Comparison and application of different empirical correlations for estimating the hydrate safety mar

Available online at www.sciencedirect.com| Joumal ofScienceDirect7I Natural GasChemistryEL SEVIERJournal of Natural Gas Chemistry 20(2011)25-33www.elsevier.com/locate/jngcComparison and application of different empirical correlationsfor estimating the hydrate safety margin of oil-baseddrilling fluids containing ethylene glycolFulong Ning1,2*，Ling Zhangl, Guosheng Jiangl*，Yunzhong Tu'，Xiang Wu'，Yibing Yu'1. Faculty of Engineering, China Universily of Geosciences, Wuhan 430074, Hubei, China;2. Department of Chemistry; Norwegian University of Science and Technology, Trondheim NO-7491, Norway;3. National Laboratory on Scientific Drilling, China University of Geosciences at Beijing, Bejing 100083, ChinaJomd[ Manuscript received October 13, 2010; revised December 19, 2010 ]AbstractAstheoiland! gas idutries contiue to increase thactivity indetp water. a | hydrate hazardsboth at present and in the future. Accurate predictids of the hyprate- free zonChd the suitable addition of salts and/or alcohols in preparingdrilling fluids are particularly important both in prevanting hydnpatcroblsmsanl decreasing the cost of drilling operations. In this paper, wecompared several empirical correlations commonly used o estipa thA ne nhibition effect of aqueous organic and electrolyte solutio$ 1using experiments with ethylene glycol (EG) al a hy ate inhiblc.1hoy that he Najibi et al. correlation (for single and mixetQ[ | 1thermodynamic inhibitors) and the Ostergaard et al. fempirial (reltoind leMt enodynamic inhibitors) are suitable for estimatin V 0|the hydrate safety margin of oil-based drilling fluids (OBDFs) il theptetepee pf thellogynamic hydrate inhibitors. According to the twocorrelations, the OBDF, composed of 1.6 L vaporizin g oil, 2% enpulityih Egqobentonite, 0.5% SP-1, 1% LP-1, 10% water and40% EG, can be safely used at a water depth of up tq 1 900 m. Hdw weryotro nccl urt et tdictions for drilling fluids, the effects of the solidphase, especially bentonite, on hydrate inhibition nled in the application of these two empirical correlations.Key wordsoil-based drilling fluid; gas hydrates; ethylene glycol; inhibition prediction1. Introduction[25,26]. In fact, hydrate formation in offshore drilling opera-tions, which involves both flow assurance and safety, is a sGas hydrate is a crystalline compound formed by waterrious concern in the petroleum industry. If shallow sedimenthat contain natural gas are encountered during deep-watmolecules) under low temperature (normally ≤300 K) anddrilling, this gas will enter the wellbore. Thus, gas hydrateshigh pressure (≥38 bar at 277 K) conditions [1]. Gas hydrates .will be formed easily under low temperature and high pres-became a major concern of the oil and gas industry, the gov-sure conditions. These hydrates will block the pipe, annularernment, and academia when it was found formed in naturalclearance or blowout preventer (BOP), obstruct the movementof the drill string and drilling fluids, and cause serious oper-experienced explosive growth [1]. So far, the research areas ofational and safety problems [27- 29]. Particularly, with thegas hydrates extensively involve flow assurance [3,4], safetyincreasing activity of the oil and gas industries in deep wa-[5], energy sources [6- 8], geological hazards [9,10], carbonter, these hazards will become more serious and challengingcycle and climate change [11,12], carbon dioxide sequestra-at present and in the future. For example, recently the BP oilspill in the Gulf of Mexico, which is the largest offshore spilland transportation [17,18], cool storage application [19,20]in the United States history and among the largest oil spillsand even outer-space hydrates [21,22]. Most of these researchin all history, was not stopped by closing the BOP which wasareas have been referredto in reviews [23,24] and monographsspeculated to be blocked by gas hydrates.“Corresponding author. E-mail: nflzx @ cug edu.cn, jianggs65 @ vip.sina.comThis work is supported by the National Natural Science Foundation (No. 50704028, 50904053), the Praiert 8623 (No 2006A A097 316) and the FundamentalResearch Funds for the Central Universities (No. CUGL1004 10), and also partially supported by the中国煤化Iatory on ScienificDrilling, China University of Geosciences at Beijing (No. NLSD200901).YHCNM HGopyright@201 1, Dalian Institute of Chemical P, ChineAcademy of Sciences. All rights reserved.doi:10.1016/S 10039953(10)60161-226Fulong Ning et al./ Joumal of Natural Gas Chemistry Vol. 20No. 1 2011ation. These works mainly focused on the thermodynamicpipelines uses certain chemical compounds called “hydrate and kinetic characteristics of gas hydrate formation and dis-inhibitors" as additives in pipeline fluids to inhibit the for-sociation in W/O emulsions, and demonstrated that they aremation of gas hydrates. These compounds can be generallydifferent from those in pure water and WBDFs. Few stud-classified as thermodynamic inhibitors and low-dosage hy-ies have been specifically aimed at estimating the inhibitiondrate inhibitors (LDHIs). The LDHIs can be further dividedeffect of thermodynamic hydrate inhibitors in OBDFs andinto kinetic hydrate inhibitors (KHIs) and anti-agglomerantsfurther predicting a suitable amount of salt and/or alcohol to(AAs) [30]. In most cases, thermodynamic inhibitors, suchuse in preparing OBDFs. In our previous work [42], we alsoas salts and alcohols, are added into drilling fluids used inused a large-volume hydrate experimental system to study thedeep- water drilling, but the effects of LDHIs are still undercharacteristics of gas hydrate formation and agglomeration ininvestigation [31-33]. Therefore, accurately predicting theOBDFs and to validate the hydrate shell formation model thathydrate-free zone and adding a suitable amount of salts and/orwas previously proposed by Palermo et al. [43] and Turneralcohols during the preparation of drilling fluids are very im-[44]. In addition, the hydrate inhibition of ethylene glycolportant for effectively preventing hydrate aggregations, and(EG) in OBDFS with low water dosage (≤30 vol%) was in-can also decrease the cost of drilling operations.vestigated under static conditions at 277 K and 20 MPa, whichIn terms of the types of drilling fluids, water-basedwould be similar to those encountered in the deep water of thedrilling fluids (WBDFs) and synthetic-based drilling fluidsSouth China Sea. However, we did not quantitatively evalu-are more often used than oil-based drilling fluids (OBDFs) inate the relationship between the hydrate inhibition effects anddeep-water drilling. For example, water-based polyol drillingvarious EG concentrations in OBDFs. In this work, using the .fluids, with a high content of salt and partially hydrolyzedsame temperature and pressure conditions used in our previ-polyacrylamide (PHPA), have been successfully used in someous work, we compare several correlations proposed for es-deep-water drilling activities [34,35]. However, even the besttimating the hydrate inhibition effect of aqueous organic andWBDFs today are not yet adequate to prevent hydrate forma-electrolyte solutions using EG as a hydrate inhibitor. We at-tion in all ultra-deep-water drilling situations, where the pres-tempt to determine which correlation is the most accurate andsure can be several hundred bars at the sea bed and the temper-suitable for predicting the hydrate inhibition effect and esatures range from 271 to 277 K [30]. OBDFs and synthetic-mating the most economical concentration of thermodynamicbased drilling fluids containing oil, which belong to the cat-inhibitors in OBDFs, and how to further refine them for futureegory of water-in-oil (W/O) emulsions, are also used in deepapplication in deep water.water because of the advantages such as strong shell inhibi-tion, high-temperature resistance and good reservoir protec-2. Empirical correlations and experimental sectiontion. Thus, since OBDFs are used in deep-water applications,potential problems should be investigated, for example, theformation characteristics present in OBDFs, the inhibitors ef-2.1. Literature review and empirical correlationsfect on preventing the formation of hydrates, the estimation ofthe hydrate inhibition of OBDFs containing thermodynamicTo accurately predict the hydrate-free zone, the phaseinhibitors and the prediction of the hydrate-free zone.equilibrium of gas hydrates in drilling fluids should be wellInitially, it was thought that the use of OBDFs insteadunderstood and correctly estimated. To date, three primaryof WBDFs could help control the formation of gas hydrates, .equilibrium prediction methods have been used to estimateas pointed out by Grigg and Lynes [36]. However, this con-the conditions (temperature and pressure) under which h:clusion may be resulted from the lack of reported problemsdrates are formed in certain systems: the empirical corre-in the field when drilling in deep water with OBDFs [33].lation method (using experiments), the statistical thermody-Later experiments proofed that hydrates could be formed innamic theory method and the molecular simulation methodW/O emulsions including OBDFs with different water cuts.[25]. The last one usually includes molecular dynamics (MD)Dalmazzone et al. [37] used differential scanning calorimetry and Monte Carlo (MC) methods. They are only used for(DSC) to measure and model the kinetics of methane hydratesome simple hydrate systems under the present computationformation in OBDFs with a 20% water cut in the pressurecapacity and speed conditions, and their prediction accuracyrange of 10-40 MPa. The results showed a strong influenceis also not high [45]. So, the molecular simulation method isof the driving force (represented by the sub-cooling degree).just a direction but not popular in the practical prediction ofLeba et al. [38] used focused beam reflectance measurement hydrate phase equilibrium. The empirical correlation meth-(FBRM) to monitor chord length distributions during the crys- ods, for instance, the gas gravity method [46] and the K-valuetallization process of W/O emulsions with a 30% water cut.method [47], are often employed to estimate the hydrate for-Sloan and coworkers [39-41] used FBRM and video mi-mation conditions. They are relatively simple and can be per-croscope (PVM) to investigate the formation characteristicsformed with a short calculation time. which is suitable forof gas hydrates in W/O emulsions with different water cuts. preliminary indust中国煤化工they are not ac-They found that the system can rapidly agglomerate with hy-curate or suitable fHCNM H Ging salts and/ordrate formation at high water cuts (>60 vol% water) and thealcohols and otherurd, oucl as hose containingwater droplet size has a major effect on hydrate agglomer-oil or condensate. Later, a statistical thermodynamic approachJournal of Natural Gas Chemistry Vol. 20 No.1 20112:nents. However, this equation is an approximation and is notVan der Waals and Platteeuw [50] extended these views andhighly accurate. Later, Yousif and Young [59] also proposed aformulated a model for hydrates with different types of cavi-correlation to predict the inhibition of gas hydrates in drillingties and guests. Their model is the foundation for all thermo-fluids with the presence of salts and glycerol. Obviously, thedynamic predictive methods of the gas hydrate equilibriuminhibition of drilling fluids is mainly affected by the types andand is considered to be the best modern example of usingconcentrations of salts or/and alcohols used therein. There-statistical thermodynamics to predict macroscopic propertiesfore, it is reasonable for drilling engineers to only consider[25]. The van der Waals and Platteeuw theory model, cou-the effect of salts and/or alcohols and to neglect the effect ofpled with the Parrish and Prausnit algorithm [51], has beenother additives on hydrate formation when evaluating the in-used to predict the hydrate phase equilibrium throughout thehibition of drilling fluids [60]. From this point of view, thelast forty years, along with many methods that were improvedOBDF can be regarded as a“black box", regardless of theupon this model during this period [52]. Many commercialtype and amount of other additives present in the drilling fluid,hydrate prediction programs have been developed accordingand the empirical correlations mainly established for aqueousto this method, such as CSMHyd by the Colorado Schoolelectrolyte or alcohol solutions may be used to evaluate theof Mines, HWHYD by Heriot- Watt University, DBRHydratehydrate inhibition of driling fluids.by DBRobinson Software Inc., Multiflash by Infochem Com-Usually, the hydrate temperature depression of aqueousputer Services and PVTsim by Calsep A/S. Ballard and Sloansolutions in the presence of salts and/or alcohols can be esti-[52,53] removed some limitations and gave a new derivationmated in three methods. The first method is dependent on theof the van der Waals and Ptteeuw model, and then used itconcentration of the salts or alcohols in the aqueous solution.in the next generation of the CSMHyd program, CSMGem.Hammerschmidt [61] firstly used this method to build the rela-Statistical thermodynamic theory methods can be used to pre-tionship between the alcohol concentrations and hydrate pointdict hydrate formation and stability in relatively complex mix-suppression. However, his equation is only suitable for typi-tures, such as pure water and multicomponent mixtures. How-cal natural gases and methanol concentrations lower than a 0.2ever, such models are relatively complicated to use and aremole fraction, and is not appropriate for high concentrationsmainly applied to predict hydrate blocks in oil and gas pipelineof inhibitor [25]. Nielsen and Bucklin [62] developed a cor-transportation. They are not convenient for use in more com- relation that is applicable at inhibitor mole fractions as highplex systems such as drilling fluids or porous media. In theseas 0.8 (88 wt%) on the basis of the Hammerschmidt equation.systems, the simple empirical correlation method of using ex-However, this correlation has not gained wide acceptance andperimentation has still been adopted.is independent of the type of the used organic inhibitor, justCompared with pure water and electrolyte or alcohol so-like the Hammerschmidt correlation. Later, Carroll [63] addedlutions, driling fluids are very complex in composition. Theyan activity coefficient term into the Nielsen and Bucklin cor-contain not only an aqueous phase, but also a solid phase thatrelation to account for the type of employed inhibitor. All ofconsists of different types of sand and clay, and many addi-these works assumed that the inhibition effect of any salt- ortives such as salts, alcohols and other polymers. Thus, the for-alcohol- based inhibitor is independent of the system pressure.mation of gas hydrates in this system is dependent not only onThen, Ostergaard et al. [64] considered the effect of systemthe gas composition, pressure and temperature of the drillingpressure on the hydrate inhibition effect and provided a gen-conditions, but also the types of drilling fluid and used ad-eral correlation for predicting the depression of the hydrateditives, the amount of solid phase present, etc. Some com-dissociation temperature in the presence of a single thermo-ponents are hydrate promoters. For example, bentonite is adynamic inhibitor that can be used for drilling fluid systemsnormal component of drilling fluids. It acts as a“thermody-and reservoir fluids.namic promoter”and can promote hydrate formation. ThisThe second method to estimate the hydrate temperaturethermodynamic P/T effect has been observed in experimentsdepression is to measure the freezing point depression of the[54- 56] and molecular dynamics simulations [57]. Someelectrolyte or alcohol solution and then calculate the hydratepolymers widely used in drilling fluids have also shown topoint depression of the solution. Dickens and Quinby-Hunthave a promotional effect on hydrate formation, such as poly-[65] combined the Pieroen equation with a similar equationoxyethylene (20) cetyl ether (Brij-58) and polyoxyethylenefor the formation of ice and obtained a simple equation for cal-(5) nonylphenyl ether (Igepal-520) [58]. Other additives, es-culating the hydrate formation temperature in electrolyte SO-pecially salts and alcohols, are hydrate inhibitors. In contrastlutions. Tohidi and coworkers [28,60] used the experimentalto hydrate promoters, inhibitors cause the equilibrium lines todata of freezing point depression of electrolyte and alcoholsshift to the left in hydrate phase diagrams. Thus, the inhibitionsolutions to develop a novel correlation that combines the cor-ability of these inhibitors can be evaluated according to therelations of Yousif and Young [59], Nielsen and Bucklin [62]shift degree, which is also referred to as the hydrate tempera-and Carroll [63] to estimate the hydrate inhibition effect ofture depression below the uninhibited conditions. Ouar et al.single and mixed thermodvnamic inhibitors. Recently, Na-[55] first established an empirical linear regression formulato jbi et al. [66,67]中国煤化工entioned worksestimate the degree of temperature drop in the hydrate forma- and proposed a siIY片. CNMH Ga rapid estima-tion of seventeen drilling fluid systems, relative to that of wa-tion of the hydratc1 vauilus ivoir fluids bothter alone, and concluded the effect of the drilling fluid compo-in the presence of salts and/or organic inhibitors, regardless28Fulong Ning et al./ Joumal of Natural Gas Chemistry Vol. 20No. 1 2011of the system pressure or hydrate structure. The reliabilityDickens and Quinby-Hunt equation [65]and robustness of this correlation were demonstrated in elec-6008ntrolyte，alcohol and Pessure solutions through experiment.[Tw~ TsAH[ 273.15~ T's 」This method is relatively simpllotmpared to the first methoddescribed above because it does not require knowledge of thehere, Ts=277.15K; for methane, AH = 54190J/mol, andsolute concentrations of the electrolyte or alcohol solutions.The last method to estimate the hydrate temperature de-Najibi et al. equation [66]pression is to use specific gravity data of electrolyte or alcoholsolutions based on the artificial neural network method [68].T,= Tw - 0.68250TpEssentially, the last two methods are just different versionswhere,△Tp=273.15-Ts.of the former because the freezing point and specific gravityof electrolyte or alcohol solutions are directly related to thesalt or alcohol concentration in the solution. In this study, five2.2. Experimental apparatus and procedureof the above-mentioned empirical correlations that correspondto the first two methods are employed regarding their uses inIn this section, a series of experiments were performed toOBDFs. The correlations are as follows:determine the EG concentrations for hydrate free in the ODBFYousif and Young correlation [59]under certain temperature and pressure conditions, and thenwe compared them with the results from the calculations ofOT =(112.3x + 2011.6x2- 6505.0x3)/1.8 (1)the above empirical correlations. The experimental apparatusis the same as that used in our previous work (shown in Fig-Mohammadi and Tohidi correlation [60]ure 1) [42]. According to the previous study [42], lower watercut in OBDF is better for use in the South China Sea. SoOT=-a[n(1-x) + b.x2 +(c1W +c2W2 +c3W3)] (2)a formula consisting of 1 .6 L vaporizing oil, 2% emulsifyingIn this study, no salt is present in the OBDFs. There-agent, 1% organobentonite, 0.5% SP-1, 1% LP-1 and 10%fore, c1, c2 and c3 are equal to zero. For EG, a= 101.47 andwater was selected as the basic OBDF for use in validatingthe above-mentioned empirical correlations. All percentagesb=- 1.9019.are expressed in mass ratios. The added amount of EG to theOstergaard et al. correlation [64]:basic OBDF was calculated through the above-mentioned cor-relations. Methane with a purity of 99.99%, provided by theOT=(c1W +c2W2 +c3W3)x (c4lnP+ cs)xWuhan Hongsheng Industry Gas Corporation, was chosen as[c6(Po- 1000)+1]the experimental gas. LP-1 was used as an agent for increas-here，P =20000kPa; for EG, c1=38.93， c2=-0.522,ing the viscosity of the OBDF. SP-1 was used as a wettingC3= 1.767x10-2, c4=3.503x10-4, Cs=5.083x10-, andagent. The water used in the experiments was deionized dis-c6= 2.65x 10tilled water.Pressure gaugeShutoff valveBooster pump unit-两八日Vacuum pumpAir bathCompressor[Autoclave]←Programmable coolerT2iPData acquisitionHigh-pressureGas ecylinder Recovery cylinderswitching deskFigure 1. Sketch of the experimental systemThe experimental temperature and pressure were set atsumes gas and releases heat, the pressure inside the autoclave277 K and 20 MPa, which were the used conditions in ourwill decrease whenfease during theprevious work and similar to the South China Sea conditionsexperiments. The中国煤化工valuate hydratewhere OBDFs will be used. The constant-volume method wasgeneration in the CYHCNMHGetsofEG.First,adopted for the experiments. Since hydrate formation con- all parts of the apparatuses were assembled and connected asGJournal of Natural Gas Chemistry Vol. 20 No.1 201129s sgyvn in Figure 1. Subsequenly, the valves, gas lines and ing Tw values were calculated using CSMHyd and HWHYD4 auaclave were checked for any leaks. Then distilled water at this pressure value. Therefore, the hydrate temperature de-was used to clean the autoclave and high-pressure lines by apression NT is:pressure-control system. The vacuum pump was started toOT=Tw- Ts= Tw-277.15intake the reaction liquids into the autoclave, and methanewas also pumped slowly into the autoclave by the boosterThen the corresponding EG concentration was calcu-pump at 2 MPa per minute until the pressure reached 20 MPa.lated by applying OT to the empirical correlations (1)-(3).When the inner pressure of the autoclave was stable, the nee-For Equation 3, Po was also calculated by CSMHyd anddle valves of the gas source and the high-pressure switchingHWHYD. For empirical correlations (5) and (6), the freezingdesk were closed to maintain the pressure. The autoclave was .point temperatures Tfs of the EG solution were acquired first._operated to sufficiently mix the gas and OBDF for about 12 h.Then the corresponding concentrations (wt%) of EG solution2 6Grwards, the programmable temperature apparatus was setwere obtained by checking the freezing point data shown into 277K, and the autoclave was cooled at a rate of aboutFigure 2. All of the calculated hydrate temperature depres-3 K/h. At the same time, the temperatures of the liquids (Ti),sions and corresponding EG concentrations (wt%) are listed infree gas (T2) in the autoclave and the outside environment (T3)Table 1. The predicted EG concentration in aqueous solutionand the autoclave pressure (P) were recorded through the data-depends on the prediction of the hydrate formation tempera-acquisition system. Because the set pressure of 20 MPa isture Tw in the pure water system under the same pressure con-higher than the operation pressure of 18 MPa when transpar-ditions. The predicted T w produced by the CSMHyd programent windows are used, the autoclave was opened to observeis higher than that of the HWHYD program. Therefore, thewhether hydrates were formed.At the scheduled time forcorresponding predicted EG concentrations are higher thanthe experiments, the temperature of the cooler was decreasedthose produced by the HWHYD program. In addition, the7 m电the autoclave was depressurized quickly to normal atmo-EG concentration predicted by the CSMHyd program and theZ spReric pressure by opening the relief valve. The process ofNajibi et al. correlation is 40%, the highest among all of theopening the autoclave to observe the hydrate was performedprediction values, while those predicted by the Yousif et al.as quickly as possible, considering gas hydrate dissociationand Mohammadi et al. correlations are relatively lower.under normal atmospheric pressure. In order to avoid the un-certainty of hydrate formation process by this experimental273system, several experiments were firstly performed on purewater and the OBDF with different water cuts in the absenceof EG under the same conditions of temperature, pressure and263duration of experiment before the above-mentioned tests. Hy-grae formation was observed in these fluids [42]. This proves2524 Brinciple that hydrates can also be observed in the OBDFwith EG if the same experimental conditions satisfy the re-243quirement of hydrate formation.2333. Results and discussionNilhetaLataiaThermodynamic model prediction by Tohidi group3.1. Comparison of empirical correlations2210203045C60EG concentration (W%)First, hydrate formation temperature Tw in the presence ofFigure 2. Freezing point data for aqueous solutions of EG [6]yu water should be predicted using an appropriate predictive2 Aohod. Here, two well-known programs, CSMHyd devel-Next, experimental examinations were performed atoped by the Sloan group [52] and HWHYD (HWHydrateGUI)277 K and 20 MPa for 20 h. First, the basic OBDF in thedeveloped by the Tohidi group [28], were used to generate theabsence of EG was tested; the temperature and pressure hadrequired data for hydrate phase boundaries in a pure water sys-obviously changed during the experimental process (The pres-tem. In this study, the desired result is no hydrate formationsure change as an instance was shown in Figure 3), and manyin the OBDF with EG under 277 K and 20 MPa. Thus, themethane hydrates were observed in the basic OBDF when thesystem equilibrium pressure (P) is 20 MPa. The correspond-cover of autoclave was opened. According to Table 1, the EGTable 1. Hydrate temperature depression and corresponding EG concentration calculated by different empirical correlationsEG concentration (wt%)7 7 Bthods Tw(K) 0T(K)Inhibitor concentration methodYousif and YoungMohammadi and Tohidi Ostergardetal中国煤化工NajibictalCSMHyd 292.64 15.4929.838.6YHCNMH G,40.0HWHPP292.3415.19. Q29.431.33038.24F50l 39.7Sysem emperaur: 277.15 k: Sytem pessre 20PR0EG concent rat i on( wt %)30Fulong Ning et al./ Joumal of Natural Gas Chemistry Vol. 20No. 1 201140% EG.30% to 40%. Thus, a mean value of 35% was selected as thecorrelation developed for reservoir fluids and Ostergaard et al.initial EG concentration in the experimental verifications. Un-correlation developed for reservoir and drlling fluids are rela-der the same conditions, no obvious changes in temperature ortively accurate in estimating the inhibition effect of OBDFs inpressure were observed in the OBDF containing 35% EG dur-the presence of thermodynamic inhibitors, and that the formering the experimental process (Figure 3). When the cover of thecorrelation could predict the thermodynamic inhibitor addi-autoclave was opened, no hydrate formation was visible on thetion that is the closest to the experimental results.surface (Figure 4). However, a few hydrate grains were dis-covered in the inner part of the OBDF (Figure 5). These grainshad various sizes and were spread throughout the OBDF. Thisobservation indicates that the EG concentrations predicted bythe Yousif and Young, and Mohammadi and Tohidi correla-tions did not accurately reflect the hydrate inhibition effect inthe OBDF. Later, the concentration of EG was increased to40%. In this case, the hydrate grains were smaller and better .dispersed than those in the OBDF with 35% EG (Figure 6), in-dicating that higher dosages of EG may have a more effectivehydrate inhibition performance and that the thermodynamicinhibitor (EG) acts as an anti-agglomerant in OBDFs to someextent [42]. At an EG concentration of 40%, a few hydrategrains were still observed in the inner part of the OBDF, butthey were very few，and also small in size. This observa-Figure 5. Hydrate grains in the inner part of the OBDFtion implies that, at this temperature and pressure, there is noneed to consider the hydrate blocking problem because severehydrate aggregation does not take place in the OBDF with20.5Abenceof EGPresence of 35% EG、20.019.519.0 IFigure 6. Smaller and more dispersed hydrate grains in the inner part of the6012080240OBDFt/ minAlthough there are other components in OBDFs, such asFigure 3. The pressure as a function of time in the absence and presence ofEG at temperature of 277 KSP-1 and LP-1 in this case, the cause of hydrate formation at40% EG concentration is mainly attributed to the presence oforganobentonite (a type of clay). It affects the hydrate inhi-bition effect of EG in OBDFs and thus contributes to the for-mation of hydrate grains [54- 57]. Our additional experimentsalso showed that the clay acts as a“thermodynamic promoter”and can promote hydrate formation. The clay particles facili-tate the nucleation of hydrate formation and increase the inter-face between water and gas. In addition, the formed hydrateparticles were easily adsorbed to the surface of clay particlesdispersed in the seawater, which promoted the hydrate growthand aggregation. With the increment of clay, these“promotioneffect" become stronger (Figure 7). Therefore, if more reli-able predictions中国煤化工Is used in deepwater, the effecthCHCN M H Gntonite, on hy-drate inhibition sucu, auiu ul Ostergaard et al.Figure 4. No hydrate was observed on the surface of the OBDFand Najibi et al.csrgyations should be modified. Referring tol;Io壹声曹:Journal of Natural Gas Chemistry Vol. 20 No.1 20113Ithe Ouar [55], Carroll [63], Ostergaard [64], Mohammadi [60]here, i represents one type of salt, because in most practicaland Najibi et al. [66] correlations, two general correlations cases, the drilling fluid contains several salts with or with-for predicting the hydrate temperature depression in drillingout an organic inhibitor; d is a constant, and Ws is the con-fluids containing thermodynamic inhibitors may be expressedcentration of the solid phase in mass percentage. These twoas follows:modified correlations will be examined in future work.Inhibitor concentration method3.2. Evaluation for the South China SeaAT = -a[n(1-x) + bx2]+ 2(c1,W; + c2,;W? +c;W})x(c4InP+cs)x [c6(P0- 1000) + 1]- dWs(7)According to the above calculations and experimental re-Freezing point depression methodsults, the Najbi et al. and Ostergaard et al. correlations arerelatively suitable for the determination of a gas hydrate safetyTs= Tw- 0.68250Tr- dWs(8)margin in OBDFs. Thus, the corresponding hydrate-free zoneand safe application water depth of OBDF with a particularEG concentration can be estimated by these two correlations,14which have important practical implications for applying this13OBDF system to deep-water drilling in the South China Sea.Here, a well with a riser, with a water depth of 1500 m anda well depth of 3000 m, in the northern South China Sea isused as an example. All relationships between the tempera-2% cI写1I2%Clayture and water depth in the drilling operations are plotted in% Clay .Figure 8. The hydrate equilibrium boundaries of 40 wt% EGE1solution were plotted jointly using the CSMHyd program, theNajibi et al. correlation and the Ostergaard et al. correlation.These correlations were found to be very similar to each other.The relationship between the water temperature and depth in101214161820the South China Sea is plotted according to the fiting equa-/htion proposed by Zeng and Zhou [69]. The average geother-Figure 7. The variations of pressure with time in the seawater with and with-mal gradient in the northern South China Sea is 3.91 K perout clay100 m [70]. Therefore, the sediment temperature below a-- Seafloor+ Najibi correlationDstergard correlation500w ater temperature一- - .1500 m scdimcnt tempcratureHydrate phase boundary2000 m sediment temperature-心- Drill pipe temperature1000。 1500x: 275.19002000 t Triangular hydrate zonc250026278028529029中国煤化工315Tempcrature (K)JYHCNMHG04 Figure 8. Gre dirbitirpo teperaur @ressre din2 diling opr4pns1b120|h32Fulong Ning et al./ Joumal of Natural Gas Chemistry Vol. 20No. 1 20111500 m water depth can be obtained. Duringnamic inhibitors) and the Najibi et al. correlation (for singlethe temperature distributions of the OBDF in the annular andand mixed thermodynamic inhibitors) which are used for es-inner drill pipe are dependent on the inlet temperature of thetimating the hydrate inhibition effect of aqueous organic anddrilling fluids, the water depth, the sediment temperature andelectrolyte solutions can also be employed to predict the hy-the mud pump displacement. Here, the distributions are plot-drate inhibition effect of salts and alcohols in an OBDF sys-ted according to the calculations reported by Gao [71].em. Furthermore, these correlations can be used to calculateObviously, the temperatures of the OBDF in the annularan economical addition of thermodynamic inhibitors. How-and inner drill pipe are far from the hydrate phase boundary,ever, the effect of a solid phase on the hydrate inhibition ofwhich means that it is safe to neglect the gas hydrate prob-thermodynamic inhibitors needs to be considered for highlylem in drilling. However, if the time of the drilling break isaccurate predictions in drilling fluids with the two empiri-more than 12 h, the inner well temperature will be the same ascal correlations. Additionally, it should be noted that hydratethat of the ambient sediment and sea water [71]. The hydrate risks are usually the greatest during drill break. That is to say,risk will increase dramatically because the well temperaturethe water depth and sediment temperature are the determin-will approach the hydrate boundary line. For this reason, hy-ing factors for the concentrations of thermodynamic inhibitorsdrate problems normally arise during the drilling break. Thepredicted by the two correlations in practical applications.longer the break time, the higher the hydrate risk. It is also forNomenclaturethis reason that the experiments were performed under statichydrate temperature depression (K)conditions. In Figure 8, the system is still in the hydrate-freemole fraction of inhibitor in aqueous solutionzone, where no hydrate formation takes place within a 1500 m .water depth. Actually, according to the Najibi et al. correla-concentration of inhibitor in liquid water phase (wt%)tion, the OBDF only needs about 38% EG at this water depth.concentration of solid phase (wt%)However, if the water depth exceeds 1900 m, a hydrate for-pressure of the system (kPa)mation zone appears. This zone is limited to a small triangu-hydrate formation pressure in the presence of pure water atlar area bordered by the hydrate phase lines, the water tem-273.15 K (kPa) .perature line and the 2000 m sediment temperature line. Inhydrate formation temperature in inhibitor solution (K)this zone, the minimum and maximum well depths are aboutTisfreezing point temperature of water with an inhibitor (K)1900 m and 2022 m, respectively. Thus, for OBDF with 40%hydrate formation temperature in pure water (K)EG, the safe water depth is 1900 m. However, if the drill breakenthalpy of hydrate dissociation to gas and pure wateris short during the drilling process, OBDF with 40% EG can .(J/mol)still be safely used at a 2000 m water depth because hydratesdo not aggregate in bulk in the OBDF according to the resultsTrfreezing point temperature depression of an aqueousof the above-mentioned experiments. It is clear that the hy-solution (K)drate zone extends with increasing water depth under the samea,b,d constantsEG concentration conditions; when the water depth is greaterthan 2000 m, the EG concentration should be increased. Thehydration numberdrilling cost will greatly increase if such a high concentrationof EG is employed in the OBDF. Using salts to replace or-ganic inhibitors can lower the drilling cost, but the additionReferencesof more salts, such as NaCl, will result in saturated solutionshaving an extremely high density, which causes operational[1] Sloan E D. Am Mineral, 2004, 89: 1155problems in the formation with low fracture gradients [30].[2] Hammerschmidt E G. Ind Eng Chem, 1934, 26: 851Therefore, the best solution is to use salt and alcohol mixtures[3] Dholabhai P D, Englezos P, Kalogerakis N, Bishnoi P R. Can Jin the OBDF. Experiments have also demonstrated that mixedChem Eng, 1991, 69: 800solutions have a better hydrate inhibition effect, and the Najibi[4] Tohidi B, Danesh A. Todd A C. Chem Eng Res Des, 1995, 73:et al. correlation can accurately predict the hydrate inhibitioneffects of mixed thermodynamic inhibitors [67].[5] Sloan E. D. Nature, 2003, 426: 353[6] Moridis G J, Kowalsky M B, Pruess K. SPE Reserv Eval Eng,2007, 10: 4584. Conclusions[7] Fan S s, Zhang Y Z, Tian GL, Liang D Q, LiD L. Energy Fuels, .2006, 20: 324Overall, although there are many components in OBDFs,[8]NingFL,WuNY,JjiangGS,ZhangL,GuanJA,YuYB,TangF L. Energies, 2010, 3, 1861such as SP-1 and LP-1 in this case, the thermodynamic in-[9] Sultan N, Cochonat P, Foucher J P, Mienert J. Mar Geol, 2004,hibitors and solid phase still play a decisive role in the estimat-213: 379ing the hydrate safety margin of offshore dilling fluids. De- [10] Rutqvist J, Mor中国煤化工, Perol Sei Eng,ducting the uncertainty and contingency during hydrate for-2009, 67: 1:DHCNMHGmation process to the experimental results, we conclude that [1] Kvenvolden K Iys, 19,01. 10Ostergaard et al. empirical correlation (for single thermody- [12] Wuebbles D J, Hayhoe K. Earth-Sci Rev, 2002, 57: 177Journal of Natural Gas Chemistry Vol. 20 No.1 2011[13] Linga P, Kumar R, Englezos P. J Hazard Mater, 2007, 149: 625 [42] Ning F L, ZhangL, Tu Y z, Jiang G s, Shi M Y. J Natur Gas[14] Kumar R, Englezos P, Ripmeester J. Proceedings of the 6th In-Chem, 2010, 19: 234ternational Conference on Gas Hydrates, Vancouver, 2008[43] Palermo T, Arla D, Borregales M, Dalmazzone C, Rousseau[15] Happel J, Hnatow M A, Meyer H. Annals of the New YorkL. Proceedings of the 5th International Conference on Gas Hy-Academy of Sciences, 1994, 715: 412drates, Trondheim, 2005[16] Zhang B Y, Wu Q. Energy Fuels, 2010, 24: 2530[44] Turner D. [PhD Dissertation]. Colorado School of Mines, 2005[17] Gudmundsson J s, Borrehaug A. The 2nd International Confer-45] Jatkar K, Lee S Y, Lee J W. AIChE 2009 Annual Meeting,ence on Nature Gas Hydrate, Toulouse, 1996Nashville: TN, 2009[18] SunZG, Wang RZ, Ma RS, GuoKH, FanSs. Energ Convers [46] Katz D L. Trans AIME, 1945, 160: 140Manage, 2003, 44: 2733. [47] Wilcox W I, Carson D B, Katz D L. Ind Eng Chem, 1941, 33:[19] Mori Y H, Mori T AIChE J, 1989, 35: 1227[20]BiYH,GuoTw,ZhuTG,FanSS,LiangDQ,ZhangL.Appl[48] Barrer R M, Stuart W I. In Proceedings of the Royal SocietyEnergy, 2004, 78: 111of London Series A-Mathematical and Physical Sciences. 1957.[21] Miller S L. Proc Natl Acad Sci USA, 1961, 47: 1798243: 172[22] Max M D, Clifford s M. J Geophys Res -Planet, 2000, 105: 4165[49] Chen G J, Guo T M. Fluid Phase Equilib, 1996, 122: 43[23] Englezos P Ind Eng Chem Res, 1993, 32: 1251[50] Van der Waals J H, PlatteeuwJ C. Adv Chem Phys, 1959, 2:[24] Sum A K, Koh C A, Sloan E D. Ind Eng Chem Res, 2009, 48:[51] Parrish W R, Prausnit J M. Ind Eng Chem Proc Des Dev, 1972,7457[25] Sloan E D, Koh C A. Clathrate Hydrates of Natural Gases, 3rd[52] Ballard A L, Sloan E D. Fluid Phase Equilib, 2002, 194: 371[53] Ballard A L, Sloan E D. Fluid Phase Equilib, 2004, 218: 15Ed. Boca Raton: CRC Press, Taylor & Francis Group, 200826] Carroll J. Natural Gas Hydrates: A Guide for Engineers, 2nd[54] ChaS B, Ouar H, Wildeman T R, Sloan E D. J Phys Chem,1988, 92: 6492Ed. Amsterdam: EIsevier, 2009[55] Ouar H, Cha S B, Wildeman T R, Sloan E D. Chem Eng Res[27] Barker J w, Gomez R K. J Perol Technol, 1989, 41: 297Des, 1992, 70: 48[28] Ostergaard K K, Tohidi B, Danesh A, Todd A C. Annals of the[56] Kotkoskie T S, Al-Ubaldi B, Wildeman T R, Sloan E D. SPENew York Academy of Sciences, 2000, 912: 411Drill Eng, 1992, 7: 130[29] Reyna E M, Stewart S R. Proceedings of the SPE/IADC Drilling[57] Park S H, Sposito G. J Phys Chem B, 2003, 107: 2281Conference, Amsterdam, 2001. SPE/IADC 67746[58] Karaslan U, Parlaktuna M. Energy Fuels, 2002, 16: 1413[30] Kelland M A. Energy Fuels, 2006, 20: 825[59] Yousif M H, Young D B. SPE/IADC Drilling Conference, Am-31] Kelland M A, Monig K, Iversen J E, Lckvam K. Energy Fuels,sterdam, 1993. SPE/IADC 25705.2008, 22: 24[32] Del Villano L, Kommedal R, Fjten M w M, Schubert ∪s,[60] Mohammadi A H, Tohidi B. Can J Chem Eng, 2005, 83: 951[61] Hammerschmidt E G. Gas, 1939, 15: 30Hoogenboom R, Kelland M A. Energy Fuels, 2009, 23: 3665[62] Nielsen R B, Bucklin R W. Hydrocarbon Process, 1983, 62: 71[33] Kelland M A, Iversen J E. Energy Fuels, 2010, 24: 3003[63] Carroll J J. Natural Gas Hydrates: A Guide for Engineers, New[34] Caenn R, Chillingar G V. J Petrol Sci Eng, 1996, 14: 221York: Gulf Professional Publishing Co., 2003[35] Hu Y L, WangJ H, Zhang Y, Wu B, Xiang X J. Offshore Oil[64] Ostergaard K K, Masoudi R, Tohidi B, Danesh A, Todd A C. J(Haiyang Shiyou), 2004, 24: 83Petrol Sci Eng,2005, 48: 70[36] Grigg R B, Lynes G L. SPE Drill Eng, 1992, 7: 32[65] Dickens G R, Quinby-Hunt M s. J Geophys Res, 1997, 102: 773[37] Dalmazzone D, Kharrat M, Fouconnier B. Proeedings of the [66 Najbi H, Mohammadi A H, Tohidi B. Ind Eng Chem Res, 2006,4th International Conference on Gas Hydrates, Yokohama: 200245: 4441[38] Leba H, Cameirao A, HerriJ M, Darbouret M. PeytavyJL, Gle- [67] Najibi H, Chapoy A, Haghighi H, Tohidi B. Fluid Phase Equi-nat P. Chem Eng Sci, 2010, 65: 5294lib, 2009, 275: 12739] Greaves D, Boxall J, Mulligan J, Sloan E D, Koh C A. Chem [68] Mohammadi A H, Richon D. Ind Eng Chem Res, 2007, 46: 3852Eng Sci, 2008, 63: 4570[69] Zeng W P, Zhou D. J Trop Oceanogr (Redai Haiyang Xuebao),40] Boxall J, Greaves D, Mulligan J, Koh C, Sloan E D. Proceedings2003, 22: 35of the 6thInternational Conference on Gias Hydrates, Vancou- [70] MiLJ, YanY s, ZhangGC, HuS B, HeLJ, Yang s C. Actaver, 2008Petrolei Sinica (Shiyou Xxuebao), 2009, 30: 2[41] Turner D J, Miller K T, Sloan E D. (a) Chem Eng Sci, 2009, 64: [71] Gao Y H. [PhD Dissertation]. Dongying: China University of3996; (b) Chem Eng Sci, 2009, 64: 5066Petroleum, 2007中国煤化工MHCNM HG