ML20027C647

Forwards Engineering Planning & Mgt,Inc Analysis of NRC Sponsored Fire Tests Evaluating 20-ft Separation Criteria

ML20027C647
Person / Time
Issue date:	09/29/1982
From:	Notley D NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
To:	Knighton G Office of Nuclear Reactor Regulation
References
NUDOCS 8210260554
Download: ML20027C647 (67)

Text

fb pa afc UNITED STATES oq(o, NUCLEAR REGULATORY COMMISSION A

o WASHINGTON, D. C. 20555 I!4. n S

E

%,,,,,*l SEP 2 9 eq.p~.,, # E 41 Gag g

7 1982 N

MEMORANDUM FOR: George W. Knighton, Chief Research and Standards Coordination Branch Division of Safety Technology, NRR FROM:

David P. Notley, Fire Protection Engineer Electrical Engineering Branch Division of Engineering Technology, RES

SUBJECT:

ENGINEERING PLANNING AND MANAGEMENT, INC. (EPM)

ANALYSIS OF NRC-SPONSORED FIRE TESTS EVALUATING THE 20-F00T SEPARATION CRITERIA Early this year while we (RES and NRR) were involved in the final plans and preparations with Sandia for conducting the 20-foot separation fire tests at Underwriters Laboratories (UL), Jim Evans, Edison Electric Institute (EEI) liaison for the EEI Fire Protection Committee, expressed a desire to participate to the fullest extent possible. One issue, particularly, seemed appropriate for EEI participation.

I suggested to Jim Evans that if they could accurately predict the course of the six full-scale tests we planned to run at'UL, many utilities would sub-stantially improve the credibility of various requests for exemption from specific requirements of Section III.G of Appendix R to 10 CRF 50 that are based upon calculations. Mr. Evans agreed with this suggestion and EEI contracted with EPM to perform such an evaluation. The enclosed report is the result of that contract. I received it on Monday, September 27, 1982 You will note that the enclosure starts with Section 3.4.3 Mr. Evans stated to me that the portions of the report that he did not send me contain information that is not pertinent to the original question, and that this represents all of the report that is of interest to NRC.

I am sending this for your information and use.

I gave a copy of the report to Bob Ferguson the same day I received it.

~

David P. Notley Fire Protection Engineer Electrical Engineering Branch Division of Engineering Technology Office of Nuclear Regulatory Research

Enclosure:

As stated cc: See attached list 0210260354 820929 PDR MISC PDR

Memo For:

G. W. Knighton dated SEP 2 9 ;gg, cc: w/ encl R. H. Vollmer, NRR P. C. Cota, NRR W. V. Johnston, NRR V. Benaroya, NRR R. L. Ferguson, NRR R. Eberly, NRR D. J. Kubicki, NRR J. F. Stang, NRR F. J. Nolan, NRR A. P. Parr, NRR V. W. Panciera, NRR J. M. Taylor, IE J. C. Stone, IE J. L. Boccio, BNL L. J. Klamerus SNL W. A. Von Riesemann, SNL LPORe s

4 i

1

v a

i 4

i a

f E

4 3

343 Preliminary Fire Experiments In Appendix III of the Sandia test program

plan, a

10-

gallon, 25 square foot heptane pool fire'is analyzed using the Harvard Fire Code (1981).

The purpose of the calculation is not clearly d e fin ed beyond that o f providing "useful input to the

(

test plan".

In this context, this subsection reviews the infor-

\\

mation described in Appendix III with two objectives in mind:

(1)

Attempt to reproduce the results of the analysis using alternative methods; (2)

Relate the results obtained through various means with the objectives and procedures of the-test plan itself.

The postulated fire analyzed by Sandia is assumed to occur within a compartment 30 feet long, 14 feet wide, and 10 feet high.

Two trays containing an unspecified quantity and type of electrical cable are placed within the room such that the edge of the 10-gallon spill is 20 feet away from the trays.

The top tray is placed "just below the ceiling" while the bottom tray is located 3 feet above the floor.

With this configuration, Sandia utilized the Harvard Fire Code (HFC) to predict the following results :

Fire Duration:

90 seconds

" Hot Layer" 0

0 Temper atur e :

1000 K (1340 F)-extinction 400 K (260 F)-200 seconds follow-ing extinction

(

Critique of.the Test Program Pag e.:3,-20

(

Radiative Flux to the Cables:

37 kW/m2 (peak) 2

>20 kW/m for approximately 40 seconds Cable temperature (Radiative Effects only) :

690 K (780 F) d 0

EPM's review of 'this data and Appendix III of the Sandia

plan, in general, was perhaps the most difficult portion of the critique.

Since the results of the analysis were presented without any discussion..of the assumptions and their

basis, the task of reproducing the results assumed onerous proportions.

Rather than endeavoring to discover and follow the path o f analy-sis y, sed by Sandia, it was decided to' relate Sandia's results to those produced by EPM's DRAGON series of computer models.

In performing analysis of the effects of exposure fires, EPM employs a series of computer models referred to as DRAGON.

These models allow the analyst to bound the anticipated thermal condi-tions resulting from an exposure fire.

If an energy-based fail-ure criteria is specified, DRAGON can perform a variety of "back

(

calculations" to define the worst case fire in terms of minimum fuel volume or quantity.

Alternatively, a " forward calculation" may be utilized to describe conditions resulting from a postula-ted fire.

DRAGON also provides the option of either yielding a "best estimate" calculation or a " licensing" calculation.

This is accomplished through the reliance on empirically-based correla-tions which provide successively more limiting results.

The l

differences in the results reflect either actual results of l

experiments

("best estimate") or bounding correlations which j

suggest far more severe conditions than that actually cbtained.

All correlations utilized in the DRAGON series rely upon actual data reported in fundamental papers widely-disseminated and com-mented on in the combustion literature.

l The appendices to this report provide a description of the a ss um ptio ns,

considerations, and empirical correlations which fo rm the basis for the EPM DRAGON models used to bound the e f-fec ts of exposure fires.

Topics covered in the appendices in-'

clude the basis for heat release rates from the fuel of interest, the e ffects o f radiation, ventilation, plume impingement, strati-fled ceiling jets and wall / corner effects.

For the configuration described in the Sandia test plan, the 4

dominant source of energy to the cables is the stratified ceiling k

layer.

Focusing on this aspect, the principal difference between

]

the

" licensing" calculation and the "best estimate" calculation s

=

Critiqua of tha Tost Program Pego 3-21

(

is the model used to describe this phenomenon.

The " licensing" calculation (DRAGON (LC)] relies upon data obtained by Newman and Hill (1981) at Factory Mutual Resesrch Corporation (FMRC).

Al-though Newman provides a correlation for his data, EPM developed a

more bounding correlation under all ventilation conditions.

This model is the one used by EPM in all of its analyses of plant fire hazards in support of requests for exemption from 10CFR50 Appendix R.

The "best estimate" calculation (DRAGON (BE)] relies upon the same plume model, but more closely models the effects of turbulent ceiling Jets.

The principal source is due to empirical work per formed by R.

L.

Alpert (1973),

also of Factory Mutual Research Corporation.

In reviewing the configuration modeled by Sandia, the fail-ure criteria employed in EPM's DRAGON codes is that calculated in this analysis for PE/PVC cable undergoing electrical failur e.

That correlation was obtained using data provided by the Sandia Radiant Heat Facility, although it should be emphasized that an earlier analysis in Section 3 4.1 indicated a wide variance in the oprrelation.

The results of that analysis yield the follow-ing criteria:

2 Critical Heat Flux:

8.03 kW/m Critical Energy for 2

Electrical Failure:

5388 kJ/m

(

Other assum stions used in this analysis include:

Heat Release Rate for 2

Heptane:

2730 kW/m Cable Height :

108 in.

Ceiling Height :

120 in.

Fuel Mass Loss Rate 2

(Steady State):

70 g/m -s Using this data, the results of DRAGON (LC) are as follows:

Maximum Heat Flux (Con-2 vective and Radiativ e) :

17,8.2 kW/m Burn Duration :

160 3 seconds Time of Cable Failure:

85 4 seconds In the "best estimate" calculation using DRAGON (BE), a more analytical approach is followed.

While less constraining than DRAGCN (LC),

many aspects o f the " licensing" calculations are k

retained.

For example, wall effects are considered using symme-try conditions.

The observation of the effects of radiation is l

e Critiqus of tha Tost Program Pago 3-22

(

ignored, allowing for maximum impact.

The convective heat flux maximized throughout the analygis by assuming the cable main-is A burn duration of 160 3 tains a constant temperature of 70 F.

seconds is predicted using the same assumptions as used in DRAGON (LC).

The results of the DRAGON (BE) analysis are as follows:

Normalized Heat Flux (Con-2 vective and Radiative):

19 09 kW/m Burn Duration:

160 3 seconds Time of Cable Failure:

Not predicted Table 3-1 presents a summary of the HFC and DRAGON code results.

A review of those results lead to the following conclu-sions:

,(1)

The results-o f DRAGON (BE) and HFC produce comparable results in terms of incident heat flux to a

cable, although the DRAGON series predicts longer exposure duration.

(2)

Only DRAGON (LC) predicts cable failure using a d amage criteria based on the Sandia data.

(

(3)

All three models indicate that the cables would be exposed to a heat flux greater than that allowed by the

" acceptance criterion" for the full scale tests.

e

[ epm

m Q

r TABLE 3-1 0

.g Results of Analysis of Preliminary Fire Experiments (10-Gallon lieptane Fire /25 Ft.2 pggy) e i

2u" EPM EPM SANDIA DRAGON (LC)

DRAGON (BE)

IlFC q

g 2

2 l

Unknown 2

i Fuel lleat Release Rate

!2730.

kW/m l 2730 kW/m I

I e

(Steady State) l l

1 1

I I

I I

I I

Fire Duration l

160,3 seconds l

160.3 secondsj 90 seconds I

i i

a i

i

!178.2 kW/m2 (peak) l l

37 kW/m2 (peak) 2 lleat Flux 19.09 kW/m l

l l

(normalized)l 1

1 I

PE/PVC Cable Failure l

Yes No No Predicted j

j j

l i

1 I

i l

i l

I Exceed Test Plan Acceptance j l

Yes Yes Yes Criteria j

I I

I I

I I

I I

I i

l N

W m

e O

Critique of the Test Program Page 3-24

(

3 4.4 Equivalency of Heptane and Cable Insulation Heat Release Appendix IV of the Sandia program plan attempts to establish an equivalency between heptane and cable insulation heat release rates. The objective of this analysis is to determine the quanti-ty of heptane to be modeled in the full scale fire tests to represent the combustion of five vertical trays of electrical cable.

The PE/PVC cables used in an earlier fire test ( fire test 76) were apparently selected for their high heat release rates.

On the basis of an HFC analysis, Sandia determined that a

30 minute fire involving five vertical cable trays is equivalent to a

16-the combustion of 25 4 gallons of heptane consumed over minute period in a 2.8 ft area.

The focus of EPM's review of Appendix IV is in the following areas:

(1)

Comparing Sandia's heat release data with results presented in the literature; a

(2)

Verifying the parameters of the

" equivalent" heptane fire; (3)

On the basis of (1) and (2),

determining the validity of the assumption of equivalency.

(

In revicwing Sandia's bounding measurement of the heat release rate from burning cable, it was apparent that fire test 76 was per formed absent knowledge of a much more comprehensive and controlled series of experiments performed at Factory Mutual Research Co rpo ration and reported by Tewarson et al.

(1979).

Rather than reviewing the original material, Sandia appears to have relied solely upon an oral presentation of a summary of that work.

While the heat of combustion ultimately obtained for PE/PVC cable by Sandia is approximately

correct, the expense associated with fire test 76 may have been saved through a more careful review of the available literature.

This point will be developed further in this section.

In addition to understanding the flammability properties of cable insulation, a literature review would also be helpful in q"entifying the heat release rate of heptane.

Ab sent that

review, it appears that the Sandia analysis may have overlooked i

an important phenomenon.

While the total energy released in the l

combustion of 25.4 gallons of heptane assumed by Sandia may be somewhat comparable under certain circumstances to the combustion l

of the five vertical cable trays in fire test 76, the rate of l

energy release is significantly affected by the stoichiometric fuel-air ratio which is itself affected by the fire diameter and the fluid dynamics of the diffusion flame.

Critique of the Tost Program Page 3-25

(

The Sandia test plan reports the pyrolysig rate for a fully on HFC

. developed heptane pool fire to be 0.053 lb/ft -s based analysis.

This rate is equal to 258.8 g/m -s.

Such a

rate conflicgs with empirical data suggesting an asymptotic limit of 70 g/m -s for heptane fires discussed in a Department of Trans-portation report by Tewarson et al.

(1979). Both Tewarson (19801 and a Brookhaven report by Pgnkel (1978) cite the same asymptotic limit (in this case, 50 g/m -s) in reproducing the results of an earlier work by Blinov and Khudiakov (1961).

The empirical results of Blinov and Khudiakov (1961) further suggest that the high mass loss rates suggested by the HFC analysis would occur only under laminar flow conditions produced in pan fires under 1

inch in diameter.

Un fo rtun a tel y, Klamerus et al. fail to identi-i fy the assumptions used in the HFC analysis, making it impossible to identify the cause of the predicted high mass loss rate.

Nevertheless, Sandia states:

This analysis does not consider any size effect of the spill area or the length of burn.

Klamerus et al. (1982), page IV-2 On this

basis, it must be inferred that Sandia carries the lamin ar conditions over into the turbulent regime leading to the

(

erroneous mass loss rate used in the analysis.

While Appendix IV concludes with a comment that the HFC analysis was merely of a scoping nature, the erroneous use of the heptane mass loss rate highlights the importance of a number of issues:

l (1)

It is important that a complete review of the litera-ture be per formed prior to initiating any research I

project.

Had such a study been performed by

Sandia, the sensitivity of the heptane mass loss rate to the spill area would have been recognized early in the process and properly addressed.

l (2)

The difficulty in relating the combustion characteris-tics of different materials should be recognized prior to attempting to establish equivalency.

A review of Tewarson (1979) would have highlighted the fact that different PE/ PVC cables exhibit different mass loss rates under the same conditions.

Hence, although know-ing the heat of combustion may describe the total energy

released, it provides little information for sizing " equivalent" heptane pan fires.

Again, a more careful literature review prior to fire test 76 would have been helpful.

(

l l

QT1 l

. Critiquo o f the Tost Program Pego 3-26 9

(

345 Analysis of Full Scale Fire Tests

~

The objective of the Sandia test program as stated in Klam-erus et al.

(1982) is to evaluate the adequacy of twenty feet of horizontal separation alone in protecting the safe shutdown capa-bility from the effects of small exposure fires.

In this re-spect, two full scale tests are discussed.

The first will employ PE/ PVC cable in a small enclosure fire while the second will utilize XPE/XPE cables qualified to IEEE-383 An

" acceptance criterion" is specified which includes a more limited heat flux criterion in addition to general phenomenological standards of ignition and cable failure.

Section 343 discusses the use of the DRAGON codes

and, especially, DRAGON (LC) in modeling the effects of exposure fires in support of requests for exemption from Appendix R.

It is shown in that section that DRAGON (LC) indicates conditions more severe than predicted by the Harvard Fire Code.

Both codes predicted failure according to the stringent " acceptance criter-ion", o f the full scale tests.

Section 3 4.4 illustrates the importance of assumptions concerning fuel flammability in model-ing pool fires.

With this information in

mind, this section utilizes such assumptions and considerations for heptane in con-junction with the DRAGON (LC) fire model to bound the thermal conditions anticipated to be present during the full scale tests.

DRAGON (BE) is also used to generate a more accurate description

(-

of actual test conditions.

In the DRAGgN(LC) analysis, five gallons of heptane is consumed in a 5 ft area.

The implication o f these conditions is that steady state fuel mass loss rates will be achieved under turbulent conditions.

On this bas the heat release rate fo r is assumed to be 2730 kW/m'ts,based on Tewarson's work on heptane behalf of the Department of Transportation (Tewarson et al.,

1979).

As in Section 3 4 3, a ceiling height of 10 feet is used

(

with cable heights of 9 feet.

The failure criteria used in Section 3 4 3-are carried forward into this analysis.

l On the basis o f such ' assumptions, DRAGON (LC) predicts a

l burn duration of 400 98 seconds with a peak heat flux of 35.63 2

kW/m.

Using the Sandia-based failure criteria,

PE/PVC cable failure (electrical failure) is achieved at 152.22 seconds into the fire.

An analysis using DRAGON (BE) under similag circum-stances calculates a normalized heat flux of 10 7 kW/m over the same fire duration.

PE/PVC cable failure is not achieved in this scenario.

These results suggest different thermal conditions at the l

location of the PE/PVC cables.

While DRAGON (LC) suggests fail-I ure and DRAGON (BE) only indicates cable

damage, both models indicate thermal conditions at the location of the cable trays

,(

which would exceed the " acceptance criterion" of the full scale W

l i

t

I Critique of tho Test Program Page 3-27 3

)

. tests for PE/PVC cables.

(DRAGON (BE)'s results would be marg-(

inal for IEEE-383 cable.)

Recognizing the likelihood that the actual tests would produce results similar to the DRAGON analy-sis, several issues are raised:

(1)

Does HFC analysis exist for the full scale test config-uration and, if it does, is the test " acceptance cri-terion" exceeded?

(2)

If HFC analysis indicates that the test will exceed the

" acceptance criterion" as does the DRAGON analysis, is the objective of the tests affected?

(3)

If both tests and diverse methods of analysis indicate that the " acceptance criterion" is exceeded, what con-clusions are to be drawn from the exercise?

Issues (1) and (2) relate to the goals and objectives of the tests which are essentially managerial and previously addressed in Section 3 3 Issue (3), however, assumes aspects of a tech-nical issue in that the conclusions derived from the tests must hav e a

technical basis in the assumptions and analysis.

To a

great extent the results upon which the conclusions are to be dra wn are dependent upon the validity of the postulated "accep-tance criterion".

Ho wev er,

as important as the criteria for acceptance is, of equal importance is the relationship of the test configuration to actual power plant situations.

In both

(

cases, the basis offered by Sandia appears weak.

The Sandia data upon which the DRAGON analysis relied has been evaluated in this report using standard statistical methods.

It was shown that additional data must be obtained in order to support a conclusion regarding the threshold for failure.

For comparison, EPRI-funded research reported by Lee (1981) and re-lying upon more data points distributed over a wider range of heat flux indicates that PE/PVC cable may be more resistant to electrical failure than t.he limited Sandia data may suggest.

This consideration highlights the need for Sandia to resolve the issue of data at the earliest opportunity if it is to be used in the test program.

Howev er the data issue is resolved,

though, it cannot be emphasized enough that the Sandia test configuration fails to l

reflect actual plant conditions.

The low ceiling heights used in the test which accentuate the effects of ceiling stratification are not commonly found in general plant areas in nuclear power 1,

plants containing safe shutdown systems, Recognition of this fact is reflected in tP.e original NRC Staff recommendation to the Commission in providing for the use of horizontal separation, automatic detection and automatic suppression as a

means of protecting the safe shutdown capability.

This issue should also be carefully considered in the test program.

A i

~. _

6--

_Em a_.-

e G

O 9

I Z

O en D

JO Z

O O

)

m

4.0 CONCLUSION

S

(

This report has reviewed the Sandia test plan from the perspective of its relationship to that of furthering the licen-sing process.

In the course of this review, a number of mana-gerial and technical issues were identified.

The comments pre-sented on these issues were directed towards highlighting those areas which may detract from the overall value of the research.

It is hoped that the comments presented herein would be viewed as constructive, relative to achieving the objectives of regulatory research and, in so doing, lead to the development of regulations which protect the public health and safety.

i l

{

i i

(

epm 4

(

REFERENCES (1)

L.J.

Klamerus, L.L.

Lukens, D.D.

Cline, and D.A.

Dube,

" Program Plan for Ev aluation.o f Twenty Foot Separation Distance", Sandia National Laboratories, February 17, 1982.

(2)

A.

Tewarson,

" Ex perimen tal Ev aluation of Flammability Parameters of Pol ymeric Materials",

FMRC J.

1.1A6R1.

Prepared for Products Research Committee, National Bureau of Standards, by Factory Mutual Research Corporation under Grant No. RP-75-1-33A, Norwood, MA, February, 1979 (3)

A.

Tewarson, J.L.

Lee, and R. F. Pion, "The Influence o f Oxygen Concentration on Fuel Parameters for Fire Modeling",

Eighteenth Symposium (International) on Combustion, The Combustion Institute, 1981.

' The Connecticut Light and Power Company et al.

v.

Nuclear (4)

Regulatory Commission", 2d Cir (1982).

(5)

A.

Tewarson and R.F.

Pion, " A Laboratory Scale Test Method for the Measurement of Flammability Parameters",

FMRC No.

22524.

Prepared for Products Research Committee by Factory Mutual Research Corporation,

under Grant No.

RP-75-1-33A,

(

Norwood, M A, Oc tob er,

1977 (6)

A.

Tewarson, J.L.

Lee, and R.F. Pion, " Categorization of Cable Flammability Part 1:

Laboratory Evaluation of Cable Flammability Parameters",

NP-1200, Part 1, Elec tric Power Research Institute, Palo Alto, CA, October, 1979 (7)

M.A.

Delic ha t sio s,

" Categorization of Cable Flammability -

Detection of Smoldering and Flaming Cable Fires",

NP-1630, Electric-Power Research Institute,

Palo Alto, CA, Nov emb er,

1980.

(8)

J.S.

Ne wman and J.P.

Hill,

" Assessment of Exposure Fire Hazards to Cable Trays",

NP-1675, Electric Power Re se arch Institute, January, 1981.

(9)

A.T. Modak, "Ignitability o f High Fire Point Liquid Spills",

NP-1731, Electric Power Research Institute,

Palo Alto, CA, March, - 19 81.

(10) J.L.

. Lee, " A Study of Damageability of Electrical Cables in Simulated Fire Environments",

NP-1767, Electric Po wer Research Institute, Palo Alto, CA, March, 1981.

(11)

A.

Tewarson,

" Fi r e Hazard Evaluation of Mine Materials",

Technical Report RC80-T-77, Factory Mutual Research

(

Corporation, Norwood, MA, October, 1980.

O ep01

=

o Re ferenc es Page R-2

(

(12) A.

Tewarson, J.L.

Lee, and R.F.

Pion, " Fire Behavior o f Tr ans fo rmer Dielectric Insulating Fluids",

DOT-TSC-1703, Transportation Systems Center, Department of Transportation, Cambridge, MA, 1979 (13) A.

Tewarson,

" Heat Release Rate in Fires",

Fire and Materials, Vol. 4 (4): 185-191, 1980.

(14) I.I.

Pinkel, Estimating Fire Hazards Within Enclosed Structures As Related to Nuclear Power Stations", BNL-NUREG-

23892, Brookhaven National Laboratory, Upton, NY, January, 1978 (15) V.I.

Blinov and G.N.

Khudiakov,

" Diffusion Burning of Liquids",

Moscow Academy of Sciences (1961).

Translated by Research Information Service, U.S.

Army Engineer Research and Development Laboratory, Fort Belvoir, VA, 1961.

(16) *H.E.

Meitler and H.W.

Emmons, " Documentation for CFC V, The Fifth Harvard Fire Code", Home Fire Project Technical Report No. 45, Harvard University, Cambridge, MA, October, 1981.

(17) R.L.

Albert,

" Turbulent Ceiling Jet Induced by Large Scale Fires",

Combustion Science and Technology, Vol.

--11: 197-221, 1973

(

(18) Office of Inspection and En fo rc emen t, " Report on a Survey by Senior NRC Management to Obtain Viewpoints on the Sa fety Lapac t of Regulatory Activities from Representative l

Utilities Operating and Constructing Nuclear Power Plants",

l U.S.

Nuclear Regulatory Commission, Washington,

D.C.,

July, 1981.

4 k

t

epn,

O 9

e O

e LLI g

OZ u.I Q.

Q.

<C

.)

= +

(

APPENDIX A Basis for Heat Release Rates This appendix provides the basis for the fuel heat release q

rates utilized in fire models described in this analysis.

The quantities reported herein and the underlying concepts are from the combustion literature and reflect the current state of know-ledge in the fire sciences.

In areas of uncertainty, conserva-tive assumptions are made so as to ensure that the integrity of the adalytical method is maintained.

The heat release rate associated with a fire is related to the fuel's mass loss rate (pyrolysis) and the heat of combustion

[Tewarson (1)] by the following relationship:

(

H QT*

"b T t tal theoretical heat release rate where QT=

99 i

s

=

mass loss rate in burning b

i t tal theoretical heat of combustion HT=

(

epmi>

Appendix A, Basis for Heat Release Rates A-2

(

The mass loss rate itself is a variable which in a realistic cense is dependent upon multiple factors such as fir e

stage, gaseous temperature and fuel type.

In general, the mass loss rate may be described by the net heat flux delivered to the fuel's surface and its heat of gasification.

.n 9n

.mb=

g where q" =

net heat flux received n

by the fuel

(

heat required to generate L

=

a unit mass of fuel vapors The dependency of the mass loss rate on the net heat flux delivered to the fuel surface and the associated feedback effects illustrates the historical difficulty of analysts to derive a

meaning ful and precise model of flame behavior.

The net heat flux itself represents a heat balance at the fuel surface and is given as the difference between the total heat flux received by the fuel and that flux lost through a variety of processes.

This balance under steady state conditions may be modified,

however,

Appendix A, Basis for Heat Release Rates A-3

(

by such factors as the relative concentration of oxygen entrained in the combustion zone, the externally applied heat flux and the optical path length of the gases.

The principal effect of these considerations becomes evident in the actual heat of combustion which reflects different oxidation reactions.

At a detailed level these multiple parameters are all inter-related.

However, it is possible to select a single parameter for the purpose of illustrating the sensitivity of the heat releas,e rate to its variation.

That single parameter would be the fraction of stoichiometric oxygen to fuel ratio given by:

A" ( a) o 2

o

=

,,, K

(

• b 0 2 where 6 fraction of stoichiometric oxygen

=

to fuel ratio fraction of oxygen entrained a

=

in combustion stoichiometric mass oxygen K

=

2 to fuel ratio mass flow rate of oxygen to M

=

02 fire vicinity k

epm

Appendix A, Basis for Heat Release Rates A-4

(

The effect of variation of this parameter on combustion may be illustrated for the case of pol ymethylmethac rylate over a

range of values of the stoichiometric oxygen / fuel fraction:

Fuel Chemical Combustion HA Condition Reactions Efficiency (kJ/g)

> 1. 0 Lean CH0

+ 60 + SC0

+ 4H O 100 24.9 5s2 2

2 2

0.81 Lean CH0

+ 4.90

+ 4C0

+ 35H O 80 19.9 sa2 2

2 2

0.63 Rich CH0

+ 3.80

+ 4C0

+ 3.5H O 60 14.9 sa2 2

2 2

+ 0.25C0

+ 0.25CH

+ 0.750 2

g 0.42 Rich C H,0

+ 2.50

• 200

+ 2H O 35 8.7 5

2 2

2 2

+ CO + CHg +C As may be evident from this table, oxygen and combustion efficiency have a significant effect on the overall heat release rate.

Moreover, it should be noted that lower combustion effi-l ciencies produce increasing amounts of carbon which lead to higher smoke rates, lower optical transmission path lengths, and l

higher soot concentrations, thereby reducing even further the effect of the released heat on a target material.

a.

^

Appendix A, Basis for Heat Release Rates A-5

(

The stoichiometric oxygen / fuel fraction affects heat release rates through its influence on the value of X

in the standard equation:

i (@)[ Hj k,,

h 7

Qi=X

\\L n

fraction of total theoretical heat where X i = release rate associated with mode i This equation and the in fluence o f X

on its results is the fundamental relationship for bounding the rate energy is released in a fire.

The remainder of this appendix will focus on each of

(

the following three elements in developing an appropriate rate for the fuels used in this analysis:

(1)

X fraction of energy i

released in mode i 4n (2) m fuel mass loss rate b*

L (3)

H actual heat of combustion A

The objective of the discussion will be to provide a scientific basis fo r selecting bounding values fo r each parameter in

(

subsequent analyses.

J i

Appendix A, Basis for Heat Release Rates A-6

(

The close relationship between parameters and the associated feedback effects was presented earlier in this appendix where the inherent difficulties in precisely modeling fires was demonstrated.

Ideally, if bounding values for X i,m and HA could be selected, then one may be assured that the heat release-rate is adequately bound through the assumption of a

suitably intense fire.

In order to achieve this goal, it is important to relate the three parameters of interest to experimental data and sensitivities.

Fo r the purpose of illustrating a general con-cept,,the case of acetone will be discussed beginning with the mass loss rate.

The mass loss rate for a liquid hydrocarbon was previously

(

given by:

q t.

[from Tewarson (1)]

mb L

gg

,ff

, il

, if

, it

, 19 where q,n

• 9

+Rfr + Rfc + 9o -El e

q"

= external heat flux incident on the fuel 4

= flame radiative heat flux incident on f#

the fuel "qfc = flame convective heat flux incident on the fuel l

l q"

= other heat flux incident on the fuel q"

= heat flux lost

(

l epm l

Appendix A, Basis for Heat Release Rates A-7

(

,o dr while for larger fires, qfp >> qfc For small fires, qfe >>

f where turbulent effects are dominant.

In the region where radiative heat flux to the fuel's sur-face is significant, it has been found on the basis of experimentation that all important parameters are independent of ox yg en concentratiop

[Tewarson (4)].

The affected parameters include:

(1)

Those parameters with slight oxygen dependency Actcal heat of combustion (H )

A CO2 yield (YC0 )

2 (2)

Those parameters which decrease with increasing oxygen concentration I

Convective heat of combustion (H )

C Convective heat flux incident on the fuel (q.fc)

CO yield (YCO)

Optical path length - fuel vapor concentration ratio (3)

Those parameters which increase with increasing oxygen concentration Radiative heat of combustion (H )

R Fuel vaporization rate (m )

Radiative heat flux incident on the fuel (q.rp)

From this important result, it is apparent that if a conservative assumption,is made for ventilation, i.e., that ideal fuel-oxygen ratios above a minimum value (>

5 mole fraction 0 )

is always 2

postulated to exist, then it is possible to bound the value fo r a

(

liquid hydrocarbon's heat release rate.

Further, one also epn1 I

Appendix A, Basis for Heat Release Rates A-8

(

obtains a s ympto tic values for the fuel steady state mass loss rate as a

function of fire area and the associated heats of combustion (radiative, convective and actual).

From this result the remaining parameter is the value of x1 The method of determination for this parameter will be illustrated for the case of acetone, although the nature of the selected hydrocarbon is unimportant.

It has been shown experimentally that the mass loss rate for most liquid hydrocarbons approaches an asymptotic limit at higher rates of 4" [Tewarson (4)],

especially for aromatic, i.e., ben-zene-like compounds [Tewarson (3)].

In particular, Tewarson (1) demonstrated that acetone, an aliphatic ketone, exhibits charac-

,teristics similar to such aromatic liquids which suggests the

(

validity of the asymptotic limit assumption for its fuel vapori-zation rate.

This characteristic limit appears to be related to the maintenance of a constant q" ratio as surface radiation achieves a dominant role in fuel vaporization.

For most hydro-carbons, this limit is bounded by vaporization rates of 40g/m2.s, a

mass flux supported by erperimental data by Tewarson

( 3, 5),

2

(

where a value of 30g/m -s is suggested, and by Blinov and Khudia-I kov (7).

The steady-state fuel vaporization rate used in this 2

analysis is 40 g/m

_3, With this parameter in mind, it is necessary at this point to focus o'n tne heat of combustion associated with the fuel; in this case, acetone.

Using a bomb calorimeter which accounts for idealized heat measurement resulting from total molecular disso-(

ciation, Weast (2) reports a theoretical heat of combustion (H,)

A

. ~.. -

n.

Appendix A, Basis for Heat Release Rates A-9

(

of 426.8 kG-cal /GMW or 30.8 kJ/g.

Turning to the experimental literature for the purpose of obtaining a value of xA, Tewarson (2) reports a value of H /L=36 for acetone while Tewarson (1)

AT reports H /L=47 48.

This suggests that XA has a laboratory value T

o f 0.76.

On this basis, the following heats of combustion may be calculated:

Actual Heat of Combustion:

23 4 kJ/g Theoretical Heat of Combustion:

30.8 kJ/s

~

These calculated values may be compared to experimental data obtained by Tewarson (6) for acetone:

Actual Heat of Combustion:

21 71 kJ/g

(

Theoretical Heat of Combusti;n:

28.49 kJ/g Recognizing the relatively consistent values obtained under dif-ferent circumstances and assumptions, this analysis utilizes the higher heats of combustion for purposes o f conservatism.

It should be noted at this point that Te war son (6) also reports the following data for acetone in the experiments per-fo rmed :

Actual Heat Release Rate :

262 kW/m2 0.762 actual

=

0 5666 convective

=

radiative luminous =

0.20 k

highly luminous =

0 37 W

Appendix A, Basis for Heat Release Rates A-10

(

It is apparent from a review of this data that a fuel vaporiza-2 c ha r ac teristi,c of the tion rate for acetone of 12.1 g/m was tests reported in Tewarson (6).

This vaporization rate may be best described as non-turbulent or transitory, a condition which would be expected to occur at lower oxygen concentrations where flame convection is the dominant mechanism for fuel vaporization.

In larger fires where flow is truly turbulent, it has been seen

[Tewarson (4)]

that radiation begins to dominate convective heat relepse.

Utilizing the higher value of 37% for the radiative component associated with highly luminous flames, the following values are assumed for acetone :

0 76

(.

ac,tual

=

radiative

=

0 37 convective

=

0 39 This yields the following results for acetone:

Heat of Combustion (kJ/g) 12.0 kJ/g convective

=

radiative

=

11.4 I

actual

=

23 4 complete combustion =

30.8 2

Vaporization Rate (g/m -sec.)

highly luminous flame

=

40.0

(

epm

Appendix A, Basis for He-at Release Rates A-11

' (

2 Heat Release Rate (kW/m )

480.0 convective

=

456.0 radiative

=

936.0 actual

=

In a similar fashion,

one may obtain heat release rate data for other fuels.

For lubricating oil, Tewarson (4) reports the t

following data as representative for typical high-temperature hydrocarbons:

Laboratory Large Scale Scale Heat of Combustion (kJ/g) convective 18.2

(

~

radiative 20.4 16 3 actual 38.6 complete combustion 46 3 2

Vaporization Rate (g/m -s) highly luminous flame 40.0 26.8 Laboratory Large Scale Scale 2

Heat Release Rate (kW/m )

convective 728 534 radiative 816 415 actual 1544 949 k

A

Appendix A, Basis for Heat Release Rates A-12

(

Tewarson (4) reports the following data for heptane:

Laboratory Large Scale Scale Heat of Combustion (kJ/g) convective 21.6 radiative 17.4 14.4 actual 39 0 complete combustion 44.6 2

Vaporization Rate (g/m _3) highly luminous flame 70 70.1 Heat Release Rate (kW/m2) convective 1512 1514 (estimated) radiative 1218 1009 (estimated) actual 2730 2523 (estimated)

This analysis utilizes the laboratory scale turbulent values for fuel vaporization rate and heat release rates in calculating the effects of exposure fires on electrical cables and plant equipment.

The impact of this practice is that this effectively assumes that the most efficient combustion achievable in the laboratory occurs in general plant areas as well.

(

l l

epm

4 Appendix A, Basis for Heat Release Rates A-13

(

References:

(1)

A.

Tewarson,

" Heat Release Rate in Fires",

Fire and Materials, V4, pp. 185-191 (1980).

(2)

R.C.

Weast, Editor,

" Handbook o f Chemistry and Physics",

61st Edition (1980-81),

Chemical Rubber Company, Cleveland,

OH, 1980.

(3)

A.

Tewarson, "Physico-Chemical and Combustion / Pyrolysis of' Polymeric Materials",

Report RC80-T-9, Prepared for U.S.

Department of Commerce, National Bureau of Standards, Center for Fire Research by Factory Mutual Research Co rpo ration,

Norwood, MA, November, 1980.

(4)

A.

Tewarson, " Fire Behavior of Transformer Dielectric Insu-

'lating Fluid s",

DOT-TSC-1703, Prepared for U.S. Department o f Transportation,

Transportation Systems Center by Factory Mutual Research Corporation, Norwood, MA, September, 1979 (5)

A.

Tewarson and R.F.

Pion, " A Laboratory-Scale Test Method for the Measurement of Flammability Parameters", FMRC 22524, Factory Mutual Research Corporation,

Norwood, MA, October, 1977

(

(6)

A.

Tewarson,

" Experimental Evaluation of Flammability Parameters o f Polymeric Materials",

Report FMRC J.1.1A6R1, Prepared for Products Research Committee, National Bureau of l

Standards by Factory Mutual Research Corporation,

February, 1979 (7)

V. I.

Blinov and G.N.

Khudiakov,

" Diffusion Burning of Liquids", Moscow Academy of Sciences (1961).

6 0

l 1

4

(

APPENDIX B Stratification The stratification model used in this section has its origins in work performed by J.S. Newman and J.P. Hill of Factory Mutual Research Co rporation on behalf of the Electric Po wer Research Institute (1). This EPRI research related the radiative and convective heat flux associated with stratified layers of hot gases developed in an enclosure fire to the room's dimensions,

?

the height above the floor,

the fuel's flammability parameters and the ventilation rate.

Data was obtained in a

series of experiments involving 14 methanol and heptane enclosure fires at elevations ranging from 30%-98% of the ceiling height for up to

(

12 room air changes per hour.

Among the general observations, FMRC scientists noted the following :

(1)

Varying the location of the pan fire within the enclosure had no appreciable effect on the measured heat fluxes or gas temperatures at any given position.

This suggests the lack of sensitivity of stratified heat flux to horizontal separation.

(2)

Differences in gas temperature or heat flux measurements at the same vertical position at different locations were, in general, inconsequential and within the variation expected from the measuring instrument.

(3)

In terms of horizontal variation, measurements indicate a

tendency for the enclosure corners to be slightly I

cooler and receive lower total heat flux es than at other locations within the enclosure.

l (4)

The ventilation rate does not appear to have a dominant e f fec t on gas temperatures or heat flux es within the enclosure, with ventilation rates below approximately one and one-half room changes per hour having virtually no effect.

(

1 l

l

J o

4 Appendix B, Stratification B-2

(

(5)

The total heat flux measured at any point in the enclosure is approximately 5-10% radiative and 90-95%

convective for all conditions investigated independent of fuel.

Since the heat flux data collected was for en exposed

sphere, this suggests predicted values which would actually be conservative for cylindrical cable bundles found in cable trays.

(6)

Because of the observed stratification, the application of these empirical results would be appropriate for any d

room shape as long as the floor area of the particular room is greater than or equal to the floor area of a

8 comparable room of the same height with dimensions of 2:1:1.

Newman and Hill reported empirical spatially dependent tran'sient and steady state heat fluxes.

Figure B-1 illustrates the course of heat flux over time following ignition.

The transient heat flux was shown to be related to t time constant unique to each fuel that was obtained by a power curve fit to the

(

~

fire diameter.

Heskestad (2) provideh the basis for such a

response in the early stages of a fire.

Correlations of the data were obtained by Newman and Hill (1) and are reproduced belcw:

l 2

h

-8 b'73 l

9ssH f

(1)

(3p ]

0.2A -

(Steady State)

=

y l

k

[h]-b

= (0.52 +

f ][ t ]0.9 (Transient) 13 (2)

H-r q

H V2 ss I

l i

l

(

l

Appendix B, Stratification B-3

(

~

These results were reviewed for accuracy against the origi-nal data in the EPRI report presented in Table 3-4 of Newman and Hill (1),

which is reproduced as Table B-1.

Plotting the r ported data onto Newman and Hill's Figure 3-2 (reproduced herein as Figure B-2) suggests that the original EPRI correlation defines a poorly behaved function with respect to the ventilation component such that with higher ventilation rates,- a refrigera-tion effect may be noted.

In reality, while higiter ventilation rates will in general have a disruptive effect on any enclosure fire to the point where some mitigation is possible, it was felt that use of the EPRI correlations would be 'non-conservative at some points.

It should be noted,

however, that for relatively small exposure fires which are not ventilation-limited,

the fire

(

severity is reduced as ventilation increases.

This point is discussed in some detail by T.Z. Harmathy (2,3).

Nevertheless, to provide assurance that the function remains well behaved in a conservative fashion and that the experimental data provides bounding

results, a

modified correlation was i

obtained as follows :

2 0.05585 0.01031 0.7854 6" D g

[0.01161 -

3-0.153.3' T H (l.193 - 1 )3

( 2.13

.h_,72 2

M H

l h

.1 H 5/2 [0.01161 - 0.01031(2.13 - h)~b]

f o) 4"s - <

Y 0.05585 f

0.7854 6" H2[

[

j-0.153.

2 T

g y H5 (1.193 - g) 2

2. H s/2 [0.01161 - 0.C1031(2.13 - h)~ ]

l

.c a

l s

t

Appendix B, Stratification B-4

(

13V 0.9 T) [0. 52 + y]

f h

.n e

.a a

(l+ )

4

=q IT)

E qg1qss ss Utilizing these revised correlations, the analysis applies

[

classical optimization techniques for non-linear functions to determine the minimum fuel volumes and associated geometries (i.e.,

fire area and spill depth) necessary to exceed the damage criteria for the cables of concern at the elevations of interest within an enclosure.

9

(

4 l

C l

l l

l l

s

~

.r_,

Appendix B, Stratification B-5

(

~

References:

(1)

J.S.

Newman and J.P.

Hill,

" Assessment o f Exposure Fire Hazards to Cable Trays",

EPRI-NP-1675, Elec tric Po wer Research Institute, Palo Alto, CA, January, 1981.

(2)

G.

Heskestad and M.A.

Delichatsius, "The Initial Convective Flow in Fire",

Report RC79-T-2, Factory Mutual Research Co rpo ration, Norwood, MA, January, 1979 (3)

T.Z.

Harmathy, "Some Overlooked Aspects of the Severity of Compartment Fires",

Fire Safety Journal, 3(1980/1981), pp.

261-271.

(3)

T.Z.

Harmathy,

" Ef fec t of the Nature of Fuel on the Characteristics of Fully Developed Compartment Fires",

Fire and Materials, V3, N3 (1979), pp. 49-60.

e 9

M

(

GAS TEMPERATURES, GAS VELOCITIES AND TOTAL HEAT FLUXES VERSUS POSITION FOR ENCLOSURE FIRE TEST EPOO8 (70 s AFTER IGNITION)

Gas Gas Total vertical Temperature Velocity HeatFgux Percent Station Position

(*C)

(m/s)

(kW/m )

Radiative 1

0.98H 387 5.0 20.4 7.9 2

458 6.4 24.9 9.4 3

429 5.1 20.5 6.5 4

457 5.3 23.1 7.9 5

406 2.8 17.1 7.1 1

0. 90H 364 1.5 12.5 6.5 -

2 356 1.9 12.2 6.8 3

328 2.1 11.8 5.2 4

342 1.9 12.5 6.0 5

385 1.4 13.4 7.1 1

0.70H 315 1.5 11.0 7.4 2

294 1.5 9.7 4.,3 3

299 1.5 10.0 7t3 4

297 1.9 11.0 7.6 5.

311 1.1 10.1 9.9

(

1 0.50H 269 2.4 10.9 8.9 2

268 2.7 10.9 9.1 3

267 1.7 9.1 5.6 4

258 1.3 7.9 3.9 5

256 0.8 7.1 5.7 1

0.30H 232 1.7 8.0 5.0 2

241 2.8 9.2 4.7 3

218 2.2 7.7 5.8 4

222 1.7 6.1 7.5 5

217 0.5 4.7 5.0 Table B-1 Reproduced from Newman, J.S.

and Hill, J.P.,

" Assessment of Exposure Fire Hazards to Cable Trays", EPRI-NP-1675, Slectric Power Research Institute, Palo Alto, CA, January, 1981 1

I s

(

Smoke Detector Activation (I)

! prinkler Activation (138 *C Link)

S 25 500

.s e

5 20 400,-

o "E 515 300 U E

e e

10 2

200

~

(.

j

[

Onset of Cable Damage

{

f (6) e Human Response To o

5 Fire Detection By 10 0 Smoke Detector o

20 O

~

O 60 120 ISO 240 300 Time From ignition (:)

Heat flux and gas temperature at ceiling (Station 4) versus time from ignition for Test EP008 Figure B-1 Reproduced from Newman, J.S.

and Hill, J.P.,

" Assessment of Exposure Fire Hazards to Cab-le Trays", EPRI-NP-1675, Electric Power Research Institute, Palo Alto. CA, January 1981.

k

I l.O

~

i

~

i

l O' uethono t O.El m dia.pon) Qh = 0.98H I

O Methanol (l.22 m dic.pon) eh=0.90H i

A Methanos (1.74 m dio.pord ch = 0.70 H II' I

7 Hepone (0.61 m dio. pan Oh = 0.50H i

Oh=0.3OH ii I

i I

I I'

.8 L g

l 1

l

!l Ii!

.h _ I ! 'l h, l il

.9 f ' H -L. _.!,_dQI

_A. I.

I li - ' -

Correlation

' +.

m_

i T_. y 1 ~ W'-

g

~

i Obtained

,, ?g i

l--

A.

L.l*ieT +

l7 i l.l',,i 2

'$ o' r

_ 6 i - r,. e

_8 i i.j '

by Plotting

' l, '- ' t5 -.! ' 1 [,.--,

'l

, i,

S' l

r ---

C' 6

i-l EPRI Data g

u o

e

-===

l i

i 6 0.I i,

I il Q

6 I

I l

1 I\\*

• 8 w

e t

I 6x ii i

I i

l t

{ll l

\\;\\>

---

\\- k l -t -t br' l

5 4

i i

i

n. T i

M i i \\,_

I i

I_j I

8 3

h 5 ki i

l l

l l

i I

! l l\\ k l

ll l

1 I I

l j:!'i*l i

l t

e l

l l

l 1

-i I

l.ll -

i i

I 'l l

! l1 l! l l'!t l

l l

i l

i j

l

-j l

1l! : !T l

i l

l.i t

l I

I l

i i

(

Q001 0.01 SCAT.ED FORCED VENTILATION RATE.

Yf I

H 5/2 ml/2.s scaled Heat Flux versus Scaled Forced Ventilation Rate Figure B -2

(

Reproduced from Newman, J.S. and Hill, J.P.,

" Assess-ment of Exposure Fire Hazards to Cable Trays", EPRI-NP-1675 Electric Power Research Institute, Palo Alto, CA, January 1981.

i

e

(

APPENDIX C Diffusion Plumes A

low-level fire in an enclosure develops a

turbulent, b uo yan t,

diffusion plume which flows upward towards the ceiling or the first horizontal surface.

Driving the upward flow of hot gases are the gravitational forces acting on the difference in density between the plume and its ambient environment, a

condition which poses a problem for the analyst to consider.

An understanding of the physics of such plumes. is esential to the modeling of the effects of such plumes on immersed materials and components.

Fortunately, recent developments as discussed in the literature allow for the prediction of the effects of such plumes.

(

The history of the modeling of turbulent buoyant diffusion plumes is fairly recent.

An early description of the flow of buoyant plumes published in 1941 is attributable to Schmidt (1).

In a

series of experiments involving convective plumes of air above small sources, Schmidt noted the tendency of buoyant plumes to exhibit conical patterns -in turbulent vertical flow.

Assuming s ymmetr y conditions

existed, Schmidt generated velocity and temperature profiles for constant ambient temperatures involving point and line sources and v erified their accuracy against experimental data.

Batchelor (2) extended Schmidt's results to both stratified and uniform environments in a manner similar to Rouse et al. (3).

These classical relationships are reproduced below:

k

Appendix C, Diffusion Plumes C-2

(

1/3 -1/3 r

U

=

F Z

fI}

a 1i 1/3 -5/3 r

=

F Z

g' fII a

2I d

=

Az where F 5

buoyancy / unit time a

source

=

2w I"U g'rdr o

do g' E buoyancy = g p a z

E height above source

('

r E

radial distance from plume axis or centerline acceleration of gravity g

=

density difference between local and ao =

ambient gas a=

ambient density o

mean vertical velocity in plume U

=

plume radius d

=

A

=

dimensionless constant k

epm

Appendix C, Diffusion Plumes C-3

(

In defining these relationships,

the forms of f (r/z) and 3

f (r/z) were initially undetermined although it may be apparent 2

that boundary conditions require that they be at once continuous and well-behaved.

This consideration was confirmed through a

series of experiments involving hot air in a large room by Rouse et al.

(3) where it was demonstrated that both the mean tempera-ture and the velocity profiles were essentially Gaussian.

On this basis, Batchelor's relationships become:

!e I 2

4 7F z

U

=

a

-I I

=

11F z

e g'

a At this

point, the development o f a theory for buoyant diffusion plumes is limited by the mixing length theories which form the basis for the similarity solution approach taken by Batchelor (2).

These assumptions imply a loss of generality of Batchelor's functions for plumes diffusing into non-uniform gas temperatures.

However, this difficulty is overcome through the use of an entrainment assumption attributable to Taylor (4) for air blast phenomena associated with nuclear detonations.

This very f und, amen tal assumption relates the mean in flo w velocity across a plume edge to the local mean vertical velocity primarily through entrainment.

Morton et al. (5) applied this asumption k

to the study of convection currents.

t epni i

a.,i

.c Appendix C, Diffusion Plumes C-4

(

As reported in Stavrianidis (6), three principal assumptions are made by Morton et al. (5).

(1)

The largest local variations of density in the field of motion are small in comparison to some chosen reference density.

(2)

The mechanics of entrainment can be represented fully by taking a mean radial inflow velocity across some-suitably defined "mean outer boundary" as proportional to the mean vertical velocity on the plume axis at that height.

Equivalently, V

=

Eo Uo where Eo =

0.1 from Stavrianidis (6) and Turner (2)

+

U mean vertical velocity on plume centerline a=

(3)

The mean profiles of longitudinal velocity, temperature and density are similar in shape at all elevations in the plume.

C These relationships apply to weakly b uo yan t plumes.

Extension of the theory to strongly buoyant plumes initially l

l leads to a redefinition of the local entrainment function due to Morton (8):

h E

=

E O a O

With this modification for the local entrainment

function, solution of the general plume conservation equations for the case of the strongly buoyant plume was shown by Morton to be essen-tially equivalent to that of the weakly buoyant plume with larger i

convective heat release rates.

Heskestad (9) subsequently l

l confirmed this generality inside the flame envelope in a series k

of experiments.

1

Appendix C, Diffusion Plumes C-5

(

With this background, it is apparent that turbulent,

buoyant, diffusion plumes could be described mathematically in terms of convective heat release rates and position above the source.

Stavrianidis (6) extended this basis in a series of experiments involving large scale hydrocarbon fires which measured the actual heat release rate in the plume.

The red e fined plume laws correlated to Stavrianidis' data

yield, independent of fuel type:

2 0.092Q (Z~Zo) e

=

AT c

2

~

i U

=

1.20Q IZ-30) e c

g c

where IT =

normalized excess temperature on plume centerline T-T a

Ta T

=

mean pl une temperature T

ambient temperature a=

l O

actual convective heat release rate c=

(

height above physical source z

=

l o=

height of virtual source above physical source l

k.

epm

Appendix C, Diffusion Plumes C-6

(

Stavrianidis demonstrated the validity of these correlations well into the flame envelope to a point of divergence noted for plume gas temperatures.

The data reveals a constant maximum value for temperature of 1235 K for

heptane, methanol, and silicone oil fires.

The point of divergence is defined as the critical height, a function solely of the convective heat release rate, and given by:

ze=

0.130

+Z c

o The determination of the height of the vertical source is given by:

1

/5 (,6

/5 8

o=

7 54F z

)

- 0.15Qc pa where F

=

Da8 fuel vaporization rate m

=

Qc a

=

c 6HT H

convective heat of combustion e=

H theoretical heat of combustion T=

S

=

stoichiometric fuel-oxygen ratio k

i

Appendix C, Diffusion Plumes C-7 With these experimentally derived relationships, it is possible to calculate a number of parameters of interest relative to the exposure fire problem, in particular:

(1)

Plume temperatures above a pool fire, (2)

Gas velocities above a pool fire, (3)

Heat flux delivered to a point above a pool fire, (4)

Radiative heat flux associated with luminous flames.

Each of these calculations is of value in the quantitative fire hazards analysis contained in this report.

This appendix will cov e'r those aspects related to the heat flux associated with diffusion plumes.

The problem of plume impingement is treated in this analysis in three distinct approaches:

(1)

Stagnation heat flux a'ssociated with direct plume impingement on a horizontal surface.

(2)

Cross flow heat flux to a cylinder (cable) associated with immersion in a turbulent buoyant plume.

(3)

Parallel flow along a plate associated with immersion in a turbulent buoyant plume.

Axis ymmetric fir e-ind uc ed flow beneath a flat horizontal surface such as a ceiling 'has been discussed in the literature i

for some time.

Early work includes that of Pickard et al. (10) and Thomas (11).

The theory,

however, did not progress to the level of generality until Alpert (12) developed a basis for the accurate prediction of turbulent ceiling jets as a function of the heat release rate and distance to the ceiling.

Alpert's analytical work, which was verified through experiments, demon-strated the validity of using small-scale models to predict the i

Appendix C, Diffusion Plumes C-8

(

behavior of large-scale ceiling jets.

The basis for Alpert's work includes the top-hat source pro files of Morton et al.

(5) and the Gaussian tempera-ture/ velocity profiles of Rouse et al. (3).

Alpert's model views the ceiling jet as a boundary layer divided into two regions:

an outer region where entrainment occurs as a result of turbulent mixing and a viscous essentially laminar sublayer at the horizon-tal sur fac e.

Data taken in Alpert's experiments indicates a

decline in entrainment by an order of magnitude 3-4 ceiling heights from the fire axis.

A significant decline in ceiling temperature as well as an increase in jet thickness is also noted 3-5 ceiling heights from the fire axis.

Finally, the stagnation region is considered to extend radially outward to a distance of

(

approximately 20% of the ceiling height prior to transitioning to a uniform stratified layer.

Semi-Gaussian profiles are assumed for the transition or turning region.

You and Faeth (13) extend Alpert's work and determine a heat flux within the stagnation region (r/h < 0.2) as a function of gas properties and the fire's he,at release rate:

4"H2 31.2Pr-3/3 Ra-I6

=

Q when Pr =

Prandtl number (-0 7)

Ra =

Rayleigh number 8 bH 8

(10 <Ra<10 ")

9 1

=

(

oC v 3 p

4 Appendix C, Diffusion Plumes C-9

(

Hr <

15 H

ceiling height H

=

H free flame' height r=

gravitational constant g

=

coefficient of coiumetric expansion 8

=

density p

=

ceiling radial velocity for the jet v

=

(

q

=

heat flux c

heat capacity p=

2 H

0.04 (h)-

3

=

for 101'0<Ra<2 x 1013 Pr-0 7 k<0.6 H

k epm

i.

Appendix C, Diffusion Plumes C-10

(

References:

(1)

W.Z.

Schmidt,

" Turbulent Propagation of a Stream of Heated Air", Z. Agnew Math. Mech.; V21, pp. 265-351, (1941).

(2)

G.K.

Batchelor,

" He at Convection and Buoyancy Effects in Fluids",

Quarterly Journal of the Royal Meteorological Society; V80, pp. 339-358, (1954).

(3)

M.

Rouse, C.S.

Yih, and H.W.

Humphreys,

" Gravitational Convection from a Boundary Source"; Tellus, V4, pp. 201-210, (1952).

(4)

G.I.

Taylor, " Dynamics of a Mass of Hot Gas Rising in Air",

U.S. Atomic Energy Commission, MCCD, 919, LADC, 276, (1945).

(5)

B.R.

Morton, G.I.

Ta ylor and J.S.

Turner,

" Turbulent Gravitational Co nv ec tion for Maintained and Instantaneous

. So urc es",

-Procedings of the Royal Society, A236, pp. 1-23, (1956).

(6)

P.

Stavrianidis, "The Behavior o f Plumes Above Pool Fires",

A Thesis Presented to the Faculty of the Department of Mechanical Engineering of Northeastern University, Bo sto n M A, August 1980.

(-

(7)

J.S.

Turner,

" Buo yanc y Ef fec ts in Fluids",

Cambridge University Press, Cambridge, England, 1973 (8)

B.R.

Morton,

" Mod eling Fire Plumes",

Tenth Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, PA, 1935 (9)

G.

Heskestad,

" Optimization o f Sprinkler Fire Protection",

FMRC Report

18972, Factory Mutual Research Co rpo ration,

Norwood, M A, 1974.

(10) 3.W.

Pickard, D.

Hird, and P.

Nash, JFR0 Note 247, Fire Research Station, Boreham Wood, Huts, England, 1957 (11) P.H.

Thom a s,

JFR0 Note ~141, Fire Research Station, Boreham Wood, Huts, Engl and, (1955).

(12)

R.L.

Alpert, " Turbulent Ceiling--Jet Induced by Large-Scale Fires";

Combustion Science and Technology, V11, pp. 197-213 (1975).

(13) H.Z.

You and G.M. Faeth, " Ceiling Heat Transfer During Fire Plume and Fire Im pin g em en t",

Fire and Materials, V3 N3, pp.

140-147, (1979).

k

S

(

APPENDIX D Radiation Radiation can be a significant contributor to the overall heat flux produced as a result of a fire and must be accounted fo r in properly modeling exposure fires in nuclear power plants.

This appendix discusses the appreach taken in this report for modeling the effects of radiation from such fires.

The combustion of organic materials such as liquid hydrocarbons is an exothermic reaction.

The energy released as a result of such reactions leads to the generation of a

high temperature turbulent buoyant diffusion-plume consisting of both gaseous byproducts of combustion and soot particles.

The energy contained within this plume is transferred to the environment through two processes:

(1) convection associated with momentum of the plume and (2) radiation from the plume.

Molecules in an excited state transfer energy via radiation principally through band emission.

For the fundamental products 2, to, H O and soot, such emission tends o f combustion,

i.e.,

CO 2

to be concentrated in the visible and infrared regions typically less than 15ju (1).

The energy transferred by radiation over these wavelengths depends on a number of parameters including average temperature of the source and its constancy.

Historically, fire models and the discipline of fire protection engineering have addressed radiation in considering the effects of an initial exposure fire.

Radiant heating has been found to be a dominant mechanism in the development o f larg-

Appendix D, Radiation D-2

(

scale conflagrations.

This focus is inherently reflected in the use of temperature as a

standard of measurement in tests determining fire resistance.

Typical of this genre are the standards published by the National Fire Protection Association for qualifying barriers and doors for commercial structures and the E-152 Test issued by the American Society for Testing and Materials (2,

3).

These tests are essentially oven tests employing radiant heaters in an attempt to model the dominant heat. transfer process in large scale con flagrations involving residential and commercial structures consisting of and containing high densities of combustible material.

The early application of classical radiative heat transfer

(

techniques to the problem of determining safe horizontal separation distances for building fires is documented in reports issued in the post-war period by British and Japanese investigators (4, 5).

These and later reports published in the 1950s and 1960s retained the concept of horizontal separation as a

principal means of protecting adjacent combustible material (i.e., neighboring buildings) from the intensive effects of major building fires where radiant heat transfer in the open air is the dominant mechanism for damage.

During this period, applications of principles for modeling radiant heating,

well known in other sc ienti fic.d isc iplin es,

were also made to such distinct problems as the effects of fire-induced flows through windows and doors on adjacent structures, effects of wind on fl ame s,

the sensitivity

(

of radiant energy to different flame temperatures and the impact 9

4 Appendix D.

Radiation D-3

(

of various wall materials.

The conclusions from such studies tended 'to emphasize the difficulty of developing generalized empirical relationships independent of scientifically based theory and the importance of understanding the effects of material flammability parameters in modeling the radiative e f fects o f fire.

At a more fundamental level, the effects of radiation may be tied to the gaseous dynamics associated with the fire plume itself.

With its dominant contributions in both the visible and in frar ed regions of the electromagnetic

spectra, the natural focus for a radiation model therefore becomes one based on the material flammability parameters and,

in particular, the height of the visible portion of the turbulent buoyant diffusion plume.

In this regard, F.R. Steward's work, (1970), assumes an important role in providing a comprehensive statement of the d ynamics of fire plumes for sub.*equent researchers (6).

Later work by Dayan and Tien (1974) builds on Steward's research in developing a radiant heat flux model which offers excellent agreement with ' experimental data (7).

This model good mixing associated with combustion conditions in the assumes burning zone so as to provide an essentially uni fo rm gaseous temperature and chemical species concentration in a cylindrical fo rm.

The use of cylindrical form does not appear to suffer a

loss of generality relative to some other shape such as one which i

is either conical or hyperbolic and, in fact, may well be a more accurate representative of average fire conditions.

Of greater k

significance than fire shape in the modeling of radiant heating epm

-?,.

1 si Appendix D,

Radiation D-4

(

is that of soot and gaseous temperature.

Soot and gaseous temperature directly affect the emissivity associated with the luminous flames of a fire.

This effect is in the following form of the Stefan-Boltzmann law:

seen 4

=

eat where 6

=

radiant energy transfer rate emissivity (dimensionless)

E

=

a Stefan-Boltzmann constant

=

absolute blackbody temperature T

=

(

The emissivity of a flame essentially determines the proportion of energy released in the form of radiant energy.

The individual components of the total emissivity may be broken up as follows :

E

=

Eg+E 3 total emissivity associated where E

=

with the fire E

emissivity of the hot gas

=

g within the burning zone E

emissivity of the luminous soot

=

3 within the burning zone 8

Appendix D, Radiation D-5

(

Felske and Tien (1973) provide an analytical basis supported by experimental data-for understanding the parametric relationships of gaseous and soot emissivity (1).

This understanding is a further development of an earlier description provided by Hottel. and Saro fim (8).

In particular, the' relationship of emissivity to spectral wavelength is given for the dominant emission species of water vapor, carbon dioxide and soot.

This relationship is strongly affected by the partial pressures of the products of combustion.

As in the case of other well-behaved spectral functions, the use of an e f fec tiv e value for emissivity is supported by the data and may be provided s

over the range of sensitivity.

This range occurs at wavelengths shorter than the 159 and for infrared band and contains over 96 per cent of the total black body radiation emitted in a fire.

Focusing on gaseous emissivity for the

moment, with the assumption of near-optimal fluid mixing and thermal conditions in a

fir e,

combustion may be assumed to involve the following typical reaction :

(CH )X + (f) X 02+XCO2+XHO 2

2 Under ideal conditions, the partial pressure of CO is 0.131 atm, 2

given a standard environment where the partial pressure of oxygen is 0.21 atm and the partial pressure of nitrogen-argon is 0 79 atm.

From Hottel and Sarofim (8) and Hadvig (10),

the gaseous emissivity is described by:

2 *}

ETg g = 600.0 (PCO L e;

l

Appendix D,

Radiation D-6

(

where Lm=

mean beam length T8=

gaseous temperature PCO partial pressure of C0g

=

2 For the case of an essentially in finite cylinder (i.e.,

an electrical cable):

Lm=

0 94D cylinder diameter where D

=

This yields the following for the emissivity of a hot gas:

(

E

[600][(0.131)(0.9hD)]

8-Tg The gaseous temperature is assumed to be a

uniform 1255 K (1800 F) based on the work of Stavrianidis (1980) using pool fires consisting of heptane and acetone as fuel (9).

As in the case of gaseous emissivity, the contribution of soot to total emissivity may also be characterized by effectively a single value.

Here again, Felske and Tien (1) develop a view consistent with earlier work by Hottel and Saro fim (8).

This view suggests that the mainstream of conditions involving the burning of liquid hydrocarbons, i.e.,

generally lower gaseous temperatures and longer volume of reaction path lengths asso-ciated with fairly efficient (energetic) combustion, the emis-(

sivity of soot may be bounded for the majority of cases.

In

O Appendix D.

Radiation D-7

(

these circumstances, a

value o f 0.1 for the soot emissivity becomes limiting.

With this perspective, a cylindrical fire podel is utilized to analyze the effects of radiant heating on the material of interest.

The burning zone is described by a

more current analytical model for turbulent buoyant diffusion plumes

strongly, supported by excellent correlations with experimental data obtained under controlled conditions involving fairly large scale acetone and heptane fires (9).

This model is described in more detail in the appendix covering diffusion plumes.

The radiant heat flux to an electrical cable from a

postulated fire is therefore given by:

(

T ' + 1.435 x 10-8 0.L12 4) 0 g" =

(5.67 x 10 g

8 where d" =

radiant heat flux incident on a cable D

=

cable diameter 0

1200 K (1800 F) g=

gaseous. temperature =

T 4

configuration factor describing the Foi

=

~

fraction of heat flux delivered to a 3

point by a radiant right cylinder i

Thi s expression is accurate to within 5%

for a

gaseous temperature range of 1000 K-1600 K.

(

epm

a :-....

Appendix D, Radiation D-8

(

References:

(1)

J.D. Felske and C.L. Tien, " Calculation o f the Emissivity o f Luminous Flames", Combustion Science and Technology, V7, pp.

25-31 (1973).

(2)

National Fire Protection Association, " Fire Tests - Building Construction and Materials", NFPA-25-1979 (3)

American Society fo r Testing and Materials,

" Standard Methods of Fire Fests of Door Assemblies", ASTM-E-152-1978.

(4)

R.C.

Bevan and C.T.

Webster,

" Investigations on BuildinE

Fires, Part IV, Radiation from Building Fires",

National Building Studies Technical Paper #5, H.M. Stationery Office, London, 1950.

(5)

'K.

Fujita,

" Fir e Spread in Japan - Fire Spread Caused by Fire Radiant Heat and Methods of Prevention",

Tokohu University, Japan, 1948.

(6)

F.R.

Steward,

" Prediction of the Height of Turbulent Di f fusion Buo yan t Fl am e s",

Combustion Science and Technology, V2, pp. 203-212 (1970).

(

(7)

A.

Dayan and C.L. Tien, " Radiant Heating from a Cylindrical Fire Column", Combustion Science and Technology, V9, pp. 41-47 (1974).

(8)

H.C.

Hottel and A.F.

Saro fim, " Radiative Transfer", McGraw Hill Book Company, New York (1967).

(9)

P.

Stavrianidis, "The Behavior of Plumes Above Pool Fires",

A Thesis Presented to the Faculty of Northeastern University, Bo sto n, MA, 1980.

(10) S.

Hadv ig,

" Gas

'Emissivit y and Ab so rptiv ity :

A Thermod yn amic Stud y",

Journal of the Institute of

Fuel, April, 1970.

l

+

i epm l

(

APPENDIX E Statistical Analysis of Sandia Cable Failure Data l

l k

epm

298 Boston Post Road.

CALCULATION SHEET CAI.C. NO.

N~

Wayland. Massachusetts 01778 3f DATE'2-N Telephone 617/358 5335 PREPARED BY PROJ. CODE Ml f'I CC/

N F

r 7

CHECMED BY DATE TA5K so.

APPvo sY DATE O:1-f[-M

SUBJECT:

$ TAT 6 tidal. AMALYSIS o d 5:fAlw N D 4 Tg.

0 SHEET I OF

... X._O WTIQ

[] ~~~_

?L L2 WFNjigof~dake. rQrf[

-. d.c &p-ogdcx 2-o4 3 law &u s. e.4.aL Ldt2.)_

,-....,..,-....-..-.._m

.e.,g_.-.

- -.. *.. - +. - _ - -. ~ +.

.. = - ~ -

p%,

.m.,

,e....g

..e

..e e

=-e em4v.

w+-M-h (0__t> TLLt 22._ LE AST sav+ cts v=sT-Ns.7nop psscasgrD_

.. LM MGOt^)VM-- >+n)0.SCHkt FSiC O.91$ $%S $ 377-38 3V E _ __

TD DE.44NI.MF_ UAJ$Ni9 2377 M ATa/Z5 9/2-(

k S TCsM MT L:_As2 DEFLUED 6Q

._... ----. --,.- p g 4 y

- wg, g y 4 fygg1xSg orr 7,ug -.

..._ _ _.._ m e fe w /zt.L _. _ _ _ _.

-]1 y = 2x.TAENN.- H 2^*T --

..._ yL "p +- fg. h z...

W A = IN tlE?-,S E Of _ _.

~

^

s k* '. w.

2,y.cg~. - E ye T-yc EM95W NSGssdef.

A

c. e -

i*o N,

as

. _. y 7 c - ( ', y ;) 3--.

t PD6' FMILt!22 g, ~,

3

. l=x-tNrdilM2PT

~~

~ ~

(z,) PEC.JDEM A ~264121SS102 - ~ AMkt-VSt3 fDe ^ 7W 2AntlWf ~

-- - - ~ - - -

- f.x e R. funJe rto N, 2., G U s a S y -

6.b

-. -... ' Q) QS.?OE.T. 225023 l

2e.fer4u.c+3*, _

(<1' J RIAuru.5 AL. LvKh5. b D. Cl W, D, A. Sver; afko694M N

V46) +M EVAEJ4 TION CP 'rthensrt-fearSLPkekTuN bstkses.

tocEr so Appes. goy fl."

^

..... Al 6'J 4.0t 24u t Aj A./ *, san)DIA Ab k71oNAt. LA 60201Df2t2 W F{GOSMH AL. R t,. Sc.H AeFRfz

• MMTH4 TAT!ML RA/WS (z.}. etm Mouckrions MEoeY PefSS, M.ScfTvME Mk; OD3) epm

. ~.

o 298 Boston Post Road.

CALCULATION SHEET

!OM ~~

CALC. NO.

Wayland. Massachusetts 01778 Telephone 617/358-5335 PREPARED Sy CATE PROJ. CODE # A I P C /)/

r CHECMED Sy DATE Appvo By DATE M 'N TASK E0.

SUBJECT:

SHEET 1 0F II E-? 8 AIL UL TiaA l..... -

-.. A..

.l

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

-.f~$JtL.](lameraS..ebal J L94A).,. Table.D Afleadix. T), ilte--

. -.. blloy1YAdb.1L.A.04tdabld.,* - _.. - -.-. - -... -.-. -... -- -.-.

..lfAST~. 5GtJARES RE

. ikOileL. EG),5sl fdtsuaj jdmXaklered)L t

N Yh.Y N N.

_.Y h. _..

c ct >ML _ _.

7 4a7407x1o--. II.-.- -

51Ml49lJIa,-....

. - _4 AsI.

S. A A 2 2 A k la-r ~S3~ ~ ~ ~ ' ~ ~ -

~ ~ - - '

-- A ~'

~ ~ ~

~ ~ ~ ~ - ~ ~ - ~ ~

f.Illl/II x10 T99

.-7 4. I H ! / 7.xla 9 9 -....

l. 20$5535xlJ G41

.,7 I

4. I1666 7.tli 30 -. - -.-.. -

I 9f 000l Al0.

A 00 I.lh03967no.

.io.

97 -.

7.or099rq9fx10 9991 l

. K = 93.2 r o

o a

i 5 g % - @ 4* f ye' m3.oro4 rg2r115').93(I.119f99hsti')

^

4'-- ( $

)*

4(19 ei) -.(? 3)L it l's!

t, I A.

-4 g

1.110 3;to 37 1.ororrrrr/

0.I61g/M/? g g(

L ArsA-M 2

_ _. 9tc

298 B st:n Pect Reed.

CALCULATION SHEET CALC. NO.

~k Wayland, Massachusetts 01778 Tr4 phone 617/358-5335 PREPARED BY g DATE DN PROJ. CODE U / O b O[

b r

CHECKED BY DATE TA3K N0.

APPVD BY DATE d N Z I

SUBJECT:

SHEET 3 0F

- - -.Cf.= <-* [.= [.19 9/996fxid'

=t--

.- 0.9i940740111o 4

V n

n

~~ ~~ ~~ K~T/6-Vl_,f,jf]-~~i~ ] ] ] _ i _

l ]

+

L

_ _ _ _ _ _ A. _ _ _ _ _._ a..

f f.a. O.f $$40740$X10 (l trr91IE?.)(97.LC-)-

_ h,. -...

. =

0 00 l 4-9 0 3 974.-

] bl. N kYb Aubo ml10$ On aA adf fal[5eS art (wHiYbel)

(

z

.YLL. --._--.h__...--_._05

?d..._.. _-. - -

L S ') - ---- (kw/m') ---

~

g

.. g.77*7777h'15y--- k 4.0ffrSTf85

%+L 9

9,, b 20* ft4x10...

N 7.1495f 71klEg 400 36

.. -...,... - - -... + + - -

.m

.. ~ -

f,J f S 7 30 3 X\\0-.-.

(4

1. S~l46 A-17 MXIO

\\7+f klfl4$4l 74*Xl0 h~

$$ kV'IVb?I0b Xl

,qg7glf su,w cr>

s.

+..

e e

a ho =.)--(1, f.= l. % S TS X lo

~ & aft 7Gyau4%WEWM-p}du/.L^y

~

i 298 8:st:n Pcat R cd, CALCULATION SHEET cat.c. wo. 4d^I "-

Wayland. Massachusetts 01778 Telephone 617/358 5335 PREPARED BY CATE PROJ. CODE Nf-lN (

r CHECKED BY DATE

(

TA$K No.

DATE M "/2 O

APPVD SY N

SUBJECT:

SHEET 4 OF

.@k kfk

_l*-.4y br4(Qaj a.p/95 are _ _

_ _ epa.s cp red J _ __

.% c_. __..__ _ ):.c-

__ _ _ yg

__ [ ]

L

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

ged{,.

. _ _ _ _ _ _t. Ifa 646fo& 'l 45. _. ku~j ^ ~ ~

~"

• I 5.SSSS~Sta Yto.

_._.t. 7 54 3 8(o v. t o'3

__3L_____

j q s g ry (,y g _x io ' ~

9(, j z_77 7 7.7 7 8*kt02..-

_ y g _.

._ _ _ t. ; 2g g55 5& q gio _*

j(og j~ }[

/ 7_999IS24rVto

5o8-5

_ _bott?7/.93 AfD

[_.

_9 h f.

M.

M

...._..4_.._..

. k#

. = = =...

..g...

. _... _ _ _....,l_.... >' L. __ _ _ _.. i = '.____** o

-,. 2, Y-c_.-.

(,,%,_7.'T 1-

_.._... $ ___g.

.. _ f. 3......

~K

)

l Y_

h h h YS 2 0 5 kto.) 93){S~O&77'95X/O

~

3(soss) "_.(.73.)'-

e g

4

-e-,

g

-e.e-w.

e 0 06S6S6bibb j,uitsi,,. gio

• 4

~

_ fgg.

t 1

-5Qc S. 08-77793m*

._- ~

g---.-

=.

3 n

/.695906433 gw'

• l

=.

I w

l

(

K=

• &{ - Y cgg

=

1

.s.

L'Ollc%~90W33--~(3 ) )(r.t itt I t t I Jua'4 ) =- - O.co,y gg3,y N

298 Bo ton P:ct R::d, CALCULATION SHEET cAtc. wo. M M 1 Wayland, Massachusetts 01778 Telephone 617/358 5335 PREPARED BY _

DATE PROJ. CODE h ( N 0 l CHECKED BY DATE b D r

3

(

~

TASK No.

APPVD BY DATEdd.U'N SHEET b CF Il

SUBJECT:

. h k.5h.b h

sh8w h408SAEMon

)

u.

_._..f c..__ __.__ A

_. N fC._

h.'*.

9. -...

_. ( u)

-4 ~[~~Q ~

Q Q y,,-r Q

~ ~ ~

~ ~.' ~ ~ ~ g^, Q

,g SEbNl926. kIO -.AI.

. f05't2TrtCGG'd!0. 16fri - -

~

(

. N((,U3o33haue 72.

_),2462SoG/4ph-t 2 G 4 2.~

~ ~ ~

3

-...{

. }.

Q ' -..

,iby ~

0

. s>I J=o. d

.,n,-

_.. %. _bCI~. [,.,

~

2 (t. 2/qf 62506/V X./ ') - - (7 2 )(3. /9 3 03 34/ X/o J) -

l

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

,O[f [// /h.

. l.9 %I7/C'-Yto Y.

= =..

=

-g 1

[=

E R 3.19 30339' m. t,955 /(,7 m' X =-

$>L_-

,39

. =,

l

. Lt%

l. $ $$'/ }/ @ k/()~'$ = " O' 0Wh0f 0W A

• l* $ $ l& I N O

~

epm o

l

298 Bost:0 Psst Rrd.

CALCULATION SHEET cate,yo,AJ0A>-d Wayland. Massachusetts 01778 Telephone 617/358 5335 PREPARED BY _

DATE -

PROJ. CODE $ 9 T O f N

E r

3 CHECKED BY DATE

{

DATE M "

fA3K NO.

APPVD BY SHEET h 0F II

SUBJECT:

b bd 4

. p..

s._

-4,

~ ~ ~ ' [

~~

~ _ ~ (.- o. o o i n f o s t 3 4 ) =p,o 3 Q,_ T

~ ~

_c o.ocoi m,3 es==.

. = * * -. * -

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

_ h '.@-

.a-

-4 m

se

.,e

.e

-., m p n zi.. a.a w Wg{fj-

~

~,~

(_ o. ooi?4ss a n )

=

,,_ 7 3

( C 0001 i l i i i / )

~

(pieiAs-Ahcpe;m i

~ ~

~ ~

p= _ -(to.oi,r s,oi 02s-y _

v.,g z %,......

o 000 / 9 3577f t')

l 2,[#c.frGd M/u.r-e b,

' d 't.

• )'

l-2.

S-(

\\

PEINc abi s 3 e8

(

mfxes i s.13 7000 2.,. vs:-

rice l

{.-

o 298 Boston P:st Rrd.

CALCULATION SHEET cai,c,,o, //dN -(

Wayland. Massachusetts 01778 Telephone 617/358 5335 PREPARED BY DATE PROJ. CODE M it-Y_dl r

3 CHECMED BY DATE

(.

b N DATEM'

~~

fAsK so.

appvo gy I!

SUBJECT:

SHEET 7 0F

.. R Ef6e.ssseog _/hve-yst s_.

. g.. { _ U. 2 1 m -(.9.? P E ft N srto n)

. _. h.z _

-L2

- Yc )...

=

. Hou12 0t%,3-hs pr_ _ A,o onethssp 1s71ysrvig

- o+= r* THt. un knect. T M r?dt M LET2 % 3 *, !$..

g gga. gy.

[ ~ M ^' % FP 3tS) l 2.

t C

~

~ ~j ; E ^

...Z..

.$_~l.~._

_3

.2.

gg. A L.IMS.AK - uAsr SQoMLSS A 7'.

(

7hiS ~ 2SI.A noAL9ftP~ Yisi.os 7HE Fcu o win 16 004cLLOen, i.e.

TMr 4t.THoop_4 _sucq A RT is passiets. R)R

.k1U4 $2.T" 0F DAT4' "Of-. E'TU'+T!0NS 'U IT't LESS TM*l.

s

2. Dhrk PotDi%. No. tingt h,0cs ugy gg est.eutersy. _. _

.. TttEREFDRE, THW t$ A)0 MSTHOD OF ~O2.TT12'HIAltMt>

.7HE Ak-02MY CF soc /t AR7" l

PS/Alc. -

@ -THtS 8 ASIS, on L Y. e 452 leQ, in volvint.;

.-.Eizreicki fn1.oet, orties someitar DkT4 612.

EEU blNAIALYSd..

dAl.8VL.kT/A)Cg S'-

2 is -

(_ b9e)2.

Oc-A..

L

~

7W-d i To. cXo rs~~

II SXog%G puA 3.Gz X,io 23 2.7779777/O'}

3.083'X io' C. Clo?.S~

22 3.re a,7x,o 7.sw <o -3 is.co zs-

-2 L

3o.

<+ o,s 801 ><<o-'

w., eio '

ur.szer i

W

=

7

e a

298 Boston Post Road.

CALCULATION SHEET

~4 CALC. NO.

Wayland Massachusetts 01778 3~

Telephone 617/358 5335 PREPARED BY -

DATE PROJ. CODE A.8,I F PC O f r

1 CHECKED BY DATE

~

APPVD BY DATE M '/

TASK NO.

SUBJECT:

SHEET OF Il 1

.6.= 3 (_ A. M 7 x to,..) -

__._.._..._7._...

S,.= S. I A-3. x l o. ---

.5 =.0 000 +69 9791. -

..ap 6

~ i~o. ftoh$ta ?Al}'

Confibbce

.Inhonts/ aboa! (f, and.

,.fki?5.Ahtl5lk

$vr Wo ] recs l l ret]bA B 09dd. _

(

Q,7.it.- gitttn f. _ 4 30 3--

.s

.s

, f..'[

5

.:r I

~

[Asare Cn,

f-(g,.-yf i-o

^

~

I~

= 0.00 A-3 7/i'M7 n

p.9. S. 7f~

o.000 \\ grr41 ( 4.101)(o.00046'4 797)(0 004171f6W)

'41 41.

//

k 3.I71103 x \\ o ' -p &~- q15-l r

\\

=

2 w1 -

~

q q

L a ).$,ST4X10 1 I.317LO?D'1

- ii

. g-L.31647Xia => E = \\$,567 n wt

298 B:cton Pcst RIsd, CALCULATION SHEET CALC.No.000~N W:ytand. M:ssachusett3 01778 Telephone 617/358-5335 PREPARED BY DATE PROJ. CODE E ( P C_O I r

CHECKED SY DATE TASK E0.

APPVD BY DATEQE.*M'db

~'

k or N

SUBJECT:

SHEET

- ~.

.._.__. A. _.

a..

g. -...

Oo...---

o, y.g 2.

. ___. b T y ).._._...._.-. _.

25 9 ;- -.

2 (,,3,,

,5.,, _ -

/

.... lI*SI WII'4.'T

(_

0 004 7H ?!Z-7--

~~

4. o ot729526

[-0.004N/3/za)'

h 25'3

-( t.tssi kw*}

l (0 001729525N ) =

14 M//

-9.32

/,L

~

~

~ (/.8574g/0%~

e e

.e

,-3 e

4.-

+. -

..g-l

(

..w

..wp

.--g*--.m*.

~

n

.e p-e.*

e-e

.w..-

i

~

298 Boston Post Road, CALCULATION SHEET cal.C. NO.

• ~N l ;

Wayland, Massachusetts 01778 Telephone 617/358-5335 PREPARED BY _

l_CATENb PROJ. CODE D N dl r

1 CHECKED BY DATE TASK NO.

APPVD BY DATE

SUBJECT:

SHEET /O OF _/ !

l

. $0)Y$$.~ 0 0

Sen me._,gs.

-. -- -- -. 4

_y. y q,,

- - - -. - + -. -. _ _... _

, I.

tk,_

(

[

[a

- -- - rue r=cuew in +nn.uizs carsrziA-

- -.6

.-._k...

+ a c 1, c m w..

-m-m.m.Gue + uw me

e v-e mp+=e 4

,.e e

v.

mm

+.

e 4

4*e w

-.-.*+>asemee-.e-W

-m-e+

+, -

amg.e.-.

M.

- ummey-e,-.y e

m e-

-+

-ie.--.m.-m-ed

(

.e._

2 2-3_7_ M 2-

..-e

- = - -_

==ame ew e -.

emm-e-..so 4

.-ew..-e.---g-

-o,.-

m.-,..

m

,se--

..i.e.mi gm, m.,e,+.

pe e

m e - -. -. -

e

. W

. +.,

.(f). THL QTHEE Mrk M %.Alor suRe ei hJ'l. 00&L-@0W e

p.

8= + - -

e

=*

,.e e

.-e es.++=--e=e.e

+=

e s*

e

• *+-

-e

=_

9

-s 4-

-.h

,, + -

e-

.p e

e==-

4 -

mme s

e e4 4

e-e.

- 6 m=

.me.

s e

e 4-.

e e

=+

=

e o-.

me

--+

ww,.eg.--e 4wh-m<

e m

NM e

$ked l' db t(

'(

b Table II A

sw Summary of Test Results Wl(//2 Time to Measured Time of Electrical Time to Weight Test Power Level Exoosure Failure Fire Loss Number (kW/m )

(Inin)

(min)

(min)

(1bs) 2 1.0 1

21 2o 30 t 30*

.1 2

11 40 (O

3 41 go 6.5 6.0 6.5 3.0 4

31 30 26.5 9.5 26.5 7.4

' O.0 5

7 g

30 0.2.

30 22.5 6

11

,o 7

23 zo 30 7.5 t 305 5.1 0.2 8

6 g-30 9

30 3o 7

4 7

2.3 6

4 6

1.2 10 29 30

• Partial electrical failure had developed at 30 minutes and it is assumed that total failure would occur if the exposure were continued.

l Note that test number 2 was run 10 minutes longer than normal.

TThermocouple readings indicated that the cables were very close to ignition temperature (600*C) and it is assumed that fire would develop if the exposure were continued.

N I-12 W:

I

_