.5 A) which apparently derives from the Ecensee's own set of heat intaneity values.

The impact of A== jag the estimated base line ====% for these sensitivity cases is quite profound, and illustrates the importance of this issue. Consider, for example, the first of the licensee sensitivity case studies (presented on pages 16 28 of the submittal). This case study involved a 3-hour, double layer, Thermo-Lag 330-1 fire barrier system with no tray cover in place and the outer surface of the Thenno-lag unpainted. The licensee has cited an estimated derating impact for the fire barrier system of393% based on a base line

ampacity of 27.5 A and a clad ampacity of 16.7 A. In contrast, if the base line ampacity is ,

j raised to 32.58 A, and the clad ampacity is maintained at 16.7 A, an ADF of48.7% is  !

j obtained. This difference is very significant. ,

I As a basis for the comparison and validation of these resuhs, consider tha.t in tests j performed by Florida Power and Ught (FPL) the ADF for a 3-hour minpe laver Thermo-Lag 330-1 cable tray fire barrier system with no upgrades and no tray cover plate was l found to be 41.4%. The impact of a 3-Lour double laver system, such as that considered in the licensee analysis, will certainly be greater than this based on the presence of an additional air gap in the system (between the two barrier layers). Hence, it is clear that the haa~ cited ADF of393% is non. conservative. The modified estimate generated by SNL,48.7%, is in aH likelihood modestly conservative due to the nature of the i I

e==ations used by the licensee in the development of the clad case thermal model (see dbcussion of modeling assumptions in Section 2 above), but is also far more mneistant with the available @. =ml resuks.

333 SNL Base une Analysis Results for the Licenea* Case Studies The licensee has implemented a total of 14 =a-4 case analyses involving different configurations ofcable trays and fire barriers. Recall that the " floating parameter" in these assessments was the cable depth offill. Hence, in order to re-assess the resuks, it is necessary to estimate the base line ampacity for a given depth of fill using the thermal .

model Using the SNL implementation of the thermal model, this is again easily I' accomplished. The approach is to simply set the depth of fill to the final value from the licensee case analysis, and to then adjust the ampacity until a match to the desired thermal conditions is c-9=' +i 19

. - . - ~ - - . - . - . - - - . _ .. - -. - ..-. - - - - - . - . - - . . _ - -. - - - - .-

i As a nominal validation cf this spyrusch, SNL calculated the ampacity limit for a 1" depth .

cf fdl. A value cf 59.5 A was obtained. In comparison, using the ICEA table 3-6 values a -

base line current limit of 58.2 A is @=M. and using the ICEA heat intensity approach a value of 59.1 A is obtained. Each of these values is in excellent agreement. This result, ,' .

coupled with the earlier results for a 2.5" fill depth which also matched the ICEA tables ..

quite well, provides a reasonable assurance that this thennal model is providing self -

(.i conaturard results when the depth of fsl is varied. ,. W The licensee's analyses resuhed in 10 speciSc values of the timhing fill depth that covered

. all 14 of the case studies (some cases yielded the same depth value). SNL has produced , ,

! corresponding base line ama=% limits for each of these depth of fill values, and has ,

j ** the impact of these modified values on the ADF as summarized in table 3.1. 'Ibe ,,-

modified SNL results can be applied directly to the case examples cited by the licensee .-

based on matching the case depth of 6111imit to that of the table. C-

{ .yp 2

Table 3.1: Sa==='y of SNL case study base line analysis ampacity limit and ====% 'l l derating resuhs for the depth of fill conditions analyzed by the licensee. :l.,

Licensee Cited Values SNL ModelResults 9.:

i M ofFH1On) Baseline ADF(%) BaseI.ine .U ADF(%) '

Ampacity Ampadty .,

i

< Limit (A) limit (A)  :..;

j 0.68 69.755 60.6 75.0 63.3 1.

i.

i 0.72 67.475 59.2 72.5 62.1  ;-

0.76 65.41 58.0 70.2 60.8 M:n i

0.83 62.117 55.7 66.6 58.7 N 0.87  :.$

60.422 54.5 64.8 57.6 sq f 0.91 58.855 53.3 63.0 56.3 (Ih ,

0.98 .56.303 51.2 60.3 54.4 l 5.5.'

j 1.06 53.722 48.8 57.4 52.1 /.I

?

1.10 52.522 47.6 56.1 51.0 :j-[

j 1.14 51.381 46.5 54.9 49.9 Yl

$$l5 l

3.3.4 Summary ofFindings and Ra~wa=aad=%

<'l ,,

W..

SNL Snds that the licensee praedce ofcomparing a calculated clad ampacity limit derived ./e from its thermal model to a base line ampacity value derived from the licensee heat MI. ,

. intensity tables is ===-:+piable. It is recommended that the licensee should be asked to  ?.1,S l implement a base line thermal analysis model fully consistent with the clad case thermal :q model 0.e., using all of the same fundamental modeling assumptions and the same heat .,

transfer correlations). The results of this base line thermal model should be used to i.,:;,

develop the derating factors.

'tl.l.

I.N.

e 20 w:e ,

':. .:e I5-

u.,_m.~~ns,- ...nn-i As an chernative, given that SNL has implemented a base line case analysis model of the

) type aceded as a part of this review, this item might nominally be considered resolved j .

provided that the licensee adopts the SNL analysis model and results directly. However, it i is recommended that some formal resolution of this point ofconcern should be undertaken. _ _ _ , _ L_ _ ___ _ .__ ___. . . _ _ _ _ _ ;_ _ _ .

3.4 Implications for the G43 Calculations

l '

i L 8 Under prior efforts , SNL had reviewed the licensee G43 Calculations for single cable I trays and conduits. Given the observations made in this review for the special -

j con 6guration calculations, SNL returned to the G43 cal ~h4== to determine if the same j observations might be applicable'.

i In the case of the conduit. calculations, the concerns identiSed in the current review have i no impact on SNLs conclusions. The conduit calculations had already been p L,-J on j a basis consistent with the "best practices" approach diern==ad here. Namely, the base line

! and clad case ampacity values were calculated using a consistent thermal model and j compared to determine the derating impact.

t i In the case of the cable tray calculations, a similar but somewhat different approach to 1 analysis was taken. That is, the licensee had utilized their own test results to establish the j

thermal resistance between the cables and the closed top, solid bottom cable trays i analyzed. This value was then =*ieminad as consistent throughout the balmacs of the calculati ;n. Hence, while the licensee did not conduct a fbil base line case thennal modeling analysis, the licensee did ensure that the base line and clad cases were treated on a consistent basis. This treatment was considered fully appropriate given the liceneaa's approach to analysis, the unique con 6guration of the licensee trays, and the unique ampacity data available to the licennae J

l Findings and Recommendations: SNL finds that the concerns identiSed in this review i~ will have no impact on the findings and conclusions doaunanted by SNL for the G43 j calculations. No actions in this regard are recommended.

i 4

l i

i

l. .

i 8

! See, for example, SNL I.etter Report of 12/20/96, "A Supplemental Review of 1

the Braidwood Station Response to the USNRC RAI of11/2/95 on Fire Barrier Ampacity

Derating," prepared under JCN J-2017, Task Order 6.

I

.; 21

)

4.0

SUMMARY

OF REVIEW FINDINGS AND RECOMMENDATIONS .

4.1 Overview .. ..

The calculations presented in BYR96 082/BRW-96-194 represent a significant expansion J.1 ,

in scope of the overall package of ficensee calculations as compared to the original package reviewed by SNL in August 1995. 'Ibese calculations cover a broad range of

.*j "special instaHations" involving both single tray inentiations and ine=11=tions with more .

, than one tray in a single protective envelope. 'l:

7:.

In general the thennal model developed by the licensee for these analyses is weH b:?!

documented, clearly defined, and based on modern and accepted thermal M-h tools. . .:t  :

Aspects of the licen=* approach to modeling are unique in SNL's experience, but overaH ' ::!;

the approach is ae jtable in principle. While SNL considered that the model does 23 include certain non-conservative assumptions, it was also noted that the model includes :s!

several other offsetting conservative assumptions. In the kahaea, it is expected that once 4l.~

the points of concern identified in Section 4.2 are resolved, the licensee analyses will '::' .

provide for a conservative atte<< ment of the ==a=W derating impact for these ,,?

?)

configurations. ' f.

y.:

4.2 Fammining Points ofConcern

$f As a resuh of this review SNL has identified four points ofconcern. It is recommended .' )

that the licensee be asked to resolve these concerns through the RAI process. Of the , ji, identified concerns, one is considered far and away the most significant. This is: . .::

.&a SNL finds that the licensee submittal is, in effect, comparing " apples and  :;.j(

oranges." In particular, the licensee assessments of base line ampacity limits .

derive from a licensee table of allowable heat intensity limits while the clad gj$

's.

case ampacities derive from the thermal model. This is considered by SNL to $$

be extremely poor practice, and hence, is na=ce-atable. SNL recommends that * ";

estimates of fire barrier ADF should universaDy be based on self consistent .'li; treatment of the clad and base line cases. In this casi, it is considered critical *:";

2 -

to assess both the clad and base line ampacity limits using a self-consistent thermal model, and the licensee has not done this. If the thermal model is used f$

-).' .

to predict the clad ama*eity limits, then a thermal model fully consistent with li!.:!.

the clad case analyses should also be used to assess the base line ampacity 'M limits as weH. It is recommended that the licensee be asked to implement a .T 1-thermal model for the analysis of the base line case ampacity that is fuBy lf.'

consistent with its clad case analyses, and to then base its final ====ei+y  :

derating assessments on a comparison of the clad and base line thermal analysis resuhs.

Y; :

Note that as a part of this review SNL did implement a base line thermal analysis model 7'g;  :.

fully consistent with the licensee's clad case thermal model. The SNL model is  ;;[:

documented in Appendix A, and has been discussed in Section 3.3 above. Results have

been presented for all of the cases considered in the licensee analysis (laeMag both the

(@

j ji.

Ub.

.22 h P..?

,.s - - -- . ., ,

i . .

. . u i l 1 1 sensitivity studies and the 14 specific case examples). The licensee's adopti:n of 6 .

analysis model for the base line case maneanmenta would fully resolve this point of concern.

i

l. M remaining three points ofconcern are all considered ofless significance, especially

assuming that the above item is resolved. These items are as follows- __,__

~

l' - M licensee has presented a table of heat intensity versus depth of fill values as '

! hesn 13 on page 13 of the Ecensee submittal. His table is in .pp a conflict

! with the heat intensity values ched by Stolpe and in the ICEA standard P l 440. h cited values appear to modestly over-state allowable best intaneity

limits, and hence, might lead to optiminic estimates of the cable ampacity

! limits. It is recommended that the licensee should be asked to establish the

! basis for how % heat intensity table was developed and how h is applied in l practice. It is also recommended that the licensee be asked to reassess its j ampacity limit calculations in light of & apparent disuWy. (A

! comparison oflicensee approach in comparison to a direct application of the l ICEA standard for a specific and well dermed case example would be very

helpful in this regard. Also, the licensee should be asked to provide the -

! supporting calculation cited in the study as the basis for this table (Calculation ESIl50-1, Rev. O, licensee ref.16).)

{

e The licensee cites in item 2 on page 12 that the base line ampacity for a 3/C, #6

AWG,600 V cable with a 2.5" depth of fill is 27.5 A. M basis for e value i is not established. SNL was unable to reproduce this limit using mandard j approaches to ampacity analysis given that the lice thermal model has cited j the ICEA definition as the basis for fill depth calculations. Two possible explanations were noted by SNL in this review, either ofwhich would appear to be inconsistent with the objectives of the analysis. It is recommended that j the licensee be asked describe in detail how & value was obtained and/or that the licensee be asked to delete references to and reliance upon 6 value as the .

• base line ====T for the cases examined.

Several of the licensee cal =1=Wmclude a reference to a "Sirremp Sheet" but the fire barrier descriptions do not include a di=~= ton of any such sheet used in the installation process. This reference may well be a " spurious" carry-over from a previous application of the thermal model. It is recommanded that the licensee be asked to clarify if and when such a material is used in its banier constructions. This is considered a minor po*mt that will not have a significant I impact on the final licensee assessments.

4.3 ImaIHiaas of the Current Finding for the G-63 Calculations SNL concludes that the Madinge of e current review will in no way impact the conclusions previously reached for the G-63 calculations. The basis of the G-63 calculations is significantly different from that used in the BYR96 092/BRW 96-194 calculations. Hence, the observations and inconsistencies noted here will not impact these l other calculations. No actions in % regard are recommended. i il l

!  : AppendrA:

A reproducGon of he Commonweakh Edson Co. metod of celaide6on for omble tot heat tansfw and barrier empecity dersung analynn. s l I l

This is besloety a reproduedon of the calcuistion methods of caladston G43 Rev. 4, and of BYR -

96 082/BRN-96194 in het al of the basic modeing eseumptes have been implemented se defined "

by he usty. This hdudes: +-

. uniform surface temperature for he ceW mass hdudng the same temperature assumed for 7, '

tie M heigit of he ende rets, and for he bottom of the soEnottom anble Way,

- same size cable: 3/C 8AWG 600V with demeter of 0.953'

-tmiform heetmg h cable mass .

- Stolpe/Holmann simpB6ed treatment of cable mass hiemel heat kansfer .

- Holmann oorrelations for oorwective heat tansfer -

- M credt for redonon/corweetion from al surfaces "' -

For tss seiend Gon, SNL has implemented a Bees une ones analyels of an iswied key My consistent 2 with the correspondng dad tray analyses pursued by he utBty in he ceicadations, especieIy toes of :jy:

gyR.06482f '

?!

This case is for a 24*x4' cable toy,2.5" dept of 5, no tre wrap, open top, said bottom.

s To start, must define a fundamental temperature unit (older version of MATHCAD). As per MathCed standard approach, we use the otherwN unused coidomb charge units and redefine his a one uriit of absolute KeMn temperature:

K := I soul CoK := 273.16 K '

i Now set some Axed physical parameters:

s ' . .. l:.

The Cables: d,,g 0.953 is aened := 3 Resbis > 0.000513 ..

i The Tray: ;p!:

w e,y := 24 is .h ,y = 4 ia dA B = 2.5 ia ye[<f. ,

A The cable mese: p.,g := 13.12 K cable mass hermal resistMty h C-ftMI as 'l.i pw @ 400 M

'S .:

The a' pproach is to set en octual empeelty, ceiciAste temperature and heat Gow values by a hermal '*

model, and noode on empecity unti e cable hot spot of 90C is obtained. This is where the Boonsee hermal model comes into pisy. SN1. has implemented his model My consistent with tie Bosneee .

,.g'.  :

i implemente6on and assump6cns, but has not included any Are benius h the analysis. *

. .l;

,':i ,h, .

Need to set some constants, cable properties and metal propertises as per tiese used in gYR96092I i?  :*

s = 5.669110 wett Stephen 4oltzmann

... e,i m'.K*  !?

.m a ,,g := 0.95 Thisis emble ammioMty ..

a ,,,,g := 033 This is the metal emmioMty, very low value assumes by CE ,.

A amble top "

• esy Asids:"2'htoy 05i P

Heat tensfer surface areas 7/;i A ,i,,1 = wany + 2 hg ,y A g boet

• may

• II Tamb " 40K + QoK Tamb =313.16 K amblert tempis 40C: 4'4 24 :n

S

?.T

• . W

. , +

Step 1: Set he empacity and oaksdate intemel oeble homeng rate:

I = 32.58 amp Start with a guess then iterate to get the right empecity NOODLE HERE TO GET 90C CABLE TEMP.

A ,,,,= wg,yd g ,

~

am= d g, m ,y,=66.064 Q ,,g ,=a,,g,a.,,gf R,,g, Q,,y,-107.921 wan ,

Step 2: Caloidste Temperature of he outside surfeos of cable mass by balancing he intemel heating rate calcidated above to be extemel heat losses by radiation and oorwecton. (Recal het top of cobie mass, side rals, and bottom panel on tray al assumed to be at same terrhe):

Seed he awface temperature value for root Gnder below:

Tsud= 32&K Recal that we mesume reis and tray bottom are al et some temp as omble mass surface 1 so een treat radiation to ambient as single equaton:  !

Qad(T.e6 -(Ts ./-T.mb ) ;(e.w^ si) + (e.sa. A e _ ,):

Conwtlen Formides:

eldes:

hga,(T.,f := 1.42 wou fd2 o --

1 ITsud- Tamb I

al K A2 i h g,,y j 1

Qeauv , side (T ,f := kgde(T ,f Agde-(T.g-Tamb)

Tsud- TembI f f hg,p(T ,f := 1.32 wu dl.2s - l adK(Ki (

• tsy j j 1

l 9 eenv_ top (T ,) = h g,p(T ,3 Asep_ bon-(T.,g- Tasnb) l Boeom: ,

fd .ns fT,,g- TunbI l h ban (T,,) := 0.61 won - l -

adKk4 K

• ksy j l

{

l QeenvboaU""O=bbon(T,,fAtopbad(Tsud- Tamb)

Total HeatTransfer. .

Q =n.i(r ,) = q 4(r ,3 + o,,,, ,;4,(T,,3 + o,,,, ,,,(T ,f . q ,,,y 3,,(T,,f , , , l l

l

I RowlowVebes:

.~.

l Q,e,si(rse ,6 =n9x , Q,ed(r.e3 = nm =su i

_ h,;d,(T.,) =.4.068 won Qeaw- side (T ,f =1.723 won .v :.

._. _ /.g _ a --

.:,7

, .u :

won

• hanp@se,f =2.416 Q,,, ,,p(r,,,) =3A7 =en ..

,g _ ,

\

won kban(r.,f=1.116 Qeen_ bon (r,,,)=1.419==u -

Nh Use ROOT to solve for metace temperatas et which total heet trenefer aquels heet generaton h anbles: 5.(' .

s; T ,,,r = not (Q ,,,,3 (T ,,,) - Q embha ,T,,,)

l .

\

T,,g=344.713 K T ,r-CaoK=71.553 eusparesunesse ". ';' ,.

l

.\

I '

Step 3: Now we have surfeos temp, need to analyze hiemel behavior of tw cable mass to get tw hot l spot temperature. As per tw Boonnee, we use tw simp 8Aed model of Stolpe/Holmerm: ..

Q embise*Pemble d g ,,"

Tomble" +T surf r

g'* tray .

l 38 1 Tenbis=363.15 K Tambie - CacK =39.99 temperan=p=== 'dy.. l t ,

• l This is the cable hot 4 pot temperature h degrees C, need to EdcBe value unti about 90:

j:[::.

Recalcurrent: .

y?.

1=32.ss amp ,,*

Final Reedt of Calculation: Uebg the unty modeEng assump6ans and correlatons, to bene Erw merent = eas,neied so2.m Tw. m h vory e ..,oeme,.w.i #w CEA vobes obtehod e.iw {'.$

tom #m tables temselves 91.77) er kom em cEA heet beenalty values Q2.44). k is, however, ': :

eigni5ceney largw that tw values eked m #w Boonoce colodeten (27.5A). The impaa on #w  ;;;.0 i

reasano empeeny dwanno femore could te wwy aimiscent. ..

• • • !'=..

i . Gil nW n

l  :::~

r..

. t. :*

l:l 5d b.5.

\

l

' 26

Y[I y;

i.

. h .

.:n

., . -m... .~--e.e>= w ow-