Travel & Places

Cement-based mixes: Shearing properties and pore pressure

Description
Cement-based mixes: Shearing properties and pore pressure
Published
of 10
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/236389431 Cement-based mixes: Shearing properties andpore pressure  ARTICLE   in  CEMENT AND CONCRETE RESEARCH · JANUARY 2012 Impact Factor: 2.86 · DOI: 10.1016/j.cemconres.2011.09.007 CITATIONS 13 READS 107 5 AUTHORS , INCLUDING:Arnaud PerrotUniversité de Bretagne Sud 76   PUBLICATIONS   326   CITATIONS   SEE PROFILE Vincent PicandetUniversité de Bretagne Sud 63   PUBLICATIONS   370   CITATIONS   SEE PROFILE Hervé BellegouUniversité de Bretagne Sud 6   PUBLICATIONS   119   CITATIONS   SEE PROFILE Sofiane AmzianeInstitute Pascal 94   PUBLICATIONS   540   CITATIONS   SEE PROFILE All in-text references underlined in blue are linked to publications on ResearchGate,letting you access and read them immediately.Available from: Sofiane AmzianeRetrieved on: 04 February 2016  Cement-based mixes: Shearing properties and pore pressure Thibaut Lecompte  a , Arnaud Perrot  a, ⁎ , Vincent Picandet  a , Hervé Bellegou  a , So fi ane Amziane  b a Laboratoire d'ingénierie des Matériaux de Bretagne, Université de Bretagne Sud, Université Européenne de Bretagne - Centre de Recherche de Saint Maudé,BP 92116, 56321 Lorient Cedex, France b Clermont Université, Université Blaise Pascal, EA 3867, Laboratoire de Mécanique et Ingénieries, BP 10448, F-63000 Clermont-Ferrand, France a b s t r a c ta r t i c l e i n f o  Article history: Received 8 April 2011Accepted 14 September 2011 Keywords: Rheology (A)Fresh concrete (A)Workability (A)Thixotropy This study presents srcinal results on the rheological measurement of concrete mixes. It focuses on how todetermine their mechanical and physical behavior under shearing stress. More speci fi cally, the in fl uence of aggregate content on shearing properties is studied. A vane rheometer was developed to characterize freshcement-based materials. In addition to the conventional concrete rheometer, a special hydraulic pressuretransducer was  fi tted to the container to monitor the pore water pressure variation while shearing the ma-terial. Experiments on cement paste, mortar, and concrete bring a new approach to help us understand thebehavior of fresh-state mixes. The results show  1 ) a correlation between water pore pressure and torque ap-plied on the vane;  2 ) a critical sand volume fraction,  ϕ c , as a limit between colloidal interaction behavior andfrictional behavior in mortars; beyond this critical fraction, a leap in yield stress and a drop in pore pressuredue to granular dilatancy are noticed;  3 ) the granular content clearly in fl uences the increase in yield stress of the cement mixes: above  ϕ c , this increase becomes negligible.© 2011 Elsevier Ltd. All rights reserved. 1. Introduction The emergence of several more complex concrete mixtures, suchas self compacting concrete (SCC), has caused concrete rheology tobecome paramount [1].Cement-based mixes are suspensions of particles in a wide range of sizes.Ithasbeendemonstratedthatthemechanismsgoverningthemate-rial's shearing properties strongly depend on particle size and volumefraction.Coussot and Ancey [2] de fi ne a classi fi cation of the rheophysical re-gimes of suspensions. At a low strain rate, they distinguish three succes-sive paramount effects depending on the suspension solid volumefractionandparticlesize:Brownianeffectsforhighly-dilutesuspensions,colloidaleffects for “ soft ” suspensions,andfrictioneffectsfor  “ hard ” sus-pensions. The transition between frictional and hydrodynamic interac-tions occurs for a critical value of solid particles volume fraction  ϕ c .Considering a suspension of aggregates in a cement paste, Yammine etal. [3] clearly show that this transition induces a leap in yield stress. Thecritical volume fraction  ϕ c  is the lower value of the solid fraction whichallows particle contact networks to exist throughout the suspension.Then, below this value, the cement paste forms layers between particlesand controls the yielding process.Mansoutre et al. [4] demonstrate for C 3 S paste that above a criticalconcentration  ϕ c  of C 3 S particles, normal forces appear during shear-ing in common rotational plate – plate geometry. The appearance of normal forces proves the existence of particles contact and samplevolume increase. At such a volume fraction, a C 3 S paste behaves inpart as a dense dilative granular material [4,5]. Such behavior is also reported for cement pastes [6].Above ϕ c ,contactthroughoutthesuspensionformsacontinuousnet-workofparticles.Yieldingisthenstronglyin fl uencedbygranularfriction.Abriak and Caron [7] show that for such a packing, shearing consists inthe formation of voids, and consequently in the concentration of localforces in a localized yield band. It gives a new resistance to the mediumand induces dilatancy. Stone and Muir Wood [8] observe that particlesize affects the thickness of localization, and consequently in fl uencesboth dilatancy (void creation) and yield behavior of granular media.Voidcreationinducesporepressurevariationsthatcanberepresentativeof the material's dilatant behavior. Amziane et al. experimentally showthat pore pressure and yield stress variations are linked [9,10]. Theirwork showed that before the Vicat initial setting time, the pore pressuredecrease seemed to be proportional to the yield stress increase.Usually,thedilatancyofgranularmediacanbemeasuredwithatriax-ialdeviceorwitha shearbox.Intheshearboxdevice,twodisplacementsensors are placed at both ends of the box to quantify the dilatancy. Forthe triaxial method, pore pressure variation is the parameter used toquantify the volume variation of the specimen under shearing.These two experimental techniques are inappropriate in the caseof fresh cement-based material, due to the high  fl uidity of the mate-rial, the range of applied shear rate and the representative volume re-quired for a representative test.Several concrete rheometers have been developed in the last threedecades [11,12]. In the case of concentrated suspensions, such as fresh cement-based mixtures, the vane geometry allows for an accurate and Cement and Concrete Research 42 (2012) 139 – 147 ⁎  Corresponding author. Tel.: +33 297874577; fax: +33 297874571. E-mail address:  arnaud.perrot@univ-ubs.fr (A. Perrot).0008-8846/$  –  see front matter © 2011 Elsevier Ltd. All rights reserved.doi:10.1016/j.cemconres.2011.09.007 Contents lists available at SciVerse ScienceDirect Cement and Concrete Research  journal homepage: http://ees.elsevier.com/CEMCON/default.asp  direct measurement of yield stress [13 – 17]. Recent studies on freshcement-basedmixesshowthatyieldstressisthemostin fl uentialpa-rameterforconcrete castingandpre-casting[[18 – 21].Withvanege-ometry,thestressgrowthtestisabletomeasuretheyieldstress[22],which is generally regarded as the transition stress between elasto-plastic solid-like behavior and viscous liquid-like behavior.The main objective of this work is to describe the transition inshearing regime for cement mixes according to aggregates contentand setting time (during the  fi rst 3 h). To reach this goal, dilatancyand yield stress have to be estimated for several aggregate contents(from no aggregate to aggregate packing) and resting times. There-fore, an innovative rheometer has been developed to measure simul-taneously shearing torque and pore pressure at the container wall.Thepresentworkbeginswithastudyperformedonself-compactingcement-basedmixes,toobservethein fl uenceofsand,gravel,andpasteage.This studyhelpsvalidate theuseofsucha device, andmakessomerecommendations.Thesecondpartfocusesonthein fl uenceofsandvol-ume fraction (aggregate content) on yield stress, structural build-up,and the drop in pressure properties. 2. Experimental procedures  2.1. Materials The different mix designs are presented in Table 1.A Portland cement (CEM I/52.5 N) is used. According to the manu-facturer, thecement is 95.5% clinker (in massfraction)and 4%gypsum.The cement has no other mineral admixture or  fi ller. The cement sizedistribution, measured in ethanol using a laser granulometer, rangesbetween 1 μ  m and 100 μ  m and the d 50  is 15 μ  m.For self-compacting concrete (SCC), common Loire river sand wasused. Sand particle size ranges from 20  μ  m to 3.15 mm and sand ab-sorption capacity is 0.9%. Gravels were 6 mm to 10 mm crushedgravels, with an absorption capacity of 3.6%. The  fi ller is a limestone,with particle sizes ranging from 0.1 to 100 μ  m (d 50 =15  μ  m). Theamount of water is corrected to take into account the water absorbedby sand and gravels. It ensures that the cement paste presents a con-stant water to cement ratio, W/C, equal to 0.35.The self-compacting cement paste (SCP) is also tested without anyaggregate.FormortarmixesM 0 toM 3 ,Frenchnormalizedsand(CECNEN196-1)was used to reduce the variability of size distribution between the sam-ples. This sand has a water absorption capacity of 0.9% in mass. Theamountofwaterisalsocorrectedtotakeintoaccountthewaterabsorbedbythesand.Itensuresthatthecementpastepresentsaconstantwatertocement ratio, W/C, equal to 0.3 for all mortars.For all mixes, a water-reducing admixture is used. It is a polycar-boxylate-based superplasticizer, a liquid containing 20% of dry ex-tract. Its recommended dosage ranges from 0.3% to 3% per weightof cement.The standard French normalized sand used for mortar mixes wasalso tested in saturated condition. The sand volume fraction of thesheared sample was 0.65.  2.2. Device Two different rheometers were used. The  fi rst, called device 1, is aconventional Anton Paar Rheolab QC rheometer equipped with avane geometry, well adapted for cement paste and mortar. The vanegeometry used in this study consisted of four blades around a cylin-drical shaft. The blade height and diameter were chosen from twotool geometries to optimize the measurement accuracy. Tool geome-tries are summarized in Table 2 with the ranges of associated mea-sured yield stress. This device was used to measure the increase inyield stress of the material at rest.Thesecond,calleddevice2,isanoriginalrheometer fi ttedwithpres-sure transducers. It was designed as depicted in Fig. 1. This rheometerconsistsofafour-bladevane,immersedinthegranular fl uidsuspensionandrotatingatacontrolledrate.Thevanedimensionsarequitelarge,soas to enable the study of concrete mixes: 120 mm height H, and120 mm diameter D. The vane is connected to the rheometer with akeyless chuck. The rheometer can be quickly set on a framework andpositionedinastandardcontainertofacilitatetestoperationandensureconsistenttestgeometry(Fig.1).Thegapbetweenbladesandcontainerwallisalwaysgreaterthantentimesthebiggestparticlesize,inordertoavoid granular effects such as arch formation or premature localizationonthecontainerwall.Inthepresentstudy,twocontainerswereused:a186 mm (inner-diameter) cylinder for cementpastes or mortars, and a  Table 1 Mix designs and performed tests. SCP C =  874 0.35 C ×   SCC C =  320 0.425 C 2 C3 C   Saturated sand 3501755   M 0 C =  1606 0.3 C   M 1 C =  12270.305 CC   M 1.5 C = 9930.309 C1.5 C   M 1.75 C =  8340.314 C1.75 C   M 1.88 C =  7720.316 C1.88 C   M 2 C =  7440.317 C2 C   M 2.5 C = 7180.318 C2.5 C   M 3 C =  6310.323 C3 C      F   r   e   n   c    h   N   o   r   m   a    l    i   z   e    d   s   a   n    d   L   o    i   r   e   r    i   v   e   r   s   a   n    d Type of sand2.5% CTested ondevice 1Tested ondevice 2Cement (kg/m 3 )WaterFillerAdmixtureSand W/C (taking into account aggregates water absorption)Gravel0.30.350.781 C1.42% C    Table 2 Vane geometries.Radius (mm) Height (mm) Stress range (Pa)T 1  20 44 100 – 5000T 2  40 60 10 – 500140  T. Lecompte et al. / Cement and Concrete Research 42 (2012) 139 – 147   500 mm (inner-diameter) cylinder for concretes. The inner surfaces of the containers were covered with sandpaper to avoid slippage.The pore pressure is measured along the container wall using apressure transducer device (see  “ 1 ”  in Fig. 2) through the  “ de-aeratorblock ”  fi lled with water ( “ 2 ”  in Fig. 2). To separate the cement pastefrom the measurement system, a  fi ltering device (compacted cotton fi bers) was used ( “ 3 ”  in Fig. 2). The balance of pressure between thewater in the chamber and the water in the paste was achieved bythe transfer of pressure through the  fi lter. Tests carried out withwater showed that the response of the measuring apparatus to thevariations of hydraulic pressure was instantaneous [23 – 26].It is important to note that pore pressure variations inside the mixsigni fi cantly depend on the material's hydraulic conductivity. Freshcement-based materials exhibit a lower hydraulic conductivity thanpure granular materials [27]. Even if pore pressure variations are expected to be greater in the shearing zone for low-permeability ma-terials such as cement-based materials, pressure variations can bepartly absorbed through the sample (between the shearing zoneand the measuring device at wall level). This implies that pore pres-sure variation amplitude is not an intrinsic parameter, but that italso depends on the device geometry.  2.3. Procedure In the fi rst part of the study self compacting cement paste (SCP), sat-uratedsandandcommonindustrialself-compactingconcrete(SCC)weretested. Self compacting cement paste behavior is dominated by colloidalinteraction, while saturated sand behavior is frictional. Comparison of test results in these two extreme cases advances our understanding of SCC shearing behavior.Inthesecondpart,mortarmixesM 0 toM 3 (Table1)weredesignedtoevaluate pressure and yield stress evolutions as the sand fraction in-creases in the mix.The last two columns of  Table 1 present tests performed on thesamples. Saturated sand and SCC were directly tested in the home-made rheometer, whereas a study of the yield stress increase linkedto thixotropy was carried out  fi rst on mortar mixes (with device 1).  2.3.1. Yield stress measurement (device 1) Fresh cement-based material rheology is quite dif  fi cult to mea-sure. These mixes exhibit time-dependent behavior due to thixotropy[1,28 – 32] andthe beginningof hydration[33].The thixotropicbehav- ior of cement-based materials is related to the coagulation, disper-sion, and re-coagulation of cement particles [29,31,34].Yield stress is the most relevant parameter to study the impact of rheology on the common casting process [18 – 21]. However, as forany fl occulating suspension,yield stress largely depends on the struc-tural build-up of the colloidal cement suspension. As a result, theyield stress of cement-based materials increases at rest as the materi-al is structured. As a result, an experimental protocol has to be foundin order to take into account the structural build-up. Roussel predictsa linear increase of theyield stress over resting time. Thisrelationshipis written as follows in [35]: τ 0  t  rest  ð Þ ¼  τ 0  0 ð Þ þ  A thix : t  rest   ð 1 Þ where t rest  is the resting time and A thix  is the structuration rate of thecement-based material in Pa/s. As a result, it is crucial to link the yieldstress to the material's resting time. Just after mixing, some of the mortar was poured into a20cm×20cm×5 cm container. A measurement stage was performedfor 120 s on the Anton Paar rheometer, to obtain the yield stress at agiven resting time, following the stress growth procedure describedbyMahautetal.[22].Alowconstantshearrateisappliedtothematerial (vane velocity 0.1 rpm). At such a low shear rate, viscosity effects arenegligible and stress growth procedure makes it possible to work outthe yield stress. Hackley and Ferraris [36] provide the yield stress stan-dard de fi nition as the critical stress below which the material behaveslike a solid. Once the torque peak value is reached, the yield stressvalue can be determined, as the material  fl ows in the sheared zone. Inconsequence, the yield stress is computed from the torque peak valueas in [37]. It is considered that when the torque value is maximal,yield stress is reachedon the sheared surface described bythevane ro-tation [38,39]. In consequence yield stress is written: τ 0  ¼  C  π : D 2 2  :  H   þ  D 3   ð 2 Þ whereCisthetorquepeakvalue,HandDarethetoolheightanddiam-eter, respectively.Every10min,thevanewasmovedinthesampletoperformanothermeasurement. The gap between each measurement location and thecontainerwallwaswideenoughtoavoidanyscalingorwallslipeffects(more than ten times the maximum particle diameter).  2.3.2. Torque and pressure measurement (device 2)  Justaftermixing forself-compactingdesigns,orafter1hformortars,mixes were poured into the container of the pressure- fi tted device tomeasure both pore pressure and torque once every 40min until an ageof about 3h. The vane (120 mm height, 120 mm diameter) was im-mersed in the tested sample at a depth of 90mm. This depth ensuresthatthepressuresensorsfacethemiddleofthevane.Vibrationswereap-pliedbyavibratingpokertoensureatotaldestructurationofthesuspen-sion(cementparticledispersion).Thepokerdiameteris30mmandwasimmersedbetweenthevaneandthecontaineratadepthof120mm.Thepoker is moved inside the sample in order to vibrate the whole sample.This dynamic excitation breaks all the reversible bonds due to  fl occula-tion or hydration between cement particles. This vibration cancelled theyield stress increase of the material at rest, due to thixotropic behavior Fig. 1.  Rheometer  fi tted with pressure sensors. Fig. 2.  Pore pressure measurement system  —  (1) pressure transducer  —  (2) de-aeratorblock  —  (3)  fi ltering device.141 T. Lecompte et al. / Cement and Concrete Research 42 (2012) 139 – 147   [28]. The fl uid suspension then underwent a stress growth test at a vanevelocity of 0.6rpm for a period of 150 s, and the pastes were allowed torestforseveralminutes.Theprotocolandanalysisofstressgrowthisde-scribed in the previous section. The procedure was renewed over 3h. Itshould be noted that vibration was applied before but not during eachstressgrowth.Atrestandaftervibration,thetorqueisnull,andhydraulicpressurecorrespondstotheweightofthematerialabovethesensor.Ma-terial homogeneity was checked after the test. No water bleeding or ag-gregates volume fraction gradient were observed. 3. Contraction and dilatancy phenomena during shearing: sandand self-compacting mixes  3.1. Sand The French normalized sand, described in Table 1, was saturated,compacted with a drop hammer and subjected to vibrations. The mea-sured solid volume fraction was 0.65. Fig. 3 shows the evolution of tor-que and pressure during shearing at a constant rotational velocity. Anincrease in torque occurs simultaneously with a drop in pore pressureduring stress growth. Then, simultaneous peaks in torque and porepressureareobserved,followedbyareturnofporepressuretoitsinitialvalue,and a torquedecreasetoa constantvalue. Thisexpected resultiswell described as the Reynold's dilatancy phenomenon [40].In the pre-peak phase, the torque and the drop in pore pressureevolve simultaneously. The effective stress on a granular medium in-creases, and the frictional effect between particles is magni fi ed. Con-sequently,theshearstressandtherecordedtorqueincreasewhilethevanerotates.Atthetestedsolidfractions,shearinginduces sandstericeffects leading to dilatancy and a measured yield stress linked to thedensity and frictional characteristics of sand contact points. Thepore volume increase induces a pore pressure decrease. It should benoted that the torque peak coincides with the minimal pore pressure.After the peak, the torque decreases sharply and remains constantdue to the frictional behavior of sand. Due to the high hydraulic con-ductivity of sand, water fl ows quickly through the sand, both porosityand pore pressure return toward the initial value. As expected, the fi nal variation of pressure after shearing is null in saturated sand, asthe water level has not changed after the test.  3.2. Self-compacting paste In any study of SCP or SCC, it should be kept in mind that the be-havior of cement-based material is time-dependent. At rest, duringthe  ‘ dormant ’  period before the material begins to set, both reversibleand irreversible bonds between cement particles are created[1,29,31,34]. This coagulation – fl occulation mode induces a structuralbuild-upofthecementpasteatrestwhichis knowntoinducea linearincrease of the paste yield stress during the  fi rst hour. Then, the hy-dration rate of the cement increases. Setting begins, water is con-sumed and paste stiffness rapidly increases as does the hydratessolid volume fraction [24,41].Fig. 4 shows an increase of pore water pressure between eachshearing test during the  fi rst 100 min of the test. At this age, the ce-ment paste  fl ow regime is not frictional, contrariwise to saturatedsand [2]. SCP behaves as a low yield stress  fl uid with a behavior dom-inated by cement particles colloidal interactions. This phenomenoncan be explained as follows: at rest, particle clusters are formed in ce-ment paste [29,31]. The bonds between particles can be reversible orirreversible, and areheld to be responsible for the material'sthixotro-pic behavior [2,29,31]. During shearing, the reversible bonds are bro-ken, leading to a shrinkage effect inside the mix, and a reduction of the particle assembly volume in the sheared band [6]. This reductionof particle assembly volume induces a slight increase in pore pres-sure. When cement particles move, they can also meet other typesof resistance (such as viscous resistance) from the surrounding ce-ment particles [42]. After the peak, if the cement paste is allowed torest, it is observed that the water pressure becomes hydrostaticagain. We note that no signi fi cant torque was observed (Fig. 4), dueto the torque sensor sensitivity and the SCP's low yield stress.After 100 min, the energy of vibrations is no longer suf  fi cient tobreak the bonds, due to thixotropy and the beginning of hydration.These new-formed bonds between particles seem to create a perco-lated network. The C – S – H have also adsorbed several layers of water molecules [43], which means that formed hydrates increasethe solid volume fraction. At this stage, cement paste begins to be-have as a dilative frictional material: shearing induces a decrease inpore water pressure. A solid cement particle network has been set,and dilatancy in the shearing band induces a suction effect on thewhole network. Each torque peak occurs simultaneously with adrop in hydraulic pressure. Actually there is a competition betweentwo phenomena:1/At rest and after vibrations, pressure represents the apparentweight of the suspension. Then, as shown in Fig. 5, the pressure glob-ally decreases over time: this is the effect of the increase of wall fric-tion stress due to structural build-up. This result has already beenobserved in recent studies on formwork pressure evolution [44 – 48].2/During shearing, a short time suction occurs when the granularmedium is dilative. Fig. 5 is a zoom on one experimental increment.The observation of both torque and pressure curves leads to a fewremarks: •  a peak in torque corresponds to a dramatic drop in pressure; •  after the stress peak, the torque is reduced and tends toward anequilibrium value while the rotation speed is kept constant [22]; 0 10 20 30 40 5005101520 Torque    T  o  r  q  u  e   (   N .  m   ) time (s) beginning of shear test 0.500.751.00 Pore pressure    P  o  r  e   P  r  e  s  s  u  r  e   /   I  n   i   t   i  a   l  p  o  r  e  p  r  e  s  s  u  r  e Fig. 3.  Saturated sand Mortar test  —  torque and relative pore pressure variation for sat-urated sand. 0 30 60 90 120 150 18002468 Torque    T  o  r  q  u  e   (   N .  m   ) time (min) 0.000.250.500.751.00 Pore pressure    P  o  r  e   P  r  e  s  s  u  r  e   /   I  n   i   t   i  a   l  p  o  r  e  p  r  e  s  s  u  r  e Fig.4. Cementpastetest — torqueandrelative porepressurevariationfor cementpasteat an early age.142  T. Lecompte et al. / Cement and Concrete Research 42 (2012) 139 – 147 
Search
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x