Download notice concerning copyright restrictions

NOTICE CONCERNING COPYRIGHT
RESTRICTIONS
This document may contain copyrighted materials. These materials have
been made available for use in research, teaching, and private study, but
may not be used for any commercial purpose. Users may not otherwise
copy, reproduce, retransmit, distribute, publish, commercially exploit or
otherwise transfer any material.
The copyright law of the United States (Title 17, United States Code)
governs the making of photocopies or other reproductions of copyrighted
material.
Under certain conditions specified in the law, libraries and archives are
authorized to furnish a photocopy or other reproduction. One of these
specific conditions is that the photocopy or reproduction is not to be "used
for any purpose other than private study, scholarship, or research." If a
user makes a request for, or later uses, a photocopy or reproduction for
purposes in excess of "fair use," that user may be liable for copyright
infringement.
This institution reserves the right to refuse to accept a copying order if, in
its judgment, fulfillment of the order would involve violation of copyright
law.
Geothermal Resources Council Transactions,Vol. 24, September 24-27, 2000
Field Test of LEAMS Drilling and Well-Test Separator
JohnFinger,’ Allan Sattler, Gary Whitlow, Ron Jacobson’, Tom Champness*,
D.B. J ~ n gW.T.
, ~ Howard3, Paul Spielman4
‘Sandia National labora~ories,
*Drill Cool Systems,
3Two-PhaseEngineering & Research,
4 COperating
~ ~ Company,
~
Inc.
ABSTRACT
The LEAMS (Low EmissionsAtmosphericMetering Separator) is a new device that improves the
separation of vapor, liquid, and entrained solids during discharge of a geothermal well. This paper
describes the first field test of the prototype, which
showed greatly reduced carryover compared to a conventional cyclone separator.
Back~round
When a geothermal well is tested or produced, the
fluid is sometimes released as a two-phase mixture of
vapor and liquid. Controlling this discharge requires
passing it through a steam separator, so that the vapor
can be vented to atmosphereand the liquid can be disposed or re-injected to the reservoir. In the conventional
and widely used cyclone separator, a significant part
of the liquid can be entrained or suspended in the vapor and borne away from the separator to fall on the
surrounding area. This “carryover” can contain silica, salts,
boron, arsenic, and, in the case of hydrogen sulfide abatement,
concen~atedcaustic and chemical by-products that may be
harmful to agriculture, equipment, or the environment and can
be expensive to remediate.
In a cyclone, or centrifugal, separator (Figure 1) fluid produced from the well enters the separator tangentially to the
cylind~calsurface through the h o ~ z o n tube
~ l about halfway
up the side of the tank. The fluid then forms a vortex inside the
separator with the steam escaping out of the open top of the
tank and the liquid collecting in the bottom of the tank, where it
drains out through a line that usually leads to a weir box for
measu~mentof the flow rate.
The LEAMS (Low Emission Atmospheric Metering Separator, Figure 2, overleaf) is a new concept in flow treatment that
uses internal baffles and diverters to reduce the amount of
carryover emitted during tes~ng.’D e v e ~ o p ~ eof
n ta LEAMS
prototype was funded by the Geothermal Drilling Organization
Figure 1. Test conf~~uration:
cyclone separator on left, LEAMS on right.
(GDO)as a cost-shared project between the U. S. Department
of Energy (through Sandia National Laboratories) and industry
(Two-Phase E n g ~ n ~Drill
~ ~Cool
g , Systems, and Cos0 Operating Company). After completion of this prototype by DCS
and TPE, Sandia managed a field test to evaluate its efficiency
at the Cos0 geothermal field.
Field test plan
The general nature of the test was to produce the same
amount of fluid through a standardcyclone separator and through
the LEAMS and to measure the amount of carryover from each
device. Canyover would be measured both directly, by using a
probe that measures particle flux in the exhaust streams and by
placing rain gauges around the separators during the tests, and
indirectly, by measuring the amount of liquid draining into the
weir box from the sep~ators.In the latter ~ e a s u ~ ~ ethen t ,
amount of brine flowing from the separators would show how
67
Finger, et. a/.
Clean Steam outlet
,
Figure 3, JonasParticle Monitor mounted in LEAMS exhaust
me~~;
(sensing element of Probe is at left end of j ~ s f ~ ~ note
anemometer above Jonas Probe).
Figure 2. Schematic layout of LEAMS (from Reference 1).
much liquid was removed from the exhaust streams and thus by
Tmt procedure and results: Test set-up began on 1 Feb 00
subtraction,because the total liquid content of the fluid is known
and continued on 2 Feb, with the first flow test on the afternoon of
from the enthalpy, the amount of liquid Iost to carryover.
2 Feb. FIow through the cyclone Separator reached a maximum
I n s t r u ~ e n ~ t and
i ~ ndata c ~ l ~ ~ ~ tData
i o ncollection
:
for
value of 2 7 7 , ~~u n dper
s hour (277 KPH) with an averagevalue
this test used the following transducers supplied by Sandia:
of 270 KPH?which produced abundant carryover but no flow
through the weir box. In fact, after more than Go hours of flow, so
* Lip pressure in the James tube (which gives totaf mass flow
little liquid had collected in the bottom of the separator that it never
rate into the separator),
reached the level of the drain line leading to the weir box. It is
* Pressure and temperature in the blooie line,
likely that the vortex inside the cyclone wauld have continued to
Pressure differential across the orifice plate (which gives
empty the liquid, no matter how long flow lasted.
total mass flow from the well,
In a more qu~lita~jve
evaluation, the ~~evajling
south-southeast
wind
blew
a
plume
of
steam
and
liquid
away
from the
Liquid level in the weir box ~ w h ~ cgives
h flow rate for liqcyclone
walking
underneath
this
plume
was
very
much
like
uid collected in the separator),
being
in
a
mode~ate
to
heavy
rain.
Because
rain
gauges
were
e FIow rate into weir box (same as above, except measured with
placed at the four cardinal directions (N,E, S, W), there was
a magne~cflow meter on the drain line into the weir box),
not a gauge directly underneath the plume, but the north gauge
Paddle meter in the overflow line into the weir box,
collected 0.8 inches of water in just over two hours of flow.
e Sound pressure level (LEAMS is designed to be quieter than
There were also large puddles of standing water on the gravel
pad
around the separators.
the cyclone),
The
conclusion was that carryover was essentially l
~
e U-tube manometer to measure internal pressure in the
that is, all the liquid produced into the separator was carried
LEAMS,
away by the steam. Given a total flow rate of 270 KPH and an
0
Anemometer in the exhaust stream, and
enthalpy of 997 ~ T U / ~ bfor
m the fluid, the liquid-fraction flow
* Data acquisition and display system based on L a b V I ~ ~ rate of 41 KPH at atmosphe~~c
pressure can be calculated. Using water density at a temperature of 2W°F (boiling point of
software.
water at the wellsite elevation of 3721 feet), the brine flow rate
Jonas, Inc. of Wilmington,Delaware, supplied a particle moni(which we take to be the minimum carryover, as discussed
tor (Figure 3) used to measure size, number, and distribution of
above)
is equivalent to 85 gallons per minute (gpm),
water droplets at specific points in the exhausts of the LEAMS
Data from the Jonas probe indicated liquid flow out the top
and the cyclone. Jts principle of operation is to measure the imof the cyclone to be 74.6 KPH, or almost twice the liquid fracpact (kinetic ener~y)of each particle on a probe and then, by
tion of the fluid at atm~sphericpressure.2 A larger amount of
using a separate measurement of velocity from an anemometer,
liquid is to be expected because the two-phase fluid is being
calculate the mass of each particle. This point mass-flow rate
discharged into cooler atmos~he~e
(ambient air tem~erature
can then be ~ n ~ ~over
~ the
t etotal
d open area of each separatur
ranged from 57-80*F during the tests), which will lead to conto get totd mass flow of liquid particles out ofthe separator (total
densation of some part of the vapor. Transient conditions and
open areas = cyclone 109.3 ft2,LEAMS - 121.O ft2).
0
-
68
~
,
Finger, et. a!.
sucked inward by the venturi effect, thus drawing cool air into
the LEAMS. Minor leaks in the LEAMS were observed both
before and during testing; which meant that measured brine flow
under-representedthe actual amount of liquid, but thc lcaks wcre
fairly small, estimated at 2-5 gpm from water level drawdown.
The LEAMS test was highly encouraging from the point
that almost no droplets could be felt in walking under the plume,
c o n ~ ~ by
ed
the fact that no measurable water collected in the
rain gauges during total flow duration of more than two hours.
The particle-size distribution data indicate very strong weighting of the emitted LEAMS particles toward smaller sizes, with
peaks at 50 and 75 microns and a rapidly decreasing tail-off at
increased particle size. The majority of emitted particles may
evaporate mostly in the immediate vicinity of the stack. The
fact that the particle probe detected a few particles in the 500micron range is inconsistent with the failure to feel any droplets
or to detect any in the rain gauge. The fact that no solid material (crystallizeddebridscale) was felt or detected in the vicinity
of the LEAMS during the five LEAMS tests strongly suggests
that any solid particulate emitted from the LEAMS is not
carryover. Despite the uncertainties described above, all of the
measured test parameters were reasonably consistent with each
other and the LEAMS was clearly much more effective than
the cyclone at removing carryover from the exhaust vapor.
Sound pressure level measure~entsshowed the LEAMS to
be somewhat noisier than the cyclone. Sound pressure levels
were 97 dB for the cyclone and 100dB for the LEAMS at equal
distances. The LEAMS was expected to be quieter than the
cyclone, because of the baffles and diverters inside the box, but
it may be that vibration in the flat-panel sides of the separator
accounted for the increased noise. This notion is supported by
the measurement of I. 1-psi internal pressure in LEAMS and by
observation of standing waves in the steam plume when it traveled down the sides of the LEAMS. The inward deflection of
the segmented rubber gasket at the James-tube entry port would
also expose a larger opening and provide a noise path. Although
the measured difference is small, the entry-port situation is
readily remedied.
After completion of the LEAMS testing, the flow line was
reattached to the cyclone (because Cos0 Operating Company
has to leave it that way) and an additional test was run in the
cyclone. The additional cyclone test was performed to get another data point for the particle monitor, because its ex~apola~on
from a point measurement to an area measurement depends on
uniform flow over the outlet area. This appeared to be reasonably accurate for the LEAMS, but was not true at all for the
cyclone. Because of the swirling effect of the fluid’s tangential
entry, most of the particles leaving the cyclone came out near
the periphery, so that the original reading taken there would
overstate the total amount of carryover. This was verified by
readings taken approximately 213 of a radius in from the edge;
values there were much less than those near the edge and were,
~ t s the LEAMS. Bein fact, less than the ~ e ~ s u r e m eabove
cause the flow varies so greatly across the exhaust area of the
cyclone, extrapolation of the Jonas probe’s point readings to
the entire area is problematic. If enough point readings could
be taken to thoroughly map the velocity distribution across the
limited instrumentation, however, make the magnitude of this
increased quantity uncertain.
Limited particle-size distribution data for the cyclone show
peaks at one and two hundred microns, with nuiiierous particles
from 500 microns to 3000 microns (Reference 2). Marble size
droplets were observed under the exhaust plume of the cyclone.
During the cyclone test, control valves from the well were
t ~ o ~ l back
e d in an attempt to reach a flow rate low enough that
no liquid was being carried from the separator. Although Jonas
probe data indicated a reduction in carryover at the lower flow
rate, wellhead pressures became high enough that the valves
could not be closed any further. Data were also incomplete
because the James tube flow measurement requires a minimum
flow rate to read properly, and at this rate a signitkant amount
of liquid was still coming from the cyclone. [It should be mentioned that cyclones are generally designed to be “as large as
practical” and the size l i ~ t a t i o nis a matter of the largest dimension that can be transported on highways by trucks. The
maximum diameter of a cyclone is usually limited to 12 feet, so
if the steam flow to be measured is larger than can be handled
by this size, the only alternative is to use more than one cyclone
in parallel.] At this point, testing on the cyclone was suspended
and test instrumentation was moved to the LEAMS.
First flow test on the LEAMS was on 3 Feb and testing continued on 4 February. The LEAMS was operated for five flow
intervals with the Jonas particle probe at different locations in
the exhaust stream. At each of the probe locations, liquid mass
flow rate was substantially reduced from the cyclone measurements. Average exhaust measurements showed liquid mass flow
of 37.3 KPH (compared to 74.6 KPH for the cyclone), but particle size was much smaller than in the cyclone, which allowed
the droplets to evaporate before falling to the ground. Total liquid output from the LEAMS was, however, greater than the
cyclone. In addition to the liquid in the exhaust, the brine output from the LEAMS ~ o u g the
h weir box was greater than 50
KPH for each of the five flow intervals. There are two major
possibilities for this result: the LEAMS created more condensation within its box than the cyclone did in its chamber or the
particle-monitor measurements were in substantial error. Possible reasons for particle-monitorerror are discussed below, but
the measurements are generally assumed to be correct.
Since the liquid fraction of the steam was approximately41
KPH, and there was no liquid output through the weir box, the
liquid output of the cyclone exhaust must have comprised entrained liquid and condensed vapor, in roughly equal amounts.
This means that all minerals or impurities in the liquid escaped
from the cyclone and were deposited in the area. The LEAMS,
however, is specifically designed to separate larger particles of
entrained liquid, so it is reasonable to believe that most of the
liquid in its exhaust stream is condensed vapor, which would
contain no impurities. This still indicates that somewhat more
liquid was condensed in the LEAMS than in the cyclone, which
could have at least three causes: the larger surface area of the
LEAMS, higher pressure inside the LEAMS (see discussion of
noise below), and the larger opening where the James tube enters the LEAMS. This entry port had a segmented (split) rubber
gasket around the James tube, but the gasket was clearly being
69
Finger, et. a/.
cyclone opening, then a weighting factor for point readings at
any position could be derived, but time and test budget did not
permit that many data points.
More resolution of rain-gauge measurements also resulted
from placement of 11 gauges around the northeast and northwest quadrants of the cyclone during its second test. Prevailing
wind was once again from the southeast, so gauges in the northeast quadrant collected no measurable water, but in the northwest
six gauges had amounts ranging from 0.2 to 1.8 inches after
only 25 minutes of flow.
Conclusions
All qualitative (perceptionof carryover from walking under
steam plume) and quantitative (rain gauges, exhaust particle,
brine flow) measurementsconfirmed that the LEAMS was more
effective than the cyclone at removing carryover. There is some
uncertainty in the test data, primarily in interpretation of the
particle flow measurements. Calculation of the particle size
and, thus, mass flow rate is clearly dependent on knowledge of
the particle velocity. In the test data, the particle velocity was
taken to be the same as the vapor velocity measured by the anemometer. Because there is no correction for slip between the
droplets and the vapor, the particle size could be understated,
and this effect would be greater with larger particles than with
smaller. This means that exhaust liquid flow from the cyclone,
where particles were larger, could be understated relative to the
LEAMS or, in other words, that the LEAMS could be an even
greater improvement in carryover reduction than indicated by
the data.
A summary of lessons learned is the following:
LEAMS Design Features
The LEAMS system is designed to be environmentally
friendly while drilling ahead with air/mist/foam drilling fluid
or while testing a well. It is based on the premise that no existing separator can perform all these functions - large flows into
existing, inadequate separators may result in very significant
amounts of carryover, leading to incidents such as spraying a
hillside or farmer’s field with brine.
The LEAMS, designed for a very high separating efficiency
in a packaged, modular design, will save time and money while
drilling and testing. Some of the LEAMS design features were
exercised during this test, others were not.
The LEAMS is designed to be:
LEAMS worked very well at reducing carryover.
LEAMS seems to have more steam condensation within the
separator than a cyclone, causing the higher liquid output.
Even though this might be simply remedied or mitigated, it
is not clear whether, from the operator’s standpoint, it is
important.
1. Multipurpose - adequate for most drilling blowdown and
flow test situations without rigging down and converting
from the “drilling” cyclone to the “testing” cyclone. If the
LEAMS is to be used in extremely large wells, additional
separator modules can be bolted on.
The well tested has relatively high enthalpy, with a steam
fraction approximately 85%. It is unclear how the LEAMS
efficacy will extrapolateto wells with higher or lower steam
fractions.
2. Stable - the combination of fluid impact and wind makes
some cyclones unstable; it is not unknown for one to blow
off the location. The energy of the well effluent slug is dissipated by the LEAMS design, low center of gravity, and
added weight from the mud pit section.
3. Portable - the LEAMS modules (designed to fit in a standard 7.5’ by 39’ shipping container) were assembled without difficulty at the test site.
4. High capacity -the LEAMS worked over a range of outputs,
with good results on a relatively large well (270 KPH)down
to a throttled flow rate that was less than half that figure.
remove particlesto
5 . Efficient at separation- the
below 50 microns in size. In this test it is likely that condensed
was emitted from the stack for both separators.There
was no physicalObservation Of
Or large drop1etsfrom
the LEAMS. The particle size distribution measurement from
the LEAMS is probably not related to its separation efficiency
and is in fact in some contradictionwith observations.
6. Quieter - its baffles and diverters should reduce noise, but
the LEAMS was X“M3what nosier than the cyclone- The
inward sucking of the segmented gasket mentioned above
might have increased the LEAMS noise.
The particle-probe method of measuring liquid content of a
mixed stream is critically dependent on knowing the velocity distribution and whether particles are uniformly scattered
across the exhaust area of the device being tested.
Steam pressure inside the LEAMS indicates caution when
feeding a larger or more energetic well into this separator.
Acknowledgements
Sandia is a multiprogram laboratory operated by Sandia
Corporation,a Lockheed Martin company,for the United States
Department of E~~~~~under contract DE-AC04-94AL85000.
Personnel from Cos0 Operating Company, Drill Cool Systems,
and Two-Phase Engineering were extremely cooperative and
helpful duringperformance of this test.
References
I . Jung, D. B., Howard, W. T.; June 2000; “LEAMS Low Emissions Atmospheric Metering Separator for Drilling and Well Testing”; Proceedings of World Geothermal Congress 2000; Beppu, Japan.
2. Jonas, 0. and Machemer, L.; February 2000; “Cyclone and LEAMS
Separation Efficiency and Moisture Carryover Test at the Cos0 Geothermal Power Site”; report for Sandia National Laboratories from Jonas
Inc., I I I3 Faun Road, -Wilmington DE 19803.
7. “Power plant friendly” - the LEAMS is designed to minimize or even eliminate any carryover.No carryover was seen.
The steam should be clean.
70