Rheological Behavior of Drilling Muds, Characterization Using MRI Visualization

Transcription

1 Oil & Gas Siene and Tehnology Rev. IFP, Vol. 59 (2004), No., pp Copyright 2004, Institut français du pétrole Dossier Solid/Liquid Dispersions in Drilling and Prodution Fluides hargés en forage et prodution pétrolière Rheologial Behavior of Drilling Muds, Charaterization Using MRI Visualization P. Coussot, F. Bertrand and B. Herzhaft 2 Laboratoire des matériaux et des strutures du génie ivil (LMSGC) 2, Allée Kepler, Champs-sur-Marne - Frane 2 Institut français du pétrole, et 4, avenue de Bois-Préau, Rueil-Malmaison Cedex - Frane philippe.oussot@lp.fr - franois.bertrand@univ-mlv.fr - benjamin.herzhaft@ifp.fr Résumé Visualisation par IRM du omportement rhéologique des boues de forage Les boues de forage sont des fluides omplexes utilisés pour le forage des puits pétroliers. Le rôle des fluides de forage est multiple : transporter les déblais roheux jusqu à la surfae, maintenir une pression au droit de la formation roheuse, lubrifier et refroidir l outil. Il y a prinipalement deux familles de boues de forage : les boues à base d huile (émulsions inverses de saumure dans une phase huile ave de nombreux additifs) et les boues à base d eau (solutions aqueuses d argiles et de polymères). Les formulations de boues à base d huile ont beauoup évolué et sont des formulations très omplexes omprenant de nombreux additifs. L huile de base peut être de différente nature et les additifs sont variés : gouttelettes de saumure, tensioatifs, argiles organophiles, visosifiants, divers solides, et. Ces additifs donnent aux boues des propriétés rhéologiques partiulières. Les boues de forage sont souvent dérites omme des fluides rhéofuidifiants et thixotropes à seuil. Du fait de leur omposition omplexe, les boues de forage ont une struture interne qui est suseptible de se modifier selon les onditions d éoulement et/ou de isaillement, pouvant mener à des phénomènes non homogènes au sein du matériau. Il est don intéressant de développer des tehniques d investigation permettant de visualiser la struture du fluide parallèlement à une mesure rhéologique. Dans et artile, nous présentons des résultats de mesures rhéologiques ouplées à de l imagerie par résonane magnétique (IRM). Ave ette tehnique, il est possible de déterminer le profil de vitesse du fluide dans un éoulement visométrique. Des mesures rhéologiques onventionnelles réalisées sur deux formulations de fluides de forage donnent des ourbes d éoulement similaires : au-dessus d un gradient ritique, elles présentent une omportement rhéofluidifiant ave seuil d éoulement apparent, en dessous de e gradient ritique, on observe un omportement visqueux simple sans seuil d éoulement. L utilisation de l IRM montre qu en fait, en dessous de e gradient ritique, un éoulement stable n est pas possible, la déformation se loalise dans une zone dont la dimension peut dépendre de la taille des éléments onstitutifs. Le omportement rhéologique apparent observé lors des mesures rhéologiques onventionnelles est don la signature de ette zone isaillée et ne représente pas le omportement du matériau dans son ensemble. Abstrat Rheologial Behavior of Drilling Muds, Charaterization Using MRI Visualization Drilling muds are very omplex fluids used to drill oil wells; their funtions are various: to arry the rok uttings to the surfae, to maintain a suffiient pressure against the rok formation, to lubriate and ool the bit. There are mainly two families of drilling muds: oil based muds (invert emulsion of brine into an oil phase with various additives) and water based muds (aqueous solutions of lays and polymers). Originally prepared from produed oil, oil based muds formulations have evolved to very omplex ompositions of

2 24 Oil & Gas Siene and Tehnology Rev. IFP, Vol. 59 (2004), No. various additives. The base oil may be of various nature, and additives are very omplex: water droplets, surfatants, organophili lays, visosifyers, various solids and others. These additives give speifi properties to the mud, partiularly regarding rheologial properties. Drilling muds are often desribed as thixotropi shear thinning fluids with a yield stress. Due to their omplex omposition, drilling muds exhibit an internal struture whih is liable to modify aording to the flowing and shear onditions, whih may lead to non homogenous phenomena. It is therefore important to develop investigation tehniques allowing to visualize the internal struture of the fluid in parallel to rheologial measurements. In this study, we present rheologial experiments oupled to magneti resonane imaging (MRI). Using this tehnique, it is possible to determine the veloity profile in a visometri flow. Conventional rheologial experiments performed on two different drilling fluids formulations give similar flow urves: beyond a ritial apparent shear rate there is a simple yielding behavior with an apparent plateau at low shear rates; below this ritial shear rate there is a simple visous behavior without yield stress. MRI experiments show that in fat, below this ritial shear rate, no stable flow an our and the deformation is loalized in a region of the sample the dimensions of whih may depend on the size of the onstitutive elements. The (marosopi) rheologial behavior observed from onventional rheometri experiments is then the signature of this sheared zone and does not represent the behavior of the bulk material. INTRODUCTION Oil well drilling operation is part of the global proess implemented to loalize and extrat from the reservoir the hydroarbons lying underground. Several wells are neessary to exploit an oilfield: exploration wells to onfirm the presene of oil, evaluation wells to estimate the eonomi viability of the projet and development wells whih lead to the prodution of the field. The lassial drilling tehnique is the rotary drilling tehnique, invented at the beginning of the 20 th entury by Luas for a drilling operation in Texas []. This tehnique onsists in applying a weight on a bit whih is put on rotation through drill pipes string. With the ombined effet of the weight and the rotation, the drill bit ut the rok produing a ertain amount of rok uttings. These uttings are then lifted off to the surfae thanks to a fluid whih is irulated downward through the drill pipe and upward through the annular spae between the rok and the pipe. Tubings are added as the drilling goes deeper and when a ertain length has been drilled, a steel asing is sealed into the bore-hole with ement. Drilling operation an ontinue with a lower diameter. Drilling fluids that are used may be lassified as omplex fluids due to their nature (these fluids may be emulsion/ suspensions of various onstituents) and to their speifi rheologial properties. These fluids, alled drilling muds are lassially shear thinning, thixotropi and strongly thermal dependent. Their internal struture may be very omplex and is liable to modify depending on the flowing and shear onditions, whih may lead to non homogenous phenomena. Charaterization of the internal struture of the fluid in parallel to rheologial measurement is therefore very important. In partiular, low shear rate rheology may have a strong effet on some drilling harateristis suh as uttings removal or barite sag. Barite sag is the rapid sedimentation of the heavy solid partiles whih are used as weighting partiles in the formulation. This phenomenon ours mainly in high angle drilling situations and may lead to atastrophi situations. A better omprehension of the low shear rate harateristis of the muds ould help in understanding and ontrolling these solid arrying problems. After a brief introdution on drilling muds, we present here results of a ollaborative study between IFP and LMSGC aiming at visualizing the veloity field of the sheared fluid with magneti resonane imaging in parallel to rheologial measurement. Conventional rheometrial measurements have been performed on two drilling mud formulations (water based and oil based) with a partiular attention to low stress behavior. The orresponding results are then analysed with regards to MRI veloity profiles obtained under similar onditions, whih show that no stable flow an our below a ritial shear in the mud sample. FUNCTION AND COMPOSITION OF DRILLING MUDS. Drilling Muds Funtions Drilling muds fulfill several purposes for the drilling operation. First the irulation of a visous mud allows to evauate rok uttings from the bottom hole to the surfae, where the solids are separated from the fluid. Muds also maintain the uttings in suspension when the irulation is stopped: indeed, in order to add tubings to the drill string, the fluid irulation is regularly stopped. During this rest time, solid uttings may sediment to the bottom hole and lead to plugging of the well. Drilling muds are therefore thixotropi and develop a gel when not irulated.

3 P Coussot et al. / Rheologial Behavior of Drilling Muds, Charaterization Using MRI Visualization 25 The fluid irulated will generate a ertain pressure (addition of the hydrostati and the fritional pressure) whih is maintained at a value superior to the pressure of the rok formation (in overbalaned drilling onditions). This helps to maintain the well walls, to prevent destabilization of the ground and to ontrol the venue of fluids from the formation. Due to this pressure differene, the mud will filtrate in the permeable rok formation and will form a filtration ake at the wall. Finally, the irulation of the mud helps to ool down and lubriate the drill bit..2 Drilling Muds ompositions Drilling muds may be lassified in three families: Water based muds are prinipally aqueous solutions of polymers and lays in brines with different types of solids and additives. Oil based muds are invert emulsions of brine into an oil phase stabilized by surfatants. various additives are added as organophili polymers, organophili surfae modified lays, solids and other additives. Underbalaned fluids are used for a speifi drilling tehnique alled underbalaned drilling where the pressure in the drilling fluid is lower than the pore pressure of the rok formation. These fluids may be gas, aerated muds (lassial muds with nitrogen) or aqueous foams. The hoie of the mud formulation will depend on the nature of the rok formation, the environmental and eonomi onstraints as well as the possibility of supplying on site. Different formulations may be used for a same drilling operation depending on the geologial nature of the different layers drilled. Oil based Muds an give better performanes but are generally more expensive and less eologially friendly. Numerous additives are added to the formulation in order to reah speifi purposes whih are sometimes ontraditory. For example, muds have to be visous in order to be able to lift the uttings to the surfae, but at the same time, visosity must not be too high in order to minimize frition pressure loss. Additives an be for example visosifiers (lays, polymers as PAC (Polyanionique Cellulose), HEC (Hydroxy Ethyl Cellulose), xanthan gum, guar), weighting agents (barytine, arbonate), filtrate reduers (starh, CMC (Carboxy Methyl Cellulose), resins), lays swelling inhibitors (KCl, glyol), plugging agent for mud losses. 2 MATERIALS AND METHODS 2. Drilling Muds Two different drilling muds have been prepared for this study. The first one is a water based mud whih is mainly an aqueous solution of bentonite and polymers. Its omposition is given in the Table. Produt TABLE Conentration (g/l) Bentonite lay 30 Visosifying polymer A 2.5 Visosifying polymer B 3 Dispersing agent 5 NaCl 30 The seond drilling mud is an oil based mud whih omposition is given in Table 2. Produt 2.2 Proedures TABLE 2 Conentration (g/l) Dearomatized oil Filtrate reduer 6 Emulsifier 8 Wetting agent 2.94 Lime 20 Organophili lay 5 Visosifyer 4.04 Brine For the onventional rheometry we used a Bohlin C/VOR rheometer with a one and plate geometry (angle: 4 ; diameter: 4 m) with smooth surfaes. We also arried out tests with parallel disks with smooth and rough surfaes: similar trends as desribed below were observed for both surfae types. The material was set up then presheared at an apparent shear rate of 300 s during 20 s and left at rest for 5 s. Then a fixed level of stress was applied during a time ranging from 40 to 400 s. After that the material was prepared again in the same way and submitted to another stress level. It was heked during the preshear that the boundary onditions and the material did not hange signifiantly between eah reep test by omparing the apparent visosity during preshear. As a MRI-rheometer we used a set-up and proedures desribed in details in [2]. Here we only reall the main aspets. We used a vertial oaxial ylinder geometry (inner ylinder radius: r i = 4 m, outer ylinder of radius r e = 6 m; length: h = m). The surfaes of the outer and inner ylinder in ontat with the fluid were overed with a sand-paper with an equivalent roughness of 200 µm. The rotation veloity of the inner ylinder is ontrolled and an be varied over a wide range but almost omplete information onerning the fluid behavior an be obtained from tests at a rate in the range [-0 rpm]. The temperature of the material in the Couette geometry ould not be ontrolled but the room

4 26 Oil & Gas Siene and Tehnology Rev. IFP, Vol. 59 (2004), No. temperature was kept between 20 and 24 (as for the tests with the laboratory rheometer). MRI was performed with a Bruker (Frane) set-up equipped with a vertial 0.5 T (Tesl magnet (Magnex Sientifi Ltd, United Kingdom) fitted with shielded gradients leaving a free bore of 25.5 m and delivering a gradient of 50 mt/m with a rise time of 500 µs. The signal was olleted within a linear birdage oil of 24 m length and 20 m diameter. We used a sequene derived from that of [3]: a D aquisition was arried out with a single spin-eho sequene where 80 RF (Radio Frequeny) seletivity was moved from the slie seletion axis to another axis in order to selet a retangular volume of interest and enhane signal to noise by aumulation. The parameters used for both materials were TR (minimum repetition time): s and TE (spin-eho duration time):.53 ms. The volume imaged was a (virtual) retangular portion of 20 mm in the axial diretion with a width (in the tangential diretion) of 5 mm and a length of 70 mm (in the radial diretion, starting from the entral axis) but only veloity data from the rheometer gap (filled with sample) are relevant and presented here. This volume was situated at the magnet enter (so as to damp the effets of field heterogeneities) and suffiiently far from the bottom and the free surfae of the rheometer for flow perturbations due to edge effets to remain negligible (this ould be heked by veloity imaging at different depths in the fluid). This approah has been validated from tests with materials rotating as a plug within a single, rotating ylinder, and from tests with a Newtonian fluid [2]. The spatial resolution of our measurements is 0.5 mm. Just after its insertion between the ylinders the material was presheared at a rotation veloity of 80 rpm during min. Then, with or without a preliminary rest, the rotation veloity was hanged (in few seonds) to a given value and NMR signal was olleted. Suessive, elementary, veloity profiles were obtained every 8 s for the oil based mud and 64 s for the water based mud, leading to respetive total times of aquisition for 28 elementary profiles of 8 min and 2 h 7 min. Limitations and possible artefats onerning tests with bentonite suspensions were disussed in [2], whose main onlusions apply for the present materials. A ritial point here onerns the possibility of the development of fluid heterogeneities during flow. Note that these effets an hardly be deteted from usual density imaging. However, if density heterogeneity developed, it should affet the veloity profile in time, espeially just after flow start up. Here no signifiant variation in time of the suessive veloity profiles was apparent, i.e. within our time resolution the differene observed between two elementary profiles an be attributed to measurement noise. 3 RESULTS AND DISCUSSION Results from reep tests are presented in Figure for oil based mud ( and water based mud (. For both materials the apparent shear rate follows different evolutions depending on the applied stress. For a stress below a ritial value it progressively dereases in time: for the oil based mud the shear rate eventually reahes an apparent plateau after several tenths of seonds; for the water based mud it goes on dereasing signifiantly even after several hundreds of seonds. For a stress beyond this ritial value the apparent shear rate inreases first then tends to reah a plateau after few seonds. These Apparent shear rate (/s) 00 0 Applied stress (P: Apparent shear rate (/s) 00 0 Applied stress (P: E Time (s) E Time (s) Figure Creep tests under different stress values with the ( oil based mud, and ( the water based mud: shear rate as a funtion of time.

5 P Coussot et al. / Rheologial Behavior of Drilling Muds, Charaterization Using MRI Visualization 27 trends are similar to those observed for various pasty materials exhibiting an apparent yield stress [4]. It thus seems that the flows of suh materials an hardly reah a stable situation when the stress is smaller than a ritial value, this fat being extremely lear with the water based mud. This differs from the usual view on this subjet: the flow urve of yield stress fluids is often determined from a more or less rapid inrease or derease of the applied stress, from whih one finds that a Hershel-Bulkley model an orretly represent the behavior in a wide range of shear rates. Aording to this modeling the steady-state flow should have been reahed before the time orresponding to the end of our reep tests. For every shear stress applied, let us note the value of the apparent shear rate at that time and plot in a diagram the orresponding shear stress vs shear rate data: we find the apparent flow urves presented in Figure 2. For both materials they are omposed of two main parts: beyond a ritial apparent shear rate they have a simple yielding behavior similar to an ideal yield stress fluid (with an apparent plateau at low shear rates and a quasi newtonian behavior at high shear rate); below this ritial shear rate they seem to have a simple visous behavior without yield stress. Similar (apparent) flow urves have been obtained on oil based muds with different rheometers in IFP as well as in the Laboratoire de rhéologie in Brest [5] with similar values of ritial shear rates. At first sight these results are troublesome. Indeed they agree with the Barnes onept [6] whih denies the existene of yield stress onsidering that the paste may flow with a high visosity at low stresses, whereas these pastes are in pratie learly apable to support a load without yielding for an indefinite time. MRI-rheometry data appear to bring some interesting elements in this debate. The results in terms of steady-state veloity profiles are presented in Figure 3. In logarithmi sale, the veloity profiles for different rotation veloities of the inner ylinder have all the same aspet and they an be well represented by a trunated power-law model, whih expresses as: τ < τ γ = 0 τ τ τ = k γn in whih k and n are two material parameters. Note that the seond equation may also be written: τ τ γ = γ in whih τ and γ are respetively the ritial stress and shear rate below whih the fluid annot flow steadily. For suh a material the veloity profile expresses as: m r r vr ()= R (4) α R where R is the radius assoiated to the interfae between the sheared and unsheared regions and m = (2/n). It may be heked that the shear stress shear rate relation dedued from (4) by using the theoretial shear rate ( γ ( / )/ and shear stress ( τ r 2 = r v r r) ) distributions in a Couette system, orresponds to a onstitutive equation n () (2) (3) 2 γ. C γ. C Shear stress (P.0 MRI data (effetive shear rate) Shear stress (P MRI No homogeneous flow Apparent shear rate (/s) Apparent shear rate (/s) 3 Figure 2 Apparent flow urves as determined by using the stress vs shear rate data at the end of the reep tests presented in Figure : ( oil based mud; ( water based mud. The MRI data, represented by the model fitted to data, are shown in the form of a ontinuous, dark line.

6 28 Oil & Gas Siene and Tehnology Rev. IFP, Vol. 59 (2004), No. Veloity (m/s) 0 - Rotation veloity (rpm): 20 5 Veloity (m/s) Trunated power-law (n = 0.2) Distane (m) rpm 20 rpm after 20 min. of rest trunated power-law (n = /7) Distane (m) Figure 3 Veloity profiles as determined by MRI in a Couette system under different imposed rotation veloities for ( the oil based mud and ( the water based mud. 0.2 Rotation veloity (rpm), after h 20 min of rest rpm 20 rpm after 20 min of rest Veloity (m/s) No apparent shear Veloity (m/s) No apparent shear Distane (m) Distane (m) Figure 4 Veloity profiles of Figure 3 at a smaller sale in order to observe the disontinuity in shear rate at the transition between the solid and the liquid phases for ( the oil based mud and ( the water based mud. of the type (3). Note that the parameters are linked by the relation: γ = α( m+) R In pratie, we fitted suh a veloity profile to our data and thus determined the parameters of the onstitutive Equations (, 2). It has been demonstrated elsewhere [7] that, when suh (5) a model an be well fitted to a MRI veloity profile over two or three deades of veloities, a Hershel-Bulkley model annot represent the same data orretly: there neessarily remains a strong disrepany between this model and the data either at high or low veloities, whatever the hosen parameters. This mainly omes from the fat that, as learly appears from Figure 4, the veloity profiles have a slope rupture at the approah of the absissa axis: the loal shear rate remains

7 P Coussot et al. / Rheologial Behavior of Drilling Muds, Charaterization Using MRI Visualization 29 almost onstant over 5 mm then drops to zero as the tangential veloity reahes zero within the unertainty of our experiments. In suh a frame only a model like for example (, 2) prediting a shear rate disontinuity at a ritial stress seems apable to represent the behavior, and not a Hershel-Bulkley model, whih predits that the shear rate ontinuously dereases to zero as the shear stress tends to the yield stress. Now, let us ompare the results obtained from MRI with those obtained from onventional rheometry: MRI data predits that the fluid may be represented by a power-law model at least up to a ertain shear rate (see Table 3) but that below γ no stable flow an our. This is preisely what we obtained from reep tests and the ritial shear rates obtained by the two tehniques almost exatly oinide (see Fig. 2). For the water based mud the results are perfetly onsistent: in reep tests, below a ritial stress, the material tends to ompletely stop flowing, in agreement with MRI data whih orrespond to steady state. For the oil based mud it is likely that at stresses below the ritial value a more or less stable loalization of shear, lose to the wall, sets up, whih leads to an apparent plateau of shear rate in time. However in that ase the flowing material signifiantly differs from the bulk material so that usual (marosopi) rheometrial data are not relevant. Suh an effet annot our in the MRI-rheometer sine the imposed stress remains signifiantly larger as long as the rotation veloity makes it possible to shear a signifiant part of the material. TABLE 3 Material Oil based mud Water based mud Rotation veloity (rpm) 20 R (m) m 3 n /5 /7 α (m/s) 3 8 γ (/s) CONCLUSIONS The rheologial behavior of two drilling muds (water based and oil based) have been studied using onventional rheometrial tests oupled with MRI/rheometrial haraterizations. These experiments learly show that the two drilling muds have similar harateristis: below a ritial shear rate, they annot reah a stable flow regime, when the imposed shear rate is lower than the ritial one, the deformation is loalized in a partiular region of the sample and there is a shear rate disontinuity at the interfae between the flowing and stati regions. The marosopi rheologial behavior of the fluid (as obtained from onventional rheologial experiments) is then the signature of the behavior of this partiular sheared area and not of the whole sample. The behavior below the ritial shear rate therefore should depend on the onstitutive elements of the fluids, partiularly when the size of the sheared area beomes of the order of magnitude of the size of the onstitutive elements of the mud (partiles, droplets, et.). This will be investigated through future studies. These results provide new insights into some drilling problems related to drilling muds, espeially onerning the low shear rate regime. Further work will also fous on the impliations of these data on the harateristis of flows of muds in annular geometries under various regimes. REFERENCES Nguyen J.P. (993) Le forage, Éditions Tehnip, Paris. 2 Raynaud, J.S., Mouheront, P., Baudez, J.C., Bertrand, F., Guilbaud, J.P. and Coussot, P. (2002) Diret determination by NMR of the thixotropi and yielding behavior of suspensions. J. Rheol., 46, Hanlon, A.D., Gibbs, S.J., Hall, L.D., Hayok, D.E., Frith, W.J. and Ablett, S. (998) Rapid MRI and veloimetry of ylindrial Couette flow. Magn. Reson. Imag., 6, Coussot, P., Nguyen, Q.D., Huynh, H.T. and Bonn, D. (2002) Visosity bifuration in thixotropi, yielding fluids. J. Rheol., 46, Coussot, P., Nguyen, Q.D., Huynh, H.T. and Bonn, D. (2002) Avalanhe behavior in yield stress fluids. Phys. Rev. Lett., 88, Coussot, P., Raynaud, J.S., Bertrand, F., Mouheront, P. Guilbaud, J.P., Huynh, H.T., Jarny, S. and Lesueur, D. (2002) Coexistene of liquid and solid phases in flowing soft-glassy materials. Phys. Rev. Lett., 88, Herzhaft, B., Rousseau, L., Neau, L., Moan, M., Bossard, F. (2003) Influene of Temperature and Clays/Emulsion Mirostruture on Oil-Based Mud Low Shear Rate Rheology. Soiety of Petroleum Engineer Journa. 6 Barnes, H.A., and Walters, K. (985) The yield stress myth? Rheologia Ata, 24, Jarny, S., and Coussot, P. (2002) Caratérisation des éoulements de pâte dans une géométrie Couette. Rhéologie, 2, Final manusript reeived in Deember 2003 Copyright 2004, Institut français du pétrole Permission to make digital or hard opies of part or all of this work for personal or lassroom use is granted without fee provided that opies are not made or distributed for profit or ommerial advantage and that opies bear this notie and the full itation on the first page. Copyrights for omponents of this work owned by others than IFP must be honored. Abstrating with redit is permitted. To opy otherwise, to republish, to post on servers, or to redistribute to lists, requires prior speifi permission and/or a fee. Request permission from Doumentation, Institut français du pétrole, fax , or revueogst@ifp.fr.