Proceedings PRO7 Self-Compacting (Gernenace Proceedings of the First International RILEM Symposium Edited by A. SKARENDAHL and 0. PETERSSON1 International RILEM Symposium on Self-Compacting Concrete Stockholm, Sweden September 13-14, 1999 Edited by A. Skarendahl and ©. Petersson Swedish Cement and Concrete Research Institute RILEM Publications S.A.R.L. The Publishing Company of RILEM This one 72LA-XFPx3dJH,Published by RILEM Publications s.a.r.1 ENS - 61 Av Pdt Wilson F-94235 Cachan Cedex France Tel: + 33147402890 Fax: +33 1474001 13 http:/www.rilem.ens-cachan.fr E-mail: sg@rilem.ens-cachan.ft © 1999 RILEM - Tous droits réservés. ISBN: 2-912143-09-8 Publisher's note: this book has been produced from camera ready copy provided by the individual contributors. The publisher makes no representation, express or implied, with regard to the accuracy of the information contained in this book and cannot accept any legal responsability or liability for any errors or omissions that may be made RILEM, The International Union of Testing and Research Laboratories for Materials and Structures, is a non profit-making, non-governmental technical association whose vocation is to contribute to progress in the construction sciences, techniques and industries, essentially by means of the communication it fosters between research and practice. RILEM's activity therefore aims at developing the knowledge of properties of materials and performance of structures, at defining the means for their assessment in laboratory and service conditions and at unifying measurement and testing methods used with this objective. RILEM was founded in 1947, and has a membership of over 900 in some 70 countries. It forms an institutional framework for cooperation by experts to : © optimise and harmonise test methods for measuring properties and performance of building and civil engineering materials and structures under laboratory and service environments, * prepare technical recommendations for testing methods, © prepare state-of-the-art reports to identify further research needs, collaborate with national or international associations in realising these objectives. RILEM members include the leading building research and testing laboratories around the world, industrial research, manufacturing and contracting interests, as well as a significant number of individual members from industry and universities. RILEM's focus is on construction materials and their use in building and civil engineering structures, covering all phases of the building process from manufacture to use and recycling of materials. RILEM meets these objectives through the work of its technical committees. Symposia, workshops and seminars are organised to facilitate the exchange of information and dissemination of knowledge. RILEM's primary output consists of technical recommendations. RILEM also publishes the journal Materials and Structures which provides a further avenue for reporting the work of its committees. Many other publications, in the form of reports, monographs, symposia and workshop proceedings are produced.Contents Preface XV Membership XVI PART | - KEYNOTE LECTURES 1__Self-compacting concrete. Development, present use and future 3 H. OKAMURA and M. OUCHI Kochi University of Technology, Japan 2 SCC is an important step towards industrialisation of the building industry 15 J. BYFORS NCC AB, Sweden PART 2 - RHEOLOGY/WORKABILITY 3__Evaluation of flow of self-compacting concrete by visualization technique 25 S. URANO', C. HASHIMOTO? and Y. TSUJP 1. Institute of Technology, Shimizu Corporation, Japan 2. University of Tokushima, Japan 3. Gumma University, Japan. 4 Three-dimensional discrete element simulation of rheology tests of self-compacting concrete 35 M. A. NOOR' and T. UOMOTO?” J. Institute of Industrial Science, University of Tokyo, Japan 2. Center for Collaborative Research, University of Tokyo, Japan 5 Fine mortar rheology in mix design of SCC 47 P. BILLBERG Swedish Cement and Concrete Research Institute, Sweden6 Rheological approach to passing ability between reinforcing bars of self-compacting concrete 59 T. NOGUCHL, S. G. OH and F. TOMOSAWA University of Tokyo, Japan 7 A flow analysis for self-compacting concrete n Y. MUROGA', T. OHSUGA', S. DATE! and A. HIRATA? 1. Ishikawajima Construction Materials Co. Lid, Japan 2. Ishikawajima-Harima Heavy Industries Co, Ltd, Japan 8 Characterization of the rheological properties of cement paste for use in self-compacting concrete 83 W. SAAK', H. M. JENNINGS" and S. P. SHAH? 1. Dept. of Materials Science and Engineering 2. Dept. of Civil Engineering NSF Center for Advanced Cement-Based Materials, Northwestern University, Evanston, Illinois, USA 9 Rheology tests for self-compacting concrete — How useful are they for the design of concrete mix for full-scale production? 95 M. EMBORG Betongindustri AB, Stockholm and Luled University of Technology, Sweden PART 3 - PROPERTIES 10 Properties of mortar for self-compacting concrete 109 P.L.J_DOMONE and J. JIN University College London, United Kingdom 11_A simple evaluation method for interaction between coarse aggregate and mortar particles in self-compacting concrete 121 M_OUCHI' and Y. EDAMATSL” 1. Kochi University of Technology, Kochi, Japan 2. Sumitomo Osaka Cement Co. Ltd, Osaka, Japan 12_Comparative study on the induced hydration, drying and deformations of self-compacting and ordinary mortars 131 G. VILLAIN, V. BAROGHEL-BOUNY and C. KOUNKOU Laboratoire des Ponts et Chaussées, Paris, France viStability of self-consolidating concrete, advantages, and potential 14 applications __ 143 K. H. KHAYAT', C. HU? and H. MONTY? Iniversité de Sherbrooke, Québec, Canada 2. Lafarge Laboratoire Central de Recherche, St Quentin Fallavier, 3. Monsanto-Kelco, San Diego. CA. USA Self-compacting concrete. stability control 153 15 EF. CUSSI GTM Construction, France Study of self-compactability of high-fluidity concrete 163 16 - [ARA' and S. NAG -- 1. Cerefo Research, Japan 2. Niigata University, Japan Microstructural features and related properties of self-compacting 17 ind Concrete Research Institute, Sweden Chloride diffusivity of self-compacting concrete 187 18 L. TANG, A. ANDALEN, J. O. JOHANSSON and S. HJELM SP Swedish National Testing and Research Institute, Sweden Strength of hardened self- compacting concrete 199 19 JC. GIBBS and W. ZHU University of Paisley, Scotland, United Kingdom 20 Properties of SCC ~ Especially early age and long term shrinkage 1. Cementa AB, Sweden 2. Scancem Research AB, Sweden Drying shrinkage of self-compacting concrete containing milled V.K. BUI and D. MONTGOMERY University of Wollongong, Australia Vil Material com direitos autorais21_Creep, shrinkage and elastic modulus of self-compacting concrete 239 Lund Institute of Technology, Lund, Sweden 22_Comparative study on properties of self compacting and high performance concrete used in precast construction 251 P. ROUGEAU'’, J. L. MAILLARD' and C. MARY-DIPPE* 1. CERIB, Epernon, France 2. CEBTP, St. Rémy-les-Chevreuse, France 23 Influence of constituents on the properties of self-compacting repair materials 263 F. J. O'FLAHERTY and P. S. MANGAT Sheffield Hallam University, England 24_Hardened SCC and its bond with reinforcement ~~~ M.SNOBI and P, J. M. BARTOS ACM Advanced Concrete and Masonry Centre, University of Paisley, Scotland 25 Structural response of self-consolidating concrete columns 291 K. H. KHAYAT, S. TREMBLAY and P. PAULTRE University of Sherbrooke, Québec, Canada PART 4 - MIX DESIGN 26 A rational mix-design method for self-compacting concrete considering interaction between coarse aggregate and mortar particles 309 Y. EDAMATSU', N. NISHIDA' and M. OUCH?’ 1. Sumitomo Osaka Cement Co., Ltd, Osaka, Japan 2. Kochi University of Technology, Kochi, Japan 27 Optimization of self-compacting concrete thanks to packing model 321 T. SEDRAN and F. DE LARRARD Laboratoire Central de Ponts et Chaussées, Nantes, France VIL28 Investigation on blocking of self-compacting concrete with different maximum aggregate size and use of viscosity agent instead of filler ©. PETERSSON and P. BILLBERG Swedish Cement and Concrete Research Institute, Sweden 29 Utility of statistical models in proportioning self-consolidating concrete K. H. KHAYAT, A. GHEZAL and M. S. HADRICHE Université de Sherbrooke, Québec, Canada 30 Toward mix design for rheology of self-compacting concrete S. G. OH, T. NOGUCHI and F. TOMOSAWA University of Tokyo, Japan 31 Mixture proportioning method for self-compacting high performance concrete with minimum paste volume V.K. BUI and D. MONTGOMERY University of Wollongong, Australia 32 Concrete mixture proportioning with optimal dry packing T. T. ROSHAVELOV Higher Institute of Construction Engineering, Sofia, Bulgaria 33 Design of self-compacting concrete for durable concrete structures F. JACOBS and F. HUNKELER TFB Technical Research and Consulting on Cement and Concrete, Switzerland PART 5 - CONSTITUENT MATERIALS IN SCC 34 Multi component polymer concrete admixtures T. A. BURGE Sika AG, Ziirich, Switzerland 35 Performance of a new-developed powder polycarboxylic acid superplasticizer K. TSUKADA, K. SOEDA, H. HAYASHI and H. ISOMURA. Taiheiyo Cement Corporation, Japan 333 345 361 373 385 397 411 42536 37 38 39 40 4 42 43 Influence of temperature on the dispersibility of polycarboxylate type superplasticizer for highly fluid concrete K. YAMADA', T. YANAGISAWA’ and S. HANEHARA' 1. Central Research Laboratory, Sakura, Taiheiyo Cement Corporation, Japan 2. Technical Research Institute, Maeda Corporation, Japan Property of new polysaccharide derivative as a viscosity agent for self-compacting concrete H. YAMAMURO. Kao Corporation, Japan New admixture to self-compacting concrete P. DAVOUST Perstorp AB, Construction Chemicals, Sweden Reduction of viscosity of concrete with modified lignosulphonate K. REKNES Borregaard Ind. Lid, Ligno Tech, Norway Shrinkage and shrinkage reduction of self-compacting concrete F. J. WOMBACHER Sika R & D, Ziérich, Switzerland The use of steel fibres in self-compacting concrete P. GROTH! and D. NEMEGEER? 1. Luleé University of Technology, Sweden 2. Bekaert AS, Belgium Testing and performance of fiber-reinforced self-consolidating concrete K. H. KHAYAT and Y. ROUSSEL Université de Sherbrooke, Québec, Canada Effect of rheological parameters on self compactability of concrete containing various mineral admixtures A. YAHIA, M. TANIMURA, A. SHIMABUKURO and Y. SHIMOYAMA Taiheiyo Cement Corporation, Central Research Laboratory, Japan 437 449 461 473 485 497 509 §23image not availableimage not availableimage not available67 AIJ recommended practice for high-fluidity concrete for building construction. 775 F. TOMOSAWA', Y. MASUDA’, I. IZUMI’ and M. HAYAKAWA‘ 1. Dept. of Architecture, University of Tokyo, Japan 2. Dept. of Construction, University of Utsumomiya, Japan 3. Institute of Technology, Takenaka Corporation, Japan 4. Technology Research Center, Taisei Corporation, Japan XIVimage not availableimage not availableimage not availableSELF-COMPACTING CONCRETE. DEVELOPMENT, PRESENT USE AND FUTURE Hajime OKAMURA and Masahiro OUCHI Kochi University of Technology, JAPAN Abstract Self-compacting concrete was first developed in 1988 in order to achieve durable concrete structures. Since then, various investigations have been carried out and the concrete has been used in practical structures in Japan, mainly by large construction companies. Investigations for establishing a rational mix-design method and self- compactability testing methods have been carried out from the viewpoint of making it a standard concrete. Recommendations and manuals for self-compacting concrete were also established. 1. Development of Self-Compacting Concrete 1.1 Purpose of development For several years beginning in 1983, Self-Compacting Concrete the problem of the dura lity of concrete structures was a major topic of interest in Japan. To make durable concrete _ structures, durable concrete structures sufficient compaction by skilled workers is required. However, the gradual reduction in the number of skilled workers in Japan's construction industry has led to a similar reduction in the quality of construction work. One solution for the achievement of durable concrete structures independent of the quality of construction work is the employment of self- Fig. 1 Necessity of Self-Compacting Concreteimage not availableimage not availableimage not availableimage not availableimage not availableimage not availableimage not availableimage not availableimage not availableimage not availableimage not availableSCC IS AN IMPORTANT STEP TOWARDS INDUSTRIALISATION OF THE BUILDING INDUSTRY Jan Byfors, NCC AB, Sweden Abstract: In this paper increased efficiency and productivity in relation to site-casted concrete is discussed. Industrialisation of construction could be divided into several components such as, direct design into a 3D-model, efficient assembly at building sites and increased co-operation between contractor, sub-contractors, designers and suppliers. Parameters for industrialisation of site-casted concrete are optimisation of formwork, exchange of traditional reinforcement, mechanisation of handling systems and use of self compacting concrete (SCC). Industrialisation of site-casted concrete including SCC could reduce the manpower by 80% compared with traditional methods. Within NCC successful full-scale testing have been performed and the use of SCC is rapidly increasing. 1. Introduction There are a lot of challenges the building sector has to meet such as: = to achieve a more customer oriented building process - to improve the architectural quality - to build sound and healthy buildings - to achieve more sustainable development - to apply life cycle approaches - to have a more efficient implementation of the information technology - to do this development in a more international perspective and finally - to increase our productivity by industrialisation of the building process In this paper the need for increased efficiency and productivity in general is discussed, particularly in relation to site-casted concrete.image not availableimage not availableimage not availableimage not availableimage not availableimage not availableimage not availableEVALUATION OF FLOW OF SELF-COMPACTING CONCRETE BY VISUALIZATION TECHNIQUE Shinji URANO Institute of Technology, Shimizu Corporation, Japan Chikanori HASHIMOTO Department of Civil Engineering, The University of Tokushima, Japan Yukikazu TSUJI Department of Civil Engineering, Gunma University, Japan Abstract Laboratory experiments are conducted to investigate the horizontal flow behavior of self- compacting concrete in two types of slab forms, set with lattice steel bars and vertical steel bars, by a visualization technique. Also investigated is vertical flow behavior in both V- funnel test apparatus and U-filling ability test apparatus, which are widely used in Japan for the evaluation of the consistency of self-compacting concrete. It is found that the average maximum shear strain rates of concrete in slabs are under 1.0/ s, and the average maximum shear strain rates of concrete in the consistency test apparatus are under 10/s. 1. Introduction Evaluation of the filling ability of self-compacting concrete is most important as it determines whether placement into a form under its own weight without vibration is possible. In order to evaluate the consistency of fresh concrete, the slump (or slump-flow) test is most widely used. In the case of self compacting concrete, however, filling ability cannot be evaluated sufficiently by a slump-flow test alone. In Japan, new consistency tests such as the V- funnel test{1][2] and the U-filling ability test{1][3] are widely used to evaluate self- compacting concrete. However, the relationship between flow behavior in the consistency tests and flow behavior in actual forms is not fully understood in terms of dynamic behavior. In this paper, in order to evaluate the flow behavior of self-compacting concrete quantitatively in terms of an index, shear strain rate, laboratory experiments to investigate the horizontal 25image not availableimage not availableimage not availablez ‘iscoity agent 2.09! g Vein 05 g 08 #0 g £ 06 #0. i i fos § 0. : £ 402 i . 5 5 <0 <0 o 5 Ww Is 20 30 0 5 0 is 2 30 Elapsed time(s) Elapsed time (s) Fig.4: Average maximum shear strain rates Fig.5: Average maximum shear strain rates of slab form set with a lattice of steel of slab form set with a lattice of steel bars (Effect of bar spacing) bars (Effect of viscosity agent) ‘Average maximum shear strain rate 0 5 10 1520 30 Elapsed time (5) Fig.6: Average maximum shear strain rates of slab form set with a lattice of steel bars (Effect of Ve/¥m) whatever the bar spacing. This can be attributed that the flow speed and the deformational speed of the concrete affected by arrangement of steel bars at the early stages only. Figure 5 shows the average maximum shear strain rates of the model concrete with the viscosity agent at levels of 0 and 2.0g/I, with Vg/Vm 0.5 concrete. It is clear that adding the viscosity agent decreases the peak level of average maximum shear strain rate, and decreases the deformational speed at the early stages. However, the quantity of viscosity agent has little effect on the shear strain rate after the peak point. Figure 6 shows the average maximum shear strain rates of the model concrete with Vg/¥m 29image not availableimage not availableimage not available4. Conclusions In this paper, in order to evaluate flow behavior of self compacting concrete quantitatively using shear strain rate as an index, laboratory experiments were conducted to investigate flow behavior in slab forms and in consistency tests by a visualization technique. The results are as follows: (1) Average maximum shear strain rates are under 1.0/s and converge on 0/s with elapsed time in the case of a slab form set with a lattice of steel bars and vertical steel bars. (2) Consistency, as controlled by adding a viscosity agent and Vg/Vm, affect on average maximum shear strain rates at the early stages, namely, that of the front of horizontal flow. (3) The V-funnel test is a test which develops different maximum shear strain rates when concrete of various deformational characteristics is tested. (4) The U-filling ability test is a test which yiclds different convergence times to O/s when concrete of various deformational characteristics is tested. (5) The average maximum shear strain rates are under 10/s in the case of the V-funnel test and the U-filling ability test. The next step in this research is to consider the shear strain rate in a wall form, which is characterized by vertical flow under concrete self-weight, and to obtain valuable information for the flow simulation[6] of self-compacting concrete by analysis. References 1, JSCE “Recommended Practice for Self-Compacting Concrete,” 1998, pp.157-176 (in Japanese) 2. Ozawa, K., Mackawa, K., and Okamura, H., "Development of the High Performance Concrete," Proceedings of the Japan Concrete Institute Vol.11, No.1, 1989, pp.699- 704 (in Japanese) 3. Kuroiwa, S., Matsuoka, Y., Hayakawa, M., and Shindoh, T., "Application of Super Workable Concrete to Construction of a 20-Story Building,” ACI SP-140 High Performance Concrete in Severe Environments, 1993, pp.147-161 4. Hashimoto, C., Maruyama, K., and Shimizu, K., "Study on Visualization Technique 33image not availableimage not availableimage not availableresearch this model was always kept active, unless a contact bond was Bond breaks present (in which case, the contact bond model behavior supersedes the slip model behavior). PFC’ allows particles to be bonded together at hen (> 0 contacts. Two bonding models are supported: a contact bond model and a parallel bond model. The presence of a contact bond, however, inactivates the slip model. Once a bond is formed at a contact between two particles, that contact Fig. 1 Constitutive behavior for contacts occurring at a continues to exist until the bond is point, (a) normal component of contact force, broken. Here, only contact bond (b) shear component of contact force. model has been used not the parallel bond model. A contact bond can be envisioned as a pair of elastic springs (or a point of glue) with constant normal and shear stiffness acting at the contact point. These two springs have specified shear and tensile normal strength. Contact bonds allow tensile forces to develop at a contact. The magnitude of the tensile normal contact force is limited by the normal contact bond strength. If the magnitude of the tensile normal contact force (Fig. I(a)) equals or exceeds the normal contact bond strength, the bond breaks, and both the normal and shear contact forces are set to zero. If the magnitude of the shear contact force (Fig. 1(b)) equals or exceeds the contact bond strength, the bond breaks, but the contact forces are not altered. As parallel bond model has not been used in this research, it is not described. Relative normal displacements, U" (overlap) Bond breaks Shear contact force Maximum shear force Total shear Gsplacement, O° 4. Model setup for simulation Packing is the first step in DEM approach to simulate the behavior of fresh concrete. This procedure resembles the actual process of placing concrete into a container. The purpose of packing is to generate the numerical model resembling the practical state of the test sample just before experiment. All published works known to the authors on computer simulations of the particle-packing problem have employed arbitrary, non- physical rules to decide upon the final particle positions. In present research, however, the complete process was simulated according to Newtonian mechanics with particle interactions controlled according to the theoretical contact mechanics. To simulate the particle deposition process, particles are randomly generated within a prescribed region and then subjected to a gravity field so that they fall as rain within defined container walls. As a consequence, particles collide with the container walls and each other and computations are continued until an equilibrium configuration of the resultant particle has been attained. At the end of the process before the particles settle down they continue moving due to inertia forces. Cycles of relaxation are needed to settle down the particles. For relaxation process some cycles were applied. After data packing simulation 37image not availableimage not availableimage not availablecombined. For concrete model an extra bond property has been introduced between mortar and aggregate. Concrete Level 6. Parameter selection procedure To select the qualitative value of the parameters like friction, contact stiffness, bond value and aggregate grading, sensitivity analysis is required. A test method has to be selected to compare the numerical result with experimental result. To see the qualitative behavior of fresh concrete, lifting sphere viscometer test has been selected. Lifting sphere viscometer is one of the popular devices to . investigate the theological properties of fresh "187 Cromelry ofthe contain! of concrete recommended by JCI [7, 8]. In this (Unit: mm) research, the lifting sphere viscometer test was chosen because it can provide the possibility of direct comparison of the experimental results, obtained from the test, with the simulated results. The type of lifting sphere viscometer device used is shown in Fig 7. It is quite similar to that devised in the literature [9]. The size, shown in the Fig. 7, however, is the half of the original size used in the simulation to reduce the number of sphere elements and simulation time. This selection procedure has nothing to do with mortar or aggregate model. This is just to get the idea of behavior of each parameter. Running time is very important in this regard. From contact constitutive model it is understood that if the normal bond breaks then the contact force resets, but if the shear bond breaks the contact forces don not reset. It goes up to the normal bond strength of the ball. During this phase slip model remains activated according to friction factor. This has been verified [10] by giving the different bond values to the program. It is proposed in Noor and Uomoto [10] that shear bond value would dictate the selection of normal bond value. Once it is selected, the same value can be provided to the normal bond. It has been found from the analysis that shear and bond strength equal to 0.01 (N) gives good qualitative 50. 50 result. 4 Donarete Simuiation Terk 7. Simulation results | RS \ From above discussion for mortar model and aggregate , model the selected parameter values are given in Table 2. 7 hy Wy The grading curve used for mortar and aggregate is given in i “4 I Fig 4, Considering all these parameters Noor and Uomoto [10] did mortar and aggregate simulation, separately. Both Dragging Force (N) the simulations [10] qualitatively simulate the force. %S Oi iis “im displacement curve for mortar and aggregate, respectively. fig g Pierce am Here, one final analysis has been conducted combining two offing sphere models to simulate the behavior of fresh concrete (Fig 8). viscometer test for For concrete simulation, mix proportion of powder type aa it: aimage not availableimage not availableimage not availableFrom the above analysis the ratio of the normal stiffness to shear stiffness has been chosen as 50, and the normal stiffness value of the aggregate has been chosen 1E+05. With these selected parameters two final simulation one for L1-box and another for U-box have been conducted and final simulation stage of U- box test has been shown Fig. 17. The value of the rising height for the L1-box is 11.5 cm and for the U-box is 12.1 cm. in simulation, which are approximately equal to 23 cm and 24.2 cm for real cases. In simulation, actual value does not have much importance but the qualitative result is important to analyze the internal behavior such as blocking of aggregate at the reinforcement etc. This confirms that flow also can be simulated using this DEM model. 10. V-funnel test As a method of evaluating the self-compactibility of concrete, a technique has been proposed for measuring deformation rate. Conventional funnel tests are used to measure the apparent viscosity of paste and grout. A funnel test is a simple mean of evaluating the ability of concrete to pass through spaces. In other “es words, the funnel tests measured different relationships Th between solid particle size and funnel size in the case > of concrete when compared with conventional relationships between powder particle and funnel size. | | Conventional funnels for concrete testing have a circular cross — section through which concrete undergoes a 3-D deformation. In actual formwork, however, concrete, deforms two-dimensionally when passing through such obstacle as reinforcing bars. Therefor, the funnel, which is shown in Fig. 13, has Fig. 18 V-funel. in simulation been used for the current simulation, as funnel forces half of this size has been concrete to deform two dimensionally. After filling the used. (Unit: mm) funnel with concrete to its top edge the discharge port was opened and the time required for the concrete to flow out is measured. One simulation has been conducted using the DEM parameter selected in the previous analyses. For funnel simulation bond has been provided and the value selected for the bond is 0.025 (N) for both normal and shear direction. The time required to flow out all the concrete from the V-funnel has been found out equal to 9.5 sec. which is closed to the actual value measured [12]. ve 5 11. Conclusions In this paper, several consistency and flow test methods have been successfully simulated. The objective of the ongoing research is to find out the relationship between 45image not availableimage not availableimage not availableParticle size analysis of cement, filler material and aggregate has been performed with laser scattering technique by using a Malvern Mastersizer. The results are shown in FIG. 1 and FIG. 2. ‘The aggregate used to simulate the fine “tail” of the total aggregate in concrete is filter sand with the commercial name Rida MO.17. The aggregate is sieved so that maximum particle size becomes 250 microns. 10 Crim F = 90 +| —2—Cements | |_| Z| s | | —e=Cement D \ 80 = = Rada 70 —— Mya 0-10 Ls a = of] —-Mya0-20 / t £ —+—Mya 0-80 2 so —+-ly ff, q —— Mya 10-70 & 40+) ——Mya30-70 - 30 | Mya : ——MyaE ak ' 20 iG + 10 i 0 o 1 10 100 1000 Particle size [microns] FIG. 1. Particle size distribution for tested cement, dolomite filler and aggregate. 49image not availableimage not availableimage not available45 “0 > Cement D 35 2 Cement § £30 # 2s 3 20 Ref Cement S 5 15 Ref Cement D 10 5 o o 10 20 30 40 50 60 40.5, median particle size of filler [microns] FIG. 5. Relationship between median particle size of filler material and yield stress. 25 20 + Cement D a J-o- Cement § é Sas 3 Ref Cement S 210 Z Ref Cement D 05 0,0 o 10 20 30 40 50 0.5, median particle size of filler {microns} FIG. 6. Relationship between median particle size of filler material and plastic viscosity. 53image not availableimage not availableimage not availableamounts corresponding t> 405 and 160 kg/m? respectively. The W/C was kept constant at 0.40. By dosing supe:plasticizer as liquid it is very clear that SP B is much more efficient. The yield stress. is almost zero for the fine mortar when SP B is added at 0.75 %. For the same low yield stress a dose of 1.75 % is needed with SP C. But by recalculating the doses not as liquid but as solid content one can se that the two different superplasticizers are almost exactly equal in efficiency. Note the measured difference in plastic viscosity once the yield stress is minimised. In the context of self-compacting concrete it can be very interesting that SP C shows higher plastic viscosity at very low yield stresses. 7 ce D=405kg cd ene = spc aa), Leeds one tse : z om ge & x g : 3 Increasing SP doses = , ° a « @ * we i Yield stress {Pa} FIG. 9. Rheological resvlts from tests of superplasticizer. 5. Concluding remarks Results have shown how measurements of fine mortar rheology are used in the mix design of self-compacting concrete. The method is sensitive and is therefore useful as instrument for measuring the different impact that cement, filler materials and superplasticizers have on the rheology of the fine mortar part of concrete 6. References 1. Barnes, H.A., Hutton, J.F., Walters, K., “An Introduction to Rheology”, Elsevier, Amsterdam, 1989. 57image not availableimage not availableimage not availableRelative viscosity PaO voy Relative thickness of excess paste Fig. 2 Relationship between relative thickness of excess paste and relative viscosity In this study, the authors semitheoreticaly elucidate the behavior of fresh concrete passing through spaces by rheologically interpreting the results of tests for passing ability through spaces and calculating the changes in the local aggregate content. This study also intends to formulate a new technique to evaluate the passing ability of self- compacting concrete through spaces by assuming fresh concrete as a two-phase material consisting of solid and liquid based on the multi-phase flow theory and using excess paste thickness theory and rheology. 3. The experimental method In this section, passing ability tests are conducted on model concretes on the assumption that fresh concrete is a two-phase material consisting of paste (matrix) and aggregate, in order to minimize the experimental scatter and grasp the effects of parameters more clearly. The model concretes used in the experiments simulating self-compacting concretes consisted of model paste and fine and coarse aggregates. 3.1 Passing ability evaluation method ‘The mechanism of fresh concrete passing through a narrow space is considered to be as follows: When fresh concrete is about to pass a narrow space, a difference arises between the velocities of paste (matrix) and aggregate, and the paste precedes. New aggregate particles flowing from behind add to the remaining particles, increasing the local aggregate content. For this reason, the relative viscosity (concrete viscosity divided by paste viscosity) of the concrete is expected to become higher than the original value when passing through a narrow space. If the relative viscosity of concrete passing through a narrow space is determined experimentally, then the relative excess paste thickness can be obtained (Fig. 2). Also, if the relative excess paste thickness 61image not availableimage not availableimage not availableimage not availableimage not availableimage not availableimage not availableline. Also, this regression curve of the V,y~V'ay relationship theoretically should pass through the origin (when V,y = 0, V‘gy = 0). Therefore this may be shaped as given in Eq. (1). The relationship between the relative clearance, C,, and the changes in the rate of increase in the aggregate volume for each clearance is plotted as shown in Fig. 10. Constants a and b is expressed by the relationship with paste viscosity as shown in Fig. 11. Accordingly, constant a is expected to be in inverse ratio to the viscosity, whereas constant b is proportional to it according to the multiphase flow theory, and the following equations were obtained. a = 0.04837," + 0.0184, b= 1.09n, (4) where (x theoretically has no unit and is in inverse ratio to the relative clearance, C, (Eq. (3)). When the clearance is infinite, i.e., when the conerete flow is not affected by the clearance, x becomes 1, and the original volume fraction of aggregate should not change while the concrete passes through the space. Conversely, when the clearance is very small, « is expected to increase infinitely, rapidly increasing the local aggregate content. In the range of these experiments, the yield values of the model concretes are not used in the process of calculating Eq. (4). Only viscosity was included in the process, as it is difficult to derive correlation from the wide scatter of yield values compared with viscosity. 5. Conclusion The behavior of fresh concrete passing through narrow spaces was discussed from the standpoint of rheology using model concretes representing self-compacting concrete as a multiphase flow, i.e., a two-phase dispersion fluid consisting of paste and aggregate (fine and coarse). The discussion led to the following findings; 1. The increase in the rheological constants when fresh concrete passes through narrow spaces can be grasped as a reduced relative excess paste thickness resulting from an increase in the volume fraction of aggregate. 2. The behavior of fresh concrete passing through narrow spaces can be macroscopically interpreted as changes in its rheological properties.. 3. A passing ability tester with an improved box-shaped obstacle is effective in rheologically analyzing the behavior of fresh concrete passing through narrow spaces. 69image not availableimage not availableimage not available3 D50mm_ 275mm 3)150mm 5 (anit: mm) Figure | : V Funnel Testing Apparatus Table 2 : Mix Proportions 4.71 (unit : kg/m’) Table 3 : Properties of Fresh Concrete and Plastic Viscosity 'V-Funnel Time(s) _[Plastic Viscosity (Pa‘s) TYPE] 160 11.4 75 TYPE2 53, 15.3 160 TYPE3 50. 21.9 45 As an example, results for TYPE2 concrete with plastic viscosity of 60Pa-s are shown in Figure 2 which illustrates relationships between outflow and its time, obtained through experiments and analyses. As is evident from Figure 2, there is a considerably good agreement between experimental results and analytical predictions. These analytical predictions for the cases with V-funnel testing apparatus of 50mm and 150mm wide, were made by using plastic viscosity predicted previously through the flow analysis for the case with V- funnel testing apparatus of 75mm wide. We consider that the agreement stated above exhibits the validity of the prediction of the plastic viscosity. 73image not availableimage not availableimage not available3.Application to a Construction Work of a Pontoon A pontoon (floating pier) is a composite structure consisting of a steel hull covered with concrete. The dimensions of width, length and thickness of the bottom reinforced concrete are 3m, 30m, and 0.16m, respectively. The interior of the pontoon constitutes a narrow space due to arrangement of rebar and shear connectors placed in the pontoon. This leads to an extreme difficulty in compaction of filled concrete. Hence, it was determined to adopt the SCC. Figure 6 shows the principal dimensions of the pontoon and the method for concrete pouring. EON Tine os —— 4 | Pouring Position sao * ls Figure 6 : Dimensions of he Pontoon and Outline of Method for Concrete Pouring 3.1 Evaluation of performance of SCC Prior to the application of the flow analysis to the construction work, a self- compactibility of SCC to be cast was evaluated. For the case of the pontoon construction, a flow distance of concrete was deemed to be Sm or more. In consideration of this flow distance, the self-compactibility and segregation related to flow distance were investigated by casting SCC into a partial model formwork to examine an aggregate distribution. The height and rebar arrangement of the model were identical to the pontoon, and the length and width of the model were 6m and Im, respectively. The investigation stated above revealed that the model formwork was completely filled up with SCC, leading to a conclusion that the SCC poured had an excellent self- compactibility. As shown in Figure 7, segregation and subsidence of aggregate were scarcely discerned even in the flow distance approximate to 6m, and thereby it became clear that the quality of hardened SCC was sufficiently homogencous. Table 7 exhibits the results of the investigation into coarse aggregate distribution. Table 7 : Coarse Aggregate Distribution Flow Distance (m) Coarse Aggregate Ratio (%) 27. 2(Design Value) 77image not availableimage not availableimage not available3.2.2 Prediction of a buoyancy and determination of a height of formwork A height of poured SCC was predicted in order to minimize the buoyancy acting on a core steel hull and height of formwork. For the case of SCC pouring at the moving speed of pouring position of 30m/h stated previously, the height of formwork was estimated at 46cm. On the occasion of the actual construction work of the pontoon, measures against the buoyancy were taken, and the formwork height of 90cm was adopted in consideration of probable variation in the moving speed of pouring position and SCC properties. 3.2.3 Comparison of results between actual construction work and analysis In the case of the actual construction work, sensors to monitor an arrival of poured SCC were mounted on the bottom of the core steel hull in order to check the degree of SCC filling. Figure 10 shows schematically the location of the sensors placed. Figure 11 is a graphical representation of the measured and predicted results for the degree of SCC filling. ‘e Measured (B, C Line) ‘A. Measured (A Line) Predicted (D Line) He Predicted(A Line) Distance from Edge to Pouring Position (m) Figure 11 : Measured and Predicted Results for the Degree of SCC Filling As is seen from Figure 11, a fair degree of agreement was attained between the results measured and those predicted by the flow analysis method. The following were achieved by the application of the flow analysis method recently developed to the actual concrete placing work, and thereby the analysis was able to contribute an establishment of a rational construction scheme. @ Determination of an adequate moving speed of pouring position @ Minimization of a height of formwork @ Avoidance of detrimental concrete segregation 81image not availableimage not availablecement paste rheology to obtain the desired flow properties. The yield stress of the paste must be above a critical minimum to insure that segregation of the aggregates does not occur when the material is at rest. In addition, the viscosity of the paste must be high enough to avoid segregation during placement, yet provide the desired fluidity. The critical minimum yield stress and viscosity of the paste will depend on the properties of the aggregates (i.e. density, particle size distribution, morphology) and the maximum shear rate expected during placement. Several studies have examined SCC mixture design procedures and a summary of some of this research is provided in reference 4. In general, the water to cement ratio is around 1.0 (by volume) with a combination of superplasticizers, mineral additives, and synthetic binders used to achieve the desired fluidity [1, 4]. Few studies have examined how admixtures influence the yield stress-viscosity relationship of cement paste. The purpose of this investigation is to establish an experimental procedure to accurately measure the rheological properties of cement paste for use in self-compacting concrete. 2. Experimental Procedure 2.1 Yield Stress Measurements Accurately measuring the rheological properties of cement paste, let alone SCC, is challenging. The most common method used to determine yield stress is by extrapolating the stress-strain rate flow curve to zero strain rate through use of an appropriate model [5]. The most widely used model is Bingham's linear approximation for flow of a viscoelastic material [6]. Although the simplicity of the Bingham model is attractive, the stress-strain rate relationship of cement paste is highly psuedoplastic and only approximates linearity as the particles become deflocculated. At best, the Bingham model provides an estimate of yield stress at low solids concentrations and loses accuracy as the solids concentration increases [7-8]. Other models more closely approximate the actual flow behavior of cement suspensions, including those of Herschel and Bulkley [9] and Cross [10]. These models require curve-fitting approximations and often only fit a limited portion of the flow curve data. Direct measurement of the stress required to initiate flow is the preferred method for determining yield stress. This type of test is performed by applying a stress ramp or by a similar type of creep-recovery experiment. Another common approach for measuring yield stress is to apply a constant rotational rate and monitor the development of stress as a function of time (Figure 1). Both methods have been shown to produce similar values for yield stress [11]. A controlled stress (or strain rate) rheometer with smooth-walled coaxial cylinders is the most common system for yield stress measurements [Figure 2 (a)]. These systems often show pronounced "wall slip" due to the displacement of the disperse phase away from the solid boundaries [12]. Slip is the result of both static effects (e.g. geometric depletion, disruption of Brownian motion) and dynamic forces (e.g. hydrodynamic, shear force gradient) [13]. 84tetris = * 2 = 300 $ g a S150 2 a of 0 100 200 300 Time (seconds) Figure 1. Stress development versus time for vane testing at an applied constant rotational rate (0.01 rad/s). The maximum shear stress corresponds to the material’s yield stress (t,). i Table 1. Rheometer Sensor Dimensions (mm) Coaxial H Cylinders | _ Vane H [r,] 10.00 11.00 r | 10.85 22.00 [| 30.00 16.00 Ts etal (b) Figure 2. (a) Concentric cylinder and (b) vane sensors. A vane is often used to avoid slip [Figure 2 (b)]. In this test yielding occurs along the localized cylindrical surface circumscribed by the vane. This prevents the possibility of slip since the material yiclds within itself. Finite element modeling and experimental data have confirmed the cylindrical conformity of the yielded surface [14]. Haimoni and Hannant showed that the yield stress of cement paste determined using a vane was 85approximately twice the value obtained when smooth walled coaxial cylinders were used [15]. This result has been confirmed in our laboratory and the influence of slip shown to be more pronounced as the solids concentration increases [16]. For this study, yield stress measurements were conducted using a rotational theometer (Haake Rheostress 150) with a vane. The dimensions of the vane are given in Table 1. A constant rotational speed of 0.01 radians/second was applied and the resultant stress recorded as a function of time (Figure 1). The peak stress in Figure | corresponds to the yield stress of the material. 2.2 Viscosity Measurements Cement paste is shear thinning (i.e. viscosity decreases with increasing shear rate) and exhibits both thixotropy and irreversible structural breakdown [5]. For this reason, the viscosity of cement paste is greatly effected by shear history. Several studies have shown that increased mixing energy decreases both the viscosity and yield stress of cement paste [17-19]. Williams et al. hypothesized that a characteristic equilibrium viscosity would be reached at high shear rates, independent of mixing method [20]. Tattersall showed that an equilibrium stress is obtained after sufficient time at a constant rate of shear [21]. In a similar manner, Lapsin et al. modeled shear stress decay versus time at a constant shear rate and confirmed their results experimentally [22]. Hattori and Izumi also predicted a minimum viscosity for a particular shear rate based on their coagulation rate theory model [23]. These studies suggest that a characteristic equilibrium viscosity exists for a particular shear rate, independent of shear history. Unlike hysteresis loop testing, the equilibrium viscosity is independent of rheometer testing conditions (e.g. stress ramp rate, mixing methods) [20]. It is important to keep in mind the shear rate of the application when measuring the viscosity of cement paste due to its shear thinning nature. The typical shear rate of pouring is between | sto 10s while that of pumping can exceed 1000 s” * (24), For this investigation, viscosity was measured at a relatively high shear rate (600 s), intermediate between the shear rate extremes of pouring and pumping. For this study the cement paste was placed in the rheometer immediately after mixing and a constant shear rate of 600s‘! was applied to the sample for 200 seconds. The equilibrium viscosity (n.) was calculated by fitting a hyperbolic function of the form; n=neto a to the viscosity versus time curve (Figure 3). At the limit where time (t) goes to infinity, the viscosity (n) equals the equilibrium viscosity (n).), where B is a constant depending on the structure of the paste. Concentric cylinders were used for the viscosity measurements [Figure 2(a)]. The dimensions of the system are given in Table 1. Unlike yield stress measurements, wall slip is not as dominant and becomes almost negligible at high shear rates [13, 25]. 86600; T — T T 500 © Experimental Data += Data fit using Equation (1) 400! 300} 200 Viscosity (mPa-s) 0 100 200 300 400 500 Time (seconds) Figure 3. Viscosity as a function of time at a shear rate of 600s". 2,3 Materials For the first stage of testing two commercially available superplasiticizers were examined, naphthalene (Type "N") and melamine (Type "M") based sodium salts . The yield stress and equilibrium viscosity were determined at addition rates of 0.0-1.0 weight % of solids. The influence of two binders commonly used in SCC, silica fume and a cellulose- based powder, were examined following the preliminarily study with the superplasticizers. The specific gravity of the solid silica fume was 2.2 and the surface area measured by BET nitrogen sorption was 23.56 m’/g. The replacement of cement by silica fume was 30% (by volume) for all experiments. The amount of Type "N" superplasticizer was adjusted between 0.60-2.50 weight % solids (cement + silica fume) for the rheology measurements. A surface treated hydroxypropyl methylcellulose powder was used as a synthetic binding agent. The surface coating provides enhanced dispersion characteristics compared to typical cellulose based products that often require thermal cycling or high shear mixing for homogenous dispersion. In water, the powder is easily dispersed at low shear rates without a significant increase in viscosity. As the pH is increased, the surface coating breaks down and the powder begins to hydrate. At the pH typical of most cementitous suspensions, hydration of the cellulose powder occurs in approximately 30 seconds after addition [26]. The cellulose powder was added at addition rates of 0.0-0.40 weight % (of solids) to cement paste containing 30 volume % silica fume and 2.5 weight % Type "N" superplasticizer. ASTM type I cement was used for all of the mixes and the ratio of the volume of solids to water was kept constant at 1.0. The yield stress and equilibrium viscosity at 600 s"' were determined following the procedures outlined in Sections 2.1 and 2.2, respectively. 872.4 Mixing Procedure All batches were mixed for 5 minutes in a mixer corresponding to ASTM C305 for the yield stress measurements [27]. First, the required amounts of water, superplasticizer, and silica fume were added to the mixing bowl. The speed was set at 140 RPM for 1 minute as the cement and cellulose powders were added. The speed was then adjusted to 285 RPM for 2.5 minutes, after which the sides of the bowl were scraped with a rubber paddle. The mixer was then set a 285 RPM for an additional 2.5 minutes. All batches for the viscosity testing were vigorously mixed by hand for 5 minutes in a small beaker with a stirring rod. The batch ingredients were introduced in the same order as mentioned above for the mechanical mixer. Based on the discussion in Section 2.2, the difference in mixing methods should not be significant since the equilibrium viscosity is evaluated at a shear rate greater than that achieved by any of our laboratory mixers [20]. 3. Results and Discussion 3.1 Superplasticizers The influence of superplasticizer dosage on the yield stress (,) of cement paste is shown in Figure 4. As expected, yield stress decreases with increasing superplasticizer content until no yield stress is observed. Surprisingly, the trends for both superplasticizers are nearly identical, with a dosage of approximately 0.75 weight % required to reduce the yield stress to zero. There appears to be a slight increase in yield stress at 0.1 weight % superplasticizer. Currently, we have no explanation for this result and more experiments are planned to determine the cause of this effect. The equilibrium viscosity (7).) as a function of superplasticizer dosage is plotted in Figure 5. The melamine product (Type "M") had a lower viscosity at all dosage levels compared to the naphthalene superplasticizer (Type "N"). In addition, Type "M* reached the minimum viscosity (approximately 154 mPa-s) at a dosage of 0.25 weight %, while Type "N" required 0.50 weight %. The data from Figures 4 and 5 were fit to a best-fit curve and plotted as yield stress versus equilibrium viscosity (Figure 6). The dashed line in Figure 6 represents plain cement paste (containing no admixtures) at different solids concentrations. As mentioned in Section 2.3, all of the mixes used for this study were at a solids concentration of 50% by volume. At a given yield stress (i.e. slump), Type "M" superplasticizer provides a material with a lower equilibrium viscosity (higher fluidity) compared to Type “N". Conversely, at a specific equilibrium viscosity, Type "N” produces a material with a lower yield stress (i.c. greater slump) than Type "M". This simple example demonstrates how the enormous potential benefits of self-compacting concrete cannot be fully realized unless both the yield stress (static rheological properties) and viscosity (dynamic rheological properties) are optimized for a particular application. 88Yield Stress (Pa) 400.>-——>-———— @ 600. é one a a @ Type "N € © Type Type "M x on 300 ” ype 8 3 400 2 200 4 3 g 8 2 > 100 7 & s 200 3 ‘ \ 2. a . 1 0 04 08 12 0 04 08 1.2 Weight % Superplasticizer Weight % Superplasiticizer Figure 4. Yield stress as a function Figure 5. Equilibrium viscosity at 600 s of superplasticizer concentration. versus superplasticizer concentration. 500 z ¢ 400 4 a ’ & : = 4 S ak é a & “Xk Z 20 Cement Paste & (Vol. % Solids = 0.50) 100 100 200 300 400 500 600 700 Equilibrium Viscosity at 600 s' (mPa-s) Figure 6. Yield stress (t,) versus equilibrium viscosity (1).) at a shear rate of 600 s for Type "M" and Type "N" superplasticizers. The dashed line represents plain cement paste (containing no admixtures) at different solids concentrations. 893.2 Silica Fume and Cellulose Binders Figure 7 shows the influence of superplasticizer dosage for cement paste with 30% of solids volume replaced by silica fume. A minimum of 0.6 weight % Type "N" superplasticizer was required to mix the sample. The yield stress decreases sharply between 0.6 and 1.0 weight % and eventually reaches zero near 2.0 weight % superplasticizer. The tend for equilibrium viscosity (Figure 8) is similar to that for yield stress. A minimum viscosity of approximately 98 mPa-s was obtained at a dosage of 1.5 weight % superplasticizer. 1500 T T @ 800 T T z S600} & 1000 1 S g a g = sob 4 a 3 Mo z 2 2 2 500 7 a 5 200} 4 e 2 = 0 ° Wo . 1. 0 1 2 3 0 1 2 3 Weight % Superplasticizer Weight % Superplasticizer Figure 7. Yield Stress as a function of Figure 8, Equilibrium viscosity at 600 s" Type superplasticizer dosage for versus Type "N" superplasticizer paste with 30 volume % replacement concentration for paste with 30 volume % with silica fume. replacement with silica fume. The results for the cellulose powder are presented in Figures 9 and 10. Both the yield stress and viscosity increase with increasing cellulose content. It is interesting to note that only 0.4 weight % of the cellulose powder is required to reach a viscosity of over 700 mPa-s. This is due to the high molecular weight of the cellulose (between 110,000 and 120,000) [26]. Figure 10 combines the data from all of the rheology measurements. At a yield stress of zero, the viscosity of the pastes containing 30 volume % silica fume is lower than that of the mixes containing only superplasticizer. This is due to the increased solids packing density and spherical morphology of the silica fume particles [28-29]. 90Yield Stress (Pa) g g 2 8 8 s 2 8 Equilibrium Visosity at 600 s! (mPa-s) 0.15 0.30 Weight % Cellulose 0! 0.0 0.45 Figure 8. Yield stress versus weight % cellulose powder for cement paste with 30 volume % replacement by silica fume and 2.5 weight % Type "N" superplasticizer. ° 06 ° 02 04 Weight % Cellulose Figure 9. Equilibrium viscosity at 600 s as a function of cellulose powder addition for cement paste containing 30 volume % silica fume and 2.5 wt. % Type "N" superplasticizer. 1000; 7 r 1 r T 30 Vol. % iL Cellulose 800 Silica Fume ae se & = 600| g a 3 400+ 3s - _ oe 200 ne x Plain Cement Paste Vol. % Solids = 0.50 a 1. 1 0 100-200-300, «400 500600 700 Equilibrium Viscosity at 600 s' (mPa-s) Figure 10. Yield stress versus equilibrium viscosity at 600 s* for different mixes.As the yield stress increases, the viscosity also increases, but less dramatically for the pastes containing silica fume and cellulose. This is beneficial for self-compacting concrete applications since a lower viscosity can be obtained in comparison to paste containing only superplasticizer at the same yield stress. Thus, at a given slump the mix has greater fluidity. 4. Conclusions Understanding the influence of various admixtures on the rheology of cement paste is critical to the continued development of self-compacting concrete (SCC). This requires accurate measurements of yield stress and viscosity in addition to understanding the range of shear rates experienced during placement. Extrapolating a flow curve via some type of model fitting procedure provides a useful estimate of yield stress. However, these values are often artificially low at the high solids concentrations typical of most cement pastes. For this reason, a direct measurement of yield stress is preferred, provided the effect of wall slip is eliminated. The viscosity of cement paste is dependent on the shear history of the sample. Tattersall showed that a characteristic equilibrium viscosity is obtained at a specific shear rate after sufficient time [21]. Evaluating the viscosity in this manner is less sensitive to the effects of shear history in comparison to typical hysteresis loop measurements. For this study, the yield stress of cement paste containing two different superplasticizers, silica fume, and cellulose powder were directly measured using a vane. The equilibrium viscosity at a shear rate of 600 s” was also determined for each mix. At the same yield stress, the melamine-based superplasticizer (Type "M") has a lower equilibrium viscosity than the naphthalene superplasticizer (Type "N" yield stress and equilibrium viscosity increase significantly for paste containing 30 volume % silica fume as the superplasticizer dosage is decreased. Yield stress and viscosity increase in a similar manner due to small additions of surface coated cellulose powder. At zero yield stress, replacement of cement by 30 volume % silica fume leads to a 36% reduction in equilibrium viscosity compared to cement paste containing only superplasticizer. This can be attributed to an increase in the solid packing density and spherical morphology of the silica fume particles. At the same yield stress, a lower viscosity is possible through adjustment of superplasticizer dosage or cellulose content. This is beneficial for SCC applications since a higher fluidity can be achieved while still avoiding segregation of the aggregates. 5. Acknowledgements The authors would like to thank The Dow Chemical Company for supplying the cellulose powder and Prof. Wesley Burghardt of the Department of Chemical Engineering for his helpful discussions. Funding for this research was provided by the 92National Science Foundation Center for the Science and Technology of Advanced Cement-Based Materials (ACBM). 6. References 1. Okamura, Hajime, 'Self-Compacting High-Performance Concrete’, Concrete International. July (1997) 50-54 2. Takada, K., Pelova, G. I. and Walraven, J. C., ‘Influence of Mixing Efficiency on the Mixture Proportions of General Purpose Self-Compacting Concrete’, International Symposium on High-Performance and Reactive Powder Concretes (Proceedings). University of Sherbrooke, Vol. 2, (1998) 19-39. 3. Saria, M., Prat E. and Labastire, J-F., ‘High Strength Self-Compacting Concrete: Original Solutions Assoicating Organic and Inorganic Admixtures’, International Symposium on High-Performance and Reactive Powder Concretes (Proceedings), University of Sherbrooke, Vol. 1, (1998) 317-329. 4, Domone, P. L. and Chai, H-W, ‘Design and Testing of Self-Compacting Concrete’, in ‘Production Methods and Workability', Proceedings of the International RILEM Conference, Paisley, Scotland, 1996 (E and FN Spon, London, 1996) 223-236. Se Tattersall, G. H. and Banfill, P. F. G., The Rheology of Fresh Concrete’ (Pitman Publishing, Inc., Boston, 1983). 6. Barnes, H. A., Hutton, J. F. and Walter, K., ‘An Introduction to Rheology’ (Elsevier, Amsterdam, 1989). b Dzuy, Nguyen Quoc and Boger, D. V., ‘Yield Suess Measurements for Concentrated Suspensions’, Journal of Rheology, 27, 4 (1983) 321-349. 8 Keentok, M. "The Measurement of the Yield Stress of Liquids", Rheologica Acta, 21 (1982) 325-332. De Whorlow, R. H. ‘Rheological Techniques’, 2nd edition (Ellis Horwood, New York, 1992). 10. Cross, Malcolm M., ‘Rheology of Non-Newtonian Fludis: A New Flow Equation For Pseudoplastic Systems’, Journal of Colloid Science, 20 (1965) 417-437. 1. Liddell, Peter V. and Boger, David V., Yield Stress Measurements with the Vane’, Journal of Non-Newtonian Fluid Mechanics, 63 (1996) 235-261. 12. Dzuy, Nguyen Q. and D. V. Boger, 'Direct Yield Stress Measurement with the Vane Method’, Journal of Rheology, 29, 3 (1985) 335-347. 13. Barnes, Howard A.,'A Review of the Slip (Wall Depletion) of Polymer Solutions, Emulsions and Particle Suspensions in Viscometers: Its Cause, Character, and Cure’, Journal of Non-Newtonian Fluid Mechanics, 56 (1995) 221-251. 14, Keentok, M., Milthorpe, J. F. and O'Donovan, E., ‘On the Shearing Zone Around Rotating Vanes in Plastic Liquids: Theory and Experiment’, Journal of Non-Newtonian Fluid Mechanics. 17 (1985) 23-25. 9320. 21. 22. 2. 24. 25. 26. 27. 28. 29. Haimoni A. and Hannaant, D. J., ‘Developments in the Shear Vane Test to Measure the Gel Strength of Oilwell Cement Slurry’, Advances in Cement Research, Vol. 1, No. 4, (1988) 221-229. Saak, Aaron W., Jennings, Hamlin M. and Shah, Surendra P. ‘Influence of Wall Slip on Yield Stress and Viscoelastic Measurements of Cement Paste’, Cement and Concrete Research (to be submitted). Roy, D. M. and Asaga, K.., ‘Rheological Properties of Cement Mixes: III. The Effect of Mixing Procedures on Viscometric Properties of Mixes Containing Superplasticizers', Cement and Concrete Research, 9 (1979) 731-739. Yang, M. and Jennings, H. M., ‘Influence of Mixing Methods on the Microstructure and Rheological Behavior of Cement Paste’, Advance Cement- Based Materials, 2 (1995) 70-78. Banfill, P. F. G., ‘Viscometric Study of Cement Pastes Containing Superplasticizers with a Note on Experimental Techniques, Magazine of Concrete Research, 33, 114 (1981) 37-47. Williams, David A., Saak, Aaron W., and Jennings, Hamlin M. ‘The Influence of Mixing on the Rheology of Fresh Cement Paste.' Cement and Concrete Research, accepted (1999). Tattersall, G. H., ‘Structural Breakdown of Cement Pastes at Constant Rate of Shear’, Nature, 175, 4447 (1955) 166. Lapasin, R., Papo, A. and Rajgelj, S., "The Phenomenological Description of the Thixotropic Behavior of Fresh Cement Paste’, Rheological Acta, 22 (1983) 410-416 Hattori, Kenichi and Izumi, Kaichi, ‘A Rheological Expression of Coagulation Rate Theory, Part 2: Combined Effect of Shear Rate and Coagulation Rate on Viscosity’, Journal of Dispersion Science and Engineering, 3, 2 (1982) 147- 167. Reed, James S., ‘Principles of Ceramics Processing’, 2nd Edition.John Wiley and Sons, Inc., New York, 1995). Lapasin, R., Papo, A. and Rajgelj, S., ‘Flow Behabvior of Fresh Cement Pastes. A Comparision of Different Rheological Instruments and Techniques’, Cement and Concrete Research, 13 (1983) 349-356. Methocel Cellulose Ethers: Technical Handbook. (Dow Chemical Company, Midland, Michingan, USA, 1997). ASTM Designation C-305-94, ‘Mechanical Mixing of Hydraulic Cement Pastes and Mortars of Plastic Consistency’, Annual Book of ASTM Standards. 04.01 (American Society of Testing Materials, Easton, Maryland, 1996) 194- 196. Roy, D. M., Scheetz, B. E, and Silsbee, M. R., Processing of Optimized Cements and Concretes Via Particle Packing’, MRS Bulletin, XVIII, No. 3 (1993) 45-49. Nedi, M., Mindess, S. and Aitcin, P-C., ‘Rheology of High-Performance Concrete: Effect of Ultrafine Particles’, Cement and Concrete Research, 28, 5, (1998) 687-697.RHEOLOGY TESTS FOR SELF-COMPACTING CONCRETE - HOW USEFUL ARE THEY FOR THE DESIGN OF CONCRETE MIX FOR FULL SCALE PRODUCTION? Mats Emborg Betongindustri AB, Stockholm and Lulea University of Technology, Sweden Abstract Within the Brite Euram project SCC (Self Compacting Concrete), several tests on rheology with the BML viscometer have been performed. At the same time, workability tests with standard methods such as slump flow and L-box were carried out. Results from the laboratory tests are utilised in pre testing at concrete plant to check whether mixes chosen are optimal at full scale production including the mixing procedure, the transport, the pumping and the casting. In addition to the general theology parameters, other important parameters for the quality of the full scale produced self compacting concrete are the homogeneity of the material in form, surface quality of the structure and the bond between concrete and reinforcement. These parameters are however difficult to examine on a laboratory scale. Concrete mix was varied in the laboratory for example by changing aggregate type; filler type and content; type and dosage of superplasticiser and viscosity agent; cement content and by changing steel fibre content. From the test results, criteria for Self Compacting Concrete can be identified. Furthermore, the correlation between parameters obtained at the rheology and workability test were studied. 1. General When developing mix design for Self Compacting Concrete (SCC) it is obvious that a large number of parameters influence the final properties of the concrete. From the literature, information can be found on how the concrete engineer arrived at or near the definite mix proportions. Several mix design methods have thus been suggested for SCC based on workability and rheology studies giving the fine aggregate volume, the water to powder ratio and the coarse aggregate content, sec ¢ g Okamura and Ozawa (1996) [1] and Sedran et al (1996) [2]. In Sweden a method for mix design has been developed including criteria of blocking, void and paste volume as well as results from 95paste rheology studies, see Pettersson et al (1996) [3] and Billberg (1997) [4]. From such methods, mix design guidelines may be formulated. Furthermore, within the Brite Euram project SCC (see Acknowledgements), besides manuals for manufacture, uansport and pouring, recommendations for mix design are within targeted project outputs. Important at mix design are also parameters which may be more or less difficult to detect in the laboratory such as resistance to the segregation, proneness in the forming of surface pores of the structure and if any additional form pressure is caused by the concrete. These parameters are however not addressed in the mix design methods found in the literature. 2. Workability and rheology studies 2.1 General The Brite Euram project has made it possible to perform basic research on workability and rheology as well as blocking tendencies of concrete mixtures for civil engineering applications and for housing. Thus, reliability and reproducibility have been investigated by various test methods including slump flow, L-box, Orimet, BTRHEOM rheometer and BML viscometer (se below). As known, rheology methods give two parameters to characterise the material; the initial shear value, g, and the viscosity, h. According to many experts, this is the most accurate way to describe the real behaviour of the fresh concrete when it is poured into the form. Furthermore, rheology test methods are less dependent on the experience and the skill of the concrete engineer than workability methods. Compare, for instance, the slump flow test method that, if used by a well experienced technician, can provide information on mobility, stickiness, separation tendency etc. However, often, in practice, only the diameter of concrete after flow is reported from the slump flow test. Rheology methods have the drawback that they are not directly intended for use at the building site which means that they are usually only utilised at the first mix design and sometimes at the pre-testing. Therefore, there is a need to find a correlation between the parameters obtained at the rheology test and the data from the workability tests which are more suitable for continuos use at concrete plant and at site. 2.2 Test equipment Betongindustri AB has together with the laboratory Scancem Research, Slite, performed tests on the rheology and workability of SCC. Equipment used was that at NTNU, Trondheim, a developed BML viscometer type WO-3 which is the latest model of the BML family, Mork (1996) [5] and Wallevik and Gjérv (1990) [6]. WO-3 is a constant rate coaxial cylindrical viscometer, see Fig 1. 96Outer rotating cylinder Inner cylinder with torque Fight cone to _-measuringcell Avis * Rotating plate for outer cylinder Outer cylinder with ribs ‘Sample with possible plug Inner cylinder with ribs Fig |. BML viscometer :ype WO-3. Radius of inner cylinder is 100 mm, the outer 200 mm and the height of the cylinders area 200 mm, from Mork (1996) [5]. The BML WO-3 viscometer has been compared with another type of BML viscometer and with the BTRHEOM apparatus, developed at LCPC, Paris, sce ¢ g Hu ct al (1996) {7}. It was shown that the results from the different types of equipment were more comparable in terms of viscosity, h, than in terms of yield stress, g. The scatter in yield stress may depend on th very small absolute values obtained as such (almost zero) implying more sensitive data. Workability behaviour of the SCC was also recorded with the standard Abrahm’s Cone ie the slump flow and a type of L-box. At the slump flow test the final spread diameter and time to reach a S00 mm diameter, T50, were documented. At the L — box testing, the vertical part of the form is filled with 12.7 litres of concrete. Thereafter a gate is lifted and the concrete will flow into the horizontal part of the form. The times for the concrete to reach 200 mm and 400 mm of the horizontal part, T20 and T40 respectively, can be documented. The main parameter to record is however the blocking after the flow has stopped which is the relation between H1 and H2 that is the concrete heights of the vertical and horizontal parts of the L box respectively. 2.3 Test series In the test series the influences of the following parameters were studied: Aggregate — grading curve, crushed or not crushed Filler - type and amount 97Cement content Water content Superplasticiser - type and dosage Viscosity agent — type and dosage The basic concrete recipe had per cubic meter a water content of 175 1; a cement content of 325 kg; a filler content of 200 kg and a superplasticiser content of about 9.5 kg. The cement was of type Std Portland Slite (Cem I). The filler was of a limestone type (Képing 500). The recipe was designed for application in housing. Rheology and workability tests were performed and separation tendency as well as general performance were noted. The purpose of the test was to discover whether appropriate values of yield stress and viscosity could be found and if there existed any correlation between these target BML-values and any corresponding data from slump flow and L-box testing. Prior to the more time-consuming BML-tests, workability studies were carried out to find possible trends for SCC and to eliminate certain bad concrete recipes. Hence, in total more than 90 mixes were studied during 1998 and the viscometer tests were limited to a number of slightly more than 30. Two BML-tests were performed on each concrete mix; at 5 minutes and at 15 minutes after the mixer had been stopped. Prior to the second test the concrete was remixed for 10 seconds. In most cases the yield stress, g, (and sometimes also the viscosity, h) had slightly increased between the first and second measurement. Results from the rheology tests are here displayed as the g-value as a relation of the h-valuc. 2.4 Results Examples of examinations with various fine aggregate (0 - 8 mm) and coarse aggregate (8 -16mm) are shown in Fig 2. The effect of a change of the fine aggregate is rather clear, compare for instance test no. | with test no. 2 in Fig 2a. Even though a larger dosage of superplasticiser was used in test no. 2c, the g— and h - values are lower than for combination no. 1 with a more suitable fine aggregate. Mix 4 contained crushed aggregate which markedly affected the rheology. An attempt was made to remedy the completely unacceptable behaviour of mix 4 by a larger water content (mix 4b). However, this was not sufficient. Fig 2b shows results from combinations of two types of coarse aggregate with the fine aggregate above. In the mixes, also the amounts of coarse aggregate content were varied between 39% and 45% related to total aggregate content (by weight). It can be seen in the figure that the effect on the rheology is rather small. 98Furthermore, by decreasig the filler content, the main effect is that the g-value will be lower, see Fig 3a, Similar tendencies are observed in Fig 3b when the cement content is decreased, compare for instance mix no. 2c with no. 9b and mix no. 1b with mix no. 5. 18 14 13 12 . a ay td No 4 outside scale —e4 (g=4,51 h=46,42) = og 2 08 g 07 a B06 + B08 _ = 2c. Natural 8 extra sp B oa 2-4. Crushed * 03 food 4b, Crushed, extra water Ee 21. Natural, A a 8-2. Natural, 8 01 0 2 4 6 8 0 2 14 16 18 2 viscosity, h (Nms) a) —— 2c. Natural 6, 42% 15 4 —*—3. Crushed, 42% 13 12 —+— 12. Natural D, 39% Eu Eos —4—19. Natural D, 45% 09 5 08 Booz ++ @ --14b. Natural C, 39% £ 06 @ os —e— 150. Natural C, 45% Bos 2 03 ++ t+ +16. Crushed, 39% 02 * —©—17. Crushed, 45% 0,1 —*—1. Natural D, 42% o 2 4 6€ 8 0 12 “4 6 18 viscosity, h (Nms) b) Fig 2. Yield and viscosity values obtained with the BML OW3-viscometer on mixes containing different types of fine aggregate, natural A and B and crushed (a) and coarse aggregate of different types (C, D and crushed) and contents (b). 99When adding extra amounts of superplasticiser, rheological behaviour changed, mainly by lowering the g — value. How large the effect will be depends on the general mix composition for instance the coarse aggregate and filler contents. In addition it can be mentioned that, as expected, addition of viscosity agent lead to an increase of h — value and as a result a more stable concrete. i 2c. F=200, C=325 ++ © ++7. Fe175, C=325 —*— 9b. F=200, C=300 —te— 106, F=175, C=300 yield stress, g (Nm) —O— 11. F=175, C=325, —*— 1b, F=200, G=325, o 2 4 6 8 10 12 14 16 18 20 viscosity, h (Nms) a) tl — 2c. C=325, F=200 g —®— 5. C=300. F=200 o g ++ te = +B, C2275, F=230 8 < @ —?#— 9%. C=300, F=200 3 = = SN —— 1b, C=325, F=200 wee —t#— 2. C=325, F=200 2 4 6 8 10 12 14 6 18 2 viscosity, h (Nms) b) Fig 3. Yield and viscosity values obtained with the BML OW3-viscometer on mixes with different filler contents, F, (a) and different cement contents, C, (b). 100Fig 4 displays the main results from the study at Betongindustri AB and Scancem Research. Although the figure may be difficult to interpret, it shows typical output from such a study. Several mixes could be excluded as possible SCC due to a too high a separation proneness, a too stiff and sticky behaviour which was not directly observed at the pre-testing using workability test methods. Here, an optimum SCC should have a viscosity of larger than 19 - 12 Nms and a shear strength lower than 0.1 - 0.2 Nm. Other target values may be defined for self compacting concretes of lower quality for less demanding castings. From the figure it can be concluded that not more than 2 or 3 mixes reach the optimum rheology values. 2.5 Correlation with workability tests In the literature, for normal types of concretes, several authors bring out the relationship between slump flow and data obtained at rheology tests. Thus, comparisons have been made in this study between BML-viscometer and slump flow, see Fig 5. It is clear that, although there is some connection between g and slump flow and h and TSO respectively, the correlation is weak. For example, a target slump flow value of 650 mm, often valid at several situations, may according to Fig Sa give yield stresses between 0 and 0.4 Nm, that is markedly outside the target values. Furthermore, the 10 - 12 Nms mentioned as suitable value of h, implies a T50 value larger than about 3.5 s, but as can be seen in Fig 5b itis likely that with a T50 of about 3.5 s also viscosities lower than 8 Nms may cccur. 15 . 14 13 12 1 a =! No 4 outside scale = 09 (g=4,51 h=46,42) > og g o7 Ss = 06 . 2 os Optimum area a % scc 03 ¥ tx 02 on o oA o 2 nr) 1° ® viscosity, h (Nms) Fig 4. Yield and viscosity values obtained with the BML OW3-viscometer. Summary of tests. 101Thus, in this example, if only slump flow value and TSO has been used, several mixes had wrongly been characterised as promising for use as SCC. Similar findings have been obtained in other tests of the Brite Euram project. In addition, the test series has shown that almost no correlation could be found between BML viscometer and the L -box method, see for example Fig 6, i e the blocking tendencies can not be detected during BML tests. Yield stress, Nm 5 580 600 620 G10” 660 680780 01 2) Slumpflow, mm 2 E Zz = 2 8 2 S 0 1 2 3 4 5 6 b) TSO, sec. Fig 5. Correlation between slump initial shear value and slump flow (a) and viscosity and T50 (b) respectively. 10209 208 Sor = go8 Bos 2o4 803 Do2 on 0 On 02 03 os OS 06 O7 shear stress, g Fig 6 Blocking ration from the L-box as a relation of the g— value from the BML. viscometer. 3. Conclusions Most mixes that have been examined in this study had a somewhat too high yield stress, g, and a too low viscosity, h, to be characterised as an excellent SCC. However, the work has been very useful in finding preliminary target values of g and h and in studying trends when concrete composition has been changed. In Fig 7 some possible influences on an original concrete mix when changing the concrete composition are outlined. Similar diagrams can be found for other types of concretes, see the literature. Important drawbacks of BML and other rheology test methods are that they are time consuming and that the 2quipment is rather expensive. Hence, there is a need to transform BML target values and trends into workability test parameters to achieve detecting method for practical use. So far, the correlations between the test methods examined are however rather weak. Concerning the slump flow, it may however be possible to supplement the diameter and T50 values with additional information that can be observed by an experienced concrete engineer. In this way, the correlation may be strengthened. In this way, rheology test methods together with modified workability methods could be very useful for basic design and for further studies and control at plant and at site. 103yield stress, g ‘natural ype A 0-8 replaced ‘with natural 0-8 type B more vise. agent Viscosity, h Fig 7. Possible influences on an original concrete mix when changing the composition. Note that the sketch is not intended to be general. Acknowledgements This study was carried out at Betongindustri AB and is part of the Brite EuRam project “Rational production and improved working environment through using self- compacting concrete” contract no. BRPR-CT96-0366. The partners in the project are: NCC AB (Sweden, Co-ordinator), Betongindustri AB (Sweden), Swedish Cement and Conerete Research Institute (Sweden), Lule University of Technology (Sweden), GTM Construction (France), LCPC (France), University of Paisley (Scotland), SIKA S.A. (Spain) and N.V. Bekaert S.A. (Belgium). References mM) Okamura H, Ozawa K ’Self Compacting High Performance Concrete’, Structural Engineering International, No 4, 1996, pp 269 - 270 (2) Sedran T, de Larrard F, Hourst F, Contamines C ‘Mix design of self- compacting concrete (SCC), Production Methods and Workability of Concrete. (ed PJ M Bartos, D L Marrs, D J Cleland), E & FN Spon, London, 1996, pp 439 - 450 B) Petterson ©, and Billberg P. Van B K 'A model for self compacting concrete’, Production Methods and Workability of Concrete. (ed PJM 104(4) (5] (6 10] Bartos, DL Marrs, D J Cleland), E & F N Spon, London, 1996, pp 483 - 492 Billberg P “The effect of mineral and chemical admixtures on fine mortar rheology, from proceedings of ‘Fifth Canmet Int Conf on Superplasticies and Other Chemical Admixtures in Concrete’, Rome 1997, Mork J H ‘A presentation of the BML viscometer’ Production Methods and Workability of Concrete. (ed PJ M Bartos, D L Marts, D J Cleland), E & F N Spon, London, 1996, pp 369 - 376 Wallevik O H and Gjérv O E ‘Development of a coaxial cylinders viscometer for fresh concrete. Proceedings of the RILEM Collogium on Properties of Fresh Concrete, (editor H J Wierig), Chapman é& Hall (quoted from [5]) Hu C, de Larrard F, Sedran T ‘A New Rheometer for High Performance Concrete. BHP 96 fourth International Symposium on the Utilization of High Strength/High Performance Concrete, Paris 29 -31 May, 1996 105PART 3 PropertiesPROPERTIES OF MORTAR FOR SELF-COMPACTING CONCRETE PL J Domone, J Jin University College London, UK Abstract The effect on the fresh properties of the mortar phase of self-compacting concrete of four different types of superplasticizers and various combinations of powders, including Portland cement, ground granulated blast furnace slag (ggbs), pulverized fuel ash (pfa), microsilica (csf) and limestone powder (Isp) have been measured. The properties include spread, V-funnel flow time and yield stress and plastic viscosity. Initial tests examined the effect of the time of addition of the superplasticizer during the mixing process. In subsequent tests, measurements were continued for two hours after mixing to assess workability retention. Ternary blends of powder have potentially very useful properties. Good correlations between spread and yield stress and between V-funnel flow time and plastic viscosity have been obtained. 1. Introduction To achieve the required fresh properties of deformability, segregation resistance and passing ability (or ‘no-blocking’), self-compacting concrete (SCC) often uses a combination of a greater number of constituent materials than in normal concrete. For example, the paste can contain one or more cement replacement materials, inert fine fillers, superplasticizers, and viscosity agents [1]. The combination of powder materials is also used to control the hardened properties, such as strength [2]. The workability and workability retention properties are of prime importance, and studies have shown that these are influenced by the properties of individual each constituent and the physical and chemical interaction between them [3,4]. This paper reports early results of an ongoing experimental study on mortars which is 109Hawa et al [1], summarising Japanese experience, have suggested that values between 250 and 280 mm for spread, and 6 and 10 secs for V-funnel flow time will produce successful SCC. Chai, who worked with UK materials, suggested a minimum spread values of 300 mm, and a V-funnel flow time between 2 and 10 seconds [7]. The required period of retention of self-compacting properties of a concrete will vary according to the demands of a particular application, but a target value of two hours is a common criterion. Similar conditions will apply to the mortar. 3. Aim and scope of experimental programme. The general aim of the programme is to use tests on mortar to evaluate the effect of a number of factors that have been identified as important in self-compacting concrete, but which have not been systematically examined. This paper reports some results from four series of tests, which have formed the first part of the programme: 1. The optimisation of mixing procedures for four different types of superplasticizers - specifically the time of addition of the superplasticizer during the mixing process. 2. The workability and workability retention properties of Portland cement mixes containing optimum dosage levels of the four superplasticizers. 3. The effect of cement replacement materials (pfa, ggbs, csf) and a fine filler (Isp) in binary combinations with Portland cement on the initial rheology and workability retention properties. 4. The extension of series 3 10 mixes with ternary blends of cement replacement materials and Portland cement. 4. Materials and mix proportions The Portland cement (PC) was a class 42.5N, with C3S, C28, CsA and C4AF contents of 56%, 13%, 8.2% and 11.2% respectively, and a specific area of 360m’/kg. The gebs had surface area of 450m'kg and a specific gravity of 2.9; the pfa had 7.2% of its particles finer than 45m with specific gravity of 2.4; the Isp had 90% of its particles finer than S0um and the specific gravity was 2.68; the csf had a specific gravity of 2.2, and was supplied as a 50% slurry with water. The sand was a river sand with a fineness modulus of 2.6 and specific gravity of 2.64 Four different types of superplasticizers were used, listed in table 1. These are typical of those used in many reported SCC applications. SP3 and SP4 are called ‘new generation’ materials by their suppliers. SP4 is distinct in that its action is a combination of electrostatic dispersion and steric hindrance from the side chains linked to the polymer backbone. The dosage rates for all four admixtures are expressed throughout this paper in terms of a percentage of admixture solids by weight of powder material in the mix. 114Table | Details of admixtures Description Solids ie = content beliead % by wt impel SPi Sulfonated naphthalene 40 K polymer up » SP2 Acrylate-methacrylate 38 dimensions co-polymer uma SP3__ Vinyl copolymer 20 SP4 Modified 20 polycarboxylic ether “2 n Figure 2 The interrupted helical impeller rheometer The proportion of sand in the mortar was kept constant for all tests at 45% by volume, a typical value for SCC. The water/powder ratio was either 0.945 or 1.26 by volume, equivalent to ratios by weight of 0.3 and 0.4 respectively for mixes in which Portland cement is the only powder. Other mix proportions were varied in each of the sets of tests. 5. Test Methods The mortar was mixed in bench mounted Hobart paddle mixer in batches of 1.6 litres. The sand and powders were first placed in the bowl before adding the mix water and admixtures. The sand was in a moist condition before mixing. Allowance was made for the free water content of the sand, the water content of the superplasticizer, and, when appropriate, the csf slurry, when calculating batch weights. The mortar properties were evaluated using three tests. 1. The spread test (figure 1a), described above. 2. The V-funnel test (figure 1b), also described above. 3. A two-point workability test. Determination of the Bingham constants of fluid concrete using a helical impeller a mixing action is a proven and well established technique [8,9]. A scaled down version of this with the dimensions shown in figure 2 was used. The system was calibrated with glycerol (a Newtonian liquid) and welan gum solutions (pseudoplastic) using the method described by Tattersall and 112Banfill [8], so that the values of yield stress and plastic viscosity in fundamental units could be obtained from the flow curves. 6. Test series 1: the effect of mixing procedures Initial tests were carried out to determine the relative performance and the optimum dosages of the four superplasticizers. A mixing procedure was used which was likely to result in the greatest efficiency of action; this is shown in figure 3. For each admixture, the effect of increasing dosage on the spread and V-funnel time for a mortar with a powder of 100% PC and a water/cement ratio of 0.3 was determined. Figure 4 show the results for SP1, which are typical. The spread results increased towards a maximum and then remained approximately constant. The mortar then showed an increased tendency to bleed and segregate. The optimum dosage was estimated to be that just before the maximum as reached, in this case 0.65 %. The V- funnel results were often more erratic than the spread results, and were not therefore used to estimate the optimum dosage. powder + sand + 80% water time (minutes) Figure 3. Mixing sequence for series | tests spread (mm) V-funnel time (secs) 350 45 | A 200 | [ —*= spread | == V-funnel 150 3 04 0.5 0.6 07 08 sp dosage (% solids by wt cement) Figure 4 Typical results of tests for optimum dosage of superplasticizers 113The optimum dosages for the four superplasticizers were found to be: SPI: 0.65%, SP2:0.6%, SP3: 0.3%, SP4: 0.15% The greater efficiency the two ‘new-generation’ superplasticizers is apparent. The mixing procedure test examined the effect of the time of addition of the superplasticizer, which was varied from direct addition at the start of the mixing to a delayed addition of up to 6 minutes. As before, with delayed additions 80% of the mix water was added at the start of mixing, and the remaining 20% added with the superplasticizer. Mixing was continued for at least 5 minutes after adding the superplasticizer. The results are shown in figure 5. Delayed addition increased the effectiveness of all the admixtures as measured by both the spread and V-funnel results. There is a addition time ‘window’ for each admixture where maximum efficiency is obtained; for SP1, SP2 and SP3 this is between about 2 and 4 minutes, and for SP4 between 0.5 and 2 minutes. This general effect is well known and has been reported for naphthalene and melamine based admixtures [10,11] but it has not previously been reported for admixtures of types SP3 and SP4. It has been suggested that the optimum time for the addition of the superplasticizers depends on the composition of cement, the temperature and the mixing procedure [10]. In particular, the change of viscosity with delayed dosage of the superplasticizer decreased with increasing C;A content of the cement [11], which suggests that the addition time ‘window’ may vary for different types or batches of cement. It has also been reported that the effectiveness of some superplasticizers based on acrylic polymers does not depend on their addition time [12]. vw aes $s spread (mm) — as ses V-funnel time (secs) em x f | \ \ \ , 3 3 t+ + + o 123 4 5 6 o 123 4 5 6 addition time of sp (mins) addition time of sp (mins) Figure 5 The effect of addition time of superplasticizersClearly this general effect has implications for both control of laboratory experiments and batching practice. All of the tests reported in the remainder of this paper were carried out with a delayed addition time of 2 minutes for SP1, SP2 and SP3, and 1 minute for SP4, at which times all the admixtures achieve maximum efficiency. 7. Test series 2: Comparison of superplasticizers in Portland cement mixes Tests were carried out on Portland cement mixes with a water/cement ratio of 0.945 by volume (0.3 by weight), with each superplasticizer added at its optimum dosage as given above. The results are shown in figure 6. The initial spread valucs are similar for all four admixtures, but the V-funnel times for SP3 and SP4 are distinctly higher than those for SP1 and SP2. This indicates that the tests are measuring two different properties; the relation between the two tests and the Bingham constants is discussed later. The rate of workability loss also varied, with SP4 having the lowest value; this again therefore proved to be the most effective superplasticizer, and it was used for the next test series. 8. Test series 3: Mixes containing binary blends of powders The mixes tested and their initial spread and V-funnel flow times are given in table 2. The SP4 dosage rate was adjusted so that a spread value between 310 and 330 mm was obtained for each mix. The pfa, ggbs and Isp mixes all required a lower dosage than the 100% pe mix, and the csf mix a much higher dose. Significantly different V-funnel flow times were obtained - the ggbs mix had the highest time, followed by mixes with pfa, Isp and csf. gis ° E10 x 3 as z E04 0 30 6 9 120 time after mixing (mins) time after mixing (mins) Figure 6 Comparison of superplasticizer performance in Portland cement mixes 115Table 2 Mix proportions and properties for series 3 tests mix powder composition (% by vol.) SP4 dosage spread V-funnel Cary wt (mm) time nope pia ggbs Isp csf powder) (secs) 1 100 0.15 310 3.9 2 60 40 0.125 325 31 3 60 40 01 325 3.8 4 60 40 0.09 325 2.5 5 85 15 0.35 320 2.2 In all mixes, fine agg. = 45% of mortar by volume, water/powder = 0.945 by volume. The workability retention results are shown in figure 7. The ggbs mix did not perform as well as the 100% Portland mix. The Isp mix showed an inferior performance based on spread results, but an improvement with V-funnel time. The pfa mix showed a greater retention than the control in both tests, whereas the csf mix had the highest retention rate of all the mixes, with little change in properties over the two hour period. The general trend is for workability retention to be improved with increased superplasticizer dosage. Similar trends were found by Punnki et.al.[3] in tests on concrete mixes containing a naphthalene based superplasticizer. Bonen and Sarkar [4], also using a naphthalene based superplasticizer, found that the workability loss increased with the ionic strength of the pore water, and that in mixes containing microsilica, the amount of superplasticizer adsorption is increased, reducing the ionic strength. The low workability loss of the microsilica mixes in figure 7 shows that similar arguments may apply to the polycarboxylic based admixture SP4. Clearly, further investigation of the mechanism of workability loss for mixes with ‘new generation’ of superplasticizer will be useful 350 g 300 2 a 5 i e E 250 2 a0 gy a & 3 200 2 & x 8 150 : gE? a & 100 +++ > 0 0 30 6 9 120 time after mixing (mins) time after mixing (mins) Figure 7 Results of tests on binary mixes (test series 3) 116It is therefore apparent that the workability retention of mortars for SCC is dependent on a combination of factors including the powder composition and the type and dosage of superplasticizer. Mixes with ggbs or Isp do not necessarily have a higher workability retention, and microsilica mixes can have high workability retention. This indicates that mixes with ternary blends of powders may provide a beneficial combination of properties, and this was therefore investigated in the next test series. 9. Test series 4: Behaviour of mixes with ternary blends of powders The mixes tested and their initial properties are given in table 3. As in test series 3, a single admixture, SP4, was used at a dosage to achieve a spread of 310-330 mm. All the mixes contained csf, and required a higher amount of superplasticizer than the 100% Portland cement mixes, but not as high as in the binary csf/Portland cement mix (table 2). The V-funnel times were lower than for the binary mixes, and again a variation was obtained, but the range was not as great as in the binary mixes. Table 3 Mix proportions and properties for series 4 mix powder composition (% by vol.) SP4 dosage read V-funnel Gyn ® time no. pce pfa_ ggbs Isp —osf. powder) (mm) (sees) 1 50 40 10 0.2 325 15 2 50 40 10 0.17 322 16 3 50 40 10 0.19 325 21 4 45 40 15 0.25 325 18 5 30 60 10 0.16 315 2.2 In all mixes, fine agg. = 45% of mortar by volume, water/powder = 0.945 by volume. 20 =p 220 3 —e— pelesf a = 300 gis {| 7M -pelestests | € —a- F £ 250 3 petee / zg = 10 s 200 a e gee a N s ct * 150 N Bote *® 2 a & serra 100 + 4 > 0+ —+ 0 30 60 90 120 0 30 60 90 120 time (mins) time (mins) Figure 8 Results for a test on a gebs/csf ternary mix 117The workability retention results for mix 4 are shown in figure 8 together with those for the relevant binary mixes. It is clear that the ternary mix retains the excellent retention properties of the binary csf/pe mix without the excessively high superplasticizer dosage, which is therefore a potential advantage of ternary mixes. A number of such mixes have been used in applications of SCC, but there is scope for work to define optimum blends for the combination of fresh, early age and hardened properties required. As mentioned above, the results from the spread and V-funnel tests often show distinctly different patterns of behaviour. In series 3 and 4, a two-point workability test with the helical impeller rheometer was also carried out at each test point, Figure 9 shows the correlations of yield stress and plastic viscosity with the spread and V-funnel results shown in figures 7 and 8. There are good correlations between spread and yield stress (figure 9a) and between V-funnel flow time and plastic viscosity (figure 9d), but the two other correlations are poor. Empirical and analytical studies on concrete have shown good correlations between slump and yield stress [¢.g.13-16], and so the good correlation in figure 9a would be expected since both slump and spread are low shear rate tests. However, the range of variables tested are limited, and the correlations may not be as good with a wider range. Nevertheless, it is clear that the two tests are measuring two different propertics which are broadly equivalent to yield stress and plastic viscosity. 150 = 150 é “. @ é | . ° g 100 e % 100 b « % 59 fo B so o ¢ 3 st 3 * © Bol . , Mm, F 0 150 200 250 300 350 0 5 10 15 2 30 S30, é . © & | . Zw o 4, B20 we g ° g § 10+ + ws 210 | ”” = % + @ Zo got & 150 200 250 300 350 8 0 5 10 15 spread (mm) V-funnel flow time (secs) Figure 9 Correlations of Bingham constants and single point test results 118image not availableBonen, D. and Sarkar, S.L., ‘The superplasticizer adsorption capacity of cement pastes, pore solution composition and parameters affecting flow loss’ Cem. and Conc. Res., 25, 7, 1995, 1423-1434 Okamura, H. and Ozawa, K., ‘Mix design for self-compacting concrete’, Conc. Lib. of Japan Soc. of Civ. Eng, 6, 1995, 107-120 : Fujiwara, H., Nagataki, S., Otsuki, N. and Endo, E., ‘Study on reducing unit powder content of high-fluidity concrete by controlling powder particle size distribution’ Conc. Lib. of Japan Soc. of Civ. Eng., 28, December 1996, 117-128 Chai, H-W., ‘Design and testing of sclf-compacting concrete’ Ph.D thesis, University of London, April 1998. Tattersall, G. H. and Banfill, P. F. G. ‘The rheology of fresh concrete’ Pitman, London, 1983 Tattersall, G. H. ‘Workability and quality control of concrete’ E&FN Spon, London, 1991 . Chiocchio, G. and Paolini, A. E., ‘Optimum time for adding superplasticizers to Portland cement pastes’ Cem. and Conc. Res., 15, 5, 1985, 901-903 . Mannonen, R. and Penttala, V. ‘The efficiency of SNF-type superplasticizers in Portland cement pastes’ Proceedings of the RILEM symposium on Production Methods and Workability of Concrete, Paisley, June 1996, E&FN Spon, London, 1996, 327-341 Collepardi, M. ‘Admixtures used to enhance placing charateristics of concrete’ Cem. and Conc. Composites 20, 1998, 103-112 . HuC., de Larrard, F, and Sedran,T. ‘A new rheometer for high performance concrete’ Proc. of 4" Int. Symp. on Utilization of High strength/High-performance Concrete, Paris, May 1996, LCPC,1996, 179-186 . Domone, P. L., Xu, Y. and Banfill, P. F. G. Mag. of Conc. Res., in press Tanigawa, Y., Mori, H. and Watanabe, K. ‘Computer simulation of consistency and rheology tests of fresh concrete by viscoplastic finite element method’ Proc of RILEM Symp on Properties of Fresh Concrete, Hanover, 1990, Chapman & Hall, London 1990, 301-308 . Jurokawa, Y., Tanigawa, Y., Mori, H. and Komura, R. ‘A study on the slump test and slump-flow test of fresh concrete’ Transactions of the Japan Concrete Institute, 16, 1994, 25-32 120A SIMPLE EVALUATION METHOD FOR INTERACTION BETWEEN COARSE AGGREGATE AND MORTAR PARTICLES IN SELF-COMPACTING CONCRETE Masahiro OUCHI Kochi University of Technology, Kochi, JAPAN Yoshinobu EDAMATSU Sumitomo Osaka Cement Co., Ltd., Osaka, JAPAN Abstract The aim of this paper is to establish a simple evaluation method for the interaction between coarse aggregate and mortar’s particles in self-compacting concrete, one of the major factor in shear deformability of fresh mortar in concrete. A simple evaluation method for the interaction, which largely depends on the physical characteristics of mortar’s particles, is indispensable since mix-design method for self-compacting concrete should be more applicable to materials with wider variation of properties. The authors newly developed a simple evaluation method for the interaction by adopting mortar funnel test, in which funnel speeds of mortar with and without model coarse aggregate are compared. It was found out that the ratio of funnel speed of mortar with model coarse aggregate to that without coarse aggregate was influenced mostly by the physical property of particles in mortar and independent on the flowability or viscosity of the mortar itself. The ratio was proposed as the index for the interaction. 1. Introduction 1.1 Limit of conventional mortar tests for self-compactabiliy Self-compactability test of fresh concrete is indispensable to mix-proportioning since self-compactability largely depends on the characteristics of materials in use. Okamura et. al proposed mortar flow and funnel tests for mix-proportioning in order to rationalize mix-proportioning, in which proper water-powder ratio and superplasticizer dosage can be obtained by estimating from relative flow area Jim and relative funnel speed Rm (Figs. 1 and 2) [1]. 121image not availablemortar’s particles. At this stage, self-compactability test for concrete is indispensable to examining the sand content in mortar when new type of sand or powder material is to be used in self-compacting concrete. However, the sand content in mortar should be evaluated with simple and easy test since materials with wider variation of properties should be used in concrete in the future and then the mix-proportioning for self-compacting concrete should be simplified. That is because simple evaluation method for the interaction between coarse aggregate and moriar’s particles was developed in this research. coarse aggregate mortar 1 difference due to ‘material or sand content in concrete mortar only 0 o Fig. 3 Increase in mortar’s shear deformability resistance t due to normal stress ogenerated when concrete deforms 2. Development of Simple Evaluation Method 2.1 Extent of this research The evaluation method developed in this paper is to be used for evaluating the sand content in mortar from the viewpoint of achieving moderate interaction between coarse aggregate and mortar’s particles on condition that the sand content in mortar has been decided from the material properties (Fig. 4). A rational method for estimating proper sand content in mortar from each material property should be established in the long run, 123and the evaluation method established in this 2.2 Outline of simple evaluation method Ozawa and Nagamoto investigated on the influence of sand content in mortar on the interaction between coarse aggregate and mortar's particles by comparing concrete and mortar test results (Fig.s 5 and 6) [2]. It was shown that higher flow area or funnel speed of mortar is required for concrete in case that the sand content in mortar is higher. It was also found out that the degree of the interaction between coarse aggregate and mortar’s particles can be evaluated quantitatively by comparing the test results between concrete and mortar, especially by using funnel test results since the interaction is more remarkable due to the existence of tapered section during flowing. conrete's flow area Ie 2 mortar’s relative flow area Jim 4 6 8 6 Fig. 5 Relationship between mortar’s and concrete’s flow areas Te = ((slump flow (cm))*—400)400 research is the first step. material property decide sand content in mortar Thoderate interaction ith coarse aggregate Fig. 4 Place of the evaluation for the interaction in the decision process of sand content in mortar et « Bos —— 2 sand content in ses 2 06) mortar : 40% eee 3 ee £ “. 2°? 49% E 2 504 poz 8 0 5 0 0s 1 15 8 mortar’s relative funnel speed Rm Fig. 6 Relationship between mortar’s and concrete’s funnel speed Re= 10/funnel time (sec) The property of coarse aggregate and mortar’s particles should be evaluated independent of each other since mix-proportioning method for self-compacting concrete by using each 124material property is indispensable, and then the property of coarse aggregate used for evaluating the property of mortar’s particles should be common. The authors adopted glass beads as the model coarse aggregate in evaluating mortar’s particles and then the evaluation is to be carried out by comparing the funnel speed of mortar with glass beads to that without glass beads. Larger interaction can result in larger decrease of funnel speed due to the existence of glass beads in mortar, which leads to larger decrease in shear deformability of mortar when concrete deforms (Fig. 7). Although this method is a kind of concrete test with model coarse aggregate, the labor is almost the same as mortar funnel test, which is effective in saving labor for experiments. > i degree of interaction | Fig. 7 Concept of the evaluation method for the interaction between coarse aggregate and mortar’s particles established in this research (The interaction is to be evaluated quantitatively by comparing the funnel speeds.) (funnel speed of mortar with oO _ coarse aggregate) Rm (funnel speed of mortar) 3. Detail of Evaluation Method 3.1 Conventional mortar funnel as testing apparatus The conventional mortar funnel for self-compacting concrete was adopted for evaluating the interaction so that water-powder ratio or superplasticizer dosage can be evaluated at the same time in mix-proportioning, which may result in saving labor for experiments Fig. 2). 3.2 Ratio of funnel speed of mortar to concrete as index for interaction The interaction should be evaluated independent of mortar’s deformability or viscosity, since it docs not need adjusting water-powder ratio or superplasticizer dosage when evaluating the interaction. It was found out that the ratio of mortar funnel speed to that 125of concrete can be almost constant independent of the mortar’s deformability or viscosity itself on condition that the effect of coarse aggregate is the same (Fig. 6). The authors defined the ratio of the mortar funnel speed with glass beads (Rms) to that without glass beads (Rm) as the index for the interaction between coarse aggregate and mortar particles, which is independent of mortar’s deformability or viscosity itself. The index is described as Rmb/Rm. Larger Rmb/Rm indicates less interaction between coarse aggregate and mortar’s particles, which results in less decrease in the shear deformability of mortar when concrete deforms. 3.3 Glass beads of 10 mm as model coarse aggregate The diameter of the glass beads in use was decided to be 10 mm. That is because the diameter in practical coarse aggregate in self-compacting concrete ranges 5 to 20 mm, and because the thickness of mortar funnel in use is 30 mm and then the diameter of the model coarse aggregate has to be less than 15 mm for fear of blocking in the funnel. 3.4 20% of glass beads content The content of glass beads should be moderate so that the blocking might not be occur during flowing and that the difference of the characteristics of particles in mortar can be clearly indicated in the funnel speed. The content of glass beads with the diameter of 10 mm was decided to be 20% by volume in total volume of mortar and model coarse aggregate since in case of 30% of glass beads content blocking occurred everytime (Fig. 8). 1 08 5 206 : S04 02 re re beads content by volume (%) Fig. 8 Relationship between glass bead content and Rmb/Rm OPC+RS, In = 3 to 5, Rm=1.4, sand content: 40% by volume 3.5 Range of mortar’s deformability and viscosity applicable to Rms/Rm The value of Rmb/Rm was obtained with variation of Zim or Rm of mortar itself so that the extent in which Rmb/Rm can be considered to be constant independent of mortar’s deformability or viscosity itself (Fig.s 9 and 10). Judging from the experimental results, the area in which Rmb/Rm can be considered to be constant is Tin of 3 to 7 and Rm of 1 to 2, which is almost the same as the range of Zim and Rm for achieving self-compactability of fresh concrete. The lower limit for Im or Rm was decided so that larger decrease of 126Rmb due to the addition of model coarse aggregate might be inhibited. The upper limit for Im was decided so that sinking glass beads into the mortar might be prevented. The upper limit for Rm was decided so that the inaccuracy of measuring funnel time can be inhibited due to too high funnel speed. Fig. 9 Relationship between Rmb/Rm and Im & Rm OPC+RS, sand content: 40% Fig. 10 Relationship between Rm and Rmb OPC+RS, sand content: 40% Rm 4. Verification of the Evaluation Method by Experiments 4.1 Mortar test Mortar’s Reb/ Rm was obtained from experiments for wide variation of powder materials (OPC: ordinary portland cement and FA: fly ash) or sand (RS: river sand, LS: land sand and CS: crushed sand) and sand content in mortar (Fig. 11). It was verified that the variation of sand content in mortar corresponds to the variation of Rmb/Rm, obtained by comparing funnel speeds, and that the physical property of materials in use also corresponds to Rmb/Rm, which had been able to be estimated qualitatively as yet. It is possible that the degree of the influence of mortar’s particles or sand content on the 127interaction can be evaluated quantitatively. 08 Fig. 11 Relationship between sand content in mortar and E07 ® © Rmb/Rm depending on the § property of sand or powder 5S os materials in use @Mopc+cs, Q@FA+CS, @FA +RS, @OPC+RS, @OPC+LS 0.2 03 04 05 06 Sand Content in Mortar 4.2 Verification of the evaluation method by concrete test The relationship between Rmb/Rm and self-compactability of fresh concrete was investigated by experiments in order to verify the testing method for the interaction mentioned above (Fig. 12). Self-compactability of fresh concrete was tested by box- shaped test, one of the recommended standard test for self-compactability in Japan (Fig. 13) [3]. The filling height of the box-shaped test shown in Fig. 12 is the maximum value obtained by adjusting water-powder ratio and superplasticizer dosage so that only the degree of the interaction can be indicated by the filling height. It was found out that the relationship between Rmb/Rm and the filling height is unique on condition that the effect of coarse aggregate on the interaction is constant. It is possible that the evaluation method “Rm/Rm” can be substitution for self-compactability test of fresh concrete for examining the sand content in mortar. 300] E & £200 o . Fig. 12 Relationship between » ° coors Rmb/Rm and degree of self- z 100) * OPCHLS compactability of concrete * BASES indicated by filling height of + FACS 0 box-shaped test 05 0.6 07 08 128open center gate 680 mm concrete obstacle 200 mm. D10 spacing 35mm Fig. 13 Box-shaped test 5. Conclusions 1. The testing method for the interaction between coarse aggregate and mortar’s particles in self-compacting concrete was proposed in which funnel speed of mortar with coarse aggregate and that without coarse aggregate are compared. 2. The conventional funnel for mortar in self-compacting concrete was adopted as the testing apparatus. Glass beads with the diameter of 10 mm was adopted as the model coarse aggregate and the content in total volume of mortar and model coarse aggregate was decided to be 20% so that blocking might not occur. 3. The index for the interaction was proposed as the ratio of the funnel speed of mortar with glass beads (Rms) to that without glass beads (Rm). It was found out that the index Rmb/Rm can be constant on condition that the deformability or viscosity of mortar itself is in the range for achieving self-compactability of fresh concrete. 4. It is possible that the evaluation method developed in this research can be the substitution for the self-compactability test of fresh concrete for examining the sand content in mortar. 6. References 1. Nawa, T., Izumi, T., and Edamatsu, Y., “State-of-the-art report on materials and design of self-compacting concrete,” Proceedings of the International Workshop on Self- 129Compacting Concrete, Kochi, Japan, August, 1998 (Kochi, Japan, 1999), 160-190. 2. Nagamoto, N. and Ozawa, K., “Mixture proportions of self-compacting high- performance concrete,” High-performance Concrete: Design and Materials and Advances in Concrete Technology, ACI SP-172 (1997), 623-636. 3. Japan Society Civil Engineers (1998), “Recommendations for construction of self- compacting concrete.” 7. Appendix 7.1 Materials Powder materials [specific graviy | fmeness___] Ordinary portland cement (OPC) 3.13 Fly ash (FA) solid volume River (RS) Land (LS) Crushed (CS) Coarse aggregate: crushed stone, specific gravity 2.7, solid volume 58.6%, F.M. 6.87, maximum size 20 mm Model coarse aggregate: glass beads, specific gravity 2.54, diameter 10 mm Superplasticizer: poly-ether-based type 7.2 Mixing and testing of mortar mixer: mortar mixer meeting Japanese Industrial Standard (JIS R 5201) mixing volume: 1.5 litter per batch mixing procedure POWDER + SAND | — mixing 30 sec — + [WATER (1) ] — mixing 120 sec — + WATER (2) + SUPERPLASTICIZER | — mixing 120 sec — end testing procedure for the interaction MORTAR | — flow & funnel tests + + [ GLASS BEADS ] — stir by hand power — [ MORTAR WITH GLASS BEADS | — funnel test — end 130COMPARATIVE STUDY ON THE INDUCED HYDRATION, DRYING AND DEFORMATIONS OF SELF-COMPACTING AND ORDINARY MORTARS G. VILLAIN, V. BAROGHEL-BOUNY and C, KOUNKOU Laboratoire des Ponts et Chaussées, Paris, France Abstract In order to facilitate casting, self-compacting concrete mixtures contain, besides superplasticizer, an important amount of water which is not completely involved in the chemical reactions and which can evaporate. Early age and long term drying can lead to deformations and to chemical changes, which can influence the durability of the structure. The aim of this paper is to perform a comparative study between a self- compacting mortar containing fly ash and an ordinary one, about hydration, drying and induced strains phenomena. Hydration and drying processes, their interactions and their consequences on pore structure and strains are registered. The long term evolution of different parameters (like degree of hydration, internal relative humidity, water content gradient, porosity, shrinkage...) is followed. The different properties have been studied as a function of the distance to the drying surface, since the surface area is particularly porous and submitted to aggressive agents. The results show that the self-compacting mortar presents no observable segregation and a good homogeneity. The bulk shrinkage of this mortar is smaller than the shrinkage of the ordinary one. Nevertheless increasing porosity and shrinkage near the surface are observed and can therefore be attributed to early age drying, which could jeopardise the durability of the materials. 1. Introduction Sclf-compacting coneretes are becoming increasingly used in the building industry as a result of their extreme ease of application without vibration (flooring slabs, base plates, etc.) and within confined spaces such as between reinforcement embedding (posts, lintels, etc.) [1,2,3]. Moreover, as is the case with all cement-based materials, their mix design calls for a certain quantity of water, a portion of which gets used for the hydration of the cement and the other for the workability of the fresh concrete, before evaporating. Concretes in general, and especially the newer self-compacting concrete 131materials, require an in-depth understanding of the hydric release mechanism through either hydration or evaporation and the associated deformations. However, drying occurs simultaneously with the chemical reactions, thereby leading to a coupling between drying and chemical reactions during the evolution of the microstructure and deformations of the materials [4,9]. ‘The objective of this research, which has focused on both a self-compacting mortar and an ordinary mortar, is to study the interactions between hydration, drying and the ensuing deformations. The measurements taken as part of this study will serve to evaluate the water content gradients, the degree of hydration and the deformations. Subsequently, the influence of these processes on both the pore structure and the material deformations, by means of a comparative analysis conducted on the entire set of results, will be highlighted. 2. The experimental program 2.1 Materials Two types of materials were studied as part of this research program. The first is a fluid mortar whose mix design includes water (approximately 270 V/m3), cement, siliceous limestone sand, a rather high proportion of fly ash and a complex system of admixtures with a superplasticizer. The water-to-binder (cement plus fly ash) ratio is high. Its 28-day compressive strength is on the order of 30 MPa. E Ctev The second material, used as reference of this study, is an ordinary mortar composed of water, siliceous limestone sand and a cement which, while different from that found in the fluid mortar, remains within the same strength category and exhibits nearly the same silicate and aluminate concentrations. This mortar contains slightly less water (approximately 250 I/m3), yet must be vibrated during casting. Its 28-day compressive strength is equal to 70 MPa. = 0,45 EL oas c The role of fly ash is first to facilitate casting thanks to their spherical shape and their lower power of water retention than cement [5]. Second, at long term, the role of fly ash is to contribute to increase the material strength. The hydration reactions begin as soon as the cement comes into contact with the water and slow down considerably beyond 7 days. In contrast, the pozzolanic reactions between the fly ash and the portlandite produced by the cement hydration start up later and continue for a much longer period. Given the large quantity of ash introduced into the fluid mortar, these pozzolanic reactions will prove to be critical. Throughout this article, the term "chemical reactions" is meant to encompass both hydraulic and chemical reactions [6]. 1322.2 Testing set-ups Measurements have been carried out on cylindrical and small-sized test specimens : 16cm in diameter and 5cm in height. These dimensions were chosen in order to correspond with the likely thickness of a building flooring slab. Two test set-ups were ultimately selected : © Set-up A: The samples are placed in an enclosed chamber at 95% relative humidity immediately upon casting. After one day, they are demolded and waterproofed. Over the ensuing six-month period, they are used for examining the autogenous chemical reactions, i.e. those reactions not displaying any moisture exchange with the outside. At the age of 6 months, with the kinetics of the reactions slowing down considerably, the upper face is uncovered to enable studying the drying effects on their own. © Set-up B : Following casting, the specimens are stored in a climate-controlled room at 50% relative humidity. After one day, they are demolded and waterproofed; the upper face however is left exposed. In this manner, the samples dry through the upper face, as is the case with all flooring slabs (on work sites). These samples are then monitored during a 12-month period in order to examine the interactions between chemical reactions and drying. To measure the amount of shrinkage by use of a retracto-meter, a series of 4x4x16 cm prisms was produced. Once demolded, three are sealed to measure the autogenous shrinkage (set-up A) during six months and then opened to measure the drying shrinkage. The three remaining prisms are submitted to the environmental conditions to measure the total shrinkage, i.c. shrinkage duc to the coupling of chemical reactions and drying (a set-up similar to B above, yet with all of the faces exposed) [7,10]. After demolding, all of the samples are kept in a climate-controlled room at a temperature of 20°+1°C and a relative humidity of 65%+3%. 2.3 The testing program The tests selected for the experimental program must enable monitoring the drying through the relative mass loss obtained both by weighing and by gamma-ray measurements (gammadensimetry). These measurements make it possible to trace the profiles of the relative density variation versus height over time as well as to assess the evolution in the water content profiles. The temperature and internal relative humidity are measured by miniature thermohygrometric probes. The degree of concrete hydration is determined by heat loss, and the evolution in the pore structure by mercury intrusion porometry, which yields the overall porosity in addition to the pore volume distribution as a function of their radius. Moreover, the samples deformations are measured on 4x4x16 cm prisms by use of a retractometer. These tests have also been intended to characterize the evolution of the microstructure with respect to both the test set-up and the distance from the sample bottom, wherein lies the justification for choosing two measurement heights (see Figure 1): at 4cm from the bottom (high), and at | cm from the bottom (low). 133Most tests are conducted at 7, 28, 90, 180 and 360 days. Figure | : Scheme of positions on sample B Gammadensimetry B Drying surface 50 mm ; 4om high 40 mm Lem low 10mm O.mm : bottom Bottom 3. Experimental results and analyses In this article, the results from the first nine months of testing on the fluid mortar are presented ; this testing regime corresponds to six months of autogenous chemical reactions for set-up A specimens, followed by three months of drying. In contrast, the results on the normal mortar only reflect 60 days of testing, which implies that the set-up A samples enable studying just the hydration of the cement. 3.1 Segregation and bleeding No segregation of aggregate was observed when examining the microstructure as a function of height for the self-compacting mortar, as opposed to the ordinary mortar (test set-up A). In Figure 2, the density profile at two days shows that the fluid mortar density is homogeneous with respect to height, whereas the normal mortar density diminishes as the height increases near the drying surface. It should be noted that the Figure 2 : Gamma-ray measurement at the age of 2 days (time reference t) 2300 =O -normalA ——g— fluid A | 250 = ae -normalB fluid B ° 2 = 2200 é 3 2150 2 S 2100 4 2050 0 10 2» 30 40 50 Height (mm) 134indicative values of density must be analyzed from a comparative standpoint as a function of height and not in absolute terms. The minor segregation of the normal mortar is due to the vibration during casting. The fluid mortar exhibits normal bleeding, given the large quantity of water in its mix design. The results from weighing the materials during the first 24-hours period allow quantifying the water loss in the oven with a 95% level of humidity (set-up A) : # The fluid mortar loses 1.5% of its total mass. The ordinary mortar loses 0.25% of its total mass. In addition, since their casting, the non-sealed samples, submitted to a relative humidity of 50%, lose an average of 2.5% of their water content. This loss causes the water-to- binder ratio to drop (to an average of 0.36). This sizable water loss undoubtedly occurs prior to setting since it does not appear to disturb the specimens homogeneity in terms of density (see Figure 2). 3.2 Drying during and after hydration The monitoring of the samples mass, such as relative density measurements by means of gamma-ray measurement (see Figure 3), indicates that the sealed specimens (set-up A) lose very little water, with a relative mass variation of less than 0.2% for both of the studied materials. Set-up A thereby enables studying the evolution in water content due solely to chemical reactions over the first six months. Hence, the gradual decrease in internal humidity corresponds to the self-desiccation process, i.e. the consumption of water for the chemical reactions. The internal humidity of the fluid mortar reaches 93% at six months, whereas that of the reference mortar, which rises to 94% after just one day, measures 93% at 28 days (see Figure 4). Figure 3a and 3b : Relative density loss for fluid and normal mortars Fluid mortar ‘Normal mortar A B ° sam —e—19 mn —a—25 mm 2 2-38 mm 45 mm Density difference (%) a Density difference (%) & 0 30 60 90 120 150 180) 0 14 28 42 56 ‘Age (days) Age (days) 135Figure 4a and 4b : Following of internal relative humidity for fluid and norma! mortars Fluid mortar Normal mortar 100 100 => 2 \e => = 90 = 90 z zB 3 3S aA = 80 E 80 —e—lowB 2 2 2 2 —a- high B 2 70 3 70 2 ee 60 + 60 ff 0 30 60 90 120 150 180) 0 14 28 42 56 ‘Age (days) Age (days) In Figures 3 and 4, for type B samples, the drop in water content displays slower kinetics during the first 28 days for the reference mortar than for the fluid mortar due to the normal mortar's lower porosity (see Section 3.3). Figure 5 : Internal relative humidity comparison of drying sample A and of both drying and hydrating sample B, for fluid mortar Dryingtime from 6 months for A (days) —e—lowB 182 196 210 224 238 | —o— high B 100 —e—lowA _ 54 a high A 2 4 = 90 ° * 3 85 ° = 2 sol a 2 2 75 t 70 4 Drying beginning at 184 days 65 | | fa . “ 0 14 28 42 56 70 ‘Age (days) At an age of six months, the chemical reactions in the self-compacting mortar are considerably slowed and thus considered as just about complete. The kinetics of the decrease of the internal relative humidity of the specimen under drying conditions (set- 136up A) are slower than those of the sample simultancously subjected to drying and chemical reactions (set-up B) (see Figure 5). The two materials differ in both their age and their microstructure ; the autogenous reactions serve to increase the density of the A-sample microstructure during six months. In Figure 6a, the relatively homogeneous reduction in density differences, and hence in water content, over the entire specimen height is characteristic of chemical reactions. The parallel curves which converge on a limit value (between 28 and 63 days) near the drying surface correspoid to a coupling of chemical reactions and drying. It can be assumed that this limit value reflects the equilibrium water content at a given external relative humidity. However, in Figure 6b, the phenomenon of a water content decreasing at the surface but not et the bottom of the sample is characteristic of drying effects alone : the slope of the curves is reversed [8]. Figures 6a and 6b : Relative density loss comparison of drying sample A and of both drying and hydrating sarnple B, for fluid mortar Hydration and dry ing (B) Drying from 6 months (A) 0 0 from time t, to f.1 Faas 1 é a M4 cays g ie -2 —o— 28 cays 2 “1 S S is a GB cys |S \ 2-3 Nad ~~ 180 days|'@ la \o. a 4 2 0 10 2 30 40 50 0 10 2 30 40 50 Height (mm) Height (mm) 3.3 Impacts on both the pore structure and deformations This drying phenomenon exerts a disruptive influence on the evolution of the fluid mortar pore structure, as revealed by the test of mercury intrusion porosimetry vs. age. First of all, the microstructure of the mortar protected from drying tends to evolve in a more classical pattern (see Figure 7a): from an initial porous mode (peak reached at 7 days between 1000 anc 1500 A), which characterizes the space initially occupied by the liquid phase in between the anhydrous cement grains, towards a second porous mode (peak reached at 28 and 90 days between 100 et 200 A), which represents the residual voids located in between the clusters of the hydrates formed. Furthermore, the microstructure of the drying mortar evolves only slightly (see Figure 7b), and the initial 137porous mode remains predominant inasmuch as the water initially present has evaporated to the detriment of the formation of hydrates [7]. Figure 7a and 7b: Pore volume distribution versus pore radius for fluid mortar, of autogenous sealed sample (A) and of drying and hydrating sample (B) Autogenous sample (A) Dry ing sample (B), high 100 100 Ze 80 =< 80 | reo E "a "Ee 60 'E 60 E E a 3° 2° Aen 8 20 $ 20 if i a 4 E 4] R OF 0 at J IE] 1E#2 1E#3 1EM4 1E+S 1E+6] IEF] 1E+2 1E+3 IEH IES 1E+6 Pore radius (A) Pore radius (A) Table 1 : Results from mercury intrusion porosimetry ‘age Sample Area where the | Fluid mortar Control mortar sample istaken | Total porosity (%) | Total porosity (%) high 20,90 14,29 packed low 21,40 13,75 7 days high 19,39 14,89 g! . 4 drying B ning low 19,42 12,66 ss high 19,51 13,30 packed A low 19,28 12,66 28 days ae high 19,60 13,18 ane low 17,94 13,04 Tigh 15,40 packed A io cr 90 days high 19,73 avinge low 17,96 Tigh 15,22 ked A packes low 14,75 180 days high 19,32 di B ce yee low 17,54The evolution of the microstructure between 7 and 28 days is delayed and quite significant in the case of set-up A ; this feature can be ascribed to the pozzolanic reaction with the fly ash. The results at 180 days show little change in the pore structure in comparison with the 90-day results for both of the test set-ups : the same phenomena continue to be applicable. In Table 1, the overall porosity results from both set-ups A and B highlight the effec: of drying on the microstructure. Figure 8a and 8b : Pore volume distribution versus pore radius for normal mortar, of autogenous sealed sampl2 (A) and of drying and hydrating sample (B) Autogenous sample (A) 7 days Drying sample (B), high 80 F | 28 days: 80 = 60 60 : te FE t : = ‘= : E40 £40 ‘ S = 3 g B20 B20 S S a a 0 0 + fA 1E+1 1E+2 1B+3 1E+4 1E+S 1Et6 1EH 1E+2 1E¥3 1EH4 IE+S 1E+6 Pore redius (A) Pore radius (A) The evolutionary mode of the reference mortar pore structure is the same as that of the fluid mortar : the pores with radii in the range of 1000 A disappear in favor of pores with radii closer to 100 A. Yet, this phenomenon begins far too early since, at 7 days, the peak at 100 A is already well pronounced in Figure 8a. Similarly, in Figure 8b, the part near the drying surface (specimen B) is impacted by water losses. However, Table I demonstrates that the overall porosity is lower for the reference mortar at 7 days and is still decreasing at 28 days. Moreover, these mercury porosimetry results confirm that the microstructure of the reference mortar is changing less significantly after 7 days than that of the fluid mortar since the reference mortar does not contain any fly ash. Despite the heavy water loss in the fluid mortar during drying, both its total shrinkage and autogenous shrinkage, as measured on the 4 x 4 x 16 cm prisms, are distinctly less than those of the reference mortar (see Figure 9) due to a greater quantity of sand compared with cement in the fluid mortar [9]. For the fluid mortar, fellowing an apparent stabilization of shrinkage around the age of 21 days, the amount of shrinkage (both autogenous and total) once again begins to rise. This phenomenon may be due to the pozzolanic reactions with the fly ash, which would 139corroborate the mercury intrusion porosimetry results in their indication of delayed reactions. Figure 10 : Deformations of fluid and normal mortars prisms [oes normal, total (By ——»—— uid, total (By a Ti, autogenous (AY _- =... normal, autogenous (A) —a— fluid, only drying (A) Dry ing time from 6 months for A (days) 180187, 194-201 208-215 222 229236 800 ' 700 Normal __| = 600 & 500 & 400 a 300 Fluid & 200 100 oly 0 7 14 21 42 49 56 28 Age (days) Figure 9 : Pore volume distribution comparison for fluid mortar at the age of 180 days for the 2 configurations A and B (high) 160 —sealed (A) _. 140 drying(B) ~< 120 4 ee Ss s ea s = ss pot DV/Dlog(t) (mm.g! 2 RX 8 2 1E+1] 1E+2 1E#3) IEH4.—IE#5—IE+6| Pore radius (A) 140After six months of protection from external conditions, the drying shrinkage of the fluid mortar is measured on a type A sample, whose pore structure is considerably changed since the largest capillaries had been clogged (see Figure 10). Total shrinkage is measured on a type B sample with a very different structure which remains macroporous. In Figure 9 however, both the total shrinkage and drying shrinkage are of the same order of magnitude, a feature which signifies that the porosity herein is of the type found between the hydrates (radii of less than 40 A and thus unmeasurable using mercury intrusion porosimetry). This porosity is similar in both of the mortars tested, yet does represent a lesser quantity in the type B specimen, which was not as advanced in age. The bulk of the total shrinkage therefore is due to drying. 4. Conclusions and outlook for further research This study has shown both the importance and the impact of drying on the evolution in the microstructure and deformations of two distinct materials. For the reference mortar, the water loss, both initially and during the first few days, exerts a real impact in that the bulk of the cement's hydration actually occurs during the first week. Protection would thus be beneficial, especially considering that this initial drying when added to a slight segregation leads to differences in the material's density as a function of height. For the self-compacting mortar, the structure is quite homogeneous following setting and is not disturbed by the initial loss of excess water. However, in the long run, the evolution of the microstructure is affected by drying. The porosity is higher and coarser, and height-related differences appear when the pozzolanic reactions are limited or even inhibited by drying. For this reason, it can be assumed that improved cement hydration in the short term is likely to enhance longer-term characteristics by increasing the rate of pozzolanic reactions. One potential solution may lie in curing the mortar at a very early age in order to avoid an overly-porous structure which could induce higher vulnerability to the penetration of aggressive agents (water, salts, carbon dioxide, etc.). Such a solution however still needs to be demonstrated. Moreover, this kind of protection would not have a significant impact on the fluid mortar's total shrinkage since the greatest portion of shrinkage is due to drying. This drying shrinkage remains the dominant influence, even on the material which had been protected from all hydric exchanges with the surrounding media for six months. Nonetheless, the fluid mortar shrinkage is sharply lower than the reference mortar shrinkage. Studying the influence of drying on the evolution in microstructure and deformations, coupled with the use of these mortars for measuring characteristics like gas permeability, does seem from our vantage point to be most necessary for ensuring the durability of these new materials, which offer such great ease of application. 141We would like to express our gratitude to Mr. Chappuis of the Lafarge Company for his contribution and support in this research effort. 5. References 6. Sedran, T., « Rhéologie et rhéométrie des bétons. Application aux bétons autonivelants », Thése de doctorat de I’Ecole Nationale des Ponts et Chaussées, 1999. Sedran, T., de Larrard, F., Hourst, F. and Cantamines, C., « Mix-design of self compacting concrete », Proceedings of the International RILEM conference on Production methods and workability of concrete, Paisley, 1996, ed. by P.J.M. Bartos, p. 439-450. Skarendahl, A., « Self compacting concrete for improved productivity, working environment and performance », IREX-mecting, Paris, february 1999. Baroghel-Bouny, V., « Caractérisation des pates de ciment et des bétons, Méthodes, Analyse, interprétations » Laboratoire Central des Ponts et Chaussées, Paris, 1994, 468 p Ozawa, K., Tangtermsirikul, S. and Maekawa, K., « Role of powder materials on the filling capacity of fresh concrete », 4" CANMET/ACI International Conference on fly ash, silica fume, slag in concrete, Istanbul, 1992, p. 121-137. Naik, T.R., Singh, S. and Ramme, B., « Mechanical properties and durability of concrete made with blended fly ash », ACI Materials Journal, july-august 1998, vol. 95, n°4, p. 454-462. Aitcin, P.C., Neville, A.M. and Acker, P., « Integrated view of shrinkage deformation », Concrete International, september 1997, vol. 19, n°94, p. 35-41. Baroghel-Bouny, V., « Experimental investigation of self-desiccation in high- performance matérials - Comparison with drying behaviour », Proceedings of the International Research Seminar “Self-desiccation and its importance in concrete technology”, Lund, Sweden, june 1997, ed. by B. Pernen and G. Faagerlund, p. 72-87. Khelidj, A., Loukili, A. and Bastian, G., « Expérimental study of the hydro- chemical coupling inside maturing concrete: effect on various types of shrinkage », Materials and structures, Vol. 31, november 1998, n°213, p.588- 594, Granger, L., Torrenti, J-M. and Acker, P., « Thoughts about drying shrinkage : scale effects and modelling », Materials and Structures, Vol. 30, march 1997, p. 96-105. 142STABILITY OF SELF-CONSOLIDATING CONCRETE, ADVANTAGES, AND POTENTIAL APPLICATIONS K.H. Khayat!, C. Hu? and H. Monty} 1 Université de Sherbrooke, Sherbrooke, QC, Canada 2 Lafarge Laboratoire Central de Recherche, St Quentin Fallavier, France 3 Monsanto-Kelco, San Diego, CA, USA Abstract Self-consolidating concrete (SCC) is a highly flowable, yet stable concrete that can spread readily into place and fill the formwork without consolidation and without undergoing any significant separation of material constituents. The use of SCC can enable the reduction in labor demand needed for vibration and surface finishing, accelerate the placement rate, and secure superior surface quality. This paper reviews the various approaches adopted to secure stability of SCC that involve the reduction in free water content, the incorporation of viscosity-enhancing admixture, and a combination of both approaches. The paper highlights the economic and technical advantages of using SCC and discusses perspectives for the development and use of SCC, mostly from a North American prespective. In the short term, the use of SCC will be mostly cost driven, involving mostly the construction of horizontal elements that are more economical and simpler to produce. The casting of concrete in areas of restricted access necessitating placement with little or no vibration is also an emerging industry for the use of SCC. In the long run, the development and use SCC will be cost and quality driven, especially as further knowledge and experience become available regarding mixture formulation, quality control, and performance. SCC should gain wider acceptance for the construction of vertical members and prefabricated elements, as well as for the casting of highly reinforced sections to reduce construction duration and enhance performance. 1. Stability of self-consolidating concrete Self-consolidating concrete (SCC) is a new category of high-performance concrete that can spread into place and fill the formwork without consolidation and with minimum risk of separation of material constituents. Special care must be taken in proportioning SCC to secure adequate deformability and high stability. The stability of fresh concrete 143describes its ability to resist bleeding, sedimentation, and segregation and depends on the cohesiveness and viscosity of the mixture. From a rheological point of view, a successful SCC is characterized by low yield value needed for high capacity of deformation, and moderate viscosity necessary to ensure uniform suspension of solid particles during casting and thereafter until the setting. The reduction in viscosity decreases the ability of the mixture to maintain homogeneous dispersion of constituents. This can cause anisotropy in the direction of casting and weaken the interface between the aggregate and hardened cement paste as well as between the paste and embedded reinforcement. Accumulation of bleed water under horizontal reinforcement and settlement of concrete around rigidly held bars weaken bond development.!.2 Surface settlement can also lead to the formation of surface cracks above upper reinforcement because of their restraining action to the settling of plastic concrete. Ensuring adequate stability is especially important when the concrete is cast through sections with high density of reinforcement. Such SCC must exhibit adequate cohesion between the aggregate and cement paste to secure uniform deformation through the restricted spacing. A lack of stability can lead to segregation near the obstacles causing blockage of the flow. Inter-particle friction among solid particles can increase the resistance to flow and the degree of interference with the spread, especially among closely spaced obstacles. The volume of coarse aggregate and sand are often limited; for example, the volume of coarse aggregate is limited to 50% of total solid volume and that of the sand to 40% of mortar volume. Reducing the free water content and increasing the concentration of fine particles can enhance the cohesion and viscosity, and hence the stability of SCC. Free (or movable) water is the total mix water minus water that is physically and chemically retained by the aggregate and powder materials as well as any water bound by chemical admixtures, such as water-retaining admixtures. The various approaches adopted to enhance the stability of SCC are reviewed below. A. Minimum free water content This approach entails the use of high content of fines to limit aggregate volume and low water content to reduce free water content. The reduction in free water was shown to decrease inter-particle friction between solids particles and yield better deformability through restricted spacing34 The resulting SCC has a typical W/CM of 0.30 to 0.35 with powder content finer than 80 jim of 500 to 600 kg/m3. Such high content of fines is necessary to increase the paste volume and enhance deformability and stability. Fine powder includes cement and supplementary cementitious materials as well as fillers that are combined to enhance the grain-size distribution, packing density, and reduce inter- particle friction to lower the water demand necessary to attain a given viscosity. The replacement of part of the cement with a less reactive powder is also necessary to limit strength gain and the heat of hydration in such rich systems.‘-7 In general, the approach of minimizing free water content to enhance stability can result in SCC mixtures with low yield value and moderate-to-high viscosity levels. The low 144water content requires a relatively high dosage of high-range water reducer (HRWR) to obtain the required deformability, especially with the lower binder contents. B. Moderate water content and medium concentration of VEA In this approach, the W/CM can be maintained at the value necessary to attain the required strength and durability (for example 0.45), and a moderate dosage of viscosity- enhancing admixture (VEA) is employed to secure the required stability. A combination of high powder factor (450 to 600 kg/m3) and water content (180 to 220 kg/m?) can be used here to increase the paste volume, hence enabling the reduction of aggregate content necessary for high deformability and low risk of flow blockage. The VEA affects the aqueous phase of the cement paste where chains of the water-soluble VEA polymer can imbibe some of the free water, thus enhancing viscosity of the paste. The use of VEA, along with adequate concentration of HRWR, can ensure high deformability and adequate stability leading to greater filling capacity and better homogeneity of hardened properties.2° Higher resistance to segregation over a wide range of fluidity, and better flowability through narrow spaces can be obtained when a suitable dosage of VEA is incorporated, !0.!1 Mixtures containing VEA exhibit shear-thinning behavior whereby the apparent viscosity decreases with the increase in shear rate. This can facilitate the deformability of the mixture during placement. Once cast in place, the shear rate decreases and the apparent viscosity increases, resulting in greater stability at rest and better capacity to retain water and suspend solid particles. C. Low water content and low concentration of VEA This approach involves the use of VEA at low concentration in mixtures with limited water content. The VEA is mainly used here to reduce the variability of the SCC that can arise from changes in material properties and placement conditions. The VEA controls bleeding and renders the concrete more robust, while the low water content provides mostly the required level of viscosity. Compared to SCC. with 0.33 W/CM without any VEA, the incorporation of a low concentration of VEA (50 g/m? of welan gum) is reported to reduce the variability in slump flow with changes in cement Blaine B18 to 342 m2/kg), fineness modulus of sand (2.08 to 3.06), and temperature of fresh concrete (10 to 30°C), !0 2. Comparison of various approaches Table 1 reviews the results of a series of experiments! carried out to compare plastic properties of SCC made with mixture proportioning approaches A and B. Two SCC mixtures made with low W/CM and no VEA, and two others incorporating moderate contents of VEA (Kelco-Crete) and higher W/CM, were prepared. The study evaluated the differences in the performance of SCC concrete made with moderate and high contents of cementitious materials (425 and 555 kg/m3, respectively). A flowing conventional concrete made with 340 kg/m} of cementitious materials and 0.50 W/CM 145was also prepared for the reference concrete. The slump flow values of the SCC and conventional mixture were approximately 650 and 440 mm, respectively). The five mixtures in Table | were prepared with a ternary binder made with Type 10 cement, and 3% silica fume and 20% of Class F fly ash replacements, by mass of mass of binder. Such a combination proved to reduce water demand and enhance stability. A riverbed siliceous sand and two nominal sizes of crushed limestone aggregate were combined to ensure adequate packing density. A naphthalene-based HRWR conforming to CSA3-A266.6-M85 having a solid content of 42% and a specific gravity of 1.21 was incorporated in all mixtures. The results indicate that a SCC made with high binder content of 555 kg/m3, low S/Pt volume of 0.60 to 0.66, and low coarse aggregate volume of 300 to 330 I/m3 is more suitable to ensure high filling capacity of densely reinforced sections than a concrete of similar slump flow (650 mm) with moderate binder content of 425 kg/m}, 0.70 to 0.85 S/Pt volume, and 375 to 405 I/m? of coarse aggregate. In both SCC systems with binder contents of 555 and 425 kg/m3, the incorporation of VEA at moderate dosage enhanced deformability and resistance to surface settlement, despite the higher W/CM. This was especially beneficial in concrete with 425 kg/m3 of cementitious materials. The increase of W/CM from 0.38 to 0.50 and incorporation of a moderate concentration of VEA resulted in a net decrease of settlement from 19 to 3.1 mm and increase in filling capacity from 19 to 42%. The flow resistance (g) and relative viscosity (h) values (determined using an IBB concrete viscometer) decreased from 1.35 to 0.29 Nm and 12.5 to 10.1 Nm.s, respectively. Such improvement in performance is due in part to the reduction in coarse aggregate volume from 405 to 375 L/m3 and S/Pt value from 0.85 to 0.70. This can decrease the extent of collision of coarse aggregate and sand particles in the vicinity of obstruction that can increase internal resistance to flow and lead to interference with spreading and leveling. The incorporation of a VEA, despite the higher W/CM, can then increase cohesion friction between the mortar and coarse aggregate to secure a more uniform flow through restricted sections with less segregation and flow blockage. As shown in Table 1, the dosage of HRWR of the conventional concrete was 2.1 I/m3 to secure a slump flow of 440 mm (slump of 225 mm). This dosage ranged between 3.1 and 5.9 I/m3 for the SCC. The 425-0.38 mixture had a high demand of HRWR of 5 V3 compared to 3.1 I/m3 for the 425-0.50/WG concrete, For mixtures with high cementitious material content, similar HRWR was needed in both mixtures, 4.8 and 4.1 m3, with the higher value associated with the VEA SCC. The material cost reported in Table | does not include delivery or profit. Compared to the conventional flowing concrete, the SCC cost 25 to 50% more. In the case of SCC containing 425 kg/m? of cementitious materials, both mixture categories, with low W/CM without VEA and with moderate W/CM and VEA, had similar costs with the additional dosage of HRWR required for the former concrete offsetting the extra cost of VEA. Given the better performance of the VEA.SCC, and similar cost, it is advantageous to use this approach in proportioning SCC with moderate content of cementitious materials, On the other hand, for mixtures made with the higher binder 146content (555 kg/m3), the VEA concrete costs an additional 10% compared to the non- VEA SCC. The VEA SCC provides marginal improvement in performance (higher filling capacity, lower flow time and lower settlement) and, although not tested here, should enhance the resistance to bleeding and result in better microstructure and homogeneity of the hardened properties. 6!3 Table 1 Mixture proportions of five evaluated concrete mixtures!2 Slump flow (mm) su + ae [so | CM content, combination W/CM, VEA AWG cra ‘. 20. } © coarse aggregate Xv(%) Fig-4 Measuring equipment of Fig-5 Result of surface resistance to surface resistance volumetric ratio of coarse aggregate theological properties at X,=30.0%, three mix ratios of viscosity-controlling admixture (HF) were used: 1.3, 2.0 and 2.7% of the weight of cement. Steel plates were used for the test. Further, in order to investigate the effects of withdrawal rate and concrete pressure on surface resistance, the withdrawal rate of the test plate and the depth to which it was inserted were taken up as factors and studied. The equipment used to measure surface resistance is shown in Fig-4. The test plate was inserted into the concrete and then drawn up at a constant speed while measuring displacement and load. Surface resistance was obtained by subtracting the value (weight of test plate-buoyancy) from the maximum value and dividing it by all the surface area of the test plate in contact with the concrete at the start of withdrawal. 3.2 Results of Experiment and Consideration All mix conditions yielded a slump flow of at least 500mm and were highly fluid. In Fig-5, the relation between volumetric ratio of coarse aggregate (X,) and surface resistance is shown. The surface resistance increased with Xy. This seems due to the increase in the amount of coarse aggregate in contact with the test plate. Regression formulae in which the surface resistance (T y:Pa) is adopted as the object variable and X,(%) as the predictor variable are given below. The other factors had no effect on the surface resistance. Using these formulae, the surface resistance of the formwork can easily be estimated from X,. T y= 0.39Xv - 0.02 (12) 167: o> § |. | 10@10 a wo} es |: £ Z z 2 s as) a | * E . 2 Qe 28 3E 36 15 —4.'5 Volumetric ratio of Unit : ca coarse aggregate Xv(%) Fig-6 Equipment for evaluatin Fig-7 Relation between Xv and internal friction angle internal friction angle 4. Evaluation of internal friction angle of concrete 4.1 Outline of experiment ‘The evaluation of internal friction angle was implemented as follows. First high-fluidity concrete was placed into a column-shaped form, and the pressure distribution in the horizontal direction on the sides of the formwork was measured. The results were compared with a calculated value of pressure assuming a liquid. As a factor indicating of concrete properties, X, was set at five levels: 26.0, 28.0, 30.0, 32.0 and 34.0. The ratio of HF was set at 2.0%. The shape of the formwork and pressure sensor locations used for evaluating the internal friction angle are shown in Fig-6. Ten pressure sensors were used to measure pressure in the transverse direction. Immediately after compaction, pressure in the transverse direction (0 4) on the side faces of the formwork was measured. The pressure of liquid pressure distribution (@ ,) was then calculated for the depth of each sensor and using the specific gravity of concrete. The internal friction angle () was obtained using formulae (2). 4.2 Experimental Results and Consideration All mix proportions yielded slump flows of more than 500mm and exhibited high- fluidity. The relation between internal friction angle of concrete (:° ) and X, is shown in Fig-7. A single regression analysis using internal friction angle as the object variable and X, as the predictor variable was carried out, yielding the following regression formulae: 168@ =0.025X,? - 1.37X, + 20.9 (13) Using this formulae, the internal friction angle can easily be estimated from X,. 5. Evaluation of resistance due to steel reinforcing bars 5.1 Experiment Planning A quantitative evaluation of the resistance imposed by steel reinforcing bars when high- fluidity concrete flows into formwork was carried out. Two cases were considered: a relatively large section in which concrete can pass all clearance between the steel reinforcing bars or between a steel reinforcing bar and a formwork surface, and a situation in which concrete is blocked as it passes between two steel reinforcing bars or between steel reinforcing bars and formwork surface. In the latter case, the fact that concrete cannot pass between steel reinforcing bars and formwork surface means the space of it causes the loss of sectional area for passing and So great resistance was expected to occur. The resistance due to steel reinforcing bars in these two cases was evaluated in separate experiments. 5.2 Experiment I : Evaluation of resistance due to steel reinforcing bars when concrete passes all openings a)Outline of experiment Test apparatus A, as shown in Fig-8, was used for the test and the clearance of steel bars and loss ratio of sectional area for passing due to steel bars were varied by altering the number of steel bars in the section and their diameter. The steel bars were arranged longitudinally at equal intervals. The left-hand chamber was filled up to the top with concrete, and the gate was opened. When the movement of concrete had stopped, the height of concrete in the left and right chambers was measured. The pressure | difference was then calculated from the 4 specific gravity of the concrete. The & pressure difference was also measured with 3 2 no steel bars in place, and this value was g i & used as the standard value. The resistance due to the steel bars was evaluated by subtracting the standard value from the 140 pressure difference under each respective 2 Le a condition. The X, value of the concrete Unit (mm) Passing section was set at four levels: 26.0, 28.0, 30.0, and 32.0%. Fig-8 Test apparatus A 169