You are on page 1of 16

Matthew Martin Recent Developments and Current Insights in Ethylene Flare Technology

Senior Product Line Manager The Impact of Burner Design, Vent Gas Variability and Ancillary
Equipment

September 13, 2016  2016 AFRC Industrial Combustion Symposium Kauai, Hawaii
© 2016 UOP LLC. A Honeywell Company All rights reserved.
1

Introduction
• Multipoint Ground Flares
(MPGFs) are used for the safe
disposal of vent gas
• The burners are generally
pressure assisted, which mixes
the fuel and air to produce
flames without smoke
• Radiation fence is used to hide
the flame and reduce the sterile
zone around the flare
• Hundreds to thousands of
burners are used in concert
2

Introduction: 40 CFR 60.18 and AMEL


• The U.S. Title 40: Protection of
Environment (40 CFR 60.18)
specifies permissible flare tip
exit velocity
• MPGF burners do not meet this
regulation
• An Alternative Means of
Emissions Limitation (AMEL)
must be requested with
evidence provided of the
efficacy

40 CFR Conforming
3

Review: NHVcz and LFLcz


• Combustion zone Net Heating
Value (NHVcz) and Combustion
zone Lower Flammability Limit
(LFLcz)
- Appear to correlate with
combustion efficiency (CE) and
destruction removal efficiency
(DRE) for 40 CFR conformant
flares
- Do not appear to correlate with
CE or DRE for pressure assisted
flares Combustion efficiency versus the ratio of NHVCZ to NHVVG-LFL

• These parameters appear to be


Source: ‘Parameters for Properly Designed and Operated Flares’, Report for Flare
interpreted as being a measure Review Panel, Prepared by U.S. EPA Office of Air Quality Planning and
Standard, 2012
of flame stability
4

Review: Burner Requirements


• Thermal Radiation
• Flame Shape
• Flame Stability
- Exit Velocity
- Vent Gas Variability
• Cross-Lighting
• Smokeless Operation
5

Review of Prior Data: NHVCZ and LFLCZ

• GLM ANOVA used to


analyze parameters
- p-value NHVCZ = 0.762
- p-value LFLCZ = 0.809
- R2 = 4.86%
• Vent gas nitrogen content
R2 = 19.2%

Pressure-assisted flare data .The blue diamonds represent stable steam or air
assisted flares. The red boxes represent unstable steam or air assisted flares. The
orange circles represent stable non-Callidus pressure assisted burners. The yellow
circles represent unstable non-Callidus pressure assisted burners. The light green
triangles represent stable Callidus MP4U burners.

Vent Gas Nitrogen Content Predicts CE Better than NHVCZ and LFLCZ
6

Test Plan and Results


• Two burner test programs
• Program A tests CE and DRE against
low calorific value vent gas
- 500 Btu/SCF flare gas mixture of C2H4
and N2
- (2) Pressure Targets
 15 Psig
 5 Psig
- (3) 20-minute duration runs at each
pressure
 One hour of high pressure
Mole Fraction of Nitrogen in Vent Gas Resulting in Unstable Flame vs
operation Mole Weight of Flammable Components
 One hour of low pressure operation 0.76

- Establish minimum heating value for 0.74

Mole Fraction of Nitrogen in Vent Gas Resulting in Unstable Flame


flame stability
- Pilot turned off after initial ignition 0.72

- Online composition verified by third 0.7

party measurements
0.68

• Program B tests cross-lighting, 0.66

stability and flame length for 0.64

conventional and new (alternate)


burner types 0.62
0 10 20 30 40 50
Mole Weight of Flammable Components in Vent Gas Considered in Absence of Nitrogen
60 70
7

Results: Stability for Low Heating Value Gas


• Stability is correlated with
Burner Type Stable Flame? NHVcz LFLcz
MP4U TRUE 481 8.44

burner type MP4U TRUE 511 7.95


MP4U TRUE 508 7.99

• The correlation is statistically MP4U TRUE 499 8.14


MP4U TRUE 497 8.16
significant MP4U TRUE 509 7.97

• Practically, the testing was not


Other FALSE 595 7.6
Other FALSE 589 7.8

carried out under the same Other


Other
FALSE
FALSE
746
650
6.6
7.6
conditions
• More research is needed
8

Physical Testing: Results


Comparison of Vent Gas at Burner Flammability Limit
Callidus MP4U Other
Compound Volume % Compound Volume %
C2H4 25.00% C3H6 28.32%
N2 75.00% N2 71.68%

LHV (Btu/SCF) 375 LHV (Btu/SCF) 618


LFLcz 11.00% LFLcz 7.06%
NHVcz (Btu/SCF) 375 NHVcz (Btu/SCF) 618
LFL 2.75% LFL 2.00%
MW 28.03 MW 32.00
NHVlfl-vg 10.31 NHVlfl-vg 12.36
NHVratio 9.09 NHVratio 14.16
9

Results: Fraction of Heat Radiated from Flame


• Statistically significant
difference in burner types
• New, alternate design
- 15% less radiation with 95%
Histogram of Normalized Fraction Heat
confidence Normal
Bu rn er Type
- Lower standard deviation in 2.0 Altern ate
Con ven tion al

radiation Mean StDev N


0.7597 0.2773 201
1 0.4272 21 2
1 .5

• Near field radiation is

Density
1 .0
anisotropic
0.5

0.0
0.0 0.3 0.6 0.9 1 .2 1 .5 1 .8 2.1
Normalized Fraction H eat
10

Results: Flame Length


• Statistically significant
difference in burner types
• New, alternate design as
significantly shorter flames Histogram of Normalize Flame Length
Normal

under the same operating 20 Bu rn er Type


Altern ate
Con ven tion al

conditions 15
Mean StDev N
0.4685 0.02279 1 0
1 0.04828 1 0

Density
10

0
0.4 0.5 0.6 0.7 0.8 0.9 1 .0 1 .1
Normalize Flame Length
11

Physical Testing: 500 Btu/SCF


12

Physical Testing: Comparison

33% C2H4 / 67% N2 100% C2H4


13

Ancillary Equipment
• The burners and fence may receive most of the design attention
• The liquid carryover from the knockout drum can cause a host of
problems
- Droplets exiting the flare result in a significant increase in radiation
- Droplets exiting the flare result in a safety hazard to the surroundings
- If the droplets are not ignited they present a contamination hazard
• The cost of a larger size knockout drum pump can be many times
less than the repair cost of a damaged flare
• The knockout drum should be close to the flare to prevent
condensation
• Consider oversizing your knockout drum pumping system

The Knockout Drum can be the Failure Point


14

Conclusions
• Burner design is shown to be statistically and practically significant
to flare performance for:
- Smokeless operation
- Thermal radiation
- Flame length
- Low heating value cross-lighting
- Low heating value flame stability
• A single strict parameter of combustion zone heating value is not
adequate to describe burner performance
• It may be possible to draw general conclusions about stability
versus inert content of the vent gas across burner types
• A new alternate burner type is shown to have potentially lower
radiation and significantly shorter flames for the same operating
conditions
15

UOP 7116-15

You might also like