June
2016
HYDROCARBON
ENGINEERING
35
The FlareSentry directly monitors flare CE, eliminating
inaccuracies associated with the current practice of monitoring
indirect parameters (heating value, velocity, waste stream
speciation, etc). Autonomous operation means the FlareSentry
needs no aiming or manual data reduction. Remote mounting
and measurement means there is no contact with corrosive
process gases. This makes the FlareSentry easy to maintain with a
low long term cost of ownership due to lower operation and
maintenance costs over the life of the system, when compared
to traditional flare monitoring technologies.
The feasibility of the FlareSentry system was first
demonstrated in a bench scale test and was recently tested on
full scale flares. The full scale experiment was conducted in
November 2014 at Zeeco, Inc.'s flare test facility near Tulsa,
Oklahoma. Three full scale fares were tested: a 16 in.
steam-assisted flare (Zeeco Model QFS), a 10 in. air-assisted
flare (Zeeco® Model AFDS) and a multi-point sonic flare (also
referred to as a pressure-assisted flare and used as a ground
flare, Zeeco® Model MPGF).
For this experiment, flares were evaluated by the FlareSentry
at a distance of 300 ft from the base of the flare stacks. In order
to validate this newmethod, simultaneous extractive sampling
was performed to evaluate the DRE and CE of the flares. For the
extractive sampling, an inductor with a sampling hood was
suspended over the flare using a crane, and a portion of the gases
captured by the inductor was extracted and transported via a
heated sampling line to a monitoring trailer. Inside the trailer, a
contracted stack tester continuously analysed samples for
combustion products carbon dioxide and carbon monoxide,
unburned hydrocarbon and oxygen. The test methods and
procedures used were consistent with standard EPA methods for
stack testing.
39 test runs were performed covering a CE range of
approximately 60 - 100%. The results from the newmethod
showed a strong agreement with the extractive methods. The
performance of this new flare monitoring technology is
demonstrated by comparing the CE measured by FlareSentry and
CE measured by the extractive method, as shown in Table 1.
As Table 1 indicates, the results from the extractive sampling
and the FlareSentry were highly correlated, where r
2
= 0.9856.
The average difference in two results is 0.50% and the
FlareSentry system showed an excellent repeatability.
The FlareSentry system also provides a metric called smoke
index (SI) to indicate the level of smoke in the flame. The
real time CE and SI outputs from the FlareSentry can be
integrated into the flare control system for closed loop
Table 1.
Flare CE validation test results
Test
no.
Flare
type
Fuel
Fuel
flow rate
(lb/hr)
Stoichiometric
air
Steam/
HC
(lb/lb)
CZNHV
(Btu/ft
3
)
CE-extractive
method
CE-new
method
CE
difference
Smoke
index
Avg. O
2
in extracted
sample
1
AFDS Propane
(100%)
7994
33.29%
259
99.94%
97.40% -2.54% 2.85 21.13%
2
AFDS Propane
(100%)
7994
33.29%
259
99.99%
98.80% -1.19% 2.46 19.45%
3
AFDS Propane
(100%)
7994
33.29%
259
99.98%
98.70% -1.28% 4.58 19.37%
4 AFDS Propane
(100%)
6670
39.89%
221
99.99%
98.80% -1.19% 2.87 17.63%
5
AFDS Propane
(100%)
6670
39.89%
221
99.97%
98.60% -1.37% 2.70 18.84%
6 AFDS Propane
(100%)
5278
50.42%
178
99.97%
99.20% -0.77% 2.66 19.83%
7
AFDS Propane
(100%)
5278
50.42%
178
99.95%
99.20% -0.75% 2.50 20.03%
8 AFDS Propane
(100%)
3063
86.87%
107
99.33%
99.00% -0.33% 0.72 20.53%
9 AFDS Propane
(100%)
3063
86.87%
107
99.77%
98.70% -1.07% 1.44 18.94%
17 QFS Propylene
(100%)
4910
0.48 1031
99.86%
99.00% -0.86% 3.99 19.93%
18 QFS Propylene
(100%)
4910
0.48 1031
99.90%
99.10% -0.80% 2.24 19.98%
21
MPGF Propane
(100%)
5079
100.00% 99.90% -0.10% 0.24 18.77%
22 MPGF Propane
(100%)
5079
100.00% 99.70% -0.30% 0.27 18.07%
23 MPGF Propylene
(100%)
4952
100.00% 99.90% -0.10% 1.41
17.92%
24 MPGF Propylene
(100%)
4952
100.00% 99.90% -0.10% 1.36 17.38%
25 MPGF Propane/
N
2
(50/50)
2448
99.97%
99.30% -0.67% 0.23 19.48%
