Report of the Committee on Explosion Protection … COMMITTEE ON EXPLOSION PROTECTION SYSTEMS Report...

EPS-1 C O M M I T T E E ON E X P L O S I O N P R O T E C T I O N SYSTEMS

Report of the Committee on Explosion Protection Systems-

F. W. Wischmeyer, Chairman

Eastman Kodak Company, Industrial Safety Dept. Rochester, NY 14650

A. R. Albrecht, Manufacturing Chemists Assn. Daniel J. MacKay, Jr., National Assn. of Fire Charles F. Averill, National Automatic Sprinkler & Equipment Distributors

Fire Control Assn. John Nag),, Mining Enforcement & Safety Admin- J. D. Dick, American Mutual Insurance Alliance istration Dean H. Gaugler, Industrial Fire Protection Sec- ...Robert W. Nelson, Industrial Risk Insurers

tion of NFPA Bernard Pesetsky, Union Carbide Corporation

George J. Grabowski, Fire Equipment Manufac- J .C. Rabuck, American Petroleum Institute turers Assn. . Anthony Santos, Factory Mutual Research Corp.

3 Dr. Waiter B. Howard, Monsanto Co.. R.F. Schwab, Allied Chemical Corporation

W. Paul Jenson, Aeroject Nuclear Co. Robert B. Ziegler, American Insurance Assn.

Alternates .

L. A. Bilker (Alternate to R.B. Ziegler) D.B. Tucker, Jr. (Alternate to R. W. Nelson)

J. F. McKenna (Alternate to J.C. Rabuck) H. Veralds (Alternate to John Nagy)

R. A. Mitchell (Alternate toJ. D. Dick)

The report of the Committee on Explosion Protection Systems is presented in two parts.

Part I presents for Reconfirmation the 1973 Edition of Standard on Explosion Prevention Systems, NFPA 69.

Part I has been submitted to letter ballot of the Committee on Explosion Protection Systems. Of 16 voting members, 15 voted affirmatively. One member (Mr. Rabuck) wished to be recorded as not voting.

Part II presents for official adoption a complete revision of the 1974 Edition of Guide for Explosion Venting, NFPA 68.

Part II has been submitted to letter ballot of the Committee on Explosion Protection Systems. Of 16 voting members, 15 voted affirmatively. One member (Mr. Rabuck) wished to be recorded as not voting.

Part I Proposed Reconfirrnation of the

Standard on

Explosion Prevention Systems NFPA 69-1973

The Committee on Explosion Protection Systems recommends that the 1973 edition of the captioned standard be reconfirmed as suitable for current use. References to existing NFPA standards will be updated in the reconfirmed edition, but no substantive change is considered necessary. Copies of NFPA 69-1973 are available in separate pamphlet .... form and the text appears in Volume 5 of the National Fire Codes, 1978.

I N T R O D U C T I O N 68-5

Part II

Guide for

Explosion Venting

NFPA 68-1978

Chapter 1 Introduct ion

1-1 Scope.

1-1.1 This guide provides useful information for the design and utilization of devices and systems to vent the gases resulting from deflagrations of dusts, gases, or mists in equipment, rooms, buildings, or other enclosures so as to minimize structural or mechanical damage.

1-1.2 This guide does not include venting devices designed to protect against overpressure of vessels contMning liquids, liquefied gases, or compressed gases under fire exposure conditions, as now covered in existing NFPA standards.

1-1.3 This guide is not intended to be used for calculating venting or emergency relief for exothermic runaway reactions nor does it cover unconfined combustion such as free air explosions or outdoor vapor cloud explosions.

1-2 Purpose. This guide is intended to provide the user with the best available criteria for venting of deflagrations.

1-3 General.

1-3.1 Vents do not prevent the occurrence of a deflagration but are intended to limit the damage from the pressure excursion generated by the deflagration.

1-3.2 This guide applies to the deflagration of combustible dusts, gases, or mists when mixed with air during manufacturing operations and storage.

1-3.3 Typical examples of industrial equipment to which this guide applies include crushers, grinders, pulverizers, sieves, screens, bolters, dust collectors and arrestors, conveyors, screw feed con-. veyors, bucket elevators, driers, ovens and furnaces, spray driers, blenders, mixers, ducts, pipes, bins, silos, spreaders, coating machines, and packaging equipment.

68-6 E X P L O S I O N VENTING GUIDE

1-3.4 This guide is not in tended to cover explosion vents on equ ipment such as o iMnsula ted transformers and switchgear or excess pressure relief devices on tanks, pressure vessels, or domestic ' appl iances in residences.

1-3.5 Explosion vent ing has not yet been reduced to an exact science. This guide only a t tempts to r ecommend what is" thought to be the best avai lable knowledge on the subject at this time.

1-4 Def in i t ions . For the purposes of this guide, t h e following terms have the meanings shown below.

A u t o - I g n i t i o n T e m p e r a t u r e . The ignit ion t empe ra tu r e of a substance, whether solid, l iquid or .gaseous, that is tl)e m i n i m u m tempera tu re requi red to_initiate or cause self-sustained combust ion independen t ly of the heat ing or heated element. T h e ignition t empera tu re of a solid is influenced by its physical condi t ion and the rate of heat ing. F!gures on ignit ion tempera tures m a y vary, depend ing upon the test haethod, since the ignition t empera tu re varies with the size, shape and mater ia l of the testing container, and other factors.

B l a n k e t i n g . The technique of main ta in ing an a tmosphere which is inert or enr iched with a fuel above a .liquid in a conta iner or

YTCO~TIglOn VC~3EI. ~Jg¢~ O ~ a l t a U l L l Oli L X ] ) t U S l O l l oys~ems', NFPA ~J.~- t ~/04

C o m b u s t i b l e . Used synonymously with the term fuel, combustible means a gas or mist or dust capable of being burned. B u r n i n g can be descr ibed as the chemical react ion of a combust ible and gaseous oxidant (normal ly the oxygen of a i r ) w i t h resultant product ion o f a flame.

W h e n a combust ib le is i n t ima te ly mixed with an ox idan t and ignited, burning in the form of def lagrat ion or de tona t ion may result.

De f l ag ra t ion . Burning which takes place at a f lame speed below the velocity of sound in the unburned medium.

D e t o n a t i o n . Burning which takes place at a f lame speed above the velocity, of sound in the unburned medium.

Dus t ( i ndus t r i a l ) . Any finely d ivided solid mater ia l 420 microns or smal ler in d iamete r (mater ia l passing a U. S. No. 40 S tandard Sieve).

Explos ion . A burst ing of a bui lding or conta iner as a result of deve lopment of internal pressure beyond the confinemeiat capabi l i ty of the bui ld ing or container .

INTRODUCTION 6 8 - 7

Fi r e Po in t . The lowest t empera tu re of l iquid in an open conta iner at which vapors are evolved fast enough to suppor t con- t inuous combust ion. I t is de te rmined by Standard Method of Test for Flash and Fire Points by Cleveland Open Cup, A S T M D92-1977.

H a m e . S p e e d or F l a m e V e l o c i t y , i s the speed at which the flame front progresses through the unburn t mixture.

F l a m m a b l e Limi ts . In the case of most f l a m m a b l e liquids, gases, dusts, and mists, a m i n i m u m concentra t ion of gas, dust, or mist in air, oxygen, or o the r oxidant below which p ropaga t ion of flame does not occur on contact with a source of ignition. Usually, there is also a m a x i m u m concentra t ion o f ' ga s in air, oxygen, or other ox idant above which p ropaga t ion of flame does not occur. These l imit mixtures of gas, dust, or mist with ox idant which, if ignited, will jus t p ropagate flame are known as " lower and uppe r f lammable l imits ." In the case of gases, limits are usual ly expressed in terms of percentage by volume of gas in oxidant : In the case of dusts, l imits are usually expressed in terms of ounces of dust per cubic foot of volume or mg/ l i te r . The lower l imit for dusts is a function of dust par t ic le size. The upper l imit for a dust usual ly canno t be well defined. T h e lower l imit for mists is a funct ion of mist part icle size; the upper l imit usually cannot be well defined.

F l a m m a b l e R a n g e . The concentra t ion range ly ing .be tween the !owe? and upper exp!osi-,e or f l ammable limits.

F l a sh Po in t . The flash point of a l iquid shall mean the m i n i m u m tempera tu re at which it gives off vapor in sufficient concent ra t ion to form an ignit ible mixture with air near the surface of the l iquid within the vessel as specified by app rop r i a t e test procedure and appa ra tus as follows:

The flash point of a l iquid having a viscosity less than 45 SUS t at 100°F (37.8°C) and a flash point below 200°F (93.4°C) shall be de te rmined in accordance with the Standard Method of Test for Flash Point by the Tag Closed "Tester, A S T M D56-752. .

The flash point of a l iquid having a viscosity of 45 S U S ' o r more at 100°F (37.8°C) or a flash point of 200°F (93.4°C) or higher shall be de te rmined in accordance with the Standard Method of "Test for Flash Point by the Pensky-Martens Closed Tester, A S T M D93-732.

ISaybolt Universal Seconds at 100°F (37.8°C). A viscosity of 45 SUS is about the viscosity of No. 4 Fuel Oil.

- 2Available from American Society for Testing and Materials, 1916 Race Street Philadelphia, PA 19103.

68-8 E X P L O S I O N V E N T I N G G U I D E

Fundamenta l B u r n i n g Velocity. The velocity of the gas normal to the flame front with which the unburnt mixture enters a flame and is chemically transformed. (This velocity is determined from laminar flow conditions in carefully controlled apparatus.)

Gas. In this guide, gas includes vapors.

I ne r t Gas. A gas which is noncombustible, nonreactive, and incapable of supporting combustion with the contents of the system being protected.

Ine r t ing . The process of rendering a combustible mixture noncombustible through the addition of an inert gas.

O p e n - V e n t Pressure . The pressure developed by a deflagration in a container having an unobstructed vent.

~: O p t i m u m Mix tu re . - A mixture in which the combustible material and oxidant are in the proper proportion to give the most violent deflagration (that is to say, the deflagration with the highest maxi- m u m rate of pressure rise). Generally, this occurs at approximately the stoichiometric proportions.

Ox idan t (oxidiz ing agent) . Any material or substance that can react with a combustible to produce burning or combustion, or a s imilar 'exothermic reaction. Oxygen in air is the most common oxidant.

P a d d i n g . . (See Blanketing.)

P u r g e Gas. A gas suitable for rendering an atmosphere noncombustible. I t m a y be inert or combustible. Air can also be used as a purge gas.

P u r g i n g . The displacement of a gaseous oxidant or gaseous combustible by another gas to render the mixture noncombustible. The purge gas may or may not be an inert gas.

Rate of P re s su re Rise. The amount of pressure rise during a particular interval of a deflagration. I t is expressed as the ratio of the increase in pressure to the time interval (dP/dt) required for that increase of pressure to occur. The "average rate of pressure rise" is the ratio of the maximum pressure to the time interval f rom the initiation of the deflagration until the max imum pressure is reached, and the ? m a x i m u m rate of pressure rise" (dPmax/d t ) is computed f rom the slope of the steepest portion of the pressure-time curve during the development of the deflagration.

Explosion Suppress ion. A technique by which burning in a confined mixture is detected and ari'ested during incipient stages, preventing development of pressure which could result in an explosion.

I N T R O D U C T I O N 68--9

Vapor . As specified in this guide, vapor means a gas.

V e n t Ratio. The relationship of the area of the rupture diaphragms or relieving panels to the volume of the equipment or room subject to internal deflagration. Vent ratio may be expressed in terms of "square feet per 100 cubic feet" or as the reciprocal of the cubic feet of vented volume per square foot of vent.

Venti la t ion. The process of supplying or removing air, by natural • or mechanical means, to or f rom any space.

Gene ra l Vent i la t ion. The removal of combustibles by moving air through the entire volume of space. (See Appendix A, Part I, NFPA 69-1978.)

Local Vent i la t ion. The removal of combustibles f rom a small portion of a space and more particularly at the immediate vicinity of emission by withdrawing air f rom that small portion of the space. (See Appendix A, Part I, NFPA 69-1978.)

68-10 E X P L O S I O N V E N T I N G G U I D E F U N D A M E N T A L S O F C O M B U S T I O N 68-11

C h a p t e r 2 F u n d a m e n t a l s o f C o m b u s t i o n

2-1 P r e r e q u i s i t e s for a Def lagra t ion. Under the .proper conditions, f lammable and combustible gas, mist, and dust mixed with or suspended in air or other o x i d a n t will burn when ignited. The following prerequisites are necessary for a deflagration to occur :

(a) fuel (mixed in the proper proport ion with the atmosphere);

(b) air (oxygen) or other oxidant ; and

(c) a source of ignition such as a flame, spark, heated surface, or glowing particle.

For an explosion to occur, in addit ion to all the above requirements, the combustion of a gas or dust must generate a pressure greater than the structural capabil i ty of the confining structure.

2-2 Fac tors A f f e c t i n g a n E x p l o s i o n or D e f l a g r a t i o n .

2-2.1 The development of a deflagration or an explosion depends upon certain conditions and factors. Those of principal concern are :

combinat ion (hybrid mixture) of these. GeneralLy, gases burn more rapidly than mists or dus~.

1 0 0

~1 ~ 6o Z 0

4o F: z 8 2o

. . . . . . . Stoichiometric / mixture (decane-air) /

, - Kerosine spray ~ /

/ Kerosine vapor / and mist

20 . 40 60 80 100 120 140

DROPLET DIAMETER, microns

0.07

.o3 g ) .02 5

.01

0 160

Figure 2-2.1.2 (a). Variation in lower limits of flammability of various combustibles in air as a function of droplet diameter. (90)

2-2.1.2 F u e l C o n c e n t r a t i o n . Most gaseous fuels have a lower and upper f lammable l imit; the concentrat ion must be within these limits for a deflagration to occur. For dusts, the upper flammable limit is not well defined. [See Figure 2-2. ?.2(b).] Some mists can be deflagrated when the temperature of the mist is such that the corresponding vapor pressure will produce a concentrat ion less than the lower f lammable limit. With fine mists and sprays, the combustible concentrat ion at the lower limit is about the same as that in uniform vapor air mixtures. As the droplet diameter increases, the lower l imit appears to decrease, as illustrated in Figure 2-2.1.2(a).

100

80

60

0..

E E x 40

b

-t

I I I lo, oo6

__ ~ e __ 8,000

- - 5,000 ~ .

t-t" Maximum Rate

o_ 4,000

r r

I l I 1.0 2.0 3.0

2,000

4.0 Dust Concentration, oz per cuf t

Figure 2-2.1.2(b). Maximum pressure and rate of pressure rise developed by deflagrations of zirconium dust in a 0.043 cu ft closed vessel. (4O)

6 8 - 1 2 EXPLOSION VENTING GUIDE FUNDAMENTALS OF COMBUSTION 68--13 O

2-2.1.3 O x i d i z e r C o n c e n t r a t i o n . The oxidizer is normal ly the oxygen present in the a tmosphere . Oxygen concentrat ions higher than 21 percent intensify the react ion rate of combust ion and increase the probabi l i ty for t ransi t ion into detonat ion. Less than 21 percent oxygen decreases the rate of reaction. (See Figure 2-2. 7. ,3.)

100 20,000

8 0

60

o

E

• ~, 4 0

2 0

D

I E

/ /

Maximum Pressure / /

• / / / / ( Maximum Rate

• / / / / / / / / /

/ / /

/ / I I l I

5 10 15 20

Oxygen Concentration, Percent

__ 16,000

_ _ 12,000

_ _ 6,000

- - 4,000

25

E

E

Figure 2-2.1.3. Effect of oxygen concentration on pressure and maximum rate of pressure rise ofO.5 oz per cu ft atomized aluminum deflagrations in a closed 0.043 cu ft vessel. (54)

2-2 .1 .4 F u n d a m e n t a l B u r n i n g V e l o c i t y a n d F l a m e S p e e d . T h e destruct ive forces of a def lagrat ion increase with the pressure and velocity developed. The speed at which a fuel-oxidant mixture burns is dependen t upon the fundamen ta l burning velocity which

is a character is t ic of the fuel consumed and other factors. F u n d a - menta l burning velocities for most gases are on the order of 10 in. (25.4 cm) to 20 in. (50.8 cm) per second; values for dusts are genera l ly lower. (See Appendix A-4.2.) Flame speed m a y be many times higher but never lower than the fundamen ta l burning velocity.

F l a m e speed is most r ap id and the highest pressures are ob ta ined when the fuel concent ra t ion is o p t i m u m and uni formly d is t r ibuted th roughout the whole vessel or confinement. The p r ima ry difference between a dust and a gas is that burning t ime may be slightly longer for the dust.

Somet imes dusts p roduce explosions more disastrous than gases. This is due, in par t , to the slower flame speed and, therefore, longer dura t ion which results in greater total impulse dur ing the burn ing process.

Dur ing the initial stages of most deflagrations, there is an induc- tion or igni t ion-lag per iod a t t r ibu ted by some authori t ies to the ini t ia t ion of a chain reaction. T h e t ime that elapses between the instant the mixture is raised to its ignit ion t empera tu re and a visible flame appears is genera l ly called the igni t ion-lag period.

2-2.1.5 I g n i t i o n S o u r c e . The rate of pressure deve lopment and the m a x i m u m pressure increase as the s t rength of the ignit ion source increases. (See Figure 2-2./.5.) Locat ion of the ignit ion source at the geometr ical center of a confined fuel -oxidant mixture results in deve lopment of the highest pressure and rate of pressure rise.

Igni t ion can result f rom a hot surface, flame, or spark. In m a n y • cases the location of the ignit ion source cannot be predicted. Simul- taneous mul t ip le ignition sources may produce high init ial t u r b u - lence. An ignit ion may occur within one piece of equ ipmen t and be conveyed by a connect ing duct to a second piece of equipment . The f lame from the duct may enter the second piece of equ ipment in a highly tu rbu len t state. This presents a large ignit ion source in the second vessel resulting in more severe combust ion than normal ly would occur from a small ignition source. T h e ignit ion source in the case of dusts may produce a cloud. (See homographs, Appendix A, for effect of ignition energy.)

In order to ignite a gas or dust, a m i n i m u m a m o u n t of energy must be available. M i n i m u m energy required for ignit ion of some mater ia ls is listed in T a b l e 2-2.1.5. Da ta listed in the table were ob ta ined with dusts which passed through No. 200 U.S. S t anda rd Sieve. These values are only intended as a guide and should not be construed as absolute:

68-14 EXPLOSION VENTING GUIDE

110

t~

(3L

t2.

90

70

50

10 I 0 0.04

Key

1 Spark Ignition 2 100 mg Guncotton 3 250 mg Guncotton 4 400 mg Flash Powder

I I 0.08 0.12 0.16

Time, sec

Figure 2-2.1.5. Effect of ignition source on the d e v e l o p m e n t of def lagrat ions of 9.4 p e r c e n t methane -a i r mixtures in a closed 1 c u f t vessel. (57)

FUNDAMENTALS OF COMBUSTION 68--15

Tabl e 2-2.1.5 Minimum Energy for Ignition (29, 50)

Material Energy, millijoule

Mixtures wi th Air

Agricultural:

Alfalfa 320 Cinnamon 30-40 Cocoa 1 O0 Cornstarch 30-40 Lycopodiun, 25-40 Rice 40 Soybean 50

. Sugar 30 Wheat flour 50

Carbonaceous:

Asphalt 40 Carbon black (Spheron, grade 9) 180 Charcoal 20 Coal, Illinois 50 Coal, Kentucky 30 Coal, Pennsylvania 50 Coal, West Virginia 60 Gilsonite 25 Pitch 20

Chemicals: *

Acetone 1.15 .\crolein 0.13 Acrylonitrile 0.16 Benzene 0.2 Carbon disulfide 0.009 Cyclohexane 0.22 Cyclopentadiene 0.67 Cyclopentane 0.54 Diethyl ether "- 0.19 Dihydropyran 0.36 Diisobutylene 0.96 Diisopropyl ether 1.14 2,2-Dimethylbutane (neohexane) 0.25 Dimethyl ether 0.29 2,2-Dimethylpropane (neopentane) 1.57 Diphenyl 20 Ethyl acetate 1.42 Ethylamine 2.4 Ethyleneimine 0.48 Fumaric acid 35 Furan 0.22 Heptane 0.24 Hexane 0.24

(con'O

6 8 - 1 6 EXPLOSION VENTING GUIDE

Material Energy, millijoule

Chemicals: (con't)

N~lethanol 0.14 Methyl ethyl ketone 0.53 Nlethylbutane (isopentane) 0.25 Methylcyclohexane 0.27 i-Octane 1.35 i-Pentane 0.21 n-Pentane 0.22 2-Pentene 0.18 i-Propyl alcohol 0.65 i-Propylamine 2.0 i-Propyl chloride 1.55 n-Propyl chloride 1.08 Propylene oxide 0.13 i-Propyl mercaptan 0.53 Sorbic acid 15 Sulfur 15 Tetrahydrofuran 0.54 Tetrahydropyran 0.22 Thiophene 0.39 Triethylarnine 0.75 2,3-Trimethylbutane 1.0 Vinyl acetate 0.7

Gases:

Acetaldehyde 1.15 Acetylene 0.017-0.018 1,3-Butadiene* 0.13 Butane* 0.25 Cyclopropane* 0.17 Ethane* 0.24 Ethylene 0.07-0.08 Ethylene oxide'* 0.06 Hydrogen 0.017-0.018 Hydrogen - - Nitric oxide 8.7 Hydrogen sulfide* 0.068 Methane* 0.28 Methane - - Nitric oxide 8.7 Methylacetylene* 0.11 Propane* 0.25 Propylene* 0.28 Vinylacetylene* 0.082

Metals:

Aluminum 10-50 Boron 60 Iron 20 Magnesium 20-80 Manganese 80-320 Titanium 10-40 Uranium 45 Zinc 100 Zirconium 5

FUNDAMENTALS OF COMBUSTION 68-17

Mater ia l Energy, miilijoule

Plastics:

Cellulose acetate 10-15 Methylmethacrylate 15-20 Nylon 20 Phenolic resin 10-25 Polycarbonate 25 Polyethylene 10-30 Polypropylene 25-30 Polystyrene 15 Polyurethane foam 15 Rayon 240 Urea formaldehyde 80

Admixtures with Oxygen

Chemicals:

Diethyl ether 0.0012 Diethyl ether with 86 vol. % nitrous oxide 0.0012

Gases:

Acetylene 0.0002 Ethane 0.0019 Ethylene 0.0009 Hydrogen 0.0012 Methane 0.0027 Propane 0.0021

*Values from "Electrostatic Hazards, Their Evaluation & Control," Heinz Haase, Verlog Chemie, NY, 1977.

Other values from References 29 and 50.

2-2.1.6 Particle Size of Dusts. Experimental data show that a sufficient concentrat ion of particles, passing a No. 40 U.S. Standard Sieve (420-micron) must be present for a dust deflagration.

2-2.1.6.1 The particle size of dust has little effect on maxi- m u m pressure, but the rate of pressure rise increases significantly with a decrease in particle size. (See Figure 2-2. 1.6. L)

2-2.1.6.2 A decrease in particle size lowers the min imum energy required to ignite dust clouds. (See Figure 2-2. I. 6.2.)

l l a

300 [ I I I I I I

200

100

8O

~- 60

oJ

40 E 2

30,000

20,000

", Maximum Pressure

"-, Maximum Rate

1

10,000

,000

6,000

4,000

g

W

E E E

i m

2 0 - - - - 2 , 0 0 0

1~ . 1,000

6 ] I I - I I I I I I 6OO 10 20 30, 40 60 80 100 200

Average Panicle Diameter, Microns

r r "

¢ 0

el"

, .Figure 2-2.1.6.1. Effect of a v e r a g e particle d iameter of a t o m i z e d a l u m i n u m on m a x i m u m p r e s s u r e a n d rate of p r e s s u r e rise d e v e l o p e d b y def lagrat ions in a c losed 0 .043 cu ft vesse l . (44)

600

500

400

300

200

100

/

I [

68-18 EXPLOSION V E N T I N G . G U I D E FUNDAMENTALS OF COMBUSTION 68-19

I I I I 0 40 80 120 160 200

Average Particle Diameter, Microns

Figure 2-2.1 .6 .2 . Effect of particle s ize on the. m i n i m u m e n e r g y r e q u i r e d to igni te dust c louds of cornstarch. (Unpublished data,,courtesy of U.S. Bureau of Mines. )

2-2.1.7 Initial Temperature and Pressure of Mixture. A change in initial pressure (absolute) of the fuel-oxidant mixture produces a proportionate change in maximum pressure resulting from deflagration. A change in initial temperature (absolute) produces an inverse change in maximum pressure resulting from deflagration. (See Figure 2-2.1.7.)

~a

68=20 E X P L O S I O N V E N T I N G G U I D E

Increase of t empera tu re in most cases results in an increase in m a x i m u m rate of pressure rise.

600

500

400 - -

300 - -

200 - -

100 - -

I I I

Po =

Po

Po = 15 psia

¢-~

E -n E x cu

o I I I 0 .04 0 .08 0 .12 0 .16 0 . 2 0

1 Reciproca l o f In i t ia l Tempe ra tu re , o R x 10 -2

F i g u r e 2 - 2 . 1 . 7 . E f f e c t o f i n i t i a l t e m p e r a t u r e on the m a x i m u m pres sure d e v e l o p e d i n a ,c losed vessel for def lagrat ions o f 9 . 9 p e r c e n t methane -a l r mixtures at severa l initial pressures . (27)

F U N D A M E N T A L S O F C O M B U S T I O N 68-21

2-2.1.7.1 Theore t ica l ly and exper imenta l ly , it can be shown tha t at a given pressure an increase of t empera tu re in most cases results in a decrease in m a x i m u m explosion pressure. This is because a d~crease in moles results in a corresponding decrease in the pressure genera ted f rom the combustion.

2-2.1.7.2 However , as a pract ical mat ter , an increase in t empera tu re general ly causes an increase in the final p ressu re because, under confinement , the initial pressure will increase as the t empera tu re increases. Fur the rmore , a decrease in the genera ted p r e s s u r e ( e v e n in the ra ther hypothet ical case of t empera tu re increase at a given pressure) approx imate ly offsets the increase in flame speed.

2 - 2 . 1 . 8 T u r b u l e n c e . Ini t ia l turbulence slightly increases

120 30,000

100

8O

aa ~. 60 E

40

20

I I I I

Maximum Pressure ( T u r b u t e n t ) ~

_ _

Maximum Pressure (Nonturbulent)

__Maximum R a ~

Maximum Rate (Nonturbulent)

25,000

20,000 ~I

15,000 ~

lo, ooo

5,000

4 6 8 10 12 14 Methane, Percent

Figure 2-2.1.8. M a x i m u m pressure and rate of pressure rise for turbulent a n d n o n t u r b u l e n t m e t h a n e - a i r mixtures i n a 1 c u f t closed vesse l . (57~

m t

F U N D A M E N T A L S O F C O M B U S T I O N 68-23

the maximum pressure, while tl/e rate of pressure rise is markedly increased, (See Figure 2-2. 1.8.) Turbulent mixing of dusts with the oxidizing medium enhances diffusion of the oxidizer to the reacting surfaces and promotes the rate of oxidation of the dust. (See Appendix A .)

2-2 .1 .9 P r e s e n c e o f A d m i x e d M o i s t u r e or I n e r t D i l u t e n t s . Moisture or inert materials in sufficient quantity will quench a dust deflagration partly through absorption of heat and radiant energy and partly by.hindering diffusion of oxygen and gases into and from the burning f u e l Moisture in dust particles raises the ignition temperature of the dust because of the heat absorbed during heating and vaporization of the moisture. The moisture in the air surround- ing a dust particle has no significant effect on the course of a deflagration once ignition has occurred. There is, however, a direct relationship between moisture content and minimum energy required for ignition, minimum explosive concentration, maximum pressure, and maximum rate of pressure rise. For example, the ignition temperature of cornstarch may increase as much as 122°F (50°C) with an increase Of moisture Content from 1.6 percent to 12.5 pe rcen t . As a practical matter, however, moisture cannot be considered an effective ignition preventive since most ignition sources provide more than .enough heat to vaporize the moisture and to ignite the dust. In order for moisture to prevent ignition of a dust by common sources, the dust would have to be so damp that a cloud could not be formed. Factors affecting inerting requirements for deflagrations are t y p e ' o f fuel and its concentration, strength of the igniting source, ignition temperature of mixture, composition of the substance used for inerting, and the particle size with respect to dust. The effect of admixed moisture and inert dust on the lower flammability limit of several dusts are shown in Figure 2-2.1.9(a).

The effect of moisture on pressures and rates of pressure rise for cornstarch deflagrations are shown in Figure 2-2.1.9(b).

For most flammable gases, the addition of water vapor to the , gas-air mixtures affects combustion properties much as .does t he addition of an inert diluent like nitrogen. Some gases are exceptions. Bone-dry carbon monoxide burns relatively slowly in bone-dry air. Addition of a little water vapor results in appreciably faster burning.


0.6

f3-

t- O 0.4

t -

O O 0J

0.: E E

t J_

O _J

0

I 1 I

~-~ . "

--. ornstarch -~nd Calcium C a r b o n a t e

~ _ ~ _ _ ~ ~ C ° r n s t a r c h and Fullers Earth

Sulfur and Fullers Earth I I I

0.4 0.8 1.2 A d m i x e d Inert, oz per cuft

1.6

t ~

O

4 ~ c -

t -

O ,

E E

u _

5

0.6

0.4

0.2

I I I

0 0.16

- - P i t t s b u ~ t ' arch

~ ~ W h L a t Gluten

I I I 0.04 ' 0.08 . 0.12

Admixed Moisture, oz per cu ft

Figure 2-2.1.9(a). Effect of admixed inert powder and moisture on the lower flammable concentration of various dusts. (55)

F U N I ) A M E N T A L S O F C O M B U S T I O N 68-25 68-24 E X P L O S I O N V E N T I N G G U I D E

200 I I I 10,000

1 6 0

120

o~

Q .

E E • ~ 80

40

- - Rate .

I I I 4 8 12

Moisture Content, Percent

8,000

6,000

4,000

2,000

0 16

r r

~6

re"

• F igure 2-2 .1 .9(b) . Effect of mois ture on p r e s s u r e s a n d rates of p r e s s u r e rise d e v e l o p e d b y 0 .5 oz per cu ft of cornstarch de f l agra t i ons in a c losed 0 .043 cu ft vesse l . (34)

2-2.1.10 H y b r i d M i x t u r e s . Mixtures conta in ing f l ammable gas and dust in air or oxidant are referred to as " h y b r i d mixtures ." T h e presence of gas may h a v e a considerable influence on the burning characteris t ics of a combust ib le dust. T h e extent of this influence depends on the type and concentra t ion of f l ammable gas.

T h e r e are concentrat ions of dusts 'which, by themselves, will not burn, even with a large or intense ignit ion source, but in the presence of small quant i t ies of f lammable gases can burn even at concentrations below the f l ammable limit of ei ther material .

F igure 2-2.1.10 shows the effect of low concentrat ions (1 to 5 percent of methane) on the m a x i m u m pressure developed by coal dust. In general , m a x i m u m pressure was increased by the addi t ion of methane with d u s t concentrat ions up to about 0.5 oz per cu ft; at higher concentrat ions (up to 2.0 oz per c u f t ) m a x i m u m pressure was decreased by the addi t ion of methane. T h e presence of a small amoun t of f l ammable gas will also d r ama t i ca l l y reduce the m i n i m u m ignition' energy required to ignite the dust mixture. Since the f l ammable gases general ly require much lower ignit ion energies, as the rat io of f lammable gas to combust ib le dust increases, the ignit ion energy required decreases.

14o[ I

1201-- i ~ 1 Percent Methane

I I

0 .

100 . - I

P (3-

E E x 80

/ 5 Percent Methane

° I 40

0

( No Methane

I I I 0.5 1.0 1.5 Dust Concentration, oz per cu f t

2.0

Figure 2-2.1 .10. Effect of low p e r c e n t a g e s ' o f m e t h a n e in the a t m o s p h e r e on p r e s s u r e d e v e l o p e d b y coal dust def lagrat ions in a c losed 0.32 cu ft vesse l . (56)

68-26 E X P L O S I O N V E N T I N G G U I D E F U N D A M E N T A L S OF C O M B U S T I O N 68-27

2-2.1.11 V o l u m e a n d S h a p e o f Enclosure . Pressure developed dur ing a deflagration results from gaseous products of combustion and from heating and expansion of the atmosphere within a vessel or confinement.

Generally, m a x i m u m pressure is unaffected by the size and shape of the vessel; however, the rate of pressure rise is markedly affected.

Increasing the volume of a vessel or enclosure produces a decrease in the.rate of pressure rise dur ing a deflagration ; the rate of pressure rise is proport ional to the ratio of the surface area of the vessel to its

1,250!

1,000

751

/

500

250

I I

Key Vessel Volume c u f t

O 0.043 • 0.32 • 1.0 A 110.0 • 905.0

t , , r~

o.

._= r r "

¢t"

E

E X

I

J 0 5 10 15 20

Vessel Surface Area, 1 Vessel Volume ft

J

Figure 2-2.1.11. Effect of vessel dimension on the rate of pressure rise developed by 9.0 percent methane-air deflagrations. (57)

volume. This is shown in Figure 2-2.1.11. The following table indicates the size and shape vessels used to obtain data for this curve :

Table 2-2.1.11

Symbol, Geometry Volume (cuft) Dimensions

O cylindrical 0.043 23/~ in. dia., 12 in. high • cylindrical 0.032 " 73/~ in. dia., 12 in. high [] cubical 1.00 1 ftx 1 ft x 1 ft

rectilinear 110 4 ft x 5 ft x 6 ft "IT spherical 905 12 ft dia.

For spherical or nearly spherical vessels, the rate of pressure rise varies with the cube root of the vessel volume. (See "cubic law," Appendix A )

tim ',,4

68-28 EXPLOSION V E N T I N G G U I D E

Chapter 3 Fundamentals of Venting

3-1 Venting Deflagrations.

3-1.I. A vent in an enclosure (building, room, or vessel) is an opening through which newly formed or expanding gases may flow. The purpose of the vent is to limit the max imum pressure resulting from a deflagration in order to limit damage to the enclosure.

Extensive destruction may result if combustion occurs within an enclosure too weak to withstand the full force of the deflagration. An ord inary building wall (8-in. brick or an 8-in. concrete block) will not withstand a sustained internal pressure as small as 1 psig (144 lbs per sq ft or 6,9 kPa).

Unless the enclosure is designed to withstand the max imum pressure resulting from a possible deflagration, venting should be considered to minimize damage due to rupture. The area of the vent opening must be sufficient to limit pressure build-up to a safe value. (See Appendix A.)

Combustion venting of an enclosure normally implies the need to vent in such a manner that the max imum pressure development is low. The max imum pressure should be lower than the pressure which the weakest building or structural member can withstand. The weakest building member may be a wall, roof or floor if the enclosure is elevated. On equipment the weakest section may 'be a .joint,

3-1.2 No data are available on actual forces or loads applied to the walls of full-size structures by different types of deflagrations. Designs must be based upon the specific equipment, the enclosure and material of construction, shock resisting ability, and the effect of vent openings on the pressure developed and duration. Pressures produced by deflagration of dust and gas-air mixtures in laboratory test chambers at atmospheric pressure are on the order of 100 psig (690 kPa). An enclosure need not be constructed to withstand such pressures if the volume is adequately vented. In most instances it is impractical to construct a buildin~ or some large equipment to withstand 100 p.sig (690 kPa).

3-2 Pressure and Rate of Pressure Rise.

3-2.1 The rate of pressure rise is an important factor in venting; it determines the time interval available for the combustion products to escape. A rapid rate of rise means that only a short time is available for venting and, conversely, a slower rate permits a longer time.

F U N D A M E N T A L S O F V E N T I N G 68-29

More vent area is required for effective venting of deflagrations having a high rate of pressure rise. Fundamenta l data on dust deflagrations in a closed vessel are shown in Appendix E, Table E-1. Pressure and corresponding flame development for deflagration in an open vented vessel is shown in Figures 3-2.1 (a) and 3-2.1 (b).

o6 1 I I . . . . . . I Vent Ratio: 10,65 sq ft/100 cuf t

0.5

0.4 ._m

~ 0 . 3

o r~

0.2

0.1

1 0 20 40 60 80 100 120 140

Time, Milliseconds

160

Figure 3-2.1(a). Pressure-t ime record of 9.4 percent methane-air deflagra- tnon i n a 0=32 c u f t vented cylindrical vessel. (Unpubl ished data, courtesy of [f.S. Bureau of Mines. )

I,,U

6 8 - 3 0 EXPLOSION VENTING GUIDE

r 2.5 In. D i a m e t e r Vent R a t i o : 1 0 . 6 5 sq f t / l O 0 cu ft

Figure 3-2.1(b). Flame deve lopment for a 9.4 percent methane-alr deflagration in a 0.32 cu ft vented cylindrical vessel. Maximum pressure of 0.4 psi occurs at 80 milliseconds. (Unpublished data. courtesy of U.S. Bureau of Mines.)

FUNDAMENTALS OF VENTING 68-31

3-2.2 T h e m a x i m u m pressure developed in a l abo ra to ry test c ha mbe r does not ful ly define the true force imposed on walls and o ther surfaces. This is because a given force app l ied to a panel

• for a long per iod m a y be more destruct ive than the same or even grea ter force appl ied for a short period. Moreover ; the pressure is not constant . The area under the pressure- t ime curve de te rmines the total impulse exer ted and hence the dynamic effect of the explosion on the structure. [See Figure 3-2.1(a).] Briefly, the effects of a def lagrat ion depend upon the m a x i m u m pressure developed, the rate of-pressure rise and the dura t ion of pressure.

Da ta . fo r me thane-a i r def lagrat ions in various size vented vessels (0.32 to 216 c u f t ) show that a l inear re la t ionship exists between m a x i m u m pressure and average rate of pressure rise.

120 1,200,

100 1,000 l

/

80 = 8oo l- x o"

E .a 60 ~ 60(]

40 ~ 400__

20 200 __

0 0

I I 1 12'°°° DUSt Concentration = 0.500 oz per cuft

~I I0,000

~ \ ~ . - - 8,000

/ ~k~k Maximum Rate

~ \ ~ Total Impulse

I I I 2 3 4 5

Rat io of Relief Area'to Volume, sq ft per 100 cu f t

6,000

4,000

2,000

o= o~

E ==

Figure 3-2.2. Variation of pressures, rates and impulses with vent ratios in magnesium deflagrations in a vented 64 cu ft vessel. (36) klL

68-32 E X P L O S I O N V E N T I N G G U I D E F U N D A M E N T A L S O F V E N T I N G 68-33 O

3-3 V e n t i n g Formulas .

3-3.1 The re are a number of empir ica l formulas for comput ing size of vent openings for a specific vessel or structure, but, unfor tu- nately, none of these can be considered ent i rely satisfactory.

Theore t ica l calculations have been made by several research organizat ions to de te rmine and evalua te the abi l i ty of various structures to wi ths tand the shock of an external pressure; both full-scale and model tests were conducted on structures and com- ponent s t ructural elements by the coordina ted act ivi ty of many governmenta l agencies and pr ivate industr ial and engineer ing organizat ions. These studies deal t with the abi l i ty of various structures and shelters to resist .external bomb blasts. Most enclosed structures, by reason of tie-ins and bracing, can usually wi ths tand more pressure f rom wi thout than f rom within. At times, this may be in the ra t io of 2 to 3 t imes greater .

3-3.2 The present knowledge of the mechanism of a large- scale def lagrat ion and the resistance of enclosures to in ternal forces does not pe rmi t precise recommenda t ions for the comput ing vent relief. T h e calculat ion procedures in this guide are given in Appen- dix A. I t is known from experience and testing that genera l ly it m a y not be prac t ica l to provide sufficient vent a rea to prevent serious damage f rom an o p t i m u m - m i x t u r e def lagrat ion in a large- volume enclosure. Experience has shown that most dust and gas combust ions do not involve a large pa r t of the total vo lume of the enclosure and tha t ignitions of vapors f requent ly occur near the limits of the f l ammable range. Consequently, such def lagrat ions are ra ther weak and vent ing can be effective.

3-4 V e n t i n g Variables . Vent ing is a measure tha t limits deve lopment of destruct ive pressures in an enclosure incapable of wi ths tanding the full force of a deflagrat ion. De te rmina t ion of vent a rea for an enclosure is mostly empir ical . General ly , the vent a rea should be made as large as prac t ica l and the remain ing enclosure constructed as s t rongly as economical ly feasible.

3-4.1 V e n t Size a n d Shape . T h e m a x i m u m pressure (as well as total impulse) in vented vessels decreases as the vent a rea increases. [See Figure 3-d. 7(a).] One large vent will relieve the pressure of a deflagrat ion in a small enclosure as effectively as several small vents whose area equals the area of the large vent. How- ever, this may not be true for large enclosures and the dis t r ibut ion

60

E x 40

20

I I I I [

• -- 0.32 cu ft Cylindrical Vessel

O - 1.0 cu ft Cubical Vessel

A - 64 cu f t Cubical Vesset

x - 216 cu f t Cubical Vessel

8 ° I ° k ~ •

1 2 3 4 5

Vent Ratio, sq f t per lOOcu f t

0

Figure 3--4.1(a). Effect of unrestricted vents on pressure developed by the deflagration of nominal 9.4 percent methane-air mixtures. (Unpublished data, courtesy of U.S. Bureau of Mines.)

of vents relat ive to the locat ion of the ini t iat ion of the def lagrat ion is impor tan t . Rec t angu la r vents are almost as effective as square vents of equivalent area for pressure relief. [See Figure 3-4. 7(b).] Vents with rounded edges to facil i tate gas flow and minimize friction are most efficient.

FUNDAMENTALS OF VENTING 68--35 6 8 - 3 4 EXPLOSION VENTING GUIDE

2 ,400 I I I

2 ,000

,.~ 1 ,600 - -

~r e~

) 1 , 2 0 0 - -

E E X

:~ 8 0 0 - -

4 0 0 [ --

/

\\ and Square Vents

Rectangular Vents > xx,

I I I t 4 8 12 16 2 0

Vent Ratio El f t per 100 cu f t

Figure 3-4.1(b). Effect of vent shape on pressures produced by deflagrations of cornstarch in a 1 cu ft vessel. (Unpublished data, courtesy of U.S. Bureau of Mines. )

3-4.2 T y p e of v e n t s . Open or unrestricted vents are most effective in relieving pressure build-up. Vents covered with a diaphragm, swinging door, bursting disc, ,some type of weak construction material or other device require inertia to be overcome; therefore, such vents are less effective ,than unrestricted (open) vents. Chapter 4 contains some illustrations Of various methods of providing vents to reduce the forces of a deflagration.

3-5 Basic R e c o m m e n d a t i o n s for V e n t i n g . Sifice venting is a complex subject on which essential information is lacking, this is only a general guide for best current practices. " Impor tan t recom- mendat ions for reducing damage by venting are summarized below.

3-5.1 Vents are generally required in buildings or enclosures containing operations or processes where dust, gas, or mist may be present in sufficient amounts to create f lammable concentrations in air or other oxidizing media.

3-5.2 The required areas of vents depends upon such characteristics as the rate of pressure rise, max imum pressure developed, the strength of the enclosure, and design of vent closure. Empirical methods and homographs may be used to determine vent area. (See Appendix A.)

3-5.3 Vents should be located as close as possible, to potential sources of ignition which may originate the deflagration. How- ever, experiments show that for spherical or cubical containers with central ignition; the shape and location of a vent is not significant. Where points of ignition cannot be determined in large enclosures, vents should be evenly distributed. Vents with no con- structional members to impede gas flow and vents with rounded edges to promote gas flow are most efficient.

,,-,,.~t vv.cncvcr posstule, vented products should be directed to a safe location outside of an enclosure to avoid injury to personnel and minimize damage to property, in congested locations, substantial ducts or diverters should be provided to direct explosive force and combustion products to a safe area.

As a precautionary measure, it may be necessary to install indoor railings along floor edges near vent panels to prevent personnel f rom falling against the panels. Warning signs should also be provided to alert personnel that panels are easily knocked loose

Vents should not be obstructed. Wherever possible, vents should be designed to minimize the accumulat ion of ice and snow and should be cleared to permit proper operation.

Where sashes are used for venting, precautions should be taken dur ing cold weather. Ice crystals may form between the venting sash and the frame due to high humidi ty in the area and produce a cementing action on the vent allowing greater pressures to build up before the vent will open; a coating of grease on the adjacent surfaces may prevent the bridging of ice crystals between the members of the vent. Corrosion and paint may also increase friction in opening a vent.

Vents, particularly those with discs, diaphragms, or other closure devices, should be located where flame, gases, or flying material cannot injure people. In addition, a vent closure such as a swing-

I 0


ing door should be designed to prevent development of a vacuum after heated gases from a deflagration have cooled. (See Chapter 4 and Appendix D for discussi'on and description of vent devices and vent closures.)

3-5.5 Structural damage can be minimized by locating hazardous operations or equipment outside of buildings and segregated from other operations. This is particularly true of dust collectors, arrestors, bucket elevators, and reactors. Multiple physical inter- connections between the ductwork system of each collector should be avoided. Furthermore, such equipment should be properly vented and a device should be provided at the inlet of the collector which will prevent a deflagration from blowing back through the ductwork and into the building or structure.

.3-5.6 Highly hazardous operating equipment should be separated into individual units by pressure resisting walls, and each unit so formed should be vented outdoors. Exterior walls may be made of heavy construction if equipped with suitable vents or adequate lightweight panels which blow out easily. Locating hazardous operations or equipment in basements or areas partially below grade should be avoided due to the difficulty of providing adequate venting.

3-5.7 When it is impractical to locate hazardous operations or equipment outdoors, they should be located adjoining outdoor walls, in a single-story building or on the top floor of a muhistory building, or in a lightly constructed penthouse and vented directly to the outside through ducts of adequate cross-sectional area.

3-5.8 Vent ducts used to conduct combustion products outdoors should be constructed to withstand the maximum pressure of a deflagration. Duct length should be minimal and bends should be avoided. I t should be realized that any duct will decrease the effectiveness of the vent in proportion to the duct length. Increas- ing duct diameter with duct length compensates, but design data are not presently available. Figure 3-5.9 shows the effect of length of ducts attached to a 1-cu ft vessel on pressures developed by deflagrations of coal dust. Under certain conditions, detonation may occur in piping or duct systems and effective venting cannot be accomplished. Additional information on vent ducts is given in Appendix A.

3-5.9 External wind pressure or suction may operate venting devices and these effects should be considered in their design. Wind pressures in severe storms may reach over 30 lbs per sq ft (1.44 kPa) and vents designed to open at a higher pressure in the event of a deflagration may not provide for building safety. There- fore, the vent design should take into consideration the local wind conditions and building safety.

F U N D A M E N T A L S OF V E N T I N G 68-37

I I I I 31500

3,000

2,500 3V~ in. Circular Du~

~2,000

g

--~ 1,500 x

in. Circular Duct

1,000

500 in. Circular Duct

I I I I 0 4 8 12 16 20

Length of Duct, ft

Figure 3-5.9. Effect of vent duct diameter on pressures developed by deflagrations of coal-dust in the 1 cu ft chamber. (Unpublished data. courtesy of [I.S. Bureau of Mines.)


Chapter 4 Descr ipt ion of V e n t s and V e n t Closures

'4-1 G e n e r a l . The vents described in this section have been designed or developed for the release of pressure from enclosures in which explosions of dusts or gases may occur. In most cases, the described vents are effective only in deflagrat ions in which the rate of pressure rise is modera te and where, in large enclosures, only a par t is involved in the deflagrat iom T h e devices descr ibed are not genera l ly sui table for protect ion of pressure vessels; which is outside the scope of this guide, nor for protect ion against pressure or shock waves produced in detonat ions of explosives.

Some types of yent closures are commerc ia l ly avai lable and may be purchased ready to install in buildings or e q u i p m e n t . . T h e following descriptions should be used as the basis for development of sui table vents and vent closures which will provide the desired protection. Examples of vents and vent closures are shown in Ap- pendix D.

4-2 O p e n or Unobs truc ted Vents . The most effective vent for the release of def lagrat ion pressure f rom enclosures is an unobstructed opening. However , there are compara t ive ly few operat ions with inherent def lagrat ion hazards that can be conducted in open equipment instal led in buildings wi thout walls.

Often some form of vent closure must be provided to protect against the weather , to conserve heat, to bar unauthor ized: entry, to preclude disseminat ion of tlle combust ib le mater ia l , or to prevent contamina t ion of the product by the ent rance of di r t or moisture from the outside.

O p e n equ ipment is r ecommended wherever a more serious deflagrat ion hazard is not created through dispersion or dissemination of the mater ia l and where closed, equ ipment is not necessary to prevent contamina t ion of the mater ia l .

4-2.1 Louver s . Al though openings containing louvers cannot be considered complete ly unobstructed vents; t h e y - d o provide- a large percentage of free space for the release of def lagrat ion pressure and have served effect i ,¢e lyas vents. They are r ecommended especially as wall vents where windows are not required to main- tain control led a tmosphere condit ions within the enclosures. Louvers can be used effectively as vents where it is necessary o r desirable .to prevent unautIf0rized entry or egress., However, compensat ion for pressure d rop must be considered.

4-2.2 H a n g a r - t y p e Doors . Large hangar - type or steel cur- tain doors installed in side walls of rooms or bui ldings c a n be opened to provide unobstructed ,)ents dur ing the opera t ion of any process or equ ipment in which there is an inherent defl~tgration hazard. Such doors can be closed to prevent unauthor ized entry

VENTS AND VENT CLOSURES 68-39

when the equ ipment is unat ter ided or not in operat ion. This type of venting has been effective and is highly recommended , but strict supervisory control is essential in cold. cl imates to insure that employees do not sacrifice safety for comfor t by keeping the doors closed dur ing operations.

4-2.3 O p e n Roof Vents . Large roof openings protec ted by weather hoods can serve as def lagrat ion vents on one-story buildings or the , top story of a mult iple-s tory building. "This type of venting is effective par t icu lar ly where l ighter - than-a i r gases may escape f rom processing equ ipment and create a hazard near the ceiling of the enclosure. In addi t ion to serving as vents for the release of pressures, such roof openings reduce the possibility of a def lagrat ion by providing a channel through which the gas can escape f rom the building.

4-3 Closed or Sealed Vents . Where large openings c a n n o t be permi t ted in a building, the most desirable a r r angemen t is an isolated single-story building. Such a bui ld ing can be most easily designed for explosion resistance and venting. Equ ipmen t which requires vent ing should be located close to outside walls so that ductwork, if necessary, can be short.

Building vent closures are necessary in a i r -condi t ioned plants or where heat is provided for the comfor t of occupants dur ing all or pa r t of the year. Vent closures are required on processing equip- , . , . , t ~v,,mcvu, , , , ~ , ~ . b ~ a t y tu l c t am uu~t oc gas ur where processes are conducted under pressure, vacuum or other control led a tmospher ic ~^ "~":~-" w n u l L I U L I ~ ,

The fundamen ta l pr inciple in the design of vent closures is that the vent will open at as low a pressure as possible. I t should have no counterweights ; counterweights add to inertia. T h e effect of various vent closures is i l lustrated in Tab le 4-3.

Table 4-3 Maximum Pressures Produced by Deflagrations in Enclosures

with Unrestricted Openings or Different Types of Vent Closures (36)

Type of Vent or Vent Opening Light Heavy

.Type of Vent Unrestricted Heavy Paper Swinging Swinging Dust Ratio Opening Oiaphragm Door Door

sq ft/ 100 cu ft Maximum Pressure, lbs/sq ft

Coal 1.56 81 292 101 - - Coal 3.52 . 29 158 36 55 Aluminum

(Atomized, fine) 3.52 71 205 • 161 232

68 -40 E X P L O S I O N V E N T I N G G U I D E

Construction of the closure should be light so that full opening can be quickly obtained; yet the structural strength must be sufficient to withstand natural" forces such as wind or snow loads.

When vents are sealed by paper, plastic, metal diaphragms, hinged panels, or other closures, the maximum pressure developed is higher than when the vent is unrestricted. Therefore, the vent areas must necessarily be larger. [See Figure d-3(a).]

Rupture of paper, plastic, or metal diaphragms during the initial stages of a deflagration is greatly facilitated by saw-toothe(t or piercing cutters along the periphery or at the center of the diaphragms. Cutters permit the use of smaller relief vents. [See Figure 4-3(b).]

I t is not possible to describe all of the devices that have been developed to serve as vent closures, but certain representative types can be grouped under separate headings.

4-3.1 B u i l d i n g or Room V e n t Closures. These type closures may be manual ly or mechanically operated, such as doors; windows, and skylights, or may have weak structural features, such as large glass areas or light wall and roof panels built or sealed in place but designed to open due to overpressure.

4-3.1.1 Doors. To serve effectively as building vent closures, doors must be installed to swing outward and have latches or locking hardware that will function automatically to permit the door to open under slight internal pressure. Friction, spring, or magnetic latches of the type designed for doors on driers and ovens are recommended. Max imum weight per unit area should be limited to 2 lbs/ft 2 (10 kg/m2).

4-3.1.2 Windows. Normally, windows installed to provide light or ventilation can frequently be arranged or adapted to serve as vent closures when they are properly .hinged to open outward. A number of different styles designed especially for this purpose are commercially available.

4-3.1.3 M o v a b l e Sash. T op or bot tom hinged movable sash or projected type, which are commercially available, have been widely used for venting. I t is usually necessary to have such sashes equipped with some form of latch or friction device to prevent undesired opening due to wind action or to prevent intrusion, but care should be taken to avoid the use of any latch or lock which is not well maintained and not always ready to operate when a deflagration occurs.

When swinging panels, windows, or other hinged devices are used, care must be taken to prevent closure of the vent opening

V E N T S A N D V E N T C L O S U R E S 68-41

after the initial positive pressure wave of the deflagration subsides. This will prevent the development of destructively high negative pressures as the remaining combustion products cool.

700

400 .Q

300 E

100

6 0 0 - -

500 - -

200 - -

I I I I

per Diaphragms

I I I ~ - 1 2 3 4 5

Ratio of Relief Area to Volume, sq ft per 100 cu ft

Figure 4-3 (a). Relative effectiveness of unrestricted openings and of heavy- rnaper diaphragms in relieving pressures from coal-dust deflagrations

itiated by electric spark. (36)


600 ] I I I

500 - - ,~ Vents with Heavy-paper Diaphragms without Cutters

400-- ~r

ff 300 - - /Diaphragms with One Cutter - -

~D / a .

E _2

2 0 0 -

100[-- Diaphragms withFour Cutters ~ I

I I I I 0 1 2 3 4

Ratio of Relief Area to Volume, sq ft per 100 cuft

Figure 4-3(b). Effectiveness of cutters to facilitate rupture of heavy-paper vent diaphragms on pressure relief of coal-dust deflagrations. (36)

4-3.1.4 Roof or Wall Panels . Such panels are more economical than fixed sash or moveable windows and can provide very effective protection against damage. In this type of venting, a portion of the roof or an exterior wall between strong partition walls is constructed to blow out readily if an explosion occurs. The panels may be of very light construction such as sheet metal, corrugated plastic paneling, roofing paper, or roofing paper supported

V E N T S AND V E N T C L O S U R E S 68-43

by coarse mesh wire. The total weight of the explosion panel assemblies should be less than 1.5 lb/f t 2 (7.3 kg/m2). In some instances, the entire roof over a room has been constructed as a panel or cover to lift or blow off if an explosion occurs. However, the roof must be securely anchored to prevent the wind from blowing it off. (The authority having jurisdiction should be consulted regarding required type of anchorage for the panels.) Metal roofs can be designed and installed with crimped edges, like can lids, that will normally hold them in place.

4-3.1.5 Skyl ights or Monitors. Such closures with moveable sash that will open outward or fixed sash containing panes of plastic that will blow out readily under pressure f rom within can be used to supplement wall vents or windows in buildings.

This makes it possible to use larger floor areas than would be permissible when only side wall vents are used. Resistance to displacement or opening of skylights or monitor windows by pressure should be as low as consistent with the requirements for structural ' strength.

4-3.2 E q u i p m e n t Vent Closures. Very few types of equipment can be operated without closures and numerous methods of providing satisfactory closures have been developed by plant operators, engineers, and equipment builders. Some of the closures described may have a very limited application, but the general principle involved can frequently be used in the design of similar devices for other specific purposes. Equipment should be vented directly to the outside of the building through short ducts of adequate cross-sectional area.

4-3.2.1 Charging Doors or Inspect ion Ports. Such doors or ports may be designed to function as automatic vent closures when their action does not endanger personnel. They can be used on totally enclosed mixers, blenders, driers, and similar equipment. Hinged doors or. covers held shut with spring latches are most frequently used for this form of protection. I t is difficult to vent equipment of this type especially if the shell, drum, or enclosure revolves, turns, or vibrates.

4-3.2.2 Ven t in g Devices . Venting devices generally used on tanks or equipment that are normally closed at all times are not intended to be used as doors or inspection ports and are expected to open only when the internal pressure exceeds a predetermined limit. The cover,of the vent opening is usually fitted with a gasket and held in place by spring action. In applications of this principle on larger equipment, on low pressure air ducts, settling chambers, etc., the travel of the cover can be restricted by chains installed at


corners or points on the periphery. Springs may be used to ~ anchor chains to absorb the deceleration forces. Magnet ic latches, where the magnetic lines of force must be ruptured, can be adjusted to release at a predetermined pressure.

4-3.2.3 Diaphragms . M a n y different kinds of d iaphragm materials have been used or are available for use as closures on explosion vents. Only a few of the many types can be described, but mention of these few will, no doubt, suggest others which may prove to be equally or more effective. The criterion in choosing a vent material is that its bursting strength will be considerably less than the walls of the enclosure.

4-3.2.3.1 Paper . Waterproofed kraft paper, building paper, and roofing paper have all been used as vent closures. The usual method of application consists of pasting or clamping a sheet of paper onto a wooden or metal f rame designed to fit over the vent opening or to slide into a vent duct. The breaking strength of the paper and the resistance to be offered to normal pressure within the enclosure will determine the best type of paper to select. Papers that weaken when exposed to water or moisture should be used only in dry places. Flame retardant paper and several different types of pl~tstie impregnated paper have been developed and used effectively under certain conditions for vent closures.

4-3.2.3.2 Cloth. Cloth impregnated with paraffin or plastic, varnished cambric, and plastic-covered cloth netting are a few of the different forms of cloth diaphragms used as explosion vent closures. These materials are generally available in rolls or sheets and the vent closures are made by cutting pieces to the proper size and gluing, tacking, clamping, or otherwise fastening them over the openings through which deflagration pressure is to be released. Such vent closures have certain properties that make them stlitable for use in places where paper diaphragms would not be satisfactory.

4-3.2.3.3 Plastic. Plastic vent closures are of two types: flexible and frangible. Both types are usually available in sheets. Pieces cut to the desired size may be installed in place of glass in window or observation port frames. The material may also be used instead of paper or cloth to seal vents.

Flexible plastic sheets are usually installed in slotted frames in such a way that pressure f rom within bulges the sheet and releases it from the holding frame. Transparent or translucent plastic sheets can be used instead of glass in certain types of window frames that will permit the sheet to function as a pull-out d iaphragm vent closure.

VE}qTS AND VENT CLOSURES 68-45

Frangible sheets are usually selected to serve as vent closures on the basis of their brittleness. I t is possible to obtain thin sheets of plastic that will crack or rupture under less pressure than single- strength glass. For this reason, it can be advantageous to use transparent or translucent plastic sheets instead of glass in window sash.

4-3.2.3.4 Metal . Metal foil is sometimes used to seal vents. This type of material can be substituted for flexible d iaphragms where the material being processed would react with the diaphragm.

Under other conditions where higher pressure is maintained within the enclosure or the vent area is very large, heavier, rigid sheets of metal may be used as vent covers. They should be designed to prevent generation of shrapnel.

4-3.2.4 D i a p h r a g m Cutters. Cutters designed to expedite the openings of vents closed or sealed with diaphragms have been used in many instances. Even a delay of a few thousandths of a second in relieving deflagrations of dusts or gases having high rates of pressure rise may cause extensive damage to equipment. To reduce this delay between the initial indication of pressure within an enclosure and the opening of the vent, some plant operators have installed saw teeth, spear points and other cutting devices designed to initiate the tearing or ruptur ing of the d iaphragm as soon as the pressure causes the least distortion from the normal position of the vent closure,


Appendix A Recommendations for Deflagration Venting of Gas and Dust Mixtures

This Appendix is not a part of this N F P A d o cumen t but is included for information purpose s only.

A-1 General Comments. At the time of the draft ing of this guide, the state of technology of deflagration venting is still only partially developed. Venting deflagrations involves many variables. Only some of the variables have been investigated to any extent. The investigations which have been conducted do allow for certain generalizations to be made. The recommended calculation •bases given below have been developed from these generalizations. The calculation bases must be recognized as approximate only. An at tempt was made to evaluate the data and correlations in recently published papers and to make use of those which were considered reliable.

The max imum pressure which can be developed during venting cah ' be significantly higher than the pressure at which the venting device releases. A number of variables affect this max imum pressure. Several of these are discussed in Chapter 2.

A-2 Venting of Deflagrative Combustion Inside Buildings. A~ discussed in c~t. . . . . . 1. . . . . . . . . • . . . . . . . . ..o ~,,a~,~, 3, most uu ,dmgs cmlnot wxmstana nlgn pressures f rom within. Damage to buildings by combustion within them can be minimized by adequate venting. As discussed in Chapter 2, the venting must be such that the max imum pressure developed will be lower, by a safety margin, than the pressure which the weakest building member desired not to break or vent can withstand. The weakest building member may be a wall, the roof, or, if the building is elevated, the floor.

A-2.1 Vent Panel Construction Material. Special consideration should be given to the material of construction of the vent panel. I t should not be of a material that tends to break into pointed shards; specifically, it should not be of glass or the cement-asbestos type board. I t is desirable that the vent panel weigh no more than 1.5 lb / fC (7.3 k g / m 2) of effective vent opening. Material which meets most of the .desired criteria fairly well is thin gage corrugated paneling made of fiberglass-reinforced plastic with a flame spread rating no greater than 25 as determined by Method of Test of Surface Burning Characteristics of Building Materials, NFPA 255-1972.

A vent panel made of light material and designed to release at a low pressure in the range of 10-30 lbf/ft 2 (0.5-1.5 k N / m 2) cannot be counted upon to hold people in case they should fall against

APPENDIX A 68--47

such panels. As pointed out in 3-5.4, it may be necessary to install indoor railings along floor edges near the vent panels. Also, as pointed out in that subsection, vents should be unobstructed from both the inside and outside of the building protected.

A-2.2 Calculating Vent Area. A number of factors com- plicate combustion inside buildings as discussed in References 41, 52, 66, and 70. References 41 and 70 discuss the "Runes" equa t ion for calculating vent area. The equation is as follows:

CL1L2 Av - (Eq 1)

Av = necessary building vent area, ft 2 or m 2 C = constant, characteristic of the fuel gas, as discussed below.

NOTE: C is dependent upon the units (English or SI) used.

La = smallest dimension of the rectangular building enclosure to be vented, feet or meters.

L2 = second smallest d imens ion-of the rectangular building enclosure to be vented, feet or meters.

P = maximum internal building pressure which can be with- stood by tlle weakest building member which is desired not to vent or break, lbf/in. 2, or k N / m 2.

For most gases such as natural gas, propane, gasoline, benzene, acetone, and many others, values of fundamenta l burning velocity are nearly the same. For such gases the value of C in Equat ion 1 is approximately 2.6 in English units and 6.8 in SI units. Note in this connection that the actual flame speed will be many times the fundamenta l burning velocity. The value of C, in fact, allows for the actual flame speed under conditions of a certain amoun t of flame turbulence increase due to (1) physical obstructions such as process equipment, piping, and building structural members; (2) flow-induced turbulence d u e to venting; and (3) turbulence induced directly by a large flame. For ethylene the suggested value for C is 4 in English units and 10.5 in SI units; for hydrogen, 6.4 in English units and 17 in SI units. The last two sets of values suggested for C have not been tested in actual building venting inci- dents. The first' set of values was calculated f rom one building venting incident (see Reference dI.) Turbulence effects, and hence flame speeds, can differ as the turbulence producing situations vary.

For venting of combustion of most organic dusts, the recommended value for C is 2.6 in English units and 6.8 in SI units. For metal dusts having high flame speeds, the recommended value of C is 4 in English units and 10.5 in SI units.

A P P E N D I X A 68-49 68-48 E X P L O S I O N V E N T I N G G U I D E

For mists of organic liquids, the recommended value of C is 2.6 in English units and 6.8 in SI units.

The following table summarizes the recommended values of C :

C for Equation C for Equation Fuel Identity in English Units in SI Units

G a s c s w i t h F l a m e S p e e d 2 .6 6 .8 like Propane

Ethylene 4 10.5

Hydrogen 6.4 17

Organic Dust 2.6 6.8

Organic Mists 2.6 5.8

High Flame Speed 4 10.5 Metal Dusts

A-2.3 Applicable Building Dimensions. It is believed that the venting equation is suitable for combustion venting of spaces in buildings having a nominal length/width (i.e., L/D) ratio up to 3. For spaces having an L / D greater than 3, the space should be subdivided into multiple units, each having an L /D of no more than 3. In a rectangular building when Lx and L2 are not equal, the effective value of D is V'-L1L2.

Wherever it is possible, combustion vents for a building should be distributed over the walls of that building rather than confined to one wall.

A-2.4 Special Points Relative to Spills Inside Buildings. Two important points relative to spill situations should be noted about the venting of combustion inside buildings. (1) Building ventilation rates even as high as one air change per minute will not necessarily prevent formation of flammable mixtures from spills inside buildings. (2) Concentrations of gas from spills

"inside buildings can vary greatly throughout the building space. Gas at a concentration above the upper flammable limit can burn rapidly because the thermal drafts caused by the initial flame will promote further mixing of air into the unburned fuel. Further- more, the initial flame raises the temperature of the unburned mixture, thereby increasing the flammability limits•

Appendix B contains a sample calculation for combustion venting for buildings.

Location of flammables-handling equipment in open structures with or without roofs obviates many of the problems of gas accumu- lation in buildings and of necessary building venting. (See the Flammable and Combustible Liquids Code, NFPA 30-1976.)

A-3 Venting of Dust Combustion Inside Vessels. The most comprehensive known design bases for venting of dust deflagrations are given in VDI Richtlinie 3673 (Reference 87). This work was published in Germany and is based on data obtained from a very extensive test program. This program involved venting tests with four dusts in containers of four sizes: 1, 10, 30, and 60 cubic meters. The homographs prepared from the data for venting of dust deflagrations are reproduced here as Nomographs A-F. The necessary venting area as a function of. the class of dust, the vessel volume and strength, and the relieving pressure of the combustion vent can be determined from these nomographs.

For the purposes of the homographs, combustible dusts were divided into three classes according to their maximum rates of pressure rise. At the present time most of the available data on rates of pressure rise for various dusts have been obtained in the Hartmann test apparatus of the Bureau of Mines, U.S. Depart- ment of Interior. Those dusts giving maximum rates of pressure rise in the Hartmann apparatus, up to 7,300 psi/see, or 50,000 kN/(m 2) (see), were designated Dust Class St-1. Those giving rates of 7,300-22,000 psi/see, or 50,000-150,000 kN/ (m ~) (sec), were designated Dust Class St-2. Those above 22,000 psi/see, or 150,000 kN/(m 2) (see), were designated Dust Class St-3. (See Table A-3(a).)

A more reliable basis for dust classification may be obtained by conducting dust flammability tests, similar to those in the Hartmann apparatus, in a larger vessel which approaches the geometry of a sphere. In recent years, for example, tests for determining rates of pressure rise for a number of dusts have been conducted in nearly spherical vessels of volume approximately 1 m a. Such data have been found to be more dependable than the Hartmann data in terms of projecting to venting requirements for equipment of commercial size.

Basic to the nomographs is the concept of what is called the "cubic law." This takes the form of the following equation for combustion inside completely closed vessels, i.e., with no venting of the deflagration.

(dp/dt)max • ( v l / 3 ) = constant where

(dp/dt)max = maximum rate of pressure rise for dust or gas combustion in a particular vessel

V = volume of the particular vessel

The constant is referred to as Kst for dusts and Ko for gases. For rate of pressure rise in bar/see and vessel volume in m a the constant is expressed in bar • m • sec -1.

• 6 8 - 5 0 E X P L O S I O N V E N T I N G G U I D E

Applicat ion of the cubic law to the venting of deflagrations is made possible by the findings of Donat (Reference 23 and 24) and Bartknecht (Reference 5), whose data indicated tJressure develop-. ment with venting did, in fact, approximately follow the cubic law. This holds, so long as the combustion is deflagration; it does not hold for detonations:

O n the basis of the cubic law the equat ion for the necessa@"vent - ing area is as follows:

FIV2VI!/8 F1V~ 2/8 F2 - or, F~ -

V1V21/8 Vz2/.~

where F, = vent area on test vessel found necessary to prevent

pressure dur ing combustion from exceeding a given value F~ = vent area which will be necessary on a second vessel to

prevent pressure dur ing combustion from exceeding the same value

VI = volume of test vessel V2 = volume of second vessel

While vent ing of dust deflagrations in vessels of various volumes may not exactly follow the cubic law, it does appear to follow this more closely than other .relations which have been used previously.

law, needs to be less than 5. When the constant is determined for dust combustion in vessels of I m 3 or more, and when the constant is expressed in bar • m • see - i , the hazard classes are as follows:

Hazard Class

Table A-3 (a) Hazard Cla'ssilicatlon of Dust Deflagrations

Kst (bar .m. sec -1) Kst (bar 'm'sec -1) Maximum for weak ignition for strong ignition rate of pressure •

source (energy source (energy ap- rise in Hartmann Approx. 10 W-sec)* prox. 10,000 W.sec ) apparatus psi/sec

St-1 ~ 100 < 200 ~ 7,300 St-2 101-200 201-300 7,300-22,000 St-3 > 200 > 300 > 22,000

*W.sec = Watt-seconds.

The following table lists some.characterist ic values of K s t for typical dusts. These are given in various papers published in the. work by Dona t and Bartknecht. It is noteworthy that the value of K s t is.affected by the dust particle size and shape as well as by its composition. Thus another dust of any one of the materials in the table may give another value of K s t under the same test conditions.

A P P E N D I X A 6 8 - 5 1

Note also the comparison of some K s t values determined in the H a r t m a n n apparatus. The H a r t m a n n apparatus values do not extrapolate as well for industr ia l -s ize .equipment as do those from tests in equipment of larger volume such as 1 m 8.

Table A-3 (b) Values of Kst for,Typical Dusts

Tests i n 1 m a or l a r g e r vesse l s , l e n g t h - t o - d i a m e t e r a b o u t 1

Chemical igni ter E ,~, 5000 W.s*

D u s t Pmax' • K S t ' • Identity bar bar .m, sec- 1

Tests in H a r t m a n n apparatus

Elec. spark 1.3L vol, E = a few W.s large L / D

P m a x ' KSt' KSt' : bar b a r . m . s e c - t b a r . m - s e c -1

Coal 7.7 85 Dextrin 8.7 200 Organic pigment 10.0 300 Aluminum 11.5 550 Flour 8.6 57 Methyl

cellulose 10 160 Starch , 10 170 .z tspoxy resin 8.2 180 Pharmaceutical

product 9 200 Polyethylene 9 200 Powdered sugar . 7.8 160

Wood - - 230

*W-s = Watt-seconds.

n o i g n i t i o n '

8.5 100 33 9.7 200 73

11.0 450 73

Tests were also conducted in France (Reference 62) with flam-. mable dust ignition in vessels of 1, 10, and 100 m 3 volume. The length-to-diameter ratio of each of the vessels was approximately three. The conclusion from these tests was that the cubic law is reasonably valid. I t was in fact found that this law leads to some overestimation of the vent area needed for large volumes. However, use of the law provides less overestimation than the_ assumpt ion that the necessary vent area=to-volume ratio found for a small volume remains constant through all sizes up- ' to large volumes. Thus, use of the cubic law prevents overexpenditure of money for fabricat ion of excessively large vent areas. This and the ability to predict vent ing areas quant i ta t ively .are the pr imary developments in recent years in the determinat ion of necessary venting areas for deflagrations.


The tests from which the nomographs were developed were conducted from initial pressures of essentially atmospheric• Most dust handling equipment operates at pressures near atmospheric. No specific correlations have been developed and extensively tested for cases of initial pressure substantially different from atmospheric. I t is believed, however, that the homographs will apply to venting deflagrations for dust handling equipment operating:at-the normal (low) gage pressures.

For large vessels, such as vertical cyclindrical storage hoppers or silos for combustible dusts, the required venting area may be as large as, or larger than, the cross section of the vessel. In this case the entire vessel roof can be made a venting area by constructing it as a weak seam roof as described in American Petroleum Institute Standards 650 and 2000.1 Space must be available above the roof for it to open sufficiently. Usually such a roof opens only partway around its periphery to vent a dust combustion. Obviously the roof thickness should be as small as possible, consistent with the strength demands upon it. Large diameter roofs of this type cannot be made self-supporting within the roof slope constraints imposed by the API Standards. Rather, internal roof supporfs will be needed• The roof sheets must not be welded or otherwise attached to the roof supports.

If the required vent area is larger than the vessel cross section, the vessel needs to be further strengthened to take a pressure consistent with the vent area that can be provided. In all cases, the total volume of the vessel should be assumed to contain the combustible dust in suspension, i.e., no credit should be taken for the vessel being partly full of settled material.

The nomographs, as presented for two different ignition energies, hold a significance for'industrial equipment. The graphs using the small value of ignition energy are intended to represent the case where ignition occurs directly within the vessel which is vented and where the ignition .energy is relatively small. This could, for example, be ignition from a hot surface or an electrical spark. For this case, then, venting areas should be chosen according to the homographs presented for this (weak ignition) case.

Another class of ignition source can occur in industrial equipment. Large pieces of dust-handling equipment are often connected b y intervening ducts. In a case such as this, an ignition could begin within one piece of large equipment and thence be conveyed by a connecting duct and enter a second large piece of.equipment. • J As discussed in 2-2.1.5, the burning dust from the duct enters the second piece of equipment as a large, highly turbulent tongue

~Available from the American Petroleum Institute, 2101 L Street, NW.. Washington, ,DC 20037.

APPENDIX A 68--53

of flame. This presents a very large ignition source in the second vessel and also increases turbulence within that second vessel. This results in much more severe combustion than that normally resulting from a small ignition source. For this latter case, the homographs for large ignition energy should be followed.

The use of vent ducts can lead to substantially increased pressure. Donat (Reference 22) cites a specific example. A vent area and vent opening pressure on a particular vessel were such that the maximum pressure developed during the combustior~ was 2.8 psig (0.2 bar). This vessel had no vent duct. When a vent duct of length 3-10 ft (1-3 m) was attached, the maximum pressure developed during venting increased to 9.9 psig (0.7 bar). With a vent length greater than 10 ft (3 m), the maximum pressure during venting increased to 24.2 psig (1.7 bar). Thus, if only a low pressure can be tolerated by a particular piece of equipment, the attachment of a vent duct must be considered very carefully. If a vent duct is absolutely. necessary, both the vessel and the duct must be designed to withstand the full pressure which can be developed dui'ing combustion venting. In order to support reasonably high pressures it is often necessary to construct a vent duct ' in circular cross section. The percentage increase in pressure as a result of addition of a vent duct is greater for dusts with high values of Kst than for dusts with low values of Kst. I t is also greater for the case of large vent areas than for small vent areas. When duct lengths exceed about 10 ft (about 3 m), gas and dust velocity in the duct during venting can be expected to become sonic. In some cases in longer vent ducts, a detonation can occur with resulting' pressures going as high as 30 bars.

In any case, vent ducts for combustion venting need to be kept as short and straight as possible•

Venting panels or "doors" are often used for venting equipment such as bag filters. Such vent panels should weigh no more than 2 lbs/ft 2 (10 kg/m 2) of effective vent opening, preferably less. Such panels need to be hinged in such a manner that they will not fly off of the primary equipment in thecourse of venting a deflagration.

In many cases, the dust-handling equipment to be vented is some form of dust collector such as a bag filter. Filter bags inside the filter vessel, and located close to combustion vents, can be expected to interfere with venting with the result that the pressure increase during the venting can be expected to be appreciably higher than that predicted by the homographs. In the general case, it is better to provide free space just inside the vent area.

Appendix B contains a sample calculation for venting dust deflagrations.


A-4 V e n t i n g of Gas Combus t ion Ins ide Vesse ls . In the usual industr ia l si tuation, the vent ing of gas combust ion inside vessels is the most difficult of the different types of venting. M a n y of the variables listed above under Chap te r 2 may come into play in a single case. I t is not a t all unusual for the gases in industr ia l equip- men t to be at a pressure above, a tmospheric . There ma9 also be a significant degree of initial turbulence and. of turbulence nonuni- for ini ty in the gas mixture before ignit ion occurs. Often the turbulence is brought about by .the mode of in t roduct ion of feed gas s t reams into the vessel. For various reasons, the vent opening pressure may have to be apprec iab ly above a tmospher ic . Geomet ry of the gas space in the vessel may be far f rom an ideal shape such as a sphere or cube. For various geometr ic reasons, the vent may have to be at a nonprefer red location. The likely locat ion of ignition source usual ly cannot be de te rmined . Hence, it usually cannot be. known whether the gases going through a vent, a f t e r -open ing , will be p r imar i ly unburned or burned or a mixture of the two.

Because of such factors the est imation of necessary combust ion vent ing area for a vessel i s c o m p l i c a t e d . Yet, the es t imat ion is fre- quent ly necessary. Because of uncertaint ies yet existent in the technology of vent ing deflagrations, the most conservative case should be assumed. Fo r example , even though a mixture of fuel gas and ox idant in a vessel.woLild normal ly be outside the f lammable limits, jt should be assumed that under the /abnormal condit ions the mixture may .de f l ag ra te and vent ing should be considered.

A-4.1 N o m o g r a p h s for V e n t i n g Combus t ion of Gases. Gas combust ion can be vented when that combust ion is a deffagration. Detonat ions cannot be vented satisfactorily. This guide relates to vent ing for deflagrat ions only.

T h e test p rog ram from which resulted the Richt l in ie (Reference 87) for vent ing dust deflagrat ions also involved a very" large number of tests of vent ing gas deflagrations. .Tests were done in vessels ranging in size f rom 1 m 3 to 60 m 3. T h ( i n i t i a l pressure in the vessels for these tests was atmospheric . Vent areas were varied, as were the pressures at which the vents opened. The vents had low values of mass per u n i t area.

Again, as in the case of combust ib le dusts ment ioned above, it was found that the cubic law holds reasonably well for the venting of gas deflagrat ions inside vessels. I t is no tewor thy tha t data in Reference 33 for vent ing of pentane combust ion in a 60 cu ft (1.7 m 3) vessel checked fair ly well with the da ta in Reference-23 for vent ing of p ropane in a 35 cu ft (1 m 3) vessel. The da ta from the gas combust ion vent ing tests in vessels of various sizes including

APPENDIX A 68--55

60 m 3 were used to construct nomographs for combust ion vent ing (Reference 2). The nomographs are included i n this Appendix . An example calculat ion is given in Append ix B. The nomographs relate to the burning characterist ics of the gases j u s t as the dust nomographs relate to the burning characterist ics of the dusts. The burn ing character is t ic of a gas is called the K~ value. For the purposes of this guide, it is expressed in units of bar • m • sec-k This is, in fact, the "cons tan t " der ived f rom the first equat ion quoted above in Section A-3. The value of KG is influenced not only by the ident i ty of the fuel gas but also by other condit ions affecting .the combust ion, such as init ial turbulence of the gas mixture and t h e type of ignit ion source. T h e value of KG which is considered to be fundamen ta l ly character is t ic of a given gas is ob ta ined by combust ion of the corresponding gas-air mixture inside an unvented vessel with the gas mixture under ini t ial ly s ta t ic condit ions and with an ignition source consisting of an electric spark of about 10 Ws energy. The following table gives character is t ic values of K a for several fuel gases. Also included in the table, for i l lustration, is a t abu la t ion for K c values for some of the gases,for the condi t ion where the gas- air mixture was highly tu rbu len t at the t ime of ignition.

Table A--4. l(a) Tal r _ _ ~ . . . .

Values determined with electric spark ignition source of about 10 Ws*

InitiaU}, static gas Initially highly tur- Fuel Gas mixture bule/at gas mixture

Methane 55 460 Propane 75 500 Hydrogen 550 1270 Propyl acetate 40 Methyl ethyl ketone 56 . Toluene 56 Methanol 66 Ethyl acetate 67

*Ws = Watt-seconds.

The mode of ignition can also have an effect on flame speed and, hence, on the value of Ko. This is shown in the table below," where:. some of the effects are not wha t might be expected.

68-56 EXPLOSION VENTING GUIDE'

Table A-4. l(b) Influence of the Type of Ignition, the Ignition Energy,

and the Degree of Turbulence on the KG Value of Propane (3)

Max pressure from

Ignition combustion in KG, Type of energy, closed vessel,

Turbulence ,gmtaon Ws* bar bar. m. sec- 1

Continuous AC Sp~k ~10 7.5 75

N o n e , Condenser S t a t i c Discharge 100 9.5 750

Pyrotechnic Ignition 10,000 9.5 280 Device

Weak

Continuous AC Spark ,~,10 8.8 370

Pyrotechnic Ignition 10,000 9.5 400 Device

Continuous Strong AC Spark ,~10 8.9 520

*Ws = Watt-seconds.

The nomographs are limited to air-gas mixtures in vessels having length-to-diameter ratios not exceeding about 3 to 5. They also pertain only to initial gas pressure, before ignition, of approximately atmospheric. They also relate only to a small ignition energy from a spark having an energy of about 10 watt-seconds.

As explained above, the value of KG can be increased by initial gas turbulence and- by high energy of ignition soui'ce. If the conditions are such as to produce a high value of Ko, a nomograph based on such high Ko values should be used. For example, if conditions are such that the Ko value for propane could be raised from 75 to about 500, it is recommended that Nomograph I be used. This was developed for hydrogen, which has a characteristic KG value of 550.

A-4.2 Effec t o f F u n d a m e n t a l B u r n i n g Ve loc i ty . The pressure developed during combustion venting is determined in part by the speed with which gases burn during deflagration. This burning speed is in turn determined at least partially by the characteristic rate at which flames proceed through different fuel gases. This characteristic rate is normally measured in terms of maximum fundamental burning velocity, tables of values for which can be found in various handbooks. Some characteristic values are given in Table A-4.2.

APPENDIX A 68--57

Table A-4.2

Maximum Fundamental Burning Velocity

Gas ft/sec m / s e e

Methane (natural gas) 1.2 0.37 Propane 1.5 0.46 Butane 1.3 0.4 Hexane 1.3 0.4 Ethylene 2.3 0.7 Acetylene 5.8 1.8 Hydrogen "11.0 3.4

NOTE: Additional values are available in Perry, R. H., et. al., Chemical Engineers' Handbook, 5th edition, McGraw-Hill, New York (1973), pp. 9-19, and also in Reference 32.

The above values are definitely not the speed with which flames move through a flammable mixture in a vessel. Flame speeds in a vessel are normally much higher than the fundamental burning velocity. However, the fundamental burning velocity is a basic measure of the flame characteristics of a fuel gas. Note that the ratios which the fundamental burning velocities bear to each other for the gases methane, propane, and hydrogen are similar to the ratios of the KG values for these respective gases tabulated earlier. By the same token, as mentioned above, the effects of turbulence on Ka values for low flame speed gases are much higher than they are for high flame speed gases.

From the table for maximum fundamental burning velocities and the table for KG values it is seen that materials having similar maximum fundamental burning velocities have similar KQ values. This observation leads to a method for estimating KG values for gases when these have not been experimentally determined. For example, it is noted that butane has a maximum fundamental burning velocity of 1.3 ft/sec (0.4 m/see). This is close to the value of 1.5 ft/see (0.46 m/see) for propane. Hence, in the absence of experimental measurement, a KG of 75 would be assumed for butane.

Significantly, the gases of most organic compounds have nearly the same fundamental burning velocity. Thus, their flame speeds, and the pressure developed during venting, will be nearly the same. The relative flame speeds for a number of gases .can be seen in the table in Appendix E. In general, the substitution of a halogen, such as chlorine, or certain other atoms in an organic molecule, retards the combustion. On the other hand, introduction of un- saturation into the molecule, as in the cases of ethylene and acetylene, speeds up the combustion:


Data f rom References 13 and 23 can be used to compare the pressures developed dur ing combust ion vent ing for p ropane-a i r and hydrogen-a i r mixtures. As ment ioned earlier, Nomographs G - J have been developed for vent ing of mixtures of air with propane, methane, hydrogen, and coke gas . . (See Reference 3.)

A-4.3 Effect of Initial Elevated Pressure. T h e effect of ini t ial pressure must be corre la ted on the basis of absolute pressures. ,The da ta f rom Reference 13 serve as a basis for correlfi t ing pressures developed dur ing vent ing as a funct ion of the init ial absolute pressure of gases in- a vessel and as a funct ion of the absolute pressure at which a vent opens. I f the rat io of vent burs t ing pressure' to init ial gas pressure in a v6ssel is kept constant , and if vessel s i z e a n d vent size is kept constant , the pressure developed dur ing the venting of p ropane combust ion will vary as app rox ima te ly the 1.5 power of the init ial pressure. The power exponent for p ropane varies from about 1.2 for larger vent ratios (6 ft2/100 ft 3) to about 1.5 for smaller vent ratios (2 ft2/100 ft3). For hydrogen, the exponent ranges f rom 1.1 to 1.2.

, I t is r ecommended tha t the 1.5 power be used in ext rapola t ing f rom N o m o g r a p h G, which is for gases having K a values close to 75. (See sample calculation, Appendix B.)

- - - - " 1 ~ - - " l l I " . . . . . 1 . . . . . . ]

1 2 0 1 llyUlUgCll, tilt; t c commcnucu cxponc~'~t for ' . . . . . . . . . ' : " ' ' mc~cascu m m m pressure is 1.2. For ethylene, it is r e commended tha t an exponent of~ A t.-t be uscu, tMs is untested. The col . . . . reiauun'- may app ly to initial pressures up to about 4 a tmospheres absolute.

A-4.4 Effect of Initial Turbulence. Ini t ia l turbulence presents special problems in app rox ima t ing combust ion venting requirements . Indus t r ia l ly it is often difficult to quant i fy turbulence inside equ ipmen t at the t ime ignit ion may occur. This is due to several factors. I f the gas in a vessel is in motion, there will be turbulence ; the turbulence m a y be nonuniform. T h e gas flows may be abnorma l at the t ime of ignition. Igni t ion may even be caused by breakage of some internal piece of the equ ipmen t and that breakage may itself result in change in the turbulence pat tern . Fo r these reasons, it is general ly assumed that any turbulence in industr ial equ ipmen t at the t ime of ignit ion of combust ib le mixture within it will be considered "h igh ly tu rbu len t . "

In those publ ished tests on the effects of turbulence, the turbulence has usually been quant i f ied in terms of opera t ion of the device used to create the turbulence. References 5, 33, and 57, for example , give da t a o n effects of tu rbulence on pressures developed dur ing combust ion. T h e tables in A-4.1 above show effects of turbulence on Ko values for selected gases. T h e effects of initial high

APPENDIX A 68- -59

turbulence are grea ter on slow flame-speed gases such as methane and p ropane than on high f lame-speed gases such as hydrogen. This in turn means corresponding effects on K a values for these gases.

A-4.5 Effect of Initial Temperature. The effect of init ial t empera tu re is discussed in this guide in 2-2.1.7. Overal l , increase in initial t empera tu re in most cases results in an increase in maxi- m u m rate of pressure rise and a decrease in the pressure genera ted from the combustion. I t is therefore believed that no ad jus tmen t in es t imated pressure deve lopment dur ing combust ion vent ing needs to be made for increase in tempera ture . The same may be true for t empera tu re decrease below ambient , but this has not been proven.

A-4.6 Effects of Combinations of Variables. At the present state of technology, there are insufficient da t a to know definitely jus t how combinat ions of variables may affect the m a x i m u m pressure developed dur ing combust ion venting. For the present, it is suggested that the effects of the variables be assumed to be cumula - tive. This pertains to the var iables discussed in A-4.1 t h r o u g h A-4.5 above.

A-4.7 Effects of Additional Variables.

A-4.7.1 Inertia of Vents. Sometimes a rup ture disc will not provide a sui table vent and a hinged panel or s imilar device must be used instead. T h e weight per uni t a rea of such a device must be as low as possible. Tonk in and Berlemont (Reference 82) presented test d a t a on effects of vent panel we igh t -per -un i t -a rea on pressure deve loped dur ing the vent ing of dust deflagrat ions. Doubl ing of the weight -per -un i t -a rea of a vent panel of fixed area resulted in a doubl ing of the pressure developed dur ing venting of deflagrations. Tr ip l ing of the weight per uni t a rea resulted in a t r ip l ing of the pressure developed dur ing the combust ion

.venting. Suggestions have been stated above for l imit ing the weight -per -un i t -a rea for vent ing devices. Normal ly , vent ing devices should have a mass / a r ea rat io less than 2 lbs / f t 2 (10 kg/m2) .

A-4.7.2 Vent Ducts. Preferably there should be no ducts in the vent ing system. If a duc t must be employed, it should have jus t as short an effective length ( including effective length of bends in producing pressure drop) as possible. Any such bends should be min imal ; the vent duc t should preferably be straight. The effect of ducts on the pressure developed dur ing combust ion vent ing of gases is the same as tha t discussed for dusts in Section A-3. Combus-


tion can take place in the vent duct itself, i.e., unburned gases may be .the first to exit from the vent. This has two implications. First, the duct should be capable of withstand!ng a pressure at least as high as that expected to develop in the vessel during venting. Second, high turbulence can develop in the duct and could possibly lead to transition of the deflagration to detonation. In that case, far higher pressure could develop in the duct. The vessel could be exposed to similarly high pressures.

A-4.7.3 Ignition Source Location. There are few data on the effects of ignition source location on the pressure developed during combustion venting. In the normal industrial case, the potential ignition source location most frequently cannot be predicted. I t is generally assumed that the ignition source will be at such a location as to generate the max imum pressure during venting of a deflagration, as stated by the nomographs.

In cases where multiple simultaneous sources of ignition are probable, it is suggested that the multiple ignition be assumed to produce the effect of initial high turbulence. (See 2-2. 1.5 for further discussion.)

A-4.7.4 Other Oxidants. Relative to the data on venting gas/a i r deflagrations, there are little data on effects of other oxidants. This is discussed in 2-2.1.3. In addition to various oxygen concentrations, other oxidants can include oxides of nitrogen and halogens. These are not covered in this guide. If direct data are not available for the system being considered, tests are recommended.

A-4.7.5 Fogs and Mists. Flames can propagate through fogs and mists at mixture temperatures below the flashpoint of the liquid involved. I t is suggested that fogs and mists be treated as gases in estimating pressures developed during venting of deflagrations.

A-5 Vent ing of Gas Combust ion Inside Air Convey ing Ducts. Most of the cases of flammable gas mixtures inside ducts of the air ventilation type occur at initial internal pressure of nearly atmospheric. The venting of combustion in such ducts is discussed in Appendix C.

" d o ~ - ' l r

i i

• !

71 -

i i ; i

i : : i

!.

W ~ 9

0 U ~ . I - - " - - I " - m

N

a . ~ " •

~-~ ~ .

, t t

A P P E N D I X A

i i e

I ' ,

V V A - - ' ~

68-61

.¢

O

v - O

Z

o

t ~

10 0,1 1 10 100 Nomograph B

Pr•=,d : O.4bar

0.6bar 0.8 bar 10bar ~:Sbor

2,0 bar

o

i COMBUSTION VENTING FOR " i~-

i_ . i . . . . . . . . . . t i ' '

. . . . . . . . . : [ ' ' I ill,,

!--I ! - i i i i . - ' ,~ - -L ..... KST [ bar . m-s " I ] I ! ~- ~--

WEAK IGNITION SOURCE

i 1 < KST<100 : St1 I 101 < KST<200 : St2 -

-'~-~-KST >201 : St 3-- .......... I - I. VENT AREA, F

10 1

Sf

I i I : . . . . . . ~ . . . . . . . . .

VENT REL~SE PRESSURE . . . . i- " - - PSTAT " 0.5 BAR GA . . . . . . . . . .

= @XIMUM PRESSURE DURING VENTING, PRED " BAR GA.

F " - : . i ~i I I . . . . . . . ~ - -~ - ~ -~" ~ . . . . . . . . . . . . .

I ,i.Li " . . . . . . . . . ii

r . . . . . r - - - ~ - I I i

I

i

J l - !

- l j

i

- " T . . . . . i

. . . . i ....... I . . . . . T ' ;

0.1 .I Nomograp1", C

i ; i l i '~T~, I - . ' i

ii~ ; J = i i r r ! ; ! _. / i VESSEL VOLUME, V I

10 100

Pred:

0,6 bar 0.8 bar 1.0 bar 1,5 bar 2.0 bar

O~

%\\

~___ ~,~F '

KsT[bo,.m.~-l] STRONG IGNITION SOURCE

1 <KsT <200: St 1

201<KsT <300" St 2

301<K~r • St3

COMBUSTION VENTING'FOR' FLAMMABLE OUSTS

VENT AREA, iF

10 1

; I I -- ~ .... L i ' I !

-:--i . . . . . . . ~ I , ; i ' l ,

. . . .

t!

,I VENT RELEASE PRESSURE

- - PsTAT ' 0 . 1 BAR GA.

__MAXIMUM PRESSURE DURING VENTING, PRED ' BAR GA.

- - ~ , ~ F r . ~ i i . . . .

.Ei APeE~ -

//

I i I

0.I 1

N'omograpb D

10

I I l i ~ri " V:,ESSEL VOLUME, V

100

I

, !1

. . ..... I

i I ,

CO~USTION VENTING FOR F ~ B L E DUSTS

i i! I

' i

KST [bor.m.s STRONG IGNIT-ION SOURCE

1 < KST < 201< KST <

301< KST

200 :St 1 _

300 : St 2 ___

: St3

VENT AREA, F 9

IJ, [ 10 1

, 1 I t

' I !

0.1 1 10 N o m o g r a p h E

100

(t bar bar bar bar bar bar

B

i

9£

TT COMBUSTION VENTING FOR + , .

~F~M~t4BLE DUSTS I I ',

'('qC~ ~ -+--- , ..... , rITI-T ., ' : , z; :i I

• ~ , . + i +~, +

+ :+, il; i+;-.t ! + +

ST,-0"NG +,T,0N

i' ,, , • i '+"I , - - ~ 4- - - I F - - L - + - - - 7 . . . . 4 t ~- ! " - ~

I : I I I | , i i ' l i ' ~l !

- ~ - - - ~ - - - ~ - - ~ - ~ ~ t " ~; t l I I i i ' i i . :~ ;

-;,,:NT ~R~,~ i:i'i ! ' ';Ira 2] i =, !:~ i

I l l [ [ l i li~ll, i I 10 1

- - - 9

10 100 Nomogra'ph F

I

M

I

L£

°

~ PRESSURE, PSTAT

PSTAT = 0.i BAR GA; PR~ ~0.2 BAR GA

PSTAT = 0.2 BAR ~; PRED ~ 0.4 BAR

PSTAT = 0.5 BAR ~; PRED ~ 0.6 BAR

!:i

i

VENT(M 2)AREA' F i

0,1 1

~0 10 1 0,1

PROPANE, SOLVENT VAPOR: K G = 75 BAR . M . SEC -1 WEAK ~GNITION STATIONARY MIXTURE AT TIME OF IGNITION±

" I i , i + o- - MAXIMUM PRESSURE DURING VENTING. PRED" BAR G A . ~ ~ I

I

i +, + I I

~- VESSEL VOLUME, V

l i l I ~ L,~ ~VI! t lF l l ; - I ! I I I l l i !

1

Nomograph G 10 100 1 0 0 0

0 .2 ' b a r

0 ,4 b a r

bQr

b a r

I

....1

I0 1

- p -nlBA °REIRE P T0 , " i ~ H r sTA~-°' ' R~-- " \ " L I , , ~J .H f \ \ "PsTAT = 0.2 BAR GA; PRED > 0

L ~ l l l f p = 0.5BARCA; P 70 \ ~ ' ~ I STITI I I RED

"K~\il I I t ~ ,',k I I

I :

I I I J I I I

I l l l l l l : ' , _ _ \ \

\ \

VENT AREA, F (M 2)

O.I 1

N o m o g r a p h H

10 100

i bar

, bar ;bar bor bor bQr bar

I000

Qo I

Oo

.¢ t~ Z

C~

C~

40

II I

VENT RELEASE PRESSURE, PSTAT

PSTAT = 0.1 BAR CA; PRED ~ 0.2 BAR CA

I ~ f PSTAT = 0.2 BAR GA; PRED ~ 0.4 BAR CA <l/ <l/

PSTAT = 0, 5 1

,,

I I

• (M2) If [

- I I

40 10 1

CA; PRED ~ 0.6 BAR CA

i I i !

VE,TA,EA.F

0.i

I I I I I I I I I I I I I l l l l l l

HYDROGEN: K G = 550 BAR" M' SEC -1 BEAK IGNITION STATIONARY MIXTURE AT TIFE OF IGNITION

1

N o m o g r a p h I

MAXIFI/~ PRESSURE DURING VENTING, PRED' BAR GA

I

t

~ ~ : : . . . . . .

I ~ I I I

_ - 1 I I I 1

I _ 1 i i i i I - - L I - - - - + -4 i i I

VESSEL VOLUME, V (M 3)

10 100

l l l l l ~ . / . . . . . . j ~

f

I i I

! i i

2bar

1000

bQr bar bar

$£

68-70 E X P L O S I O N - V E N T I N G G U I D E A P P E N D I X B 6 8 - 7 1

t'~

O "

i t - : I £ == t i l l t / , / Y l - I i i ~ 1 1 = , _ - , _ , _ ,._ ' L I . - ~ - -

.- I I I I I I :1 ~ "- "- " - ' . J ' , Z - , ,K -~ ;~ - - I I I I I ~ - I <~ -= -

llllt]

O Q o

o 9

9

e~

0

O

0

Appendix B Sample Calculations

This Appendix is not a part of this N F I 'A document, but it is included for information purpose.~ only.

Part I

Sample Calculations for Combustion Venting for Buildings

B-1 ings is given in A-2.2 as follows:

CLtL2 A v =

There is an important constraint:

L3 _< 3 ~/L~E~

La = longest building dimension, ft or m

Summary. The equation for combustion venting for build-

(E n 1)

B-2 Other Important Considerations.

B-2.1 M a x i m u m Allowable O v e r p r e s s u r e , ' P . The purpose o f b u i l d i n g ven t i ng is to prevent serious structural damage and production of lethal projectiles. The max imum allowable overpressure, P, is the static loading which the weakest member of the structure can withstand.

Care must be taken to be sure the weakest member is recognized. All members of the structure, walls, windows, floors, ceilings, and roofs need to be considered. I t is important to keep in mind that floors and roofs are not often designed to take much loading from beneath.

A qualified structural designer is needed for making the evaluation. His analysis must be based on the actual design and condition of the structure to be protected.

B-2.2 The Constant, C. The value of the constant depends on the type of f lammable gas o/- dust present and wl~ether the calculation is made in English units (ft, lbf/in. 2) or in SI units (m, kN/m~). For gases such as natural gas, pentane, benzene, acetone, or vinyl acetate, and for most f lammable mists and dusts, the recommended value for the constant is 2.6 in English units, 6.8 in SI units.

Gases like ethylene, butadiene and hydrogen burn faster than aliphatic or aromatic hydrocarbons. The recommended value of the constant for ethylene or butadiene is 4 in English uni t s , 10.5 in SI units; for hydrogen 6.4 in English units, 17 in SI units.

/

t ~


B-2.3 V e n t Area . As ment ioned earlier, the procedure is based on the assumption of open, unobs t ruc ted vents. See Append ix A for vent panel design cri teria which permit vent ing per formance to approach that of open vents.

B-2.4 B u i l d i n g Re in fo rc ing . I t is not always feasible to obta in the vent a rea required to protect the integri ty of a bui ld ing unless some of the weaker members of the structure are reinforced. When re inforcement is required, the expected overpressure needs to be calculated, using the actual vent area. Weak members must be reinforced to withstand the expected overpressure.

Pe = ( P ) ( A / A a ) 2 (Eq 2)

Pe = expected overpressure, lbf/in.2, i.e., psi, or kN/."m 2 P = overpressure used in calculat ing required vent area,

lbf / in . 2 or k N / m = A = requi red vent area, ft 2 or m = Aa = vent a rea actual ly obta inable , ft 2 or m =

B-3 Sample Ca lcu la t ion . Consider the following bui ld ing:

INot to ~ l e ) I " 170 ft (51.8 m} L I

: ....... ... ...... t30,, _ 1 2 , 8 o ,

60 ft 118.3 •

End Wall l End Wall II

. . . . . . . . . . . . . . . . . . . . . . . . . . . . t ...... T 20 ft {6.1 m~, i . . . . . .

[= 6Oft ~ 5 0 f t

19:15 m)

B-3.1_ Calcula te the requi red vent area according to equat ion (1) in Section B-1. Before the calculat ions can be made, the shape of the bui lding should be "no rma l i zed , " a n d the bui ld ing considered as several separate parts. Normal iz ing consists of combining ,several i r regular portions o f a bui lding into a simple rec tangula r s tructure o f equivalt~nt surface area and volume. The bui ld ing shown here can be divided into three par ts and each par t normal ized as follows:

APPENDIX B 6 8 - 7 3

B-3.1.1 Consider the "foot" of the " L " as one structure, called Par t I in the i l lustration. Normal ize the height to the area of End Wal l I.

Wal l I a rea = (60 x 20) + (0.5 x 30 x 10) = 1,350 ft 2 (125.4 m 2)

Normal ized height = 1,350 fC/60 = 22.5 ft (6.88 m)

T h e vent area requi red for Par t I will be the same as for a s tructure with :

L1 = 22.5 ft (6.86 m) L2 = 50 ft (15.25 m) La = 60 ft (18.30 m)

B-3.1.2 T h e " leg" of the " L " is normal ized on the basis of the area of End Wal l I I .

Wal l I I a rea = (30 x 20) = (0.5 x 30 x 10) = 750 ft 2 (69.7 m 2)

Normal ized height = 750/30 = 25 ft (7.62 m)

Therefore, L1 = 25 ft (7.62 m) L2 = 30 ft (9.15 m)

The m a x i m u m permissible length to which Lx and L2 can be appl ied is:

La = 3v 'L-~2 = 3 x 27.4 = 82.2 ft (25.1 m)

T h e actual length of the leg is 170 ft -- 50 ft = 120 ft (36.6 m). Therefore, the leg needs to be considered as two parts. These are designated Par t I I and Par t I I I in the i l lustration. Par t I I and Pa r t [ I I require the same vent area according to the calculat ion procedure. The bounda ry between Par t I I and Par t I I I should be located in a way which permits symmetr ical dis t r ibut ion of vent a rea over the entire building. Sometimes it is not necessary to locate the bounda ry to make calculations. Sometimes it is. Both cases will be i l lustrated.

B-3.2 Calculate the Vent Area for Normal ized Parts I, II, and III. Let us assume for purpose of i l lustration that the maxi- m u m al lowable overpressure for the bui lding has been es t imated by a s t ructural engineer to be 0.5 psi (3..45 k N / m 2) and tha t the f l ammable mater ia l in the bui ld ing is a typical a l iphat ic or a romat ic hydrocarbon for which the C value is 2.6 in English units, 6.8 in SI units. The vent a rea required for each par t of the bui ld ing can be calculated as follows:

4~


B-3.2.1 For Part I:

Ll = 22.5 ft (6.86 m) L2 = 50 ft (15.25 m) A1 = 2.6 x 22.5 x 50/(0.5) l/s = 4,137 fC, or A1 = 6.8 x 6.86 X 15.25/(1.857) 1/2 = 384.3 m 2

I t is recommended practice to locate vents only in roofs and outside walls. Inside walls and the imaginary boundaries used to

fac i l i t a te calculation of vent areas cannot be included as vents. Consequently, for Part I the max imum area available for. locating vents is:

Wall surface --- (60 x 20) -k- (0.5 x 30 x 10) + ( 5 0 x 2 0 x 2 ) Jr- ( 5 0 x 1 0 ) n t- 30 x 2 0 = 4,450 ft s (413.4 m s)

Roof surface ~ 50 x 60 = 3,000 ft 2 (278.7 m 2)

The total max imum available surface ~ 7,450 ft s (692.1 m2). For Part I, the vents can be confined to the walls only if virtually

all the wall surfaces are available. In actual practice, wall surfaces are rarely totally available for the location of explosion vents. I n a real case some of the vent area would have to be located in the roof. If roof locations were not available, some members of the structure would have to be reinforced to prevent serious damage.

P--a 2 2 For Parts ir ~_a TTT.

L1 = 25 ft (7.62 m) L2 = 30 ft (9.15 m) A2 = 2.6 x 25 x 30/(0.5) 1/2 = 2,758 fC, or As = 6.8 x 7.62 x 9.15/(3.45) 1/2 = 256.2 m s

The total vent area required for Parts I I and II1 combined is:

2A2 = 5,516 fC (512.4 m 2)

This is the vent area required to protect the "leg" of the "L" of our illustration.

The total wall surface available on the "leg" is:

750 q- 1 2 0 x ( 3 0 q -20) = 6,750 ft s (627.1 m s)

The roof surface is about :

30 x 120 = 3,600 ft 2 (384.4 m 2)

The required vent area needs to be distributed as symmetrically as possible over the available surfaces. ..

Since there is plenty of available vent area, location o f the boundary between Part II and Part I l l is not required in this example.

APPENDIX B 68- -75

B-3.3 Est imate the Effect of Actual V e n t Area on Expec ted Overpressure . I t is not always possible to get sufficient-vent area to assure that the m a x i m u m allowable overpressure will not be exceeded. For example, let us assume that the leg of our "L" shaped building has certain venting restrictions; namely, that only the end and front walls are available for venting. The back wall and roof cannot be used. The available wall surface in this example is :

750 + 120 x 30 = 4,350 ft s (404.1 m s)

This is not enough vent area to keep the m a x i m u m expected overpressure below 0.5 ps i (3.45 kN/m2). Serious explosion damage could be expected unless the structure is strengthened. There are two ways to proceed.

B-3.3.1 Consider the leg as a single structure. F rom equation (2) the expected overpressure with the limited vent area is :

Pe = (P) ( A / a a ) 2 = (0.5) (5,516/4,350) 2 = 0.80 psi (5.5 k N / m 2)

The members of the structure, walls, ceilings, roof, etc., which cannot withstand this amoun t of overpressure will need to be strengthened so that they can withstand the expected overpressure.

B-3.3.2 Consider the leg as two parts (Part I I and Part I I I ) . By proper location of the imaginary boundary, Par t I I can be sized with sufficient vent a rea- to keep its expected overpressure to 0.5 .psi. I n this case the length, La, of Part I I needs to be enough for the required vent area of 2,758 fC (256.2 m s) to be provided.

750 ft s -t- 30 La = 2,758 ft 2 La -- (2,758 -- 750)/30 = 67 ft (20.44 m)

Since this dimension is greater than L2 and less than 3x,/LIL2, it meets the criteria for length and results in acceptable proportions for Part II. With these dimensions, Part I I would not need to be reinforced. The length of Part I I I is consequently 120 -- 67 = 53 ft (16.2 m). This length likewise falls between L2 and 3 v/-L-xLs for Part I I I , so tha t . the proportions of Part I I I are acceptable. However, the area available for vent ing is only 30 x 53 =- 1,590 ft s (147.7 m2). The expected overpressure is then Pe = (0.5) (2,758/1,590) 2 -- 1.5 psi (10.3 kN/m2). The structural members which comprise Part I I I would therefore need to be strengthened to withstand this overpressure.


B-3.3.3 Final ly, we can consider the case where a ma jo r pa r t of the bui ld ing cannot be vented except into another section of the building. Assume, for example, tha t Par t I can only vent into Parts I I and I I I . In this case the effective vent area for Par t I is only 750 ft 2 (69.6 m2), and the expected overpressure is:

Pe = (0.5)(4,137/750) 2 = 15.2 psi (104.9 k N / m 2)

S t ruc tura l fai lure will a lmost cer ta inly occur in Par t I. T h e case illustrates the need for symmetr ica l dis t r ibut ion of

vents. I t shows that if vents are not symmetr ica l ly d is t r ibuted significant overpressures can develop in par ts of a building. Severe d a m a g e will occur.

B-4 C o m p a r t m e n t s . Buildings are often compar tmen ta l i zed into several rooms and levels. Usual ly each c o m p a r t m e n t must be t rea ted as an independen t unit for purposes of eva lua t ing its explosion protect ion requirement . I t is not r ecommended pract ice to vent one c o m p a r t m e n t into another ; vent ing needs to be outside.

Consider, for example, a fully enclosed room with dimensions 15 x 20 x 30 ft located on the second floor of a building. T h e dimensions of the outside wall avai lable for vent ing are 15 x 20 ft. Fo r a typical hydroca rbon combust ion the expected overpressure, by r ea r r angemen t of equation (1) is:

P = (2.6 L1L~/A) 2 ---- [2.6 x (15 x 20) / (15 x 20)] 2 = 6.8 psi (46.9 k N / m 2)

T h e inside walls, doors, windows, duc t seals, floor, ceiling, etc., must all be able to withstand this overpressure in order to prevent s t ructura l fai lure of the rooms. A thorough investigation of the room and its components is requi red to b e sure every member is reinforced as necessary.

Part II

Sample Calculations for Combustion Venting for Dust Deflagratlons Inside Equipment

B-5 I n t r o d u c t i o n . The object of vent ing is to allow for the relief of the products of combust ion of a dust before a pressure can develop which will cause d a m a g e to the conta in ing vessel. Closed vessel combust ion of most dusts with air, at a tmospher ic pressure, can develop a m a x i m u m pressure of about 100 psig (7 ba r ga.). A vent which has low mass per unit area, opens at a low pressure, and has sufficient a rea can reduce the m a x i m u m pressure developed to a lower value.


The required vent area, for a vent of low mass /a rea , is a function of vessel volume, dust class, vent release pressure, ignit ion energy, and the m a x i m u m pressure not to be exceeded dur ing the dust deflagration. These variables and their effects are discussed in Chapters 2 and 3, and in Append ix A. Nomographs A - F , a t tached to Ap- pendix A, can be used to de te rmine the v e n t a rea requi red for a vessel. Units of length are expressed in the nornographs as meters, units of pressure as bars gage.

B-5.1 Vessel Volume (m3). For vent ing calculations, the worst case must be assumed. This is the full volume of the vessel.

B-5.2 Dust Class (St- l , St-2, St-3). "Dus t Class" is re la ted to the m a x i m u m rate of pressure rise dur ing a dust combust ion. This must be de te rmined in a closed (unvented) test vessel of sufficient pressure capabi l i ty and equ ipped with high-speed press.ure recording equipment . T h e recorder gives a g raph of pressure versus time. F r o m this g raph the m a x i m u m slope can be de te rmined . This is called (dp/dt)max and can be expressed in units of bar / sec . This value is then mul t ip l ied by the cube root of the volume of the test vessel, V */a. Vessel volume can be expressed in cubic meters. The p roduc t of the two terms thus de te rmined gives the value, Ks t , for the dust :

(dp/dt)max • V t/a = K s t

The units of Ks t , thus ca lcu la ted , are ba r • meter • sec - l . The following are the ranges of K s t for the various dust classes.

Dust Class Kst, bar . m -sec -~

St-I _< 100 St-2 101-200 St-3 < 200

B-5.3 Vent Release Pressure (bar ga.; Pstad. Open ing of a vent dur ing combust ion will reduce the pressure development . T h e lower the pressure at which the vent opens, the lower will be the m a x i m u m pressure developed dur ing the combust ion. Vent release pressure should be as low as possible wi thout being so low that normal pressure variat ions in the vessel will open the vent.

The vent ing device needs to have a low inertia, below 2 lb / sq ft (10 kg/m2) . A thin, rup tu re - type membrane device is avai lable for this purpose. In some cases a vent ing panel, not exceeding the m a x i m u m al lowable inertia, is used. T h e design must prevent bu i ldup of deposits on the inside of the vent ing device. I t must also prevent malfunct ion due to snow or ice on the outside. The vent ing

68-78 E X P L O S t O N V E N T I N G G U I D E A P P E N D I X B 68-79

device must not impose hazard to people or equipment when it opens. For example, a vefft panel will need suitable hinging. Similarly, the large ball of flaming dust that comes out during venting must not impose hazard to people o r equipment.

B-5.4 Ignition Energy (Weak, Strong). The nomographs for dusts (Nomographs A-F) provide for two types of ignition of a dust deflagration. A "weak" ignition occurs within the vessel, for instance, f rom a hot surface or a static spark. A "s t rong" ignition occurs when a burning cloud of dust enters the vessel through a duct, or if an open fire exists when the.dust cloud is formed in the vessel. I f there is any doubt about potential ignition sources, assume a strong ignition.

B-5.5 Desired Maximum Pressure (bar ga.; Pr~a). When a dust deflagration occurs, the max imum pressure to be reached during venting should be no more than two-thirds of the pressure which will cause the weakest part of the vented vessel to break. This location is usually a welded joint at the roof or at a -bo t tom c o n e .

B-5.6 Sample Calculations.

B-5.6.1 Assume:

(a) Vessel has a f l ammable 'mix ture of air with a dust of class St-1.

(b) Vessel volume is 20 m3; vessel L / D is 2,

.(c) Initial pressure is atmospheric.

(d) Ignit ion energy is assumed to be large.

(e) Vent release pressure is 0,1 bar ga.; venting device has mass per unit area less than 2 lb / fC (less than 10 kg/m~).

Determine the vent area to prevent pressure from exceeding 0.2 bar ga .

Use the dust venting homograph, Nomograph D, for vent release pressure of 0.1 bar ga. and Istrong ignition source. On the right hand side of the homograph draw a line vertically upward to 20 m 3 to intersect the 0 .2 bar line for Pr~a. From the intercept draw a horizontal line to the left to intersect the line for dust class St-1. From that intercept draw a line vertically downward to the line showing vent area. T h e necessary vent area is 2.3 m s.

See Appendix A for discussion of the need to vent to the outdoors and the limitations of vent ducts.

Part III

Sample Calculation for Venting Gas Deflagrations Inside Equipment

B-6 Introduction. The object of venting is to allow for the relief of the products of combustion of a gas before a pressure can develop which will cause damage to the containing vessel. A non- vented deflagration of most gases with air, beginning with an initial pressure of atmospheric, can develop a pressure of about 100 psig (7 bar ga.). A vent which has low mass pe r unit area, opens at a low pressure, and has sufficient area can reduce the m a x i m u m pressure developed to a lower value.

The required vent area, for a vent of low mass/area, is a function of a vessel volume, gas .Ko value, vent release pressure, amount and type of igni t ion energy, initial turbulence, and the maximum pressure not to be exceeded during the gas deflagration. These variables and their effects are discussed in Chapters 2 and 3, and in Appendix A. Nomographs G - J in Appendix A can be used to determine the vent area required for a vessel. Units of length are expressed in the nomographs as meters, units of pressure as bars gage.

B-6.1 Vessel Vo lume (m3). For venting calculations, the worst case must be assumed. This is the full volume of the vessel.

B-6.2 Gas KG Value . The KG value for a gas is a function of the fundamental burning velocity of that gas. Th e KQ value is related to the max imum rate of pressure rise during a gas deflagration. This must be determined in a closed (unrented) test vessel of sufficient pressure capability and equipped with high-speed pressure recording equipment. The recorder gives a graph of pressure versus time. F rom this graph the max imum slope can be determined. This is called (dp/dt)max and can be expressed in units of bar/see. This value is then multiplied by the cube root of the volume of the test vessel, V l/8. Vessel volume can be expressed in cubic meters. The product of the two terms thus determined gives the value, KG, for the gas:

(dp/dt)m~x • V 1/3 = KG

The units of Ko, thus calculated, are bar • meter • sec-~. Gases of most organic compounds have similar fundamenta l

burning velocities and hence have similar values of KG. The Ka value for such materials is about the same as that for propane, i.e., 75. This includes Ka values for materials like butane, benzene, gasoline, acetone, ethyl acetate, and many others. , L


As shown in Table A-4.1(a), hydrogen has a Ko of 550. Certain other gases also have higher KG values than 75. If there is substantial doubt about the Ko value for a particular gas, that gas should be tested as described above for determining Ko values.

B-6.3 Vent Release Pressure (bar ga.; Pstat). Opening of a vent during combustion will reduce the pressure development. The lower the pressure at which the vent opens, the lower will be the maximum pressure developed during the combustion. Vent release pressure should be as low as possible without being so low that normal pressure variations in the vessel will open the vent.

The venting device needs to have a low inertia, below 2 lb/sq ft (10 kg/m~). Since operating pressures in gas handling equipment are often substantially above atmospheric, the venting devices are frequently some form of rupture disc. The design must prevent buildup of deposits on the inside of the venting device. It must also prevent malfunction due to snow or ice on the outside. The venting device must not impose hazard to people or equipment when it opens. Similarly the large tongue of flaming gas issuing from the vent must not impose hazard to people or equipment.

B-6.4 Ignition Energy (Amount and Type). The homographs for gases (Nomographs G-J) are based on a small or weak ignition energy, about 10 Ws (watt-seconds), from a small, continuous electric discharge. Other forms of ignition can result in high flame speeds with the consequent need for larger vent area. For example, the normal Ka for propane is 75. As shown in Table A-4.1(b), the Ko for propane changes to about 750 when the ignition is by 100 Ws of energy from a condenser discharge in a spark. If ignition in the practical case could give a high KG value for propane or other similar Ko-value gas, .the nomographs for hydrogen (Nomograph I, KG = 550) can be used instead of that for propane (Nomograph G). A slight extrapolation of required vent area may be advisable on 'account of the 750 value for propane Ko versus the 550 value for hydrogen.

B-6.5 Initial Turbulence. Nomographs G-J are for fuel gas mixtures with air which are quiescent at the time of ignition. The test vessels also contained, no internal obstructions that would lead to turbulence increase during combustion. Turbulence leads to increase of KG for the gas. Thus, as shown in Table A-4.1(a), initial high turbulence increases the Ko for propane from 75 to 500. For such a case the nomograph for hydrogen, Ko = 550 (Nomograph I), can be used.

Note that, as shown in Table A-4.1(a), the effects of initial turbulence vary with the basic KG value for the gas.


B-6.6 Desired Maximum Pressure (bar ga.; Pred). When a gas deflagration occurs, the maximum pressure to be reached during venting should be no more than two-thirds of the pressure which will cause the weakest part of the vented vessel to break.

B-6.7 Initial Elevated Pressure. Quite often, industrial equipment which must be vented for combustion is operated at pressures significantly above atmospheric. Nomographs G- J are for gases at initial pressure of atmospheric. In A-4.3 the calculation for effect of initial elevated pressure is discussed. For gases having a Kc of about 75 the pressure resulting from venting of a vessel at initial elevated pressure will vary approximately as the 1.5 power of the absolute pressure of gas in the vessel before ignition. The exponent for gases having higher Kn values is discussed in A-4.3.

B-6.8 Sample Calculations.

B-6.8.1 Assume:

(a) Vessel has flammable mixture of air and acetone. (b) Vessel volume is 20 m3; vessel L /D is 2. (c) Initial pressure in vessel at time o f ignition is 2 bar

absolute.

(d) Ignition energy is small, say 10 Ws or less; no initial turbulence and no internal obstructions to cause turbulence generation.

(e) Vent release pressure is 2.2 'bar absolute; vent device has low mass/area, less than 1.5 lb/fC (less than 7.5 kg/m2).

Calculate the pressure resulting from venting with a vent area of 2.0 m 2.

Assume the flame speed in air-acetone mixture will be approximately the same as that for propane. On this basis, assume Ko is 75 and use Nomograph G.

Vent release absolute pressure/initial pressure = 2.2/2 -- 1.1. Therefore read from Nomograph G where vent release pressure = 0.1 bar ga. Enter the nomograph on the left hand side at vent area of 2 m s. Draw a line vertically upward to intersect the slanting line for P~tat of 0.1 bar. From the intercept draw a horizontal line to the family of lines for Pred. Also draw a line vertically upward at 20 m ~ to intersect the horizontal line just drawn. These two lines intersect at the slanting line for 0.6 bar ga.

The 0.6 bar ga. equals 1.6 bar absolute. On the basis of an initial pressure of 2 bar absolute, the peak pressure reached during venting will be (2 bar abs/1 bar abs) 1"5 x 1.6 bar abs. The calculation result is a final pressure of 2.83 x 1.6, or 4.53 bar abs, or 3.53 bar ga. The 3.53 bar go. equals 51.2 lb/in, go., or 353 kN/m 2.

t:


As an al ternate , a series of .tests could be made in a smaller vessel of, say, 1 m 3 volume. The tests could be conduc ted with a p ropane-a i r mixtt lre in which the p ropane concentra t ion is 10 percent higher than stoichiometric. In i t ia l p ressure would be 2.0 ba r abs. Vents of low mass / a r ea and of various areas would then be tested. T o achieve, with the 20 m 3 vessel, the same pressure as tha t found in a test on the 1 m 3 vessel with a given vent area, tha t vent a rea would be ca lcu la ted over the 20 m 3 vo lume on the basis of the cubic law. The equat ion to use is:

FtV2 2/3 F2

V 1 2 / 3 .

For definit ion of t e rms see Section A-3 of this guide.

APPENDIX C 68- -83

Appendix C Exp los ion Rel ie fs for Ducts and Elongated Vesse ls (31)

This Appendix is not a part of this N F P A document but is included for information purposes only.

Wherever f l ammable vapors and gases occur inside an industr ia l plant , there is the danger of a gaseous explosion. The main precaut ion taken to avoid an explosion is to prevent the concentra t ion of the f lammable gas or vapor f rom falling within the f l ammable limits in air. Thus, when pure me thane ' i s passed th rough a duc t there is no danger of explosion unless an accident occurs which results in an approx imate ly seven-fold di lut ion with air of the methane in the duct ; this would bring the concentra t ion of me thane down to the upper explosive limit. Conversely, in o ther systems precaut ions can be taken to reduce the concentra t ion of f l ammable gas to well below the lower explosive limit. For many processes, it is not possible to be s u r e that at all times the concentra t ion of f lammable gas will be outside the f l ammable limits. U n d e r these conditions, the p lant has to be designed so tha t if an explosion were to occur the m i n i m u m amoun t of damage would ensue. One of the ways in which this is done is to use explosion reliefs. These are provided on the side of the piece of equ ipmen t concerned and are designed to open very early in an explosion and allow the harmless release oI the products of combust ion of the explosion. T h e a rea of these vents should be large enough to relieve the explosion gases sufficiently quickly to prevent the m a x i m u m pressure f rom reaching a value greater than the pressure the conta iner can withstand.

Plants i n which f l ammable gases and vapors are hand led in industry vary widely in size and shape and duct systems of differ- ing degrees of complex i ty a re used to connect i tems for p lants in which various processes are carr ied out. I n fo rma t ion on the pro- v i s i o n of explosion reliefs for containers approx ima te ly cubical in shape have been publ ished elsewhere following the work of the Gas Council on Explosion Reliefs for Dry ing Ovens (Reference 71). In this note, design da t a for explosion relief for ducts and e longated vessels are provided. These da t a are based mainly on work carr ied out at the Jo in t F i re Research Organ iza t ion on the vent ing of gaseous explosions in ducts (References 81 and 82). T h e ducts varied in dimensions f rom 3 in. d iamete r to 12 in. square section and f rom 6 to 30 ft l o n g . . In most experiments, p r o p a n e / a i r or p e n t a n e / a i r mixtures, were used as the explosive gas, a l though a few exper iments were carr ied out with e thy lene /a i r and m e t h a n e / air mixtures. The corre la t ion of the results of this work and also the inclusion Of other sources of informat ion do allow, however, the results to be ex t rapola ted to ducts with diameters u p to about 2 ft 6 in. and to a number of other gases.


Scope. This guide may be used to design explosion reliefs for ducts and

elongated vessels where L /D is equal to or greater than 6 and D does not exceed.2 ft 6 in. The basic formula given in the section entitled "Size and Spacing of Explosion Reliefs for Stationary Gases or Gases Moving at Speeds of less than 10 ft/s" will provide design data for straight, unobstructed ducts containing propane/air mixtures moving at velocities of less than 10 ft/s.

If the ducts are not straight or contain obstacles, additional relief is required and this may be calculated by using the information given in the section entitled "Size and Spacing of Explosion Reliefs forSta t ionary Gases or Gases Moving at Speeds of less than 10 ft/s."

The section entitled "Size and Spacing of Explosion Reliefs for Gases Moving at Speeds of 10-60 ft/s" deals with propane/air mixtures which are moving at velocities of between 10 ft/s and 60 ft/s.

Correction factors given in the section entitled "Data for Gases other than Propane" should be employed when vessels containing gases other than propane are to be protected.

Where L /D is less than 6, the design data given by Simmonds and Cubbage (Reference 85) may be used, but since this work was carried out on vessels where L /D did not exceed 3, it may lead to the provision of explosion relief with an increased factor of safety as L /D approaches 6.

Principle of Relief Venting for Ducts and Long Vessels.

When a gas is ignited at the center of a long vessel, the products of combustion can first expand freely until the flame reaches the vessel walls. Thereafter, the products of combustion expand in two directions along the length of the vessel. During this period, if the flammable gas is hydrocarbon vapor, the flames travel initially at a speed of about 10-20 ft/s. The expansion of the burnt combustion products behind the flame in the duct causes a motion of the unburnt gas ahead of the flame. After a short time this moving unburnt gas becomes turbulent and one of the consequences of this is that the rate of combustion at the flame front is increased. This process may result in the continued accele'ration of the flame to very high speeds. Shock waves associated with the acceleration of the flame may also give rise to a large increase in .pressure both in front and behind the flame and may also play a vital part in the .eventual transition to a detonating combustion. Under the latter conditions, flame speeds of the order of 6,000 ft/s and pressures of several hundred lbf/in. ~ may be obtained.

APPENDIX C 68--85

If the gas is initially in rapid motion, the initial propagation of flame is also faster and, other conditions being constant, a more violent explosion occurs than when the gas is initially stationary. Increases in the rate of pressure rise may also be caused by obstacles in the duct. These create local pockets of intense turbulence in the moving unburnt gas and may bring about a very rapid increase in the rate of combustion.

As a general principle, relief vents should be sited so that burnt gas close behind a flame is expelled from the vents; this would minimize the effect of the expansion of this gas on the motion .of unburnt gas ahead of the flame. A relief vent should, therefore, be placed wherever there is likely to be a source of ignition. If there is a chance that ignition may occur at any point along the vessel or duct, it follows that relief vents should be installed along the whole length of the duct. It is also necessary that explosion reliefs, particularly those behind and near the flame, should open at a very early stage in the explosion, otherwise high flame speeds and an increased motion in the unburnt gas may quickly result.

In the following sections, information, is provided based on size and the assumption that the gas may become ignited at any point in the duct and that the particular mixture of flammable gas and air is that .which would give the most violent explosion. The information is for propane/air mixtures except where otherwise indicated, although figures for propane/air mixtures will apply with little modification to many other gases and to most industrial flammable solvents. In general, the requirements are given in terms of a relation between the maximum pressure that may be expected during an explosion, the length End diameter of the duct or vessel, and the size and separation of the explosion reliefs used. The size of the vents is usually expressed by a K factor which is the ratio of the cross-sectional area of the duct to the area of the vent. Thus, K = 1 indicates an explosion relief of area equal to the cross- sectional area of the duct and K = 2 an explosion relief equal to half the cross-sectional area. The term duct or elongated vessel applies to any vessel with a ratio of length L to mean hydraulic diameter D > 6; the mean hydraulic diameter being four times the cross-sectional area divided by the perimeter.

The maximum design explosion pressures apply only when an appropriate vent closure is used. Thus, for example, the use of covers heavier than those recommended may result in explosion pressures greater than the maximum design values:

68 -86 EXPLOSION VENTING GUIDE

Size and Spac ing of Explos ion Rel ie fs for Stat ionary Gases or Gases M o v i n g at Speeds of less than 10 ft /s .

Straight unobstructed ducts L/D less than 30. The provision of only one opening as an explosion relief is general ly sufficient if L / D , the length of the vessel to the mean hydraul ic d iamete r of the vessel, is less than 30, but not less than 6. The m a x i m u m pressure will be given by one of the two following formulae :

where K - 1 ; P = 0 . 0 7 L / D [Eq (i)a]

where .K is between 2 and 32; p = 1.8K [Eq (i)b]

For K between 1 and 2 the m a x i m u m pressure may be taken as the " mean of those given by equations (i)a and (i)b, i.e.,

P = 0.035 L / D + 0 .9K

P = m a x i m u m pressure in ibf/ in.2

K = rat io of cross section of duct to area of vent

I f only a single vent is used it should be placed as near as possible to the most likely posit ion of a source of ignition. I f no such position can be ascertained it should be placed as near to the center of the vessel as possible. Equat ions (i)a and (i)b give the m a x i m u m pressure for the most unfavorable relat ive position of the explosion relief and ignition source.

Wi th this vent system, covers weighing 2 l b / f t 2 of vent a rea and held by magnets or springs may be used. A burst ing disc failing at a pressure not higher than half the designed explosion relief pressure given by equations (i)a and (i)b may be used.

Example 1: A react ion vessel is 20 ft long x 2 ft in d iamete r and an explosion may occur dur ing the empty ing or filling of this vessel. W h a t size of explosion relief is required if the m a x i m u m pressure tha t can be allowed is 10 lbf/in.~?

Answer: F r o m equat ion (i)b the value of K corresponding to a m a x i m u m pressure of P of 10 lbf / in . 2 is 5.5; therefore, the cross- sectional area of the vessel divided by the area of the vent = 5.5, giving a vent of d iamete r 10.2 in. Wi th this a burst ing disc designed to fail at a pressure of 5 lbf / in . 2 may be used. This vent must, of course, be pu t in the end of the vessel.

E x a m p l e 2: A flare stack comple te ly open at the top and with water- seal at the bo t tom is 40 ft high and 2 ft in d iameter and discharges hydroca rbon vapors t o the atmosphere. W h a t is the m a x i m u m pressure if a fue l / a i r mixture is ignited in the stack?

A n s w e r : App ly equat ion ( i)a; L / D = 20 gives a m a x i m u m pressure of 1.4 lbf / in . 2

APPENDIX C 6 8 - 8 7

Straight unobstructed ducts L/D greater than 30. For ducts with an L / D rat io greater than 30, it is necessary to provide more than one explo- . sion relief. Even if the L / D rat io is less than 30, a given area of explosion relief is more efficient if it is d is t r ibuted a long the length of the duct. The m a x i m u m distance apa r t at which vents should be placed and the m a x i m u m pressure which would result f rom an explosion depend on the size of the vents and are given in Tab le C-1.

Table C-1 Maximum Distance Between Explosion Reliefs and

- Maximum Pressures for a Long Length of Unobstructed Duct

M a x i m u m . p r e s s u r e f o r

Size of vents Maximum Formulae giving greates't (K factor for distance maximum pJressures spacmg each vent) apart (lbf/in. ~) (lbf/in. 2)

l 60 D 04 L1/D 2.4 2 30 D .06 Ll/D + 0.1 -1.9 4 20 D .07 L1/D + 0.2 1.6 8 15 D .08 L1/D + 0.3 1.5

NOTE: L~ = distance apart of explosion reliefs.

In designing explosion reliefs for long ducts, an open end of a duc t may be regarded as an explosion relief of size K = 1. Fo r this purpose, an. open end m a y be defined as ei ther an end leading without restriction into the open a tmosphere or leading to a vessel which itself is adequa te ly provided with explosion reliefs or leading into a room of 200 times greater volume~than the volume of the duct . I f the ends of a duct are not open, or may be closed some of the time, an explosion relief should be placed near as possible to these ends.

Vent covers weighing not more than 2 l b / f t ~ should be used. T h e y should be held in position by magnets or springs. I f heavier covers cannot be avoided, then explosion reliefs will need to be closer than indicated in Tab le C-1, i f-pressures are to be kept b e l o w 2 lbf / in . 2 I f la ter informat ion on moving gases is followed in this respect, then any error will be in the direct ion of increased safety.

E x a m p l e 3v. I t is necessary to protect a s traight duct 300 ft long and 2 ft in d iameter with explosion reliefs. One end of the duct is open, the other end is often closed or par t ia l ly closed. How many explo-

68 -88 EXPLOSION VENTING GUIDE

sion reliefs are required if the size of each vent is (a) equal to the cross-sectional area of the duct, or (b) ~ of the cross-sectional area of the duct? T h e duct can withstand a m a x i m u m pressure of 1 lbf / in . 2

Answer: The open end of a duct is a vent of size K = 1 and a vent of equal size should be placed at the closed end of the duct . Accord- ing to Tab le C- l , when K = 1 and the m a x i m u m pressure is 1 lbf / in . ~, the dis tance apar t of explosion reliefs should be 50 ft . This would give a total of 5 explosion reliefs in addi t ion to the reliefs at each end. Also, when K = 8, the distance a p a r t of explosion reliefs should be 17.5 ft. This gives 16 openings plus those at each end.

Vessels and ducts containing obstacles. A single obstacle in a duct may increase the m a x i m u m pressure in an explosion. Even for an obstacle blocking only 5 percent of the cross-sectional a rea of a duct , an increase in pressure by a factor of 2 to 3 may be ob ta ined ; for obstacles such as sharp r ight-angled tees or elbows and for orifices or strips blocking about 30 percent of the cross-sectional a rea of the duct , the factor is about 10. There is insufficient informat ion to give deta i led vent ing relationships for obstacles of various kinds. Exper iments have shown, however, that to reduce m a x i m u m pressures in ducts containing an obstacle of the above type to 2 lbf / in . 2, explosion relief equal to the cross section of the duct needs to be sited every 6 diameters . I t is essential that an explosion relief also be p laced near the obstacle. For a long s t raight duct connected to an obstacle, e.g., a tee-piece or orifice, an explosion relief should be placed as close as possible to the obstacle and also at 6 d iameters on either side of the obstacle. Thereaf ter , explosion reliefs should follow as with a s t raight unobst ructed duct . Any bend sharper than a long sweep smooth bend and any obst ruct ion obscur ing more than 5 percent of the cross section of the duc t should be regarded as an obstacle. For obsicacles within these limits, it is still advisable that an explosion relief should be sited near the obstacle. Vent closures weighing not more than 3 l b / fC should be used. T h e y should be held in position by magnets or

• 'springs.

Size and Spac ing of Explos ion Rel ie fs for Gases M o v i n g at Speeds of 10-60 ft /s .

Unobstructed ducts. Vent systems are given in Tab le C-2 which are designed to l imit the m a x i m u m explosion pressure to 2 lbf / in . 2 for explosions in duc t systems containing f lammable mixture moving with velocities up to 60 ft /s . Wi th both systems, covers held by magnets or springs should be used. The m a x i m u m permi t ted

APPENDIX C 6 8 - 8 9

weight of the covers themselves varies with the velocity of the f lammable gas. For gases moving with a velocity of 25 f t /s , the closures should not weigh more than 10 l b / f t 2 of vent a rea and for gases moving with velocity 25-60 ft/s, not more than 5 l b / fC of vent area.

Table C-2 Space and Size of Vents Along Ducts Containing

Moving Gases

L/D (Ratio of distance K (Ratio of cross between consecutive

section of duct vents to hydraulic Duct diameter to area of vent) diameter of duct)

1 12 Up to1 f t6 in . 2 6

1 ft 6 in. to J 1 9 2 ft 6 in. ~ 2 5

Ducts containing obstacles. Ducts conta in ing obstacles require a greater venting ai~ea in the ne ighborhood of the obstacle. I n fo rma- tion is avai lable only for ducts up to 1.5 ft in d iameter , for which there should be a vent equal to the cross-sectional area of the duct on each side. of the obstacle at a distance equal to 3 duct diameters , followed by a fur ther vent for each side spaced at a distance equal to 6 duct diameters . T h e r ema inde r of the duct should be t rea ted as unobst ructed ducts.

T h e weight of the 6 covers nearest to the obstacles should not exceed 3 l b / f t 2.of vent area for gases moving with velocities below 25 ft/s, and 1 ~ l b / f t 2 of vent a rea for gases moving with velocities of 25--60 ft /s .

These covers may be held by magnets or springs.

V e n t Closures.

In the ma jo r i ty of applicat ions, vent closures must be leak-t ight , robust, and designed in such a way that na tu ra l de ter iora t ion and lack of main tenance will not result in an increase in the maxi- m u m pressure ob ta ined dur ing an explosion.

The use of burst ing discs for vent ing explosions in duc t systems is somewhat restricted. These general ly are more suitable for higher explosion pressures and, when used with ducts conta ining obstacles, may considerably increase the explosion pressure. T h e y may, however, be used in s traight and unobstructed ducts ; re levant


data on bursting pressures, mounting, etc., may be found elsewhere (References 58 and 75). They are commercially available either as ready-made units or in the form of materials for fabrica- tion. The majority of disc materials are metals, but graphite is being used where low bursting pressures are required.

If the vents have to withstand a high temperature, asbestos millboard may be used as a bursting material. There are no reliable design data for this material and bursting pressures of a given batch need to be determined experimentally by subjecting a panel to a static test with compressed air. As a rough guide, panels of asbestos millboard 12 in. x 12 in. x 1/~ in. thick fail at a pressure of approximately 1.4 lbf/in. 2 in a static test; the failure pressure is approximately proportional to the thickness and in- versely proportional to the linear dimension. Figure C-1 shows a method of constructing an asbestos millboard closure.

Clamping flanges

in. R ad iu s . . ~ . . . . ~ ~

g in. Radius

Asbestos mill --board panel

@@ @ @ @@

@@ @ @

@@ Figure C,-1. Asbestos milIboard panel closure.

Panels may also be used which are sufficiently strong to withstand rough handling, but which are clamped or retained in such a way that the whole panel is easily pushed out if there is an explosiom It is important that the panel should be light so that the inertia of moving the panel is reduced to a minimum and the panel itself does not become a dangerous missile. The weight of the panels should not exceed the weight given in the guide. A common way

A P P E N D I X C 68-91

of providing a seal is to retain the panel by light friction between 2 surfaces bearing on a strip of the panel about ~ in. thick at the edge. The drawback of this method is that the pressure at which the friction will be oyercome and the panel will fly is un- certain. A more reliable method is to use magnets or springs to clamp the panel at the edge. In this method, the force holding the panel may be controlled according to the strength of the magnets or the springs. For the applications covered in this guide, this force plus the weight of the panel should not normally exceed 30 lb/ft 2 of vent area. If, however, there is a permanent slight positive pressure within the duct, the magnetic force or the force of the springs on the vent cover could be correspondingly increased to avoid leakage and displacement of the cover up to a maximum value of 50 lbf/ft 2, but the weight of the cover must not be increased.

A method of constructing vent closures held by magnets is shown in Figure C-2. Essentially, it consists Of a cover and a magnet assembly. The cover may be constructed from a variety of light and dimensionally stable materials. Fiberboard is one of these. The metal plates matching the magnets may be screwed or riveted to the panel. There is no point in making covers heavier than required for strength and heat insulation purposes. Magnets are located on the periphery of the vent and they are held in position by screws or rivets or any other convenient means. The seal may be obtained by the .use of soft rubber or other suitable materials.

The total force required to dislodge the panel may be affected by the goodness of fit between the steel plates and the magnets and needs to be ascertained by trial. This may be done by loading the cover with weights or pulling it off with a spring balance.

A method of retaining the cover by the use of spring clips is shown in Figure C-3. The spring must be designed in such a way that the restraining force is no longer active after the cover has traveled 3/~ in. This distance is shown on the drawing.

Hinged doors may be used with advantage instead of covers" since, if adequately anchored, they do not become dangerous mis- siles even if they are quite heavy. The weight data given for covers held by springs or magnets also apply to hinged doors. These, however, have to be arranged in such a way that they open to an angle not less than 45 °. In order to obtain a good seal the end op- posite the hinges may be held by magnets or spring clips.

68-92 EXPLOSION VENTING GUIDE . APPENDIX C 68-93

Corner detail of cover showing interior of resinated kraft paper o r ~ a l u m i n i u m honeycomb

Mild steel plates

fWra°°den ~ K%\\\\\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \~ Permanent me ~ ~ ~ ~ ~ magnets

! in. Clearance ----#~,~" 8

in. Clearance

Sectional detail of recessed magnet showing ~ in. allowance for compression of rubber to ensure dust-tight closure

Figure C-2. A duct fitted with an explosion vent cover held in place by magnets. The field strength of the magnets ma~" be varied according to the strength o£ the plant and the pressure within it. (This is an experimental design and may be modified where necessary.)

Adjusted to disengage after cover has moved more than ¾ in. |

_ _ . . ~ T h i s end anchored

Meta) bracket. . ] ~ f / ] anchored to duct / ~1~ " ~ / /

C o v e r - ~ - - ~ /[ DETAIL OF SPRING ~,~ ,~ '~, ~ ,~

Duct

IL/DUCT WITH THE CLOSURE IN PLACE

Figure C-3. Closure held by springs.

Data for Gases Other than Propane.

I t is possible to extrapolate the data given above to ducts containing a f lammable mixture of gases other than propane in air. This extrapolation is based on the fundamental burning velocity. (See also 2-2. 7.4.) A list of burning velocities of some common vapors and gases is given in Tables A-4.2 and C-3. Conversion of the da ta to a gas other than propane is accomplished by the use of equation (ii):

S * P2 = - - Px [Eq (ii)]

2.2 where

Px = the max imum design pressure for propane

P2 = the max imum design pressure for t hegas under consideration, and

S 2 = the fundamental burning velocity of gas in ft/s

68-94 EXPLOSION VENTING GUIDE APPENDIX D 68--95

O n the other hand, if the distance between neighboring vents is L1 for a given m a x i m u m pressure with a p ropane /a i r mixture, then for a different gas the new distance L2, to give the same maxi- m u m pressure, would be given by

2.2 L~ = - - L, [Eq (iii)]

S ~

The m a x i m u m fundamenta l bhrn ing velocity occurs with mixtures near the stoichiometric composition and these velocities are the values which should be used with any ratio of f lammable gas with air. An increase in the temperature of a given gaseous mixture increases the fundamenta l bu rn ing velocity by a factor approximately proportional to the 1.5 power of the absolute temperature. O n this basis, the m a x i m u m pressure should be proport ional approximately to the cube of the absolute temperature.

Equat ions (ii) and (iii) may overestimate the•ventirig required for gas mixtures very much lighter than a p ropane /a i r mixture since they do not take into account the effect of gas density on the inertia and frictional resistance of the gases.

Mixtures of gases with oxygen or air enriched with oxygen can give substantial ly higher max imum pressure dur ing a n explosion, but there are no data to give an estimate of this increase.

Table C,-3 Maximum Fundamental Burning Velocities*

Burning velocity Distance between Gas Mixture " ft/s vents (propane = 1)

Methane/air 1.2 1.5 Propane/air 1.5 1.0 Butane/air 1.3 1.3 Hexane/air 1.3 1.3 Ethylene/air 2.3 0.4

**Town gas/air 3.7 0.16 Acetylene/air 5.8 - - Hydrogen/air 11.0 - -

*Additional information on burning velocities can be found in Perry's Chemkal Engineers' Handbook, 5th ed., McGraw-Hill Book Co., 1973, pp. 9-20.

**Town gas containing 63 percent hydrogen• ~-

/

A p p e n d i x D I l l u s t r a t i ons of V e n t C losures

f ~ £ £ L SA~W

, v'/ :F

2" -fl r--V-T--T

U~Hr ~UL~/Ne ACRO~ l I ] TO KEEP OUT RAIAI I i I ]

CLIPS ON I& ° C £ N 1 ~ ~tl • ~'3t0£30F.~LSH ONLY

E4CH 3451"1 TIED TO ,JAMB ,~ANO MULLION BY TY~

FOOT LON# CHAINS

-1- I

1 I

"'r" MULLIOM WITH TEE REVEE3ED FOR THIS IM37ALLATIOM

3 £ C T I O N " A - A "

U~,,.L',NU ~,~£i.

1 i -j S E C T I O N " B - B *~

I

M U L L I O N C L I P J A M B C L I P

Figure D-1. Method of converting standard steel sash to deflagration venting type.

68-96 EXPLOSION VENTING GUIDE APPENDIX D 68-97 b~

A

~ ~ A u H=r "

.Stock DJ~cl~r#*

//

TYPICAL ,I~ILVERIZED ,.~I"~K PNEMMA~PI,~ HANDLING & 5EPARA~PING ,SY3TEIW .SNOWING EIrPLOJION VEHT3 & P~OT£CTIVE BY-~3 .~ AItRANGEMENT

F--X/~.O-clON VENT WITH DIAlaNRAGiW CUTTER & ¢,UPPORTING GRII.L

(~Gction thru A-A) /

~a D

~d

• %

D

J1 8tiCKET ELEVATOR

PROTECTION

5F•. d I<

"d,#e \

Mr,.=. r~.ce VA}

L__[ II I I/or/re VAt,

DRUM BLENDER ~1~ OTECTIO N

Figure D-2. Venting methods for dust conveyor ducts and separators. Figure D-3. Venting methods for dust blenders and elevators.

68-98 , EXPLOSION VENTING GUIDE APPENDIX D 68--99

• JP ' - I

t . ' i . 0

"I3

r. . . . n

~o

.H I

o

.i

!

~o

-8 O

H et

el l i

"i

o ~

• L o r . ~ r / N ~

--~-~~1% / \ / \ / \ VI-~~. ~'~-~

~4a.~,y yApON /~OOr D I ~ p ~ , e Y 6 ~ ' J _~ .~ Io o_ o o o o Ol

I°FL! ........... -- I! 'oI i1111 Io ~--~-5-~--o:'~

. I I I I I ~Z3L- G # 3 x # r P L R N I ~ . ~ l l t " ~l I~l" l~.~v.q~.exl-r i i ° '~ ' ' , "~R~' ' n"n

~ h ~VA r / o ~ C~OKN31~ Trek DE#R/#. OF D/ *qPH~G~

E x p x o ~ l o N O I A P H ~ R G ~ ; f ' ~ J O L V E N T a e ~ ' c o v E ~ y ~.~X,~T~-M

Figure D-5. Vent ha tches for coating machines; ven t d i aph ragm for solvent recovery system.

LII l l i

68-100 ' EXPLOSION VENTING GUIDE

Appendix E Experimental Data

E-1 Combustion Characteristics of Various Dusts.

E-I.1 The pressures and rates of pressure rise produced in dust deflagrations depend mainly on the fineness of the dust particles, dust concentration, source of ignition, size and shape of the vessel or enclosure, and the uniformity of the dust cloud. The data in Table E-1 were obtained from laboratory experiments with dust passing through a No. 200 U.S. Standard Sieve, at a concentration of 0.50 oz per cu ft as described in U.S. Bureau of Mines g.I . 5624. (25)

Table E-1

Combustion Characteristics of Various Dusts (29)

Type of Dust

Min. Min. Spark Explo- Max.

Ignition Energy sive Explo- Temp. o f for Concen- sion

Dust Ignition, tration, Pres- Cloud Milli- oz per sure, Deg. C joules cu ft psig

Max. Rate of

Pres- Dust sure, Classi-

ps i /sec fi ~ation

20,000+ ST-3

2,200 ST-1

15,000 ST-2 4,900 ST-1

2,400 ST-1 4,400 ST-1

1,700 ST-1 6,000 ST-1

1,700 ST-1 9,500 ST-2

7,500 ST-2 3,600 ST-1 4,300 ST-1 4,700 ST-1 2,000 ST-1 6,000 ST-1

Metal Powders: Aluminum, Flake

A422 610 10 0.045 127 .Iron, Hydrogen

Reduced 320 80 0.120 61 Magnesium, Milled,

Grade B 560 40 0.030 116 Manganese 460 305 0.125 53 Silicon, Milled

(96% Si) 780 960 0.160 87 Tantalum 630 120 <0.200 55 Tin, Atomized (96 %

Sn, 2% Pb) 630 80 0.190 48 Titanium 330 25 0.045 70 Zinc, Condensed

(97% Zn, 2% Pb) 690 960 0.460 50 Zirconium 20 15 0.040 59

Plastics: Allyl alcohol

derivative 510 20 0.035 91 Cellulose acetate 420 15 0.040 85 Cellulose triacetate 430 30 . 0.040 107 Cellulose propionate 460 60 0.025 107 Chlorinated phenol 570 60 0.040 70 Epoxy (No catalyst) 540 15 0.020 94

APPENDIX E 68 -101

Type of Dust

Min. Min. Spark Ex.plo- Max.

Ignition Energy slve Ex.plo- Max. Temp. of for Concert- slon Rate of

Dust Ignition, tration, Pres- Pres- Dust Cloud Milli- oz per sure, sure, Classi-

Deg. C joules c u f t psig psi/sec fication

Plastics: (con't) Lignin resin 450 Methyl methacrylate

resin 480 Nylon 500 Petroleum resin 510 Phenol formaldehyde 580 Polycarbonate 710 Polyurethane foam 550 Polyethylene 450 Polypropylene 420 Polystyrene latex 500 Polyvinyl acetate 550 Rayon 520 Rosin, DK 390 Rubber, Syn. 320 Shellac 400 Urea molding

compound 460

Agricultural Products: Alfalfa 530 Cinnamon 440 Cocoa 510 Coconut shell 470 Coffee 720 Corn 400 Cornstarch 400 Cork 460 Cottonseed meal 540 Garlic, dehydrated 360 Guar seed 500 Gum, Arabic 500 Malt barley 400 Milk, Skimmed 490 Pea flour 560 Peanut hull 460 Peat, Sphagnum 460 Potato starch,

Dex trinated 440

20 0.040 102 5,000 ST-1

20 0.030 84 2,000 ST-1 20 0.030 95 4,000 ST-1 25 0.025 94 4,800 STol 15 0.025 77 3,500 ST-1 25 0.025 96 4,700 ST-1 15 0.025 96 3,700 ST-1 10 0.020 80 7,500 ST°2 30 0.020 76 5,500 ST-1 15 0.020 100 7,000 ST-I

160 0.040 69 1,000 ST-1 240 0.055 107 1,700 ST-1

10 0.015 87 12,000 ST-2 30 0.030 93 3,100 ST-1 10 0.020 73 3,600 ST-1

80 0.085 89 3,600 ST-1

320 0.105 66 1,100 STol 30 0.060 121 3,900 ST-1

100 0.075 68 1,200 ST-1 60 0.035 115 4,200 ST-1

160 0.085 38 150 ST-1 40 0.055 113 6,000 ST-1 40 0.045 106 7,500 ST-2 35 0.035 96 7,500 ST-2 80 0.055 104 2,200 ST-1

240 0.100 57 1,300 ST°I 60 0.040 70 1,200 ST-1

100 0.060 84 1,500 ST-1 35 0.055 95 4,400 ST-1 50 0.050 95 2,300 ST-1 40 0.050 68 1,900 ST-1 50 0.045 116 8,000 ST-2 50 0.045 104 2,200 ST-1

25 0.045 120 8,000 ST-2

(con' O


(Table E-1 con't)

Type of Dust

Min. Min. Spark Explo- Max.

Ignition Energy sive Explo- Max. Temp. of for Concen- sion Rate of

Dust Ignition, tration, Pres- Pres- Dust Cloud MiUi- oz per sure, sure, Classi-

Deg. C joules cu ft psig psi/sec fication

Agricultural Products: (con't.)

Rice 510 100 0.085 47 700 ST-1 Rice, Hull 450 0.05 0.055 109 4,000 ST-1 Soy protein 540 60 0.050 98 6,500 ST-1 Sugar, Powdered 370 30 0.045 109 5,000 ST-1 Wheat flour 440 60 0.050 97 2,800 ST-I Wood flour, Pine 470 40 0.035 113 5,500 ST-2 Yeast, Torula 520 50 0.050 123 3,500 ST-1

Drugs: Aspirin 660 25 0.050 88 10,000-I- ST-2 Nitropyridone 430 35 0.045 111 10,000 + ST-2 Vitamin Bl 360 60 0.035 101 6,000 ST-1

Chemicals: Adipic acid 550 60 0.060 84 2,700 ST-1 Aryl nitroso methyl

amide 490 i5 0.050 142 8,500 ST-2 Benzoic acid 620 20 0.030 76 5,500 • ST-1 Bis-Phenol A 570 15 0.020 89 8,500 ST-2 Sulfur 190 15 0.035 78 4,700 ST-1

APPENDIX F 68--103

A p p e n d i x F B i b l i o g r a p h i c a l R e f e r e n c e s

1. Bartknecht, W., "Application of Explosion Pressure Relief to Protect Apparatus in Industrial Production Facilities," Part II, Loss Prevention, Volume 11 (1977), American Institute of Chemical Engineers.

2. Bartknecht, W., "Bericht ilber Versuche zur Frage der Drucken- tlastung von Grossbeh~iltern und Fabrikationsbauten im Falle von L6sungsmitteldampfexplosionen," Ciba-Geigy Ltd., Cen- tral Safety Service (May 1973) (personal communication).

3. Bartknecht, W., Ciba-Geigy, Basel, Switzerland (personal communication).

4. Bartknecht, W., "Explosion Protection Measures on Fluidized Bed Spray Granulators and Fluidized Bed Driers," Ciba- Geigy, Ltd., Basle Central Safety Service (1975).

5. Bartknecht, W., and Kuhnen, G., Forschungsbericht F 45, Bundesinstitut Fi.ir Arbeitsschutz (1971).

6. Bonyun, M. E., "Protecting Pressure Vessels with Rupture Discs," Chemical and Metallurgical Engineering, Volume 42 (May 1945), pp. 260-263.

7. Brown, Hylton, "Design of Explosion Pressure Vents," Engi- neering News-Record (October 3, 1946).

pp. 261-76. 9. Chappe!l, W. O., "Calculating a Prcssure-Time Diagram for

an Explosion Vented Space," Loss Prevention, Volume 11 (1977), AIChE. Coffee, R. D., "Dust Explosions: An Approach to Protection Design," Fire Technology, Volume 4, No. 2 (May 1968), pp. 81- 87. Coffee, R. D., Raymond, C. L., Crouch, H. W., "A Linear Variable Differential Transformer as a Transducer," Eastman Kodak Company, Kodak Park Division (1950). Cousins, E. W., and Cotton, P: E., "The Protection of Closed Vessels Against Internal Explosions," Ainerican Society of Mechanical Engineers (1951), Paper No. 51-PRI-2. Cousins, E. W., and Cotton, P. E., Chemical Engineering, No. 8 (1951), pp. 133-137. Coward and Hersey, "Accuracy of Manometry of Explosions," Bureau of Mines Report of Investigations 3274 (1935). Coward and Jones, "Limits of Flammability of Gases and Vapors," U.S. Bureau of Mines, Bulletin 503 (1952). Creech, M. D., "Combustion Explosions in Pressure Vessels Protected with Rupture Discs," American Society of Mechan- ical Engineers, Transactions, Volume 63, No. 7.

10.

11.

12.

13.

14.

15.

16.

" i


17. Creech~ M. D., "Study of Combustion Explosion in Pressure Vessels," Black, Syvalls and'Bryson, Inc., Kansas City, MO (February 1940).

18.. Crouch, H. W., Chapman, C. H., Raymond, C. L.,Wischmeyer, F. W., "Maximum Pressures and Rates of Pressure Rise Due to Explosions of Various Solvents," Eastman Kodak Company, Kodak Park Division (1952).

19. Cubbage, P..A., and Marshall, M. R., "Explosion Relief Protection for Industrial Plants of Intermediate Strength," Institution of Chemical Engineers Symposium Series 39 (April 1974).

20. Cubbage, P. A., "Flame Traps for Use with Town. Gas/Air Mixtures," Gas Council Research Communication, GC63, London (1959).

21. Dicken, Clinton O., "Dust Explosion Hazards in the Con- fectionery Industry," National Confectioners' Association, 1 LaSalle Street, Chicago, IL.

22. Donat, Claus, "Application of Explosion Pressure Relief as a Protective Measure for Industrial Plant Equipment," Part I, Loss Prevention, Volume 11 (1977), American Institute of Chemical Engineering.

23. Donat, C., "Release of the Pressure of an Explosion with Rupture Discs and Explosion Valves," paper presented at Achema 73, Frankfurt, West Germany.

24. Donat, C., "Staub-Reinhaltung der Luft" (April 1971), 31 (4) pp. 154-160.

25. Dorsett, Jacobson, Nagy, and Williams, "Laboratory Equip- ment and Test Procedures for Evaluating Explosibility of Dusts," U. S. Bureau of Mines Report of Investigations 5624 (1960).

26. "Dust Explosions in Factories," Health and Safety Executive Series 22 (1975), Her Majesty's Stationery Office, London.

27. Fenning, R. W., "Gaseous Combustion at Medium Pressures," Phil. Trans. Roy. Soc., London, Serial A, Volume 225 (1926).

28. Fiock, E. F., "Measurement of Burning Velocity," High Speed Aerodynamics and Jet Propulsion, Volume IX, pp. 409-438, Ox- ford: University Press (1955).

29. Fire Protection Handbook, Fourteenth Edition, Boston: National Fire Protection Association (1976).

30. Freeston, H. G., Roberts, J. P., and Thomas, A., Proc. Instn. Mech. Engrs. (1956), 170 (24), pp. 811-862.

31. Guide to the Use of Flame Arresters and Explosion Reliefs, Ministry of Labour, New Series No. 34, Her Majesty's Stationery Office, London, (1965).

32. Hajek, J. D., and Ludwig, E. E., Petroleum Engineer, 32, No. 7, C-44 - - C-51 (July 1960).

33. Harris, G. F. P., and Briscoe, P. G. Combustion and Flamel 1 1 (4) (August 1967), pp. 329-338.

APPENDIX F 6 8 - 1 0 5

34. Hartmann, Cooper, and Jacobson, "Recent Studies of" the Explosibility of Corn Starch," U.S. Bureau of Mines Report of Investigation 4725 (1950).

35. Hartmann, Jacobson, and Williams, "Laboratory Explosi- bility Study of American Coals," U.S. Bureau of Mines Report. of Investigations 5052 (1954).

36. Hartmann and Nagy, "Effect of Relief Vents on Reduction of Pressures Developed by Dust Explosions," U.S. Bureau of Mines Report of Investigation 3924 (1946).

37. Hartmann and Nagy, "Inflammability and Explosibility of Powders Used in the Plastics Industry," U.S. Bureau of Mines Report of Investigation 3751 (1944).

38. Hart/nann and Nagy, "The Explosibility of Starch Dust," Chemical Engineering News, Volume 27 (July 18, 1949), p. 2071.

39. Hartmann, Nagy, and Brown, "Inflammability and Explosi- bility of Metal Powders," U.S. Bureau of Mines Report of Investigation 3722 (1943).

40. Hartmann, Nagy, and Jacobson, "Explosive Characteristics of Titanium, Zirconium, Thorium, Uranium and Their Hydrides," U.S. Bureau of Mines Report of Investigation 4835 (1951).

41. Howard, W. B., Loss Prevention, Volume 6, a CEP Technical Manual, Am. Inst., Chem. Engrs., 1972, pp. 68-73.

42. Howard, W. B., and Russell, W. W., "A Procedure for Design- ing Gas Combustion Venting Systems Chemical Process Haz- ards with Specific Reference to Plant Design V," Institution of Chemical Engineers Symposium Series 39, April 1974.

43. Hulsberg, F., Rev. Industr. rain. (1957), 39, pp. 373-376. 44. Jacobson, Cooper, and Nagy, "Explosibility of Metal ~Pow-

ders," U.S. Bureau of Mines Report of Investigations 6516 (1964).

45. Jones, w. M., "Determination of Dust Explosion Possibilities," Factory Insurance Association (1940), Special Hazard Study No. 4.

46. Jones, W. M., "Prevention and Minimizing the Effects of Dust Explosions in Manufacturing Plants," Factory Insurance Association (1940), Special Hazard Study No. 5.

47. Jones, Harris, and Beattie, "Protection of Equipment Con- taining Explosive Acetone-Air Mixtures by the Use of Dia- phragms," U.S. Bureau of Mines Technical Paper 553 (1933).

48. Jost, Wilhelm, "Explosion and Combustion Pressures in Gases" (1946).

49. Kaplan, K., Gabrielsen, B. L., and Van Horn, W. H., "Wall Failures from Explosive Forces," Loss Prevention, Volume 9 (1975), American Institute of Chemical Engineers.

50. Litchfield, E. L., "Minimum Energy Concept and Its Applica- tion to Safety Engineering," U.S. Bureau of Mines Report of Investigations 5671 (1960).


51. Loison, R., Chaineaux, L., and Delclaux, J., "Study of Some Safety Problems in Fire Damp Drainage," Eighth International Conference of Directors of Safety in Mines Research, Paper No. 37, Dortmund-Derne, 1954, Safety in Mines Research Establish-. ment.

52. Mainstone, R. J., Current Paper 26/71, Building Research Station, Garston, Watford WD2 7JR, England (1971).

53. Nagy, J., Conn, J. W., and Verakis, H. C., "Explosion Devel- opment in a Spherical Vessel," U.S. Bureau of Mines Report of Investigations 7279 (1969).

54. Nagy, Cooper, and Stupar, "Pressure Development in Labora- tory Dust Explosions," U.S. Bureau of Mines Report of In- vestigations 6561 (1964).

55. Nagy, Dorsett, and Jacobson, "Preventing Ignition of Dust Dispersions by Inerting," U.S. Bureau of Mines Report of Investigations 6543 (1964).

56. Nagy and Portman, "Explosibility of Coal Dust in an Atmos- phere Containing a Low Percentage of Methane," U.S. Bureau of Mines Report of Investigations 5815 (1961).

57. Nagy, Seiler, Conn, and Verakis, "Explosion Development in Closed Vessels," U.S. Bureau of Mines Report of Investigations 7507 (1971).

58. Nagy, Zeilinger, and Hartmann, "Pressure-Relieving Capaci- ties of Diaphragms and Other Devices for Ventitag Dust Ex- plosions," U.S. Bureau of Mines Report of Investigations 4636 (1950).

59. Palmer, K. N., "Explosion Protection of a Dust Extraction System," Institution of Chemical Engineers Symposium Series 39 (April 1974).

60. Palmer, K. N.,'Fire Research Note No. 830, "Dust Explosion Venting - - A Reassessment of the Data" (August 1970), Fire Research Station, Boreham Wood, Herts., England.

61. Philpott, J. E., Engng., Mater. & Des. (1963), 6 (1) pp. 24-29. 62. Pineau, J., Oiltaire, M., and Dangreaux, J., "Efficacit~ des

events d'explosion," Note No. 881-74-74, Cahiers de Notes Documentaries, No. 74, ler Trimestre 1974.

63. Potter, A. E., "Flame Quenching," Progress in Combustion Science and Technology, Volume I, pp. 145-181, Oxford: Per- gamon Press, 1960.

64. Quinton, P. G., Brit. Chem. Engng. (1962), 7 (12), pp. 914-921. 65. Radier, H. H., J: Inst. Petrol (1939), 25, pp. 37-81. 66. Rasbash, D. J., The Structural Engineer, 47 (10), pp. 404-407

(October 1969). 67. Rasbash, D. J. and Rogowski, Z.W., Combust. & Flame (1960),

4 (4), pp. 301-312.

APPENDIX F 68 - -107

68. Rasbash, D. J., and Rogowski, Z. W., "Relief of Explosions in Dust Systems," Symposium on Chemical Process Hazards with Special Reference to Plant Design, Institution of Chemical Engi- neers, 1961, pp. 58-69.

69. Rasbash, D. J., and Rogowski, Z. W., "Relief of Explosions in Propane/Air Mixtures Moving in a Straight Unobstructed Duct," Second Symposium on Chemical Process Hazards with Special Reference to Plant Design, Institution of Chemical Engi- neers, 1964.

70. Runes, E., Loss Prevention, Volume 6, a CEP Technical Manual, Am. Inst., Chem. Engrs. (1972), pp. 63-67.

71. Simmonds, W. A., and Cubbage, P. A., "The Design of Ex- plosion Reliefs for Industrial Drying Ovens," Symposium on Chemical Process Hazards with Special Reference to Plant Design, Institution of Chemical Engineers (1961), pp. 69-77.

72. Schmidt, H., Haberl, K., and Reckling Hausen, M. K., Tech. Uberwach. (1955), 7 (12), pp. 432-439.

73. Schwab, R. F., and Othmer, D. F., "Dust Explosions," Chem. & Proc. Eng. (April 1964).

74. Smith, J. B.,"Explosion Pressures in Industrial Piping System," Factory Mutual Insurance Association (1949).

75. Stecher, "Fire Prevention and Protection Fundamentals," The Spectator (1953).

76. Stretch, K. L., The Structural Engineer, 47 (10), pp. 408-411 tvactooer l vov).

77. "Symposium on Bursting Discs," Trans. Instn. Chem. Engrs. ( ~ a ) , 31 tz).

78. Technical Manual No. 5-1300, Dept. of the Army, "Structures to Resist the Effects of Accidental Explosions" (1969).

79. Thompson and Cousins, "A Direct-Reading Explosion Effect Gage," Instruments (April 1947).

80. Thompson, N. J., and Cousins, E., "Explosion Tests on Glass Windows: Effect on Glass Breakage of Varying the Rate of Pressure Application," Journal of the American Ceramic Society, Volume 32, No. 10 (October 1949).

81. Thompson and Cousins, "Measuring Pressures of Industrial Explosions," Electronics (November 1947).

82. Tonkin, P. S., and Berlemont, C. F. J., Fire Research Note No. 942, "Dust Explosions in a Large Scale Cyclone Plant" (July 1972), Fire Research Station, Boreham Wood, Herts., England.

83. Underwriters' Laboratories, Inc., "A New Type of Bomb for Investigation of Pressures Developed by Dust Explosions," Bulletin of Research No. 30 (March 1944).

84. Underwriters '• Laboratories, Inc., "The Lower Limit of Flam- mability and the Autogeneous Ignition Temperatures of Certain Common Solvent Vapors Encountered in Ovens," Bulletin of Research No. 43. t,A


85. Underwriters' Laboratories, Inc., "The Spontaneous Ignition and Dust Explosion Hazards of Certain Soybean Products," Bulletin No. 47.

86. Valentine and Merrill, "Dust Control in the Plastics Industry," Transactions of the American Institute of Chemical Engineers, Volume 38, No. 4 (August 1942).

87. VDI Richtlinien, Druckentlastung yon Staubexplosionen, VDI 3673 Draft, June 1975, published by Verein Deutscher Engenieure.

88. Yao, "Explosion Venting of Low Strength Equipment and Structures," Technical Paper Presented at AIChE Loss Pre- vention Symposium, Philadelphia, PA (November 1973), FMRC, Norwood, MA.

89. Yao, deRis, Bajpai, and Buckley, "Evaluation of Protection from Explosion Overpressure in AEC Gloveboxes," for U.S.A.E.C., Chicago-Operations Office, Argonne, IL. AEC Contract AT (11-1 ) 1393, F.M.R.C. Serial No. 16215.1, Factory Mutual Research Corp. (December 1969).

90. Zabetakis, Michael D., "Flammability Characteristics of Com- bustible Gases and Vapors," U.S. Bureau of Mines Bulletin 627 (1965).

NFPA Publications. Many standards, fire reports, and other publications containing

information relating to various phases of explosion prevention and protection have been published by the NFPA. A complete list of publications will be mailed on application to the National Fire Protection Association, 470 Atlantic Avenue, Boston, M A 02210.

Report of the Committee on Explosion Protection … COMMITTEE ON EXPLOSION PROTECTION SYSTEMS Report...

Documents

Transcript of Report of the Committee on Explosion Protection … COMMITTEE ON EXPLOSION PROTECTION SYSTEMS Report...