International Conference on Gas Hydrates (ICGH) (6th : 2008)

CRITICAL DESCRIPTORS FOR HYDRATE PROPERTIES OF OILS: COMPOSITIONAL FEATURES Borgund, Anna E.; Høiland, Sylvi; Barth, Tanja; Fotland, Per; Kini, Ramesh A.; Larsen, Roar 2008-07-31

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata


59278-5574.pdf [ 151.97kB ]
JSON: 59278-1.0041096.json
JSON-LD: 59278-1.0041096-ld.json
RDF/XML (Pretty): 59278-1.0041096-rdf.xml
RDF/JSON: 59278-1.0041096-rdf.json
Turtle: 59278-1.0041096-turtle.txt
N-Triples: 59278-1.0041096-rdf-ntriples.txt
Original Record: 59278-1.0041096-source.json
Full Text

Full Text

Proceedings of the 6th International Conference on Gas Hydrates (ICGH 2008), Vancouver, British Columbia, CANADA, July 6-10, 2008CRITICAL DESCRIPTORS FOR HYDRATE PROPERTIES OF OILS:COMPOSITIONAL FEATURESAnna E. Borgund ;1;2, Sylvi H?iland 2, Tanja Barth 1, Per Fotland 3,Ramesh A. Kini 4 and Roar Larsen 21 University of Bergen, Department of Chemistry, Bergen, Norway2 SINTEF Petroleum Research, Norway3 StatoilHydro R&D, Bergen, Norway4 Chevron Energy Technology Company, Houston, TX, USAABSTRACTIn petroleum production systems, hydrate morphology is observed to be influenced by the crudeoil composition. This work is aimed at identifying which crude oil compositional parameters thatneed to be determined in order to evaluate natural anti-agglomerating properties of crude oils, i.e. thecritical compositional descriptors. The compositional features of 22 crude oils have been studied,and multivariate data analysis has been used to investigate the possibility for correlations betweenseveral crude oil properties. The results show that biodegradation together with a relatively largeamount of acids are characteristic for non-plugging crude oils, while excess of basic compounds ischaracteristic for plugging crude oils. The multivariate data analysis shows a division of the non-biodegraded oils, which are all plugging, and the biodegraded oils. In addition, the biodegradedoils seem to be divided into two groups, one with plugging oils and one with mostly non-pluggingoils. The results show that the wettability can be predicted from the variables biodegradation level,density, asphaltene content and TAN.Keywords: crude oil compositional features, multivariate data analysis, hydrate plugging tendencyNOMENCLATUREGPC Gel permeation chromatographyPCA Principal component analysisPLS Partial least squareTAN Total acid numberTBN Total base numberINTRODUCTIONHydrates can cause problems in petroleum produc-tion lines, e.g. plugging of pipelines. Reliable phys-ical models for prediction of hydrate formation areavailable [1]. However, these models do not de-scribe the morphology of the hydrate particles, i.e.whether the particles agglomerate and grow into a Corresponding author: Phone:+47 55 54 39 12 Fax: +4755 54 39 05 Email: anna.borgund@iku.sintef.noplug or remain as a dispersion of small hydrate par-ticles in oils. Some crude oils have been observedto have lower tendencies to form hydrate plugs inpetroleum production lines than others when oper-ated within the thermodynamic conditions for hy-drate formation. Hence, the morphology of the hy-drates is influenced by crude oil composition. Theoils with low tendencies to form hydrate plugs prob-ably contain natural inhibiting components that pre-vent hydrate plugging [2?6]. A possible mechanismfor formation of a dispersion is adsorption of specialcompound types onto the hydrate surface, prevent-ing the small hydrate particles from agglomeratinginto large plugs.In a previous work [5], standard compositional pa-rameters alone, like asphaltene content, and totalacid- and base numbers, were found insufficient indescribing the black oil hydrate behavior. Althoughsome properties are found to be more important thanothers, e.g. the acid profile, the complex chemicalinterplay between interfacial active components andbulk oil most likely excludes the possibility of iden-tifying one single compositional parameter that ex-plains the hydrate behavior for all crude oil systems.Acid fractions extracted from certain non-pluggingcrude oils were previously found to be able to al-ter hydrate behavior from plugging to non-pluggingat operationally relevant PT conditions [7]. Otheracid fractions, however, fail to do so, even thoughthey too have been extracted from non-plugging oils.In order to develop a fundamental understanding ofnatural anti-agglomerating behavior of crude oils interms of inhibiting mechanisms and critical compo-sitional features, more knowledge of the interactionbetween the acid fraction and the bulk oil is crucial.Understanding these aspects is important for assess-ment of hydrate plug risk as well as for searchingfor ways to obtain non-plugging hydrate behaviorthrough modification of the chemical composition ofthe system.In this work, the wettability of freon hydrates incrude oil/brine emulsions is used to evaluate theplugging tendencies of crude oils [5]. Chemicalcharacterization with special focus on the petroleumacids is performed. Multivariate techniques are uti-lized to find interaction effects among the parame-ters. The evaluation procedure for limiting the num-ber of compositional parameters necessary for pre-dicting the black oil hydrate behavior is described,and the critical descriptors are presented.MATERIALS AND METHODSMaterialsThe data set consists of 22 crude oils, spanning fromheavy biodegraded oils enriched in asphaltenes tolight non-biodegraded oils and condensates. Two ofof the oils were supplied by Chevron ETC and therest by StatoilHydro ASA. The oils supplied by Sta-toilHydro have been investigated for several years,and published data have been assembled [5?11]. Inthis data set, the crude oils are marked with a letter,B - biodegraded oils and S - sweet, non-biodegradedoils, followed by a number indicating productionfield and a letter denoting different wells or differ-ent batches within one field.Experimental methodsThe crude oils are characterized with respect to thefollowing properties and compositional features:Plugging tendencies in crude oil/water/gas sys-temsIn order to characterize whether a crude oil will havea tendency to form hydrate plugs or not, both fieldexperience and laboratory tests performed at fieldconditions in a stirred, high pressure sapphire cellare considered [2, 5, 6]. In this paper, oils that formdispersed hydrates are termed non-plugging crudeoils, and oils with high tendencies to form hydrateplugs are termed plugging oils.WettabilityThe wettability of freon hydrates in crude oil/brineemulsions is directly correlated to the plugging ten-dency of the crude oil, see H?iland et al. [5] fordetails of the emulsion method. In the emulsionmethod, phase inversion in oil/brine/hydrate emul-sion systems is used for evaluating the wetting stateof the system [12]. The wetting state, termed as?wettability? in this paper, is a direct measure ofthe expected morphology in terms of whether thehydrates generated in the system tend to agglom-erate or not. The wettability is thus related to theanti-agglomerating behavior. The wettability pa-rameter (Dj ) is given as a number between - 1and 1, where large positive values are correlatedto non-agglomerating systems, and non-pluggingcrude oils. Negative values and values close to zeroare correlated to agglomerating systems, and plug-ging crude oils.Biodegradation level, density and asphaltenecontentThe biodegradation level of the crude oils are deter-mined using the Peters and Moldowan scale [13].The densities of the crude oils are determined at20  C using an Anton Paar DMS60 densitometerconnected to an Anton Paar DMA602HT measuringcell.The amount of asphaltenes is found from reflux-ing a portion of the oil with 40 times excess ofhexane for 6 hours, filtering through a WhatmanGF/F glass fibre filter, and solving the residue indichloromethane:methanol 93:7.Titration: TAN and TBNA Metrohm autotitrator (model 798 MPT Titrino)connected to a Metohm Solvotrode combined LLPH glass electrode (model 6.0229.100) is used forthe titrations.The amount of titratable acids (TAN, total acidnumber) is determined by the standard methodASTM664-89 [14], and is defined as the amount(mg) of potassium hydroxide (KOH) necessary totitrate 1 g of sample to a well-defined inflectionpoint.The amount of titratable bases (TBN, total basenumber) is determined by the standard methodASTM2896-88 [15], with modifications accordingto Dubey and Doe [16], and is defined as the amount(mg) KOH necessary to titrate 1 g sample to a well-defined inflection point.Extraction of acids and analysis by GPCThe acids have been extracted by the use of ion ex-change material, as described by Mediaas et al. [17].The molecular weights of the acid fractions havebeen determined by gel permeation chromatography(GPC). In this technique molecules are separated ac-cording to their size and shape, and standards withknown molecular weights are used to make a cal-ibration curve for determination of the molecularweight of the samples. More information can befound in Borgund et al. [9].Multivariate data analysisCharacterization of crude oil is performed by manydifferent analytical procedures, and correlations aresometimes found between two compositional pa-rameters. In some cases a combination of severalvariables can correlate to other variables, and multi-variate data analysis can be used to investigate howseveral variables affect each other at the same time.A systematization of a large data set in terms of in-ternal correlations can in some cases reveal that ap-proximately the same information can be obtainedfrom fewer analyses.Principal component analysis (PCA)Principal component analysis is used to extract sys-tematic information from large data sets [18]. Thesystematic variation in the data set can be describedby principal components that describe common in-formation in several variables. The first principalcomponent (PC1) is a linear combination of the orig-inal variables that explains as much as possible ofthe variance in the data set. PC2 is a vector that ex-plains most of the variance that is not explained byPC1, and that is orthogonal to PC1. PC3 can be ex-tracted in the same way, and explains the variancethat is not described by PC1 and PC2. Principalcomponents can be extracted until all the variancein the data set is explained. However, the purposeof PCA is a reduction of variables, and two or threeprincipal components are the optimal to extract for adata set [19].The samples and the variables in a data set are pro-jected in a plane defined by the principal compo-nents [20], see Figure 1.Figure 1: The samples in a data set are projected in aplane defined by the principal components, adaptedfrom Wold [20].A plot with both samples and variables is called a bi-plot. The distance between the points is a measure ofthe similarities between the samples. Two samplesthat are close to each other in the plot are similar toeach other. Variables that are close to each other arepositively correlated, and variables that are situatedin opposite direction (through origo) are negativelycorrelated. Variables that are located in an orthog-onal position to each other, when centered throughorigo, are independent of each other.Partial least squares (PLS)In the PLS technique one set of variables (X) is usedto describe another set of variables (Y), e.g. only thethe information in X that is relevant for Y is used toexplain Y [19]. From the calculations, a model thatpredicts results can be obtained.Details of the statistical techniques can be found inthe cited literature [18?21].RESULTS AND DISCUSSIONReview of previously reported dataAs mentioned above, acids extracted from crude oilshave previously shown to hold inhibiting proper-ties [5?7]. From the collected data set, a moderatelyhigh TAN value seems to be characteristic for non-plugging crude oils. However, not all the oils withhigh TAN values are non-plugging. In addition tothe TAN value, the amount of extractable acids isimportant.The biodegradation level has previously been shownto be important, in the sense that all the non-biodegraded oils are plugging and all the non-plugging oils are biodegraded. However, not allthe biodegraded oils are non-plugging, so other fac-tors must be considered (biodegradation seems tobe a necessary, but not sufficient criterion for non-plugging crude oils) [5,6].Multivariate data analysis was also previously usedon these oils, and the TBN value and the amount ofasphaltene were shown to be characteristic param-eters for the biodegraded, plugging crude oils [22].Most of the biodegraded, plugging oils in this dataset contain significantly more asphaltenes comparedto the non-plugging oils.The amount of bases relative to the acids was pre-viously indicated as important [8]. From inspectingthe collected data we find that most of the pluggingoils have excess base (TBN is higher than TAN), andfor the plugging, biodegraded oils the TBN numberis significantly higher than the TAN value. In fact,for most of the non-plugging oils, the TAN value ishigher than the TBN value (excess acid), see Figure2.Evaluation of the plugging tendency of newoilsIn this present work, two new oils, B5a and B6a,have been characterized in the same way as the otheroils. Both oils are found to be biodegraded, andthus from the current know-how, they might be non-plugging. Quite large amount of titratable acids arefound in the oils, i.e. more than 2 mg KOH/g oil forboth oils.Both oils also contain a significant amount of ex-tractable acids, but B5a contains considerable moreextractable acids compared to B6a. From combin-ing GPC analysis of the acids with TAN results,B5a is found to mostly contain compounds with oneacid group, while B6a contains a high level of com-pounds with more than one acid group.The TBN values of the oils are quite high, i.e. morethan 2 mg KOH/g oil for both oils. A comparisonwith the TAN value shows that B5a has a small ex-cess of acids, while B6a has an excess of bases.These results indicate that B5a is a non-plugging oil,and that B6a is a plugging oil, see Figure 2. How-ever, it should be emphasized that the estimation ofplugging tendency from this criteria alone is far fromcertain.The amount of asphaltenes is clearly lower in B5acompared to B6a. This is another indication that B5ais a non-plugging oil and B6a is a plugging oil, seeFigure 3.Results from multivariate data analysisThe different crude oil properties are correlated withthe plugging tendency by the use of multivariate dataanalysis in order to define critical descriptors for hy-drate plugging.A PCA (principal component analysis) plot of thedata comprising both new and collected data fromliterature, is shown in Figure 4. A similar analysiswas previously presented by H?iland et al. [22], al-though without the new oils. The different oil sam-ples are shown in red (coded: B - biodegraded oilsand S - sweet, non-biodegraded), and the variables inblue. The first principal component (x-axis) explainsmuch of the variance in this data set (56.4 %). Onthis axis the biodegraded oils are separated from thenon-biodegraded oils. All the biodegraded oils canbe found to the right in the plot (inside dark-greensquare), and the non-biodegraded oils are found toFigure 2: TBN subtracted from TAN for the crude oils. The black bars represent non-plugging crude oils andthe grey bars represent plugging crude oils. The new oils, B5a and B6a, are marked with white bars.the left in the plot (inside green square). The secondprincipal component (y-axis) explains 26.4 % of thevariance in the data set. The biodegraded oils areseparated into two groups on this axis. In the lowerright corner of the plot plugging oils are found, andhigher in the plot a group of B2 and B4 oils are foundin addition to the B5a oil. Most of these oils arenon-plugging. However the B4b oil have tendenciesto form hydrate plugs. The wettability seems to bean important parameter for the non-plugging crudeoils, while the amount of asphaltene (Asph.cont) andthe base number (TBN) seem to be important for theplugging, biodegraded crude oils, corresponding toresults previously presented by H?iland et al. [22].The oil B5a is situated amongst the non-pluggingcrude oils, and the B6a oil is situated with the plug-ging, biodegraded crude oils. The variable TAN-TBN (the TBN value is subtracted from the TANvalue) is found in the same direction as the wetta-bility variable, but does not seem to have a directcorrelation with any of the other variables.The results from the multivariate analysis show thata high total base number and large amount of as-phaltenes are indicative of plugging systems. How-ever, low amount of bases and asphaltenes do notdirectly correspond to non-plugging systems. In or-der to separate the non-plugging oils from pluggingbiodegraded oils with low amounts of asphaltenes(B2 and B4 oils, in top circle) more information isneeded.Figure 3: The amount of asphaltenes in the biode-graded crude oils. The black bars are non-pluggingoils and the grey bars are plugging oils. The newoils, B5a and B6a, are marked with white bars.Selected variables are used to make a regressionmodel (PLS analysis), in which the wettabilityis predicted from other variables. The variablesbiodegradation level, density, asphaltene contentand the TAN value, are used for predicting the wet-tability, see Figure 5. Only 13 samples are usedfor this regression model, because the other sampleslack data for the wettability or a reliable value forbiodegradation.The R value for the regression equation (R = 0.808)is not very good, but a clear trend can be seenfrom the plot in Figure 5. The analysis indicatesthat biodegradation level, density, asphaltene con-tent and the TAN value are important variables forthe wettability of a system, and hence the pluggingFigure 4: A PCA plot of all the samples (red) and the variables (blue). The non-biodegraded oils (plugging)are found in the small green square to the left, and the biodegraded oils are found in the large dark green squareto the right. The biodegraded oils are separated into two groups: circle in the lower right corner - pluggingoils and higher circle - mostly non-plugging oils.tendency. It is unexpected that the density is impor-tant for the wettability of the system. The densityvalues for the different oils alone do not reveal anycorrelation with plugging tendency. Thus, the vari-ables must be combined in order to obtain a predic-tion of the plugging tendency of crude oils.For a validity check, the oil B4a was removed fromthe data set, and a new model for prediction wasmade. The wettability of the oil was predicted fromthe new model, and a value of 0.32 was obtainedcompared to 0.35 which is the measured value. Thesame was performed with the oil Sb2, and the pre-dicted value was -0.06 compared to the measuredvalue at -0.12. The model did not give a good pre-diction for the B4b oil. Thus, the model might notbe very adequate for biodegraded oils with relativelyFigure 5: Predicted versus measured for the wetta-bility, based on 13 samples (PLS analysis). The wet-tability is predicted from using biodegradation level,density, asphaltene content and the TAN value. Theregression model equation is omitted due to restric-tion from industry partners.low amounts of asphaltenes.By using the model for prediction of the wettability,the B5a oil is predicted to 0.30, which indicate anoil-wet system and non-plugging tendencies. B6a ispredicted to -0.19, which indicate a water-wet sys-tem and plugging tendencies. The accuracy of theprediction is yet to be verified.To summarize, the results from the multivariateanalysis give a group with plugging, biodegradedoils. The oil B6a can be found in this group. Thisoil also obtain a negative value of wettability fromthe prediction model. Thus, this oil is most likely aplugging oil. The B5a oil is situated in a group in thePCA-plot that mostly contain non-plugging, biode-graded oils. The prediction model also gives a pos-itive value, indicating a non-plugging oil. However,this group in the PCA-plot also contains a pluggingoil, which was predicted wrongly by the predictionmodel.CONCLUSIONSThe plugging tendency of crude oils can be pre-dicted from information of the composition of theoil. Biodegradation and moderately high amount ofacids seem to be necessary for non-plugging crudeoils. An excess of acids relative to the bases in thecrude oil (TAN - TBN), as well as low amount of as-phaltenes for biodegraded oils, also seem to be im-portant for having non-plugging systems.The PCA analysis confirmed that large amountsof basic compounds and asphaltenes are connectedwith plugging, biodegraded oils, and that the non-plugging oils are connected with the wettability pa-rameter.From regression analysis it is shown that in mostcases it is possible to predict the wettability of crudeoils, based on the variables biodegradation level,density, asphaltene content and the TAN value.ACKNOWLEDGMENTSStatoilHydro and Chevron ETC are acknowledgedfor funding and permission to publish data. TheNorwegian Research Council, the Petromaks pro-gram, is thanked for funding.The HYADES project group consists of experiencedresearch personnel from both university, research in-stitution and industry, that actively participates inplanning of activities and discussion of results. Thein-kind contributions from both university and in-dustry partners are of essential value for the qualityof the project, and are thus highly appreciated. Theproject group consists of the following persons:  SINTEF Petroleum Research: Roar Larsen,David Arla, Sylvi H?iland, and Jon HaraldKaspersen.  University of Bergen: Tanja Barth, Alex Hoff-mann, Pawel Kosinski, Anna E. Borgund, GuroAspenes (PhD student), Boris Balakin (PhDstudent), Ziya Kilinc (MSc student), and H? onPedersen (MSc student).  Chevron ETC (Houston): Ramesh Kini, LeeRhyne, and Hari Subramani.  StatoilHydro: Per Fotland and Kjell M. Askvik.REFERENCES[1] Sloan ED. Clathrate hydrates of natural gases.New York: Marcel Dekker, Inc., Second edi-tion, 1998.[2] Fadnes FH. Natural hydrate inhibiting com-ponents in crude oils. Fluid Phase Equilibria1996;117:186-192.[3] Gaillard C, Monfort JP and Peytavy JL Inves-tigation of methane hydrate formaion in a re-circulating flow loop: Modeling of the kinet-ics and tests of efficiency of chemical additiveson hydrate inhibition. Oil and Gas Technology1999;54:365-374.[4] Bergfl?dt L. Influence of Crude Oil Based Sur-face Active Components and Synthetic Surfac-tants on Gas Hydrate Behaviour. PhD thesis,University of Bergen, Norway, 2001.[5] H?iland S, Askvik KM, Fotland P, Alagic E,Barth T and Fadnes F. Wettability of freon hy-drates in crude oil/brine emulsions. Journal ofColloid and Interface Science 2005;287:217-225.[6] Borgund AE. Crude oil components with affin-ity for gas hydrates in petroleum production.PhD thesis, University of Bergen, Norway,2007.[7] H?iland S, Borgund AE, Barth T, Fotland Pand Askvik KM. Wettability of freon hydratesin crude oil/brine emulsions: the effect ofchemical additives. In: Proceedings of the5th International Conference on Gas Hydrates,Trondheim, 2005.[8] Barth T, H?iland S, Fotland P, Askvik KM,Pedersen BS and Borgund AE. Acidic com-pounds in biodegraded petroleum. OrganicGeochemistry 2004;35:1513-1525.[9] Borgund AE, Erstad K and Barth T. Frac-tionation of crude oil acids by HPLC andcharacterization of their properties and effectson gas hydrate surfaces. Energy and Fuels2007;21:2816-2826.[10] Barth T, H?iland S, Fotland P, AskvikKM, Myklebust R and Erstad, K. Relation-ship between the content of asphaltenes andbases in some crude oils. Energy and Fuels2005;19:1624-1630.[11] Genov G, Nodland E, Skaare BB and Barth T.Comparing biodegradation levels and gas hy-drate plugging potential of crude oils using FT-IR spectroscopy and multi-component analy-sis. Poster at the 23rd International Meetingon Organic Geochemistry, Torquay England,2007.[12] Fotland P and Akvik KM. Some aspects of hy-drate formation and wetting. Journal of Col-loid and Interface Science, accepted for publi-cation 2008.[13] Peters KE and Moldowan JM. The Biomarkerguide Prentice-Hall, New Jersey, 1993.p.252-265.[14] ASTM664-89 Standard test method for acidnumber of petroleum products by potentiomet-ric titration. In: Annual Book of ASTM Stan-dards, Section 5, American Society for TestingMaterials, Philadelphia, 1989.[15] ASTM664-89 Standard test method for basenumber of petroleum products by potentiomet-ric perchloric acid titration. In: Annual Bookof ASTM Standards, Section 5, American So-ciety for Testing Materials, Philadelphia 1988.[16] Dubey ST and Doe PH. Base number and wet-ting properties of crude oils. SPE RE (August)1993;195-200.[17] Mediaas H, Grande KV, Hudstad BM, RaschA, Ruesl? HG and Vindstad JE. The Acid-IER Method - a Method for Selective Isolationof Carboxylic Acids from Crude Oils and OtherOrganic Solvents. Society of Petroleum Engi-neers, paper 80404, 2003.[18] Isaksson T and N?s T. Prinsipal komponentanalyse - en metode for ? tolke multivariatedata. In: Nortvedt R, editor. Anvendelse avkjemometri innen forskning og industri. Nor-way: Tidsskriftforlaget Kjemi AS, 1996.p.145-152.[19] Grung B. Det matematiske grunnlaget for la-tente variable metoder. In: Nortvedt R, edi-tor. Anvendelse av kjemometri innen forskningog industri. Norway: Tidsskriftforlaget KjemiAS, 1996.p.121-128.[20] Wold S. Kemometri - historik och filosofi.In: Nortvedt R, editor. Anvendelse av kje-mometri innen forskning og industri. Norway:Tidsskriftforlaget Kjemi AS, 1996.p.33-51.[21] Carlson R. Design and optimization in organicsynthesis, Data handling in science and tech-nology - volume 8. Amsterdam: Elsevier Sci-ence Publishers, 1992.[22] H?iland S, Borgund AE, Aspenes G and Fot-land P. Hydrate agglomeration and depositionstudies. In: Proceedings of the Oil Field Chem-istry Symposium, Geilo, Norway, 2008.


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items