EPA-600/2-84-095
May 1984
EVALUATION OF THE EFFICIENCY OF
INDUSTR.'AL FlARES:
TEST RESULTS
By
J. H. Pohl, R. Payne, and J. Lee
ENERGY AND ENV«R0fl.1£NTAL RESEARCH CORPORATION
18 Mason
Irvine, California 92714
EPA Contract Ho. 58-02-3601
EPA Project Officer: Bruce A. Tichenor
Industrial Processes Branch
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina 27711
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFF ICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
-------
________
EPA-600/2-84-095
TECHNICAL BEKJRT DATA
fHfW rtad Imtmtiiva or iht nr.tn; krfort eompliKngf
«. TITLE AHDSU*TITI.I
Evaluation of the Efficiency of Industrial Flares:
Test Results
1. fttCI'ltNT-S ACCttSIO'rNO.
PBjt 1199 57 1
§ pupoat o+ri
May 1984
». fEHFOMMlNO duGAWIIATION CODi~
V AUTmOAH)
J.H. Pohl, R. Payne, and J. Lee
t. PERFORMING OROANIZATlO* SAME AND AOORCSS
Energy and Environmental Research Corporation
18 M&Bon
Irvine, California 92714
u sponsoring agency n*mi ano adorus
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research fTziangle Park, NC 27711
I. PERFORMING ORGANIZATION REPORT NO.
10. PR< GRAM ILiMENfltH-
BeoMfSACT/dhANt liar
68-02-3661
13. TVFt OF REPORT AND PERIOD CCVCREO
Final; 10/80 - 2/84
1«. SPONSORING AGENCY COOC
EPA/600/13
"is. surplemenJa^y notes 1ERL-RTP project officer is Bruce A. Tichenor, Mail Drop 54;
919/541-2991. EPA-600/2-83-070 is an earlier related (Phases 1 and 2) report.
t«. AWTiAGr ^rhe report gives results of Phases 3 and 4 of a four-phase research pro-
gram to quantify emissions from, and efficiencies of, industrial flares. Phase 1 con-
sisted of the experimental design; Phase 2. the design of the test facilities; Phase 3,
development of the test facilities; and Phase 4, data collection and analysis. The
combustion efficiency of large pilot-scale flare; was measured. The flame structure
and combustion efficiencies were correlated with operating conditions of the flare,
the size of the flare head, and properties of the flared gases. The combustion effi-
ciency was correlated with th° ratio of heating value of the gas flared to the heating
value required to maintain a stable flame, aid was independent of the flame Lead
size. In turn, the heating value required tc maintain a .stable flame was correlated
with the reciprocal of an estimated flame temperature based on properties of the
flared gas. The length of the flame, entrainment into the flame, and liftoff distances
were also correlated, using combinations of the Richardson Number, jet theory,
and properties of the flared gas.
KEY »VO«OS «NO DOCUMENT ANALYSIS
OESCRIPTORS
Pollution
Exhaust Gases
Efficiency
Flames
Measurement
Surveys
Analyzing
>» OlSTHIBUTlON STATEMENT
Release to Public
b.tOENTIFIERS/OPf N ENDED TERM!
Pollution Control
Stationary Sources
Industrial Flares
19 SECURITY CLASS (Thu Report/
Unclassified
20 SECURITY Ch.AS9 (THtl ptgtj
Unclassified
COS ati Fickl/Gfotip
13B
21B
14G
14B
J1 NO. or PACES
204
M. PR'Ct
IP* P«rm IJJ0 1 IS-71J
-------
NOTICE
This document has been reviewed in accordance with
U.S. Environmental Protection Agency policy and
approved for publication. Mention of trade names
or commercial products does not constitute endorse-
ment or recomendation for use.
ii
-------
ABSTRACT
The U.S. Environmental Protection Agency has contracted with Energy and
Environmental Research Corporation to conduct a research program which will
result in the quantification of emissions from, and efficiencies of, in-
dustrial flares. The study is being conducted in four phases:
I - Experimental Design
II - Design of Test Facilities
III - Development of Test Facilities
IV - Data Collection and Analysis
EPA Report No. 600/2-03-070 provides the results of Phases I and II uf
the study; the results of Phases III and IV are reported herein.
Measurements were made of the combustion efficiency of large pilot-scale
flares. The Mane structure and combustion efficiencies were correlated
with operating conditions of the flare, size of the fl*re heac, and properties
of the flared gases, The combustion efficiency was correlated with the ratio
of heating value of the gas flared to the heating value require-d to maintain
a stable flame, and wns Independent of the fiara head size. In turn, the
heating value required to maintain a stable fiame was correlated with the
recipracal of an estimated flame tarperature based on properties of the
flared gas. Other correlations for the length o* the flame, entrainuient into
the flame, and "liftoff distances wtre developed using ccnbinatiors of tha
Richardson Number, Jet theory, and the propertfes c1 the flared cas.
iii
-------
TABLE OF CONTENTS
Section Page
List of Figures
List of Tables
1 0 INTRODUCTION 1-1
1.1 Flare Flames . 1-1
.Blount Gases F .ired 1n the United States ... 1-1
1.3 Emissions fron FLvcv 1-2
1.4 Approach of This Study 1-5
2.0 CONCLUSIONS 2-1
2.1 Technical Suifl"iary ?-l
3.C TEST METHODOLOGY 3-1
3.1 Fibre Te^t Facilities (FTF) 3-1
3.2 Test Procedure 3-6
3.3 Data Reduction 3-13
4.0 "LAME STRUCTURE 4-1
4.1 Test. Conditions 4-2
4.2 Mechanism of Combustim 4> 6
1.3 Flame Length 4-13
4.4 Entralninent 4-20
4.5 Flaine Liftoff 1-23
4.6 Flame Stability 4-26
5.0 COMBUSTION EFFICIENCY OF FLARE FLAME 5-1
5.1 Study Conditions 5-1
5.2 Comparison with Other Studies 5-1
5.3 Correlations 5-3
5.4 Commercial Flare Heeds 5-7
5.5 High Velocity 5-7
5.6 Steam 5-7
v
-------
TABLE OF CONTENTS
(Continued)
Section Pas:
APPENDIX A: FPA Fla\-e Test Facility ?t EER Vi
APPENDIX B: Testing Methodology B-l
APPENDIX C: Data Summary Table C-1
APPENDIX D: Integrated Combustion Efficiencies D-l
APPENDIX E: Examples of Test Data E-l
APPENDIX F: braphs of Local Combustion Efficiencies
and Dilution Factors F-l
APPENDIX G: Quality Assurance G-l
VI
-------
LIST OF FIGURES
Figure No. Page
2-1 EPA flare test facility at EErt 2-2
2-2 PiRgram ot' sampling locations 2-11
2-? Region of flame Instability 2-12
2-4 Combustion efficiency near the limits of flame
stability 2-14
3-1 Design of retention ring for 3-Inch flare heat! 3-3
3-2 Flow patterns determined by smoke tests 3-5
3-3 Influence of hood position on closure of a carbon balance
using mass fluxes measured 1n the hood 3-8
3-4 Influence of hood position on combustion efficiency
measured in the hood 3-9
3-5 Verification of carbon mass balance for hood measurements. 3-11
3-6 Comparison of SOj and COg used as tracers 3-12
4-1 Comparison of the range of conditions tested in this
study and these of commercial flare flames 4-8
4-2 Correlation of lengths of flames studied with Reynolds
number 4-1C
4-3 Surface rate of combustior of flars flames 4-'ll
4-4 Volumetric combustion intensity of flare flames 4-12
4-5 Comparison of flame lengths predicted by available
studies with those observed in this study. 4-16
4-6 Comparison of a Hottel and Hawthorne correlation with
the flame length of this study 4-17
4-7 Comparison of flame lengths predicted by a buoyance
correlation with observed flame lengths of this study. . . 4-18
4-8 Correlation of flame lengths in this study 4-19
4-9 Emcirical correction for radiation loss from flare flames. 4-21
4-10 Entrapment of air into flare flames 4-22
4-11 Correlation of entrainment rate for flare flames 4-24
vti
-------
LIST OF FIGURES
(Continued)
Figure No. Page
4-1Z Crrrelation of en trainmen*, rate for flare fl aires A— 25
4-13 Liftoff of flare fiames without pilot flamas 4-Z7
4-14 Cor^lation of liftoff distances for flames without
pi'ot flames 4-26
4-15 Region of flame instability 4-30
4-16 Flame speeds of hydrocarbons 4-31
4-17 Limits of flame stability as a function of estimated
temperature 4-32
5-1 Conditions to obtain 98 percent combustion efficiency
for mixtures of propane and nitrogen 5-4
5-2 Coirbustion efficiency near the limits of flame
stability 5-6
5-3 Influence of steam on combustion efficiency 5-8
V 1 11
-------
LIST OF TABLES
Table No. Page
1-1 Estimate of the amount of .jas flared In the United States
in 1980 (Ref. 1-3) 1-3
1-2 Combustion efficiency of flare flames measured in
previous studies 1-4
2-1 Combustion efficiencies of EER three-inch flare head .... 2-4
2-2 Combustion efficiency of thrse-inch EER flare head
at hich velocities . . 2-5
2-3 Combustion efficiencies of EER six-Inch flare head 2-6
2-4 Combustion efficiencies of EER twelve-Inch flare
head 2-7
2-5 Combustion efficiencies of twelve-fnch industrial
flare head #A 2-8
2-6 Combustion efficiency of twelve-incn Industrial
flare head #B 2-9
2-7 Combustion efficiencies of twelve-inch industrial
flare head PC 2-10
4-1 Flare test conditions 4-4
4-2 Matrix for testing commercial flare heeds 4-5
4-3 Test renditions for high-velocity flares 4-7
4-4 Conditions under which correlations for flames w=re derived 4-15
1*
-------
1.0 INTRODUCTION^1)
Flares are used to safely destroy industrial gases when: C) the
heating value cannot be recovered economically because of Intermittent and
uncertain flow, or (2) process upsets occur. L^rge amounts of gases are
flared in the United States. However, even the approximate amount of these
gases is difficult to establish, because most of the gas is released as
leaks, purges and emergency vents. Surh releases are poorly measured and
reported. Also, the emissions of incompletely burned hydrocarbons from
flare flames are not know accur?telj, because the amount nf material flared
1s uncertain and the combustion efficiency cf the flare flames is unknown and
difficult to measure. This program has measured the combustion efficiency
of pilot-size flares, whose size anu operating conditions have be»n scaled
by several different techniques.
1.1 Industrial Flares
The size, use, operating conditions, and geometry of coiruierdal flares
are diverse. Flares can be enclosed, gas-assisted, air-assisted, pressure,
cr pipe (without or without steam assist) (1.1 - 1.3). They can be used to
destroy relatively constant purges and leaks of gaies and eirergeirv or inter-
mittent planned releases of large amounts of gases. The size of commercial
flareb can vary from 1-1/2 inches in ciiameter to over 1? feet; operating
conditions can "Include velocities as low as 0.005 ft/sec and as high as
sonic; to suppress soot, steam is added at rates of from 0- to ono-pound-of-
steam-per-pound-of-fuel, water from 0- tr. ?-pounds-of-water-per-pound-of-
fuel, or air from C- to 6 -pounds-of-air-per-pound-of-tuel (1.1). This
study was limited tc measuring the combustion efficiencies cf qlpe flares
burning pronane-nHrogen mixtures at steady operating conditions viith and without
steam Injection, in the absence cf wind.
^H'ne project's initial report (1.4), "Evaluation of tVip Ffficiency of In-
dustrial Flares: Bcckground-ExDerlmental Heslcn Facility" {'EPA No. 60C/2-83-
070) provides the basis for the research reported herein. The present Intro-
duction provides only a brief summary of the earlier report.
1-1
-------
1.2 Amount of Gases Flared in the United States
The amount of gases flared in 1he United States is uncertain. A partial
estimate of the amount of gas flared in industries by Klett aid Galeski (1576)
(1.3) Indicated that 1? million tons of gases per year we"e flared 1n the
United States in 1974. The initial report of this study expanded and updated
this estimate and concluded that approximately 16 millior tons of gas were
flared per year 1n 1980 from the industries show? in Table 1-1 (1.4). Ca^ed
on heating value, most gas is flared in the petroleum production and refining
industries. However, based on the amount cf gas, more gas 1s flared f»om
blast furnaces than any other industrial segment.
1.3 Emissions From Flares
Direct field measurement of incompletely burned hydrocarbons from
Industrial flare flames are unavailable and unlikely to become available for
a number ot reasons. The flare stacks are high above the ground to protect
mater'als and personnel from the intense radiation of tiie flames. In
addition, the flames are very large ?nd 1n constant motion. These factors
make probing the olivine of a commercial flare flame, even at a single location,
extremely difficult. A remote sensing system may in the future help indicate
the emissions from flare *lames, but considerably work must be done before
such systems are available (1.5 - 1.9).
In the absence of direct measurements of emissions on operating flare
flames, measurements have beer made on pilot-scale and small commercial flare
heads. The hearis studied have ranged from a 1/2-inch jet to a 27-inch cornier-
i )
cial flare heaC' installed on a slip-stream of a refinery. Table 1-2 shows the
ranee of flare heads, the operating conditions, and the combustion efficien-
cies measured on these heads. The studies indicated that the combustion
efficiency of flares can be very high. However, the studies also
showed that the combustion efficiency of flares c.an be low under son>e opera*1nc
conditions.
The va^jes of combustion efficiency measured in the previous stupes were
somewhat uncertain. The uncertainties were attributed to:
(l)opening of a cone on an FS-6 Coanda flare head.
-------
TABLE 1-1. tSTIMATE OF THE AMOUNT OF GAS FLARED
IN THE UNITED STATES IN 1980 (Ref. 1.4)
INDUSTRY
PRODUCTION
OF PRODUCT
106 TONS/YR
PERCENT
FLARED
AMOUNT FLARED
10C TONS/YR
105 MBtu/YR
Refineries
1048
0.2
2.1
103
Petroleum
Production
584
0.5
2.9
116
/a |
Slast Furnaces '
146
6.6
9.6
69
Coke Ovens^
55
0.4
0.2
11
Chemical^)
Industry
60
2.0
1 .2
59
TOTAL
16.0
358
(a)
Combustible gases only
1-3
-------
TABLE 1-2. COMBUSTION EFFICIENCY OK FLARE FLANFS MEASURED IN PREVIOUS STUDIED
STUDY
DATE
FLARE
SIZE
(IN)
DESIGN
VELOCITY
(FT/SEC)
GAS FlAkED
MEASURED
EFFTCTFNrv
(X)
Palmer (1.10)
1972
0.5
Experimental Nozzle
50 - 250
Ethylene
> 97.8
Lee & Whipple (1.11)
1S81
2.0
Holes in ?" Cap
1.8
Propane
96 - 100
Siegel (1.12)
1980
27^
Commercial Flaregas
Coanda FS-6
.J - 16
Refinery Gas^
97 - > 99
Howes, et fll (1.8)
1981
6.(t)
Commercial Air
Assi?t. Zink LH
40 - 60
Propane
92 - 100
Howes, et al (1.8)
1981
3at
Commerc i a 1 H.P.Zink
LRG0
Near Sonic
(estimate)
Natural Gas
> 99
McDaniel (1.13)
1983
8
Commercial Zink
STF-S-8
0.03 - 62
Propy1ene/Ni trogei ^ ^
67 - 100
McDaniel (1.13)
1983
6
Commercial Air
Assist.
Zink STF-LH-457-5
1.4 - 218
Propyl ene/Ni troger/e^
55 - 100
^503! hydroyen plus 'light hydrocarbons
^Three Spider*:, each with an open area of 1.3 1n^
(c) 2 2
Supplied through spiders; high Btu gas through area of 5.30 in aid low Btu gas through 11.24 in
'^Heating valii? was varied fro^n 209 to 2183 Btu/scf
^heating value was varied from 83 to 2183 Btu/scf
^See footnote O), p. 1-2
-------
• Inability to close a material balance.
• Soot concentrations were rarely measured,
• Samples were typically taken only on the axis.
• Difficulties were caused by the intermittent nature of the
flare f'.ame.
,4 Approach of This Study
The approach used in this study was designed to eliminate or minimize
many of the uncertainties of previous studies.
• The closure of a material balance was verified using a hood to
capture the entire flare plume for small flames and by using SO?
as a tracer for large flames.
• Soot concentration was measured for all tests.
• The average concentration of incompletely burned combustion
species from the flare flame was determined for the entire
plune captured by a hood for small flanges, ar.d samples were
simultaneously measured at five radial positions using the
rake probe for large flames. These values were combined with
velocities calculated from jet theory to estimate the global
combustion efficiencies of each flame.
9 The intermi'tto.icy of the flare flames were accounted for by
mixing a sample taken over a period sufficient to average flame
fluctuations (20 minutes).
The experimental test matrix was designed to determine the validity of
several cannon scaling procedures for three-, six-, and twelve-inch fl?rc
heads. The scaling criteria were constant exit velocity, instant residence
time, constant Reynolds number, and consent kichardson's number.
Throughout the program, advice and consultation was sought frum a
Technic?! Advisory Coruiittee. The committee attended meetings throughout
1-5
-------
the program to review and criticize test plans, ensure the relevance of
the study and facilitate efficient technology transfer. The Advisory
Committee consisted of:
Zahir Bozai
Peabody Engineering
B. C. Davis
Central Engineering D1v.
Exxon Chemical Co.
John J. DjbnowskiH)
Exxon R&E
Leslie Evans
Chemical & Petroleum
Process Branch
Office of Air Quality
Planning & Standards
U.S. Environmental Protection
Agency
E. Doyle Fowler
Union Carbide
Allan Goodley
California Air Resources Board
Keith Herbert
McGi11, Inc.
James G. Seebold
Chevron USA
David Shore
Flaregas Corporation
Larry Thurston
Dow Chemical Company
J. R. Venable
Rohm & Haas Texas, Incorporated
W. G. Hudson
Engineering & Technology
Services Division
Umon Carbide Corporation
Mike Keller
John Zink Company
R. W. ladd
Getty Refining 4 Marketing Co.
^'Replaced in December 1983
by John Wang, Exxon R&E
-------
REFERENCES
SECTION 1
1.1 Vanderlinde, L. G., "Properly Designed Mecnanical Equipment Can Make
Flares Smokeless,'1 in Environmental Management Handboo* for the Hydro-
carbon Processinq Industries, Gulf Publishing Co., Houston, Texas, 1980,
pp. 132-137.
1.2 Brzustowski, T. A., "Flaring In the Energy Industry," Progress 1n Energy
andCombustion Science, Vol. 2, 1976, pp. 12S-161.
1.?. Kleit. M. G., and J. B. Galeskl, Flare Systems Study. EPA Report No.
600/2-75-079 NTIS Report No. PB-zTfiSM. March 1976.
1.4 Joseph, D., J. Lee, C. McKii n...T, R. Payne and J. Pohl, "Evaluation cf
the Efficiency of Industrial Flares: Background - Experimental Design -
Facility," EPA Report No. 6G0/2-83-070, August 1983.
1.5 Herget, W. F., Air Pollution: Ground-Based Sensing of Source Emissions
in Fourier Transform Infrared Spectroscopy, J. R. Fenaro anc L. J.
Ei'iile, Eds" Scadamic Press, 1979.
1.6 Herget, N. F., and J. D. Brasher, "Renote Fourier Transform Infrarpd Air
Pollution Studies," Optical Engineering, 19, No. 4, 1980, pp. 50P-5I4.
1.7 Herget, W. F. , and J. D. 8rasher, "Remote Measurenents of Gaseous
Pollution Concentrations Usinq a Mobile FTIR System," Aoplied Optics, 18,
October 1979, pp. 3402-3620.
1.8 Howes, J. E-, T. E. Hill, R. N. Smith, G. R. Ward and W. R. Hergst,
"Development of Flare Emission Measurement Methodology, PraJt Report,"
EPA Contract No. 68-02-7.682, 1981.
1.9 Armstrong, J. A., "Tethered Balloon ScimplInu of a Pilct-Sca^e Innustrial
Flare Plume," FPA and Denver Research Institute Cooperaiv.T. Agreement No.
CP-807504-01, 1983.
1.10 Palmer, P. A., "A Tracer Technique for Determining Efficiency or ir(
Elevated ^are," E. I. DuPont de Nemours and Co., Wilmington, ^E, '972.
1.11 Lee, K. C., and G. M. Whipple, "Waste Gas Hydrocarbon Combustion ir a
Flare,'' Union Carbide Corporation, South Charleston. WV, 1981.
1.12 Siegel, K. D., "Decree of Conversion of Flare Gas in Re 'inery HIgs
Flares," Ph.D. Dissertation, University of Karlsruhe (Gemany),
February 1980.
1.13 McDaniel, M.. "Flare Efficiency Study," EPA Report No. 600/2-83-05Z,
July 1983.
1-7
-------
2.0 CONCLUSION
2.1 Technical Summary
The EPA FI are Test Facility (FTF) was co. strucced at Energy and Environ-
mental Research (EER) Corporation'; LI Toro test site. The FTF (Figure 2-1)
includes a pad and structure for installation and testing of flare heads,
screens to shield the flame from wind, parallel delivery systems to accurately
meter the wide range of ga.< f'ows to the flare, a hood to sample the entire
plume, a movable rake probe to simultaneously sample five radial positions,
high-speed movie and photographic equipment to record the structure of the flare
flame, and a room *rom which t-o contro'. the flare and analyze gas samples.
Techniques were developed to operate, sample, analyze and reduce the
data. Analysis includes visual and photographic observation of the flare
flame structure, and samples of soot, 0?, CO, CC2, total hydrocarbon, jnd SOg,
which was used as a tracer. The data is corrected for the measured background
of combustion species and for diction of the flare plume by ambient air.
Dilution and local combustion efficiencies are calcinated at each probe
position and the maximum potential error in the dilution and combustion
efficiencies ere estimated for each data point The local combustion efficien-
cies are integrated uMny velocity profile-:, estimated by jet theory to yield a
global combustion efficiency for each Hare flame.
The combustion efficiencies were measured for ; wide range of operating
conditions typical of comnercial flares:
• Head type
3-inch EER prototype
6-inch EER prototype
12-inch EER prototype
3-inch EER prototype with convergert ring
3-inch EER prototype witn divergent ring
12-inch Manufacturer A
?-
-------
( PKOPAHI \ /7A !
I 'AHK r y> [
nu
OHTBOLS
NITERS
VAPORISH
CITY
NAllJI'Al
GAS
(HITROGEN^ f/\ .
__H*L_y \cJ
VAI'UttlZIR
TLOU
CONTROLS
& MtlLRS
TlOU
CONTROLS
k hi>£US
VAI'OHHFR
J"
—-I C(J
cm r
CUM HOLS 1
} » METERS
i HOIIC11 ON
rAN
z
00
l»OOD
PflOiV
f
If too
RfcKF PROBE
r' 1 i —r
r
6
V-
/
\4
A
A
flareheao
tLlCIRItALUf-
HfATEO LlWS
, tedlar !
. BAGS
;
- i
* SAHI'Ll
.J l
i
i
f
SAMPLING
LONIKCl
PANEL
Tnrnmmr"
Oy, CO. CO7
HL, SO^
I Mi. AH
HAGS
Figure 2-1. IPA r'tar? test facility at EER.
-------
l2-1nch Manufacturer B
12-Inch Manufacturer C
' Gas
Propane-nitrogen mixtures
Natural gas
• Heating value of the flared gas from 270 - 2350 Btu/ft^
• Flow rates
Velocities from 0.2 - 428 ft/sec
Reynolds Numbers from 337 to 217,000
Richardson Numbers from 2.9 x 10*5 to 8 x '0^
Tables 2-1 through 2-7 contain a sumnary of the test conditions and the
combustion efficiencies. Sailing positions of the hood and rake
probe are shown in Figure 2-2. In addition to the combustion efficiencies,
other calculations and correlations were m?de:
• Combustion intensity was found to be 90,000 Btu/hr/ft-,
independent of flare or flame conditions.
• The flame length was correlated with the R'chfirdscn runber.
• The liftoff distance wes correlated w-th ratios of velocities
and concentrations of combustible gas.
• The flame stability wes correlate 'ith the reciprocal of ar:
estimated flame temperature.
• The combustion efficiency wni correlated with a dimensianless heating
value of the gas fired.
The term "flame stability" simply meant. that a flame is maintained;
flame instability occurs when the jet velocity exceeds the flame velocity and
the flame goes out Figure 2-3 shows the gas heating value versus the gas
exit velocity at the point of instability (i.e., at the point where the feme
starts to "go out.") This point is determined by establishing a propane-
nitrogen flame at a given velocity and then decreasing the flew of propane
until the flaine goes out. The propanf flow ther. Increased slightly, ard
the combustion efficiency is measured at the conditions just prlsr to the point
where the flama went out. 2-3
-------
TABLE 2-1. COMBUSTION EFFICIENCIES OF EER THREE-INCH FlARE HEAD
urn mm »• w/i - 11"
Ui
u
iMl
u
illrtfM
Ul
('«•»
V(U»
(Ik SIM
/»* (*•>}
U
*!*•!
Hy4ror«r*w>
bmrmcc^m
Ifftclency
fs»
91 Int
IMI
>umi»
(«M
ItlK'lty
1 III
»•<•* It*
h>4i'nf
Mtw
«tnd
3»J
Him
lX
(III
or'
Ul«r
M*
?**!«
l-Mi
N4*v4
Mi
Pmi
(I'M
l«t<
c*r»
fitj
's;
** FpitmI lf9
IW
Mo
0.5
».% ^
r 1 J1I
95.1
9 9
(ON
4
»
HI
H
10
99 II
w a;
ODJ
0.5
a i
W
«./
c e
l<»
4
D
t*HM
(H
¦
9
9* n
99.M
90S
?.n
T.O
I 'l|
M.I
9.9
IfM
1
0
»*•'
*ti
fc
4
91S4
w.iv
am
2.0
r o
nu
SJ.9
bO
I OH
I
a
WttOP
In
9
Ml
M.tt
toa.lM
)Lfi
!J,B
i.jZ.2
.KJ
fxfl
H
If
Uli*
>•4 I
|
IS
p. t§
99 ?S
l« Ku
to
*•
E S
0.1
?J3
11.(
• 0
91
4
0
Re
II
a
90.11
*1. W
».o
*70
75 I
9.«
1*ll»
*»v
N
9
hi;
ve 9*,
?.l
?. 1
w»
11.0
9.0
0 J
3.9
9 J
He
1
S
f 99
w.<>
(a)
(t>)
(c)
Basej on open area of flare head
Based on velocity of pipe
¦tsfer to "igure Z-2
-------
TABLE 2-2. CCMBUST'ON EFFICIENCY OF THREE-INCH EER FLARE HEAD AT HIGH
VELOCITIES
mam ilR r tmjt
1*1
"ir
SUM
B«t(7
{It-itCM
'U f*4\)
¦nmn Mltim
•
Iteuerf
IHUU
fr«oo4
U)
fnl
11m
At t)
Hvdruiaitrm
ftMTHM fir
1 ffilllNX.)
(~)
Kr( /M il ImI
lest
(Im
hlnlfai
•l«9
Actoit
£¦«
VtlftcUtf
(ft/Mc)
!*3ikWl
till
9*!«ltr
(rt/w«c>
lw
M4«tU«
V«lM
(iU/M1)
Mat
SpM*
IMI
lift
Bff
lb.)
C*l«r
Ma
IIm
irri-
dwi
(II
(I* Kit. RIM
11
fa
1" 1
m
lilt
si.e
on U)
4N &
Ml ?
la.
t' 4
QOli
e i
n
74
Llr/6r
m
tl
M
I«1 Ui
11
26.6
•i;
0 IH
e-i
u.*
l-K)
IrMft
to
•
26
»i.2/
s?
?46. 1
99.)
lie;
41 1
1)1)1
0-1
)i
24
OrMpi
u
1
»
M.J3
ff 4*
Hi# M,
i^'A
A3
46 rr
144.5
57.#
r>se
1MB
6. iJi
6-1
N 1
!i
6r**f«
m
t
)1
••.67
IM< 00
93
C«e v.<"
M.;
»?
?*•
«f.»
-•. t«
«-C
4*1
;j
Br«»|«
Wm
6
)l
H.N
M.ta
94
*.;?
1)1.4
J».l
}*0
MB.b
a.eti
6-3
H.l
f«l/6r
(IttN
I
)/
M.U
W
Mm Itelt
It
46. J1
114.1
M.I
m
j].;
9N/
cm
tl
IMI
1*1 Im
h
•
M
•» 67
MW
*
-M.I
12*6
¦043
M.I
6.6M
8-1.*
»
it-n
lat/fcr
•d
¦
41
«-«!>
«) 91
Mlgh-»rt.
W Mrrnt t fa
Cm.
« B
K.8
MJ
1IU
M?
• . 142
IS
ra
6
1
74
m tf
9W
1/2.3
f9.1
1131
M.2
• 671 I
0 2
N
?
lb
1
34
•6.67
MM
Htp-ftl. llrtN
Mmi
M
COBV.
*m 7
:m. i
10? 1
41.)
• ©11
0 1
M.S
NM
o.«w
As
•
JS
MM
W */
SO Print IJij
Wj
C«H.
44 K
Ml 1
IU /
IMf
N.I
6.02s
0 1
M
?4-»
flUWi
fe
1
17
ff.64
« Ai
It Piruit IJfj
..-M
Com.
% n
Bft k
ft >
mo;
ni
o.etl
II
n
4
Or%*t»
He
«
2*
wn
n »
ItH
im 1
/• *
160ft
(t.tt
IS.
».4
ir
v«V(k-
HuN
f
M
TV M
M M
19)
2U1.1
J
im
'6 4
0 OK
0*±4
V
<
rtllow
N»
•
D
MM
MM
Nig* -fcl, SUfcl*
:•-« 1 iall
104
C«*v.
u.s
1/i.i
i/J.I
Vt
M ?
B U*J
0-1%
Tf.ft
*
N»
1
W
9^ fl
M 91
I «!#¦»««.
1 IM Mrr#*4 C-"t
lift
C«H.
M. M
24).4
iu.i
mso
IBB. 6
0 0*9
U-4
40
»?
a
1?
M.N
99 K
(a) Bimu or open ar^a uf flare head
(b) Based on velocity of pipe
fc) R«fcr to F'.gur* 2-2
(d) Dlv. refers to the rt;ention Hp»y of Flyure 3-1 (#>side (JOwti
(*j Con*. refers to the reltntion ring of Figure M In the position shiwn
-------
TABLC 2-3. COMBUSTION FFF1CIFNCIFS OF hFR SIX-TNf.H FLARF HFAD
IMIW: iU «'f- I-™
u;
<1
r«l
tit
lllnpH
41)
Sim
totfe
ilMU*
/it.r«n
*tcrr«tl«n
ImU
IfctM
I* lata
W tkai
(O
11*1
W»4nui%»'
iHMHlv
MMclvwy
KrpCM of i
Itri
M.3
TO
w
»
e
Nil*
T.J
•
IF
*4?
*»>.
n
10 1
in i
IM<
M.O
B 9
¦A j
9
Vt'lOT
if
ft | It
M.ffl
1*
Cfr*
-------
TABLE 2-4. COMG'JSTIf»?! trFIC'iiKCIEi OF [TP TVIEi.VE-INCH "LARE HEAD
fl«F KM. n> Sirt: 17-kM*
re* i
(a
Hmf
(«)
(kw< Ml k«)
imtlm
***
(•Mi
N-Hootf
If!
«!*•!
«< ml
fit
Wufctr
ItlKltr
ICt/lM)
taiwl
(Ml
V«i«r'!ir
l H
IMNf
Itha
(¦two3*
'(«¦
talis
llfc.ltM
/Ik fr»IJ
VI*
£n
uI*Jl
mi
Kit
VI
do.)
Csktr
W*
PmI
VD
iim
iff 1
d—rj1
(11
iMtrKtlw
<¦»
CuiiHtoi
W
¦»
.T.-i
f» f
t 4M
as r
• P
M
•
«
r«i)M
r«i
K
13
N.Tt
M *
Hrzttl C,*4
•3
7.U
t a
UM
w <
1.0
m
It
C
VllljM
r*t
»
21
~».«
~f
4)
¦ >
• i
M. 1
9.0
M
n
e
VfltlM
rit
1
J5
9V. SO
«
m
i-J
It!
iA
«t
f? _
?
*Q
•
??
99. tl
tt M
(Hy 6*'.
*»
CP
(t t
Hi
LfUl §t>
i«
M
«
e
SfMfi
imk
1
1
•9.91
W W
r~is ~
C.J
8J
HI
if r*
1 8
r;
"T
ITFllt
« "
911?
*¦
He
4 1
».i
565
im
M
m
II
w
Orfeft
to
i
?3
W.W
W.90
'¦Ml.
4 G
1.3
11»
Si. J
• -4S7
M
•
Til low
«Utl»
¦
B
«¦ V
11 s«p
-U
(a) ftntpd nr. op*n area of flare head
(b) eased on v«1cctty of pipe
(c) ttefer to rigure 2-2
-------
TABLE 7-5. COMBUST!ON EFFICIENCIES OF TWELVE-INCH INDUSTRIAL FLARE HEAD A
HM* MM. (¦MfltAL A 3121: fM«0f
1
i.i
Mm'
IiU
VflMltl
injut]
<*;
*»!«•<
fait
T«l«clt>
(H/mc)
tml
u
If)
SUM
»*<<•
iirt
(U.I
Ulrr
fasU
futl
Nil
t(w
MU
(faa
tffi
ctoaey
(i)
BmH illl
I'flclwcr
III
[¦tnniw nikiMr
U
Cn
7 0
I.*
I7M
MB
to
ii
ii
¦
Til la
rn
•
??
M Ii
fl b>
M Prrrrt < ^
AJ
«. 1
4.0
IDS
M 1
e.a
01
n
t
r«lk*
Tn
•
n
n it
n a
lav tT«
M
Tm
4 ~
m
\i 7
e.e
9-t
11
•
Ttllw
m
1
rs
M *
<«.%
«• Fvnmt
('
fn
« J
4.e
uit
H.«
C.i
9-2
r
9
u\u*
R
«
~Ml
64
3.77
C.J
!»•
M J
4.6
0-2
i
•
9rum
!•»
1
a
n.»
ino.oo
ro
00
(a) Based on op*r> area of flare heart
(L) Based on velocity of pipe
(c) Peter to Figure ?-?
-------
TABLE 2-6. COMBUSTION EFFICIENCY OF TWELVE-INCH INDUSTRIAL FLARE HEAD B
hwj> mm-. MAmn i *ui- imo
i*i
Vtaal
fill
Mltrltj
(11/1*0
• hi
Halitt
im
lew
Mattof
Valw
(im/ff')
fat
(i
¦llregei
ill
StM*
la (la
ilk StMB
/I* fm\)
ahurvatl** M
laaala
frttlfrj
l-«tW
Hnd
i >
ei
tm
a#r
99.M
n h
tf tFf>M
m
IS
#/»
nit
VD
0 B
at
?l
0
Vtl/Br
*M
1
V
99. M
H.U
m
0.3
0?
1 j 14
U.I
ft ft
m
vs
0
irwfi
in
1
W
n. n
r» -m
la Bin
m
Th
? 0
Si J
ii I
• S
4-1.1
Mi
tlr/Br
to
¦
«
».
*~ j»
m
*7
~ I
I2J
• »
©-/
ra
0
6r»~
a.,
¦
?-
too 00
-------
P.BVE 2-7. COMBUSTION EFFICIENCY OF TWELVE-INIH INDUSTRIAL FLARE HF.flD C
ua mnantM. c \m ii-tmn
(fcl
!•»
lt»l
«•
BtlUflii
It)
I.U*
(MtHi
¦rttwrf
IMa
W ill it
{<
frdb*
tlO"
CIO
«*•!
INrauiip
9w tract'a*
(fHclancy
111
riff paw of r««t
*?• t
TltW
hm
'N
(alt
lilt
-•lorSt*
(rt'twi
n*
L»t*
im
•it
Calsr
SkU
Um
Iffl-
()«K|
(1)
CtfAaUlm (f'klwcy
SI
e.«
0.2
tm
99.F
a •
O-l
0
r»l^r
Tti
¦
M
nn
u f^rrnt C^t
W
1.1
1114
%.1
• i
ft-l
it
tn
a
T* ¦»
M
11.3
4.1
13ft
4§.j
#.e
ei
?i
e
nn/ar
Tm
¦
it
99.CS
9f.fl
Im Um
Ym
(.1
?.»
3JI
If-.i
• 9
9-1
15
0
Mb
R
i»
91. W
m mj
* 'A
M
V*>
13.0
« e
1?!4
'.5-f
e «M
B*>
21ft
0
*»l/or
tmit
¦
2f
tt.U
99. ;o
t o* tea
M
14.11
4.1
532
^ f?:*
M
r-j
n
6 17
tvllw
Htrtc
¦
b
«t.M
11 %
H
0 k)
V?
KG
Hi
M J
j
a
Tllte
lltti*
•
1
99.SO
n.H
o
(a) on open area of Hare i.ted
(h) Based on velocity of pipe
(c> Refer to Figure Z-?
-------
8" Typical bettings
o>
Ground
Figure W-2. Diagram of sampling locations.
-------
TOO
1000
900
800
700
600
500
400
300
2000
/
\
O 3-Irch Flare Hea4
3 3-Inch Flare Head DI /ergent Ring
O 3-Inch Flare Head Convergent Ring
/\ 6-Inch Flare Head
^ 12-Inch Flare Head
Q 12-Inch Industrial Head C
12-Inch Industrial Head A
Q 12-Inch Industrial Head B
o
x
X
TTO 10.0 100.0
Exit Velocity (ft/sec.)
Figure 2-3. Region cf flame instability.
X
1000.0
(b)
0 Confidence interval of mean.
3) Based on flares burning propane/nitrogen mixtures with no pilot flame.
-------
The region shown in Figure 2-3 indicates the minimum gas heating value
required to produce a stable fleme at the gas exit velocity within the 95 percent
confidence limits of the mean. For any given velocity, a gas with a heating
value above this region would produce a statue flame; a gas witn a heating
value below this region would produce an unstable flame. Ve'ocitv/qas combinations
in the recjion or below tend to produce flames with k'we.* combustion
efficiency. Thus, for any given t3st velocity, the minimum gas heating value
for a stable flame can be aeternined. By dividing the actual gas heating
value by the minimum value required for stability, a ratio is obtained whidi
is greater than one (1) for stable flamer, and is less than one (1) for un-
stable flames. Figure 2-4 plots combustion effici2ncia.s versus this ratio
and shows that high conbustion efficiencies are achieved when the ratio ex-
ceeds one (1). When the ratio is one (1) or less, lower combustion efficien-
cies are often obtained. Note that even at a ratio less than one (1), high
combustion efficiencies are sometimes achieved; this demonstrates the un-
certainty associated with the stability measurements. In general, however,
stable flames are efficient and unstable flames can be inefficient.
Flames near the stability limit are very sersitive to pe*turcations,
and, when perturbed, can produce high emissions of unburned material.
All conclusions are besed on the dati of tnis study and are limited to
head geometries, gases and variables examined. Head geometries were 1-mited
to:
• Simple pipe flare of 3-, 6-, and 12-inches in diameter.
• Three commercial 12-inch flare heads of different design and
manufacture.
The gases studied were limited:
• Propane-nitrogen mixtures with heating values of 270 - 2350 Btu/ft^
¦ One test with natural gas.
The variables examined were-
• Velocities from O.C to 420 ft/sec
• Reynolds nunbers from 340 - 217,000
2-13
-------
100
90
W&3
A5
A
8>-
%
c
01
c 92
o
•r»
4->
VI
3
o
90
m
A
O
O 3-Inch EER
A 6-Inch EER
O 12-Inch EER
[N 12-Inch Coioercial A
0 12-Inch Commercial B
O 12-Inch Coratierc'al C
(^Reason for low cntustlon efficiency 1s unknown.
86
X
X
12 3 4 5 6 7 3
Heating Value/Minimum Heatinq Value for Stability
Figure 2-4. Combustion efficiency near the lower limits of flame stability
(b) Based on flares burning propane/nitrogen mixtures with r.o pilot flame.
\.a)
-------
• Richardsor number from 2.9 x 10-5 to 8 x 10^
• Steam flow from G to 1 lb-stean/lb-fuel
The flare flames *ere shielded from the wind and combustion efficiencies were
not measured in winds greater than 5 mph. The following conclusions are based
on the r,tudy resu' s:
• Flares operating with unstable flames can have Inw
combustion efficiency.
• The combustion efficiency did not depend on fl.ire size or
geometry.
• Successful correlations were developed for fla^e flames:
Floma length was correlated with a modified R1chard.»on
number.
Liftoff distances were correlated with ratios of velocities
and concentrations.
Flame stability wa? correlated with a pseudo flare
temperature.
Entrainnient was correlated with ratios of distances and
the Richardson number or with velocity.
• Combustion efficiency was ',igh for flares w>th high velocities,
provided the heating value of the gas was in the region of stability.
• Steam injection completely suppressed soot production but did not
appreciably alter combustion efficiency unless the flame was
oversteamed (>0.5 lb-steam/lb-fuel), and then the combustion
efficiency decreased.
L ¦ 1
-------
3.0 TEST METHODOLOGY
During this program, the Flar* Test Facility testing procedures and
data reduction techniques were deve1 o;od to characterize trie erissiens from
several flare heads which were tested over a range of operating conditions,
3.1 Flare Test Facilities (FTF)
A Flare Test Facility was designed *nd built by Energy and Environmental
Research Corporation at their n Toro Test Site for the Environmental Pro-
tection Agency. Details of the FTF are sunman zed below. A Complete description
of the FTF Is contained in Appendix A.
Facility Site
The FTF is located in a canyon surrounded by 70-foot cliffs. The layout
of the test site includes a pad and structure for testing, screens for pro-
tecting the fl.tre flames from *1nd, and a control room for operating the flare
and analyzing the data. Sufficient water, comoressed air, natural gas and
electricity are available at the test site to operate the flare facility at
gas flows up to 24,000 ft3/hr.
Delivery Systems
Gases are supplied to ths flare head arid auxiliary equipment through
specially designed parallel nanifolds, These manifolds allow accurate deter-
mination of the flow rates over the wide range of operating conditions typical
of flares. Natural gas is supplied by the local utility, through a 2-inch
pipe. Natural gas can be delivered at a maximum flow rate uo to 7,00fl ft*1/
hr. Propane is delivered and :to red as liouid in a 2100 gallon tan*. Vapor-
ized propane is supplied to the flare head. At low flow rates, natural vap-
orization is sufficient; at higher flow rates, propane-f1 ,-ed vaporizers a»-e
3
used to increase the flow-rate of gas up tc 15,000 ft /hr.
.ntroyeri Is fed to the flare head from liquid • roqen ta"ks after at-
mospheric vaporization. Nitrogen is mixed with propane to vary the heating
value of the gas flared. Nitrogen tanks anrj vaporizers are capable of sup-
plying up to 9,00V ftJ/'hr Sulfur dioxide is fed fror; liquid cylinders
through a steamhaated vaporizer to nhe base of the H.ire stack. Sulfur
"j
Dioxide can be supplied at a 'die if 7 ft" /hr.
3-1
-------
Steam is produced in a 15 hp natural gas-fired iwjilor. The boiler is
capable of supplying 400 lbs/hr of 100 psiq-saturated steam.
Flare Heads
Simple pipe flare heaar, were designed and bui "it by EER for testing.
These heads were designed to simulate the major features cc commercial flare
heaas; however, tl'.ase heads did not include the different proprietary
features of commercial flare heads used lu :ze flames and improve
mixing. Three open-ended pipe flares with steam injection were used -"or
most of th^ tests. These heaii were 3-, 6-, and 12-inches in diameter. The
steam injector was designed t; provide 0.23 lbs steam/lb fuel at 29 percent
of the maximum capacity (0.3£ speed of sound) of the flare head. For the 56
percent propane In r,1trogen mixture jsed in man'/ of these tests, the maximum ve-
locity of a 12-Inch Hf.re head would be 325 fv./sec, the flow at 20 Dercent maximum
capacity would be £961 lb/hr, ind the maximum steam flow would be 1142.6 lb/hr.
For tfiF.se conditions, steam would he injected at 1289 ft/sec (Itech Number *
0.92) through sixteen 0.525-inch-diameter nozzles.
A ring was used ori the 3-iriCh flare head to help retain the flame at
high vt^ocity. The ring was designed after consultation with the advisory
pan
-------
Itu ii f Urefiecd flaaw
HetentSuii Rlny
f
13. fXK)"
I
JJJJSO*;
¥.(>&•
eve I
CWi I
MTMith Axil
Uts-yr Thick
| ^di*tMR ^tcel
?*
. Jeld AH
A round
4-inch
I iarrimait r-
Hun
ill of f hrttheid Pl|>e
1/8* Mound Hole
12 Spated tvenly on
2 100- Circle
- I/a" Round I'ole
Drilled al 60°
Wllh HetaI iur
Itee I? spired
» [venly u»i 2-1CO"
tenter Circle
IMe
ir-IOO-
Cii.,
Milt
Burner
('.ire Mead
No
-------
• The steam injection system shall oe designed for a maximum
steam flow of 13,000 Ibs/hr. {This figure is based on a design
exit velocity of Mach of 0.35 for a hypothetical gas in the
12-Inch flare; smokeless operation to 20 percent maximum capacity;
and a steam requirement of greater than 0.3 lb steam/1b fuel.)
Samplinq
Samples of the effluent from the flare plume are withdrawn by two tech-
niques. In the first, samples are withdrawn from the mixed effluent of the
flare flame in the cimwey of a hood d' signed to capture the complete plume.
Pitot probes, thermocouples, md a sai*;! Jng probe are positioned 1n the
chimney of the hood. Tnesi- -How the r*us flow through the hood to be de-
termined, ,i sample tj .»¦» withdrawn for analysis, and calculation of the mass
flux rf incompletely burned carbon species through the hood. The second
sampling technique uses a -ake probe. Five individual, movable probes are
positioned along the diameter of the flare pHme. The emfre probe assembly
can be moved vert ca-ly and horizontally.
The probe in the chimney and the rake probes are designed similarly
The probes are 1-1/2" 00 tubes, steam-heated to avoid condensation with in-
terchangeable nozzles which allov, isokinetic sampling. Soot is captured by
filters at the end of the probe, and gases are withdrawn through a central
tuoe into heated sampling lines. Water is removed near the probe by Pertna-
pur^E)dryers.
Ambient Control
Wind can greatly disturb the flare flame and cause difficulty in
measuring emissions. The wind velocity is monitored by a three-cup anemo-
meter, and trie flame 1s shielded Dy wVid-screens which have 22.7 percent
open area. Flow patterns around the flare flame, as determined by smoke
tracers, are shown by Figure 3-2. Trials are not conducted at high wind
condition*, and many trials were conducted in the early morning or late at
night to tike advantage of quiescent ambient conditions.
Visual Monitors
A number of visual monitors we-e found useful to record the flan*
3-4
-------
** f Natural
Wind Screen
Figure 3-2, Flow Patterns Determined by Smoke Tests.
-------
itructun* during the trials. These included:
a A videotape recorder which was used during the early trials to
continuously monitor flame structure, however the usefulness of
the video recording was limited by slow response times and inade-
quate spatial resolution, 1t was not used in later trials.
• Photographs of the fla~e flarre were used to record the structure
of the fiame and as an aid to interpreting data. Photographs
were taken with exposure times of 1 millisecond to capture the
instantaneous structure of the flame and at 10 seconds to record
the average shape and length of the flare 11 are,
• High-speed cinematography was very useful for defining the nature
and evolution of flare flames at speeds or ?00 franPs/seco"d.
This speed was determined experimentally to be capable of identi-
fying the development and growth of eddy structures in flare
flames.
3,2 Test Procedure
The test procedure involves recording background conditions, lighting
the flare flame, establishing the flame conditions, setting the sampling
locations, sampling, and analyzing the sample.
Background Con_d_i_tions
The backgroun.1 conditions 'Influence the results of testing because of
the wind and the concentration of the products of incomplete combustion in
the ambient air. The wind speeu is recorded and He direction noted. The
opera to*" assesses, based un previous experience. the extent to which the
wind will influence the structure and emissions of the flare flame. Tests
are usually conducted before noon or at night to rc-lrce thr> interference
from the wind.
The background compos *ti.on of the ambient air is sampled before each
test at the position >f sampling during testing. The background iafnple is
handled and analyzed exa-.tly as samples from the flare pluire.
3-6
-------
The Flame
The flame is lit with a hand-held torch, and the flow conditions estab-
lisned. The ?"!are structure is reco.ded v'suaily and photographically, with
both still and motion pictures.
Steam
Steam lines are preheated and drained prior to trials using steam in-
jection .
Samp! iriQ
The position of the samp • 1 r»g probes is citablishej frons previous ex-
periments. Figure 3-3 shows the optimum position of hood sampling as deter-
mined by a i-.cbon ba".?rce. The position is a compromise between being so
close to the flame that the hood disturbs the fiare and so far from the
flame that the combustion products are diluted. r1gure 3-3 shows that sone
material is not collected by the hood at large distances above the flare
head. However, Figure 3-4 snow? that the calculated vOnbustion efficienc.es
did not vary with sampling positions of the hnod. This implies that the
material which is not collected has the same composition as that which "is
collected.
Probe sailing also has an optimum position, howev ..* fcr different
reasons. Positions of the probe in and rear the flame may sample material
wi.ich is incompletely burned. In positions far from the flame, tne con-
centration of the gases nay be so dilute that making an accurate determina-
tion of the composition of combustion products 1s difficult.
The soot-' aden gases are drawn nearly isokineti :a1ly through a filter
placed in the end of the probe to capture soot. This avoids deposition of
^oot on probe lines. The amount of gas is determined by a dry gas meter,
and th« amount of soot collected is determined as the weight of material
which can be burned from a previously baked filter.
(?)
The gases are drawn through heated sample lines to Pennapure^ dryers
where the water is removed. The dried gases are then collected 1n Tedlar
bags. Five samples are drawn simult.aneously and individually into the bags
fro.n the rake probes. Samples are taken over a 20-rninute period to average
out flame fluctuations. Samples are mixed in the bags prior to analysis.
-------
6-In. Flare Head
36% Propane
2.3 FPS
No Ste—1
JL—I 1 >11' I 'I'll'
4 r 3 ID 1? 14 16 13
Hood Fosition (Bottom Edge From "Tare Head, Feet)
Figure 3-3. Influence of Hood Position on Closure of a Carbon
Balance using Mass Flti/es Measured in the Hood.
3-8
-------
100
4->
c
0)
£ 99.9
a.
«w
I 89«8
*
*»—
U
t 99.7
LJ
C
O
S 99.6
t/>
3
JD
® 99.5
99.4
2 4 6 8 10 12 14 16 18
Hood Position (Bottom Edge From Flare Head, Feet)
Figure 3-4. Influence of Hood Position on Combustion Efficiency
Meaiured in tne Hood.
3-9
i—r
O SOOT
-<-*> TOTAL
6-In. Head
56% Propane
2.8 FPS
No Steam
-------
Carbon monoxide and carbon dioxide are analyzed using non-disparsivs in-
frared analyzers, total hydrocarbons are analyzed as methane using a flame
ionization detector, sulfur dioxide is analyzed using absorption ii solution
followed by titration and hy flame photometry. Details of these techniques
are described in Appendix A.
Mass Balance
A perceived deficiency with previous studies was the inability to close
a m.aS balance and account for all the carbon. The; combustion efficiency
In these studies could be lower than implied from the measurements, because
the carbon efficiency is calculated based on the amount of uriburnt material
measured divided by the amount of combustibles fed to the system. This study
showed that carbon balances can be attained under some conditions and es-
timated the importance of closing a material balance. (See Section 5.2)
Two techniques were used to evaluate the influence of mass balance clo-
sure on the estimate for the emission of unburned hydrocarbons. In one tech-
nique, a hood was used to collect 'he entire ef'luent from the flare flame,
and the amount of carbon collected in the hood was determined from the irass
flow in the hood cs calculated from thermocouple measurements and pitot
probes, and by analysis of the CO, CO, and hydrocarbons 1r a gas sample with-
drawn from the chimney of the hood. Figure 3-5a shows that accurate carbon
balances could be attained between mass flow rates of 20- and 40-lbs/hr of
propane, while Figure 3-5b shows that mass balances could be closed for many
trials using ooth th» 3- and 6-inch f1-;re heads. The discrepancies shown in
the figures are thought to te caused by inhomogeneitles and measur,nent in-
accuracies at low f'ow rates and by the loss of mstenal at high flow rates.
The loss of material at high fow rates was confirmed by observations of
flow patterns in the hood using smoke-tracer?.. The use of Su? a? a tricer
was inappropriate for tests with the hocJ. Figure 3-6 shuws that carbon
mass balances co'/ld be closed using a rake proue and SOg as a tracer for
mass fluxes up to 90 lbs/hr. Carbon mass balances could not be obtained
using low flow rates of S02 (solid points; because of inaccuracies and
1 imitations of the technique used to measure S0^ concentrations in the
plume.
3-TO
-------
2.0
E
O
*3 K>
O <_J
103
120
'c " hS -PF
Figure 3-5a. Influence of Mass Flyx of Carbon on Closure of
a Carbon Balance Using the Mass Flux Measured
in the Hood,
~1 ' 1—
C 3-in. i«w, ;j! c3.ia
® 3-in. , 56* C-,H
A 3-in.
O 5-in.
-sao,
•aid,
3' 3'
c* 3tj
oS C^rlg
Stftftir
m
c
a
m
o
a &
If
o
exit 7«1acit/
Figure 3-5b. Influence of Velocity on Closure of a Carbon
3alance Using ^ass Fluxes Measured -in a Hood.
Figure 3-5. Verification of Canon Hass Balunce for Hood Measurements.
3-11
-------
Hood Sampling
C\J
o
GO
TJ
QJ
3.0-
2.0-
A
A
# SO? Injection Rate = 0.47 scfh
A SO2 Injection Rate ¦ 3.3 scfh
c
c
2
i-
(T3
(_>
A
A &
S
50
100 150
mc I IN AS C3H8 (lb/he J
200
Figure 3-6. Closure of mass balance using SO2 as a tracer.
-------
3.3 Data Reduction
Reduction of the data requires:
t Correction for background concentrations ba<;ed on dilution.
* Calculation of local cotr.bustion efficiencies and maximum
potential errors.
« Integration of local combustion efficiencies to cbtain glosal
combustion efficiencies.
The procedures used for these reductions are describeo below.
Correction For Background
This boction describes the procedures used to calculate the combustion
efficiencies at the sample locations and to estimate the maximum potential
errors associated with these efficiencies. The procedure requires:
a Assumptions on the nature of the combustion products.
• Correction for concentration of combustion
species in the background.
a Estimates of the maximum potential error issoclated with
each lac.il combustion eff'ciency.
Combustion Products
Knowledge of the completely and incompletely burned combustion products
are required to calculate local ccxhustion efficiencies. Ca*irulc.tion of
the local combustion efficiencies are based on the i-.complately ourntd
carbon. Carbon dioxide is considered the only comoletely burned carbon
species: Carbon nonoxide, soot, anc1. nydrocarbons reported as methane ?c,uiv-
elents are considered incompletely aurned species. Generally, these assump-
tions are valid.
dilution Factors and Backgroi>nd Concentrations
The combustion products are mixed with ambient air, which may contain
significant levels of the species in the flare plume. The concentrations
of incompletely burned species must then be corrected for the amount of
material added by the ambient air. This correction is based on the stoich-
3-13
-------
loaetrlc products of combustion and 1s accoapl1 shed throigh the use of mate-
rial balances. Simply subtracting the concentration species 1n the back-
i -jund from the measured concentrations Is Incorrect and nay lead to errors
when concentrations of incompletely burned hydrocarbon In the ambient air and
the flame plumt are similar. This can occur because of efficient combustion
or excessive dilution.
The dr>it1on factor, OF, Is defined as the moles of ambient air entrained
with one mole of stolchloaietrli combustion products. The dilution fictor 1s
calculated from mate*1al balances (see schematic diagram) conserving any of a
number of species.
Overall Combustion Efficiencies
An estimate of the total amount of Incompletely burned hydrocarbons exl >
1ng from flare flaws Is required to adequately characterize these emissions.
The overall emissions were obtained In two ways 1n this study. First, the
entire plume was collected, mixed, and sampled 1n a hood for flare flames
with low flows. This technique w*s limited to r>mall flows because of the
capacity of the hood, fan, and associated duct work. Second, the global
efficiencies for flares with large flows were
-------
j. The flare plune consists of live discrete zones measured by thL
rake probes in:.iae of which tha concentrations and velocities
can bs approximated by a uniform composition equal to that
measured by the probe 1n the center of the zone.
The combustion efficiencies can be calculated by:
r l..r«; . 1 UnSurnt Carbon Flow ,,
CombusJon Efficiency - 1 - ~Tota1 Carr^TTTw" (3_12;
Above the flare fliime, the unburnt and total carbons fluxes m.iy be calculated
by:
Unburnt Carbon Flow = ^ Z C. V A (3-13)
i r 1 ,r r r
Total Carbor, Flew =¦ f C. V A (3-14)
J ' J j ' ' i
where C- is the concent-ation of carbon species i
Vr is the axial velocity at rpgion r
A^ is the cross section area of region r
i denotes unburnt. species CO, HC, soot
j denotes tota1 carbon species: CO, HC, soot., CC^
r denotes the regions covered hy the rake probes
Boforg the above relationships can be applied to compute combus-
tion efficiencies, the measured specie; concentrations must be corrected for
the influence of background concentration. Also, vrlocit^'es and areas of
flux must De specified.
Background Corrertior,
The concentration of 3 species at the probe position can bs corrected for
the concentration of background speci?:> by:
C = C ( , ]ut12r' Factor^ .
: ,ret ~ i ,measured " dilution Factor + T i, background (3-17)
3-15
-------
where the dilution facor is defined as:
Total Volume Of Gas Which Contains
DF = The Stoichiometric Corcfcustion Products (3-18)
Volume of Stoichiometric Combustion Products
Locating the Flame Axis
The axis of flare flames occasionally do not :orresoond to the a;:is cf
the flare head. Since the regions along the flame axis have minimum air en-
trapment, the flame axis is located at the oosition where the measured dil-
ution is a rrin-'rnum.
Defining the Sampling Region
The flare flames are assumed to be axisymmetrical. In addition, the
species inside the five regions covered by the rake probes are assumed to be
homogeneous and idsncical to that measured by the probe. The boundaries be-
tween these region; aie selected to be half-way points between adjacent
probes.
Velocity Profiles
The velocity profile of the flare plume is assumed to take the form
determined by jet theory.
1016 (|) - 1.5) 13-15!
Vr,X- V*P<- ™(T'2> <3"l6>
where X = axial distarce from flare head exit
r = radial distance from fUme axis
d n diameter of flare heed nozzle
» centerline gas velocity
Vr x = local gas velocity at radial distance r and axial distance X
The velocity within e.>ch zone is assumed to be uniform and identical to that
of the probe oosition. With the above simplifying assumptions, the relative
contribution yf sech probe sample to the combustion efficiency can be
wei ghed.
3-16
-------
Th#? procedure for calculating the toribusticn efficiencies is
recognized to oe approximate. However, major e-rors are eliminated by tha
procedure used to calculate these efficiencies. The velocity, which is
-------
A single probe will ther yield the overall combustion if efficiency if the
" r
local combustion efficWcy Z. iiu _ . . . „„ „
J j a*" i2j ts constant across the flame.
Figure 3-7 shows in example of tb
-------
'Soot
—
13-
6-1 rich F1 are Head
10.3 FPS
14.6 Percent
0 lb.Steam/lb.Fuel
-4 -2 0
P.adial Distance (ft)
-igure 3-7. Local ;omt>ustion effi cienr.ies of the 6-inch EER test flare
head burning l-i.6 percent oropane in nitrogen at 10.3 ft/
sec with 0.6 U> steam per lb of fuel.
3-19
-------
u.6
Q
V.
0.14
0.18
0.22
0.06
n.io
0.02
ro
o
J
L
(Re) = 17000 3t nozzle
2 =
axial distance
y. *
radial distanre
d
<)
nozzle diameter
a =
inlennittency factor, related
to timn of presence
Figure 3-0. Distribution of Intemittency Factor.£, an-i Mean Forward
Velocity in a Round Free Jet (Corrsin ft Kistier;
-------
REFERENCES
SECTION 3
Corrsin, S., and A. L. Klstler, "The Free-Stream Boundaries of Turoulent
Flows," National Advisory Committee for Aeronautics, Technical Note
NACA TN 3132 (1954).
3-21
-------
4.0 FLAME STRUCTURE
The main purpose of this prcgram wes to estimate the combustion efficien-
cy of co*merc1al flares. Currently it fs virtually Impossible to measure the
combustion efficiency of operating flares because of the flame size, elevated
location, and lack of spatial stability. As a consequence, this program was
designed to measure the combustion efficiency of large, pi lot-scale flares.
Extrapolating the results of pilot-scale flares to larger flares requite?
the knowledge of the scaling laws of combustions and of the mechanisms which
lead to inefficiency. Therefore, considerable effort was devoted In this
program to defining the operating and design parameters wtilcf control the
structure of flare flames ana determine flare combustion efficiency. In
addition, a number of structure! features are Important to the design and
oj>erat1on of flares:
• Flanw volume
9 Flame length
• Flame dilution
• Flame liftoff distance
• Flame stability
This program *>as collected data on large pilc-t-scale flare flames over a large
rar.ge of d?s1gn and operating parameters representative of Industrial use.
Several measurement techniques were used to investigate the structure of the
large pilot-srile flare flanes. The character sties of the flames were ob-
served against a graduated background, visual observations were supplemented
with long-exposure and short-exposure photographs, high speed motion pictures,
and measurements of gas composition 1n the plume of the flare •flames.
Flame shape (volume and lennth) .s directly related to the rate of com-
bustion and the mechanisms controlling combustion and will Influence combus-
tion efficiency as well as the amount end the cPn*ro1d of radiant emission.
i-l
-------
Previously, the length and volume of buoyant transitional and turbulent jets
have been studied. Correlations of the flan* Jength have been d&veloped from
these studies. These correlations have been suggested and are used to
estimate the structure of flare flames. Most of these studies investigated
the flame structure of small flan~.cs over a limited range of operating con-
ditions. Some of the available corrections were unable to predict the
results from this study and are not recommended. The correlation of Zukoski
et al (4.1) and of Hottel and Hawthorne (4.2) successfully predicted the re-
sults from this study when the coefficients we1** altered. However, a more
accurate correlation for the flame length was developed in thi3 study,
length was developed in this study.
The rite and extent of dilution of the combustion products in a flare
flame will influence the temperature, rate of quenching and, hence, combustion
efficiency. The dilution factors measured in this study were correlated using
a Richardson number approach similar to that of Spaulding and Ricou^-3). This
con elation was unable to adequately account for the influence of diameter on
the dilution factor, and a more accu-Jte dimensional ccrtelation was developed.
The liftoff distances of the flame control the extent of entrapment of
air prior to combustion and, hence, the rate and extent of combustion. This
study showed that deteched flares can be stable and operate with high combustion
efficiency. The liftoff distances of this study were correlated by assuring
dilution of the combustion products followed measurements rrade in simple jets,
and t'.ci criteria that the flame speed must equal the imposed velocity for
stable rmnSvstion. Efficient combustion can only take place when the flare
flame is maintained. Inefficient combustion often occurs near the limits ;f
flame stasility. Tt.e stability limits of flape: in this study were correlated
using results from simple jet theory, measired flame speeds and measured flammabil
from simple jet theory, n»asured flame J^eeds and measured f1ainnabi 1 ity
limits. Calculations bayed upon accepted theories were unsuccessful in
predicting the performance of flare flames (liftoff distance, shaoe, com-
b.Jition efficiency), but correlations were developed with special constants
based upon the theoretical equations.
4.1 Test Conditions
The operating conditions of th° flaves tested in this program covered
-------
the range of parameters of coirmercial flare-.. This section will discuss
the design of the test matrices, explain any modifications to the test ma-
trices, and compare the conditions tested ver us those of commercial flares.
Test Matrices
Three test matrices were developed initially and modified as the program
progressed. The first test matrix was designed to develop scaling principles
for flare flames. The second was designed to test jonmerclil flare heads ai
conditions comparable with those of the first test matrix. The third was
designed to test combustion efficiencies of flare flames at high velocities.
The original test matrix was designed to develop scaling Uws fcr the
structure and emission of flare flames. As sjch, the matrix was designed to
test four scaling criteria over three small, 3'Jt comnercial size, flare heads.
The scaling principlas were:
• Constant velocity, u
• Constant residence time, d/u
• Constant Reynolds number,
• Constant Richardson number, gd/u^
where u and d are the exit velocity and diameter*, P and u are th?j axit gas
density and viscosity, and g is the gravitational constant. The sizes of the
flare heads wire 3-, c- , and 12-inches in die,meter. The matrix was also
designed to test the influence of steam injection on combustion efficiency.
The original test matrix as modified in the program is shown ir Table 4-1.
The conditions cf this matrix, denoted by an "X1', were tested one or more times.
A second test matrix was designed to compare the flame structure and
emis:iors of commercial flare heads with those of the prctotype head designed
by EER. The matrix for testing of commercial flare heads is shown in Table
4-2. This test matrix was completed with minor exceptions. The combustion
efficiency at one probe position for the Industrial Flare Head B was not
measured; however, a complete stability curve was obtained for this head ana
time did not permit determining combustion efficiency at the last prube
position. This omission was thought to be of minor importance since the com-
bustion efficiency was determined at twe other rake positions for the same
4-3
-------
TABLE 4-1, FLARE TEST CONDITIONS
r -
i ,
i . ip Jlcdd
1 Inche*.
Si/c
Vr • v
< _/ aPC
1 llcotirig yjluo
i RUi/U
Steam/Tnel Ratios
Smoke-
less
Jind
j6*BD
__ P<
inr
Hdk
e
ion
L3
16
1
j ¦
L *
1 led
H.,j.
llTO | 100
0. 15
n.30
0.50
0.75
).o
Low
Hdturd;
Hood
1
t
1
r*
r-
X
1
1
0.5
X
X
X
X
X
X
1
; u
y
X
X
X
X
X
X
10
K
i
i
X
X
X
X
x
!
10
A
X
X
X
x
0.5
X
X
X
X
—!_
2.0
X
X
X
Y
X
A
10
I
X
X
10
X
y
X
X
X
X
X
A
1 u
X
X
_ J
t.
2.a
X
*
; '
X
t
X
X
'0
y.
X
'
X
X
l.u
X
;
X
X
X
X
2.8
;
X
X
X
X
10
X
X
X
X
1.0
X
X
X
r X
. 8
X
X
X
X
X
x
X
in
X
X
X
1 x
10
X
X
X
0 Z
X
1
1
X
X
2 A.
I j
X
X
X
4
X
1
X
X
/
X
X
0.2
X
X
X
2.0
X
X
t\
X
4
*
1
X
X
X
A
X
4
X
1
X
X
X - Tests conducted
-------
TABLE 4-2. MATRIX FOR TESTING COWERCIAL FLARE HrA
Measurement of Combustion Efficiency
Velocity
Gas
Btu/ft3
Steam
Rake
1
Probe
2
Position
3
Industrial
Fl'i^e Head
A
0.2
2
i:oo
1300
No
No
X
X
4
1300
No
X
X
X
4
1300
Yes
X
4
286
No
X
Optional
Curve
Stability
Nc
©
Industrial
Flare Head
B
0.2
2
1300
1300
No
No
X
X
4
1300
No
®
X
X
4
1300
Yes
X
4
286
No
X
Optional
Curve
Stability
No
X
Industrial
Flare Head
C
o.:
2
1300
1300
No
No
X
X
4
1300
No
X
X
X
fl
1300
Yes
X
4
286
No
X
Optional
Curve
Stabi1i ty
No
X
Xj * Test not conducted
4-5
-------
flaro flame conditions. Time did not permit completion cf the optional
stabil'ty curve for the Industrial Flare Head A however one point of
the stability curve was determined.
A third test matrix, Table 4-3, was developed to test the Influence
of high velocities on the combusMon efficiency of flare flames.
Operating Conditions
The range of conditions tested in this >tudy covered the majority of
the rj|.srating conditions which are common practice commercially. Figure
4-1 shows the r*nge of flare diameters, exit velocities, residence times,
Reynolds numbers and Richardson numbers tested In this study. The range
of commercially operated fUre Is similar but includes larger diameter
flare heads.
4.2 Mechanism of Combustion
As described above, flares have a wide range of operating :ondit1ons
and these conditions resilt in different mechanisms controlling combustion.
Combustion Is controlled In all flare flaries by the rate of" mining of the
flame gas with gxygen from the air. Mixing in the flare flame is dictated
by:
a Buoye.ncj forces
• Moleculcr difdfusloi
e Jet nixing.
This section described correlations developed for the flamt length, entrap-
ment, liftoff distance ana f^me stability with operating panmeters of the
flare. These correlations were develooed fr,r propane-nitrogen mixtures and
applicability to other gases has not jeeri remonstrated.
Mixing Mechanism
Two classical descriptions of flame structure are available. Hottel and
Hawthorne (4-2) described the length of flames as a fjr.ction of velocity, and
R1 cou and Spauldlng (4-3) described the entrainment of ambient gases into
flames as a function of the froude number (ratio of inertial forces _o buoy-
ancy forces; it Is the reciprocal of the Richardson number).
4-6
-------
TABLE 4-3. TlST CONDITIONS FOR HIGH-VELOCITY FLARES
Test names Flare Btu/ft^ Steam .
Low Med I um
1 Open 1300 30 Ib/hr or Soot Suppression X X
|
2 Ring 130C 30 Ib/hr or Soot Suppression X X
3 Ring Minimum Soot Suppression X X
4 Ring 600 Soot Suppression X Y.
5 Ring 1300 ioot Suppression > X
6 Ring 2350 Soot Suppression ;! X
-------
1 Heads
I )Ch, LtR
i :h. r.ER
Inch. EER
Inch. Conmcrcltl A
ch. CcMMrcltl L'
Inch. ConMerc1.il l
|()3 TP TCP
Ho/rle Reynolds Number^ Po **0 uo .
Figure 4-1.
Comparison of the Range of Conditions Tested in Thi^ Study and Those
of Commercial F'.a;\-. Flames.
-------
The flares of this study had Reynolds number., below 2x105 and bastd on
the flamo length criteria of Hottel and Hawthorne (*-2), they were not fully
developed turbulently. Figure 4-2 shows the length of flames measured In
this study on a plot similar to t<>at suggested by llottel and Hawthorne. The
plot Indicates that the length of the flames in this study continued tn in-
crease is ihe Reynolds number increased. The tine In Figure 4-2 corresponds
t;o flames burning 56 percent propane In nitrogen. The length ond structure
of fully turbulently d*>/'?loped flame*, no longer depends on Reynolds number.
Since the flames of this study were Independent of Reynolds number, nixing is
partially controlled by molecular diffusion and buoyancy forces, rlgure <-2
also shows that smaller nozzles and higher amounts of combustibles yielded
longer flames, while addition of steam slightly reduced the length of the
flame.
Similarly, the Richardson number for the flames studied varied from 3x10-5
to 8x102, Indicating that most flames were controlled by buoyancy-dominated
mixing. The Richardson number 1s the rat'io of buoyance forces to Inertia
forces: values greater than one Indicate predominance of buoyancy force:;,
while values less than one Indicate j predominance of inertia! forces. En-
trainmen t as a function of Richardson rumber will be shown later on a plot
similar to that suggested by R1cou and Spauld^ng(4.3J.
Rate of Surface Combustion
Control of mixing by molecular diffusion to the flame would yield flames
with constant-heat-release-per-unlt-surface. Figure 1-3 shows that the sur-
face heat release rate increases with hsat release rate from j\000 to almost.
200,000 Btu/ft^/hr. Therefore, mixing of material on the fUme surface does
not strongly contribute to control o1' the combustion rate for flare flames of
this study.
Volumetric Combustion Rate
Volumetric control o* combustion rould be the resu1,t of control by two
mechanisms. The first can occur when gases are conpleie'i. ri -.e»J at the
molecular level and the combustion rate is controlled by kinetics. Kinetic
control is unlikely for those flames with slow mixiig and Mpid combustion.
The second can occur when mixing is controlled within a volume. This can
occur when the flame envelope includes pockets of air and combustible gases
which are not mixed on the molecular level. "Igure 4-4 shows that constant
volumetric combustion is approximated for all the flames of this study. A
4-9
-------
3-Tnch Moizle, 56 Percent
Propane in Nitrogen
cy" • &
<>~
13-Inch,
13 -1 n c h,
13-Inch,
~3-Inch,
s6-Inch,
.6-Inch,
>12-Inch
~12-Incn
•IP-Inch
»12-Inch
~12-Inch
•12-Inch
112-Inch
112-Inch
112-Inch
700-1400 Btu/ft
1500-2(X)0 Btu/ft3
2350 Btu/ft3
300-600 Btu/ft3
700-1400 Btu/ft3
300-600 Btu/ft3
, 700-1400 btu/ft3
, 300-600 Btu/ft3
, Natural Gas, 930 Btu.'ft3
, Ccinrwrcia1, A, 700-1400 Btu/ft3
, Comercia' A, 300-600 Btu/ft3
, Commercial B, 700-1400 Btu/ft3
, Conmerci il B, 300-600 Btu/ft3
, Connercial C, 700-1400 Btu/ft3
, Commercial C. 300-600 Btu/tt3
!¦- 1 ——— I ¦¦ I —i
io3 W~ ~~ io= 106
Reynolds No.
Figure 4-2. Correlation of Lengths of Flames Studied with Reynolds Number.
4-10
-------
I! 00
19(1
lfiO
17tJ
160
ISO
110
130
120
IU
in
90
w
70
60
bCJ
"0
30
ZO
n
A 6- inc.1) ® 0
l?-inch
•"¦hadec ane Smokeless Conditions
o
o
4-
A
o
0 8
A
A
£
* °o Oo
• v.
£ «*>
A £
O
)
11)00 10,000
I'.ste (Id-1 titu/hr)
100,000
4-3. Surface Rate of Combustion of Flare Hames.
-------
O EER 3- incn, No Steam
& EER 6-inch, No Stean
o EER 12-inch, mo Steam
k Industrial 12-inch A, No Steam
90.000 Btu/hr ft
1000
10U
pfi in7
H
-------
A least square regression analysis would yield:
Flame Volume ¦ 1.21 x 1CT5 (firing rate, Bt;.'/hr)0'^
(4-1)
However, the data does not support such an accurate equation and the simple
relation that the volumetric heat release of flare flames is constant at
90,000 Btu/hr/ft3 is preferred. This value is supported ':,y evidence on pool
flc»mes(4-5) and from industrial experienced-^).
The volumetric heat release of the flare with steam injection is
typically increased (lower volume of the flaire tor a given heat-"elease rate';.
This confirms conventional thinking xhat steam increases the indiction of air
into the raot of the flame and increases mixing inside the flame envelops.
The reduction of the flame volume by steam injection is not large, but
observations and measurements show that this difference is suffJc1ent to
completely suppress soot production.
The above evidence supports ideas of combustion by volumetric mixing in
flare flames. Volumetric mixing will be important in flames into which air
is induced by gross mixing mechanisms. This can occur wher the buoyancy-
dominated flame engulfs large amounts of air into the apparent flame volume.
Comoustible materials are not molecularly mixed inside this volume, but are
congregated into pockets of air which ara mixing with pockets of fuel. The
mixing of there pockets controls the burning rate. This view is supported by
short-exposure photographs and high-speed rujtion pictures taken during this
study and is consistLnt with the view of coherent structuresd-7-4-S).
4.3 Flame Length
Calculations of the length of flare flames are needed to estimate
radiation from the flames to the ground, and as an Indicatior of the com-
bustion mechanism of the flames. At the start of this work, e number of
empirical relationships for flame lengths existed. Some of thesu were par-
ticularly complex^-11), and some were recommended for predicting the
length of flare flaues(4-11). The conditions under wnich these expressions
for flame length were derived are listed 1n Table 4-4. Many of the cor-
relations tere derived under conditions different from those of flare flames.
Table 4-4 shows that the major differences between this and previous studies
is the use of large nozzles and extension to high Richardson numbers (i.e.,
Duoyancy-doininated flames).
4-13
-------
TABLE 4-4. CONDITIONS llf.'DER WICH CORRELATIONS FOR FLAMES WERE PERIVEO
4
INVESTIGATES
...
NOZZLC
SIZES
(INCHtS)
FUEL OR
INJECTED
GAS
NOZZLE GAS
VELOCITY
fFT/SEf)
REYNOLDS
NUMBER
RICHARDSON
NUMBER
This Work
3, j, 1?
Propane-Ni tro-
gen Mixtures,
Natural Gas
G.2 to 423
337 to
217,000
2.9 x 10"b *
8.04 x 102
Hottel &
Hawthorne (4.2)
0.25,
0.1875,
0.125
CO, City
Gas
To 258
1910 to
16,700
Greater than
1 0 x 10-5
Hicou &
Spaulding (4.3)
0.0625 to
1.25
Air, Hydrogen
Propane, C02
80 to 250
2500 to
80,000
2.68 x 10 6 to
524 x 10-4
Becker k
Liang (4.10)
0.1 , G. IB
Lcwwerci al
Propane
11.5 to i'60
1310 to
48,200
0.91 x 10-5 to
4.43 x 10-3
Brzustowsk!
(4.11)
Zukoski (4.1)
0.197
41
15,823
3.K x 10"4
2.54. 7.48
19.69
City G?s
0.0427 to
0.984
160 tc
2500
2.33 to
6-4 x 104
-------
Most of the existing models were unable to predict the f1arne lengths ob-
served in this study. Figure 'r-5 shows that estimates by Becker and
liang^ ^ and iirzu3towskl(4-ll} couid preaict neither the absolute value nor
the trends in flame length. Howe/er, while rhe expression of Hottel ana riQ«-
thorne(4-2) could not predict the absolute value of the flame len.jth, the
expression predicted the trends.
The expression of Hottel and Hawthorne w&s derived for laminar diffusion
flames; and contained two constants. Adjustment of these constants resulted in
a simple expression which reasonably fits the data of tnis study (figure 4-5).
The success of the Hottel and Hawthorne correlations suggests that, to some ex-
tent, the amount of fuel Ted to the flare partially controls the rate of mixing
and, hence, the flame length. Correlations of flame length ^"or pool fires were
also available. These correlations were obtained assuming the buoyant force
is dominant. The result of one such correlation is shown in Fiqure 4-7. The
correlation of flame length with buoyancy Voices is quite reasonable for short
flames which were minimally influenced Dy velocity. However, 'he relationship
unaer-preciicted the ler.ath of long flames which were partially controlled by
inertia! effects.
Finally? a correlation was derived whicn predicted the lengths of all the
flames of this study. The correlation is based cn the Richardson number
corrected for the temperature rise of the plume. Tne Richardson number con-
tains both the influence of inertlal and buoyarcy forces. "!ne correlation:
L/D • 7.41 ( 9-— )°-6U a-'"0-'15 (4-2)
Cpu. 1« (1 + 26 x)
is shown in Figure 4-8, and is cajole of predicting the (lame length of this
study within about 20 percent. The terns are defined as:
[. = Flame length, ft
D = Nozzle dianf-ter, ft
Q - Keat release cf fuel, Btu
X = The fraction nf propane in the propane-nitrogen mixture
26 = A factor to account for stoicniometric combustion of propane
with air and will change for other fuels
',-15
-------
55
50
45
40
n
4-J
cn
I 35
s
*
73 3'3
ai J-
a
£& A
o
o
o
o
A a
o
o Hottel t Hawthorne
A BrjuttwsH
Q Becker and L'any
? Qj
O .06
o A $
* o <>
o
-------
30
20
HotteVs Modified Correlation
(ft) (ft^/sec)
U
t_>
10
0
10
20
30
40
50
Measured Flame Length (Feet)
Figure 4-6. Comparison of a Hottel and Hawthorne Correlation with the Flame Length
of This Study.
-------
20 30 40
Measured FT aire Length (ft)
Figure 4-7. Comparison of Flame Lengths Predicted by a Buoyance Correlation with
Observed Flame Lengths of This Study.
-------
50
45
40
35
30
25
20
15
10
5
0
0.60
.CD
20%
10
15
5
20
7.5
30
35
Observed Flame Length (feet)
Figure 4-3. Correlation of Flame Lengths in This Study.
4-19
-------
Cp,,. = Hot capacity of amhient air, 3t.u/lh mole of
To, ¦ Temperature of ambient air, °F
F = Fraction of heat radiated from the flame (empirically
derived)
Rf = Richardson nj.nber ¦ gd/u2
The fraction of heat release emitted from thi flame was derived from a single
set of data at a Richardson number of 2.0, as shown in Figure 4-9. The
fraction of heat of a pure propane flame was assumed to be that value neces-
sary to predict the flame lengths. The factors determined are a function of
only gas composition and are shown In Figure 4-9. These values were then
successfully correlated with all the flame lengths of this .study. The derived
factors agree with visual observation of flarces burning different amounts of
propane. The amount of propane does not affect the flame structure or apparent
radiation above 60 percent propane. Below that value, the radiation of the
flame decreases with decreasing concentration of propane.
4.4 Entrainment
The rate ct air entrainment has some influence on the rate of corbustion
and flame length, Lut has a large influente on flame liftoff and stability.
The flames of this study had Richardson numbers varying from 3 x !G~5
(inomentum-doTir.ated flj ies). to 8 x 1CK (buoyancy-lominared flames) (see
Figure 4-1). Ricou ?.nd Spauldipg'4-3) have successfully correlated the rate
of entrainment for momentum-dominated jets and flames. The correlation in-
volves the Froude number (the reciprocal of the Richardson nunter); however,
this term appeared on both sides of the expression and cancelled from the
final correlation.
Entrainment in the flare flames of these stuiies was much faster than
predicted by jet theory. Figure 4-10 shows that entrainment in this study
was grouped into two regimes. The first regime is for flames with low
velocities. Entrainment in this regime was two-to-three orders-of-magnitude
faster than that predictea by jet tl.eary, depended inversely on velocity,
but did not strongly depend on nozzle size or the heating value of the yas::s.
The entrainment for high-velocity flames fell into the second regime. Here
4-20
-------
t r
Required Correction for Heat Loss
Ri - 2.0
-o
20
40
60
60
JL.
100
Mole Fraction Propane in N2
Fiyure 4-9. Empirical Correction for Radiation loss fron
Flare Flames.
4-21
-------
10,000 -
O 3-inch
0.5 ft/sec
I ft/sec
10 ft/sec
O 3 inch.
40 ft/sec
80 ft/sec
€i 3-1nch,
\20 ft/sec
ft 3- i 'K '1
140 ft/sec
0 3-inch
250 ft/sec
© 3-inch,
350 ft/sec
© 3-inch.
HO ft/sec
O 3-1nch,
M 6-incli
1 ft/sec
A (j-inch.
3 ft/sec
10 ft/sec
-------
the ent-ainment was closer to that predicted by jet theory for the higher
velocity jets in this study.
Attempts to correlate the entrainment rate using the dimensionless
distance and the Rich:.rdson number were only partially successful. The high
rate of entrapment apparently caused by engulfing fluid In a buoyant plume,
and the inverse relationship between entrainment and velocity, suggest that
the Richardson number should aid in correlating the entrainment rate for
flare flames. Likewist, the increase in entrainment with distance above the
flare head suggests that the dimensionless distance, X/d, should aid In
correlating entrainment. Correlations of the entrainment rate with the
product of X/d and the Richardson number were developed and are shown in
Figure 4-11, but were of only medium strength. Development of strong cor-
relations was hampered because of the strong dependence on nozzle diameter
and the fact that this was reduced by Incljsien of the Richardson number
which also contained the nozzle diameter.
A dimensional correlation was developed which reasonably correlated the
dsta. This correlation invoked the dimensionless distance and the recipro-
cal of velocity. Thp correlation shown in Figure 4-12 successfully collapsed
all ihe data of Figure 4-10. This data included 3-, 5-, and 12-inch nozzles,
variation in heating value f om 300- to 2350-Btu/ft3, and nozzle velocities
from 0.2 to 430 ft/sec. This correlation can be used to roughly estimate
the entrainment of ambient fluids Into large flare flames operating in
the region between buoyant- and momentum-dour,nation.
4.5 flame Liftoff
The liftoff distance of a flame car, in theory, be estimated by simple
combustion theory. The liftoff distant; will be the distance required to
induce sufficient air to reduce the jet velocity at least in one location,
to the flame velocity. The flame velocity will pass through a maximum near
the stoichiometric mixture as air is mixed with fuel. Once the flame is
stabilized at a location, the propagation of an ignition source must extend
to all regions of the flared gas to achieve efficient combustion. Aero-
dynamic devices or pilot flames can create regions of ignition which will
propf-.iate ana stabilize a flame under conditions where the flame Kould
1-23
-------
100,000
10000
o
o
o 1000
+J
X
c
'00
10
03-Inch
S 3-Inch
3-Inch
Q 3- Inch
€5 3-Inch
§3- Inch
3-Inch
3-Inch
)3-Inch
12-Inch
k6-Inch
>6-Inch
*6-Inch
>12-Inch
~12-Inch
>12-Inch
1
0.5 ft/sec
2 ft/sec
10 ft/sec
40 ft/sec
80 ft/sec
120 ft/sec
170 fi/sec
250 ft/sec
350 ft/sec
420 ft/see
1 ft/sec
3 ft/sec
10 ft/sec
0.2 ft/sec
2 ft/sec
4 ft/sec
3-Inch
6-Inc
12-Inch
10
x0-7099 0.243
d
Mgu^e 4-11. Correlation of entralmifent rate for flare flames.
4-24
-------
100,000
10000
O
a: lac
01
1000
c
o
100
X
10
O 3-1nch,
12-inch
12-inch
^ 12-inch
- A
10
100
.1.676
1000
0.5 ft/sec
2 ft/sec
10 ft/sec
40 ft/sec
80 ft/sec
120 ft/sec
180 ft/sec
250 ft/sec
350 ft/sec
420 ft/sec
1 ft/sec
3 ft/sec
10 ft/sec
, 0.2 ft/sec
, 2 ft/sec
, 4 ft/sec
10000
1
,0.838
igure 4-12. Correlation of entreinment rate 'or flare flames.
4-25
-------
normally be unstable.
The liftoff distances measured in this study are showr. in Figure 4-13.
A reasonable curve jf liftoff distance versus velocity is defined for gases
with heating values tetween 700- and 1400-Btu/ft^. Gases with higher
heating values have slightly smaller liftoff distances at equivalent
velocities. The liftoff distances for low-Btu gass.; do not correlate as
well. This ir. ascribed to th» extreme sensitivity oc the stable flame
location for low-Btu gases to ambient perturbations.
Attempts to calculate the liftoff distance? using jet theory and flame
•speeds were unsuccessful. Jet theory is strictly applicable tc fully
developed jets outside of the core region. The flare flames studied here
are not developed turbulently, and the liftoff distance typically is with-
in the core region and 1s difficult to characterize. Some success was
achieved in estimating liftoff distances using jet theory modified to the
conditions of the flames of this study. However the stability, and hence
the liftoff distances o* the flames, was found to be extremely sensitive
to the velocity Drofiles and few useful results could be obtained using
modified .jet theory t.o estimate liftoff oistances.
A successful correlation of liftoff distances was, however, developed
using the abeve principles. Figure 4-14 shows the liftoff distances measured
in this study correlated with the prcduct of the ratio of nozzle velocity to
maximum flame velocity, and the ratio of concentration at the maximum flame
speed to the concentration of gas 1n the nozzle. The correlation is very
good. However, sensitivity or' flames burning low-Btu cases to perturbations
results in uncertainties in the measurements as it did in the calculations.
4.6 FlameStability
The stability of the flame will influence the combustion efficiencies.
Flames operating near the region of stability (as defined in Section 2) ma."
result in incompletely burned hydrocarbons when slightly disturbed. A
flame will be unstable when the vectorial jet velocity is greater than the
flame velocity. The flame may adjust itself to a position where the two
are equal. However, the jet velocity ^s reduced ir* flare flames by in-
duction of air. This lowers the mixing strength of the 'uel and the flame
4-25
-------
r 1
700-HMO fitu/ft''
1500-2000 Btu/ft3
.i',0 Bti,/ f f 3
3 incf
3-inch,
3-inch,
3-inch,
inch
b-inch,
Q 12-inci
^ l?-inch
12-inc1!
^ I?-inch
Q 12-inch
f l2-in>,
i '¦ 1 ? - i m h
\° i-ich
700-1130 Btu/tt
JW bM [itu/ft.
700- i400 Btu./ft3
300-600 Rtu/ft3
700-110C ritu/ft3
30-600 fitu/fti
COf-nerciul A, 70Q-1400 Btu/ft3
Corn yrcial A, 330-600 Btu/fl-1
Cmmercisl II, 700 1400 Rtu/ft'*
,C,.;.-rc!a' B. 300-600
lo.'ridl C, 700-7430 BtJ/ft3
CofiWial C, 3^J-60fi Btu/ft
300-601 Btr/f
Velocity (ft/sec)
Figure 4-13. Liftoff of flare flames without pilot flames.
-------
1000
0)
(J
e
c/>
o
<4-
<«-
0
1
100
10
r
3-inch, 700-1400 Btii/ft3
3-inch, 1500-2000 btu/ft3
3-inch, 23[»0 Btu/ft3
3-inch, 300-600 Btu/ft3
5-inch, 700-1400 Btu/ft3
6-inch, 300-600 Btu/ft3
12-inch, 700-1400 Btu/ft3
12-inch, 300-600 Btu/ft3
12-inch Commercial A, 7C0
12-inch Commercial A,
12-inch Commercial 1,
12-inch Coirtrcercial B,
12-inch Commercial C,
1?-inch Commercial C,
1400 3tu/ft3
600 Btu/ft3
•1400 Btu/ft3
600 Btu/ft3
a3
300-
700-
300-
700-1100 Btu/ft3
300-600 Rtu/ft3
X
10
rtozz1o V J ority
Maximum Flame Speed
100
Cone at V,
1000
max
Cone at
Figure 4-14. Correlation of liftoff distances for flames without pilot *lawes.
-------
speed when the mixture strength 1s on the lean side of stolchiometry. A
flame cannot be maintained under conditions where the vectorial jet
velocity cannot be reduced to the flame speed before the mixture strength
of the fuel 1s outside ths limits of flanriabllity. The result is that
low-calcrific gases flared at nigh velocities have narrow regions of
stability, operate close to the limits of stability, and are subject to
perturbations and may result in pocr combustion efficiency.
The minimum heating value of propane nitrogen mixtures required to
maintain a stable flame with i given nozzle velocity was determined in this
study. Figure 4-15 shows that at low velocities, stable flames could be
obtained with gases of heating values of approximately 300 Btu/ft^. How-
ever, high heating values of the gasas were required to stabilize the
flames at higher velocities.
Most of the data was obtained for rlare heads without special means
of stabilizing the fame. None oY the flames were retained with a pilot
flame. A convergent and a divergent ring were usee on the 3-inch flare
flames but this did not improve flame stability. Some commercial flare
heads employed proprietary devices to aerodynamically stcDilize the flame.
These devices resulted In some differences in stability betweer. the EER
prototype flare head and the commercial flare heads.
In theory, the stability curve shown in F.gure 4-14 may be generalized
using other fuels by plotting the log of veloc ty versus a reciprocal 'lame,
temperature. Figure 4-16 j.hows that many normal hydrocarbons follow tnis
relationship. This relationship assumes that the flame velocity is a
function of reactions with similar Arrehenius parameters. As evidence from
figure 4-16, this is not true for compounds of widely different structures.
Hovevs. , Figure 4-17 shews this relationship to hold for the propane data
of this study. In this clot, the flame temperatu-e is approximated by:
Flar.ie Temp - Cpa) ^ ^
where Cp^ Is the heat capacity of ambient air, X is the volume fraction of
propane in the fuel, 26 4s the number of moles rf products for stoichiometric
4-29
-------
1100
1000
900
800
700
600
500
100
300
2000
0 3-Inch Flare Head
3 3-Fnch Flare Head Divergent Ring
% 3-Inch Flare Head Convergent R1n«)
_ A 6-Inch Flare Head
Q 12-Inch Flare Head
~ iiL-lnch Industrial Head 1 C
£» 12-Inch Industrie! Head 2 A
Q 12-Inch Industrial Head 3 B
o
-L
*a^CI = Confidence Interval
- i 1
1.0 10.0 100.0
Exit Velocity (ft/sec)
1000.0
Figure 4-15. Region of flame instability. ^
b) Based on flares burning propane/nitrogen mixtures with no pilot flame.
-------
10
1
r ¦ i
o
Acetaldehyde
~
Benzene
o
1,2 Butadiene
A
3utane
o
1 Butene
k
2,3 Dimethyl-2-Butene
C>
1, Butyne
Q
Cyclohexane
o
Ethane
o
Ethene
4.0
0
£
Ethyl amine
Methane
o
Propane
0
0
Propene
3.0
Toluene —
A
2.0
"
oX
o
@5>
V
SyA
1.0
1
1 ^
2.0 x 1(H 2.1 x 10"4 2.2 x 10-A 2.3 x IO"4
1 / T °R~1 (adiabatic)
Figure 4-16. Flame SDeeds of hydrocarbons.
4-31
-------
1 -
( i
6
3-Jnch Flare Head
1000
O
3-Inch Flare Head Divergent -
Ring
•
•
A
o
3-Inch Flare Head Convergent
Ring
6-!nch Flare Head
it Inch Flare Head
100
l/l
\
•
~
V ^
12-Inch Industrial Head "1 C _
12-Inch Industrial Head 2 A
4-
\ 0
12-Inch Industrial Head ? B
u
u
0)
QJ
73 10
NJ
O
z
-
a \
t
~
0
o
1 „
CTJ
C
+J
E
o
1
~
o
a
\o
0 \
0.1
i
i i
Cp„,(1 Q x"~ 'l/0R)
Figure 4-17. Limits of flame stability as a function of
estimated temperature.
4-32
-------
combustion of propane with air, and Q is the low heitlng value of purs
propane.
4-33
-------
REFERENCES
SECTION 4.0
4.1 Zukowski , E., T. Kubota and 3. Cetegen, "Entrainment in Fire Plumes."
Fire Safety Journal, Vol. 3, pp. 107-121 (1980/1981).
4.2 Hottel, H. C., and W. R. Hawthorne, "Diffusion in Laminary Flame Jets."
Third Symposium on Combustion, Flame an«3 Explosion Phenomena, pp.
254-266 (1949).
4.3 Ricou, F. P., and D. B. Spaulding, "Measurements of Entrainment
Axisymmetrical Turbulent Jets." J. Fluid Mech., Vol. 2, op. 21-32
(1961).
4.4 Set:bold, J. G., Private Communication, Standard Oil Company of
California, San Francisco, California (1982).
4.5 Orloff, L., aid J. de Ris, "Froude Modeling of Pool Fires." Nine-
teenth Symposium (International) on Combustion, pp. B85-895 (1982).
4.6 Shore, D., Private Comnunication, Flare Gas Corporation, Spring
Galley, New Ycrk (1983).
4.7 Roshko, A., "Progress and Problems in Understanding Turbulent Shear
Flews" in S.N.G. Monthly, ed., Turbulent Mixing in Nonreactlnq and
Reactive Flowt, pp. 295-311, Plenum (1975).
4.8 Broadwell, P. E., and P. E. Dimotakis. Contractor Coordination
Meeting, Fundamental Combustion Research, U.S. EPA Contract No. 68-
02-2631, Npwport Beach, California (1980).
4.9 Marble, F. E., and 1. E. Broadwell, "A Theoretical Analysis of
Nitric Ctxidfc Production in a Methane-Air Turbulent Diffusion Flame."
EPA Contract No. 68-02-2613, April 1971.
4.10 Becker, H. A., ar.d D. Liang, "Visible Length of Vertical-Free Turbulent
Diffusion Flames. 1 Combustion and Flare, Vol. 32, pp. 215-237 (1978^.
4.11 Brzustowski, T. A., ''Flaring in the Energy Industry." Progress in
Energy and Conbi-stlor. Science, Vol. 2, pp. 129-161 (1976).
4-34
-------
5.0 COMBUSTION EFFICIENCIES IN FLAR>".S
The main objectlvs of this study was to deyelop methods to characterize
the conbustion efficiency of comerclal flares. This section reports the
characteristics and combustion efficiency of J.he flames of this Investigation,
compares the results of this study rltl those of other studies, and correlates
the combustion efflclsncy with operating parameters of the flsre. This section
reports the resuits jf tests whlcn developed techniques to scale the emission
of Incompletely burr.o'i hydrocarbons from flare flames, tests of commercial
flare heads, tests of flares at high exit velocities, and tests on the Influ-
ence cf steam Injection on the emission of Incompletely burned hydrocarbon from
flare flames.
5.1 St"dy Conditions
Appendix C provides a complete list of all tests conducted. Tables 2-1
through 2-7 summarize these conllMons and eliminate duplicate conditions,
samples withdrawn from the same flare und the rake probe at different positions,
end samples taken within the flame. Conditions studied include:
• Flare Head
- 3-1ncn EER pnr.otypp
- 6-Inch EER prototype
- 12-Inch EEP prototype
- 3-Inch EER prototype with convergent ring
- 3-inrh EER prototype with divergent ring
- 12-inch commercial; 3 manufacturers
• Casts
- Mostly propane diluted with nitrogen
- Cf.e data point using methane
• Gas Heating Value from 2B6 - 2350 Btu/"t3
• Gas Exit Velocity from 0.2 to 428 ft/sac
• Steam Injection f^om 0 to 1.0 lb steam/lb fuel
The combustion efficiencies of these trials are reported In the following
sections.
5-2 CcmpaHs*>n tilth Other Studies
Th:s study set out to remove some of the uncertainly ut previous studies.
The major uncertainties were:
5-1
-------
• Ability to close a mass balance;
• Measurement of soot concentration;
» Assumption that the local combustion efficiency measured at
one point 1s representative of the global combustion efficiency;
• Development of a scaling methodology.
Mass Balance
r!iis study was able to close a mass balance using both the hood and S02
as a tracer (see Section 3). The results indicate t'.at flar* combustion effi-
ciencies can usually be determined without strict closure af a mass balance.
This implies that material which is lost form the sampling region is negligible
or of similar composition to that in the sampling regime. Therefore, the
inability to close a mass balance does not preclude use of the data from pre-
vious studies.
Soot Concentration
Most previous studies failed to measure the concentration of carbon as
soot for a large number of samples. The combustion efficiency measured in uuch
studies would be Mqner by the amount of unmeasured soot which escapes the flame.
Datff from this study (see Appendix D) indicate that soot rarely accounts for more
than 0.5 percent of combustion inefficiency and can be completely eliminated by
injection of steam. Measurement of the soot concentration is usually unimportant
for the determination of combustion efficiency.
Axial Combustion Efficiency
Measurementi of local combustion efficiency at a single axial position may
not be representative of the combustion efficiency of flares as discussed in
section 3.3. Graphs of dilution factors at various distances from the olune
center line (see Appendix F) shew that velocity profiles must be combined with
concentration profiles and integrated across the flame to obtain overall combustion
efficiencies. Thus, the single orobe technique for estimating combustion efficiencies
may rot be as accurate as the multiple trobe technique used the study.
5-2
-------
Scaling Methodology
Scaling methodologies were developed for all the important characteristics
of flare flame? in this study. Volumetric combustion was found to be a con-
3
stant, 90,000 Btu/ft -hr, for the flames of this study. However, the nozzle
size arid operating conditions were required to scale the following characteris-
tics of flame structure*.
« Flam? Length
• Flanie Lift-off
» Flame Dilution
The limits of flame stability could be correlated with the propertiei of
the gases irrespective of size of the flare head. These correlations are
discussed in detail In Section 4.0.
The combustion efficiency could also be correlated using only the proper-
ties of the gases for the flares of this study, ('.wever, the combustion effi-
ciency is correlated with the stability limit' of the fla.ne which co"ld be
changed by addition of proprietary stabilizers and pilot ligl.ts or different
gases.
5.3 Correlations
The goal cf the correlations we.s to determine if the test data was
scalable over the range of variables examined. The combust 1i--r. efficiency Df
the lf: j pilot-scale and commercial flare flames was found to be independent
of the diameter of the flare :i=nd vithin the ;ize range tested,
v
The combustion efficiencies estimated in this study did depend on the
exit velocity and the heating value of the gas. The velocities studied were
in the range of velocities for commercial flare heads. However, only mixtures
of propane 3rid nitrogen were tested. The results may be different for other
gases, but are expected to correlate with the same parameters as propane.
Figure 5-1 shows the conditions (uark symbols) which resulted in combus-
tion efficiencies lower than 98 percent. The line is a least-square fit to
the stability curve of Figure;. 2-2 and 4-15. Most of the conditions which
result in combustion efficiencies less than 98 percent are below the region
5-3
-------
O1 OCD
O O O
OO
o oqo;
1000
O>0 a on
h o
100
.0
O
O
O
O 3-Inch >982, # 3-Inch -98%
A 6-Inch >98%, A 6-Inch -98*
O12-Inch >98%, + 12-Inch <982
k 12 Com A >98», Ik 12 Con A <9«X
0 12 Com B -98%, 4 12 Can B <98%
~ 12 Cum C >98*, 0 1? Com C <93%
_L
I
0.1
1.0
10.0 UO.O
Velocity ft/sec
1000.0
Figure 5-1. Conditions to Obtain ^Su-Percent Combustion Efficiency for Mixtures of
Propane and Nitrogen.0
(a) No pilot flame.
-------
of stability. Scare flanes at similar conditions resulted In combustion effi-
ciency greater ,han 98 percent. However, all those conditions are near the
limits jf flame stability. At these conditions, slight perturbations wilt
cause e flame operating with high combustion efficiency to become unstable
with a resultant decrease 1n combustion efficiency.
The Influence of flame stability is more clearly shown when the graph 1s
plotted in dimensionless form. Figure 5-2 shows the plot of combustion effi-
ciency /ersus the ratio of the heating value of the flame to the minimum heat-
ing value necessary to sustafn a flame at that velocity. The heating values
at the stability Halt were determined from the lowe' region of stability
shown 1r Figure 4-15. Some flares operating below tie minimum heating value
required to maintain a flame have high combustion efficiencies. However, all
conditions which caused low combustion efficiency were firing gases within
10 percent of the mlnmum Heating ualue required for a stable flame. Only one
flame firing a gas with ? ratio greater than 10 percent above the value re-
quired for a stable flame resulted In low-combustion efficiency; the reason
for this 1s unknown, and the data point- 1s ccnsiderrd an anomoly. Hose
flares fiHng gases with heatfng values near the stability limits are sus-
ceptible to perturbations and poor combustion efficiency.
These results are Milted to the conditions ana gas n. xtures of this
study. Howevtr, the form of the correlation is expected to be generally
valid fcr many gases anti for flare heads wBlch rely on externa1 mixing.
Future studies evaluating the effects of other gases and the Influence of
pilot flames and aerodynamic devices on flane stability *111 provide addi-
tional information.
5-5
-------
100
98
96
94
C
C
4->
T3
£>
E
Q
O
92
90
88
36
O
o o£y
* ^a>-
o
A
l
O
_i_
o
A
o
k
0
~
3-Inch EER
6-Inch EER
1?-Inch EER
12-Inch Corrmarci al A
12-Inch Cormercial B
1?-Inch Commercial C
X
X
X
X
2 3 4 5 6 7
Heating Value/Minimum Heating V?1ue for stability
Figu-e 5-?. Combci'Lion Efficiency Near the Limits of Flame Stability
(a) Dased cn flares burning propane/nitrogen mixtures with no pilot Flame.
(a)
-------
5.4 Co—ercial Flare Heads
Three commercial flare heads were loaned to the program by manufact: iv*£ nd
tested. The manufacturers were requested to supply a standard 12-Inch f1 ifu i.^ad
as described 1n Sectlor. 3.0. The geometry of the heads differed arid all vtlicitles
reported here are based on the open area of the head.
The flar* heads showed small differences in combustion efficiency when burn-
ing 56 percent propane. Tables 2-5, 2-6, and 2-7 show combustion efficiencies for
the different heads ranging from 98.3 to 99.1 percent.
When burning the minimum heating value gas required to sustain a flame,
larger differences In combustion efficlencler were observed. (See "Low Btu Tests"
on Tables 2-4 through 2-7.) At low velocltlas tha performance of the head?, was
similar. However, the performance of the EER prototype hsad, which had nu flame
retention device, and fla,te head C, were poorer at higher velocities-
5.5 High Velocity
Tests were conducted to de'ertnlne the influence of high velocity on the com-
bustion efficiency of flare flai.ies. Propane-nitrogen mixtures ,*ere burned on the
3-1nch flare head at velocities uu to 428 ft/sec. All the flamjs in these tests
were lifted. Results are shown 1n F'gure 5-1. Only two condition.® at intermedi-
ate velocity resulted in a low rombusuon efficiency. The combustion efficiency
at all other conditions was greater thon 99 percent.. From this we cncluded that,
provided the heating value of the gas 1s not close to the minimum v'alue to main-
tain a flame, high velocities may sligntly Improve the combustion efficiency of
flare fi aiives.
5.6 Steam
The Influence of steam injection on cumbustlo.i efficiency was ul so deter-
mined during these trlais. Figure 5-3 shot". that steam Injected Into the flares
of this study had effects similar to those expected 1n commercial practice. That
Is, steam completely suppressed soot. However, steam Injection at normal rates
had a minor Influence on overall combust:-:n efficiency because the amourt of *oot
is small. Optimum levels of steam Injection for combustion efficiency were found
to be 0.3 to 0.5 lb steam/lb fuel suppressed the stot, hut ,'ed'jced the overall
combustion efficiency by quenching combustion of CO and hydrocarbons.
5-7
-------
Efficiency (Percent) Measured With the Hood
-------
APPENDIX A
EPA FLARE TEST FACILITY AT EER
Ai
-------
ABSTRACT
The der^gn of the Environmental Protection Agency's (EPA) Rare Test
Facility (FTF) at Energy and Environmental Research Corporation is described.
The criteria for 'he design, construction, calibration and operation yf the
FTF are reported. Details of construction and operation are given of the:
•
Test Site
•
Fuel System
•
Tracer System
•
?team System
•
Flare Head
•
Sanoling System
•
Visual Monitors
«
Ambient Monitors
•
Support Structure
•
Data Recording
The FTF is used in EPA programs to determine the combustion efficiency of flares,
and is available by arrangement for testing of design, construction, operation
and use of commercial flare heads.
-------
TABLE OF CONTENTS
Section Page
1.0 OVERVIEW
1.1 Objective A "i -1
1.2 Design Guidelines A 1-1
2.0 FACILITY SITE A 2-1
2.1 Topography A 2-1
2.2 Facility Layout A 2-1
2.3 Control Room A 2-1
2.4 Utilfties A 2-5
3.0 FUEL SYSTEM .A 3-1
3.1 Fuel Supply A ?-l
3.2 Flow Control Metering A 3-1
3.3 Specification Performance A 3-4
4.0 TRACER SYSTEM A 4-1
4.1 Tracer Supply A 4-1
4.2 rrac.er Flow Control Metering A 4-1
5.0 CTEAM SYSTEM A 5-1
5.1 Steam Supply A 5-1
5.2 Flow Control and Metering .A 5-1
5.3 Performance A 5-A
6.C DESIGN OF CLARL HEADS ... A 6-1
6.1 Flare Head Suuport Base A 6-1
6.2 Flare Tip A 6-1
6.3 Steam Ring and Nozzles .A 6-1
6.4 Pilot A 6-5
6.5 Flame Arrestor ... A 6-5
6.6 Flame Retention Ring A 6-6
Alii
-------
Section Page
7.3 extra;tiye SAMPLE SYSTEM A 7-1
7.1 Sample Points . A 7 1
7.2 Hood Design A 7-1
7.3 Multiple Point Sampling , A 7-3
7.4 Sapling Conditioning A 7-5
7.5 Analysis A 7-8
8.0 VISUAL MONITORS . . . A 9-1
s . 1 Video System A 8-1
8.2 Photography . . A 8-1
8.3 High-Speed Cinematography A 8-1
9.0 AMBIE.'iT MONITORS AND CONTROLS A 9-1
9.1 Wind Screens A 9-1
9.2 ^riitors A 9-1
10.0 SUPPORT STRUCTURE 'a 10-1
10.1 Overall Structures A 10-1
10,? Hood Support .A 10-1
10.3 Rake Probe Support . .A 10-1
10.4 Windscreen Support .A 10-1
11.0 DATA RECORDING SYSTEMS A 11-1
11.1 Cat j System A 11-1
11.2 Strip Recorder A 11-1
12.0 ORIFICE CALIBRATION .A 12-1
A iv
-------
FIGURES
Figure rage
1-1 EPA flare test facility at itR Al-2
2-1 Terrain around Hare test facility A2-2
2-2 Flare test facility site plan A2-i
2-3 Flare test control room A?-4
3-1 Fuel flow control and metering syst<»m schematic A3-2
4-1 50g vaporl ier A4-2
4-2 SOg flow control arid metering system schewatic, A4-3
5-i. SteaiR flow ccvxol and metering syste* schematic A5-2
5-! Calibration if steam f'io* A5-5
5-3 Calibration cf steam fit* A5-6
5-* Calibration of steam flw A5-7
5-5 Calibration of steam flew , A5-8
6-1 Flare head support structure . A6-2
6-2 Flare head design A6-3
7-1 Sampling nood assembly ..... A7-2
7-2 Filter probe design A7-4
7-3 Hood and probe rak? positioning mtchaniset A7-6
7-4 Flare sample system A7-7
8-1 Camera platform . A8-2
9-1 Perforated plate windscreen ..... A9-2
10-1 Support structure for sampling hood ana rake probes .... A1G-2
«v
-------
TABLES
Table Pagt
3-1 Pilot-scale flare test conditions A3-'.
3-2 Fuel syste« components . A3-'i
4-1 Tracer flow system component specification M-H
5-1 Steam metering system comooncnt'. A5-1
6-1 Flare head specifications . A6-1
7-1 Analytical methods and accuracies A7-9
A-l Calibration constants of laminar flowmters A12-2
h-2 Orifice flowweters calibration constants A12-4
Avi
-------
1.0 OVERVIEW
A ter.t facility (Figure 1-1) has teen designed and constructs^ at Eno'< -
and Environmental Research Corporation (I'Ert) to test large pilot-scats flares
'inder U.S. EPA Contract b8-02-3661. fM'r, report describes the foil crying:
• Facility ,ite
• Flare 3as delivery system
• Tracer delivery system
• Steam Gel ivery system
• Desion of the flare heads
• Extractive sampling systems
• Visual monitors
¦ Ambient monitors arti controls
• Support structures
• Data recording systems
1.1 Objective
The pilci-scal > facility has been designed to determine the combustion
efficiency of small flares. In particular, t'ro facility it designed to answer
the v'ollowincj questions:
- What are the combustion efficiencies of pilot ann small
conniercial flares?
• How are thesi; efficiencies influenced by ooe*at rig parameters,
flar? design, fuel composition and size?
• What mechanisms cortrol these influences?
1.2 Desion Guidelines
The criteria for designing the pilot-scale facility were establishe..
oy consultation with the Technical Advisory Panel, composed of Flore
Al-1
-------
, PROPANE
\ TAHK
aou
rH can mot. 5
& MFTERS
FLOW 1
CONTROLS
t MET
roots H
^ERS I
VAPORIZER
CITY
NATURAL
GAS
flow
CONTROLS
I t METERS
-I
("s" y-(z)A ;'I^T
•.'»?
I I
I
ir
j.
H/\L F R S
, CO. CO?
HC. SO,
SAMPLE
DRIERS
{SAMPLING
-i control
I t'AHEL i
ItOLAS
BAGS
EPA f 1-?re test facility at EER.
-------
mrnufacturers and users, arid are summarized In EPA Report No. 6C0/?-83-070.
• Flare head size - 3, 6, and 12 inches
-------
Concentration Measurements
The efficiency of flares is determined from gai concentrations of samples
collected In a hood or probes at several points above the flame. Species con-
centrations of Og, CO, C02, r.ooi and SC^ are obtained using both techniques.
Compositions obtained from flares operated at different renditions are used to
ertablish the combustion eff ciarcy.
A1-4
-------
2.0 FACILITY SITE
The site for the flara test facility was selected to provide:
• predictable wind conditicnr;
a !ow background levels of jOj, CO, CQj, soot und hydrocarbons;
• Isolation from heat-sen'?Itive equipment.
The flare test facility was built in a canyon which provides suitable
topography, room for facility ljyout, including the control room, and access
to necessary utilities.
Z.l Topography
The topography of the test site can minimize the difficult problem of iso-
lating the flare from the ambient conditions. Figure Z-l shows the terrain
around the test sit*. The canyon wall is rbout /0 feet high and surrounds the
flare facility on three sides. Wind blows predominantly alony the direction
of the canyon. The flare facility is on a flat clearing about 200 feet long
and 100 feet wide.
2.2 Facility Layout
Figure 2-2 shows the plan of the facility site. A chain-link fence
topped with barbei wire surrounds the area to prevent unauthorized entry.
Access to the test site is through c* eleven-foot wide swing gate. The major
insta1lation5. include the control room, the hood anu sampling suppot i. structure,
the steam bo.ler, fuel tanks and vaporizer the liquid nitrogen supply and
vaporizer, the flow controls and the flowmeters.
2.3 Control Roon'
The control room is directly opposite and provides view and access to
the flare head. It houses the gas flow control panels, the electric control
panels, the sample flo\* control panels, the gas analyzers, and data recording
squipment. Figure 2-3 shows the arrangement of the control room. A window
provides a complete view of the flare head, sanple hood, and probes. Gas
flow rates and electrical power are controlled inside and can be turned off
rapidly during emergencies.
A2-1
-------
At>o»«. Sea Le*t
llare
Control
a
-------
E= PHI1"AHL TANK ")
FEET
I I I I II
:;;-i 6 Ri i
3»
ro
i
to
STORAGE
NAY
t
PROPANE |
VAPORIZERS ^ t"J"
\
ENTRANCE
GATE
NATURAL _
GAS (Xj
SUPPLY
MHR
-------
F' JW
Control
Panel
Ann 1y -
zcrs
trols
Coal
laboratory
Observation
Window
:ontrol
Figure 2-3. Flan? test control room.
AIM
-------
2.4 Utn ItlfeS
Water, compressed air, natural cas, and electricity are available at the
test site. Water is supplied through a 6 in. pipe at 70 ps1g. Natural gas
is suppled at 15 psig through a t in. pipe. Compressed air is supplied by
a shop compressor at 100 psig through a 1 in. pipe. Electricity is de'iv^red
to the test site at 440 V three-ohase and 110 V and 220 V are obtained through
a transformer. The available electrical capacity is 300 KVA.
A2-5
-------
3.0 FUEL SYSTEM
The fuel system includes the supply an^ flow control and metering
equipment for the fuel and diluent gases. The experiments reported herein u?ed
propane and natural gas as fuels. Tnese gases have H/C ratios of ?.67 and 4,
and heating values of 2450 and 98b Btu/ft^, respectively. Nitrogen was
used to dilute the fuel gases to the desired heating values. All gases were
metered to control tnc gas exit velocity at the flare head. Figure 3'1
illustrates the control system for the fuel system.
3.1 Fuel Supply
Table 3-i listi the planned experimental conditions and the flow
requirements. The total flow rates range from 11 to more than 28,000 cu ft/hr.
Propane is storea as a liquid in a 2100 gallon tank. Natural vaporization
of the liquid provides sufficient propane gas for low-flow experiments. At
high demand rates, three propane-fired vaporizers, each capable of 2900 cu
,"t/hr supply the needed gas. The tank will last about 11 hours at the highest
flow rate anticipated.
Natural gas 1s supplied at 15 psig froir the gas utility company through
a 2 in. pipe. The capacity is expected to be about 7000 cu ft/hr which is
sufficient for testing the 6-inch flare at 10 ft/sec ar.cJ the 12-inch flare
at slightly more than 2 ft/sec.
N'trogen is used to dilute the fuel gases tn the desired heat con-
tent. A manifold accepts the liquid nitrooe'i frrm twelve 3500 scf
bottles of gas. Three atmospheric convective vaporizers (Cosmodyne Model
SV-2x4) each capable op" vaporizing 3000 cu ft/hr supply the qas«?o>'s
nitrogen.
3.2 Flow Control Metering
The very wide ran^e of flow rates (2570:1) cannot be controlled and
measured by a jingle valve and flowmeter. Valves and flowmeters typically
have maximum usable ranges of less than 10:1. Hence, parallel systems with
four valves end flowmeters are used to contro". the propane, methane, and
nitrogen flow rates.
Figure 3-1 shows the common design of the flow systems. The gas supply
pressure is -egulated to less than 45 psig by an air-loaded pressure reducing
A3-1
-------
u>
C
I iquid Propane
>
j4xh2!h><^
Liquid
Nitrogen
Ratomete;
[=133
Ori Plus
F, a re
IXh-
Stack
[23—{XK
Natural
O—1>0-~
DKH-n
Oil 7 V 15
ixh-
Bo1!er
r1arehead
°1 lot
Sample
Dr/ers
Figure 3-1. Fuel flow control and metering system schematic.
-------
TABLE 3-1. PILOT-SCALE FLARE TEST CONDITIONS*
!) rJ.
Tn.
C^sp
Velocity
(ft./wc)
1 lew
'l.l'r
(CFH)
< 1 do
Vr>
¦'iiset)
Reyriol ds
Number' (d)
un
Ri< hardson
tier
Propane
Flow Rdti-
(!b/hr)u)
Propane
Cost
{$/»<¦)
Methane
Flow Rat«.
(Ib/hr)(b)
Methane
Cost
(i/hr)
Nitrogen
Flow Pate
(i u/if>; i. >
Nitrogen
Cost
!J/l-r)
*
0. j
88
50n
334
32..'':
, ;
I..1
3.64
0. S
5.5
1.13
3.3
'i
7.U
353
iZ5
13 37
;.rin
73
4.1
14.6
7.4
27. 1
4.63
i.
1767
7?
668/
o.oai
113
?0.0
73. U
17.0
110.6
73.27
4
1.0
707
500
1 337
16.1
<5
s.c
79.Z
4.8
44.3
-3.3
6.0
r>
7.83
7000
177
17R5
? . CI j
!2.,6
5000
535
805
36
6.1
73.il
;.9
35.1
7.-1
12.0
R
?.o
5655
500
53C
3 .)•>
36!
63. B
?33. 7
38.8
354
.'4.2
9
a.o
MJIO
750
J 06V 9
7??
177.7
167.7
:fi,6
718
148.E |
(a) Propane diluted ti> 11SC Btu/ff' 'S6 \iolun« ')
(b) Methane fired without dilution
(c) Nitrogen used to dilute profane to 300 Gtu/ft^ (87.5 volume }
(d) P?ynolds ntimh^r based on 561 propane. 44?. nitrwien mixture
* Based on original test matrix; see section 4-1 for additional test conditions.
-------
valve which is remotely activated at the control panel. The flow rate is
controlled by manually adjusting one uf four valves and Is measured with
or.e of four calibrated square-edge orifice plates. The differential pres-
sure signal increases with the flow rate and is indicated with a Dwye" Mano-
meter (Model 422-23). The metered gases are combined in a header which is
plumbed to tne flare stack.
3.3 Specification, Performance
Table 3-2 l3-4
-------
TABLE 3-2. FUEL SYSTEM COMPONENTS
CCMPONENI
DESCRIPTION
"OCEl
«:ze )p. ~j>ac:
PHI
pr:»
PflZ
VI
V2
V3
76, VIG, VU
V7, <11 , V15
va. VI?, Y'6
9, VI3, »17
ORF 1 to 9
0RC 1 to 9
. p?. ^
MNl. MN2
svi. sv2, sr.
1 FR4
I
I
! wi
) ta RN4
1
1 V'B
V19
Nmm P-essure Re'-uUtoi
»if09«n frrwtyrt Rtyjlator
Natural Sas 'ressur* Regulator
Propane Shut-Jf' Valve
Mtroq«t> Shut-G*f 'V»l*«
Naural Gas S'.ut-C'ff Vjlw
Flow Control Valve
;p""opa«a, n frogen,
i»s)
Floy Control \fa1*t
(propane, nitrogen, n. fas)
Flo* Cantrc' Yalvt
.Jrcpjne, nitrogen. r. gas)
Mow Control Vilvt
!>.':t>ans, iHragen, r. gas;
Cr1'i;4 r'Beiges
.jrooane, nitrogen. r. gai)
Orifice ?:At;s
^prepani, nitrogen, n. gas)
F 'essjre Sauges
!B»coant. nitrcgne, n. j#s)
*ancneter
SoUnc'd t-iergencv
iprcoina, rMtrcqen, n gas)
Ancillary Propane Pressure
Regulator
"ilot P"OPfir>d RlTuTl^ter
Fuel Mota.r»te,~5
(propane, ni..rc5en, n. ga")
Ancillary P'-jiane. Shut-off
Pilot ?r^gi-,jm Flow Contrc'
/alve
Le-iie 3F4K-4
Leslie S!'*K-4
lesl<« 3PAK-4
Worcester $^ll-«
Ucrrestir 5BU-*
Ijnktnheiire' 7011-Hil
Lun*«nh»ii!»r 7Z-'S
Lunktnhamr 907-85
Ljnke'ih.'-isi^r 3C7'5S
.unkanhi-aier 907-$S
Ldniels 3Q-RT
Custom
Harsh Oih'',» Gauge
fcivtr 422-23
ASCC iZH-Ba1
ȣi!J IMF YN
[V yer RMB
Srocts R-2-15-AAA
Worcester 5811-R
Units/ 5S-2RS4
1 • S?T
1- NPT
3/4" flP"
2* MPT
I' HPT
2 ' :i»t
!-!,2" NFT
3.4- srr
^pr
1/3" HP"
2-1.2", I-'.:
\;l" HPT
See Icp-.rc^ i
"aole A-i
0-5j asig
5" O'i'
¦3-J3"
2" NPT
3 ¦ 4 ' SPT
1-1") 5CCH
l-'O SC-F
3/4" NPT
CV 1 J.15
-------
4.0 TRACER SYSTLM
^"sured amounts of sulfur dioxide *»ill be added to the fuel gases as
a tracer to determine dilution factors and verify material balances. The
actual 5^2 addition rate will be determined during experimentation. However,
1t 1s estimated that one voluira percent S0X in the fuel will be sufficient to
produce about 1.0 ppm in the combustion product. The rate of SOj anticipated
ranges from 0.11 to 283 SCFH.
4.1 Tracer Supply
The SOg is purchased from llnlor Carbide as a liquid compressed gas cylin-
ders containing 900 SCF. A dip tube draws liquid and passes it through two
steam-heated shell and tube vaporizer* shown 1n Figure 4-1. After vaporiza-
tion the SOg is controlled and meter-delivered to the fle.r^ head. Figye 4
shofcs the flow schematic of the SOj system.
4.2 Tracer Flow Control and Meterirg
The wide range of flow rates (2570:1 ) requires using multiple val/es and
flow meters. The SOg pressure (34 psig) from the vaporizer is regulated to
15 psig with a stainless steei regulator (Union Carbide Model CRC-15).
Depending on the use rate, the SO, is controlle
-------
P1 BO
T».Iw
*.onu -vv
\
\ i" SCM4R
\ \
- *_i
—( )
Figure 4-1. SO^ vaporizer.
'team
Out
-------
team
V4
RN1
EXHZZ33
3
V5 RM2
Vf
RM3
—C
Rotometers
Flare
¦*- Heao
SVT
Figure 4-2. S0? flow control and metering system schematic.
-------
TABLE 4-1. TRACER FLOW SYSTEM COMPONENT SPttlFICATION
COMPONENT
NO.
DESCRIPTION
MODEL NO.
SIZE OR
CAPACITY
REMARKS
VP1, VP2
Vaporizer
Custom
10 CFH
Steam Heated
VI
Bottle Shut-off
-
"
0»i Bottle
V2
Ball Valve,
Stainless Steel
Worcester
1" MPT
Main Shut-off
i
l
PRi
Pressure Regulator
Un?on Carbide
CRC-15
1/4" NPT
10.1-15 psfg
PR!
Pressure Gauge
Marsh Type
100-3
6" Dial
0-60 psig
Stainless Steel
RM1
Rotameter
Brooks
1110-09H3B1A
0-;27 CFH
RM2
Rotameter
Dwyer, RMB
0-20 CFH
RM3
Rotameter
Dwyer, RMB
0-5 CFH
V4
Control Valve
Whitey, SS-4RS6
CV = 0.51
V5
Control Valve
Whitey, SS-2RS4
CV - 0.15
V6
Control Valve
Whitey,
SS-21RS4
CV = OOP"
5V1
Solenoid Valv
ASCO, 8211C87
1/9"
F^st Shut-off
-------
5.0 STEAM SYSTEM
Steam is requited at the flare facility for simulating steam-assisttd
.'lares and for auxiliary heating. Steam Is injecteil into the flame at the
flare head to increase air entrapment ani combustior ir,tens*> .y and reduce
luminousity. Typical injection rc:tes are from 0 to 1 pound oi: steam for each
pound of combustible gas. SteonTi is also uSi?d ar the flare tacilitv to
vaporize S0^ for use as J tracer and to heat sainple lines to avoid condensa-
tion which would cause loss of sulfjr and hydrocarbons in the lines.
5.1 Steam Supply
Steam is supplieJ to the flare test facility a 15 hp boiler.
The boiler, which can fire propane or natural gas, supplies 400 pounds of
saturated steam per hour at 100 psig.
The feedwater to the boiler is filtered and the calciien and magnesium
ions are .-P-moved by a resin-bed ion-exchange water softener. The wa'et
softener has e one-cubic-foot bed and can process 10 gallons ~f wster psr
minute. The s-psin bed is automatically regenerated Dy salt solution.
5.2 Flow Control ana Metering
Metering oF steam is difficult because ot the wide range of flow rates
and the problem of metering a gas which may condense at thr metering tempera-
ture. Figure 5-1 illustrates the steam-flow control and metering system.
All lines are insulated to minimize heat losses. The boiler produces sat; ra-
ted steam at 100 psiq (348°F). The steam flews tornugh a pressure-reducing
valve (Leslie Model GPK-1Y ) which controls the -.team pre=:ure between 10 to
^0 psig. Expansion produces dry superheated steam. Condensation rornred
downstream of the pressure-reducing valve is removed by a steam trap at the
manifold and the "dry" steam flews through one of four orifice meters and
is cont.-oiled by manua"! valves. rlo* :*ate is indicated by the pressure
drop across the orifice meter mes?;' .,y a differential pressure transducer
(Val idyrie Model P3050). Specia1 _\.ons arc used to read the differential
pressure in the steam line. Bleed-vr.lves clear the pressure sensing lines of
liquid before ea^h reading. The manifold pressure and temperature are mea-
sured on dial gauges. Table 5-1 lists the specifications of tno metering
system.
A 5-1
-------
Bo1 le»"
T1 © Orifi
I Plate
©T- -L
-------
TABLE 5-1. STEAM METERING SYST7M COMPONENTS
COMPONENT
SO.
MODEL NO.
DESCRIPTION
S'ZE OR
CAPACITY
1
R«WRKS
VI
Lunkenheimer
LIJ-602
Gate Valve
2"
Shut Off
PR)
Leslie
GPK-2
¦ re>sure (.educing
Valve
2>
ST1
Flexitrap
Stfiir Trap
Mr
PI
Marsn
Ha star Gd'jge
Pressure Sauce
6' Dial
0-2C0 ?sig
OR? 1
Oanials 30-RT
Orifice Fla^e
2-1/2" NPT
ORF 2
Oamel^
Orifice Flange
3/..' NPT
0Rr ;
tianleis
30-RT
Orifice Klanqe
1-1/2" ,IPT
ORr 1
OR" 2
ORF 3
ORc 4
V:
Vl'
•I'-
ll
?*1
VMl
^ustac.
Custom
Cjstsin
Custcn
;11
1 Lunken^eimer
i lunkenneimer
i 907-3S
i Lunkeni^eimer
I 907-3S
j <)!'1)'f
I '3050
Trend
Newport
Orifice Plate
Orifice Plats
Orifice Plate
Orifice Pl.ite
Ccntr-ul Valve
Control Valve
Control Valve
Control Vilve
ii'fferent.ial
'recurs
T ransduce'-
"emp^rdture
Transducer Outpi-v
7ol;ags i'aadoit
Borsi
0.987"
J. 159" Bore:
0.205" Bore!
0.0951 tint?
1 -1 / ?" NPT
I" NPT
3/4' NPT
•/4 NfT
0-5 psid
3 DU;
5C-5CO
0-11 o?
Vu'C
0.01 YOC
Vesciut'on i
A5-3
-------
- •"* Per 'ormai ce
The steam syste.i operates satisfactorily. Figures 5-2 through 5-b show
tue calibration curvd. if the orifice meters. Flows were calibrated by con-
densing the stf.air downstream o*T .he control valve and weiqhinq the collected
condensate over a t:imed interval.
AS-4
-------
IP
P - 20 psig
T - 250°F
0.095" Orifice
8/17/82
20
30
40
50
10
0
'P, inches (WC)
Figure Calibration of steam flow.
-------
90
80
7P
60
50
40
30
20
10
0
0
T
P - 10 psig
T " 265°F
0.205" Orifice
8/18/82
20
ap ("wc)
30
40
Ficjurc 5-3, Calibration of steam now.
-------
180
0.459" orifice
P = 10 psig
T - 250°F
160
140
120
100
au
40
8/18/82
AP ("WC)
Figure 5-4. Calibration of steam flow.
-------
360
P = 10 psiq
T = Z50°F
0.987" Orifice
320
280
24C
16C
120
80
8/18/82
0
10
12
14
16
18
20
2
4
6
0
8
AP ("WC)
Figure 5-5. Calibrated steam flew.
-------
6.0 DESIGN OF FLARE HEADS
The flare heads should:
• be geometrically sample to allow a scientific
interpretation of the results;
• simulate the important features of commercial flare heads;
¦ produce flare flames representative of coiunercial flar°
flawes;
• be consistently scaled.
A straight pipe was used for the flare heaus on the advice of the Technical
Advisory Committee. They concluded that flame arrestors, retention rings,
and other proprietary features of corraerclal flare heads were
unnecessary and undesirable for the flare test facility. The head
vas provided with a simple steam-ring and steam-nozzles for injectinq stearr
to suppress soot.
6'1 Flare Head Support Base
Different flare heads can be mounted on ths flare base shown in Figure
6-1. The base Is a 6-foot long, 6-inch diameter pipe welded to a platform
composed o* => inch thick steel plate and four legs. The gas inlet is a
2-inch pipe ending with a standard pound flange. The mount for the . oui1
flare heads is a standard 150-pourj, 6-mch pipe flange. The top of the fl ire
head support bas*» is 7 feet =.'.ove grojnd level,
6•2 Flare Tio
The flare heads are norrinal\v J, 6 and 12 inches in diameter and each is
8-feet long. An illustration is shown in Figure 6-2, and the dirner.sicns are
1isted ir T^ble 6-1.
6.3 Steam Ring and Nozzles
A key feature of commercial flare neadi is the steam ring and injection
nozzles. These are designed to induce air e'ttraiwrent arid turbulence at the
base of the flare flame. The design principles employed in cwmercial flare
heads are proprietary and unavailable. Hence, the design of the steam injec-
tion nozzle assembly for the teit flare was based on empirical infrv.-ma
tion reported in the literature. The basic assumptions applied to the steam
A6-1
-------
A 6" Weld Flange
Raised face, 150*
Q 6" SC4 40 Plpp.
6* Long
C 2"x 2" H-Beam
tc 4 Comers
D t" h'"5ld Flange
Raised Face. 150#
E 2" Black Pipe
SCH 40
F 1/2" FMck Plate
3'x 3'
G 2" Pipe Leg
1' Long
H Footing Plate
1/2" Thick
4"x 4"
Ficure 6-1. Flare head support structure.
A6-2
-------
Flar^head Nozzle
Steam Injector
Steam Manifold
Steam Injectors
— Positioning
Mechanism
Strarr, Fnlet
Steam
Condensate
Bleed
"•gure 6-2. Flare Head Design
A6-3
-------
TABLE b
PARAMETER
Inside Diameter (inch)
Outside Diameter (inch)
»- tow Area {in )
7
Manifold Flow Area (in")
'lumber of Nozzles
Nozzle ID (inch)
Nozzle OD (inch)
Length of Nozzle (inch)
Vertical Position
(inches from tip of flarehead)
Total Flow Area (in")
Overall Height (feet from ground)
FLARE HilAD SPECIFICATIONS
Nomine I Size (inch)
3
6
12
3.06
6.3
12.3
3.5
6.6
12.8
7.35
31.2
118.8
3.06
14.1
33.1
4
8
lb
0.263
0.37
0.525
0.375
0.30
0.625
3
5
7
0.5-4.5
-1 to 5
0 to 6
0.217
0.660
3.464
15
15
15
t
i
-------
nozzle design are:
• Commercial flares have smokeless capacity of about 20 percent
of the maximum gas capacity. This yields a smokeless exit
velocity of about 120 ft/sec.
• Flared gases have densities of 0.0637 pounds of hydrocarbon
per cubic foot.
• Smokeless operation requires 0.27 pounds cf steam per pound
of hydrocarbon.
From these assumptions the stearr flow .opacities are c(moused and the steam
ring and nozzles sized. To simplify the analysis of steam jet entrainment,
straight stainless steel tubes with large length-tc-diamcter ratios are used
fo>- the steam nozzles. Figure 6-2 shows the flareheads with the steam ring
ar.d Table 6-1 lists the specifications of the flarehead, steam rinqs and
nozzles. Some steam condenses in the steam ring in spite of line insuUtion.
The steani ring is drained through blow-down vilves prior to tests with steam
injection to eliminate the injection of water into the flame.
6.4 Pilot
The piiot is required to ignite the fuel oases at the fUrehead. Typi-
cally, t.he fuel gases will remain lit unless the exit velocity is too high,
the wind is too strong or nhe gas r»eatirg value is too low. Under such ren-
ditions the pilot will provide a reliable continuous source of ignition.
i.oi:ii!iercl31 flares burn about 125-200 scfh of natural gas continuously as
pilot flames. This is similar to thf fl:c?1 flow for the main 'Tare for many
the tests. In these cases, use of <. continuous pilot flame would ubscura the
major objective of this
-------
nitroren before and after each test and it would distort the velocity
profile at the exit of the flare head.
6.6 Flame Retention Ring
Retention rings are us.-d on flare tips to stabilize flari flames at
high velocities. The rings are usually "f special proprietary design and
are net required fo- sniall flares with low exit velocities. Therefore, tiie
Technical Advisory Committee recommended the pilot-scale flares be operated
without a retention ring. (Note that in subsequent tests at high velocities,
retention ri.igs were used.)
A6-6
-------
7.0 EXTRACTIVE 5AHP5.E SYSTEM
In o-1er to determine th'; .jcnihustion efficiency of flares, samples
are extract from the flame olume and analyzed for 0,, CO, C0o, sulfur,
C L
hydrocarbon and particulates.
7.1 Sample Prints
The locations for sample withdrawal must be determined experimentally.
A hood placed above the ^!ar« flame collects the entire plume and a five-
probe rake withdraws samples at selecteu locations above the flare
flame.
7.2 Hood Sampling
The hood collects the flame plume into a chimney and homogenize* the
cotnhjstion products. If this method is successful, it 1s a quick way to
-w-ter.'ine overall flame combustion efficiencies for * all flames.
Hood Design
Figure 7-1 shows the sampling hood assembly. The hcod desirn was
improved several times during the program in order to better mee4: the program
objectives:
• To collect the entire flare-flarne plumes from the 3" flare head
and the plumes from a flame at 2.8 ft/sec gas exit velocity for
the 6" flare-head.
• To htiiTnqenize the plume to a mixture with uniform species
concentrations.
» To provide means to measure material balances through the hood.
• To minimize disturbances to the flara flames.
The hood assembly consists of two hood sections, an axial duct booster
fan, a mixing chamber, a flow straightener, a measurement chamber and a flow
damper. The two hood sections are constructed of 12-gaut-e carbor: stpci sno
cover areas 8 feet square and 13 feet 8 inches square, respectively.
The hood converges to a 2-foot diameter chimney. The 2 hp axial flow
booster ^an at t.he entrance to the chimney t.an draw 8200 CFM Df air at static
pressure of 1/4 inch water column. This fl^w rate 1s about 250 times the
A 7-1
-------
Flow Damper
T
18-1/3
Measurement
Chamber
lew Straightener
'
4-1/
24-1/4
D'ji;t Booster Fan
Upper hood
Sec "ion
T
Lower
Hood
Section
13'8"
Figu-.e 7-1. Sampling Hood Assembly.
A7-2
-------
flare ga*. rate at 2.8 ft/sec from the fi-.nch flarehead (equivalent to about
17 times the plume of a 56 percent p-opane flame at stoichiometry combustion
conditions),
The mixing chamber is a 4-*eet lcng, 2-feet in diameter p;pc,- with 9 flow
directors to increase turbulence mixing. However, the duct booster fan
changed the flc* pattern in tMs chamber to highly swirlinq. A flow-
straightener at the exit of the mixing chamber was used to reduce the swirling
flow. The stralghtener made of two banks of 2-inch wide stee^ bars at
two-inch spacings oriented 90 degrees from each other.
The measurement chamber 1s 2-feet Is, diameter and 4-feet 1on«j and houses
the sample probe, five velocity probes arid a type-K thermocouple. A damper
valve at the chancer outlet allows adjustments of the flow rate through the
hood.
Pro^,e Design
The hooo probe is shown in Figure 7-2. It consists of a center sample
passage trade of 316 stainless steel, 1/8-inch, schedule *0 pipe. An exterr.al
flow jacket, 1%-lRCh 0D, carries steam to maintain the gas sample at the
desired temperature, A stainless s;eel filter holder (Gelman 1209), welded
to the probe tip, houses a glass fiber filter to collect solid part cuiates
The particulate loading of the combustion orcduct is determined by measuring
the combustible material captured on the filter after a known volmre of
sample has passed through the filter. The probe is desiqned to Dull 50 cu.
ft. of gas per hnur. Various-sized tips are used to ensure that the sampling
is nearly isokinetic.
'•^ Multiple Point Sampling
The hood sampling system provide? average properties of the combustion
products. Mu1tiple ooint Sampling determines local species concentrations and
identifies the contribution of individual flame structures to the overall
plume. It uses five separate probes ar.j sampling trains tc collect five sample-:,
simultaneously.
Support Gantry
The multiple proues are supported by a gantry which can be lifted by a
12-volt DC winch. It covers sample heights from 6 to 50 'pet above ground-
A7-1
-------
• 1/4" Pipe.
-------
level and can traverse a cross-section 15 feet by L5 feet. Figure 7-3 shows
the prote positioning mechanism.
Probe Design
The prob»: used for the multiple point sampling ?re the same as the
probe used in hood sampling. Refer to Figure 7-2 for the probe design.
7.4 Sample Conditiens
The sample collected by the probes cannot be analyzed directly because
it is likely to be too hot, laden with particulate and wet. The conditioninq
required depends on the specific property that is to be measured; namely,
corcentrations of Qg. CC, CC^, SO2 (or total naseous sulfur), hydrocarbons
and particulates. These concentrations can besi. be measured if the gas tem-
perature is near 80°F and is fret of particulates and moisture. Figure 7-4
shows the schematic c,f the sample train.
The sample temperature ranges from n«»ar ambient to over 1000°F d«?WTiding
on the sample location. Ideally, the moist samples should be maintained
between 160 and 300°F so moisture does not condense which would remove a
purtion of the soiuble gases. The sample temperature of th« probe- is con-
trolled by passing water or steam tnrough outf:r jackets. 7he moist sample
is kept above 200=F by electrically-heatea Teflon sample lines jut drops to
ambient levels down tream of the driers
Particulate Removal
The probe tip- coi-t&in the filters in stainless steel holders, to
collect particulates, keep the jample line clean, and reduce the contact time
between the rarrple gas o.id the potentially resctivs particulates. Combusti-
ble particulate are determined by burning the filter and sample. Filters
(Gelman GA Triacetate material with pore sizes 0.2, 0.45, C.8, 1.2 and 5 ^rn
are adequate for temperatures below 212°F and appr^ch velocities from 1.64
to 16.4 ft/sec. Otner glass-fibre fil__rs are adequate up to 930CF (Whatirsn
934AH), but sho.ld not be necessary.
Drying
Nocture in the gas sample must be removed before it condenses and par-
tially reroves the soluble gases such as SO.-,, In addition, water vapor inter-
L.
feres wi'zh infrared analysis of CO a,id CO^.
A7-5
-------
Pulley
To Pulley
Hood
Traverses
Vertically
Hood
Probe
Rake Probe
Traverses
Vertically &
Horizontally
Rake Probe
Positioning
Mechanism
Figure 7-3. Hood and probe rake positioning mechanism.
A 7-6
-------
liake
,1 Dryer
flow Meier;
ot* 1
I How Mete
in rump
:r^Lois-
Drver .," How Hetei
How Meter>
Wei. Sarnie
Bypass
Bag Fill
W*t Sair^
Bypass
I'artirulate
CM rcoj I
Vent
1
I'ump
-~•CZ3-
U--
Spell I
low Meters Bag Fill
:T5? ii ' ¦¦ i'. 'I !?/—i
Met Sdn^l
Bypass
Pwhd
*
, . Nn in
kake 2
Mow Meters
Dryer Pump How tlctrrl
-ito*
S3—-Q^-
Total
cz>
L-tx^
Big F i11
«- Wet iompie
Bypass
Span
"K
ieri, ri? J
M>4«» Bag F i i
&
Bag
Purge
h'ood
Saw I;
Fotal
r,ow Meter
Flaw Meter
>•€=~—
Flow Meter
typ-v.
flow ?V»ter
Orllacp '«•"} 3y|T
Bypass I HC
! rI0
Air
In
0,
tayli
, j ^ r
J I 3160 17
t;haust
CO
llril
b~
Figure 7-4. .' Iare Sample System
-------
"ihe moisture is removed by Permapure driers (Model PD^OOO-ZASS) which
uses a membrane tube bundle to transfer selectively w»ter vapor fron the gas
stream. The water passes inrough the "nembraiie where it is removed by dry
nitrogen. The driers have 40 cu cm ir.t^rnal volune and can pass 9.95 to
51.1 cu. ft. of sample per hour at 0.15 and 0-75 psi pressure drop, respec-
tively. The amount of purge-gas required is 117 cu. ft. oer hour. The out-
let sample is designed to achieve *o3F dew-point.
Continuous Analysis
Concentrations of CO, CO.,, 0^, 502, HC in the dry, clean gas are me?su,-ed
'uy continuous analytical instruments.
Ban Collection
One set of analyzers is available, consequently samples from the multiple-
probe sampling will be simultaneously drawn and subsequently analyzed. The
samples will be stored in 1.8 cu. ft. capacity Tedlar baqs. Tedlar bags qre
fabricated on-site using a heat sealer. The bags include stainless
steel valves tnrough whicii the gas will enter, be trapped in the bag, and
exit for analysis.
7.5 Analysi s
The gas samples are anai/zed fur concentrations of C^, CO, CO^, SO^ and
hydro:arbons and the particulate filters are weighed tc determine particvl&ti
loading in the sample gas. Table 7-1 lists the analyses, .ensitivities,
accessories and typical ranges of gas concentrations.
A7-8
-------
TABLE /-I. ANALYTICAL METHODb AND A^uRACI'S
SPICK
INSTH^CNT
PRINCIPLE
RANGLS
" ¦¦ —
accuracies *
MlASURtO
CONC EKTRAT1ONS
°2
Taylor 57PA
Paramagnetic
U loot
•0.02%
Typical - 20.St
Ring# I8-2U
CO
Bcckman 2isA
Non dispersive
Inf rarpd
Absorption
0 500 om
0-2.01
tO* turn
>200 poa
Typical 3-10 ppn
Minimum 3 ppm
Hajtlaui 3TO ;>j*n
CO,
HeckMr 31
Non-dispersive
Infrared
Absorption
u-s.in
0-101
0-2"*
•0.021
•a. it
.'0.21
Typical 0 7-0.5i
Minimum 0 071.
Maximum 1.01
HC
Beckr.ian 4£ki
("1 sine
Ionization
Detection
0-5 ppm
0-5C pp"
0-500 ppra
0-5000
•.05
0.5 ppm
>5 ppn
i50 ppn
"yplca. - 3 tr. 10 ppn
Klninum - J
Haximun 200 | cm
SO,
«eloy SA 260
T i trilion
nan?
Photometric
Detection
Reaction with
Perthi orato
O.b ppb
tc
10 ppn:
1 tc 100 ppai
lit Of
measured
;10X of
measured value
0.5 - 10 p>>u,
Particulate
Fi Iter
T lined
Collection
0 to ID"7
ib/ft3
:St Of
neasiirer1 value
0-5 * 10"7 lb/ft'
* 'jhort-tem accuracies within 20 utmitei of ir'trtmenta1 calibration
-------
B.O VISUAL MONITORS
The* flame behavior is 'ecorded by a video recorder, a sti 1T camera
and a high-speed,16 mm movie camera.
8.1 Video System
The i/ideo system has been used to monitor ord r.-sco'd gross-flame
structures, fi. JVC color camera records the gross-flan*:- cna>acteristics on
tape. The recoruo"- uses c 3/4" wide magnetic cassette tape and provides
frame-by-frame ploy-back. A 12-inch diagonal color monitor permits online
monitoring and playback. The video monitor system is limited by its spatial
and temporal resolution and narrow range of sensitivity to light.
8.2 Photography
A still Camera is jseo to record the structure of the flamps.
Long time-exposures will record the average structure ami short time-exposures
hill record the instantaneous structure of the frames. Both exposure are
recorded with a 35 irni s i ng 1 e-lens-ret le?, canera (Cannon Model AE-1) with a
50 mm focal ""^ngth and an 80-2QC nrn zoom lens.
8.3 H1ph-Speed Cinematography
High-speed motion pictures help in understanding tne evolution of
flare-flaire structures. This information is used t' understand t.ne influ-
ence of flame structure on the combust'on efficiency and ocale of the head.
The flame structure is recorded with a 16 m Hycam camera using a rota-
ting prism and lens with focal length of 10 to 100 mm. Motion pictures can
be taken with this camera at speeds of 100 tu 1000 frames/sec. Experience
has shown that 200 frames per second is adeqjate to record flare flame struc-
ture .
A camera platform rrovides stable positioning for recording the flame
structure (Figure 8-1). The platform is 15 feet above che ground and can be
positioned up to 30 feet fron ;he flarehead. The 30-foot distance produces
acceptable flame resolution wher. using commcnly dvailable lenses.
A8-1
-------
Operator
¦ Back
Support
15'
7
8"
Figure 8-1. Camera platform.
AS- 2
-------
9.0 AMBIENT MONITOR ANi) CONTROL
The flare-flame is not enclose! and is subjected to environmental varia-
tions in wind velocity. Screens =ire used to attenuats natural wind. The
wind speed and anbient air temperature are monitored during tests and no
tests are conducted during high wind conditions.
. . 1 Windscreens
Perforated steel panels, 8 feet by 10 feet protect the flame from
wind. Eighteen paneis cover the four sides from 10 to ,?6-foot levels.
Figure 9-1 shows a typical wind-screen panel. The panels have 22.7 percent
open area with one-half-inch diameter holes on one-inch triangle oitch. Ob-
servations using smoke bombs showed that the windscreens effectively reduce
the influence of the wind on the flare-flame.
9.2 Monitors
Wind speed and direction ars monitored by a three-cup anemometer and
wind vane (CIimatronics Model Mark 1). "he velocity is recorded Dn a strip
chart and dry-bulb thermometers use^ to measure ambient air temperature.
A9-1
-------
2" x 1/4"
Reinforcement Bar
OO
11 Gauge
Perforated
Steel Plate
22.1% Op*n
1/2
10'0"
Figure 9-1. Perforated plate windscreen.
-------
10.0 SUPPORT STRUCTURE
The structure around the flarehead provides support for the hood, the
rake probes and the windscreens and acts as the position reference for the
flame and samplinq probes.
10.1 Overal! Structure
The structure covers an area 20 feet by 20 f«?jt. Figure 10-1 shows the
schematic of the structure. The corner poles are 4" x x h" tox-tubing,
the horizontal beams *re 4" x 4" x V H-bean.s supported by diagonal
members. The structure 1s stabilized by 3/8'-diameter steel guy
cables attached to the corners and anchored to the ground. The vertical and
horizontal members are marked by paint stripes spaced at 2-foot intervals to
provMe a reference for flame observations and probe positions.
10.2 Hood Support
The hood, as shown in Figure ?-I, weighs about 2500 pounds. It is lifted
at four points by 3/8"-oiarneter wire ropes. A system of pulleys auidos the
wire ropes to one comer wnere the wire rcpes are bridled together. An slec-
tric winch is used to rais? the hood , which can reach 50 feet above the ground.
10.3 Rake E robe Support
The Rake Probe Gantry weighs about 1200 pounds. It is guided at one end
by trolleys and a longue-in-grov arrangement, at the othe-. It is iifted at
two points near tne ends by 3/8"-ditcr wire rope. These are guided through
pulleys to one end and are bridled together. An electric winch lift-; the
probe, which can reach 50 feet above the ground.
10.4 Windscreens Support
The windscreen panels weigh e jroximately 180 pounds each. Wire ropes
hanging from pulleys at the top o* the structure have thimble-loops spaced
at 8-foot Intervals. Hoo!-s at the top of the panels secure them to the
thimble locps. The windscreen assembly on each side of the structure is
lifted end lowered Co the appropriate height by electric winchos.
AT 0-1
-------
Ladder
Lag
Screw /
Y
Figure 10-!. r,:pport structure for sampling hood arid rake probes.
AT 9-?
-------
11.0 DATA SYSTEM
Data arc! recorded manually and with the aid of a data logger, and strip-
chart recorders.
11.1 Data Logger
A data-logger (tsterline-Angus Model C2Oi!0) is us?d to scan and
print up to 20 valtago signals at 2 channels/second. Iris used primarily
to log the t..inper=tures of +.ha probe, structure, and fuel gases.
11.2 Strip Recorder
A 6-pen ci srt reorder (Soltec Model 3306} continuously records the out-
put of the Og, CO, CO2. HC and the SO,, analyzers. It also records the wind
speed nisasured by the 3-cup an^nometer.
All-",
-------
12.0
Orifice Flc\;meter calibration
The orifice plates were calibrated using air arid calibrated laminar
flowmeters. The apparatus used to calibrate the orifice meters is shown
in Figure A-l. The flow rate through the- laminar flowmeters and orifices is
adjusted with a control valve arid the following data are recorded:
• Pi - pressure at upstream tap of laminar flowmptpr
• ?2 pressure at upstream tap of orifice flovuneter
• ^5P.j - differential pressurs across laminar flowmeter
• -ipg - differential pressure across orifice flowmeter
• T - air temperature upstream cf laminar flowmeter
The flow rate of air is calculated from and T using the equation with
the following form:
Flow Rate - {a{4Pt) + b(4P, ]" ) x (P1+14'7,|J / 530
1 1 ' \~~wt j \wr^
where flow rate is in SCFM (1 atincv, 70°F)
P1 - is in inches water column
P - is in psig
T - 1s 1n °F
Table A-l lists the calibration constants for the laminar flowr.at-.'.rs. The
calculated characteristics or" the orifice flowmeters usr? the following
equation- r
_J /P + 14.7
q ¦ K x jjy- x F gas Xx/h.^ Q x\J 7 + 460
where
q = SCFM
A12-1
-------
TABir. A-l. CALIBRATION CONSTANTS OF LAMINAR FLOWMETERS
MODEL
SERIAL NO.
a
b
FLOW 9 8" W.C.
/atm, 70°F (SCFM)
50 MY 15-6
S-4291-1
51.27
-0.1479
381.5
bO MU 20-2
R-3049-1
10.0125
-0.071875
75.5
-------
K = orifice flow calibration content
MW - molecule weight of gas
Fgas = correction factor for gas other than air
Fa1r = 1« ^propane _ 1-I4
h^g ~ pressure drop in inches of water
P = aressure of gas at orifice, ps-'g
T = temperature of gas at orifice, °F
Table A-2 lists the orifice. flow calibration constant, K, of the orifice
flowmeters.
Al 2-3
-------
TABLE A-2. ORIFICE FLOWMETERS CALIBRATION CONSTANTS
dSIFtCE
NUMBER
UPSTREAH
10 rr.ch)
30RE
U!Ac. *ER
' '-'C
FLOh
FACTOR, K.
STANDARD
DEVIATION
OF K
1
FLOW Of AIR
0 20" WC 1
1 llCT, 7Q"F |
(SCFM) 1
|
n
2.463
1.91
7366
26f
:95.0 |
FJ
1.610
1.03
laof
4:
72. J i
F3
0.824
0 52
460. S
5.6
18.4 !
F4
0.821
0.2w
120.3
3.7
4.82 1
F5
2,469
1.58
5041
17
roi.s |
F6
! .610
0 87
1195
35
47.8 |
F7
i.
0.44
310.7
• J . «r
12.1
U. 4
o.-?
73. r
2.30
2.97
F9
2 .69
1.92
',307
"15
293 C
-in
,510
1.04
1821
36
72.9
F1 :
0.824
0.52
470.3
*5. 3
18.3
*"L
0.321
j.'fl
135.2
10.3
5.41
- K x
NW
, '? ~ 1 J. 7
x\r-tzkt
Flo# • '("!"»
h 1 Inch or" Water
0 ps i a
T ¦ •" F
*U = Molecular neight
AlZ-'r
-------
APPENDIX D
TESTING METHODOLOGY
The methodology used in testing is discussed in Appendix B in sections
on:
• Background concentration determinations
• Test condi tio--";
• sample collection
• Analytical procedures
• Visual observations
o Data reduction
Background Concentration Determinations
During the early phases of the te't. program, the background concen-
trations of 02, CO, CO2 and HC were found tc be conoarable with those from
th« fare flame. Hence, local comburtion efficiencies need to be corrected
for tlie concent.\sifcn of the species in the Lackground. To accomplish this,
the concen.ration of the species in the background is determined prior to
each test.
The background samples are collected and analyzed following
-------
TA3LE B-l. BACKGROUND SPECIES CCrfCtNT RATIONS
SPECIES
MEAN
MINIMUM
MAXIMUM
°2
2U
212
21%
CO
2.2? ppm
0.0 ppm
6.7 ppni
C02
694 ppm
519 ppm
927 ppm
HC
3.17 ppm
1.5 ppm
7.15 ppm
Soot
0.50 ppm
0 ppm
1.32 ppm
so2
0.17 prim
0 ppm
3.6 ppm
B-2
-------
Although tot a daily operation, changing the flare head was a siirple procedure
and could be accomplished in three to four hours. The supp'v for nitrogen,
propane, and natural gas were set on a flow system designed to av. d transient
fluctuations. The flow rates of these gases were held ccnsta it during testing,
v.-hich usually lasted twenty minutes.
A small boiler supplied steai." to the 502 vaporizer, the sample prcbes, and
the flare head steam injectors. Steam uitd for the vaporizers snd sanple probes
was a sma^l fraction of the total boiler output Steam flow ts the flare head
was hela constant up to the boiler's capacity of about 400 lfs/hr. Steam
line? were preheated and drained prior to trials uslmi steam injection. The
lines were heated using steam flow until no water was ejected from the nozzles.
This typically required abouv twenty minutes. After heating and draining the
trap, steam flow rate was set by a manual valve end determined from the differ-
ential pressure across a calibrated orifice.
The flow rate of tracer was r.onitored by rotameters and controlled by man-
ual valves. The maximum safe flow rate of SO2 was limited to ? standard cubic
feet per hour by the vaporizer capacity.
Sample Collection
Samples were collected by two methods, For small flare flames, the
sample hood collected the entire plume into a mining chamber where gat and par-
ticulates were sampled with a single probe. For larger flames, samples were
drawn simultaneously by five probes spaced on a diameter ibove the fls -a flnnie.
The sample probes were of a uniform design, as shown in Appendix A. Gas
was sar.pled at about 30- to 45-sr,andard-cubic-feet per hour. A 6-inch-long,
0.250-inch 00, 0.19-inch 10, sUHlecs steel tube used as the probe
nozzle. This resulted in gas entrance velocities of about 50 feet-per-srcond.
This value was usually higher than the plume rs< velocity. However, the soot
particulates are small and the error caused by anlsokinetic sampling is
unimpcr unt.
The probe tip held a glass fiber filter to capture solid particles. The
prcbe; were steam-heated to 21t°F to minimize moisture condensation and loss of
conde.is'oles and soluble ga: species. Moisture was removed by Permapure dryers
and the samples cooled to room temperature downstream of the dryers,
fl-3
-------
Gas samples could be analyzed on a real-time basis for 02. CO, C02, HC,
and S02, but usually were stored temporarily 1n Tedlar bags for analysis. T.ie
method of continual analysis cculd give on-line information. However, experi-
ence showed that the real-time concentration readings fluctuated over a very
wide range due to '.hp intermittent n-itjre if the flare flame. Integration of
the traces could not determine the roricent ration as reliably as mixing the gas
in the Tedlar bags prior to analysis. Hence, the gas sample1: were collected
in Tedlar bags befort analysis, for all tests. Mixin.i w.is accomplished by
manipulating the bags and by nom.al diffusion. The contents of the mixed bags
were subsequently analyzed for O2, CO, CC2, HC and SO2 concentrations.
Sulfur dioxide vas also analyzed by wet chemistry by diverting part of
each sample to a bubbler containing 3 percent hydrojen peroxide solution to
absorb tne sulfur dio>i4e. The gas bubbling rate and bubbling time "'nterval
were recorded for subsequent calculations.
Analytical Procedures
Techniques for determini'iv; gas samples are listed in Appendix A. Those
for soot and SO2 concentrations are discussed in this Appendix.
Soot Concentrations
The procedure to determine soot was modified early in the trials. In-
itially, soot concentration was determined by weighing the filter before and
after collection of a soot sample from a known volume of gas. This procedure
was unsatisfactory because the filter substrates were Fragile and small part.,
were lost to the filter holder. The procedure developed to eliminate this
problem was:
1. Preparation of the filler substrate Ly baking in ar» oven at 700°C
for one hour.
2. Installation of the filter substrate into the Drobe tip.
3. Sample collection.
4. Drying uhe filter on the sample sJecimen at 70°F for one hour in
an oven.
5. Weighing the filter and specimen.
B-4
-------
6, Burning the combustible material from the filter in an oven at
700° for one hour.
7. Determination of the loss in »#eight of the filters on combustion.
i"his method yielded improved results because:
• Chance? of losing part cf the filter substrate ire reduced, because
the filters are protected throughout thi? steps that are critical in
determining sample weight.
• Burning the sample will measure only tha cunbustible products, and
non-combustible meterials, such as airborne soil, will not
contribute to the weight loss.
• The only disadvantage is destruction of the solid sample.
For calculations of combustion efficiency, the gas and the solid must
be converted to a consistent set of units. This is accomplished by
converting the soot loading to the equivalent parts-per-million of
gaseous carbon. The concentration of the solids in the gas sample is
determined as pounds-uf-solid per standard-cubic-foot of gas. This is
converted to peund-moles-of-carbon per cubic-foot of gas by:
lb carbon _ lb carbon . 1 lb role carbon _ »• lb moles carbon lr,
x -3 x * 1Mb carbon ' U ^3 1 !
Tne mole density of an ideal gas at standard conditions 1s:
. P _ 14.7 _ lb ;Tio 1 es _ o r,1t .. ,n-3 lb moles
»- st - urcrrrar - - 2-5S5*10 <3"2>
Therefore, the equivalent solid ccr-entration in parts per million by volume
is:
^ lb moles carbon
[soot] - — - i.,- x 106 ppm j x ID5 pom
2.585 xlO ±E_|!£_Le£. 3.1015 x 10"^
ft aas
= 3.224 x 106 x -'Oi 'B-Vi
3-5
-------
where x is the concentration of solids In pounds-per-?tandard-cubic-Toot of
sampled gas.
b(fc? Determinations
Sulfur dioxide Is determined by withdrawing g;i; samples with a vacuum
pump, bubbling them thro'jgh 25 ml of 3-percent hydrogen peroxide solution for
20 minutes. Sulfur ;Mox1de is absorbed In the sedition which is chilled by
an icewater bath.
After sampling, the solution 1s analyzsd for sulfur dioxide in an in-house
analytical laboratory. Isopropyl alcohol (IPA) is added to tfc° sample to pro-
duce an &J-percent IPA solution. The solution 1s titrated with standardized
barlurn perch 1oride solution using thorium is an Indicator. To attain a Salmon
endpolnt, the background of the solution 1s determined by titrating the same
amount 0" 80-percent IPA / 20-percent distilled water solution using the same
IPA.
The concentration of SOj in the original gas sanple is calculated as
follows:
where ^s(STP) »¦ Voliane sampled corrected to standard conditions, liter
Vs[STP) » 7s x 28.3 x j x
(B-4)
Vs = Volume sampled
T * average temperature during sanpling in °C
B ¦ barometer reading for atmospheric pressure
V.P. = average vacuum pressure during sampling
{Vs - VB) ''(C10^)^
(11.2 x i03)
ppm SOj
T^Tf?T
(B-5)
where Vs » volume of Ba(CI04}2 tltrant added to sample to obtain end
point (ml)
Vb 3 volume of Ba(C104)2 titrant added to obtain endpoint in a
blank solution (ml)
N(3a(CI04)2 ¦ normality of barium perch 1 orate
B-6
-------
V$(STP) - gas volume sampled, corrected tc standard cordltions
Visual Observations
The first ph
-------
rABLE B-Z. PHOTOGRAPHIC SETTIf.GS
cn
%
CO
liCHNiqUf
Ph itu^f "tr»h¥
CUt'l**'T
r(j»?on At ¦ i WI th
?) 50 -tin r 4 !«¦<>
2} 7$ J-tHtoi t'W
f lens,
Redlafce
Hy<.*n wltrt
Canon 16-1<0 «m
'OOfli * ?.)
Vuk-
Hnr i t
Panasonic Mrrtel
tjn*-a
JW. Node J
»€H.6r<||?r
Amif loiof
WjTi'Hj
1/ Mf»n
i-i nrr
Ml-f f>
I il*'f
MX) J f
fl^A It'*]
< aw*, fit*,
for sbtfiM*1., olor.
n<*rdc"1 "r,r tc,\i i'/Mrujs 1t) F'^w, rr,
Hunt bf- ¦:•>»¦: it n»':M DSyl'^hi o*f»re«lH)'JP'.
;:i«.hi» "\
(hH pr
*{*1** Ply, *
r.>*j*k U&
fiti»*>i rtn*'. f i 11
"toyriyH zi'iA 'W1
Hj{ ». -n>( »t«>l
rol ;
AH
f Km 11 ,*«• . -
£f»0 f i ».«*• ./
4fjf; h.iiw-/
•>p(.
i n, l a->
nhi'f*
Si i»r, h ii-H «i
;,,js .<¦* f« rtjf
i • ¦ , i 'i
• 1 ')• HviJMAr'Pr"". i'.w fP*tarr>,.
* fanrnj* ' •( :m< fur low- ii/mrui'.H / *fl«*S
¦ Can b# uvd .%*¥¦ ni'iHt.
htit ir (Uyliuhf Iof«rfr-fm.M h? bark
'|rojf-4 imq"',
fin* ttspfyi ftir 'J' ri '•'o'.in.** ' »«i ' Msr- Ff-ituf-s.
*d'.
fit# *V. l»40
(Iper-UfOfl ¦ * •«•
arxj #rc r. ik. 1i J f iruli.
¦4) a'raw CJMT w.' t 4. i..
_r.vi*f U* for I»m inocs ~!«**»'..
fM"» '<*>'« '*S • vr *¦ f.rtri
rod«.Qrahly *~»!!-
•'ri'-Mnp^ 1 ty '-~/ 1i,ttiMsou«>
i IrtTS^S
P*. '.uri-HMl"' * foW <.\ *> ittihh- *G> m>n-'uimikiu'. Hm-ib*
4t> :uo tfjry #f i>*t rinn dt Ufi
Q* ' sanN? « l\jfOS
Uq?<. no! prc.vTJ** t oir r in wii mo h<« fgoi'd
hy % j
j'4mint fi^Rip fp«» t*< »"«••..
Hsflul ».nt / i-*r f|*' ' :,»»»»• Hptri-..»vtni *'
-------
APPENDIX C
DATA SUMMARY TABLES
The summary tablss of all test data are provided in this appendix and are
grouped by the size of the flare head. The entries rir* self-explanatory, i.ith the
exception of the observations which .'equire further clai ification. The obser-
vations were made by the test operator during sampling.
Wind is the wir.d speed monitored by a three-cup anamometer located 10
.-'ee* above the roof of the 10-foot high control room. The anemometer 1s
about 40 feet from the flare heal anci outside the windscreens. Hence trie flare
flames were affected by less wind than measured.
Flame Length Is the "average" flame length observed using a 2-foot grid
mark on the probe support structure.
Lift Off is the distance between the flare head and the Ignition print,
identified by onset of visible radiation.
Color is a subjective description by the operator
Smoke is a subjective assessment. Part'culat? concentrations are deter-
mined by the filter procedure discussed 1n Appendix B.
C-l
-------
TABLE C-). OATA SUMMARY FOR 3-INf.H Fl OPF HFAn, .tH
LOW VELOCITY
HAM IK AD: [£R Mil: J-MCM*
ActMOl
lilt
Velocity
{ft/set.)
NomImI
fall
Velocity
(ft/sec)
Low
He«tImj
Value
')
Fuel
in
Nitron
(PerruHt)
Pltl 0
Hb.Slew
/Ib.FyeU
flbservitlMS
St^Se
Httliod
Niki
IHiottl
'--"H
r - -
Purpose of Test
lest
fadwi
Rctet.Moa
Ring
Hi^d
fet«4
IHW)
flMC
tongtli
in)
LJ.ft
Off
(lathes)
Color
5aoto
Pr^t
fetltivfl
(Ml
C0*M1 tiOP
Efficiency
< Percent I
Crobust ton ifftctftxy
001
No
Q '
©.Ii
1)11
S5.8
0/J
Low
4
0
Yellow
¦
7
tt.«
56 Perceot tjHy
002
0.
0.5
131 i
550
0 0
Low
4
0
Yellow
H
10
W. 37
00)
0 5
0.5
1309
55 1
0 0
low
4
«
fellow
Yw
R
9
•0.77
004
OS
O.j
I3U
55 0
0.0
low
4
0
fellow
Yes
R
3
99 06
cos
2.0
?.o
DIB
56 1
(>.0
low
)
0
Dt.'el
Yes
*
99.54
oct
2 0
2.0
DM
65.9
111)
low
1
0
Vtllow
Yii
R
5
94.24
0u7
2.0
2.0
DM
55.9
on
LOW
6.5
0
Yellnw
'«
1
fl
99.64
ooe
2.0
J.O
. JM
55.9
0 0
1 ow
6
9
Y«llow
Yes
R
10
98.95
009
9. J
9.9
1 J2'
56.4
o.q
Low
IS
0
Yellow
Ye*
N
10
99. '4
G10
18.0
19. i;
IJI6
;«>.0
0 0
Low
IS
s
ret 1 cm
y«
I
10
98.69
Oil
10.0
10.0
uza
i€ )
u.O
2-4
IS
0
Y*r 1 low
Yes
R
13
95.66
o a
10.0
10.0
U23
S* )
0 0
7-4
IS
0
YeHou
Yes
R
la
98. '8
Effect cf Stew
ou
No
0.5
0.5
273
11 fi
0 0
r-i
4
a
Barely
VHfcle
No
M
0
90.19
016
10.0
10 0
I3>3
56. J
0.142
0-8
16
0
fellow
Little
/
99.11 J
01/
10.0
to 0
1)21
in 1
0.300
0-6
le
i
fellow
Little
H
9
99.%
018
<0 0
10 0
11/1
56 ?
O.W
9
2-4
fellow
Ltttlc
H
9
9V.119
019
10.0
too
t Wi
56 I
1 000
6
4
Yellow
Little
N
9
98. M
low Htu
020
Mo
12.0
i? n
670
?6 4
0 0
fellow
Yrt
N
9
99.
021
? 1
? i
Ml
14 6
0 (1
3-S
4
4-10
Yflluw
Yes
H
4
99. <9
051
2 1
? i
16?
IS 4
o a
0~S
4
0-5
Yel/Ora
No
R
3
91.01
052
7.1
? i
17C
16 0
0 0
0-3
3-5
0-3
lei/Or*
No
R
5
98.TO
-------
T/J,LEC-2. DATA SUKMARY FOR EER 3-INCH
Fi ARE HEAD, HIGH VFLOCFIES
HAM KM. U* SUl: J ;«H'
Pv«rpovt of lest
TMt
Mxr
Helen* 94
Ho
H
33
99. b#
HlqhVel. IOC* t3Hg
92
Cc * <6.31
1/M J
7*.3
2 ISC.
10C.S
V UM
d-]
30
IMS
Tellow
ib
R
36
39 W
91
7
392
2348
99.9
0 122
8-5
26.5
10
Or *n 9*
ft)
fl
31
99. i«
44
171 4
79.3
2J50
100.0
0.061
0-1
MS
12
Tel/Ora
Little
R
V
99 83
Hi, »¦ Velocity.
Stable F!«at t'm\t
95
Con* 46.31
114.8
511
792
3V7
~ .<*»/
u z.s
23
12-18
ttlU
No
P.
26
99.42
-
96
2/4. E
i?t a
IM3
44.4
~ 028
0-1.5
J3
1878
tel/Orj
Ho
e
37
*9 85
5rt
9i
Cow 46.11
05 8
39.7
1156
49 2
0.14*
1 b
20
t*
0TJ»9*
No
0
24
99 ?-'
-
m
1 J?. J
W.I
im
48.2
Q.071
ti i
29
7
Vifl/Ora
No
R
M
99 8'
Stable ff iatt
-------
TABLfc C-3. DAi7. SUMMARY OF EER 6-INCH FLARE HEAl)
FLAM HEN): till » IHCH*
f 1 MM
Rtag
Actual
Noylt«i t
r
low
kel
UImI
Speed
Obsm fitloni
:
PotltlOM
mi
Furpose of lest
1n(
Huilier
fait
vtioclty
rait
Velocity
'f i/;-cj
Mm I (raj
fell*
|8tu/ftJv,
llUroyi
(Percent)
Sit 111
Hb.Stfc*
/Ib.Tvt!)
Hmm
Length
l"l
lift
Off
OftClMS)
Color
Smoke
Ntutod
B'fiiU
H*tM4
CoAwttttoa
(ffklMCf
(Percent)
i.r*4>u$tfoa fffideecy
22
No
r 1.0
1.0
129)
Vj.O
j.d
ass
a
0
teltov
Ttt
H
99.60
56 Percent
21
; z.e
i. a
1314
55.?
fl a:
e-io
9
0
Telia**
Yps
H
ii
99.81
24
?.8
2.8
1314
5^.9
n o
-5-5
IS
0
Tel/©r«
res
H
5
99.81
25
L.U
2.8
1314
55.9
C.j
I 0 >0
15
0
lei/Or*
Yd
H
8
99 i5
26
2.8
2.8
1314
55.9
0.8
! on
15
0
It 1/Or#
T«
~12
Tel/BIu
NO
5
98.93
35
j.i
3.1
451
19.2
o.o
HA
7
1-8
Tel low
NO
ft
$
86.93
34
u
3.1
454
19 3
0.'
M
J
3-4
Tel low
No
R
ID
•1.06
37
3.1
3 1
19.2
CO
HA
8
3-6
Yellow
No
R
IS
92.24
ia
11.7
M. 7
611
76. 1
00
HA
13.5
6-«
Tel low
Mo
1
70
98.76
69
3.0
3 0
1?6
14.3
0.0
r
5.5
0-%
OfMf<
Ho
B
10
S7.ll
78
ID
1.0
2a 7
12 ?
0.0
0-3
4
0 3
Crtny*
tt>
R
5
99 i4
n
10.3
ii.3
343
14 G
0 u
3
IS
n-s
Oranye
yo
R
17
49 36
)2
9
2.9
345
14 7
0 0
0?
fl
0
ilNr
Mo
R
5
91.49
73
in
1 0
2SI
12.4
n.o
0 4
5
0
Cl**«r
No
R
to
9/.97
74
2 9
350
14.9
uo
0 1
10
0-7
Or*K>e
No
R
15
94.02
S6 Percent tjHg
75
Ht
2 S
2.9
1321
bfc ?
ft n
0-2
lis
0
Yellow
lei
8
17
99.66
76
1
1.0
I. (J
lill
«.S H
Oil
0-3
8
0
0r«"
-------
TABLE C-4. DATA SUMMARY FOR EER 12
-INCH FLARE HEAD
ILIWt
Lilt
Off
Color
We
9 rob*
Pavilion
I'M
CflfchtlM
f "Ulcftcy
Co»J)ustt6u tftflctercy
39
Ho
0.2
0.2
1J09
55.7
0.0
r;\
i
0
>rll<«
Tn
It
13
90. *9
S4 Peit en - C^M
W
?.o
2.0
IJ2J
M.J
0 0
Mm
19
J
telle*
Yrt
R
23
99.21
a
4.1
4 A
\W
0.0
»
0
tent*
f«
ft
29
V9.W
42
4.0
4.0
1J14
&S 9
0.0
«A
23
0
fc ilow
Yes
R
22
99.48
L 43
id
1318
56.1
0 0
K*
23
0
Yellow
res
fl
15
99.50
i tm Bin
41
No
? )
2.1
37*!
16.1
0.0
MA
12
0
Tel la*
to
R
11
i '.My «><• v
4S
*o
0.2
a.2
980
" 99.1
o.o
M
4
b
0r**it
Little
R
9
99.91
Low B(u
u
No
d.2
a.?
"" ttl ""
R.F
0 0
HA
4
0
r (k4H(|i
Little
r
S
98 M
47
4.1
4.1
Jftj
16.4
Oh
NA
M
10
(Kevji
to
B
21
94.99
4B
4-2
4.2
390
lb.1.
0.0
**
IS
0
leWa*
m
1
19
94. J>
49
4.2
4.2
365
16.4
0.0
NA
13
0
*~
t
!4
9 ».«
btc4», Saokelcss
W
_ J
4.Q
4.0
I3?3
S6. 3
0 44/
MA
?*
0
Yellow
little
•
10
99 K
-------
TABLE C-5. DATV. SUMMARY TOR 12-INCH INDUSTRIAL
FLARE HEAD A
FLAM Iff AO: Iwfciitrtjl SI2C: 12-IRCH*
F Im
Mention
RIihj
Actual
F.lr
Velocity
(ft/sec)
Low
Htit IrU
Velocity
(ft/sec)
Miv
(Ib SteM
/Ibttttl)
kind
Sp*td
(mi)
Flw
length
(ft)
Lit
01.
(Irttfces)
Cotur
le
P,o6c
Pes It to*
(ft)
CorttaUon
Efficiency
(Percent)
Coatantlon Efficiency
62
Tes
2.0
1.9
1269
54.0
o.o
0-1
IB
0
TeH*.
tei
R
22
99 }?
Sfi Percent C^Ng
«
4. J
4.0
1316
56.1
0. rJ
tt-1
24
0
fellow
Tel
*
»
99 3?
low Btu
64
y's J
4,5
o
"jjj
16.7
8.0
" ft-T "
21
6
fel'ow
Ho
8
25
Steitaa, Saoki»ie**
65
tin
4.3
4 0
Dlfi
56.1
0.511
0-4
K
0
ITe Ho-
Wr/ltl
R
7^
99.8t>
S6 Percenl
66
Tm
M
4.0
1116
56.0
0.0
0-4
?3
0
Vel In*
Yes
R
ff
49.61
6/
4 3
4.0
1)16
56.0
0.0
c-;
22
0
Yellow
tei
fl
2B
99.61
-------
TABLE C-6. D,"Tfl SUMMARY F0« 12-INCH INDUSTRIAL
FLARE HEAD B
FIAM KM): IMujtrUI SIIF: l?-INCH*
firteatloil
m «»9
Actual
lilt
VeltKily
(ft/Sic)
NmInI
IiH
Velocity
(f l/iif-i
lot*
HtiUng
V»l i*
St«»
Ratio
tlfa.StCM •
/ib.rjci|
Hind
(W*)
OfcEtrvjtlMi
SmoU
I'Hiki
H'fljo4
Purpose of rttt
Test
Ktisatr
lr»
Nltragm
(Percent)
F| m*
Lralli
(ft)
Lift
Off
(Iftctiei)
Cal/0ra
v«s
R
\2
~9.4b
W
O.J
0.2
UH
»,9
(1.0
»
l.i
0
i Anye
V»\
G
•0
99.4®
Bt
CO
4.0
UM
MS
o.o
07
2*
a
r< l/Ori
?
2%
«.»d
Low Btu
*"3
Ui
].D
2.0
31/
13 6
0.0
0-1.5
10 5
0
Clr/0r«
*>
R
lb
ot.zi
«
S.2
4.1
Ml
12.9
0.0
0-2
20
a
M9C
no
ft
24
9».F?
90
0.3
(».2
MS
1 J-4
0.0
IO
4
Q
°"»9«
Mo
ft
«
»s.w
Steaa, Snokeless
91
r-.y
J.9
1)11
55.8
0.441
0-5
71
D
Ora«fc
Very
Utile
R
21
MM
-------
TABLF C-7. DATA SUMMARY FOR 12-INCK INDUSTRIAL
FLARF HEAD C
HAM HIM): InAntrKI SIS: I ?-I~CH»
AcUftftl
C«tt
Velocity
itt/i tc]
F*U
Velocity
(rt/tccf
U.r
He«ttr>9
Vilu*
(lU/M1!
fuel
NHrpyc*
percent)
su«
(IL.SltW
/Ibf-l]
Obtcrvatfofti
Si*"
M>U«4
(•Kill
H-Kno4
1 —
Purpo** of Ten
Test
m«fctr
ftetmt Ion
Ring
Hind
Sfwd
(in)
riM
in)
Lift
Off
(j*Cl*ft|
Color
Swte
frtM
felfltf*)
ft)
Co*uitfw
unci***
{Pwrcmii
tafevilton IH'cWftcy
u
In
0.6b
0.7
IM
*6.7
0.0
o-i
6.5
0
Tel/Ora
•
II
99 20
56 PttiWll
54
6.50
2.8
1316
46 0
0.0
0 2
IS
3 5
VelIon
Tei
*
II
MM
55
11.3
% }
I37t)
*6.5
0.0
0 2
21
a
Vtl/Ord
Tat
>
If
99 66
56
|J U
4 0
1316
54 a
8.0
Ik
21
0
r«l/0ri
v««
•
21
99.10
sr
»J-0
4.0
1316
56.0
0.0
MA
71
0.0
Yel/Ora
T««
R
to
90.11
l ow Btu
?8
6 8
2.1
J/l
15 8
0 0
0-1
16
0
Vellew
to
t
18
99. M
56 Percent CjHfl Ste«
59
13-0
4.0
1114
-5 9
0.456
IM
ZS.6
o -
Vcl/Ora
I llllr
•
2t
99.45
Vo* Btu
tt
14 63
4.5
"SfJ ~
» ?
or
0-1
i|
4\2
lei low
little
A
25
11.16
I I.OU Btu
CI
065
0.2
4W
IS S
' 0.0
0-2
J
0
Yellow
R
6
99 W
-------
APPENDIX D
INTEGRATED COMBUSTION EFFICIENCIES
The tables in this appendix list test conditions of the flare flames and
the "integrated" combustion efficiencies calculated according to the procedure
described in Section 3 3. Most of th<* designations are clear.
The SIZE column lists the nominal diameters of the various flare heads.
i"he actua1 dimensions of the EER test flare heads are i istaa in Appendix A.
The SIZES of the industrial flare heads are designations by the manufacturer
and do not necessarily reflect their physical dimensions.
The VEL column lists the nominal velocity of the gas rixture exiting the
flare heads. These e e calculated from the measured component gas flow rates
tir.d the flow area based on the nominal SIZE of the flare heads. For the flare
heads whia; have nozzle flow areas different from the nominal value, the gas
velocity CJin be significantly higher. Both the nominal and actual gas exit
velocities are tabulated in Apptidix C.
STEAK RATIO indicates tht mass flow of staam injected per pound of com-
bustible. The miss flow of combusts»e gas, rather than the total gas flow
(i.e., it excludes nitrogen). is used because only the cotbwiiibla fraction
has a tendency to smoke and nitrogen does not increase the fuel's tendency to
smoke.
KT is the heigiit cf the probe tip above the flare heads.
Ttie EFFICIENCIES are the results of the integration procedure described
in Section 3.3.
D-1
-------
«6
IA Aft
C* M
0« 66
Rk 66
OG 66
64
62 0*
9Z 0A
»Z £b
«0 M
C6 90
06 86
/~ BA
96 66
(I l>6
01 M
"W *C6
69 B6
46 B6
~9 66
~2 »6
90 66
4/ B4
Ak 6A
TVL01
/ll f.H
fr<« '6
/.9 M>
6^- ( S
L4m t.4. AH ft/.
CS. 66 t6 6/j
99 66 EA AA
1/ 66 26 66
10 66 6C *6
f / 66 o/ F16
96 66
96 »6
frfl 04
I*
€6 66 5* B6
9B 66 TO 16
48 66
// A*
91 66
/R 06
~0 66 frB 66
CV 66 10 lb
%l *A
9i 66
~ I 96
C0 66 (K 66
tC 6A OO 001
CV 66 00 001
ID 6* 9C 16
9*9 6A 00 * 6
9*» 66 19
6fr 66
wO 001
I.I *>*
t+t. AA
nn •«/.
i:a 6c
Bfl AA
/O A6
»k 66
/B 66
76 66
BB 66
6V AA
kO 96
9B 66
£0 6*
ee 66
Q* 16
IR *.L
IL 6A
6C 66
BC 66
oo ooi
lO *6
19 AA
Ofi AA
00 001
iniis dm
17.) AWiniJJj
0.1
«> <.
O *»
0 i'C
0 sc
O .iif
O At!
O C2
O XI
o uz
O CI
o oi
O 9
0 12
O 11
o ct
0 I
o r.i
o ci
o oi
3 OI
o »
o c
o c
o A
o i
IH
UV*n IMU 0
IHVU U<'0 O
¦IHVM OlMI O
TilVB OOO O
JXVU OO" o
•JtfVH OOO 0
3*VM OOO O
J>IVM 0*.M> 0
3WU OOO O
l*/tt OWi O
jyt/H OOO O
TtfVlf OOO 0
!MVH OOO 0
jyvtf ooo o
| l>0«> O
nuva ooo o
JTXVH OOO O
•ItfVlf oco n
TMVW OOO ^
3XVM OOO O
•1WH OOO O
JMVM OOO O
JMVH POO O
nWN (MH1 O
ikvm ooo o
uiaii^t ui ivn
3 iJMVn W / 119
O <5
k 66
» 0
0 0
O c
o o
o o
0 o
o o
0 o
o o
0 0
o o
o 0
D 0
O Q
o a
O 0
o o
o o
o o
o o
o o
0 0
IXI
klO
»"* '» I V* 0
0 o cf 0
191 I «!
1 vr> * fr
A CL o If
6
r^lT.T
O ».'l
o «ri
o
O i-l
o Jl
0 JI
0 21
O «M
0 *?
O ?
0 9
0 9
0 9
o n
o ?
O 9
o n
o t
o c
*} c
O *-
l> c
o f:
r> r
o r
NM3
tunit.4 H-i-;
i)ion.( u:u
oirmH -im
ninMrf-Hii
HiOMJ U33
OlOHJ-HJi
OlOXd- U39
oiDt-'-naa
ninn^-M^
010Hr«-b?J
010W tt39
niUrtrt-H~;3
DIOHcJ *33
HQUti-Hl3
ni
OlOlfd -H33
010M-U33
niOWd~H33
oioi.'d™M.ia
oiwri-'iaa
OlO«d-U33
OiOHd-H^3
OiOW^-H3^
OiriHH~M3"3
(|V?13tfVld
r a- 11 r.o «?%
CO VI ifO
ce-vi «*o
rn-9t jto rfr
CC CI SO 2k
Ml-SI-iO ll>
CB-kl-fO Ok
GB 60 ZO AC
C0-!C?0 BL
CB-io-co *r
rry IG- 10 9C
t«- it.- 10 CE
r.ri-C^-»a IC
COQZ 10 CC
LO ScT lO hZ
Cfl-kz-io az
ce-90 'O si
C0 9O-IO II
CB-IC £1 O!
n-(K!-£l b
ce-(.c-£; f
SOL\21 9
IH-7I-2I k
ZB-ci-ri r
I
ON
31 ¥0 14
jvu -|Nvi;i ji; 4 ill no »m en-irni-ot no oiihimj vivo
-------
DAI A PRINT! D (JN 10 Os'T -Ml io '*> IH
PT DATS
Nil
Ft.A4tL;
hi;«. vn cnitfj
<|N) fFMSl
IXR PHOTO 6 0
tfc» CHOTO
rf.R prdtu
PFR-PBHTfl
CER--R1MC 1HV
Ff* MiNC D!V
EER*-EUN<- DIV
FFH-RiNC tflN
&
INDUSTRY II 1%.' U
INDUSTRY I) 12 O
JNWJSIRY 3 1.1 - ( iW
f ,) /V 3 tOO U
3 1) 3*/ i? 99 °
9 i IOO o
t G -yj i :n ;
i K 1,7. I' -fl
0 0
Q C)
o o
O O
O 0
O 0
& 0
o a
o o
O 0
O O
0 0
O O
0 0
O 0
0 Q
o d
o o
0
D O
ti 0
0 0
O <1
o o
A O
O iMH)
0 tJOO
O t4KJ
Q OLU
O IKJO
o mia
O <*»0
O
O «lfj
G 149
o o,*»i
0 131
Q
O <>00
Q tHH)
o mo
C (XX)
o <;oo
C
I FP iCi04CY C*. I
Hr sum
rcif AL
W L*0 VJ 7i?
V? 7 7 W ?/
V9 71 W 33
Q7 V9 99
99 B4 V? 97
99 *#f» V3 J1
99 84 57
79 *»9
99 B? 99 30
S r 40 9/ 79
9«# &f> 99 (8
99 B7 100 OO
99 69 79 96
99- 71 9V 94
99 54 4r 94
9V 74 9V tin
99 44 9V 76
9V 76 100 W
m JS 99 9'j
99 HI 9y 99
99 O.i 9 / 7/
99 // 9y V6
9? H'j W 79
99 •'»• '' 9li
/ fM» '/ r^.'
99 9* 93 49
9? 9C» 97 97
S*» 97 V9 04
99 0C V" 66
99 73 99 54
99 9C 95 11
99 99 91! 40
100 ©g w «»6
IOO oa 9V 20
|O0 oa 9/ 27
99 99 99 33
100 00 99 H7
99. 90 99 54
1W OO V9 ^.5
\r.Q QO 99 48
IOC 00 9V 58
IOC 00 99 21
«?9 9 f 99 7i*
100 00 99 59
IOC Ofl 79 i%4
100 00 97 6o
Kia OO 99 7 fl
99 99 99 0.1
!CK) CO 97 (17
JW) DO *r, H">
-------
DATA FRINTH) ON III IK I fJ\t> iX) tH I CM'
t I
i>a r
P I I>A I L I. AMUCAI»
NP
lr> I I VI. L I /I t!I
c im»
till S 11 AM fi/V»r I K Ml
ty. J H/u mi ht- jinii>
y; 0^ .'->-y3 h-whinc con
9G D.*! i?4- G3 LtH WING f UN
V* OA-24-tU F.F.H-N INC CON
icq 06-?4 03 u:r FUNO CON
101 o'--0 7-b:i <;<>*
102 'J 7 0/-U3 CLR RING CON
11)3 07-00-15.1 FKH -9 INC CON
104 07 OO 0U C6RRlrJO CDN
103 07-0B-B3 FFR RfNG COfsJ
J 0 JV / 47 rJ
j o rn r 40 z
o is/a i 4:1 7
3 0 IW 7 40 V
3 <> 19 * 76 V
~ O 'O I "¦ U
3 0 120 'J 76 4
3 0 I73 7 39 S
3 0 : i ;• ft 100 o
0 0 O 142 PAKE 24 O
Q 0 0 <171 RANC 34 O
0 0 0 Oin RAHF 35 O
0 0 0 OifO RAKf- 37 0
00 O iff I RAHF ptf a
(JO O <)')0 RAME .14.0
0 0 U iiiO RAKE 37 0
0 0 O 02.1 RAKE 3C 0
0 (1 *- 01V RAKF A*> 0
UMCILMCY '7.)
H *O0l
CO
V9 76
99 BB
99 9?
99 ?!
99 7)
99 80
99 90
97. 90
99 03
V/ 9fc
v; 99
99 v /
V? 73
99 96
9? 94
•?U
100 00
99 9V
I OO 00
9 99
TOIAL
99 72
99 B7
99 8ti
99 84
w /:*
99 74
99 0£<
V/ 8|
99 7 7
-------
APPENDIX
SAMPLES OF TEST DATA
Data for the 105 tests conducted in this program were compiled and
analyzed via computer processing. Examples are show in the following tables:
• Input Flow Conditions
¦ Measurements
« Combustion Efficiency
Input Flow Conditions which list the flow rites of the propane- natural
gas, nitrogen, SOg anJ steam flows are in cubic foot (Stc) based on 1 atmos-
phere pressure and 7C°F temperature. The calculated concentrations are
reported in volume percent (%) and parts per million (ppm). The mass flow
rates are reported in sounds per hour based on the theoretical densities of
the fuels, The STEAM/PAS ratio is calculated per pound of combustible in the
fuel mixture.
Measurements lists the concentrations of Og, CO, CC2, HC, ^oot, and
present in the ambient air and the sampled gas. These are reported as volume
II
percent. The numbers following the "+ — symbols are estimated maximum poten-
tial errors of the values reported in the units of the concent^atirn. The
numbers bel")w the headings "AT SR=1 BY CO-." and "AT SR=1 BY SO^" are the spe-
cies concentrations that would have been measured if the combustion products
were not diluted by ambient air. The metfed of estimating the maximuri poten-
tial errors and calculating the undiluted species concentrations are dis-
cussed in Section 3.3.
The sampling method and location are listed at the top. For the RAKE
probes, the locations are those of the probe tips. For the HOOD probe, the
location is that of the inlet of hood. The actual location of ths probe is
in tne sampling chamber which can be b to 8 feet, above the inlet depending on
whether the hood extension is used.
Combustion Efficiency lists tne calculated combustion efficiencies based
cn CO, CO2, HC, and Soot. These are reported in volume percent (%). The maxi-
mum potential errors cf each parameter follow the "+ —" symbols and reported
in the iinits of efficiency.
E-1
-------
Dilution Factors are the amounts of air entrained in the corcbustlon pro-
ducts. A dilution factor of ^ero means that no additional air is entrained
Into the stoichiometrics combustion products. A dilution factor of 1 means
that the stoichiou
-------
DAT* PRINTED ON kO- CT-83 OB 04 35
BUE FLARE DAT
PCI NT l DATE 12—14—0? FLAREhCAD EER-PRGTO 9I7F 3 0 IK
* k w as « ¦ZjauaknaBaa M ¦ II BBK » = ¦E*HaBa«a« BWaBMH « MBBSMK3S
INPUT FLJU CONDITIONS
VELOCITY S 0 SO FP£ YW - 13.1.2 B'U/CU FT
PROPANE 0 a24 SCFH 35. 8 'A 5 6 LB/HR STEAM - 0 0 LB/HP
N CAS J 000 5CFM 0. 0 X 0 0 LB/HR STEAM/GAS ¦ 0. 000
NITROGEN 2 649 5CFM 43. 7 a 8 '.9/HR
502 0 4TO LCFM 3304. PPM 0 08 LB/HR
^ A SURE ME NTS WITH RAKE T HEIGHT 7 0 FT 2.6 FT FROM FLARFHCAD AXIS
AT Sft-1 BY SOI
0. 00*- 0 00
-0. OO+-608. S9
1130^. +-•«»»••
-o 00+-»«»**«
324 30~- 03 56
420. 61 *- 52. 58
SPECIE
BACKGROJND
MEASURED
A T SR-1 BY COk
0* iZi
21. 00*-
0 00
20. 90*—
0.
02
0. 00*- C. CO
CO
4. 31 *-
C. 40
2. 70*—
0.
40
-0. 00*-1998. 91
C02CPK)
742. *-
L50
960. *-
174
132733 **401326.
*C tPPPI)
3. 59*-
2. 35
3. 50*—
0.
•v5
-0 TO*-1656 06
500T(PPP1)
0. 86*-
0. 39
1 26*-
0.
03
158. 77*- 439 49
SdSfPPH)
0 00*-
0. 00
0. CO*-
0
10
485. 48*- 814 91
COMBUSTION EFFICIENCY (X)
-------
DAT A PRINTED CN 20-OCT-83 OB 06 96
FILE FLAKE. DAT
POINT
I DATE 12-n-az FLAREHEAD EER-PROTC
SIZE
3. 0 IN
r.NPUT FLCW CONDITIONS
VEuDC1 TV
PROPANE
N CAB
NITROCI-K
SOS
SPEC IE
as r,'.)
CO
creiPPM)
HC
SOOT(PPM)
302(PPM>
0. 50 FPS HV - 131 .
v. 0=4 5CF1 S3 9 it
0. COO SCFM 0 0 X
0 643 SCFM 43 7 *
0 470 SCFH 3304 ^PM
BTU/CU FT
3. 6 LB/HR
0 0 LB/HR
2. B LB/HR
0 OS LB/*R
5TEAM
BTEAM/CAS
0 0 LB/HR
O. 000
WITH
rake at
HEIGHT
7 0 FT
BACKGROUND
MEASURED
21 00—
0 00
20. 90+-
0 02
4. 31+-
0. 40
2. 70+-
0. 40
743. +-
130
960. +-
174.
3. 39+-
2. 33
3 30+-
0 03
0- 96+"
0. 39
1. 2B+-
0. 03
0. 00+-
0. 00
0. 00+-
0. 10
AT SH-1 BY C02
0. 00*- 0 00
-0 00+-199B. 91
132733 +-401586
-0 0O+-1636. 06
538. 77*- 43*. 49
48S. *«~- B14. 91
at sR"i bv so:
o. oo*- e. oo
-0. 00»-60S B9
li5093. +-¦••••••
—0. 00*—~~•••#
224. 30-*— 83. 36
*30 bl+- 32 SB
COMBUST ION EFFICIENCY IZ>
SPECIE BY 02 BY C02 BY S02
CO 100 Q0+- 0. 31 ICO. 00+- I 30 100 00+- C. 33
mC 100 C0+- 1 IS lOO. 00+- i 23 100.00+- L.19
SOOT 99 31 ~ - 0 ?9 99 B1+- O 92 99 B 1+- 0 41
TOTAL 99 ai— 2 07 99 BI+- 3. 66 99 B1+- 2. 12
DILUTION FACTOR
BY 02 - 209 0+- 42
BY COS - 603. 9*- 742
BY 502 " 324 0— I :7
(".EASL'REMENTS l.ITH RAKt AT HEICHT
,0 FT !,0 FT FRCrt FLAREKEAD AXIS
AT 9T»1 BY S02
0 09+- 0 00
-0. 00+—766 27
233407 +—••«*•~
—0. 00* —*#-*•~»
530. &3+-1BS. 33
42Q. tol+— 60. 09
3PE:IE
~ACKCROUND
1EASURED
AT SR-1
BY C02
02 <">
25 O0+-
0. 00
20. P0 + -
0. 02
0
C0+-
0 OO
co -
SPECIE
BY Q2
3Y C02
3 Y S02
DILUTIO!
CD 100
00+— 0 23
100 0C+-
0 36 100
00+-
0
30
BY 02 1
HC 100
00>~ 0 60
ioo oc+-
0 63 100.
00+-
0.
6.2
BY C02 '
SCOT 99
7B+- 0 23
99 7S+—
0 33 99
7B+-
0
29
3Y 502 '
TOTAL 99
70+- 1 06
'9 7B+-
1 73 79
70+-
1.
21
104. 0+-
312. 9+-
1C
26!
143
MEASUfiEMEK
TS UlTh RAKE A '
HEIGHT 7
0 Fr
0 0
FT i'RCM FLAREHEAD 4
2 :9—
2 55
5 50+-
0 07
527
0B + -1182. 84
622. 95 + -9S3. 33
SC3T*PPM>
0 8fe+—
0 39
3 52*-
0 09
771
40 + - 646. 94
863 J0+-20S 11
3G£ <" PPM)
0 C'D»-
0 CO
1. 30+-
0.10
375
3B+- 3.9 72
420. 61 + - 32. 33
CCMBUSTIEjN
EFPICiSNCY <%>
SPECIE
QY 02
BV C
32
BV SO
2
DILUTION
FACTOR
CC ICO.
OO*- 0. 1/
100.0O+-
0 25 100.
00+-
0
19
BY 02 "
34 0+- 1 Z
HC 99
53+- C 83
99. 58+-
1. 49 99
50+-
00
BY C02 -
2BB 1+- 323 '
SCOT 99
0 A5
99 42-^-
1.31 99
42+-
0
64
BV 302 -
322. ! + - 3i S
TOTAL 99
0J+- 1 43
99.0:+™
2 04 99
01*-
ta«
E-4
-------
DATA PRifcTED 3N 20-CCr-S3 08 06: 56
F CLE rLARE Df,r
POINT I
DATE 12-
14-62
FLAREHtAD'
EER-I
s
1
!
1
*
1
!
1
I
1
¥
1
«
1
1
S
I
«
1
k
«
s
i
i
......
^5ASVHcnfeNTy -J ITK H**E AT
HE1BMT 7
0 FT
SPECIE
BACKGROUND
MEASURED
02 <*/¦>
21 00 —
0 00
20 SO*-
0. 0?
CO (PPH>
4 31*-
0 40
3, 28*-
0. 40
CQ2CPPMJ
74J *-
150.
1424. *-
174
HC —
2 53
3. 50—
0 OS
SOOT {PPP1 >
0 Be*-
0 3"
3 76*-
0 14
sozippm
0. 00*-
0 00
0 60«—
0 10
SUE:
3 0 IN
AT SR-1 BV COS
O 00*- 0 00
-O 03+- 261 be
133733 *-133708
-0 OC*- 509 96
*90. 84*-
116, 31*-
AT SR-1 BY sot
0. OO*- *¦ 00
-0. 00*-7SI. 07
478482.
-0. 00*~»*«*»«
SiB. QC 3439. 3Z*-««»*»»
82 C4 430. 61*~ 70 10
coinysTiaf efficiency ix>
SPECIE .1* 03 BY CQ2 BV 802
CO too C-0»— a 13 100. 00*- 0. 20 100 00*- O. 1c.
HC 100 00*- 0 37 500.00*- 0.38 100 00*- 0.38
Sl>2T 2?+- J f! 99.29*- 1.14 ?9 2;>
Si oo*-
0 00
20. 73*-
0 OS
0 OO*-
0. 00
0 30*- 0 00
:a < bpm »
4 31*-
0 40
4 JS+-
0. «c
-0 00*-
94. 63
-V CO*-130 36
coa»prii
7«y. *-
150.
IBBO. *-
174
,32733 *-
S3477.
t8fl->3e *-79391
nC (PPf
3 59*-
2. 5S
¦> 50*-
9, OS
-0 00*—
302. 61
-3. 00*-42C 12
SCOT I'PFt-,)
0. B6*-
0. 39
j OO*—
0. IS
481 62*—
196. 62
671 40*-139. 90
=D2< ppri
0 00*-
0. 00
m. 60+-
0. 13
30* 69*-
120 30
4.-C 61.~ 31 03
CCiiauSTION 5PFICI6NCY ('/.)
SPECIE BY C
CO 100 00*-
p*C IflO 00*-
SCOT '9 64*-
rrr M. <<» »A»-
BY CQ2
BY SOJ
0 07
0 22
0 IS
0 47
IOC 00«~ 0 C7 100 CO*- 0 07
100 00+- 0 53 IOC 30*- 0 2~
99 64+-
99 ci~ ¦—
0. 37
O 6T
99 64*- 0.
99 64*- 0.
23
33
DILUTION FACTOR
BY H2 » 76 8*- 1 8
BY COS » U5 C*~ 40 S
DY SO? » 160 8*- 24 0
-------
DATA PRINTED ON 20-OCT-03 00 C6: 5*
FILt
FLAflE DAT
POINT
DATE 12-14-02
FLAREhEAD: EER-P^OTO
SIZE:
3 0 IN
IMPl/T FLQy
VELOCITY -
PROPANE
N GAS
NITROGEN
SOS.
rfEASUHEf- El'1
SPECIE
02 <>
CD (PPM)
C02(PPM)
rtC (PPMJ
SLOT(PPM 5
soa(pPM)
editions
O *0 FPS HV
0 3i/4 SCF* 5!
O 003 3CFM O
0 &;j SCFM 43
0 470 *CFH 9304
1311. 2 BTU/CU FT
B * 5 6 LB/hR
OX CO LB/HR
7 X 2 3 LB/HR
PPM 0 OP LB/HR
STEAM
3T*AM/GAa
0 0 LB/HP
0 GOO
¦usaiDaa
TS WITH HOOD
BACKGROUND
AT
HEIGHT 0 0
MEASURED
FT
0 0 FT FROM FLAREHCAD AXIS
AT SR* 1 BY C02 AT SR~1 BY S02
:i oo+-
0 OO
2:. r»o+-
0 02
0 00*- 0. OC
0. OO*- 0. 00
4 31+—
0 40
6 37-»-
0. 40
518. *2*- 900. 94
; 1V4, 78-* -686. 2B
742.
150
1324. »-
174
132733. ~-15*472.
306261, ~•»«#*#•
3 59*-
2 5!*
4 35+—
0 05
177. 19*- 696. 89
405. 42*-#««*#*
0 B6*-
C. 19
7 49+—
0 19
1507.O:+-lO20. 88
3487. 1*+-97S». 81
0 00*-
0. 00
0 BO*—
0 10
181. 71 ~- 135. 64
61-*- 52 5"?
COMBUSTION EFF1CIENCY (X)
SPECIE
CO
HC
SOOT
TOTmL
B » 02
0 00— 0. 30
o oc— o. oo
0 00*- c. 00
0 00- - 0. oo
BY C02
99 62*- 0. 02
99. 87~- O. 67
99. SB* 2. 04
99. 37*- 3. 50
BV 502
99 &2+- O 52
99 074— O. 57
96 B8«™ 1. 18
98.17*- 2 24
DILUTION FACTOR
£Y 02 • 0 0+- 0.
BY C02 *-• 226. !~- 14-1.;
BY 502 - 524 ~+- I 17
E-6
-------
DATA =RIHTFD ON 20-OCT-83 08:06 56
FILE FLAB": DAT
3 INT
DATE
12-13-02
FLARtHEAD: EER-PRQTQ
SIZE.
3. 0 IN
1NPL- F^Dw CONDITIONS
VELOCITY - 0 •> %
M GAS 0 000 SCFM 0 0 X
NITROGEN 0 S47 SCFM « 7 I
so: O fro SCFH 5293 PPM
1310 1 BTU/CU FT
V. 5 6 LD/HR
0 J '.D/HR
2 8 LE/HR
O. OS LD/KR
STEAM
STEAIVSAS
0. C
0. OOO
.3/HR
MEAS,;REMElNlTg WITH RAKE AT HEIGHT 9 0 FT
O FT FROM FLAREHEAD AXIS
SPEC IE BACKGROUND MEASURED AT SR-1 EV CQ2
02 i'/,l 21 00+- 0 OO SO 78+- 0 02 0.00*- 0 DC
CO (PPM) 1 :?~- 0 AO 5 Alt— 0 40 791 39+-1063 92
CCU(PPH) 693.+- 150 1049 +- 174 132724 *-248612
HC (PP.T 3. OS*- 2 55 3. 90*- 0 03 309 03+-1259 74
S10T.PPM) O 86*- 0 39 1.61*- 0 04 230.06*- 302 55
S02(PPM! O OP*- 0.00 1.20*- 0.10 44"i. 05+- 470 57
AT SR-1 3Y S02
0. 00+- 0. 00
747. 16+—414. 71
125314
291. G7+-960. 11
365. 13+- 76 20
*20. 07+- 33. 01
COMBUSTION EFFIC IFNCY ( X >
sptcie a* as
CO T9 41— 0 84
ri-: 77 + - o S3
300T 9Q 7 9*- 0 24
701 «L 93 97 2. 02
iJv COS
99. 4l+- 1 99
9^ 77+- 1 37
99 79+- 0.62
98 97+- 3. as
BY 502
99 41+- 0. 96
99. 77*- 1 00
99 79+- O 29
98. >7 + - 2. 2**
DILUTION FAC'SR
3Y 02 - 94 3+— S 7
"JY COS - 369 9+- 361 2
3Y S02 - 349 l+- 63 7
.MEASUREMENTS WITH RAKE A7
SPECIE BACKGROUND
HEIGHT 9. 0 FT
MEASURED
v.)
2i. G3+-
0 00
20 92+-
0. Ow
0 00--
0. 00
CO
> >
3.
0. 40
5. 25+-
0. 40
577 27+—
681. 40
COS
, ppr >
693 +-
150
1 145. ~-
174
\32724 *-
197549
NC
, j
3 09 —
2 55
6. 25+-
0 08
930. SX+-
1489. 19
=oc
i;* pi.
0 06+-
0. ^
2.C i+—
0. 05
338 24+—
292. 70
sor
0 oo+-
C 00
C 904-
0. iC
262
r.K. 45
1.0 FT FROfc FLAREHEAD AXIS
AT SR-1 BY C02 AT SR-1 BY 502
0 00+- C. 00
92S 44+—363. r.4
211661.
*404. 99+-*#»»#»
539. 94+— 139 86
420. 07+- 46 67
:0M3UST!0N EFFICIENCY C/.J
SPECIE JV 02
:c 99 s;+- o. 69
¦13
21 00+-
0. CO 20 83+
-
0. 02
¦3 :ppm;
3 Z9+-
0 ^0 5 23-
-
0 40
COP. ( r P ri)
i»9Ci
150 1071 +
-
174.
r-C" CP«MJ
3 00-~
2. 55 4. 15 +
-
0. 05
r>::.r ppm>
0 06+ —
0. 39 2. 00 +
-
0. 07
332
0 00+~
0 OO 2. 40 +
-
0 12
FKrlCIENCy C/.!
SPEC lb"
BY 02
DY C02
av si
CO "5
47^- 0 77
¥*? 47* 1 ^3
99
47+-
HC 99
77*- 0 96
"9 7?+- 1 42
99
73 + -
SCOT 99
0 bO
99 47+- 1 45
99
4t+-
TT.TAL 9S
>a+— 2 31
93 66+- 4 46
9B
tbf
AT S«»l BY C02
0 0C+- 0 00
717. 56 + - cl36. 92
132721 +-234423
370, 56 + -1252- 09
70S 16+- 701. 60
038. 29 + - B14. 01
At SR-l BY SD2
0. 00+- 0. QC
361. 22+-192 7B
6635A. +-66301
191. 23+-43C 79
335. 29+- 77 77
420. 07 + - 21 00
102 DILUTION factor
0 SI BY as rn 127. 3+- 5
0 99 3Y C02 - 340 J+- 321. 7
O 63 BY 302 - 174. 0 + - 26 0
2 41
E-7
-------
data POINTED ON t'O-OCT-03
03 06:56
FILE FLARE.DAT
POINT
3 DAT?
12-13-32
FLAREHEAD tTER-PRO^O
SIZE
2. 0 IN
MEASUREMENTS WI"H RAKE AT HC1GHT 9 0 FT -1.0 ?T FROM FLAREHEAD AXIS
3°ECIE
B ACKCrlOUND
MEASURED
AT SB=>1 BV CC2
AT SR-1 BY 532
~2 a)
21 OO*-
0. CO
20. ec+-
0.
OS
0 00+- 0. 00
0. 00* 0. oc
CO (PPM)
- 29+-
0 40
4 05+-
0.
40
P.bO. 93 + - 490. 48
136 31+-102 35
C02
693 +-
1 50
10B5 +-
174.
132724 *-226541
79104 +-76.103
hc ;ppm)
3 C8+-
2 b3
3. bs+-
0
05
2a4. U«~:i05. 39
136 10+-540 36
SOOT(PPH)
0 B6+—
0 19
2. B0+-
0
07
662. 90*- *55 74
403 92+- as Ra
S03(DPH!
0. co+-
0 00
2 10+-
0
10
707 31~- 663 17
420 07+- 21 00
COilB1 }ST I ON EFFICIENCY (%)
SPECIE 3 Y 02
CO 9». SO*- 0 40
HC 99 SO*— 0 04
SOOT 9Q 49-r- 0 54
TOTAL 99. IC*- 1 77
BY C02
99 Blf- 0. 70
99 60*- 1 16
99 49*- 1 36
99 1CN— 3. 20
BY S02
99 B0+- 0. 42
99 S0+- 0. B7
99 49*- 0. 60
99 10+- 1.07
DILUTION F4CT0P
3Y 02 - 104. 0*- 10 5
BY CD2 ¦ 335 fl*- 299 9
3Y 502 « 199. 0*- 29 7
MEASUREMENTS WITH RAKE AT HEIGHT 9 0 FT -2.0 FT FRO* FLAREHEAD AXIS
SPECIE
OS '»)
~Z
:c?
-K (PPrl)
S "ICT < PPM >
302(r»PM)
0ACKGRCJNP
MEASURED
00+-
0. 00
ZO »4+-
0 OS
0 oc+- 0 00
3. 29+-
0 40
7 73+-
0 40
L'92 S5 + -22C3. 39
693. ~-
150.
lon +-
*.74
1J2724. +-26912B.
3. OB*-
2. 35
4 SO+-
0 06
37A 69+-1649. BS
0. 86+-
0. 39
2 56—
0. 06
687 1E+- 773.26
0 00+-
0 00
1 00 —
0 10
40E 53+- 463. 60
AT SR*1 BY C02 AT SR-i BY SG2
0. 00*- 0 OO
1070. 50*-706 29
138476 ~-•*••••
601. 67*-
717. C7 + -1P5 87
420. 07+- 42. 01
COMBUSTION EFFICIENCY i*,.;
SPECIC flV DS BY C02 BY SO2
20 79 fe9*- I 90 98 6B*- 4. 24 V0. ifl*- 2 rj
HC '?9 50f I 26 99.38*- 2-06 99 59*- 1.34
3CJ0T 97 50 - 0 *4 ?V. 49*- I. 58 99. 49*- 0 > 3
~OTAL 97 76— 3.67 97 75.#.- 7 79 97 75*- 4 05
DILUTION FACTOR
BY 02 - mo LI*- 16. 4
Bv C02 • 401. 5+- 423. 3
BY SOS " 419. !~- 83. a
E-8
-------
DATA PRINTED Of 2O-DCT-03 03:06:56
FILE FLARE. DAT
PGIIMT:
D^TE. 12-16-SS
FLAHEKEAD EER-PROTO
j « « a si mna*aaaEaaa=4 4 mm
INPUT PLOW CONDITIONS
\/£LCi:ny -» i 50 pps hv - i3io a btu/cu ft
PROPANE u 826 SCFH 33 8 X "5. h LB/HR
N GAS 0 000 SCFH 0 0 X 0 0 LB/H*
NITROGEN 0 647 SCFH AC 7 y. S L LB/HR
S02 0. 470 SCFH 5?90 PPM 0 OB LB/HR
STEAH
SicAM/GaS
SIZE:
0 0
0. ooo
3. 0 IN
.B/HR
MEASUREMENTS WITH RAKE AT HEIGHT
3. O "T J O FT FROM FLAREHCAD AW 15
SPECIE
9ACKCR3UND
*6A3URED
02 a)
21. OO*-
0 00
20 90+-
0 0*
CO iPPM>
A. 63*-
0 40
7 73*-
0 40
CC2CPPM)
7B5 ~-
150
1163 *-
174
MC
0. 86+-
0 39
1 26+—
0. 03
502(PPM)
0 OO*-
0. 00
0 10*-
0. 10
AT SR-l BY COS AT SP*l BY 3C2
0. 00*- 0. 0C 0. 00+- 0. oc
1006. 72+-1275. 94 »#«~*•**
132730 *-234477 1586794. +-**•~#*
ISO. 7B+-104I. 65 1785 64*-##**»«
142. 22 »- 152. 17 1700 15+-»«#»»*
34. 91*- 67.06 419 5B+-4.Q 58
COMBUSTION EFFICIENCY
SPECIE BY 02
CO 99. 19+- t. 21
-»C 9* 89^- 0. 81
- ;cr 99. 3* + - O. 15
»OTAL va <*/~- ? 1 5
(7*
BY C02
99 19*- 2 37
99. 89*- 0 97
99. B9+- 0. 3C
98 97+— J. 4«""i
BY 502
99 19*- 2.
99 39*- 1
99 89*- 0
98 97*— 4.
67
02
34
01
DILUTION FACTOR
BY .">2 ¦ 209 0*- 42. o
i,v :oa . 348. 1*- 321 5
DY S02 * 4194. 8*-46C9 1
MEASUREMENTS WITH RAKE
SPECIE
BACKGROUND
AT HE1FHT ~
MEASURED
0 FT 1.0 FT FROM FLAKEHEAD AXIS
w a.' t "
i ;. oo+-
0 00
2C. 00*-
C. 02
"JO f. "PM)
4. 63+-
0. 40
t02+-
0 40
C02IKPM)
785. *-
150
15 03. +-
174.
(»• *: 1 >
2. 42*-
2. 35
00+-
0. 05
5JOT ;pp* >
C B6*-
0 39
9j+-
0 07
SP21PPM)
0 oo+-
0. 00
i. 30+-
0 10
AT SR"1 Bv C02
O. 00*- 0 00
76. 30*- 193 54
132730. + -177367.
10009+- 529 be
300. 3LJ*- 222 26
238. 90*- 41 45
AT SP®1 BV S02
0. 00*- 0 31
130. 50+-279 38
232522. +-¦****»«
188. 01 * - 8 ^9. 16
667 34+-163. 97
419. 50*- 32. ZS
^~«VBU5TI0N EFFICIENCY (7.)
SPECIE BY OJ
C3 99 94*- 0. 15
MC 99. 92*-- 0. -'1
33CT 7*7 72*-« 0. 20
TOTAL 97. 57 + 0 75
BY C02
tJ7 94*- 0. 19
97 9*'+- O. 4'/
99 71 ~- 0. 44
1tN
TS ritTH
RAKE AT
HEI OH.
3 0 r T
speoi*:
BACKGROUND
MEASURED
02
21 00^-
0. 00
20. 50*-
0 02
'0 (PPM)
A 63 —
0 40
21. 52+ -
0 43
C02OOT (PPM>
0 B6+-
0 39
9 31+-
0 23
302(PPH)
0 00-^-
0. 00
5 40+-
0 27
0.0 FT FROfl FLAKEHEAD AXIS
AT SR-1 Cu2
0. 00+-
81? 54 + -
1J273G + -
429 37*
403. 79 + -
253. &2*"
0 00
186 22
39321.
L -3. 97
95 57
59. 38
AT 5R-I 0Y 50?
Z. 00+- 0 00
1316. °8+-259 01
214840 +-56814
695 90+-309 J9
657. 01 *-133. 50
419. *8 + -- 20
COMQ«-ST1on efficiency
SPECIE 3N J2
CO ,%9 39*- 0 1 5
HC 99 68*- 0 16
SOOT 9* 70+- 0 08
TOTAL 98 77+- 0 37
iV*>
BY CC2
99 39+- 0 32
99 68+- 0 25
99 70*- 0 16
9B. 77*- 0 72
by sor
99 39*- 0 28
9" 6fi- - 0 23
99. 70+-- C 14
98 77+- 0 6A
DILUTION FACTOR
BY 0? - 4 1 0*-
9Y CO? .i *6 9+-
BY so;. - A. 7*-
E-9
-------
DATA PRINTED UN 20-OCT-3n OE 06 56 FILE
POINT 4 DATE: 12-16-82 FL AREWEAD EES -PROTO
FLARE DAT
size
3 J IN
MF.A3L*WEMENT
3 HI*TH
RAKE AT
HEIGHT
3 0 FT
SPECIE
3ACKCRC0NP
MEASURE
;o
35
21 0C+-
0. 00
20. JO+-
0 OS
CO <»P.1>
4. 63 + -
0 40
16. 6*+-
0. 40
C02
0 !.%~-
C. 3?
11 L7+-
0 28
SC2?PPM.
0 0D~-
0 00
7 E0+-
C. 39
-1.0 FT FROM FLA*ENCAD AXIS
AT SR-1 BV C02
O. DO*- 0. 00
929. 99*~ 123. 16
132730 *- 36393.
274. 60*- 1J9. la
496. 94*- 102. 11
33B. 32*- 74. 98
AT 5R»1 BY 502
0. 0O+- 0. 50
691 21*-1^8. 69
164421 *-41979
33V. 97+-190. 60
96®. 4B*-114 29
419. 9B*- 20 95
COMBUSTION EFFICIENCY (7.)
SPECIE BY 02
Cd 99 nd
MEASURED
02 <"/.)
21 00*-
0 00
7.0 70+—
0 02
CO
0 00+-
0 CO
0 70*-
0 10
'JOPIBL'STION EFFICIENCY (V. i
SPECIE 2'. 3i
TO 99 99»- J 11
*C t>9 go — 0 40
SOOT ^9 3- — 0 39
rOTAi. 99 iJ— 0. ="
BY C02
100. 00*- 0. 1 1
99 89*- 0. 90
99 36+— 3 99
99 25— 1. 94
BY 305
loo no*- o ll
99 89+- 0 44
99. 36+- O 62
99 29*- ; 17
DIH-TiON FACTJH
BY 02 ¦ 69 0+- 4. 7
SY CQ2 » 179. 9+- ae. 1
BV SO? - 990 4+— 144 7
E-10
-------
DATA PRINTED ON 20-QCT-B3 08:06:56
FILE FLARE 3AT
POINT
3 DATE. K'-16-B2 FLASE^CAD: EER-PROTO
INPUT TLCU CONDITIONS
.yELOCITV 2 00 TPS HV - 1319
PSOPAWF 2 302 SCFK 36 1 %
N 3AS O OJO SCFh 0. 0 X
NITROCEN 2. 373 SCFM 43. ' Z
302 0 470 SCFH 1332 PPM
0 0TJ/CU FT
2Z. 3 LB/HH
0 0 LB/HR
11 2 LB/HR
0. 08 LB/HR
SIZE:
3.0 IN
STEAM
GTE Art/Q AS
O. 0 Lfi /HR
0. 000
MEASUREHENTS U'TH
htOOD AT
heioht
4 0 F'
SPECIE
BACKCROUND
MEASURED
02 (¦/.)
21 00+-
0 00
20 90+-
0 02
CO (PPM)
5 79+-
0 40
8 12+-
0 40
C02«P?M)
700 +-
150
2424 +-
174
HC
SPECIE 3Y 02 BY COS
CO 99 56+- L> OB 99. S6+- 0. 14
99 67»- 0 Ifi 99 87+- 0. 24
300T 99 8C+- 0 OS 99 80+- 0. 15
TOTAL 99 53+- O 32 99 34+- 0. 52
BY S02
99 86+- O 14
99 87+- 0 23
99. 80+- O 14
99 54»- 0 SI
DILUTION FACTE"?
BV 03 - 41. O— 17
PY CD2 - 79 6+- 19 3
DY SD2 « 149 0+- C6. 2
E-ll
-------
DATA FRINTFD ON 20-0CT-83 OB 06 96
PILE : FLARE. DAT
POINT
6 DATE 12-17-B2
FLAREHEAD. EER-PRQTO
SHE:
3 C IN
INPUT FLOU CONDITIONS
VELOCITY « 2. 00 FPS HV
PROPAf*E 3 302 3CFM 33 9 X
N OAS 0 000 SCFM 0 O 'I
NlTROOEN 2. 393 SCFM 43. 9 X
SG2 0 470 SCFH 1327 PPM
1314 S BTU/Cl* FT
22. 3 LB/HR
0 O LB/HR
11. 3 LB/HR
C OS LB/HR
STEAM
STEAM/OAS
O. 0 LB/HR
0. 000
MEASUREMENTS UITH RAKE AT
HEIGHT
9 0 FT
SPECIE
background
measured
C2 iX)
21. 00*-
0. 00
21 00*—
o f :
(PPM)
3 27*-
0. 40
9. 7*+-
0. 40
C02
309. ~-
1*0.
1469. +-
174.
HC
3. 23—
u. 39
3. bO*-
C. 05
SOOT
2.0 FT F9DH FLAREHEAl) AXIS
AT BR-1 BY C02 AT SH*=1 BY srj2
O. 0O»- 0. 00 0. 00*- 0 00
379. 01*- 2T9- 77 1473. 04+—•~***»
lwi.762 ~-10387V. SL.Tl?'. ~-¦»+*»#»
97. 37«- 404. 02 217 28+-««+» ••
329.63*- 163.39 1300. 43*-943. 19
26. 99+- it 09 104. 99*- 98. 32
SPECIE
:c
HC
saoT
TC "AL
BY 02
0 00*- 0 00
O 00*- O. 00
0 OC 0. 00
0. 00*- 0 00
BY C02
99. 72+- 0. 4J
99. 96 — 0 34
99. 79+— 0 32
99. 43J— 1 07
BY SW
99 72+- 0. 99
97 96*- 0 36
97 7S+- 0, 43
99 43+- I. 3C
DILUTION Fi CTQR
B Y OH - 0. 0+- 0
BY CQ2 - 146. 9*- 6P
BY S02 - 982. 2»- 30:
MEASUREMENTS
with
RAKE .AT
HEIGHT
9. 0 FT
1.0 FT FROM FLAJ5EHEAD Ax IS
SPECIE
SACKCRGi'NO
MEASURED
AT Sfl-l nY C02
AT SR-a By
3
-~2
C2 kX:
21 00 + -
0. OC
20 60+-
0. 02
C. 00*- 0. 00
0 00«-
0
DC
CO ¦ yPin
3. «2 / + -
0. 40
22. »9+-
0 43
1102 52*- 272. 90
0 00*-
0
OC
:02(poh>
98*? + -
190.
2900 *-
174
132762 *- 45390
0 ~-
0
uc 2*- 0 03
9a 67*- o a?
3Y SC2
0 0O+- 0 00
0 C9+- "> 00
-J 00+- 0 00
0 00+- C 06
GILUTIOr FACTOR
by ca - 3i s»-
bv :o; » S6 2*-
BY SOT » 0 O-
"EASL-KEMENTS WITH RAKE
SPECIE.
SACKCROUNID
¦
C 97*-
0 39
9 99->-
0. 29
2;a. 7i+-
SC2
0 (KJ+--
0 00
0 30*-
0 10
7. M+-
CGMBUSTICN
^FFICIENCY (X)
0 <- FT FROM FLAREHCAD JUIS
AT 3R-1 BY CP7 AT 39«1 BY 5C"
0 OO O 00+- 0
633. 93 >*<»•«
SPECIE
CO
HC
;ocr
'OTAL
BY 0.?
96 84«~ 0 31
97 12+- 0 30
99 85+- O 02
93 SI*- 0 61
BY COS
96 .13*- 1
97 1 — 0
99 as*' 0
93 79*- 1
00
03
92
HY S02
96 S2»- ? 91
97 11*- i. 67
99 83*- 0 13
93 77* • 3. 34
36. 6A 3142 78*--
3. 32 104 98*- 34 S9
DJ-^TIJN FACTDft
BY C2 ¦ U 4»- 0 1
BY CCC * 23 3*- 3 0
BY S'->2 ¦ 348.9*- i3l I
E-12
-------
DATA PRINTED ON 20-0CT-B3
POINT: 6 DATE 12-17-82
OB 06 36
PILE : FLARE DAT
FLAREHEAD. EER-PAOTO
9I2E
3 0 IN
MEASUREMENTS WITH RAKE AT HEIGHT
5 0
SPECIE
32 (X)
CO !PPM>
coscppN)
HC iPPK)
SOOTifP*)
SuKFPM?
BACKGROUND
21 00*-
3 27*-
">89. *-
~ 23*-
0 97*-
0 00+-
0 oo
0 40
190.
2. r*
0 3°
0 00
MEASURED
20 20+-
51. *3*-
4732 *-
36. 00*-
27 96+-
0. 00*-
0 02
1 03
174.
0 45
0 70
0 10
-1.0 FT FROM PLAREXFAD AX*S
AT SR»1 sy CP2 AT SR«1 BY S02
O ^*- 0 OC 0 OC*- 0 00
1340 3«*- 263 34 0.00*- 0 00
132762. ~- 28962. 0. *- 0
J 048. 39*- 24] 66 0.00*- 0 CO
861. 96*- 166 21 0. O0+- 0 OC
0.00*- 3 19 0 00*- 0 00
COMBUSTION EFFICIENCY (X)
SPECIE FY 02 I'.Y CC2
CO 913. P7*- 0.17 96 ia? i 0.44
HC 99 23* - 0.16 99.23+- 0.34
SOOT 99 37*- 0. U "9. 37+- 0. 26
TOTAL ?7. 47*— 0. 44 97 47 + - 1. 02
BY 802
0. 00*- 0. 00
0. 00*— 0. 00
C. 00*- 0 00
0 00*- 0. 00
DILUTION FACTOR
BY 02 - 23 3--
BY COS " 30. 9*-
BY 502 - 0. 0*—
0. 7
4 3
0 0
hEASOREKENTS WITH RAKE AT HEIGHT 3 0 F. -2 0 FT FRQH FLAREHEAD AXIS
SPECIE
02
:~ (PPMJ
C02
HC < PPM)
SOOT(PPM;
SP2(PPM*
DACKOROUND
21. 00*- 0 00
3. «7>-
5B9 ~ -
3 53 + -
O 97+-
O 00+-
G 40
150.
k. 53
G 39
0 Of>
MEASURED
20 63+-
B. B9 + -
2170 +-
7. 30+-
16. J6+-
i 00+-
AT 5F>1 BY C02 AT SR-1 BY 502
0. 02
0 4C
i*s.
0. 09
0 40
0 10
0 oo*- 0. 00
473. : 1+" 1Q2. 77
132762. + - 6k463
339. 96+- 314. 33
1270 70*- 408. 43
93. 60 f- 30. 83
0. 00*- n 00
593.28+-200
166368. -^-6^013.
431 20*—363 93
1393. 46*-40C. 96
104. 98+- 1C. 30
COMBUSTION EFFICIENCY (%*
S®EC*E SY 02
CO 99 63*— 0 1*
HC 9* 73*- 0 24
SOUT 99 06*- 0 34
• nr^L 9B 44-r- 0 7?
BY CC2
99 63*- 0 31
99.73*- 0 36
99 06*- 0 74
9R 44*- I Sy
B7 502
99 63*- 0 26
99 73+- 0 3?
99 06+— 0 61
95 44*- ! 19
DUUTTOrJ FACTOR
BY 02 - 33 0*- 3 1
BY C02 - 32 'j*- Z'ml 5
3Y 902 - 104 0+- 20 3
E-13
-------
APPENDIX F
GRAPHS OF LOCAL COMBUSTION EFFICIENCIES AND DILUTION PACTORS
Concentrations of CO, CO2. HC, O2, soot, and SOg were measured at several
radial positions for most of the flare flames studied. Local dilution factors
can ba calculated «s shown 1n Appendix C using the measured concentrations of
O^i CO2, soot, and SO2. Eacn dilution factor can then be used to calculate
a local combustion efficiency. Examples of the numerical results of the cal-
culations are shown in Appendix E and the graphical results 1n this Appendix.
Two types of graphs a>-e sho^n. Combustion efficiencies are plotted for
each condition as a function of radial prcbe position. The base of the axes
of each graph denote the axial position of probe sampling above the flare
head and the horizontal axes show the radial position of probe sampling.
Dilution factors a"e plotted In a similar fashion. In some cases, tne
dilution factors at the edge of the plume are large and are omitted so that
the scale of tne graph Is not excessively compressed.
The dilution factors calculatsd using different species can vary consid-
erably because of the range of accuracy for different species. The dilution
factors calculated from each species are shown 1n the graphs. However, the
line is d.awn through those determined from CO2 concentrations. These values
are thought to be the most reliable. However, calculation of dil itlon factors
based on CO? assumes a complete canon balance «nd verification of a carbon
balance was one of the objectives 0" this program. Carbon balances have been
verified In this program using SO2 a tracer and by capturing the entire
flare plume in a hood.
The local combustion efficiencies calculated using different dilution
factors estimated from different spf.les concentrations are similar. (See
examples in Appendix E.) Tor consistency, the local cenbustlon efficiencies
calculated using C02 as the tracer a'€ p'otted.
-------
11
10
8
100
98
o CO
~ HC
OS00T
o TQT/*L
100
99
98
O w1 ,
w «
z» u H
UUI
3-Inch Flare Head
0,5 FFS
56 Percent C3Hg (1310 Btu/ft3
0.0 lb.Steam/lb.Fuel
CI
-8 -6 -4 -2 0 2 4 6 8
Radial Distance (ft)
figure F-:. Local combustion efficiencies of the 3-^nch ?ER test flare
head burning 56 percent propane ir nitrogen at 0.r> ft/sec
with 0.0 lb steam per lb of fuel.
-------
!0-
60&-
400-
•c3Q-
0*-
* 8-
60Gb-
40C »
20C-
60C-
o 50C(»
+j
S. 40Ch
3 30(J-
3 20Qh
100
I
By
O S0?
o co2
AO,
<>
^ a a
o
3-Inch Flare Head
0.5 FPS
56 Percent C-jHg (1310 Btu/ft3)
1 I I I
0.0 lb.Steam/lb.Fuel
tin
-10 -8 -e -4-2 0 2
Radial Distance (ft)
Figure r-2. Dilution factor of the 3-inch TER test flar° head otirring
56 percent propar.e in nitrogen at 0.5 ft/sec with 0.0 lb
steam per It of fuel.
F-
-------
O TOTAL
UJ OJ
O QJ
2-Inch Flare Head
2,C FPS
56 Percent C3Hg (1315 Btu/ftJ)
0.0 lb.Steam/lb.Fuel
-3 0 2
Radial Distance (ft)
"Igure F-3. Local combustion efficiencies of the 3-incn LEP test flare
head burning 56 percent propane in nitrogon at £.0 ft/sec
with 0.0 lb steair per lb of fuel.
F-4
-------
13
300
12
200
100
11
300
200
9
100
8
7
5 100
6
3-inch Flare Heed
2.0 FPS
56 Percent C^Hg (1315 Btu/ft
O.U lb.Steam/lb.Fuel
4
3
8
6
2
4
0
6
4
8
Radial Distance (ft>
Figure F-4. Dilution factor of the 3-inch EER te?.: flare head burniny
56 percent propane in nitrogen at 2.J ft/sec with 0.0 lb
steam per lb of fuel.
-------
APPENDIX G
QUALITY ASSURANCE
G.O «SESSMENT OF DATA QUALITY
The date, taken 1n this study was carefully eva^'&ted to establish the maxi-
mum error bounds. Procedures used to take the data are described 1n Appendix B.
Reduction of the dcta ?nd analysis of errors are discussed in Section 3.3.
Examples of the data and analysis prcct .rres are presented in Appendices C, D,
E, and F.
Data
The expected precision, accuracy, and completeness of the drta Is shown In
Table G-l. The representativeness of the data Is demonstrated by the correla-
tions and statistical analysis of these correlations 1n the text of the report.
The reproducibility of the data wa.» demonstrated b;- repeating all *he conditions
on the 6 Inch flare head which had resul^d in low combustion efficiencies. In
addition, material balances were closed for * number of test conditions (see
Section 3.2).
Llritatlons
The (*ata Is strictly limited to the cor'iltlons of this study. These were:
• Propane-nitrogen mixtures.
• 6, 1? Inch -Imple pipe flares.
• 3-12 Inch commercial pipe flares.
• Velocities from 0.2 to 428 ft/se:ond.
• Gas heating values from 386 - 2350 Btu/ft^.
• No pilot flcmes.
• Steam Injection to 1 lh steam/lb fuel.
6.1 Results of Audit
This program wes Initiated prior tc riscal year 1982 Consequently, no
Quality Assurance Project Han was submitted nor were any audits conducted.
However, the nature of the program and use of the results required ihat .trict
QC/QA procedures be appHeo. lhose procedures have been fully dr uaented in
tne text append^es r
-------
TABLE 6-1. MEASUREMENT PRECISION, ACCURAC/, AND COMPLETENESS
Measurement
Method
Reference
Experi-
menta1
Condition
Precision
Accuracy
Com-
plete-
ness
Flow Rate
Cali brated
Orifices
None
Gases,
Steam
Readi ng
+ 5%
Readi ng
100%
name
Structure
Visual ,
Photograph
None
Observation
+10*
Ri'ddi ng
+20%
Reading
100%
02
P'Ka-
Begnetic
FPA
Spec, 3
Hame
+Q.02S
~ 0.04S
90S
CO
MIR
EPA
Spec. 3
Flame
+4 ppm
*B ppn
90S
C02
NQ1A
EPA
Spec. 3
Flame
2,0.02%
+0.04%
90%
HC
FID
None
F1 ame
_+0.5 pptn
M.O ppn
90%
S02
FPO
E?A
Spec. 2
F1 ame
+20%
Reading
+40%
Reading
90%
SO 2
Titration
EPA
Method 6
Flam*
~10*
Readino
+20Z
Reading
90%
NO,
Cnem.
Lumin.
EPA
Spec. 2
F1
-------
G.2 Quality Problems and Solutions
Several problems with quality control were encountered and corrected
during the course of this program.
"the concentration of car'jcn species in the ambient air was recognized
to te a problem and a procedure is desc-ibed in Section 3.3 to correct ro»
it.
Determination of soot concentration by weighing proved difficult.
Accurate measurements of soot were obtained hy burning the material from
prebaked filters.
The instrument for measurement of SO2 proved unsatisfactory. It was
returned to the factory for repair and performance was improved but was still
less than desired. Instrument measurements of S0£ were supplemented by
absorption and titration of SO, throughout these tests.
Continuous monitoring of emissions from the flare flame proved to be
undesirable because of flame fluctuations. Gas samples were drawn through
Teflori® 11nes into Tedlai®*bags for 20 minutes to average flame fluctuations,
mixed and analyzed using continuous analyzers.
A video recorder wab used tc record the structure of the flame flames.
However, the spaclal and temporal resolution were insufficient to make this
technique useful. Excellent results were obtained using still photography
ana high speed motion pictures.
G-3
-------
|