I

SPE 29487 Downhole Emulsion Breaker Injection into the Lift Gas Stream

6iilti
m

$oclety

ofPetrdsun

Engifwor8

David L. Lagerlefi, Scott T. Sindelar*, William G. Mc Lelland*, and David J. Blumer ARCO Alaska, Inc. *SPE Members ~mht 1SS5, of PetroleumEnginwa, SooIetY
lno. This paper wea p+wpwed praaentafion the ProducfbnOperationsSymposiumheld in OklahomaCHY,OK, U.S.A., 24 A@ 1SS5 for at TM paper wss aslaotadfor praaantat&nby an SPE ProgramCofnmittaafollowfng review of inforntatbnoorttainad an abatreotsubmittedby the aUth0@3). in COnWntS the Pakwr, Of = FIMSII~t fMVW bean retimvd by ~ SIX@ ~ ~um ~t EndU@ ~ SUN* m oorraotion the author(s).The matafiaf,ss praaantad does not mcSaaadtymflecf by . at MY-~ of tfISSOObIY P@daum Enginaars,Its offioara,w mambara Papersptaaantad SPE meetingsera subjaotto publbafkmreviewby EdifcdalCommitteesof the Sooiaty of d~um EWiH. ~Mti~k~wm~dti Hti W-.lll~ maynotbacopkd. Tha~ahouldomtainoon@mta dmddnmd of wltere and by whom the fMPWk fwaaanfad.Write Librarian,SPE, P.O. SOX S2SSSS, Rkhardaon, lX 7S0@2SS6, U.S.A. Telex, 1SS245SPEUT.

A program is in operation at Prudhoe Bay, Alaska which injects emulsion breaker downhole with lift gas to enhance production. This paper addresses the reasons emulsions can limit production in a gas lifted well and the benefits that can be realized by breaking the emulsion downhole. A discussion of the simple but effective produced fluids emuiskm SCitXXl@ test and the field implementation of downhole emulsion breaker in wells at Prudhoe Bay are included in this paper. The emulsion breaker inhibits the formation of highly viscous emulsions by acting downhole. Along with increased production of nearly 3000 bopd for 22 wells, gas lift efilciency has been enhanced, heading of these wells has decreased, and less back pressure in the produced fluid gathering lines to the processing facility has been observed. The processing facilities have also experienced a reduction in the amount of emulsion breaker usage and a reduction in the volume of unbreakable emulsion pads going to produced water disposal. Wells that are gas lifted, produce small fines, and generally are above SO%water cut may be susceptitiIe to production of high viscosity emulsions. Downhole emulsion breaker injection can result in lower friction losses in the tubing string and other benefits in the entire production system. Downhole emulsion breaker injection can provide system wide benefits when tight emulsions are being produced.

The Prudhoe Bay Eastern Operating Area (EOA) has 305 water flood/enhanced oil recovery (EOR) producers. Most wells are gas lifted and produce between 100 to 3500 bopd at water cuts ranging from 4 to 97% and total liquid rates ranging from 200 to 16,300 blpd. The wells all penetrate over 2000 feet of frozen water-laden grmmd (permafrost) before kicking off toward the bottom-hole target which can result in wellbore deviations from O to 65 degrees. The wells flow from a wellhead shelter into a flow line, approximately 1400 feet long, to a manifold building. The manifold building houses a test separator, production headers, and lift gas supply headers, among other things. The individual well production is transported to the processing facility (flow station) after being commingled with production from other wells into a common line. Lift gas from the supply header is &livemd from the manifold building and rate controlled into individual supply lines, approximately 1400 feet long, to each well at approximately 1950 psi. Operational problems with emulsion producers have created difficulties for surveillance engineers, field engineers, drill site operators, and flow station plant wide operators. fluctuations in water cut, gas lift design difficulties, tool running problems, and cold flowing temperatures that often result in paraffin plugging and hydrate formation. The test separator problems were prevalent even after Production Operations began continuously injecting emulsion breaker into the test separator. Downstream of the manifold building, flow station separation facilities have had to contend with emulsion
~tl~ 110 f%%hl= hn.ra WVQ avnmr; ranr WApAw.sed . ..-

References and figures at end of paper. 427

Downhole Emulsion Breaker Injection into the Gss Lift Stream

SPE29487

pads. Emulsion breaker must be added to the incoming three-phase production stream at the slug catcher inlet and occasionally strong unbreakable emulsion pads are diverted to produced water disposal. In late 1993, while taking multiple samples from various wells to evaluate a new water cut measuring device, production fluid characteristics of certain wells were found to be adverse to efficient flow behavior. The samples, taken off the test separator oil leg, were a thick and foamy emulsion. It was difficult to fiil a shake-out tube; the samples were well aerated. The viscosity was apparently very high since the fluid sample piled up in the sink. Produced fluid samples with similar characteristics were obtained from the header upstream of the test separator to determine that the emulsion was not created in the test separator. The evidence suggested that emulsion was forming in the wellbore. Investigation of whether downhole emulsion formation could substantially impede the optimal draw down on the wells and if anything could be done to reduce or prevent emulsion from occurring then followed.

loss), is especially prevalent in Prudhoe Bay wells due to heat loss caused by permafrost extending over 2000 feet below ground level. All three conditions play an important role in creating a tight emulsion. Evidence suggests lift gas energy introduced into the wellbore may be partially responsible for emulsion formation. A wind storm in December 1993 damaged a lift gas supply line at Drill Site 16, inadvertently providing performance data on Well 16-31 without lift gas. The well did flow without lift gas, at a much reduced rate, and the tested water cut dropped from around 80% to around 55% (Figure 1). This change in water cut can be attributed to a reduction in emulsion production which allowed increased efficiency in oil/water separation within the test separator. Emulsion samples for two of the suspect emulsion producing weUs, taken from the production header and the test separator oil leg, were sent to the Prudhoe Bay lab for viscosity analysis. The lab report indicates that the Well 16-22 sample had a viscosity of 345 cp and the Well 16-30 sample had a viscosity of 569 cp. These measurements are two orders of magnitude greater than the typical viscosity of the oil or water components making up the emulsion. Hy@mlic modeling was done to simulate the effect of producing a highly viscous fluid. The modeling -. .,i~ ~ ~ ..timnte n= Wo.lu.w.w nf ~~~ draw ~QW~ ~~d d ~1 o v~ueu . -. ..., production benefit from breaking the emulsion downhole. The live oil viscosity was modified to simulate increases caused by emuision formation. Various tubing performance curves were generated to determine the theoretical bottom-hole pressures while varying emulsion (live oil) viscosity and maintaining a constant lift gas rate and liquid rate. The theoretical liquid rate that can be obtained by decreasing the viscosity from that assumed for each tubing nmfimnan% ~-...,..__.v . aurve m ().6 cp was determined by using the tubing performance curve for 0.6 cp to determine the rate at the higher bottom-hole flowing pressure. The hydraulic modeling results indicate that friction is reduced enough to increase the oil rate 100 bopd due to reduced draw down in wells that have a fluid viscosity of over 100 cp which can be reduced to 0.6 cp. (Figure 2). The lab analysis and modeling results indicated emulsion production was a likely cause of reduced production in the wells. Since it appeared possible to obtain a benefit of 100 bopd (370 blpd in a 73% water cut well, 667 blpd in an 85% water cut well) a field test was proposed.

Emulsion formation in a producing well requires one or more of the foiiowing condiaons~?~~: 1) Two immiscible fluids 2) An emulsifying agent (solids, acids, suIfactants) 3) Energy change (at the lift point, at the perforations, at restrictions) The wells identified as emulsion producers all have water cuts exceeding 50%. The wells are tested in a three-phase ~~~~-at~r at l-- A-11:4- k..+ A..- +* UIGLU 11131LG emulsion pads flowing out the oil leg, the test results often indicate an artificially high water cut. This occurs because the capacitance probe, used to measum the water cut in the oil leg, shorts out and records 100% water cut when emulsion pads containing more than 40% water cut pass the instrument. Emulsion samples contain small amounts of formation sand and approximately 6% asphaltene.
UUL3 UUG LU

Factors that can cause an emulsion to stabilize are often present in the wellbore. Any oil well with a high water cut is likely to have a continuous phase viscosity emulsion and large dillerences in specific gravity of the phases--two of the conditions that promote emulsion stabilization. A third condition, aging (time and heat

428

. SPE 29487 David L. Lagerlef, Scott T. Sindelar, William G. Mc Lelkand, David J. Blumer

Most drill sites with lift gas being supplied to the wells have an existing chemical storage, pump, and computerized delivery system centrally located in the ~,allifOk! !Xlih!hlg. This systsm. had origina~~y been installed for injecting corrosion inhibitor into lift gas. Stainless steel lines from the pump/tank/valve switching skid to the lift gas supply line of every well are still in place. The system had been mothballed several years ago after it was discovered that occasional corrosion inhibition tubing displacements were more effective in reducing tubing corrosion over the entire string. Another option for getting the emulsion breaker into the lift gas was a skid mounted injection system which the Corrosion Group occasionally used for testing continuous injection of various chemicals into pipelines. This skid system couki be located at the W-ellhead for the eintikkm hr.amlrarn;ar.iinn A n;lnt tect u~~~~g ; ~fi~ sfid S~Stern was attempted prior to incurring costs to recommission the injection system located in the manifold building.
WA W-U4 SUJWWUU. . . y- . m....

Seltzer bottle. After 30 minutes of observation at 180F, 5 ml of water broke out in the graduated cylinder. A centrifuged sample of the oil near the interface contained about 809i0water. A potential problem with injection of chemicals into lift gas systems in general is flashing of the solvent which results in solid residues that can cause a variety of problems. It was earlier learned that injection of neat scale inhibitor into the dry Prudhoe lift gas reduced the effectiveness of the treatments and resulted in plugging of the otifice gas lift valve. Mixing the scale inhibitor with a large portion of diesel eliminated these problems. A thermodynamic equation of state phase behavior -ntlal ..,0 ,,~~~t W.JAU.L. t~t ern.l~~i~nbreaker in the n ~n.fi~ -A Uluuul was -u Prudhoe lift gas would have similar transport characteristics and would not flash off before reaching the gas lift valve, resuiting in reduced treating efficiency. The model work investigated the effects of v~~~~g ~hemai~al solutions, using a wide range of . either diesel or ethylene glycol, with a flash of the solvent. The diesel composition was input into the simulator and a high molecular weight component added to account for the emulsion breaker molecule. Dilution concentration, lift gas rates and pressures, DEBI injection rates, and bottom-hole conditions were investigated to provide a measure of the sensitivities to important parameters. The model results indicate that complete flashing of the solvent occurred prior to reaching the orifice gas lift valve until 2160 gpd of 90% diluted chemical was injected. The results indicate solvent flashing was very sensitive to solvent composition. An increase born 0.3% to 1% of a C20C26 fraction prevented diesel flashing down to 40 gpd injection rates. Emulsion breaker A contains a petroleum based solvent enough heavy components were present to prevent flashing undermost anticipated operating conditions. In addition, the flashing behavior is highly dependent on the surface and Due to varying lift gas bottom-hole pressures. saturation conditions in DEBI wells, use of a slightly higher molecular weight solvent or the addition of a heavier secondary solvent to improve emulsion breaker transport in the lift gas was recommended. Solids formation was not predicted using either solvent, Also, neither diesel nor ethylene glycol dilution induced separation of the chemical over the range of 50% - 90% dilution. Laboratory testing also confirmed these conclusions. Given the modeling and laboratory results and availability of diesel at Prudhoe,

STING
Well 16-22 was chosen as the first well to have what was christened Downhole Emulsion Breaker Injection (DEBI). The skid system was delivered to a location adjacent to the Well 16-22 wellhead shelter in November 1993. A wooden framework, covered with plastic, was erected around the skid to protect the equipment from the harsh North Slope winter weather. A light plant and a generator were also placed on location. Stainless steel tubing connected the pump outlet to the chemical injection port on the lift gas supply line just upstream of the casing valve on the wellhead. The same emulsion breaker (A) which is being injected into the inlet of the flow station slug catchers and used in the drill site test separators was used for DEBI initially. This emulsion breaker was specially formulated for use in the slug catchers, however, bottle testing indicated it might be effective in breaking emulsions downhole in Well 16-22. A series of emulsion tests were performed using a Seltzer bottle techniques which allows simulation of emulsion formation under high gas shear regime and high pressure drop at a downhole temperature. A blank test without addition of any emulsion breaker, f o re~l~~+~d.~m,nrm ntinn r-if ~ nnn-finw+n ~~.=&9 ~~~c~,~~ XwLlfiaub*uxa u. *~u~~ *.emulsion which had to be manually scraped from the Seltzer bottle. With addition of 100 ppm of emulsion breaker A, the sample (50 ml of ernulsiiled oil and 50 ml produced water) did flow spontaneously out of the

429

Downhole Emulsion Breaker Injection into the Gss L& Stream

SPE 29487

the skid systems 500 gallon tank was filled with a mixture of 109oemulsion breaker A and 90% diesel. Prior to beginning the DEBI testing on Well 16-22, the well tests indicated it was producing approximately 230 bopd, 3300 blpd, 93% water cut at a wellhead temperate of 100 F. The skid system pump rate was set at 160 gpd to achieve a 115 ppm injection of emulsion breaker. This concentration of emulsion breaker was considerably higher than the 8-10 ppm used at the flow station. However, since this initial effort was intended to prove the concept, a high chemical concentration was selected to begin the program; lower concentrations were tested after the concept was proven. DEBI was initiated at Well 16-22 on late November 1993. The well was placed into the test separator in anticipation of monitoring any change in well production petionnance. Although the theoretical time for the emulsion breaker mix to enter the tubing through the otilce gas lift valve was much less than a day, no response was seen for the first few days. This delayed response was expected because of previous experience with injecting scale inhibitor via gas lift. The monitoring of scaie inhibitor returns through the tubing revealed there was a delay of up to fourteen days from the start of injection. At one point, there was speculation that the delay was caused as the casing volume below the gas lift valve filled with inhibitor. Fluid levels on wells taken before DEBI was initiated indicate the casing annuli was liquid packed. Yet, similar time delays were observed. It is now believed the time delay occurs because the casing annulus and the external tubing surface must become coated before the emulsion breaker mixture can pass through the gas lift valve. The exact time from beginning of injection on Well 1622 to well performance change was missed because the test separator had to be utilized for other purposes. The first indication of a change was noted on day 5 when the well was placed into the test separator again. The liquid rate increased almost 500 blpd (15%) and the tested water cut dropped to approximately 89%. Although additional well tests indicated the total liquid rate was higher than past tests, the 15% increase in liquid rate was not duplicated while injecting at the wellhead. This may have been due to difficulties in maintaining a constant injection rate since the pumps on the injection skid were not sized appropriately for the low injection rates. However, the wellhead injection testing results were encouraging enough to justify moving DEBI to the manifold building.

Injection of emulsion breaker into Well 16-22 resumed from the manifold building in early January 1994. The well tests almost immediately provided encouraging data. The well tests indicate a change in flowing wellhead temperature of almost 40 F, a liquid rate increase of almost 1500 (a 46% increase), and a drop in the tested water cut from 93% to 7070. The oil rate, adjusted for a constant water cut of 70%, increased 350 bopd (Figure 3). The positive response of Well 16-22 to DEBI led to an economic analysis of the field testing prior to implementing DEBI at other areas of the Prudhoe Bay The chemical costs, including chemical Field. deliveries and increased water handling costs (mainly additional corrosion inhibitor in the common line due to increased water production), were approximately $450 a day per well (assuming 350 bopd incremental production) or $1.30 per bbl of incremental oil. The chemical cost per barrel was clearly below the breakeven benefit per barrel, and it was expected that the costs could be lowered by reducing the concentration of emulsion breaker and/or by using a more effective emulsion breaker. Non-chemical costs for DEBI, when operating the injection skid inside the manifold
L.-:1 ..1. m.mding, W~i~iIMJ -. ___ :-..A.. A A :tl.ua CmSml., (?;c

llluluueu

111 LUG

CZMLy

iw.

F~pANDFD ApHJcA~oN OF~~ PR~~D~JRE

The manifold building skid at Drill Site 16 was injecting into several wells by February 1994. Not every well responded as favorably as Well 16-22. Wells were taken off DEBI when they did not respond. Initially, all wells with a positive change in oil rate and totai iiquici were ieft on DE131in order to gather dati under various conditions. At one point 22 wells were on DEBI for an approximate 3300 bopd increase (Table 1). Thus, a considerable amount of data was gathered in an effort to study the affects of DEBI on the many variables affecting production performance. The oil and liquid rate increases provide more than accelerated oil production. The increased off-take from wells that are in enhanced oil recovery patterns also results in additional recovery. Prudhoes enhanced oil recovery project is dominated by gravity effects; therefore, increased flood rate improves the vertical sweep efficiency by contacting more of the reservoir before the solvent rises to the top. Evaluation with the COBRAS waterflood and EOR forecasting model resulted in approximately 100 MSTBO additional recovery per well with an average total liquid rate increase of 10%.

430

SPE 29487

David L. Lagerlef, Swtt T. Sindelar, William G. Mc Lelland, David J. Bhmer

It had been theorized that DEBI would not only assist in increasing liquid production from reducing friction in the tubing, but would also improve iift gas performance, reduce pressure loss due to friction in common lines transporting wells producing emulsions, and improve slug catcher performance. The use of DEBI did improve lift gas performance for some of the wells. In general, the casing pressure dropped in all wells on DEBI, in some cases by as much as 200 psi. The slugging and heading was reduced substantially in some wells with DEBI, as in well 17-9 (Figure 4). Lift gas is utilized more efficiently with DEBI since the lift rates were not modified but, due to the incnmses in liquid production, the gas liquid ratio decreased. Outlet pressure was reduced and the surface productivity index increased at a steeper rate over time in some of the common lines transporting DEBI enhanced production to the flow station (Figure 5). The common line friction pressure loss lowered back pnxsure on all the wells producing to those iines, improving production potentiai. One of the flow stations had been using 285 gpd emulsion breaker at the slug catchers prior to DEBI. While DEBI injection improves slug catcher performance in terms of fewer upsets and operational problems, it is still necessary to add some emulsion breaker at the facility to consistently meet basic sand and water specifications. The rate of emulsion breaker injection at the flow station has been reduced to 40-60 gpd with multiple wells on DEBI. In addition to direct production improvement and cost saving benefits, there are indirect benefits. One of the most pronounced has been improvement in three-phase test separator. The wells which have benefited the most from DEBI all experienced a drop in the tested water cut. The least drop in water cut is 2% while the highest is 20%. It is not necessary to add emulsion breaker into the test separator and the drill site operators have fewer test separator control problems requiring their attention and time. These benefits become increasingly important as wells approach their economic water cut limit. Accurate weii test data is essential to determine if a well should be abandoned. Another indirect benefit of DEBI is an increase in flowing tubing temperature in reducing paraffin deposition in the tubing. Flowing wellhead temperature increases of 20 F and higher are common for the wells which have responded to DEBI. There has been no plugging of orifice valves and examination of two valves pulled for well work reasons showed no sign of solids buildup or liquid cutting.

Although there have not been any major problems with DEBI, there are costs. Maintenance on the high pressure pumps in the manifold buiMing skids is higher than anticipated, in part due to age. Funding is approved to replace these with a more reliable model The high daily usage of emulsion pump. breaker/diesel mix has required special storage and mixing facilities at the central warehouse. Frequent delivery and transfer of the DEBI mix at the drill sites is also necessary. Also, optimization efforts required a fairly significant number of engineering hours initially. The oil rate increase of wells with DEBI resulted in a substantial water production increase since most of the wells have water cuts well above 75%. Concerns about the affect of DEBI on pipeline corrosion rates are being investigated. The first concern investigated by the Corrosion Group was whether there has been any negative direct chemical interaction between the corrosion inhibitor being used and the emulsion breaker. To test the effect of chemical interaction, tank and pump skid, similar to the one described earlier, was installed on a well that had not previously received DEBI. The well was chosen because it did not produce stable emulsion and should not respond to DEBI, allowing the variables to be tested separately. The tank was initially filled with corrosion inhibitor, normally used in the common lines, and injection into the wells flow line was initiated at 15 gpd. Corrosion rate was monitored with 5 flush-mounted electrical resistance probes installed at the 6 oclock position and spaced at intervals along the 2300 ft. long flow line. The probe resistance increases as metal is corroded away from the probe element and the slope of the metal loss vs. time plot (sampled hourly) indicates the corrosion rate. The base line corrosion rate was Oroils per year. Emulsion breaker A was then introduced into the gas lift system well as monitoring of the corrosion rate and corrosion inhibitor injection continued. No production or temperature . increase response was noted when the . weii was part oi DEB1, as expected. N-ochange in any probe response was detected during the 18 days of monitoring. Thus, the fluids corrosion inhibition characteristics were not directly impacted by the DEBI introduction. The Corrosion Group also plans to monitor whether pipeline corrosion rates increase due to additional water breakout running along the bottom of the line and whether there is any increase or decrease in wellbore corrosion rates or corrosion treatment effectiveness due

431

.
6

Downhole Emulsion Bresker Injection into the Gas Lift Stream

SPE 29487

to

temperature increases.

Three optimizing programs were initiated as additional wells were put on DEBI. One involved developing a candidate selection method, another was to determine the most effective DEBI chemical composition, and the third was to determine if there was another approach to . arresting erntilslon fmmatlcxl. The original candidate selection process was based upon selecting wells which had a history of well test or This criteria was somewhat gas lift problems. subjective. There was concern it could be costly to determine if DEBI worked on all the candidates since it took at least 4 days of injection before results were known. It was also estimated to be costiy and not always accurate to obtain the fluid sample viscosity for each candidate. Thus, a screening procedure was developed utilizing a Marsh Funnel. The procedure involves obtaining a one liter sample of emulsion by repeatedly filling the Marsh Funnel and draining the free water off the bottom. Once the funnel is full and no additional free water drains out the metered orifice, the time for the emulsion to pass through the funnel and fill a graduated cup is recorded. The flow line temperature and pressure are also noted. The time to flow one liter of emulsion through the Marsh Funnel provides a relative measure of the viscosity. For example, one liter of low viscosity crude or pure water will take less than 30 seconds to pass through the funnel. Some of the emulsion from DEBI candidates took up to 10 minutes or longer for the same volume to pass through the funnel (Table 2). Because oii characteristics can vary across Prudiioe Bay, it was important to determine the water cut of the emulsion to insure the high viscosity was due to an emulsion instead of low API oil. Emulsion samples were placed in a graduated test tube with emulsion breaker and spun in a centrifuge to determine the water cut. Water cuts in the emulsions varied from 5% to 85%. The wells with the highest times and significant emulsion sample water cuts are targeted for DEBI. In an effort to maximize production performance of wells on DEBI and to lower the costs, tests were conducted with different emulsion breaker concentrations, different combinations of emulsion breaker and carrier fluid mixtures, and different emulsion breaking chemicals. The testing program was initiated on Wells 16-22 and 16-20. A separate tank and pump was dedicated to these wells. The first testing resulted in successfidly reducing the volume of

diesel carrier fluid. A pump rate of 60 gpd and a mixture ratio of 25910 emulsion breaker/7570 diesel was subsequently initiated for all wells on DEBI. Additional testing resulted in an emulsion breaker concentration reduction from 125 ppm to 40 ppm in Well 16-22. Based upon this data, the pump rate on all DEBI wells was lowered to 40 gpd. Concurrent with these field tests, various emulsion breaker suppliers were informed of the DEBI project. They were invited to conduct bottle tests and submit suggestions for improving DEBI effectiveness. Seltzer bottle tests were conducted with two additional emulsion breakers on emulsified fluids from the well previously tested with emulsion breaker A. Both B and C performed significantly better than emulsion bnnker A, with B experiencing a 64% water drop and .= ._ -:$it QOZ....t=. U! Wy &ar -- qhmtt?!i -~rnm -* qf) -------c expGIlcllwllga 90 -m w awl 180F. The separated fluids had low viscosity and flowed easily, in dramatic contrast to the untreated produced emulsion and the emulsion treated with emulsion breaker A. To date, both B and C emulsion breakers have been field tested. Both are effective in the 10-20 ppm concentration range. The production improvements in Well 16-22 with one of the tested chemicals is presented in Figure 3. Supporting data from this well and two others is provided in Table 3. One component of the energy change which contributes to the formation of emulsions may be the loss of energy through turbulence created by an inadequate perforation density. This hypotheses was tested by re-perforating wells that have responded on DEBI. The DEBI supply to these Wells 16-22 and 1624 was shut-in and the well performance monitored (Figure 3 and Figure 6). The effective perforation density in Well 16-22 was increased from 5.5 shots per foot to 7 shots per foot, however, a rate increase was not observed. Lockes peroration modelb indicates a productivity increase of less than 10% for this well. Figure 3 shows the total liquid rate of this well drops significantly after being taken off DEBI--before the shots were added and after. Perforation density was aiso increased in Wdi i6-24. The effective perforation density was increased from 1.5 shots per foot to 4.5 shots per foot. A 19% rate improvement was experienced; less than the 35% productivity increase predicted by the perforation model. Unfortunately, this well has experienced several problems with a subsurface safety valve making it difficult to determine the true productivity increase. It is important to note that although the production rate has fallen since the well was taken off

432

kPE 29487

David L. Lagerlef, Scott T. Sindebx, William G. Mc IAIand, David J. Blumer

DEB I, it has not fallen off as significantly as before. * rnl L... :+:. +finam-l<rn !3Qct=sc + The we~,is kk on DhDl, ULLL lL 1S it appears that LIIG Shut Uwlso. effectc .-. curren~y, J w..increasing the shot density had httle affect on the formation of emulsions as both wells continue to have benefits associated with i3EBi.
LUU -., W --6L m hfi ~-Ici~v

4. ~~rrntie whether uipeline corrosion rates increase and whether there is any increase or decrease in wellbore corrosion rates or corrosion treatment effectiveness due to temperature increases.

c OCCLUSIONS 1. Downhole emulsion breaker injection into the lift gas supply can improve off-take rate from wells producing tight emulsions. Downhole emulsion breaker injection reduces produced fluid viscosity and friction loss in the tubing. The emulsion breaker chemical must be formulated for specific needs in order to obtain optimum results. It took more than 4 days for effects of the downhole emulsion breaker injection to become apparent in the performance of the weils.
h&S llo~ Downhole enmlsicm hn=nker. . inkc~~~ . .. . .... . . created any adverse effects to gas lift equipment, tubing, common lines, or production facilities.

NoMENcJ.A~RE
DEBI = Downhole Emulsion Breaker Injection EGA = Prudhoe Bay Eastern Operating Area AmR 9 S NO~4

2.

The techniques and/or conclusions contained within this paper tie those of the authors and not necessarily shared by Prudhoe Bay Unit Working Interest Owners. OWLF.DGMENT The authors wish to thank the ARCO Alaska, Inc. Prudhoe Bay Production Group, Operations Engineering, the Corrosion Group, and Trinidad .-: ...:--L n La UG ant . .... Flores for contrmuung to tIIG J-.,-I ,uloPmulAc nf tht= information presented.

3.

4.

5.

6.

Modeling work was helpful in confirming the ability of Prudhoe lift gas to transport the emulsion breaker. A portion of emulsion production in some wells may be the result of inadequate petioration density. Increasing perforation density may improve :-..; 2 ,-.S4 +m~~ ...a!aw.~-..~s~zm~ po.m~~~ d Lqu,u ,*--Q k., ~l;mkaiinu , emulsion production, but add~tional work must be completed to fully assess this. ZOMMRNDATIONfl

1. Bansbach, P. L.: The How and Why of Emulsions, Oil and Gas Journal, September 7, 1970. 2. Bansbach, P. L. and Bessler, D. U.: Cold Treating of Oilfield Emulsions, Presented at the Southwestern Petroleum Short Course, Lubbock, Texas, April 17-18, 1975. 3. Lissant, K. J.: Demulsification - Industrial Applications, Vol. 13, Sutiactant Science Series, Marcel Decker, Inc., New York, New York (1983), Chapters 4 and 5. 4. Grahman, N. A.: Personal Communication, Champion Technology, Inc., October 23,1991. 5. Wingard, J. S. and Redman, R. S.: A Full-Field Forecasting Tool for the Combined Water/Miscible Gas Flood at Prudhoe Bay, paper SPE 28632, Presented at the SPE 69th Annual Technical Conference and Exhibition, New Orleans, LA, Sept. 25-28, 1994. 6. Locke, S.: An Advanced Method for Predicting the Productivity Ratio of a Perforated Well, JP?f December 1981.

7.

8.

1. Analyze perforation density in wells with tight emulsions. Mechanical fixes are cheaper than chemical means. 2. Select an emulsion breaker that is optimal for the DEBI application. 3. Evaluate all high water cut wells while on DEBI when calculating the economic viability of the we]l~e

433

Downhole Emulsion Bresker Injection into the Gss Lift Stresm

SPE 29487

S1Metric Conversion Factors bbl X 1.589873 ft X3.048* OF (oF.32)/l .8 gal X3.785412 in. x 2.54* psi x 6.894757 E-01 E-01 = rn3 =m = Oc E-03 . m3 E+OO = cm E+OO = kPa

*Conversion factor is exact.

434

EFFECTS

OF GAS LIFT ON EMULSION WELL TEST DATA FOR

FORMATION

16-30

100

96 80

70

Full Time Gas Lfi Reinetaf ~1 % -- Wk,0 ..m.. ... \ XL i-he b n..+ Wil+fyuaar 12/3i193 /

60 50
t

zi2?3/W
I I Temporary Gas Lfl Started 1114194

40

1000

! :::::::::::::::::::::::::::::
1
~ Liquid Rate ~ 1 Figure

:! 30

Water Cut

Effect

of

Reducing Viscosity of Emuisions Weli 16-22

4 ~

Liqutd rate reeultmg from reduction m vmxsty

10000

1000 Reduction

100 In Viecoeity
Claurn .= --2

10

(OP)

435

DEBI Response - Well 16-22

oEBIEmhklrl
BreakerB
000 800 600 400 200 000 800 F 600 8 400 ! 200 = 000 00 00 00 00 0

6000 5600 5200 4600 -4400 g ~ ~ ~ ~ g 4000 3600 3200 2600 2400 2000 1600 1200 600 400 0

Tost Datom ~
I TOTAL .- _____ LIO -+-C4RATF.._ -_ .. ~ ml RATFmr-m+ -.%....----. .. 711u WC . . .. . . .

Flgum 3

Pre-DEB1 w

17-9 Well

Performance

1123194

1124194

112S194

1126194

1127194

1128194

1129194

1863,

Post

DEBI

Well

Performmtce I

n,J

1
m. . . . .. . . . . .

Sle8194

51e9194

SI1OI94

6111194

5112\94

5113194

S114194

Fi~re
426

Effaots of DEBI on Common Line 16C Surface ProductlvltY Index

4000

3500

3000

1000
500,

nnn

Ugum

Reaponeeto

RepaffOrtilng -DEBIWelllS-24

20000 10000 1Sooo

~ @ ~ s g + Sooo 6000 14000

1
I
d-L

K-v:k
U* ~
I

4000 1
3s00

ChanOoeJl] andK.VahmRun K.V81V0

eorBfidga Plug
OEBISta

Repoli

r#J
K-Vdw
II

\
I 1 al I

3200

2s00 2400 F 8 2000 t 1s00 = o

1200 800

12000 10000

-b
If

I
q q

I I
m u

II

\l

VI

2000 0

400

-1

,.-

Well
4-38 9-23 9-46 9-46 12-14 13-12 13-13 13-27 13-29 14-22 14-30 16-8 16-15 16-20 16-21 16-22 16-24 16-26 16-31 17-3 17-9 17-20

Date DEBI Started

Jun-94 Jun-94 Jun-94 Jun-94 May-94 Apr-94 Jun-94 Jun-94 Apr-94 Apr-94 Jun-94 Jun-94 Jun-94 Feb-94 Jun-94 Jan-94 Mar-94 Mar-94 Jun-94 Apr-94 Apr-94 May-94

TABLE 1 Resulte (as of 7/11/94) 011 Liquid Rate Percent Increaee Increase Liquid Rate New Increaee Water Cut (BOPD (BLPD) DEB I
360
2550

Comments

10%
25h 33% 25%

80%

72
357

2000 325 150 700 400 450 300 100 500

1190 450 3100 2730 1500 4000 126 309 400 450 1000

15%
24% zg~o 18% ll%

4% 15% 35?. 15~o 60% 35% 46% 4170 25% 28% 4470 17% 25%

86% 80% 80% 84?0 90% 80% 827. 73% 67?40 89% 76% 78% 92 Y. 88% 75% 93% 70% 71% 80% 78% 90V0

400 65 24 70 80 81 81 33 55 286 99 248 328 375 280 38 90 80 99 100 3340 Btg change in WC from 92% to 75% No response yet

No well test yet

Big change in WC from 82% to 67%

Was very thick tar locking, now loosened up Btg change in WC from 89% to 71%

otal increase

in oil rate to date

Marsh

Funnel
Fill Time

TABLE 2 Testing Data, 011 Wet/

Water Wet

DEBI
011/

Candidates

I I

Test

On
DE81 M

Sample
Point

Foamln9
Tendanc-y

1000 Cc
(Seconds)

Emulslon
WC (%) Comments

Well
09-02 16-08 16-15 16-18 16-20 16-22 16-22 16-22 16-24 16-26 16-26 16-27 16-27 16-27 16-30 16-30

Date 24-Jun

Header Header Header Header Header Header Header Header Header Header Header Header Header Header Header Header Little Little Little

48
60

oil

2 ?40

1 16-30

26-May 26-May 3-Jun 3-Jun 2-Jun 1-JuI 8-JUI 3-Jun 3-Jun l-Ju I 26-May 2-Jun I-Ju I 26-May 21 -Jun 1 -Ju I

Fb b lb Yse Yes Yes Yae Yea Yea Yea Yse Yea Yea f$b Yes Yea

Yea Yea Very .l. Yea Little Yea

300 60 90 720 400 900 60 00000 300 600+ 300+ 240 300+ 550
120

oil

6% so% 12?10 11% 48% 54% 54%

5%

oil oil oil

oil TABLE 3 Emulsion

36 sec for 750cc Visually doesnt look too thick or foamy Steady slow throughput Just oil emulsion, drained water off before test Actually, 30 sec for 350 cc of oil emui (6% WC] 30 sec for first 500 cc of oily water 3 min for 450cc, free water still breaking out 3 min for 200CC, Lost 3cc during shake out test Very foamy, separated off free H20 (lots of it) nl.-b -- -m, ,Pte *Arcm }fi Iah .an-Iml -n-m, .W.-w UIaem ~w, 3 min for 60000, H20 still dropping out while testing Not foamy just thick Slight foam, very thick
Foamy on top Very thick Black Cream 1 min & 30 eeo for 7S0 cc, H20 still breaking out

Performance DOSAGE 115

40 20 10

of Tested

Breakers

in

BOPD CHEMICAL C WELL 16-22 NOTTESTED

375 375 315

- PPM

CHEMICAL A WELL 16-22 350

375 0 NOT TESTED

CHEMICAL B WELL 09-46 NoTTEsrED

400 400 375