Offshore Oily-Water Separation

Earl S. Snavely Jr., Mobil R&D COIl'.

Henry F. Yarbrough, Mobil R&D Corp.

Summary
Various oily-water separation devices were evaluated for for the application of several types of separators for off-
deoiling drilling-rig deck drainage. A loose-media shore oil waters.
coalescer selected for the application averaged 92 % These projects resulted in some effective modifica-
removal of free oil over a I-year period of operation tions to commercial separators and also resulted in a sim-
when it was fitted for solids removal and media cleaning. ple effective test for the anticipated effectiveness of
Dispersed-air flotation cells were found to remove the dispersed gas flotation devices.
same amount of oil from water that could be removed by
filtration with a filter retaining particles with diameters Experimental Setups
greater than 8 microns. Filtration is a siml'ie and ac- Experimental setups of equipment are described for each
curate method for determining the oil-removal efficiency type of equipment tested under the Results section.
that can be expected from a flotation separator for a par- The oil content of waste waters was measured by ex-
ticular application. traction and infrared spectrophotometry. A 100-mL sam-
Introduction ple was acidified with hydrochloric acid to pH 3.0 or less
and was shaken vigorously with 10 mL of carbon
Many types of equipment are available for the separation tetrachloride. The solvent was allowed to settle, and an
of oil from waste water, and the selection of a process to aliquot sufficient to fill an absorption cell (3 mL) was
attain a prescribed separation efficiency for a particular withdrawn with a syringe. If the extract was cloudy, it
application can be very difficult. 1 It is not possible in was filtered before the cell was filled. The absorption at
many cases to predict the effectiveness of an oily-water 3.3 microns was measured by a portable infrared spec-
separator, even if the prediction is based on case trophotometer. The spectrophotometer was calibrated
histories of similar applications. with dehydrated samples of the actual oil present in the
The applicability of an oily-water separator usually samples whenever possible.
can be judged by the performance of laboratory or field
tests 2 ; however, the tests require the availability of the
actual water to be treated. Laboratory tests that simulate
Results
field conditions 3,4 have demonstrated the parameters in- Drilling-Rig Deck Drainage
volved in oily-water separation and have been useful in The separation of oil from drilling-rig deck drainage is
the design of separators and the selection of chemical an imposing problem because of the variety of oils that
aids for separation. might enter the system, the presence of mud solids and
A thorough investigation of oily-water separation by cuttings fines, and the possible entry of surface-active
gas flotation 5 has shown that separation efficiency is a agents into the system. Fortunately, the average volume
function of oil drop size, oil concentration, gas concen- of drainage to be treated is small (about 1,000 BWPD),
tration, gas bubble size, electrolyte concentration, and and the input of oil is small but variable. The input of oil
type of oil. These results demonstrate that a given can be held to very low volumes by good housekeeping
separation process cannot achieve consistent results for and the use of strategically placed drip pans. Water to be
different waste waters. This paper gives results obtained treated by the deck-drainage system is usually fresh (rain
0149-213618310001-9581 $00.25
and wash water), but it can be salty from sea spray and
Copyright 1983 Society of Petroleum Engineers of AIME from the fire system.

234 JOURNAL OF PETROLEUM TECHNOLOGY

 TABLE 1-TEST RESULTS FOR A MULTIPLE-CELL TABLE 2-TEST RESULTS FOR A FIXED-MEDIA COALESCER
FLOTATION DEVICE
Oil Content
Flow Rate (ppm)
Oil Content
(ppm) (gal/min) In Out Remar~s
Flow Rate - -
(gal/min) In Out Remarks 5.5 135 5
- 4.5 60 3
1.6 140 56 wash engine room, no mud 2.8 84 3 prefilter plugged after 140 gal; prefilter
0.3 160 156 removed
4.0 41 26 diesel spill 8.2 50 2
2.5 86 81 deck washed with cleaning compound 7.8 50 4
2.5 98 103 7.8 57 11 effluent cloudy after 245 gal without
1.0 59 50 pulling drillstring prefilter
7.8 57 18
7.8 57 29

Areas of the drilling rig that might discharge oily water subject to frequent plugging of its prefilter by mud
are the propulsion room, pipe rack, drill floor, and crane solids. The coalescer commonly removed 96% of the in-
decks. Therefore, the oily inputs are expected to be fluent oil when the prefilter was operational; however,
diesel fuel, hydraulic oil, engine lubricant, compressor the prefilter plugged after only 140 gal total throughput
lubricant, soluble oil, oily mud additives, and crude oil. and was removed. With the prefilter removed, the
To avoid complications caused by surface-active agents coalescer continued effective separation for a time but
in cleaning compounds, cleanup with absorbent deteriorated to only 50 % oil removal after 245 gal of
materials must be encouraged. The oily absorbents then operation without the prefilter. We concluded that the
are stored and hauled to shore for disposal. decline in separation efficiency of the coalescer was
Mud solids, cement, and cuttings fines are a severe caused by oil absorbed on solid particles and that the
problem to a deck-drain treating system. These solids coalescer still was removing suspended oil droplets
can plug piping and other equipment associated with oil effectively.
removal, and they also act as oil absorbents.
The factors described made it imperative that some Operational Deck-Drain Treating
testing of representative classes of oily-water separators The drilling-rig deck-drain treating system is shown in
be done to detennine their ability to cope with the prob- Fig. 1 along with other discharges and their treatments.
lems of drilling-rig deck drainage. The types selected on Here we discuss only the treatment of drainage from the
the basis of size effectiveness and ability to cope with drill floor, pipe rack, and mud pump room. This
surges in volume and oil content were dispersed-air drainage was passed first through a coarse screen to re-
flotation cells, fixed-media coalescers, and loose-media tain debris, such as gloves, rope, and wood splinters. It
coalescers. It would be desirable to precede the oily- then entered the bottom of a 23-bbl cone-bottom tank
water separator with a large-volume (1,OOO-bbl) reten- through the inner of two concentric pipes. As we first
tion vessel for solids removal; however, this is not possi- had conceived, the bottom fluids were circulated by a
ble on most offshore drilling rigs because of space and 25-hp pump through a hydraulic cyclone and then to the
ballasting problems. top portion of the cone-bottom tank. This method of
removing solids was abandoned in favor of simple gravi-
Tests of Flotation Cells and ty settling because the hydraulic cyclone was not very ef-
Fixed-Media Coalescers fective in removing solids and because it discharged too
Oily-water separators tested under drilling conditions much untreated water. From the cone-bottom tank, the
were a multiple-cell, dispersed-air flotation device and a water entered a 20-bbl loose-media coalescer that com-
fixed-media coalescer with a prefilter. The flotation prised an outer open partition for preliminary oil skim-
device was a recycle type that pumped part of the treated ming and an inner vessel that contained a quantity of
water back to the cells through venturis to provide the loose polypropylene packing roughly similar to small,
dispersed air. The fixed-media coalescer and its prefilter spoked wheels. The water passed from the outer open
comprised replaceable cartridge elements that consisted part of the vessel downward through the plastic coalesc-
of media supported on a cylindrical wire mesh. The units ing media before being discharged through a weir box.
were susceptible to limited backwashing but were not The weir box allowed the discharge flow to be measured
equipped for it, so backwashing was not attempted. and also provided a source for samples and convenient
The flotation device was not able to cope with drilling- visual inspection of the discharge. An oil sheen was
rig deck drainage (Table 1), even at flow rates of only never observed in the weir box.
5 % of its rated capacity. The best oil-removal efficiency This system was operated for 1 year during drilling but
observed during the tests was 60%, which was not con- not without problems that were solved successfully. The
sidered to be adequate. Other difficulties encountered first alteration was made to the slop oil drum that col-
were excessive foaming when deck cleaners were used lected too much water. It was fitted with a small pump so
and excessive water entering the slop oil when the rig that the collected water could be returned to the inlet of
rolled. the cone-bottom tank. Second, the packed portion of the
The fixed-media coalescer was very effective in coalescer became plugged with mud solids and bacteria
removing oil from the deck drainage (Table 2) but was growth after 2 weeks of operation. It required two men
JANUARY 1983 235
 PIPE RACK HELIPORT
AND QUARTERS BILGE a
SHALE MUD PUMP MACHINE LAUNDRY SUBDECKS PROPULSION
SHAKER
CUTTINGS
SAMPLE
ROOM ROOM
DRIP
PANS
SUMP
FLoooj
GALLEY SANITARY NON OILY AREAS -r--.:..:.RO=O=M=----

OIL REMOVED
FROM SURFACE
OVERBOARD (ABSORBENT)
PUMP SUMP
SCREEN SAMPLE
ABSOtBENT
TANK, 23 BBL TO
OVERBOARD SHORE

CHLORINE WATER
OVERBOARD
COALESCER SAMPLE

SAMPLE OVER- OVERBOARD

BOARD

SEPARATOR
OVERFLOW
PUMP, 25 HP

OVER-
BOARD
Fig. 1-Waste discharges, treatment, and sampling for an offshore drilling rig.

about 4 hours to remove the packing, clean it, and ceeded 40 mg/L in a I-year period; total oil exceeded 50
replace it in the coalescer. This problem was simplified mg/L only twice, and this occurred during upsets.
greatly by stringing all the packing on 100-lbf test The results described indicated that oil contained in the
fishing line. The packing then was cleaned by simply solids was rather tightly bound. To confirm this, several
pulling it out of the vessel as it was hosed down. This experiments were done in which oil-laden solids were
operation, requiring only 30 minutes, was repeated once removed from the original deck-drainage water and were
a week, and plugging was prevented. shaken with fresh water. Only traces of oil could be
Results of deck-drain deoiling with the system detected in the fresh water after vigorous shaking with
previously described are given in Table 3. These results the solids.
are only a portion of the data taken over a I-year period
because for a large part of the time only traces of oil were Simple Test for Anticipated Performance
measured in the coalescer inlet; also, the flow entering A simple test was devised and evaluated for disclosing
the coalescer varied from 0 to 25 gal/min. The oil and the expected performance of oily-water separators. This
grease concentrations shown in Table 3 were measured test requires only the measurement of the oil content of
separately in the water phase and in the suspended filtered and unfiltered water samples. It is well known
solids. This was done by allowing the solids to settle in a that the drop-size distribution of suspended oil droplets
separatory funnel and then removing them and extracting in contaminated water has a dominant effect in determin-
oil from the water and solid phases separately. The solids ing the performance of oily-water separators. Further-
content of the coalescer inlet ranged from 11 to 40,000 more, water in contact with many oils, especially crude
mg/L. The highest solids contents were observed when oil, will dissolve some chemical species-e.g., organic
the drill string was being pulled, especially if it was wet. acids-from the oil. Gravity-type oily-water separators
Table 3 shows that the coalescer was very effective in cannot separate dissolved species; therefore, the effluent
removing free oil (average 92 % removal) but was less from a gravity-type separator will contain suspended oil
effective in removing oil attached to solids (average 64 % droplets that are too small to be separated by that par-
removal). The free-oil content of the discharge never ex- ticular device, and dissolved species extracted by the
236 JOURNAL OF PETROLEUM TECHNOLOGY
 280

240
TABLE 3-RESUL TS FOR OFFSHORE DRILLING-RIG
DECK-DRAIN COALESCER 200

Oil and Grease Oil Removed

(mg/L) (%) 160
OIL
Sample Free Solids Total Free Solids CONTENT
--- -- PPM
120
in 5.4 9.6 15.0
out <5 <5 <10
in 9.0 18.2 27.2 80 --8--------- - x-----
out 7.8 <5 12.8 >13 72 '-- FIL TERED SAMPLE
in 1,604 29.7 1,634
out 37.1 21.9 59 98 26 40
in 84.3 82.2 166.5
out 9.0 12.4 21.4 89 85
in 260.4 111 371.4 INLET 3 4 OUTLET
CELL NUMBER
out <5 <5 <10 98 95
Averages Fig. 2-Performance of 38,000-BWPD separator. The flow
in 135 27.1 189 92 64 rate is 20,000 BWPD. X represents no chemical; 0,
out 10 9.7 19.4 chemical aid; and D, filtered sample.

280
water from the oil it originally contacted. If a filter is
selected that does not absorb dissolved oily species and
240
that passes oil droplets of the same diameter as the
minimum diameter that can be separated by the
200
separator, the oil content of the effluent from such a filter
should be indicative of the performance that can be ex-
160
pected of the separator. This approach was tested on two OIL
CONTENT
multiple-cell, induced-air flotation separators that had
PPM 120
rated capacities of 38,000 and 26,000 BWPD and that
were treating produced water that had contacted Califor-
80
nia crude with an API gravity of 12.7. These flotation
separators were fitted with special Ys-in. stainless steel
gravity-fed sample lines placed so that the inlet and 40

outlet of each cell and the outlet of the separator could be

sampled. The objective of sampling each cell was to INLET 3 4 OUTLET FIL TERED
determine whether the motor-driven rotor that induced --C-EL-L-N-U-M-BE-R-- SAMPLES

and dispersed air in each cell was shearing the oil Fig. 3-Performance of a 26,000-BWPD separator at two flow
droplets, causing them to be more difficult to remove. rates. 0=10,000 BWPD; X=20,000 BWPD.
Results of these tests (Figs. 2 through 4) demonstrated
that filtration precisely indicated the performance of
flotation cells, at least for this particular system. The
minimum drop size capable of separation by induced-air 280
flotation is probably 10 to 15 microns. These tests used a
Whatman No.1 ™ filter, which is rated by the manufac- 240
turer to separate particles down to diameters of 8
microns. 200
Fig. 2 compares the performance of the 38,000-
BWPD flotation cell with and without a chemical floc- 160
OIL
culant added to the inlet stream. Comparison of the inlet CONTENT
PPM
and outlet oil concentrations indicates an oil-removal ef- 120
ficiency of 67 %, whether or not chemical flocculant is
added. However, a filtered sample of the inlet had the
same oil content as the treated water effluent, and the oil
80 ----x-x
content of the effluent could not be lowered by filtration; 40
therefore, the separator was 100% efficient in removing
oil that was susceptible to removal.
INLET
Figs. 3 and 4 demonstrate that the separation efficien- -'-----,--=-2----,--:-:-::=3-=_4.:.. OUTLET FILTERED
CELL NUMBER SAMPLES
cies of the two flotation devices were independent of
flow rate up to 20,000 BWPD, which was the maximum Fig. 4-Performance of 38,000-BWPD separator at two flow
that could be obtained from the source. Results for the rates. 0=10,000 BWPD; X=20,000 BWPD.

JANUARY 1983 237

oil content of effluent from each cell did not indicate any References
shearing of the oil drops, because each cell continued the 1. Boyd, J.L., Shell, G.L., and Dahlstrom, D.A.: "Treatment of
removal of oil as expected. Oily Wastewaters to Meet Regulatory Standards," AIChE Synip.
Series, Water (1971) 393-401.
2. Ellis, M.M. and Fischer, P.W.: "Clarifying Oilfield and Refinery
Conclusions Waste Waters by Gas Flotation," J. Pet. Tech. (April 1973)
1. Oily-water separators must be tested and adapted to 426-30.
a particular application. Published separation efficien- 3. Wamstedt, H.C.: "Studies on Deoiling of Petroleum-Containing
Brines by Sparging with Fine Bubbles," Erdoel, Erdgas Z. (April
cies are not applicable to all separation problems. 1972) 88, 115-29.
2. Loose-media coalescers provide effective oil 4. Collins, G.L. and Jameson, G.L.: "Experiments on the Flotation
separation from drilling-rig deck drainage if they are of Fine Particles," Chern. Eng. Sci. (1976) 31, 985-91.
equipped for backwashing and if provision is made for 5. Strickland, W.T. Jr.: "Laboratory Results of Cleaning Produced
solids removal from the separator inlet. Water by Gas Flotation," Soc. Pet. Eng. 1. (1980) 20, 175-90.
3. Drilling-fluid solids and cuttings fines retain oil that SI Metric Conversion Factors
cannot be removed by coalescers or by washing the
solids with water. bbl x 1.589 873 E-Ol m3
4. The oil content of water after filtration through a gal x 3.785412 E-03 m3
filter with a 8-micron pore size is an accurate indication hp x 7.46* E-Ol kW
of the oil content of the effluent from a multiple-cell, in. x 2.54* E+OO cm
induced-air flotation separator. lbf x 4.448222 E+OO N
L x 1.0* E+OO dm 3
micron x 1.0* E+OO ftm
Acknowledgment mL x 1.0* E+OO cm 3
We thank Mobil R&D Corp. for permission to publish *Conversion factor is exact. JPT
this paper. We also thank R.L. Morris (Mobil R&D Original manuscript received in Society of Petroleum Engineers office Jan. 15, 1981.
Corp.) and Paul L. Caldwell (Mobil Oil Corp.) for their Paper accepted for publication April 5, 1982. Revised manuscript received Oct. 26,
1982. Paper (SPE 9581) first presented at the 1981 SPE Middle East Oil Technical
assistance in obtaining the results presented. Conference held in Manama, Bahrain, March 9-12.

238 JOURNAL OF PETROLEUM TECHNOLOGY