NUREG/CR-0636

SAND079-0263
RP

Nuclear Power Plant Fire Protection -

Ventilation (Subsystems Study Task 1)

John A. Milloy

Printed August 1979

Sandia Laboratories

Prepared for
U.S. NUCLEAR REGULATORY COMMISSION
 L11

NOTICE

This report was prepared as an account of work sponsored by

an agency of the United States Government. Neither the
United States Government nor any agency thereof, or any of
their employees, makes any warranty, expressed or implied,
or assumes any legal liability or responsibility for any third
party's use, or the results of such use, or any information,
apparatus, product or process disclosed in this report, or
represents that its use by such third party would not infringe
privately owned rights.

The views expressed in this report are not necessarily those

of the U.S. Nuclear Regulatory Commission.

Available from
National Technical Information Service
Springfield, VA 22161
 NUREG/CR-0636
SAND7 9-0263
RP

NUCLEAR POWER PLANT FIRE PROTECTION - VENTILATION

(SUBSYSTEMS STUDY TASK 1)

.John A. Milloy

Date Published: September 1979

Sandia Laboratories
Albuquerque, New Mexico 87185
operated by
Sandia Corporation
for the
U.S. Department of Energy

Prepared for
Engineering Methodology Standards Branch
Office of Standards Development
U.S. Nuclear Regulatory Commission
Washington, DC 20555
Under Interagency Agreement DOE 40-50-75
NRC FIN No. A-1080
14
 ABSTRACT

This report examines the role of compartment ventila-

tion from the standpoint of nuclear power plant fire
protection safety. Based on a review of numnerous
design standards which consider the influence of
controlled ventilation on reducing fire severity, the
report concludes that current standards and regulatory
guidelines inadequately define criteria for design of
ventilation systems and their operation under fire
emergencies. To resolve this deficiency, the report
evaluates four candidate design bases for ventilation
systems: smoke removal, smoke dispersion control,
fire spread control, and fire temperature control. It
is concluded that the lack of existing fire technology
precludes the implementation of all but one of these
criteria--fire temperature control. On this basis the
report presents an example design calculation for
applying the temperature control criterion.

5
 CONTENTS

Page
Introduction 9
Bac kground 9
Task 1 Description 10
Technical Approach 10
Review of Existing and Proposed.Guidelines and Standards 11
Documents Considered 11
Guidance for Ventilation System Design 12
Candidate Technical Bases 14
Smoke Removal 14
Smoke Control 20
Fire Control. 22
Temperature Control 26
Relative Merits of Candidate Design Bases 29
Design of Heat Venting Stacks 31
Results and Recommendations 34
APPENDIX -- Sizing of Heat Venting Stacks for Natural Convection 37
References 52

7
 NUCLEAR POWER PLANT FIRE PROTECTION - VENTILATION
(SUBSYSTEMS STUDY TASK 1)

Introduction

Background

An earlier Sandia Laboratories fire protection studyl surveyed the

guidelines and standards pertaining to nuclear power plant fire protection
and the investigative reports which followed in the aftermath of the
Brown's Ferry Nuclear Power Plant fire of March 22, 1975. The purpose of
that "survey was to establish a firm basis for future activities in
assessing the adequacy and development of improved design criteria for
nuclear power plant fire protection systems. One of the conclusions
reached in the study was that further work should be undertaken to develop
more compr ehensive guidanc~e for the design of ventilation systems.

Based on this and several other considerations, the NRC Office of

Standards Development funded a new program to carry out a more detailed
investigation. in particular, the following tasks were identified for
study:

Task 1 Ventilation Systems

Task 2 Fire Detection Systems
Task 3 Fire Barriers
Task 4 Fire Hazards Analysis

This report addresses the first of these subjects, Ventilation

Systems.

9
Task 1 Description

It was the objective of this task to examine the role of compartment

ventilation as it affects nuclear power plant fire protection safety. To
do this, the following general approach was used:

" Review and compare existing standards for ventilation

systems to evaluate the adequacy of the guidance provided;

" Develop technical bases for ventilation system functions and

performance in fire emergencies-and identify topics requiring
further investigation or testing; and

" Recommend changes or additions to existing guidance to clarify

intent and define design criteria.

Technical Approach

The existing guidelines and standards, as they apply to the effect

of ventilation systems on fire protection in nuclear power plants, were
reviewed from the point of view of a design organization attempting to
bring a facility into compliance. Specific criteria were listed and
examined to determine if they were adequate to evaluate and specify system
designs.

Current literature in the fire protection field was surveyed to

locate investigative research reports on ventilation-related aspects of
fire phenomenology. Particular attention was directed to reports dealing
with the effects of variable ventilation rates on the growth of
compartment fires and burning rates in fully developed compartment fires.

Information gathered from the review of the standards and the

literature search was used in formulating and evaluating four candidate
technical bases for ventilation system design. Each candidate basis was
examined to determine whether or not it fully met' the intent of the
guidelines and standards. The question of feasibility was also addressed
in each case, with regard to equipment design, plant layout implications,

10
and the availability of data upon which to base design parameters. From
the four candidates, a single design approach was selected and examples of
its application are given.

Review of Existing and Proposed Guidelines and Standards

Documents Considered

The following documents were considered representative of the

available guidance which exert substantial influence upon the nuclear
power industry in the area of fire protection. They'represent the
position of national and international regulatory agencies or are
statements of requirements for insurability:

1. Proposed Regulatory Guide 1.120, "Fire-Protection

Guidelines for Nuclear Power Plant," U.S. Nuclear
Regulatory Commission Office of Standards Development.
Several revisions of this document dating from June
1976, through November 1977, were included in the
review. (RG 1.120)

2. "International Guidelines for the Fire Protection of

Nuclear Power Plants," National Nuclear Risks Insurance
Pools and Associates, February 1974. (IGL)

3. "Property Loss Prevention Standards for Nuclear

Generating Stations," Nuclear Mutual Limited,
September 1975. (NML)

4. "Basic Fire Protection for Nuclear Power Plants,"

Nuclear Energy Liability Property Insurance Association,
April 1976. (NEL-PIA)

5. "Safety Guide on Fire Protection in Nuclear Power

Plants," International Atomic Energy Agency, January 1976.
(IAEA)

6. ANSI N 18.10, "Generic-Requirements for Nucl ear Power

Plant Fire Protection," American National Standards
Institute, February 1977. (ANSI N 18.10)

7. NFPA 803 "Standard for Fire Protection for Nuclear

Power Plants," National Fire Protection Association,
1977. (NFPA 803)

11
 Guidance for Ventilation System Design

The review of the above documents indicates that the ventilation

system as a subsystem in the total fire protection program in a nuclear.
power plant should function:

" To assure accessiblity to fire-involved areas for manual

fire fighting operations

" To prevent the spread of smoke and toxic gases to other areas
of the plant

" To limit damage to electrical equipment and structural features

caused by corrosive products of combustion
" To assure continued habitability of areas normally occupied by
operating personnel

* To prevent smoke infiltration into routes of emergency egress

" To remove bieat which could result in serious structural damage

* To control the release of radioactive contaminants from fire-

involved controlled areas.

The IGL includes the most extensive section dealing with ventilation
system design. It not only states the underlying philosophy in each of
the areas of guidance, but, in most cases) gives specific minimum
standards for compliance with the stated requirements. In some areas in
which specific criteria are given, alternative courses of action are also
given.

En general, however, the existing guidance for ventilation systems is

brief, nonspecific, and lacking in sufficient detail to judge the adequacy
of a particular system or to specify system designs for compliance with
the standards. En those documents where specific criteria do appear, no
basis for the criteria is given nor is its source referenced. For
instance, RG 1.120 states that smoke and heat vents should be provided for
the control room, cable spreading rooms, diesel oil storage areas, switch-
gear rooms, and other areas where the potential exists for heavy smoke

12
conditions. RG 1.120 then references NFPA 204, "Guide for Smoke and Heat
Venting," for additional guidance. However, neither RG 1.120 nor NFPA 204
(see page 19) provides sufficient design information or design bases for
judging the adequacy of a particular smoke venting system design. Further-
more, no information is provided to identify "... areas where the poten-
tial exists for heavy smoke conditions.". Although some specific locations
are identified in RG 1.120, system-design guidance for these locations is
lacking.

A preference for separate smoke and heat venting facilities is

stated in several of the documents. if design criteria are stated for
these vents, they are generally given as a ratio of vent to floor area of
the compartment. Since the capacity to remove smoke and heat is the true
measure of adequacy of such a system, as opposed to the physical size of a
vent opening, it is conceivable that a system fully capable of functional
compliance could be rejected because of its physical measurements.

No consideration is given to the extent of modification that might be

required in existing facilities to bring them into compliance with newly
promulgated standards. Because of the capacities needed to meet the
stated ventilation objectives, the physical size of the equipment required
could preclude their installation in a plant which has many internal
compartments. In such cases, the only practical alternative is to waive
ventilation standards in favor of detection and suppression systems of a
specified reliability level.

En the following section, four candidate approaches for the design of

ventilation systems to meet the intent of existing guidance are presented.
The discussion is intended to develop an understanding of the significance
of ventilation in fire development in the light of current technology. A
design criterion is proposed which could aid both in system design and.
evaluation of compliance with the guidelines.

13
 Candida .te Technical Bases

Smoke Removal

Specific reference is made in three of the guidelines (RG 1.120, IGL,

ANSI N 18.10) to the use of ventilation as a means of assuring access for
manual fire fighting. From a practical point of view, regardless of
whether .or not a fixed fire suppression system is provided, access to the
fire area must eventually be gained to allow mop up and damage assessment.

Access to a fire area could be precluded by three factors: toxicity

of smoke and gases, heat, and visibility. The hazard posed to fire fight-
ing personnel by toxic products of combustion is acknowledged in all of
the guidelines reviewed by the requirement for self-contained breathing
apparatus and the training of the fire brigade in its proper use. The re-
moval of heat from the fire area is discussed under Temperature Control
below. The impairment of visibility by dense smoke is the primary concern
of this section.

The reduction of visual range by dense smoke in a fire area presents

an obvious impairment to manual fire fighting efforts. In order to
effectively extinguish a fire, the suppression agent must be delivered to
the source of the fire. If smoke is so dense that it obscures the
location of visible flame, then efficient delivery of the suppressant
becomes impossible. The difficulty of directing the manual fire fighting
effort is further increased if the involved area possesses a high degree
of mechanical congestion, which is typical of many areas in nuclear power
plants. In such areas, the obstructing machinery makes it necessary to
approach the fire closely for direct application of the hose stream.
These obstructions also serve to mask the fire location and present
obstacles which impede the progress of the fire brigade. Since early.
arrest of fire development is an important factor in limiting the extent
of involvement, impaired visibility can severely limit fire fighting
effectiveness.

14
 Besides acting as an impediment to effective manual fire fighting,
smoke and gases represent potential damage mechanisms which can cause
losses beyond the area of actual fire involvement. The deposition of
heavy soot in Adjacent areas obviously exacerbates postfire cleanup and
salvage problems. The presence of corrosive chemicals in the products of
combustion can also propagate equipment failures beyond the damaging range
of the flames. Electrical control equipment, such as relay racks and
switchgear, are particularly susceptible to damage caused by corrosive
chemicals evolved by cable fires. Therefore, the prompt removal of smoke
can serve to limit the extent of damage and the cost of salvage
operations.

It should also be recognized that smoke production is a time

dependent phenomenon. A practical smoke venting system might be designed
for a given fire area which would be capable of ensuring adequate
visibility in the early stages of fire development, but which would be
totally incapable of coping with a fully developed fire. Similarly, a
smoke venting system could be designed to satisfactorily vent smoke and
ga ses after successful operation of a fixed suppression system and allow.
manual mop up operations and damage assessment, but which would be totally
inadequate if the suppression system should failý: Thus, the design basis
for smoke removal is closely coupled with the speed and reliability of
detection, the expected response time of the fire brigade given detection
if manual suppression is the primary defense, or the effectiveness And
reliability of a fixed suppression system.

Smoke Production -- No definitive means of quantifying smoke

production from a fire is currently available. NFPA 258 sets forth a
standardized. test method and apparatus for the measurement of smoke
generated by solid materials.2 A. F. Robertson presents an analytical
approach for the application of data derived from this test to the
estimation of smoke production in fires within buildings. 3 The latter
reference discusses in considerable detail the criticality of assumptions
made in smoke production analysis, and the uncertainties in the
applicability of laboratory smoke measurements to the prediction of smoke
production under full-scale fire conditions.

15
 Several observations made by Robertson are pertinent to this
discussion. One of the most important uncertainties encountered in this
analysis was the relationship between smoke production and the rate and
extent of fire involvement. Although total smoke potential may be
indicated by the test performed per NFPA 258, the optical density within
the confined area as a function of time will be highly sensitive to how
fast the fire develops. The rate of developmnent is also influenced by the
ventilation rate which, in turn, fosters open flaming conditions under
which smoke can be consumed by the fire in the same fashion that an
afterburner functions to reduce air pollution. This type of combustion is
typical of cellulosic materials or simple hydrocarbons. However,
materials which contain inorganics such as fire retardants* often release
aerosols that are noncombustible and are, therefore, not consumed.
Finally, Robertson observes that even a small fire in a confined cell can
rapidly reduce the visual range to a few feet. This leads him to conclude
that the most effective means of limiting smoke production involve the
confinement, control, and early extinguishment of fires.

Some of the questions rai sed by Robertson have been investigated in

the work of E. E. Smith.4 -6 The data derived from tests such as NFPA 258
for smoke production and ASTM E-84 for flame spread rating are intended to
be used for comparison purposes in arriving at appropriate choices of
construction and furnishing materials. The data are taken under carefully
prescribed conditions of thermal flux exposure, but generally at a single
exposure level. Smith suggests that the measurement of a single parameter
at a single exposure level does not provide sufficient information for
design purposes. Further, it is pointed out that cellulosic and non-
cellulosic materials exhibit very different hazard pr~oducing properties
when exposed at levels other than those specified in the standardized test
procedures. In Reference 4, Smith has proposed a test procedure in which
the rates of heat, smoke, and toxic gas release are measured over a range
of exposures. Smoke release rates at various exposures are reported for
both cellulosic and noncellulosic materials. These results show that the
smoke production rates from synthetic materials, in this case PVC and ABS
pipe, are much higher than those of red oak at the same exposure levels.

16
Since synthetics dominate the fuel loading in many power plant settings,
particularly those which include electrical cables, smoke could be
produced at rates which practically preclude the maintenance of
visibility.

Methods of Smoke Removal -- The means of smoke removal suggested by

the guidelines and standards include dedicated smoke vents (both natural
convection and power assisted), use of the normal HVAC system, and
portable smoke ejectors. The smoke and gases are to be removed to a"'safe
location," which generally means outdoors. However, in those cases where
the potential for radioactive contamination exists, the venting must be
monito red and, if permissible radiological maximums are exceeded, the
smoke must be directed to the gas treatment system or released at a
controlled rate. Of the three suggested options for smoke removal, the
dedicated or separate smoke venting system is listed as the preferred
approach by most of the documents, although normal HVAC systems may be
used provided appropriate design precautions are taken to ensure that:

1. HVAC equipment temperature limitations are not exceeded;

2. HVAC filters are demonstrated to be compatible with the

expected quantity and quality of vented smoke or are
appropriately bypassed during smoke venting operations;

3. HVAC ducting shared with other plant areas is isolated*

to prevent the spread of heat and smoke.

The generally accepted design basis for power plant HVAC systems is
the control of normal equipment and personnel heat loads. Consequently,
air-change rates range from one to perhaps four changes per hour in all
but the most heavily heat loaded areas of the plant. By comparison, for
smoke removal purposes, the IGL suggests air change rates of 5 to 10
changes per hour with a note that up to 50 changes may be required if the
start of smoke venting is delayed. it can be conc luded that these
criteria .and the observations made by Smith4 of smoke produced by burning
synthetics that normal HVAC systems: (1) could be capable of efficient
smoke 'removal from incipient fires; (2) would be of little use as a means
of smoke removal from well-established fires, or during active fire

17
suppression activities, when smoke production is at a maximum; and (3)
could be useful in conjunction with portable smoke ejection for clearing a
space of smoke-following fire extinguishment.

There are also requirements for fire area isolation which preclude
the use of normal ventilation systems for smoke venting. The duct systems
must be equipped with rated fire and smoke dampers to prevent the spread
of fire and smoke to other areas through the ducting. These dampers are
generally of the type which are actuated by excess heat. in the duct by
means of a fusible link. Once closed, these dampers must be manually
reset. Since closed dampers would normally be within the fire area, the
normal ventilation system would be unavailable for smoke venting of the
fire area.

In compartments equipped with gas flooding extinguishing systems, the

area must be sealed off before and during system actuation to allow the
required agent concentrations to be reached and maintained for adequate
soak times. 'Electrically closed datpers are usually installed in the
ductwork for this purpose with no provision for remote reopening.

Additional complications are encounteredein those plant areas where

radioactive materials may be entrained by smoke. For these areas a
reliance on smoke venting to facilitate manual fire fighting or control
fire severity seems ill-advised. Venting from these areas would require
some means of preventing high-efficiency radiological filters from being
plugged with dense smoke and of monitoring radiological effluent
concentrations to prevent excessive discharges to the environment. In
addition, smoke leakage to other plant areas would need to be controlled
during times when fire fighters gain access to the contaminated area.

Under these conditions, a more reasonable fire protection philosophy

would emphasize the use of automatic detection and suppression systems,
while isolating and monitoring the burning area. Through remote mo'nitors
(e.g., video, radiologic-al, or temperature devices), the success of the
automatic extinguishing systems can be judged by the fire'brigade and
plant operators without entering the fire are a. Only after the fire has

18
been extinguished or the automatic extinguishing systems prove inadequate
would the fire area be entered by a fire brigade. These and similar
concerns are being considered in detail by a combined committee of the
American Nuclear Society (ANS .59.2) and the American National Standards
Institute (ANSI N189) in conjunction with drafting a new standard,
entitled "HVAC'Systems, Important to Safety, Located Outside Primary
Containment ."

The use of portable smoke ejectors will generally serve to reduce

smoke density in a fire area by spreading the smoke to other areas of the
building unless a duct system or direct access is provided to route the
products to outdoor areas. Since high concentrations of corrosive gases
can be attained in fire areas, ejection to other parts of the building
could result in extensive damage to equipment not involved in the fire.

As mentioned earlier, separate venting systems are preferred by the

guidelines. The IAEA document states that "fire venting shall predomi-
nantly be of the underpressure-venting type where a separate smoke/gas fan
is provided to evacuate the smoke." IGL, NML, NEL-PIA, and NFPA 803 give
vent-to-floor-area ratios for various areas of the plant which range from
0.005 to 0.04, depending upon combustible loading. NFPA 204 is also
referenced as a source of smoke and heat venting design criteria. This
document seems to be the source of the vent area ratios specified, but the
criteria by which the ratios were determined are neither stated nor
referenced in the text. It is interesting to note in the introductory.
paragraph of NFPA 204 that the document was inspired by a trend toward
construction of large single-area, one-story buildings of light construc-
tion. In the section entitled "Application and Scope," it is further
stated that many of the features suggested by the guide would be difficult
or impractical if applied to multistory buildings.- This is understandable
since the proposed vent configurations are intended, at least in part, to
provide access for hose streams to fight the fire from above.

Since a range of vent area ratios is stated by the standards, some

criterion of adequacy is implied. The IGL publication alone provides
insight into what the design basis might be by associating volume change

19
rates with the venting process. The stated venting capacity is to be
determined by the smoke production of the materials within the fire area
and the number of air changes per unit time needed to sweep the smoke from
the volume enclosed. Rather than a given ratio of vent-to-floor area, the
basic measure of venting adequacy is clearly stated to be volumetric
capacity of the system that can be brought to bear upon the fire area.

G. T. Tamura and C. Y. Shaw have proposed a design basis for smoke

venting shafts in multistory buildings.7 Their paper describes the
mechanism of "stack action" in tall buildings by which smoke is spread
throughout the upper floors through ex~isting service shafts. A method is
developed for the design of smoke venting shafts to prevent this from
occurring. The calculations presented are based on the venting of a low
temperature fire with a temperature difference of 75*F between interior
and outside air providing the stack draft. Losses due to friction,
velocity pressure, and leakage are considered. A table of minimum smoke.
shaft sizes for a range of floor areas and building heights at several
leakage rates is also given. The minimum sizes are indicative of volume
change rates of two to four per hour at the bottom floor of the building.

The work of Tamura and Shaw provides insight into the design of
natural convection smoke vents for multistory buildings, presumably of
residential or office occupancy. The selection of a low temperature fire
and a relatively low air change rate would be indicative of this type of
combustible loading. To be applicable to nuclear power plants, the shaft
sizes would have to be increased to provide much higher flow rates to
accommodate greater smoke production from synthetic materials.

Smoke Control

As used here, smoke control is defined as the prevention of smoke

migration from the involved fire area to other parts of the building. The
IGL lists the confinement of smoke and gases as a specific point for
considerat-ion in the design of enclosures for fire areas. Several of the
standards specifically list the stairwells as areas for which special
measures should be taken to prevent smoke infiltration, generally by means

20
of overpressure ventilation, to assure a positive pressure inside the
escape routes relative to other building areas.

The movement of combustion products toward the upper floors of high

rise buildings-has been recognized as cause for concern because these
.gases can result in serious visibility and toxicity hazards long before
fire reaches the area. The use of differential pressure to control
infiltration has been applied successfully as a solution to the problem.
Fung has reported the results of full-scale smoke movement experiments
conducted in the 36 story Seattle Federal Building and the 42 story
Chicago Federal Building.8 Both facilities employ the "systematic
pressurization" concept in which the structures are divided into vertical
zones, of several floors each, served by separate air handling systems.
In a fire emergency, the systems can be switched to smoke control mode by
simultaneously placing the fire zone in full exhaust condition and the
adjacent zones in full supply operation. Thus air is forced to flow
toward the fire from all directions.

An approach such as this could conceivably be adapted for use in a

nuclear power plant, but only at the expense of severely complicating HVAC
system design. In the high rise situation, the fire zones are arranged
vertically with entire floors being considered single fire areas. To
.consider a single floor or group of floors as a separate air distribution
zone follows naturally from the high rise configuration. In the power
plant, however, the arrangement is primarily one of horizontal expanse
with much subdivision of a single level. An arrangement of ductwork and
remotely controlled dampers which would allow any given area to be aligned
for exhaust and all surrounding compartments to supply 'Modewould pose
substantial problems both in design and operational reliability.

Typically nuclear power p lanit layouts incorporate a great deal of

totally internal compartmentalization with very few or no exterior
windows. Large open areas are also included in buildings, such as the
turbine hall, which have considerable vertical extent, unsealed floor
openings, and extensive-6pen grating. The combustible loading and smoke
production potential of many plant areas is also much higher than the

21
typical office and residential occupancies of the buildings in which this
system has been successfully implemented.

The air handling facilities for power plants generally incorporate

both supply and exhaust fans. It might, therefore, be opted to seal off
the involved fire area with fire dampers, and shut down the exhaust fans
but continue operating the supply fans of adjacent compartments.. This
approach may provide sufficient differential pressure to confine smoke
during the very early stages of fire developmnent. But unless some means
of venting the fire's combustion products to'the outside air is provided,
pressure will rise very rapidly in the heated compartment and soon over-
come the differential.

it is therefore concluded that for differential pressurization to be

effectively and economically used as a means of smoke control in power
plants, it must be accompanied by a properly designed smoke venting system
to provide pressure relief. The option to independently control supply
and exhaust fans could, however, prove to be an asset during fire
emergencies.

Fire Control

Intuition and Fire Phenomenology -- The concept of "control" over a

phenomenon can have many different meanings depending-largely upon the
nature and intended application of the event being observed. For example,
to many motorists, the term "controlled skid" may seem to be a contra-
dictory use of words. Yet when viewed from the premise that a skid has
occurred and the problem is now that of limiting undesirable consequences,
the fact that an effective degree of control can be exerted on the
situation can be significant to the eventual outcome.

The term "fire control" can be viewed in a similar way. Given that
fire, in the context of this work, is an undesirable event, one "control"
might be the elimination of fire as a possiblity. Another course might be
to provide for the immediate and decisive extinguishment of any fire
occurrence so that no fire can get "out of'control." The difficulty

22
inherent in these approaches is that neither can be applied with absolute
certainty. one is therefore forced to assume that a potentially destruc-
tive fire has occurred and to investigate practical means by which the ex-
tent of damage might be limited. More specifically, in this section, the
discussion is concerned with the role of ventilation in the development of
the fire and its influence upon the eventual confinement of damage.

Most people have acquired some familiarity with fire phenomena.

However, because of the potential for destruction, most first-hand
knowledge is based on observation of fire in controlled circumstances.
If, for instance, a fire is not physically confined, we have learned that
its size c~an be controlled by regulating the amount and geometry of the
fuel. For confined fires, say in a wood stove, the amount and arrangement
becomes less important since the supply of oxygen can be regulated to
control the burning rate. In either case, it is necessary to deal with
the question of ventilation, particularly in the initial stages of
development, since we become intuitively aware of the difficulties of
starting a fire in a strong breeze. on the other hand, we have also
encountered the situation in which, because of the arrangement or moisture
content of the fuel, it is necessary to direct a stream of air to the fire
in order to sustain burning.

It is perfectly natural to extrapolate acquired intuition from these

controlled small-scale encounters and apply the same principles to large-
scale destructive fires. Thus, a forest fire could be viewed as a large-
scale campfire, and a fully developed compartment fire as a larger version
of one in ,a wood stove. The fact is, however, that for all of our long-
time association with fire, the actual phenomenology is, at best, poorly
understood. The complexity of burning and the multiplicity of parameters
involved preclude all but a cursory understanding of the events which
occur during a fire. Because of this, the use of various scaling theories
which have been developed to predict large-scale fire phenomena from small-
scale experiments appear to be of questionable value and, therefore, the
study of fire phenomenology becomes a complex and expensive endeavor
involving large-scale experiments whose results, often, violate intuition.

23
 Confinement of Compartment Fires -- The work of T. Z. Harmathy is
noteworthy, not only for reporting the effect of ventilation rates on
compartment fires but also as a source of detailed description of the
processes and parameters influencing the course of destructive fires.91
A basic premise of this work is that if a fire can be confined to the
compartment of origin, then the goal of effective fi re safety design has
been met. Thus, if buildings could be designed so that a fire cannot
propagate beyond the first set of barriers, even in the absence of active
intervention, then the ultimate limits of destruct ion have been set.
Active intervention could then be expected to further limit fire loss and
provide an extra measure of protection.

A potential application of this philosophy would be in nuclear

facilities in which redundant safety systems are physically separated from
each other by fire barriers. If this were the case, and if compartment
barriers could be shown to be the ultimate limit of propagation, then fire
fighting, whether manual or automatic, could be directed toward limiting
the extent of damage within the cell of origin. Fire would therefore pose
no threat to the ultimate safety of the reactor system.

It must, however, be pointed out that Harmathy' s work applies to

relatively light fire loadings, as would be encountered in office and
residential occupancies. The compartments modeled are typically limited
in depth and ventilated through windows on at least one side. Further,
the expressions developed are based on results of tests employing only
cellulosic materials. These qualifications preclude the direct
application of the results to the power plant setting. The theoretical
framework and the method of analysis should, nevertheless, be valid for
any compartment fire, given an appropriate experimental data base. In
some regards, the analysis of the power plant problem by this technique
may be more tractable than the one addressed in the referenced material.
Since the ventilation rate is of primary importance in the analysis, a
closed compartment with a well defined ventilation system should be more
easily modeled. The class of combustibles available for involvement in a
power plant is much more limited and more clearly defined than those in
residential and office applications. Therefore a more coherent set of

24
experimental data should be obtainable for evaluation of' the various
parameters. Finally, a power plant is designed as a single-use facility,
and by its very nature is a more carefully controlled environment during
operation. The uncertainties inherent in the analysis of human
occupancies would thus be minimized.

The main thrust of Harmathy's work is toward the development of a set

of expressions, which are sufficiently comprehensive to match theoretical
predictions to existing experimental data. Since the objective was to
arrive at a measure of fire severity that could be applied in the
evaluation of the barriers which form the boundaries of the compartment,
only fully developed fires. were considered. The burning rate was found to
be a function of the flow rate of air and the free surface area of the
combustible materials. A critical value of air flow rate is defined below
in which the rate of burning is determined by the air flow rate and the
fire is said to be "ventilation controlled." At air flows above this
value, the rate of burning is determined by the'available free surface of
the involved fuel and is termed "fuel surface controlled." Conclusions
reached for cellulosics indicate that fires in the ventilationý-controlled
regime (i.e., low ventilation rates) tend to be of longer duration and
reach higher peak compartment temperatures than those which are surface-
controlled. This can be attributed, in part, to the cooling effect of the
excess air at high ventilation rates. The ventilation rate, therefore,
not only governs the rate of burning of the involved combustibles., but it
also determines the duration and peak temperature attained in fully
developed compartment fires.

H1armathy' s approach to the analysis of compartment fires involves

consideration of the fire and its confining barriers as elements of a
total fire system. By formally accounting for the differences between
unconfined and confined burn test results and fitting the expressions
developed to compartments of various geometries, a method of interpreting
burn test data in the context of compartment hazard analysis is provided.
The analysis also provides insight into the relative importance of the
various parameters-in the total fire cell chronology. This information
could therefore be used to structure a fire testing program for the
acquisition of the requisite data in specific applications.

Unfortunately, many of the conclusions reached by Harmathy apply to

cellulosic-fueled fires in relatively light combustible load settings.
These are unlike the conditions typically occurring in nuclear power
plants and, therefore, a program of carefully designed confirmatory
compartment burn tests would probably be required. It should also be
noted that this approach to fire safety requires that, as a last resort
once a compartment fire has developed, the fire is allowed to run its
course, thus consuming the entire contents of the compartment. In many
existing plants where spatial separation is used as a means of limiting
the extent of fire damage, such a premise could not be tolerated.

Temperature Control

Where separate venting is discussed in the guidelines, the

terminology used is generally "smoke and heat venting." The design bases
for smoke venting anSI heat venting are quite different, so the two
functions are considered separately in this report. This section deals
with temperature control, which can be equated with "heat venting."

The work of Tamura and Shaw,7 which was discussed under Smoke
Removal, is concerned with smoke removal by means of separate shafts. The
motivation for considering separate smoke venting systems is at least
twofold: (1) The capacity of normal HVAC systems is not sufficient to
support the required flow rates, and (2) the high particulate concentra-
tion of smoke rapidly degrades the capacity of the air handling system
because it clogs the exhaust filters.

As was the case for smoke venting, the normal HVAC system is
gemer-ally of insufficient-ca-pacity to handle the flow rates required for
heat venting. In this case, however, the design criteria imposed on the
heat exhaust system are much more demanding than those for the normal air
handling function. The temperatures that could be attained by the
effluent from a compartment on fire are very much higher than those of

26
normal exhaust air. The ductwork must therefore not only be noncombus-
tible, but must be capable of retaining structural integrity at elevated
temperatures. The high-temperature gases also threaten the integrity of
exhaust filter systems and exhaust fans. Indeed, the charcoal filters
used. to filter radioactive effluent are a secondary fire hazard when
exposed to high temperature.

Harmathy has proposed the use of "fire drainage systems" in connec-

tion with his studies of fully developed compartment fires.1 0 These
systems are actually heat vents designed to remove combustion products by
natural convection at flow rates compatible with the fuel surface con-
trolled fire regime. This method of ventilation was proposed to contend
with large areas or deep compartments in which ventilation through broken
peripheral windows could not be relied upon.

A compartment fire model has been presented by J. Quintiere which

deals with the early stages of fire development, namely that period from
ignition up to room flashover.12 Flashover can be defined as the critical
point in fire growth, within a cell which marks the difference between a
fire that stays relatively small and confined to its initial surroundings
and a fire that reaches its fully developed state, i.e., total room
involvement. Quintiere' s work is therefore complementary to Harmathy' s in
that the two models taken together address the entire history of a com-
partment fire. The two models are also functionally complementary since
the same course of corrective action is indicated by both models,
specifically the design of effective heat venting systems.

Quintiere has modeled conditions within a door-ventilated compartment

with a centrally located fire source. The model is quasi steady state in
nature, primarily because the history of turbulent mixing within the cell
is not well defined. Because of this limitation, the predicted behavior
is probably more accurate in the early stages of fire development when
mixing has less influence on room conditions.. The ventilation rate is
determined by the size and proportions of the door opening in a fashion
similar to Harmathy's use of exterior windows. The features considered
which are involved in the rate of fire spread are the fire plume itself

27
and a layer of hot gases confined in the upper portion of the room. The
inclusion of this gas layer was inspired by the observations of Waterman
from a series of experiments performed to determine the conditions sup-
porting room flashover. 1 3 These experiments led him to observe that the
ability of fire to jump the intervening space between widely separated
fuel packages is traceable to the heating of the remote combustibles by
convection and radiation from the upper portion of the room, not propa-
gation directly from the source fire. J. B. Fung has observed average
temperatures in this upper gas layer in excess of 600%C (1110*F) resulting
from confined wood crib fires.' 4 The model suggests that compartment
flashover and the onset of fully developed fire could be significantly
delayed, or perhaps precluded, if the maximum temperature of this hot gas
layer could be limited. It can also be observed that the rate of fire
growth in the early stages following ignition is dependent upon thermal
feedback to the source fuel. Thus, if the immediate environment of the
source fire is cooled by ventilation, radiative and convective losses from
the source fire increase, thereby slowing its growth.'

It should be noted that, if the combustible loading is spatially

continuous, the fire can spread along the material directly. Even in this
case, however, the growth rate can be slowed by ventilation since the
materials are not preheated to the same extent by convection. In fuel
configurations which have a vertical arrangement, materials near the
ceiling can be heated to their ignition temperatures by envelopment in the
hot gas layer. The extent of this envelopment may also be limited by
removal of the hot gases.

In summary, by venting heat from a burning room to reduce compartment

temperatures several benefits can be realized. These include:

e Slowing the growth rate of a fire in early stages of

deve lo pment

e Limiting the extent o~f fire involvement by delaying or

preventing flashover to remote fuel packages by radiative
or convective heating from the upper portion of the room

28
 *Reducing peak temperatures within the compartment with
consequent reduction of threat to the integrity of struc-
tural members and fire barriers

* Controlling the spread of combustion products to other

areas by reduction of pressure within the fire cell,
whether by natural convection or power assisted ejection

* Improving accessibility for manual fire fighting by reduction

of both compartment temperature and smoke concentrations.

Relative Merits of Candidate Design Bases

The preceding sections have discussed the role of ventilation as a

subsystem of fire protection in nuclear power plants from four different
points of view. The objective of this discussion has been to consider
each of these viewpoints in the light of existing guidance and current
fire protection technology to determine which approach or combination of
approaches could be sufficiently quantified to generate design criteria.
Taken in the order of appearance in the above discussion, the relative
merits of the candidates are addressed in this section.

Smoke Removal -- Considered as a separate design goal, the~removal of

smoke to facilitate manual fire fighting is poorly defined. Minimum
requirements for visibility and maximum allowable temperature in the
involved compartment are not known. Even if they were, no definitive data
is available to predict either the total amount or the rate of smoke
production. Therefore, no specific design criteria can be identified to
establish firm design bases.

Smoke Control -- The use of a systematic pressurization technique to

confi ne smoke and gases to the involved fire area by realignment of the
normal HVAC system is not a-practical design basis unless some means of
pressure relief is available in the fire cell. When considered in con--
junction with a separate venting system, however, the option to continue
the operation of supply fans to surrounding compartments following closure
of fire dampers in the ductwork serving the fire area could be used to
advantage to assure flow of air toward the fire source.

29
 Fire Control -- This concept of allowing a compartment to reach full
fire development while supplying excess air to limit temperatures and in-
sure survival of compartment barriers is currently developed only for
cellulosic combustibles in relatively light fire loading occupancies.
Since many nuclear power plant areas contain high concentrations of non-
cellulosic combustibles, it is questionable whether sufficient air can be
supplied to a room to achieve a surface-controlled, fully developed.fire.
Without first reaching surface-controlled conditions, additional air will
increase, instead of limit, fire temperatures. However, since additional
air will reduce a fire's duration, it may be possible to balance increased
room temperatures and reduced fire durations in a manner compatible with
compartment barrier survival.

Temperature Control -- The concept of heat removal from a burning

compartment embodies desirable features of all of the proposed technical
bases. Heat removal obviously will remove smoke as well as heat and will
lower the pressure within the involved compartment, thereby controlling
the spread of smoke and gases by differential pressurization. In addi-
tion, a properly designed path for heat removal will ensure fire control
by balancing peak room temperatures and fire durations to the capabilities
of installed fire barriers.

it is therefore concluded that temperature control is the most

reasonable and effective technical basis for designing a ventilation fire
protection subsystem, because existing fire technology insufficiently
quantifies design criteria for smoke removal, smoke control, and fire
control. Although temperature control appears to represent the best basis
for integrating controlled ventilation into an overall fire protection
scheme, no evaluation has been made in this or any other study of the
importance of controlled ventilation relative to other fire protection mea-
sures (e.g., automatic suppression, automatic detection, or se paration).
Such a relative evaluation needs to be made on the basis of both the
benefits and detriments resulting from the use of a temperature control
ventilation system in nuclear power plants. Further consideration of a
controlled ventilation scheme in this report presupposes that the relative
merit of a temperature control system will be demonstrated before any

30
serious consideration is given to its implementation. The task of demon-
strating the relative merit of fire protection options lies outside the
scope of this present study.

Design of Heat Venting Stacks

On the basis of the conclusion reached in the preceding section, that

temperature control represents the best technical basis for designing a
ventilation fire protection subsystem, it was decided to outline here one
possible design approach for accomplishing temperature control. It should.
be recognized, as stated earlier, that the relative merit of a temperature
control ventilation system to overall fire safety should be demonstrated
before any serious consideration is given to implementing the design tech-
nique presented in this section.

The venting of heat from a fire can be accomplished by the use of a

heat venting stack designed specifically to channel the gases directly
from the compartment involved to the atmosphere. Since gases flowing in
this duct could reach high temperatures, material used in its fabrication
must be capable of withstanding a high temperature environment for extend-
ed periods of.time and maintaining its structural integrity for the dura-
tion of the fire. For the reasons discussed earlier, it is unlikely that
the normal HVAC ductwork, filters, and fans would have the capacity or the
high temperature capability required. Therefore, a vent stack dedicated
to this purpose should be considered.

The driving force to carry the gases to the outdoors could be either
natural or power-assisted convection. The use of natural convection is
attractive in that no external power need be supplied during the fire
emergency. The disadvantages of the passive system are that a temperature
differential must exist for the stack to function and, since the veloci-
ties created by the buoyant forces are relatively low, fairly large cross-
sectional areas are required to provide the necessary flow rates. Since
the purpose of the stack is the removal of heat, the temperature differen-
tial will exist when the system is called upon. However, if the fire is a

31
relatively low-temperature, smoldering fire with heavy smoke production,
.the smoke removal function could be too slow to maintain good visibility.
This problem could be overcome by the attachment of exhaust blowers at the
roof level.

The size of the stack must be determined by the maximum flow rate
that it must provide to fulfill its intended function. The starting point
for the analysis would be to determine the maximum heat generation rate of
the compartment served-by the stack. The maximum permissible temperature
must then be specified for the compartment. The flow of air necessary to
carry away sufficient heat to limit the temperature to the acceptable
value can then be calculated. The required flow up the stack would then
be the calculated air flow rate at that temperature. Provision, of
course, must be made to assure that an equivalent mass flow of air at
ambient temperature can be supplied to the compartment during the venting.

Applying this calculational sequence, a design equation was developed

as described in the Appendix. This equation relates the stack's flow
rate, diameter, and height to the temperatures in a burning compartment.
By solving this equation for a variety of conditions (see Appendix) it was
found that, for specified venting rates and temperature limits, the re-
quired stack diameter varies considerably with the availability of make-up
air entering the burning compartment or with the extent of air leakage
along t-he stack length. If room openings are too few or too small, make-
up air to the burning room will be restricted. Also,. if any air leaks
into a stack through dampers from other rooms, the natural draft charac-
teristics of the stack will be reduced. Either of these shortcomings
would require an increase in stack diameter to ensure adequate flow rates
for temperature control._

In addition to these findings, it was shown for an example case (a

5 ft2 gasoline fire in a 37 500 ft3 compartment) that the stack size
required to limit peak temperature to a relatively low value is indeed
reasonable: a 3-ft diameter stack would limit the peak to 400 0F. It was
also observed in this same example that the required airflow rates were
quite high compared to those normally handled by an IIVAC system. This

32
supports the statement made earlier that normal plant air handling systems
lack the capacity to provide adequate heat removal. Finally, it was found
that the compartment air change.rates, particularly for the low-
temperature cases, are high enough to remove appreciable amounts of smoke
from the fire area. Therefore, such a system would also serve to facili-
tate manual fire fighting.

Although a number of factors which could require that vents be made

larger were neglected in the calculations, including resistances to make-
up airflow occurring upstream of the compartment opening, other factors
should be mentioned that could have the opposite effect: no heat losses,
to the contents or w~alls of the compartment were included. These losses
would have the effect of reducing the required air flows. No mention has
been made of the role of suppression systems in the removal of heat. A
fixed water suppression system is capable of delivering water at a rate of
about 0.25 gal/ft2 -min or 2.0 lb/mmn-ft 2 . Given the heat of vaporization
at atmospheric pressure of 1150 Btu/lb, the heat removal capability of a
sprinkler system could be 2300'Btu/ft 2 -min if all water were vaporized.
2 compartment used in the examples, this represents a heat
In the 2500 ft
removal capability of 5.75 3t 106 Btu/min. In actuality, total vaporiza-
tion of the water is neither achieved nor desirable in sprinkler system
design. But the heat removal capability of a sprinkler system could
substantially reduce the required capacity of the heat venting system.

Another point which must be considered in the design of heat venting

facilities is the handling of radioactive effluent. Smoke and gases from
areas of the plant which have potential for radioactive contamination must
be monitored to make sure nuclide concentrat Iions do not exceed permissible
limits. If limiting concentrations are detected, the effluent must be
directed to the normal plant gas treatment system or the rate of release
controlled. This suggests that heat venting and normal RVAC system dam-
pers in the ductwork servicing radiologically controlled areas must be
capable of continued remote operation. If both types of venting capa-
bility were available, it could be possible to limit the heat and smoke
load imposed on the gas treatment-system.

33
 Results and Recommendations

As stated in the section titled Technical Approach, the technical

position taken in this study was that of the designer whose goal is to
bring the ventilation system of a nuclear power plant into compliance with
the referenced standards. By virtue of the current state of the art in
fire protection technology, this approach leads to the conclusion that the
technical design basis for the fire protection design of ventilation
systems should be that of heat removal from the involved fire area for
purposes of controlling fire temperatures. Examination of the ventilation
system design parameters generated by this approach reveals that all but
one of the roles listed under Guidance for Ventilation System Design would
be fulfilled. The single exception is control of radioactive release from
controlled areas of the plant.

The control of entrained radioactive substances presents a ventila-

tion design problem in that the normal HVAC system, which is equipped to
remove these substances, is generally incapable of handling the partic-
ulate concentrations and temperatures associated with fire generated
effluent. NFPA 803 states that, for radioactive substances, the options
are either confinement or release under controlled conditions. The IGL
suggests using the normal gas-treatment system, provided that filters are
suitably protected by the installation of prefilters or scrubbers and
demisters to cope with smoke, heat, and corrosive gases. Presumiably, the
IGL suggestion meets the "controlled conditions" requirement of NFPA 803;
however, another option might be the regulation of the removal rate
through the vent which could lower the entrainment rate if the radioactive
substances are being picked up by high-velocity flow patterns within the
compartment. The confinement of smoke and gases, however, is clearly not
a viable option unless the fire is arrested in its early stages of growth.

If a system of prefilters were added to the normal filter banks

serving the controlled areas, the particulate concentration of contam-
inated smoke could be reduced to manageable levels. The addition of an

34
upstream water curtain and demister, as suggested by IGL, would reduce
temperatures and corrosive properties before filtration. The design or
backfitting of all of the associated exhaust ductwork to withstand- high-
temperature gas flows remains a formidable problem.

As previously mentioned, the extinguishment of a fire in its early

stages and the cooling ef fect of fixed water suppression systems greatly
reduces the smoke and heat load on the ventilation system. Therefore the
practical approach to fire protection in plant areas with high potential
for radioactive contamination may be increased emphasis on early detection
and automatic fixed suppression systems with attendant relaxation-of
ventilation requirements. Plant areas in this category might include the
reactor building,.radiation waste treatment areas, fuel handling facili-
ties, and some emergency pump rooms. A potential-problem with this
approach arises in the case of the BWR, for which the turbine building and
main steam tunnels are included as controlled areas. It is most probable,
even in this case, that radioactive concentrations in these areas would be
within acceptable limits for direct venting during fire emergencies in the
absence of the simultaneous occurrence of a pipe break.

The selection of "heat removal for temperature control" as the fire

ventilation system design basis implies a close systems relationship with
other phases of the fire protection program. Sizing of venting facilities
is determined by potential heat production rates identified for fire areas
in the fire hazards analysis. The cooling capabilities of fixed suppres-
sion systems can also enter into the ventilation requirements as well as
the performance capabilities of the fire detection system. The maximum
permissible temperatures for which the venting system must be designed are
governed by the ignition temperatures of the materials confined Within the
vented area, and by the thermal and structural characteristics of the fire
barriers which form the boundaries of the compartment.

The results of this investigation can be summarized in the following

conclusions and recommendations:

35
 " An evaluation needs to be made of the benefits and detriments
of a heat removal fire ventilation system relative to other
available fire protection measures (e.g., automatic suppression,
automatic detection, or separation). This effort should be
completed before serious consideration is given to implementing a
temperature control ventilation scheme in nuclear power plants.

" If the use of a heat removal fire ventilation system is evaluated

as worthwhile, the technical design basis for the fire venting
system should be the required rate of heat removal from involved
fire areas.

" Existing guidelines and standards are generally lacking in suffi-

cient detail to function as criteria for the design of ventilation
systems as an integral part of the fire protection system.

" Current fire protection research activities are directed primarily

toward the solution of light fire loadings, which are not typical
of all areas of a nuclear power plant setting, Experimental
programs should be proposed to provide basic fire performance
data on combustibles normally found in critical areas of power
plants.

" In areas of the plant which involve high probability for

entrainment of radioactive containments in the smoke and gases
and in backfitting of existing facilities, emphasis should be
placed on the design and reliability analysis of fire detection
and suppression systems with accompanying deemphasis on venting
requirements.

* To allow sufficient flexibility of operation during fire emergen-

cies to adequately control the spread of smoke and provide makeup
air for fire vented compartments, the fans, isolation dampers,
and their associated power supply and control cables should be
protected from fire damage. Manual remote operation capability
should be provided so that regulation and realignment of the
systems can be accomplished as the particular-fire situation
demands.

36.
 APPENDIX

Sizing of Heat Venting 'Stacks for Natural Convection

The design of heat venting stacks can be approached by considering

the stack operation as a steady flow process. A general energy balance
for the system between state 1 at the stack entrance and state 2 at the
5
stack exit then takes the form1

(2- PI) + (K2 -K 1) + (Wf2 -Wfl) + (U2 -U 1) Q-W, (1

where

(2- P1 .)=potential energy change

(K2 - K1 ) =kinetic energy change

(Wf2 - Wf1 ) system flow work

(U2 -U 1) internal energy change

Q =heat input to the fluid

W =work done by the fluid.

In order to evaluate this energy expression, it is necessary to cast

each term in a farm appropriate to the physical. system being considered.
Accordingly, it can be observed that the gases flowing in a stack produce
no useful work, thus eliminating the work term. The potential energy
change is defined in terms of the force required to overcome the gravita-
tional field in changing the elevation of the fluid from points 1 to 2.
This requires accounting for the buoyancy of the flue gases. For a given
volumetric flow rate F, a unit volume of air is displaced by each unit
volume of flue gas. The bouyant mass flow then can be written as

W
w=F (Y - Ya) - 1 9(Y 9- ya)
g a y g

37
where

3
Ygl Ya =specific weight of flue gas and air, lb/ft
w, wg =mass flow rate, lb/win

F =volumetric flow rate, ft3 /min.

The potential energy term then becomes

w
(2- PI) -I}, (Yg - Ya (Z2 - Z1 )

where

Zl'2 =elevation of points 1 and 2, in feet.

The change in kinetic energy in the stack is then written in terms of the
velocities

where

go= standard gravitational acceleration, ft/sec2

=lvvelocity at points 1 and 2, ft/sec.

Flow work is a consequence of the movement of the gases against pressure

at the two reference points in the system. It is expressed as

(Wf -2(
Wf1 w g\ Y
p2
2
2 T1,

38
where

2
PP2=pressure at points 1 and 2, lb/ft .

Substitution in Eq (1) gives

Wg Y - Y) (z2 - z) + V
2 Y

+ Wg (U2 7
*-U 1) Q .(2)

This is the expression for the general energy balance of the stack under
steady flow..

Some further conditions can be imposed on Eq (2) to put it in a form

that will be useful for stack design. The expression is for steady state
operation of the stack, which implies that stable flow has been estab-
lished and that initial transients have died out. Thus, it is reasonable
to assume that the wall temperature of the stack has approached that of
the flue gases. It can also be assumed that the stack would not be exceed-
ingly high and that the'time required for the gases to pass through the
stack is therefore relatively short. The heat loss from the flue gases to
the walls should consequently be small compared to the total heat carried
by the gases. Accordingly, the heat transfer term, Q, can be set to zero.
The same arguments applied to the third term lead to the assumption that
the density change of the gases between points 1 and 2 is minimal, Thus,
Y can be substituted for Yand Y2 . This is equivalent to an assumption
g
of incompressible flow in the stack. Equation (2) can then be recast in
this form:

(Y ,Y ) (Z2 - Z) +- -- (v2 _ v2) +(p-p

1/ (2 1l
g a 2 1 2g0 \2

+yY U2 -U 1 ) =0. (3)

39
 The change in internal energy can be attributed to friction losses
within the stack. In order to obtain a practical expression for friction
losses as well as to include consideration of other configuration-related
losses of the stack system, it is appropriate to introduce some of the
techniques used in the design of duct systems into Eq (3). 16 In duct
design, friction losses are usually expressed in terms of the velocity of
the flowing fluid and the Moody friction factor:

Pf 'g")

where

2
Pf = pressure loss due to duct friction, lb/ft

f = Moody friction factor, dimensionless

L = length of duct, ft

D = diameter of duct, ft

pv=velocity pressure, lb/ft2

The Moody friction factor is a function of the re-lative roughness of the

duct and the Reynolds number indicative of the degree of turbulence of the
flow. its value can be found by reference to a generalized flow resis-
17
tance diagram.

Duct configuration losses, such as bends and entrance and exit

losses, are accounted for by inclusion of an empirical loss coefficient
for each fitting. The product of this coefficient and the local velocity
pressure gives the pressure loss due to flow through the feature being
considered. Chapter 31 of the ASHRAE 1977 'Fundamentals Handbook 18
tabulates a number of fitting loss coefficients, including those appli-
cable to the stack design problem. In applying the loss coefficients it
is necessary to carefully observe the velocity to which each given
coefficient is referred. For the purposes of this work, it is assumed
that the stack is one of constant cross section, which, in turn, implies

40
Lhat a constant gas velocity applies from entrance to exit. The appli-
cable ASHRAE coefficients are such that the reference velocity is that of
the stack gases. Consequently, a combined friction and system configura-
tion pressure loss expression can be written as

= v : + f ,(4)

where

2
PL combined pressure loss, lb/ft

Ci individual configuration loss coefficients.

The second term of Eq (3) represents the change in kinetic energy

during stack transit. If the velocity of the gases in the vented com-
partment at some point near the entrance to the stack is assumed to be
small compared to the velocity in the stack, then v, can be taken as zero
and v2 becomes the gas velocity in the stack. Thus, the term represents
the exit pressure loss with an implied coefficient of unity which is
appropriate for a stack exit without a tapered diffuser section.
.Combining the kinetic and internal energy terms of Eq (3) in the form of
Eq (4), and noting that.(Z2 -zl) and L are equal to the stack height,
yields

.2
H gY y 2g (3C (5)
-P 2gy
--- + (2 -l 1 05

where

H =stack height above entrance, ft.

T'his result is more conveniently applied to the present problem if the

stack velocity is stated in terms of the volumetric flow, rate, using

41
 g 1517D2

where

vg flue gas velocity, ft/sec

Qg stack flow rate, ft3 /min

D = stack diameter, ft~.

Substitution in Eq (5) and rearranging gives

2 -p I{( Yg 2
g0D5 \''~ + (6)
- -4O17
/H

Solving this expression for the volumetric flow rate results in:

1/2
40r2 90D5
(7)
Qg(y = [( Yg (P2 - P

Evaluation of this equation is inconvenient because of the presence

of the gas densities. These can be expressed in terms of temperature if
the air and flue gases are considered ideal gases by invoking the Ideal
5
Gas Law in the form'

Y=p
=RT

.where

p = gas pressure, lb/ft2

R = gas constant, ft-lb/lb-*R

T = absolute gas temperature, *R.

42
The gas pressures throughout the system are essentially atmospheric. It
is further noted that the equivalent molecular weights of air and flue gas
are very nearly identical, 1 9 which allows the use of a single gas constant
for both constituents. Making the substitutions for Ya and Y9in Eq (7)
gives the following:

t45O 2r gOD5 rR(T

Lg(ZC - T)a
Ta RT A(
Pa 2 -. 1, / 9 8

where

Pa = atmospheric pressure, lb/ft2 .

Now examine the significance of the second term in the bracketed

expression. Reference to Eq (2) shows that p1 and P2 are the pressures at
the inlet and outlet of the stack, respectively, during steady flow
conditions. If, by some means, the fire compartment were to be supplied
with air by a completely unrestricted path, then both the inlet and outlet
of the stack would be at atmospheric pressure and the term would vanish.
This could be done, for example, by supplying air at exactly the right
flow rate through a fan powered input duct system. A more likely
scenario, however, would involve drawing the air into the compartment friom
an adjacent plant area through an open door or some other opening. In
this case, the pressure in the compartment would be subatmospheric, i.e.,
Pl< P2 , and the system flow rate would be reduced by this term.

Assuming a single opening into the compartment and considering it a

square-edged orifice, the pressure drop acros.s the opening can be written
as

(2- - y(v)
go

43
where

C0 = orifice coefficient referred to the velocity through

the orifice

Va = velocity of air through the orifice, ft/sec.

Since the underlying premise of the heat venting system is that sufficient
excess air is provided to the fire compartment to limit peak temperatures,
it will be assumed that the flow rate of combustion air into the fire is
small compared to the total air flow. The mass flow rate into the compart-
ment must then equal the outflow and

Qa QTa Qg (10)
ga Qg

Now expressing the orifice velocity in terms of the stack flow rate and
the orifice area gives:

C p TQ2
(p 2 -PI1 2 92'
7200A g0 RT

where

A0 area of air supply opening, ft.

Substituting this result in Eq (8) and solving again for the volumetric
flow rate of the stack gases yields the final expression:

450v290 D 5H(T 9.- T a)/

Qg{T[(D7 a2 (1T

44
where

H = stack height, ft

Qg = stack flow rate, ft 3 /min

2
go= standard gravitational acceleration, ft/sec
D =stack diameter, ft

TaTg =absolute air and flue gas temperatures, OR

f =Moody friction factor, dimensionless

Co orifice coefficient of compartment air supply

opening, dimensionless

C. individual configuration loss coefficients,

1dimensionless

A= area of compartment air supply opening, ft.

Evaluation of Eq (11) for a range of stack heights and diameters,

stack configuration, compartment air supply conditions, and ambient and
peak temperatures produces a corresponding range of stack flow rates.
Estimates of the potential heat production in a fire area can then be used
to determine the stack flow rate required to limit peak temperatures to
acceptable value-s. The appropriate combination of stack parameters can
then be selected, based on the required flow rate. The presence of the
Moody friction factor in the equation requires that some degree of con-
servatism. be introduced by selection of a suitably high value for this
parameter. A check on the validity of the assumed value can be made as a
secondary calculation once the velocity of the stack gases is determined
for a given design situation.

Tables 1 and 2 give the stack flow rates resulting from evaluation of
Eq (11) for stack diameters from 1 to 10 ft, stack heights from 50 to
200 ft, and gas temperatures from 2000 to 800*F. Table I represents the
case in which make-up air is somehow supplied to the compartment at a rate
which maintains the compartment at atmospheric pressure. As such, the
flow rates are the maxima attainable by natural stack action. The values
in Table 2 include the assumption that air is being drawn into the room b y

45
stack action through an opening of 20 ft2 , approximately that of a fully
opened door. The expression is quite sensitive to the inlet area assump-
tion so the tabulated results should be viewed only as indicative of
passive air supply. A constant friction factor, ,f= 0.025, was assumed
for both sets of calculations. over the range of stack flows obtained,
this is a conservatively'high value representative of a stack liner of
relatively rough concrete. The assumed configuration is that of a cen-
trally located stack with dampered openings of the same cross-sectional
area as the stack into each of the compartments served. The loss co-
efficients were taken from Reference 16 and the orifice coefficient for
the doorway is from Reference 18.

It should be noted that no leakage losses were included in the

calculations. Tamura and Shaw have shown that leakage can reduce the flow
rates by a factor of two if the total leakage fraction is as little as 3%
of the stack flow. 7 If a central stack were to be designed to vent
several compartments, only-the damper in the fire cell would-be actuated,
while all others would remain in their normally closed condition. All of
these dampers must be very carefully designed to assure minimum leakage.

46
 Table 1

Predicted Smoke Vent Flow Rates*

(Unlimited Inlet Area)

Stack Height = 50 ft Stack Height = 100 ft

Di am Flaw Rate (cubic feet/minute) at Di am Flow Rate (cubic feet/minute) at

(ft) 2000 F 4ooOF 600OF 800OF (ft) 200OF 400OF 600OF 800*F

1 634 1 035 1 320 1 553 1 781 1 276 1.627 2 105

2 2 765 4 515 5 756 6 773 2 3 587 5 859 7 468 10 001
3 6 426 10 494 13 377 15 741 3 8 535 13 938 17 768 24 218
4 11 620 18 975 24 189 28 463 4 15 642 25 543 32 562 44 866
5 18 348 29 962 38 194 44 943 5 24 913 40 684 51 862 71 98.2
6 26 609 43 453 55 392 65 180 6 36 352 59 363 75 673 105 585
7 36 405 59 450 75 784 89 175 7 49 959 81 583 103 998 145 683
8 47 735 77 952 99 370 116 928 8 65 734 107 344 136 837 192 280
9 60 600 98 959 126 149 148 439 9 83 679 136 647 174 191 245 379
10 74 998 122 472 156 122 183 708 10 103 7.92 169 492 216 -061 304 983

Stack Height = 150 ft Stack Height = 200 ft

Di am Flow Rate (cubic feet/minute) at Dian Flow Rate (cubic feet/minute)- at

(ft) 2000 F 400OF 600OF 800OF (ft) 200OF 400*F 600OF 800OF

1 859 1 403 1 789 2 105 1 908 1 483 1 .891 2 225

2 4 083 6 667 8 499 10 001 2 4 422 7 221 9 205 10 832
3 9 887 16 145 20 581 24 218 3 10 1U358 17 731 22 603 26 597
4 18, 316 29 910 38 128 44 866 4 20 296 33 .143 42 249 49 715
29 386 47 988 61 173 71 982 5 32 766 53 507 68 209 80 261
6 43 105 70 390 89 730 105 585 6 48 283 78 847 100 118 271
7 59 474 97 122 123 807 145 683 7 66 855 109 175 139 171 163 76.2
8 78 498 128 186 163 406 192 280 8 88 486 144 497 184 199 216 746
9 100 175 163 586 208 532 245 379 9 113 178 184 820 235 600 277 230
10 124 509 203 322 259 186 304 983 10 140 934 230 145 293 378 345 218

*Amnbient Ai-rTemperature = 80*F; Friction Factor = 0.025

Loss Coefficients: Entrance = 0.5; Exit = 1.0; Stack Tee' = 1.2

47
 Table 2

Predicted Smoke Vent Flow Rates*

(Limited Inlet Area)

Stack Height = 50 ft Stack Height = 100 ft

Diam Flow Rate (cubic feet/minute) at Di am Flow Rate (cubic feet/minute) at

(ft) 2000 F 4000 F 600OF 800OF (ft) 200*F, 400OF 600OF 800OF

1 634 1 035 1 319 1 553 1 781 1 276 1 627 2 105

2 2 744 4 489 5 729 6 746 2 3 565 5 830 7 439 9 972
3 6 178 10 178 13 048 15 413 3 8 242 13 567 17 381 23 818
4 10 319 17 281 22 393 26 654 4 14 031 23 451 30 348 42 464
5 14 205 24 359 32 096 38 680 5 19 597 33 523 44 090 63 149
6 17 156 30 138 40 451 49 466 6 23 909 41 906 56 144 82 347
7 19 105 34 214 46 658 57 826 7 26 791 47 897 65 222 97 643
8 20 311 36 863 50 865 63 700 8 28 583 51 817 71 424 108 583
9 21 047 38 535 53 600 67 622 9 29 677 54 296 75 469 115 953
10 21 502 39 591 55 362 70 197 10 30 353 55 '861 78 077 120 807

Stack Height = 150 ft Stack Height = 200 ft

Di am Flow Rate (cubic feet/minute) at Diam Flow Rate (cubic feet/minute) at

(ft) 2000 F 400OF 600OF '8000 F (ft) 200OF 400OF 600OF 800OF

1 859 1 403 1 789 2 105 1 908 1 483 1 891 2 225

2 4 060 6 639 8 470 9 972 2 4 401 7 194 9 178 10 805
3 9 582 15, ,759 20 179 23 818 3 10 553 17 346 22 202 26 198
4 16 569 27 648 35 738 42 464 4 18 493 30 814 39 793 47 249
.5 23 442 40 010 52 537 63 149 5 26 465 45 078 59 105 70 966
6 28 868 50 491 67 531 82 347 6 32 875 57 384 76 630 93 325
7 32 540 58 080 78 978 97 643 7 37 267 66 412 90 189 ill 376
8 34 836 63 083 86 865 108 583 8 40 032 72 411 99 611 124 406
9 36 241 66 255 92 028 115 953 9 41 726 76 227 105 807 133 230
10 37 108 68 259 95 362 120 807 10 42 771 78 639 109 812 139 054

*Ambient Air Temperature = 80'F; Friction Factor = 0.025

Loss Coefficients: Entrance = 0.5; Exit = 1.0; Stack Tee = 1.2

Compartment Air Inlet: Area = 20 ft; Orifice Coefficient = 2.5

48
Examples of Vent Stack Applications

To illustrate the use of Tables 1 and 2 and to investigate the feas~i-

bility of heat removal by stack action, some sample calculations were made
for a postulated fire, the results of which are presented here. The heat
removal capability of air can be expressed as

q wc p (T 2 -Tl)

where

q = heat removal capability, Btu/min

cp= specific heat of air at constant pressure, Btu/lb-*R

Tl, T2 =initial and final air temperatures, *R

w =mass flow rate, lb/min.

The specific heat is defined by

h 2-h1
-2 T2 1,

where

=~h enthalpy at T2 and TI, Btu/lb.

Making this substitution and applying the previously stated form of the
Ideal Gas Law, the following expression for the required flow in the stack
can be written as follows:

Qf + RT (12
Qg P a(h 2- h)1(2

49,
where

Qg required stack flow rate, ft /min

Qf = heat evolved by postulated fire, Btu/min

R = gas constant, ft-lb/lb-*R

TF= assumed allowable maximum temperature, OR

2
Pa =atmospheric pressure, lb/ft

h2l= enthalpy of air at maximum and ambient te mperatures,

Btu/ lb.

For purposes of estimating stack size requirements, Eq (12) can be simpli-

fied by assuming standard atmospheric pressure and R = 53.3 which gives

Qg = (2.5186 x 102) qfT (13)

The values for enthalpy of air at various temperatures can be found in

standard gas tbe.0Equation (10) can be used to estimate the required
make-up airflow rate. Equation (9) will give the value of negative
differential pressure in the compartment unde?"atrl tckatin we
air is supplied through an assumed opening for a given supply air velocity
through the inlet opening. Again using the standard air conditions stated
above and an orifice coefficient of 2.5, Eq (9) become s

2
Ap -(8.2334 x 10 ) _! (14)
T
a

where

Va =inlet air velocity, ft/mmn.

Using these expressions, consider a 5 ft2 gasoline fire in a

compartment 50 ft square by 15 ft high. NFPA 204 gives a pool-burning

50
heat evolution rate of 10 000 Btu/ft 2 -min for gasoline, which represents a
heat input to the compartment of 50 000 Btu/min. A stack height of 100 ft
was assumed. Stack sizes were taken from Table 2 which implies an ambient
temperature of 800 F and an air inlet opening of 20 ft2 . The results of
these calculations are shown in Table 3.

Table 3

2Example Heat Venting Results

(5ft2 gasoline fire in 37 500 ft3 compartment)

Assumed maximum permissible 200 400 600

temperature, OF
Enthalpy change, Btu/lb 29 76 126

Required gas flow in vent 28 660 14 250 10 594

stack, ft3 /min

Required air flow into 23 450 8 950 5 397

compartment, ft3/min

Minimum stack diameter 8.0 3.1 2.5

from Table 2, ft

Velocity in stack, ft/mmn 570 1 888 2 158

Velocity through 20 ft2 1172 447 269

compartment air inlet,
ft/mmn

Air change rate at 80*F, 37.5 14.3 8.6

vol/hr

Compartment differential -0.210 -0.031 -0.011

pressure, in; H20

As might be expected, very large flow rates, and hence large stacks,
would be required to maintain low peak temperatures. In this example, a
peak permissible temperature of 400OF results in a reasonable stack
diameter and flow rates. If a stack height of 50 ft had been assumed,
Table 2 suggests that the 200*F temperature could not be maintained.
Reference to the flow rates in Table 1 for the 50-ft stacks shows that
this latter case is governed by the limitations of the compartment air
inlet.

51
 Re ferences

1. E. A. Bernard and G. L. Cano, Report on Task I - Fire Protection

Study, Sandia Laboratories, SAND76-0630 (NUREG 766516), February
1977.

2. Standard Test Method-for Measuring the Smoke Generated by

Solid Materials, National Fire Protection Association, NFPA 258,
1976.

3. A. F. Robertson, "Estimating Smoke Production During Building

Fires," Fire Technology, Vol. 11, No. 2 (pp. 80-93) May 1975.

4. E. E. Smith, "Measuring Rate of Heat, Smoke and Toxic Gas

Release," Fire Technology, Vol. 8, No. 3 (pp. 237-245), August
1972.

5. E. E. Smith, "Application of Release Data to Hazard Load

Calculations," Fire Technology, Vol. 10, No. 3 (pp. 181-186),
August 1974.

6. E. E. Smith, "Relation of Performance Tests to Actual Fires,"

Fire Technology, Vol. 12, No. 1 (pp. 49-54), February 1976.

7. G. T. Tamura and C. Y. Shaw, "Basis for the Design of Smoke

Shafts," Fire Technology, Vol. 9, No. 3 (pp. 209-222), August
1973.

8. F. C. W. Fung, "Smoke Coutrol by Systematic Pressurization,"

Fire Technology, Vol. 11, No. 4 (pp. 261-268), November 1975.

9. T. Z. Harmathy, " A New Look at Compartment Fires, Parts I and

II," Fire Technology, Vol. 8, No. 3, (pp. 196-217), August 1972
and Vol. 8,No. 4 pp. 326-351), November, 1972.

10. T. Z. Harmathy, "Design of Buildings for Fire Safety, Parts I

and II," Fire Technology, Vol. 12, No. 2, (pp. 95-108), May 1976,
and Vol. 12, No.3 (pp. 219-236 , August 1976.

11. T. Z. Harmathy, "Fire Resistance versus Flame Spread Resistance,"

Fire Technology, Vol. 12, No. 4 (pp. 290-302), November 1976.

12. J. Quintiere, "Gro wth of Fire in Building Compartments," Fire

Standards and Safety, ASTM STP 614, A. F. Robertson, Ed.,
American Society for Testing and Materials (pp. 131-167), 1977.

13. T. E. Waterman, "Room Flashover--Criteria and Synthesis,"~

Fire Technology, Vol. 4, No. 1 (pp. 25-31), February 1968.

52
14. 3. B. Fung, "Fire Buildup in a Room and the Role of Interior
Finish Materials" NBS Technical Note 879, National Bureau of
Standards, Washington, DC, June 1975.

15. V. M. Faires, Thermodynamics, Fifth Edition, MacMillan Co.,

London, 1970.

16. ASHRAE Handbook of Fundamentals, 1977 Edition, American

Society of Heating, Refrigerating, and Air-Conditioning
Engineers, Inc., New York, Chapter 31.

17. Marks' Standard Handbook for Mechanical Engineers, Seventh

Edition, T. Baumeister, Editor, McGraw-Hill Book Co.., New
York, p 3-60, 1967.

18. ASHRAE Handbook of Fundamentals, 1972 Edition, American

Society of Heating, Refrigerating, and Air-Conditioning
Engineers, Inc., New York, Chapter 25.

19. Steam, Its Generation and Use, 38th Edition, Babcock &
Wilcox Co., New York, p 17-2, 1975.

20. P. J. Potter, Power Plant Theory and Design, Second

Edition, The Ronald Press Co., New York, 1959.

53
DISTRIBUTION:
U. S. Nuclear Regulatory Commission (RP')
David P. Notley, USNRC (50)
Sandia Laboratories
P. 0. Box 5800
Albuquerque, NM 87185
Attn: J. E. Christy (50)

54
 NRC FORM 3W3 .. NCERRGUAOYCMISO 1. REPORT N UMBER (Asignedby DDCj
(7-77) US ULA EUAOYCMISO

BIBLIOGRAPHIC DATA SHEET NUREG/CR-0636.

4. TITLE AND SUBTITLE (Add Volumne No., if appropriare) 2 (Leave blank)
Nuclear Power Plant Fire Protection - Ventilation
(Subsystems Study Task 1) 3 RECIPIENT'S ACCESSION NO

7. AUTHOR (S) 5 DATE REPORT COMPLETED

John A. Milloy MNHAugust YER1979

9. PERFORMING ORGANIZATION NAME AND MAILING ADDRESS (Include Zip Code) DATE REPORT ISSUED

Sandia Laboratories MO979IEA

Albuquerque, New Mexico 87185 September YEAR
6 (Leave blank)

8 (Leave blank)

12 SPONSORING ORGANIZATION NAME AND MAILING ADDRESS (Include Zip Code) 1 RJC/AKWR NTN
Engineering Methodology Standards Branch PRET/AKWRUTN
Office of-Standards Development 11 CONTRACT NO
U.S. Nuclear Regulatory Conmmission
Washington, D. C. 20555 FIN No. A-1080
13 TYPE OF REPORT PERIOD COVERED (Inclusive dates)
Final TechnicalI

16 ABSTRACT 000O Words Or less)

This report examines the role of compartment ventilation from the standpoint of nuclear
power plant fire protection safety. Based on a review of numerous design standards
-which consider the influence of controlled ventilation on reducing fire severity, the
report concludes that current standards and regulatory guidelines inadequately define
criteria for design of ventilation systems and their operation under fire emergencies.
To resolve this deficiency, the report evaluates four candidate design bases for
ventilation systems: smoke removal, smoke dispersion control, fire spread control,
and fire temperature control. It is concluded that the lack of existing fire technology
precludes the implementation of all but one of these criteria--fire temperature control.
On this basis the report presents an example design calculation for applying the
temperature control criterion.

17 KEY WORDS AND DOCUMENT ANALYSIS 17a DESCRIPTORS

Fire
Ventilation

17b IDENTIFIERSIOPEN-ENDED TERMS

18. AVAILABILITY STATEMENT 19 SECURITY CLASS (This report) 21 NO OF PAGES

Unlmied20 SECURITY CLASS ('This page) 22 PRICE

IUnrlagifiprF S
NRIC FORM 335 (7-77)