ICOLD

COMMITTEE ON CONCRETE DAMS
THE PHYSICAL PROPERTIES OF HARDENED CONVENTIONAL CONCRETE IN DAMS

Draft for ICOLD Review

MARCH 2008

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams

AVERTISSEMENT – EXONÉRATION DE RESPONSABILITÉ :

Les informations, analyses et conclusions contenues dans cet ouvrage n'ont pas force de Loi et ne doivent pas être considérées comme un substitut aux réglementations officielles imposées par la Loi. Elles sont uniquement destinées à un public de Professionnels Avertis, seuls aptes à en apprécier et à en déterminer la valeur et la portée et à en appliquer avec précision les recommandations à chaque cas particulier. Malgré tout le soin apporté à la rédaction de cet ouvrage, compte tenu de l'évolution des techniques et de la science, nous ne pouvons en garantir l'exhaustivité. Nous déclinons expressément toute responsabilité quant à l'interprétation et l'application éventuelles (y compris les dommages éventuels en résultant ou liés) du contenu de cet ouvrage. En poursuivant la lecture de cet ouvrage, vous acceptez de façon expresse cette condition. NOTICE – DISCLAIMER : The information, analyses and conclusions in this document have no legal force and must not be considered as substituting for legally-enforceable official regulations. They are intended for the use of experienced professionals who are alone equipped to judge their pertinence and applicability and to apply accurately the recommendations to any particular case. This document has been drafted with the greatest care but, in view of the pace of change in science and technology, we cannot guarantee that it covers all aspects of the topics discussed. We decline all responsibility whatsoever for how the information herein is interpreted and used and will accept no liability for any loss or damage arising therefrom. Do not read on unless you accept this disclaimer without reservation.

As submitted for ICOLD review, March 2008

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams

COMMITTEE ON CONCRETE DAMS COMITÉ DES BARRAGES EN BÉTON (1991-2008)
Chairman/Président USA/Etats-Unis USA/Etats-Unis/Canada/Canada Vice Chairman/Vice Président France/France Members/Membres Australia/Australie Austria/Autriche M. Pegg (to 1996) B. Forbes (since 1997) H. Huber (1996 - 2000) G. Heigarth (2001) W. Pichler (since 2002) F.R. Andriolo (to 2001) J. Marques Filho (since 2002) R.G. Charlwood (1996 - 2006) P. Ko (since 2007) Shen Conggang (to 1998) Jia Jinsheng (since 1999) V. Ukrainczyk (to 2006) J. Launay (to 1999) M. Guerinet (since 2000) M.R.H. Dunstan B.J. Parmar M.E. Omran (to 1997) A.A. Ramazanianpour (1998 – 2006) M. R. Jabarooti (since 2007)
P. Bertacchi (to 1995)

J.R. Graham (1992 – 1997) R.G. Charlwood (since 1998) J. Launay (since 1998)

Brazil/Brésil Canada/Canada China/Chine Croatia/Croatie France/France Great Britain/Grande-Bretagne India/Inde Iran/Iran

Italy/Italie Japan/Japon

M. Berra (since 1996) S. Nagataki (to 1997) O. Arai (1998-2002) T. Uesaka (since 2002) O.J. Berthelsen A. Camelo i

Norway/Norvège Portugal/Portugal As submitted for ICOLD review, March 2008

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Russia/Russie South Africa/Afrique du Sud Spain/Espagne Sweden/Suède Switzerland/Suisse USA/Etats-Unis A.D. Osipov (to 2000) G. Kostyrya (since 2000) J. Kroon (to 2004) J. Geringer (since 2005) J.M. Gaztanaga (to 1999) J. Buil Sanz (since 2000) J. Alemo (to 2006) T Ekstrom (since 2007) H. Kreuzer (to 2007) M. Conrad (since 2007) V. Zipparro (1999-2000) G. Mass (2001) R.G. Charlwood (since 2002)

As submitted for ICOLD review, March 2008

ii

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams

ACKNOWLEDGEMENTS

This Bulletin was drafted under the auspices of the ICOLD Committee on Concrete Dams. The Bulletin was initiated under the Committee Chairmanship of J.R. Graham (USA), and completed under the Chairmanship of R.G. Charlwood (USA/Canada). The lead author of the Bulletin initially was P. Bertacchi (Italy) and since 1996 has been M. Berra (Italy) who has developed and completed the document. H. Kreuzer (Switzerland) provided a precious assistance in reviewing and editing. The initial drafting team included P. Bertacchi, M. Berra, G. Ferrara & P. Morabito (Italy), V. Ukrainczyk (Croatia), R. Charlwood (Canada), M. Machado (Portugal), J. Alemo (Sweden) and T. Liu (USA). The final Bulletin was drafted as follows: Section 1 - Introduction, was prepared by P. Bertacchi and M. Berra (Italy). Section 2 – Strength Properties, was initially drafted by V. Ukrainczuk (Croatia) and subsequently redrafted by H. Kreuzer (Switzerland) with inputs regarding sonic measurements from M. Berra (Italy). Section 3 – Elastic Properties, was initially drafted by R. Charlwood (USA/Canada) and subsequently completed by M. Berra (Italy) with inputs from H. Kreuzer and M. Conrad (Switzerland) and J. Buil Sanz (Spain). Section 4 – Creep Properties, was drafted by T.C. Liu (USA) and completed with contributions from M. Berra (Italy) and H. Kreuzer ((Switzerland). Section 5 – Shrinkage, was drafted by M. Berra (Italy). Section 6 – Thermal Properties, was drafted by P. Morabito (Italy) and completed with contributions of M. Berra (Italy) and H. Kreuzer ((Switzerland). Section 7 – Water Permeability, was initially drafted by J. Alemo (Sweden) and subsequently redrafted by T. Ekstrom (Sweden) with contributions from H. Kreuzer (Switzerland) and M. Berra (Italy). Section 8 – Frost Resistance, was drafted by M. Berra (Italy) with inputs from H. Kreuzer (Switzerland). Appendix A – Fracture Energy: application for dams, was drafted by G. Ferrara and G.A. Plizzari (Italy) with inputs from H. Kreuzer (Switzerland) on size effects and biaxial testing. Appendix B – Physical Properties of Construction Joints in Concrete Dams, was drafted by F. R. Andriolo and W.A. Pacelli (Brazil), G.S. Sarkaria (USA) with inputs from V. Zipparrro (USA), M. Berra (Italy), M. Guerinet (France) and J. Buil Sanz (Spain). Appendix C – Physical Properties of Concrete Subjected to Expansion Phenomena such as AAR in Dams, was drafted by R. Charlwood (USA/Canada) with inputs from D. Curtis (Canada) and subsequently completed and edited by M. Berra (Italy).

As submitted for ICOLD review, March 2008

iii

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams

THE PHYSICAL PROPERTIES OF HARDENED CONVENTIONAL CONCRETE IN DAMS
Table of contents Section Committee on Concrete for Dams Acknowledgements Table of contents FOREWORD 1. INTRODUCTION 1.1 Purpose of the Bulletin 1.2 Scope 1.3 Relationship to earlier Bulletins 1.4 Definitions 1.5 Content of the Bulletin 2. STRENGTH PROPERTIES 2.1 General 2.2 Concrete testing 2.3 Compressive strength 2.4 Tensile strength 2.4 Shear strength 2.6 Dynamic strength 2.7 References 3. ELASTIC PROPERTIES 3.1 General 3.2 Static modulus of elasticity 3.3 Dynamic modulus of elasticity 3.4 Poisson’s ratio 3.5 Significance of elastic properties on dam behaviour 3.6 Use of elastic properties in mathematical models for dam structural analysis 3.7 References As submitted for ICOLD review, March 2008 iv Page i iii iv vii

6 Coefficient of thermal expansion of hardened concrete 6.5 Thermal diffusivity of hardened concrete 6.2 Factors influencing the concrete water permeability 7.3 Tests method for water permeability 7.3 Measurements of creep 4.6 Effect of shrinkage on cracking 5.8 References 7.7 In situ measurements of concrete temperatures 6.1 General 4.1 General 5.1 General 6.4 Modelling water permeability in saturated concrete 7.4 Factors affecting drying shrinkage 5. THERMAL PROPERTIES 6.7 Effect of shrinkage on concrete dams 5.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams 4.3 Causes and mechanism of drying shrinkage 5.5 References As submitted for ICOLD review.5 Estimation of drying shrinkage 5. SHRINKAGE 5.8 References 6.4 Significance of creep properties on dam behaviour 4.6 References 5.1 Introduction 7. March 2008 v . CREEP PROPERTIES 4.2 Factors influencing creep 4.2 Types of shrinkage 5.5 Modelling creep in structural analysis 4.3 Thermal conductivity of hardened concrete 6.4 Specific heat of hardened concrete 6. WATER PERMEABILITY 7.2 Temperature rise of young concrete during the hydration of cementitious materials 6.

4 Methods for experimental determination of the concrete frost resistance 8.1 General B.2 The nature of the ASR in concrete dams C.5 References As submitted for ICOLD review. PHYSICAL PROPERTIES OF CONSTRUCTION JOINTS IN DAMS B.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams 8.3 Factors affecting the frost resistance 8.5 References C. FROST RESISTANCE 8.3 Damage due to ASR in concrete dams C.5 Applicability of linear elastic fracture mechanics (LEFM) A.1 General A.2 Fracture parameters A.3 Performance of construction joints and investigations B. FRACTURE ENERGY: APPLICATIONS FOR DAMS A.4 Case histories on construction joint investigations B.2 Frost resistance.7 References APPENDICES A.1 Introduction. mechanism and effects 8.1 General 8. C.6 Freezing and thawing on concrete dams 8.2 Typical requirements and methods for joint treatment B.4 The physical properties of ASR affected concrete in Dams C. PHYSICAL PROPERTIES OF CONCRETE SUBJECTED TO EXPANSION PHENOMENA SUCH AS AAR IN DAMS C.5 Laboratory diagnostic investigations 8. March 2008 vi .6 References B.3 Determination of fracture parameters A.4 Factors influencing the factor energy A.

1960). Berra (Italy). Charlwood (CANADA/USA). to show typical behaviour. Robin G. This specific topic may be the subject of a separate Bulletin at a later date. refreshing the partial information contained in some out-of-date ICOLD Bulletins. These include strength. and frost resistance. Bertacchi (Italy) and subsequently M. permeability. prepared this Bulletin. March 2008 vii . creep and temperature on concrete for large dams . methods to introduce the properties in mathematical models to be utilized both for design and observation. chaired initially by P. The main body of this Bulletin addresses the physical properties of the mass concrete used most frequently in design and analyses of concrete dams and appurtenant structures. and subsequently under the Chairmanship. The purpose of this Bulletin is to provide a comprehensive treatise on the physical properties of hardened conventional concrete for dams. the properties of construction joints and the properties of expanding concrete in dams. factors influencing it. for each property considered. then under the Chairmanship of J. the International Commission on Large Dams (ICOLD) directed that its Technical Committee on Concrete for Dams. in particular concrete in dams subject to alkali-silica reaction although this latter topic will be addressed in a new Bulletin. This document supersedes the ICOLD Bulletins: n° 15 (Frost resistance of concrete . and. elastic. The scope of the Bulletin is. Charlwood Chairman Committee on Concrete Dams March 2008 As submitted for ICOLD review.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams FOREWORD In 1991. shrinkage and thermal properties. The Appendices address newer and more advanced concepts including the application of fracture energy to concrete dams.G. giving an account of new approaches and examining also some important properties not included in the previous Bulletins.1976). last but not least. to undertake the preparation of a comprehensive report on the physical properties of concrete in dams.R. A sub-Committee of the ICOLD Committee on Concrete for Dams. n° 25 (Extensibility of concrete for large dams . of R. This Bulletin does not specifically address the properties of hardened concrete of Roller Compacted Concrete (RCC) dams which are partially dealt in the Bulletin n°75 – Roller Compacted Concrete for Gravity Dams and in Bulletin n° 126 Roller-Compacted Concrete Dams. creep. methods for experimental determination. Graham (USA).1976) and n° 26 (Methods of determining effects on shrinkage.

.... in particular concrete in dams subjected to alkali-aggregate reactions. drains or have cavities like cracks or honey combs...................................................1 1.................1976) and n° 26 (Methods of determining effects on shrinkage....1976)........ n°25 (Extensibility of concrete for large dams .......1960)...1)... comprehensive and update treatise on the physical properties of hardened conventional concrete for dams....... often have manmade structural components and imperfections as for example construction and contraction joints..... In particular this new document serves to supersede the following out-of-date ICOLD Bulletins: n°15 (Frost resistance of concrete ........... B and C............... injected sealing materials).. Some information on the physical properties of “manmade discontinuities” and cracked concrete is presented in the Appendices A..............................3 CONTENT OF THE BULLETIN ........ As submitted for ICOLD review.......... drying shrinkage and thermal properties...... creep........ These include strength...4 1..............g.............. refreshing the partial information contained in some out-of-date ICOLD Bulletins.............. elastic properties............... waterstops............. creep and temperature on concrete for large dams ......ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 1 (Introduction) 1 INTRODUCTION 1 INTRODUCTION .... 1.................... the properties of construction joints and some physical properties of expanding concrete..........................3 1.............................4 1. giving an account of new approaches and examining also some important properties not included in the previous Bulletins.......2 1....... such as dams.....3 DEFINITIONS ............1 1. and durability.... It is understood that mass concrete material represents the intact concrete inside the mass of the dam and not the more general mass concrete dam that include also structural components and imperfections (Fig... Chapters 2 through 8 focus on issues of the “mass concrete material”.........1 PURPOSE OF THE BULLETIN The purpose of this Bulletin is to provide a unique.................. The main body of this Bulletin addresses the physical fundamental properties of the mass concrete material used most frequently in design and analyses of concrete dams and appurtenant structures.. Furthermore the Appendices address newer and more advanced concepts including the application of fracture mechanics to concrete dams.......................................................5 PURPOSE OF THE BULLETIN...............3 RELATIONSHIP TO EARLIER BULLETINS ......... interfaces with other materials (e.................. mass concrete material can be thought of as “defect-free” whereas mass concrete structures..... In fact...1 SCOPE..... march 2008 Section 1-1 ........ water permeability.....

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 1 (Introduction) Mass concrete material Construction joint Fig.1 – Mass concrete material in a concrete dam This Bulletin does not specifically address the properties of Roller Compacted Concrete (RCC) dams which are partially dealt in the Bulletin n°75 – Roller Compacted Concrete for Gravity Dams and in the recent Bulletin n° 126 (Roller-Compacted Concrete Dams – State of the art and case histories). n° 79 .Ageing of dams and appurtenant works. n° 93 .Alkali-Aggregate Reaction in concrete dams. Chemical reactions and cracking resistance are also intentionally excluded since they are already extensively treated in other recent ICOLD Bulletins (n° 71 . The concrete erosion resistance is extensively dealt with in the ACI Report 210 R-93 “Erosion of concrete in hydraulic structures”.Exposure of dam concrete to special aggressive waters. 1. march 2008 Section 1-2 . Bulletin n° 107 .Control and treatment of cracks in concrete dams). The properties of fresh concrete are not considered. As submitted for ICOLD review.

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 1 (Introduction) 1. 1. Bulletin n° 26 “Methods of determining effects on shrinkage. for each property considered. creep and temperature on concrete for large dams” (1976) was focused mainly at the consequence on the structure. The physical properties of concrete must be addressed in a developing sense. the long list of papers has been omitted and substituted with a more recent one. caused by shrinkage. Standard curing conditions are necessary for comparative reference. factors influencing it. march 2008 Section 1-3 . methods for experimental determination. methods to introduce the properties in mathematical models to be utilized both for design and back analysis. and a deeper knowledge of creep behaviour. 1. to show typical behaviour. The adjective “conventional” is added to “hardened concrete” just to differentiate from roller compacted concrete that may need a separate treatment in some respects. starting just after initial set of cement. both from point of view of design and observation of dams.). humidity etc.4 DEFINITIONS The term “hardened concrete” refers to the properties of the concrete. of the air content and of the spacing factor were not considered. the influence of the degree of saturation. Bulletin n° 25 “Extensibility of concrete for large dams” (1976) is today at all outdate: new concept of fracture mechanics. and. The influence of chemical composition of cement on frost resistance was emphasized. As submitted for ICOLD review. When analysing a particular stress situation at a specific age. creep and temperature. In some cases an “equivalent age” could be defined to represent the “maturity” of a concrete. give now a totally different approach. but don’t represent a realistic situation of maturity. A first part containing these general aspects has been saved. updated and introduced as general considerations. the disruption mechanism due to freezing of the water in the capillaries. last but not least. we must know how the concerned properties have been developed up to that time and in that specific curing condition. It contains the results obtained in different laboratories more than 40 years ago.3 RELATIONSHIP TO EARLIER BULLETINS There are significant differences in objective and approach between the previous Bulletins and this new one: Bulletin n° 15 “Frost resistance of concrete” (1960) is really outdated. the evolution of a property in fact starts from initial set and depends on age and on curing conditions (temperature.2 SCOPE The scope of the Bulletin is.

both under uniaxial and biaxial tensile and compressive stress conditions (see Appendix A). durability and frost resistance). The type of loading excites different types of strength: compressive-tensile-shear strengths are of main importance. drying shrinkage. The structure of each chapter is generally as follows: „ Definition of the property.5 CONTENT OF THE BULLETIN The physical properties of hardened conventional concrete for dams have been addressed in seven Chapters dealing with all frequently used engineering parameters (strengths. Typically for dams are strength requirements at high maturity and the rather high scatter of strength values as compared to common civil structures. as reported by the experience acquired. creep. both for static and dynamic (earthquake) loading. „ Factors influencing the property. both in laboratory and in situ. water permeability. in the design phase. temperature. time. As submitted for ICOLD review. elastic properties. Specific additional information on advanced topics is added in the three Appendices. principal author(s) and content of each Chapter and Appendix are as follows: . „ Methods for experimental determination of the property.Strength properties (Section 2) Concrete strength is defined as the maximum stress recorded during the load testing of specimens carried out to failure. ageing related to physical and chemical processes. thermal properties. New testing techniques taking advantage of knowledge in fracture mechanics provide most valuable information about post-peak stress-strain behaviour. „ Typical behaviour. etc. Information on tensile and shear strength of construction joints is reported in Appendix B.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 1 (Introduction) 1. compressive and tensile strength degradation due to ageing caused by expansion phenomena in Appendix C. These are: the application of fracture mechanics concepts. humidity. Expected effects on the property. the performance of construction joints. just after construction or during the normal service life of the dam. such as: components and composition of the concrete. stresses. for the total behaviour of the mass concrete dam. march 2008 Section 1-4 . and the behaviour of concrete subject to expansion phenomena (particularly alkali-aggregate reactions) in dams. The subject. „ Transfer of acquired parameters of the mass concrete material to be utilised in mathematical models.

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 1 (Introduction) . can be used. Dynamic Modulus of Elasticity through non destructive techniques are presented and used both to evaluate the elastic properties and to determine the concrete quality and integrity.Drying Shrinkage (Section 5) Volume changes in concrete can be caused by mechanical.Elastic Properties (Section 3) The stress-strain concrete behaviour is quite complex and it was variously theoretically characterised. . As submitted for ICOLD review. creep deformation can also be explained partly in terms of viscoelastic deformation of the cement paste and to the gradual transfer of load from cement paste to aggregate. physical and chemical processes: in this Chapter only volume changes due to the moisture variations in concrete and the consequent drying shrinkage are dealt with. .Creep properties (Section 4) Creep is time-dependent deformation due to sustained load. * the environmental variation of temperature under the normal service conditions. Finally the significance of elastic properties on dam behaviour and the use of elastic properties in mathematical models for dam structural analysis are treated.50 m. march 2008 Section 1-5 . Creep properties are of particular relevance to understand the mechanism leading to the prediction of potential thermal cracking of mass concrete: the most extensive use of creep data is in thermal stress analysis for concrete dams. However for most of usual applications the classic and simplest constitutive model. Two distinct phenomena are considered: * the hydration of the cement which causes an increase of temperature during the hardening phase of concrete. . The two parameters of Static Modulus of Elasticity (E) and Poisson’s ratio (ν) have been examined with reference to the main factors affecting them and to the proposed correlation with the compressive strength.Thermal properties (Section 6) Temperature has a very important effect on concrete dams. based on linear elasticity in an isotropic homogeneous material. specific heat and coefficient of thermal expansion. In both cases the analysis of the temperature distribution in the dam and of the consequent induced stresses need data on the conductivity and diffusivity. The effect of drying shrinkage of concrete in large dams is discussed: it reduces rapidly with the thickness and become negligible at a depth of about 0. It is generally accepted that concrete creep is a rheological phenomenon associated with the gel-like structure of the cement paste.

They are presented through a series of case histories. Therefore. Forecasting the development and propagation of cracks can now be studied through fracture mechanics with a new mechanical property of concrete: the fracture energy.Water Permeability (Section 7) The flow of water or moisture through a concrete dam can affect its general performance and particularly its durability. it is necessary to prepare new concrete lifts in such a manner that the joint would have the same properties as the monolithic concrete. can become a discontinuity or a plane of weakness in the concrete mass. describing or interpreting cracking in concrete structures. . Information on water permeability of construction joints have also been reported in Appendix B.Frost resistance (Section 8) The deterioration of the concrete in a dam can be ascribed to a series of chemical and physical causes (both internal and external).AAR) and As submitted for ICOLD review.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 1 (Introduction) . Main test to evaluate the frost resistance are considered to prevent the phenomenon and specific diagnostic investigations are suggested to identify it in frost affected concrete dams to be repaired.strain curve in tensile tests. whether planned or unexpected. Freezing and thawing is the most common external physical attack. External exposure conditions and concrete quality are of main importance as they determine the resistance of concrete to freezing and thawing cycles.Physical Properties of Concrete Subjected to Expansion Phenomena such as AAR in Dams (Appendix C) Many existing large dams have been found to be subject to chemical reaction induced concrete expansion phenomena (mainly due to Alkali-Aggregate Reaction . To this aim specific construction joint investigations are required. . defined as the energy to create an unitary crack. . dams in particular. defined as tensile strain capacity to failure (see former Bulletin nº 25) is not able to adequately describe the tensile behaviour of concrete and therefore it is not useful for cracking analysis.Fracture Energy: application for Dams (Appendix A) The classical method of analyzing tension in structures appears to be inadequate for forecasting. Furthermore models describing the mobility of water in a porous material such as concrete have been developed through the years and briefly presented. Extensibility. if untreated. . march 2008 Section 1-6 . it is evaluated from the area subtended by the complete stress . Permeability and moisture coefficients can be estimated through suitable test methods.Physical Properties of Construction Joints in Concrete Dams (Appendix B) A construction joint.

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 1 (Introduction) the number is growing as awareness of the phenomena and the use of high resolution instrumentation increases. This phenomenon can cause serious problems to the operation of gated outlet structures and on the structural safety of the dam. through the analysis of testing results from AAR-affected dams. diagnosis procedure for the existing expansive dams and effective remedial measures are of big concern. The issue has already been the subject of the ICOLD bulletin n° 79 (Alkali-Aggregate Reaction in concrete dams -Review and Recommendations) published in 1991. mainly derived from the Bureau of Reclamation experience in USA. Furthermore a new updated Bulletin on chemical expansion of concrete in dams & hydro-electric projects have also been planned. As submitted for ICOLD review. The compressive and tensile strength. march 2008 Section 1-7 . the modulus of elasticity and creep phenomena are taken into consideration. Suitable prevention measures for new constructions. This Appendix just refers to the effects of Alkali-Aggregate Reaction on some physical properties of concrete.

...........33 2......2......2 2.....2...4.............................2 2.......................3................................5 Evaluation of Strength Testing ....................1 2.......................................8.......2 Effects of maximum size of aggregate (MSA) .5 Tensile strain capacity in dams ..............6.........2.....4.................................4 TENSILE STRENGTH.......2...............5 SHEAR STRENGTH ..............31 2...29 2..............32 2......................................6.....................................1 Effects of porosity and w/c-ratio on concrete strength .........................................................4.......4 2..........ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 2 STRENGTH PROPERTIES 2 STRENGTH PROPERTIES .............8...................................................................................................8 Use of tensile strength in mathematical models..... 8 Semi-destructive or non-destructive methods...2....................................... 37 2.........................1........................................................................................................1 Linear elastic analysis ..............................................................16 2.....6 Effect of sustained loading .3 COMPRESSIVE STRENGTH ..7 Multiaxial compressive stress domain in a dam ...26 2.....................................42 2........2.............6..............2 2.....................46 As submitted for ICOLD review..........3................2......... 7 2..2 Type of tensile testing ................................2.................1 Testing of concrete strength for new dams .....................1 Dynamic Compressive Strength....................................2 2.3 Tensile strength ratios......6 Testing of concrete strength in existing dams ................................................................6 DYNAMIC STRENGTH ..........20 2................................43 2........ 2 Testing program......32 2...................................2................................................................................................................................41 2..........2.............1......22 2..28 2.................. 5 Wet screening .....................4............................2.............4..........................................16 2..........3................................3 2.......3 Dynamic Shear Strength .......4................2...................1..............23 2.......................................2................................45 2.................................3 Effects of aggregate properties ................................................................................................... 9 Length-to-diameter ratio........ 3 Sample Sizes .............................39 2.........................................3............ 11 2.................3..........................................5 Effect of mineral admixtures and blended cements ..2..................................1........................................5 2.............................................7 REFERENCES ....4........................................3..............2........................4..2...........1 2.............................................. march 2008 Section 2-1 ..4 Test Equipment .....2..............3 Nominal age of concrete strength ...............................2.................29 2.........20 2.......................2 Non linear analysis .......................................................18 2..4..........................36 2.................................................................................................................................30 2........................25 2............. 1 2.........34 2........6 Generic Uncertainty in Test Procedures ................................................... 36 2.......................6 Tensile strength at joints ..............................2.................................................2 2.....................2 2.............................................................................2......................................................................................4 Effects of curing ..............................................................................................3.........................12 2................4...2....24 2..........8 General .3 2.. 10 Core vs.................4 Size effect ............. 10 Core size.....1 Tensile loading in the dam ...................................4 Testing in stages......7 Tensile strength criteria............................................................................................... laboratory strength ................................2 CONCRETE TESTING...............2.................. 8 Ratio between core diameter and MSA ......1 GENERAL ..............2 Dynamic Tensile Strength ..............................................................

march 2008 Section 2-2 . Types of strength important for dams are compression.1 GENERAL Strength is the most commonly considered parameter in structural design of new dams and for monitoring of existing dams. As submitted for ICOLD review. It is perhaps unfortunate that it has become the custom to relate structural safety mainly on compression although no dam ever failed by excessive compressive stresses. high strength concrete may experience less ultimate strain capacity and thus lead to earlier incipient cracking. blending for cementitious material). coefficient of variation. Standard testing does respect this physical fact. This Section therefore tries to emphasize the importance to include scatter (standard deviation. concrete in the dam is exposed to more variables.1.2. It is then the relevant material strength. which counteracts a specific failure mode to happen. The use of compression as the predominant strength indicator is also supported by the common assumption that improving compressive strength will improve other concrete parameters as well. This test stage is intended to set limits for design mixes and will be used as a guide-line for the subsequent test stages. This corresponds to common loading combinations which give rise to different modes of possible failure. Strength is not a deterministic but a random parameter. both under static and dynamic loading. However. references to test results not in the same manner. for example.1 Testing in stages Testing for new dams is done in stages. confidence limits) as an integral part of reports on concrete strength. compressive strength is easy to test and easy to monitor. The approach is for simple tests about main strength parameters with values from the literature for review and comparison. 2. coping with the increasing level of desired knowledge about material parameters. the effects of ageing and size effects. several types of tension and shear strength. Correlating the two conditions by sound inductive reasoning based on project-specific data is imperative.1 Testing of concrete strength for new dams 2.2 CONCRETE TESTING 2. Factors influencing strength are the concrete constituents. Such a general assumption may be misleading as.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 2. Three stages are common for concrete testing during final project phases: • STAGE I: preliminary testing programs to curtail several potential sources of constituents (borrow areas for aggregates. curing.2. the test conditions. Concrete strength is tested on small specimens under controlled laboratory conditions. This Section will therefore also focus on issues where other than compressive strength criteria are of importance.

It is evident that uncertainty levels of any applied conservatism can be reduced drastically if reliable and sufficient test results are available. 2. Examples are bleeding. 2. pre-construction testing in not only a safety but also an economic issue.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) • STAGE II: main testing program before construction commences to refine the results of STAGE I with aggregates and cementitious material which are expected to be used for construction STAGE III: quality control during construction. Thus. creep. as required. fly ash etc. march 2008 Section 2-3 . Only one component shall be changed at the time and its influence on fresh and hardened concrete properties being tested. wedge splitting tests.2. • Concrete testing of existing dams will be discussed in Section 2. Tab. As submitted for ICOLD review.1 shows an example of a STAGE II test program for hardened concrete properties. The results should allow the following: • • • • Obtain required strength values to satisfy the result of the static/dynamic stress/stability analysis Assess the maximum content of blended material (pozzolans.) Obtain satisfactory workability and permeability Assess the benefit of admixtures.2.1. It is rarely a benefit to curtail pre-construction programs for economic reasons. Alkali-Aggregate Reaction (AAR). dynamic properties.2 Testing program The general set-up of test programs comprises the variation of the most significant concrete constituents and may be grouped in several parts: • • • • • • • Variation of cement content/content of cementitious material Variation of type of cementitious material and its blending Variation of borrow area Variation of aggregate grading and influence of wet-screening Investigation about minimum content of cementitious material Investigations with different admixtures to investigate required water-cement ratios and workability Special testing.2.

grading curve and MSA. air entraining agent. As submitted for ICOLD review.1 – Example of test Program for hardened concrete properties Mass Concrete Testing Part 1 Variation of Components Variation of cement content 150 kg/m3 175 kg/m3 200kg/m3 225kg/m3 Variation of cement type (if an option) 175 kg/m3 225kg/m3 Variation of blending (cement+puzzolans.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) Tab. march 2008 Section 2-4 . quarried aggregates vs. workability (W/C ratio is then a result of acceptable workability). Therefore this strength value migth support to reduce cement content 2) as a measure for general durability for facing concrete also at sites without temperatures below zero. StrengthPermeability Freezing and Compressive Strength [days] 1) 2) Thawing 7 28 90 180 360 28 90 • • • • • • • • • • • • • • • • • • • • • • • • Comments Constant: aggregate source. river deposits 4 • • • • • • • • • • • • • • • • 5 • • • • • • • • • • • • • • • • • • • • Comments: The table is an example. Important is that preconstruction testing is done well ahead of construction to profit from the knowledge of strength development with time and thus prepare a reliable mix design. 2. The scope depends on the variables under discussion. fly 120+30=150 kg/m3 100+50=150 kg/m3 160+40=200kg/m3 140+60=200kg/m3 Variation of borrow area 150 kg/m3 225kg/m3 Minimum content of 3) cementitious material 100+30=130 kg/m3 75+75=150 kg/m3 Facing concrete: 250kg/m3 borrow area 1 borrow area 2 4) Summary of Investigations on Hardened Concrete Density Flex.g. 1)nominal strength for large dams can be extended to the age of 1 year. Pre-testing with different admixtures • • • • • • • • • • 2a • • • • • • • • • • • • • • • • • • 2b • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • can replace Part 1 if the use of blended material is preferable from the onset 3 • • • • • • • • • • • • • • • • e. 3) if reduction of cementitious material seems possible based on results of 2b.

E. 2. cubes or the broken halves of prism from flexural strength testing. march 2008 Section 2-5 .2. It is the relation between these two values.1]. which merits particular attention.2] [2. As submitted for ICOLD review. is a value which (a) has a low variability. So. Common cylinder sizes are 150x300mm (Φ=150 mm and h=300 mm). why then concentrate on tests to obtain E? The literature provides ample values and ways to predict reliable moduli [2.1. 300mm cubes and 75x75x300mm prisms. Tab. Smaller samples sizes than mentioned above would result in unduly influence of the large aggregate sizes on the test result. Reliability on structural safety depends on obtaining satisfactory strength in the dam relating it to laboratory test results. 2.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) When setting up test programs any investigator should be flexible enough to scrutinize his/her needs: what is the impact of a particular concrete parameter on costs and safety? . It is not reasonable to set up tests for parameters of marginal significance just because they are needed for analysis or it is common practice to test them.2 lists some conversion factors for common sample sizes used for dam concrete testing [2. a change of E by multiples of its base value will cause a proportionally minor change in stresses. Standard test for compressive strength use cylinders.3].and to define the scope of the program accordingly. Lets take an example. Even in highly stressed arch dams. The modulus of elasticity of concrete. The ensuing bias of strength values has to be given special attention. In Austria samples sizes for compressive strength are 30 cm cubes and for flexural and pure tensile testing 20x20x60cm prisms are used. and (b) is only marginally important for the stress level in a dam.3 Sample Sizes Typical for dam concrete is the large Maximum Size Aggregate (MSA) with respect to commonly used sample sizes for strength testing.

They are easier to test because they need no capping or grinding.8 f 30 f15/30=0.86.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) Tab.2] [8] [8] 1) for a range of 25 to 35 MPa with lower relative strength for the lower absolute compressive strength Reference sample size: 6"x12"cylinder (15x30cm)=100% 3"x6" cylinder 8"x16"cylinder 12"x24" cylinder 18"x36" cylinder f3/6=1. Japan.2 – Conversion Factors for Different Sample Sizes Reference sample size: 30cm cubes=100% Conversion to other sizes Size 20cm cubes 15x30cm (6"x12") cylinders 15x30cm (6"x12") cylinders 30x45cm (12"x18") cylinders Conversion formula for MPa f20=1. The use of cubes has the advantage that no grinding and capping of the surfaces is necessary and that the scatter in strength values is generally less due to less imperfections in the production of cubes.f 6/12 f12/24=0.6) Relative strength % 110 80 78 to 83 78 to 82 1) 1) Ref.f 6/12 f8/16=0. In Switzerland some testing for dam concrete used 40 cm cubes which are then sawed into 8 cubes with size of 20 cm for testing. Canada. Cylinders are the common sample form in the US.2) f30/45=0.f 6/12 106 96 91 86 [2] [2] [2.(f 30 -3.98. it is a more cumbersome procedure. However.(f 30 -5. France.06.f 6/12 f18/36=0. Benefits and shortcomings of cylinders vs.96 . On the other hand. This is considered as very meaningful. Cubes are generally used in Continental Europe (except France) and UK. ISO 2736 [2. As submitted for ICOLD review. Latin America and Australia. 2.91. cylinders are generally preferred in research because they better represent true strength due to less influence of stress singularities from sample surfaces. cubes are widely discussed in the literature. Another advantage is that cubes can be tested parallel and perpendicular to the direction of casting to indicate any difference due to casting direction.91. It allows using the full mix and has the advantage to get 8 specimens for the mean strength value of the full mix and the sample-to sample scatter is small. march 2008 Section 2-6 .1 f 30 f15/30=0. It eliminates the conversion from wet-screened to full mix.3] [2] [2] The commonly required ratio between smaller dimension of the specimen and maximum size aggregate (MSA) in excess of 3 also applies for mass concrete with large size aggregates.

77 Results from wet-screened laboratory testing are therefore non-conservative.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 2. Correlation tests between the strength of wet screened concrete tested at 28 days and test specimens containing the full mass concrete are recommended to be part of a STAGE II testing program (section 2.3 – Compressive strength ratio between full and wet – screened mix [2.2.65 to 0. 2. march 2008 Section 2-7 . Age at testing the wet-screened samples in days 28 90 180 Compressive strength ratio full mix (150 mm)/ wet-screened mix (38 mm) 0.65 to 0.8 0. 2.4]. Similar results are obtained by Portuguese research with a larger scatter of test results due to comparing full mix cylinders with wet-screened cubes [2.1).1.4].75 0.3 reports a range of compressive strength ratio for two types of concrete.1.2. The US Bureau of Reclamation in its Designation 33 reports conversion factors for wet-screening of the 38 mm to 150 mm sizes which are depending on age.7 to 0.4]. As submitted for ICOLD review. Tab. the lower values for blended and the higher values for unblended cement [2. Wet screened concrete needs conversion factors in order to assess the full-mix strength.4 Wet screening Wet-screening is a common measure to overcome the discrepancy between a mix with full MSA and dimensions of common sample sizes [2. Tab. This because the strength test for full size concrete is actually a large-scale test and it may not be possible to do it on site (STAGE III).5].

Among those are: • • • to evaluate strength gain at ages in excess of sample age in laboratory testing to compare laboratory strength with in-situ strength to evaluate loss of strength from ageing.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 2. The latter is more commonly used.2 Testing of concrete strength in existing dams 2.2. different sample size and MSA (Maximum Size of Aggregate). multiaxial). By means of tomography techniques it is possible to obtain a detailed map of elastic wave velocities inside the sections of the dam and to evaluate the elastic dynamic characteristics and locally the state of integrity of the concrete. 2. In Italy this technique is widely used and several concrete dams have been checked by means of the sonic tomography [2. A fine summary of implications with core extraction and testing is also reported in [2. It is therefore not surprising that laboratory strength is different than strength in the completed structure.g. different state of loading (stress free vs. The extensive French experience is summarized in a comprehensive report by LERM [2. They are based on the principle that the characteristics of the impulsive sonic signal transmitted across the concrete can be related to the elastic properties and soundness of concrete itself. different initial stress implications and different curing.2 Semi-destructive or non-destructive methods There is a whole range of semi-destructive or non-destructive methods of testing concrete strength in situ. e. demonstrated by unstable cracking or movements. with rebound hammer.9] As submitted for ICOLD review. measuring the velocity of sonic or ultrasonic pulses through concrete. • Testing is generally either by non-destructive methods or by core drilling.6]. or chemical alteration (AAR or sulphate attack). or loss by stress concentrations in certain parts of the dam. march 2008 Section 2-8 . Core testing for highly stressed dams is justified.7] by Adam Neville. with frequencies of 5-30 kHz are generated and received by special electro-acoustic transducers located inside or on the surface of the dam. The investigation involves taking sonic logging measurements along boreholes and direct velocity measurements between pairs of adjacent holes or surfaces.2. Non-destructive sonic methods can be used for diagnostic investigations on concrete in existing dams. What is essential is to be aware and interpret these differences reasonably.2. pull-out tests. This has to be accepted. In practice trains of elastic waves.1 General Testing of concrete strength on aged dams during operation may have several objectives. to evaluate long-term creep.8]. Consider the many differences between laboratory and field conditions: different compaction.2.2. break-off tests and penetration tests [2.

with the 200 mm cores only for smaller MSA. Fig.Sonic tomography of a dam – Map for sonic velocities (1) and for concrete compression strengths (2) [2. on traditional destructive compressive tests on samples cored from the structure.10]. Smaller than 200 mm cores should not be used in any case. the ratio between core diameter and MSA (Maximum Size of Aggregate) should be close to 3 [2.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) It is also possible to obtain a correlation between velocity and concrete strength. As submitted for ICOLD review.1). the number of samples for calculating a mean strength should be increased.3 Ratio between core diameter and MSA Similar to the condition for laboratory test cylinders.2. say less than 80 mm. 2. march 2008 Section 2-9 . 2.1 . 2. 200 to 300 mm cores are generally used drilled from dam galleries or downstream surfaces. In this case a detailed map of concrete strength can be derived (Fig. When core drilling is known to be done it is useful to mark locations between reinforcing bars in the dam galleries already during construction. for each dam investigated.2.9]. When the desired ratio of 3 in not maintained. But it should be based.

For example.94 0.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 2.87 BS 1881 1. disturbed from the drilling process.2.11]. 2. reducing the diameter from 150 mm to 75 mm would result in a 50% reduction of tested strength (results in [2.00 ASTM C42(1990) 1. 2. However. But.4 Length-to-diameter ratio Strength from cores depends on the length-to-diameter ratio (l/d).98 0. cores from dams are occasionally difficult to extract with this ratio. The core diameter has a significant effect on strength.00 0.00 1.93 0. 2.2. The larger the size of the core. the influence of skin effects (particularly the disturbed bonding by cutting through coarse aggregates) governs the strength reduction with decreasing core diameter.96 0. This is further researched in [2.2.92 Already the comparison of correction factors for low l/d ratios between these two standards indicate that for l/d-ratios approaching unity the uncertainty of strength prediction increases rapidly. As submitted for ICOLD review.75 1. This is not recommended for dam concrete although sometimes smaller diameter holes are used because of lack of equipment.2. The reason is that the thickness of the surface zone. Conversion factors have then to applied for assessing the length-to-diameter influence: Conversion factors as suggested by ASTM and British Standards are (Tab.00 0.4 – Conversion factors suggested by ASTM and British Standards l/d-ratio 2.5.98 0. 2.10].50 1. Thus.5 Core size Another influence comes from the core size if smaller than 300 mm drillholes are used. the higher is the compressive strength [2.2.25 1. unlike molded cylinders. is constant and thus becomes increasingly influential at small core diameters1.4): Tab. It is therefore difficult to conclude reasonably on the real strength from small diameter core sampling because of the joint influence of the small core diameter and the heterogeneous concrete structure dominated by the presence or not of coarse aggregates in the core.96 0.11] from a 19 mm MSA concrete). it is recommended not to use cylinders with l/d<1. 1 This finding seems to contradict the inverse relation between diameter of test cylinder and strength as shown in Tab. march 2008 Section 2-10 . Standard cylinders have a length-to diameter ratio of 2.

strength from cores. i. and (b) the influence of field-cured cores with less gain at high ages as one would obtain from laboratory test results.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 2.2). Moreover. Fig. The two series have different gradients indicating the superposition of two influences. with excessive bleeding of mixing water. Another reason has its origin in the fact that dry cores (not even oven dried) show higher strength than cores from wet concrete due to the influence of pore pressure. 2. the scale effect is larger in small than in large cores. However. the scatter is larger. core strength higher than laboratory strength. The above range of depends on quality of drilling.2.3] As submitted for ICOLD review. curing and age. where the reverse relation.2.3] (Fig. 10 to 15% higher strength values might be expected.2 – Development of compressive strength: laboratory vs. is found for dam concrete. there are rare test cases. march 2008 Section 2-11 . i. if cores are extracted outside the saturated upstream zone of a dam.e. field sampling [2. which then reduces the w/c ratio leading to higher core strength. It is reported that such reverse relations result from overvibrated concrete. Thus. laboratory strength Of interest is to compare laboratory compressive strength with specimen of the same geometrical characteristics and maturity has generally a lower strength than laboratory made and cured to 30%). The difference in dry/wet strength has then to be modified by the influence of a scale effect. 2.6 Core vs. namely (a) the known decreasing rate in gain of strength with increasing age (here demonstrated by the kink at 90 days). USA [2. A drilled from a structure concrete specimen (10 10 to 30% very much These influences are corroborated by an interesting observation comparing laboratory with field test series form Shasta Dam.e.

For dam concrete. Otherwise the core strength tends to be unrealistically low. As submitted for ICOLD review. 2.16].2. 2. where a damaged matrix can expand and thus reduce core strength. This is important because the difference in compressive core strength between the 2 curing methods can be 20 to 35%.12] and EN 12504 [2.3 Nominal age of concrete strength In large hydraulic structures. at least for large. if curing is done properly.20]. the correlation factors at 1 year may range as shown on Tab. 2. This last one is defined in the following section 2.5. If the concrete to be tested is assumed to be in a steady-state dry condition then the cores should not be wet-cured.5 and Fig. Other comparative test results for structural concrete (28 days) demonstrate 90% to 76% for core as compared to laboratory strength [2. march 2008 Section 2-12 .18]. as occasionally specified for dam concrete. soaking is more appropriate for a saturated concrete mass and it should be long enough to eliminate moisture gradients in the sample (curbing the influence of pore water pressure). especially where pozzolans are applied. which stipulates that field-cured test cylinders should achieve at least 85% of the strength from laboratory-cured cylinders [2. Unless unusual circumstances prevail. highly stressed dams. Percentages depend on stress level: it decreases from close to 100% for low strength (∼20 MPa) to 70% for high strength concrete (∼60 MPa) [2. with the air dried being the larger of the two values [2. it should be mentioned that checking the compressive strength from drilled dam cores is only successful if drilling is done with much care. ACI 207 reports that design strength is on ages between 90 days and sometimes up to two years [2. and if pre-test conditions are representative and realistically interpreted. However. So can ACI 318. structural elements are rarely required to withstand substantial stress at early age.13] can be used as a reference for drilled cores.15]. This can be particularly influential in Alkali-Aggregate Reaction (AAR)-damaged concrete.17].3 from [2. In California 1 year is accepted too [2. the Committee recommends an age in excess of 180 days as basis for evaluating the characteristic strength. For most dams at least 180 days seem reasonable. Using cored strength values in a back-analysis would provide more reliable safety margins. It is utterly conservative to use 28 or even 90 days for assessing strength safety margins.14].19]. ASTM C 42 [2. ACI 318 defines two field curing conditions depending on the expected condition in the structure: one is air drying. The above dichotomy leads recommending to core dam concrete as a rule and not only if unsatisfactory laboratory tests are obtained.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) Another factor contributing to the difference between core and laboratory strength is the relief of the multiaxial stress condition after coring. the other soaking of the specimens. Accounting for the continued strength development beyond 90 days.2.

3 24.34 1.0 37.7 27 28.8 17.0 28.06 1.1 24.0 58.09 Legend: 1) none….6 22.8 30. As submitted for ICOLD review.3 32.3 32 29.8 45.5 26.12 1.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) Tab. shale fly ash none fly ash fly ash fly ash fly ash none none none pozzolan none 33% 25% 25% 30% 30% 34% 22.8 31.53 1. The mean strength values are generally from construction and the mean of a large sample population.0 33.9 Strength ratio 365/90 1.5 22.31 1. 1) Dam/Country Dworshak/USA Glen Canyon/USA Las Portas/Spain Valparaiso/Spain Llauset/Spain Flaming Gorge/USA Puylaurent/France Hoover/USA Libby/USA Baserca/Spain Yellowtail/USA Katse/Lesotho Grand Coulee/USA Sambuco/Switzerland Morrow Point/USA Val Gallina/Italy Schlegeis/Austria Grand Dixence/Switzerland Year 1972 1963 1974 1988 1962 1995 1935 1972 1965 1995 1942 1954 1967 1950 1971 1955 II CEM III IV II II II II + IV II II IP CEM III IV OPC II II Blending Pozzolan Quantity fly ash 25% Pumicite fly ash fly ash fly ash calc. 2.2 41.3 24. march 2008 Section 2-13 .6 36.39 1.17 1. Ordinary Portland cement (OPC) only.26 1.6 38.5 21.7 20. of indicated ASTM or CEM type quantity is % of total cementitious material.32 1.6 21..5 – Gain on Compressive Strength with Age [2.16 1.1 47.0 29.30 1.9 42.20].23 1.1 52.29 1.4 42.6 31.12 1.10 1.5-30% 25-30% 25% 34% 40% Mean compressive strength [MPa] 90 days 365 days 14.4 47.4 46.35 1.7 41.2 35.31 1.

As submitted for ICOLD review. march 2008 Section 2-14 .3 – Gain of compressive strength with age [2.20].ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) Fig. 2.

1 Valle di Lei Strength 90 days [MPa] 15.2 Strength ratio 180/90 1.06 1.5 20.2 14.06 1.7: Tab.7 Strength 180 days [MPa] 16.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) Tab.1 28.7 – Core strength from Italian dams [2.6 15.0 From Italian dams.8 44.3 46.29 1.2 13.0 20. march 2008 Section 2-15 .0 40.6 – Core strength from Austrian dams [2.06 1. 2.09 1. more readily available [2. Some are reported on Tab.4 31.21] Dam Age of core in years Schlegeis Drossen Mooser Limberg 12 26 26 32 Core strength (φ 200 mm) [MPa] 32.3 41. 2.06 As submitted for ICOLD review.1 44.6 shows long-term core strengths from Austrian dams: Tab.7 44.1 Suviana Goillet Santa Giustina Pian Telessio Campo Moro n.4 40.3 41.5 90 day’s cube strength [MPa] 20.22] are ratios between 90 and 180 days strength.22] Dam Cignana n. 2.06 1.7 37. 2.6 30.4 20.

production. The factors contributing to variability are concrete constituents. the steel platens and the concrete specimen. with the consequent effect on strength. 3951 [2. As mentioned. in a soft machine it will follow without delay. 2. CEB – FIP Model Code [2. In a stress-strain diagram this will result in lower peak stress for a soft as compared to a hard machine. The consequence is that in a soft machine the energy stored in the machine is released and acts on the specimen when it starts to fail.2.4]. for dams the characteristic age should be 180 days or.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 2. The reader is encouraged to consult the numerous publications. The effect is more pronounced for specimens with lower l/d-ratios and it is also the reason why cube strength is higher than cylinder strength. as for example set by the Japanese Industrial Standards (JIS). results in lateral restraint forces in the concrete near the platen. ACI 214 [2. They are not discussed here. the compressive load induces lateral tensile strains in both. • One is related to the strength gain with age. The literature of concrete testing has therefore developed a host of theories and methodologies to evaluate scatter of test results and its application for dam concrete.2. Test machines should therefore be low in frictional resistance between platen and specimen.27] and reference [2. testing and ageing. During the test. for larger dams (longer construction). The mismatch between the elastic modules of steel and concrete and the friction between the two.24]. These machines are classified as hard (rigid) and soft (less rigid) depending how the head of the machine is following the deformation of the specimen. march 2008 . This additional energy will cause more extensive crack propagation and indicate failure at lower loads than with a hard machine.25] and 5479 [2.5 Evaluation of Strength Testing Strength parameters of a composite material like concrete can never be defined without uncertainty.28]. demonstrated by a shear-type failure pattern. due to Poisson effect. The concrete specimen is therefore locally in a triaxial stress state. Furthermore. Strength follows a statistical distribution and is not a deterministic value. the corresponding ISO Standards 2859 [2. one year. This end-effect results in a higher strength than for specimen which are exposed to pure compression (failure pattern § parallel cracks).26]. This restrains the lateral end-expansion of the specimens and introduces shear stresses. Loading range of the testing machine shall not exceed 90 % of the maximum range capacity applied during the verification test. This means that one-year strength data should be available when concrete Section 2-16 As submitted for ICOLD review. Some particularities with respect to dam concrete testing are added. in order to improve the reliability of measured values a minimum level of 20% is preferable.23]. among them the Bureau's of Reclamation Concrete Manual [2.4 Test Equipment The type of the testing machine (press) may have an effect on the obtained results. In a hard machine the head has difficulties to follow rapid deformations. Most compressive test machines in field laboratories are not free from friction between specimen and platen. This effect is larger for cubes and smaller for cylinders or prisms.

23]: fm = fc + k.σ the mean required strength for which the mix has to be fm designed fc the characteristic strength (specified design strength to meet required safety against stresses) k a factor derived from the assumed strength distribution which depends on a stipulated proportion of tests (fractile) to fall below the level of fc. which is uneconomic and. and on the number of tests carried out. In order to overcome the impasse. and shear/tensile strength is generally taken as a percentage of compressive strength. as aside effect. As submitted for ICOLD review. march 2008 Section 2-17 . 2. taking into account the scale of the structure and the times of construction and service life.g. and preliminary series with similar than the final mix design from the dam in question provide a host of strength-gain values to judge strength gain up to the age of the characteristic strength. contributes to unfavourable thermal conditions. generally n≥30) are reported in Tab. What this all means is that engineering judgement and peer review. mostly along local zones. Certainly. such a judgement has to be accompanied by assessing the consequences of an anticipated error between assessed and later measured characteristic strength: What. Research. However the characteristic age have to be always related to the specific dam under examination. • Evaluating strength data. discipline and is often not done because of unknown final concrete constituents. based on an increasing data-base. published data from case histories. it is often decided to take 28/90 day as characteristic strength or to conservatively extrapolate these early strength data. Commonly used k values (for high number of tests.8. However. σ standard deviation of strength test series The k-factor is derived from fractiles of a Normal distribution. This Bulletin would like to encourage the dam profession to also view the problem from a more pragmatic standpoint. This can be beneficial for economy and safety. The shortcut of immediate conservative solutions is not necessarily the only safe way to go. e. Of course. is based on the following formula [2. not measured per se. This percentage is highly speculative (particularly for dynamic loading) and uncertain in a similar order of magnitude than speculating a particular strength gain between 90 days and one year. during extreme temperatures. often is a remedy to justify acceptance of uncertainties. This is not always possible. the obvious remedy is to start early with concrete testing. In most cases. this needs planning. floods or an earthquake? In many cases the limiting strength is not compression but shear or tension.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) placement starts. the most common design criterion.8 also acceptable? Can a local redistribution of stresses be assumed beneficial in mitigating a lacking safety margin? How much is the gain of strength between one year and the age when the calculated stresses are really occur. such decisions lead to excessive cement content. if the required strength is not met? Are the stipulated factors of safety (FS) still acceptable? But: isn't FS=3 just an arbitrary chosen number and FS=2.

should succeed to finding information about test conditions. and thus on safety factors. • 2.6 Generic Uncertainty in Test Procedures As mentioned previously. 2.2. is considerable [2.33 When comparing laboratory with core strength. An example is the water-cement ratio (w/c ratio). • For statistical evaluation bounded or curtailed distributions (instead of an unbounded Normal distribution) are actually the correct choice because they represent more realistically the physics of statistical scatter. As submitted for ICOLD review. details of the mix design. resorting to test documents.8 – Examples of k values Percentage of tests falling outside the limits ± k. In a similar way goes the decision about the impact of parameters. This is unfortunate.4) and. which are cumbersome to be tested or which are commonly not tested at all although their influence of strength is significant. 2. in the case of concrete strength evaluation from cores of existing structures.29] as the influence on reliability indices. the variability of test results and the like. the choice of k should be different for both types of tests because of the larger inherent variability of the latter.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) Tab.σ σ 20 10 5 2 Percentage of tests falling below fc (fractile in %) 10 5 2. because its share on the standard deviation for compressive strength is considerable (Fig. Dams may show signs of distress years and decades after construction.28 1. in setting up test programs one needs to first assess the impact of a parameter on safety and its costs of testing and then decide if testing is meaningful or if it suffices to rely on published values. fallen into oblivion. for good reasons.65 1. march 2008 Section 2-18 . not measured. the scatter of compressive strength could certainly be reduced [2. were it measured.5 1 Chances of tests falling below the lower limit 1 in 10 1 in 20 1 in 40 1 in 100 k 1. sources of constituents. Then. Although most codes still refer to Normal distributions there is a tendency for a change [2.96 2.31]. number of samples. Reporting and archiving of test results is essential.30]. In mass concrete field testing the w/c ratio is. The procedure have also to be adapted.

with the impediment of high scatter if n is small. 2. there is the dichotomy of simple tests with high repeatability as against complicated tests with high dependability. Certainly. Any tested strength value has therefore to be recognized as an interpreted estimate of the unknown true strength. As submitted for ICOLD review. 2.5). in MPa Share of cement variation 3. if carefully executed more closely approaches reality.4 .0 Share of test variation 4. The latter.0 10 20 30 Compressive strength in MPa 40 50 Fig.0 Share of variation of W/C ratio 2. It is therefore suggested to incorporate tensile testing in the technical specifications. the second are delicate to execute and evaluate. The first have their limitations in the declining statistical effectiveness with increasing sample size (Fig. however. dev.0 0.31] Then. the conflict ensues either having to deal with a high uncertainty of the mean value (formula-based assessment) or with a high scatter around a more reliable mean (tensile testing).Shares of standard deviation during testing of compressive strength: above without and below with control of w/c ratio [2.0 Std. the latter is preferable. Here.0 1. An example is the decision either to assess tensile strength from formulas relating tensile to compressive strength or to rely on the more delicate (pure) tensile strength testing. What this all means is that the true strength in the dam is a random variable affected by many influencing factors. march 2008 Section 2-19 .ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 5. conditions and testing traditions.

advantage is taken to reduce the w/c-ratio by the use of water-reducing and plasticizing admixtures.3. Concrete dams are designed to carry compressive stresses and to minimize tensile stresses. Porosity and degree of hydration have a fundamental relation to strength of concrete. because this zone (about 50 µm wide) is the weakest link in the matrix. For dam concrete w/c ratios are rarely specified and are predominantly determined experimentally to achieve satisfactory workability.1 Effects of porosity and w/c-ratio on concrete strength The strength of ordinary concrete is governed by the porosity of the cement paste. Hence.2 0. 2.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 1.0 Multiplier of std. deviation. Typical relations between the water/cement ratios for an age of 28 days. 2. but they are not easy to measure. using several theoretical formula and empirical evaluations are shown in Fig. as cracking initiates within it.6. m 99% confidence 0. In this context. these parameters are replaced by the free water/cement ratio and the age of concrete when relating them to strength measurement. its placement and curing.m m= t n ft … strength of corresponding confidence lim σ … std. deviation t … constant depending on the required confidence limit n … number of tests/samples m … multiplier of std. for practical engineering applications.5 .4 0. march 2008 Section 2-20 .0 0 10 20 30 40 50 Number of tests.Reduction of uncertainty with increasing numbers of sampling (inverse proportional to √n) 2. which influences strength. It is primarily the porosity of the transition zone between the cement paste and the aggregate. which in turn depends on the water/cement ratio and on the degree of hydration.8 95% ft = fmean+σ. As discussed in the following. 2. compressive strength is influenced by a wealth of conditions of fresh concrete. deviation 0. n Fig.6 50%(mean) 90% € 85% 0.3 COMPRESSIVE STRENGTH Compressive strength is the predominant property for dam concrete. The As submitted for ICOLD review.

4 0. by means of its pozzolanic activity index) [2.34] with 200kg/m 3 cement and 4% air 20 10 0 0. plasticizing or superplasticizing agents as a potential means to reduce the water content in the mix. Besides the basic goal of higher strength (at equal content of cementitious material). a reduced peak temperature and thus less thermal cracking.g.3 0.[2.7 0. march 2008 Section 2-21 .6(72-60. 2.6(84-72.w/c Swiss 1: f c = 0.8 0. Ref.w/c). Abrams 70 Swiss 1 Swiss 2 PCAmax PCAmin ACI 211 Popovics 60 Constant: Age 28 days Sample size 6"x12" (15x30cm) cylinders Air entrained concrete OPC (no blended cement) 50 40 30 Abrams: f c = 100/4 1. As submitted for ICOLD review.5 .5 0.w/c). Ref. or the like. 2. whereby the higher w/c ratios are typical for older dams where admixtures are limited to an air entraining agent. slag.32] Swiss 2: f c = 0. less shrinkage and thus less shrinkage cracking and. [2. however appropriate because it makes sure that the denominator also includes all cementitious ingredients such as fly ash. if ice is used for cooling.[2.9 Fig.33] Popovics: equ.6 can be interpreted in this sense: the range with lower w/c ratios is mainly for concrete in more recent dams where admixtures are more widely used. a clumsy term. it is now common practice to use the term "water-cementitious material ratio.35].32] PCA(max/min) see Ref. pozzolans.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) curves in Fig. these are a reduced permeability and thus higher durability (less porosity of the hydrated cement paste). French standards use the expression "liant équivalent" and even provide a factored coefficient for “cm” which considers the amount of strength participation of cementitious ingredients (e. [2. To lower the water content has several advantages. (w/cm)".6 w/c ratio 0. One needs to recognise the beneficial effect of water-reducing.14 in [2. which participate in strength development together with Ordinary Portland Cement OPC.6 – Compressive Strength vs water/cement ratio.20] ACI see Ref. With respect to the denominator of the w/c-ratio.

only the lower third of Fig. For modern dams. To achieve the greatest cement efficiency. 3 say.7 and the large scatter band of the Bureau's test series (not shown).ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 2. Larger or lower sizes are the exception for large dams. The results in [2.36] are from test series for Grand Coulee and Clear Creek dams by the Bureau of Reclamation (USA). an optimum maximum size for each compressive strength level has to be obtained with a given MSA and cement content. 2. march 2008 Section 2-22 . The larger the aggregate size the less cement is required for a given volume of concrete to produce enough cement paste necessary to cover the aggregate surface. 2. as shown in Fig.7 is applicable because cement content rarely exceeds. Other investigations confirm the low gradients for cement contents below 250 kg/m3 [2.37]. the shown relations may differ from dam to dam.7. Fig.2 Effects of maximum size of aggregate (MSA) For the main mass of interior concrete a MSA between 120 mm and 150 mm is commonly used (80 mm for older dams).36] As submitted for ICOLD review. 250 kg/m . however. Given the low gradient of the curves in the lower part of Fig. 2.7 – Effect on Maximum Size Aggregates on Cement Content [2.3. This rather small range is the result of a long lasting experience balancing to reduce cement content and to achieve a good workability. 2.

Lets take an example: reducing the MSA from 6" to 3" for a 200 kg/m cement content would – according to Fig.8 – Compressive strength vs type of aggregates [2. rocks with high mica content. even more. As submitted for ICOLD review.39].05 in the w/c-ratio for equal workability which.3 Effects of aggregate properties Compressive strength is governed by the strength of the aggregates and their water requirement for satisfactory workability. Austrian experience is that the content of schist in the sand fraction (0 to 4mm) should be kept low. i.e.reduce the strength by roughly 5 MPa. say less than 5 %.38] Essential for the acceptance of marginal aggregates is an extensive test program of both the aggregate parameters and then the concrete. This influence can be large. 2. The difference between.3. say sandstone and limestone as source of aggregates can be in the same range. 3 Fig. march 2008 Section 2-23 . A common detrimental constituent in the aggregates is schist. Due to the fact that aggregates of marginal qualification are used now more frequently. to know the influence of aggregate quality becomes increasingly important.g. in turn. as shown in the examples of Fig.7 . 2. Aggregates with 15% content of schist for the concrete of the Kölnbrein dam resulted in an 18% decrease of compressive strength as compared to aggregates free of schist [2. 2. Experiments demonstrated that an increase of schist by 10% in the sand required an increase of 0.8. reduced the compressive strength by 12%. e. for small gravity dams with a low stress level.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 2.

0° C+40% RH) for 1 day and fog room curing for 7 days thereafter (23. in ACI 308 [2. For the Zillergründl dam in Austria. Curing requirements are treated in ICOLD Bulletins n° 47 (Quality Control of Concrete).40] or [2. Curing of test specimens is well defined in standards. 2. prone to thermal cracking. strict adherence to curing as stipulated in standardized test methods is mandatory in order not to add another source of uncertainty. march 2008 Section 2-24 . Reference [2. the higher loss being for blended (fly ash) concrete. or after removing the formwork. Lift joints are discontinuities in the concrete mass and they are exposed to high temperature difference.3.g.4] and [2.41]). Such test would seem to simulate inadequate dam curing. a curing compound was successfully sprayed on the lift surfaces.41] reports of test results on 100x200mm cylinders comparing: • • exposure to environmental conditions (23. its weakest links. n° 76 (Conventional Methods in Dam Construction) and a recent Bulletin under publication on “The Specification and Quality Control of Concrete for Dams”.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 2. Curing at site is unfortunately often taken less serious than laboratory curing. which is still labour intensive and therefore prone to lack of care. its stipulated quality and uniformity is at risk being impaired by deficient curing. The loss of 28-days compressive strength for 7 types of concrete (with Ordinary Portland Cement and blended cements) between the two curing conditions was between 10% and 40%. Dry dam and block surfaces are a common view at construction sites although simple sprinkler devices (such as hoses with holes) are easy to install and to operate. It is essential for large dam projects that the contractor provides a fairly uniform product with a specified strength of reasonable scatter as confirmed by periodic testing. immediately after pouring. This also enhances adhesion between lifts. A loss of compressive strength also means a loss of shear and tensile strength at lift joints.4 Effects of curing Here we have to distinguish between curing of test specimens and curing during dam construction.0° C+95% RH). Therefore a loss of strength at this particular location can be particularly detrimental for structural integrity and water tightness. Interrupted curing will largely affect the compressive strength development as indicated by the test results of Fig.4]. e. Given the many factors influencing strength results. what counts is that the impact of inadequate curing hits the dam at the joints. This Section therefore only covers some additional aspects to what is said in the above two Bulletins. However. However. as soon as the product of this mechanized process leaves the mixer and is placed. One may argue now that for dam concrete such a loss of strength is less significant than for structural concrete due to its limited impact on a large mass. and occasionally also exposed to uplift. As submitted for ICOLD review.9 (from [2. and exposure to environmental conditions throughout (no fog room exposure). generally caused by intermittent or no moistening of lift and block surfaces. It is a pure matter of discipline and site supervision to overcome this simple problem.

march 2008 Section 2-25 .3.5 Effect of mineral admixtures and blended cements Blending of Ordinary Portland Cement (OPC) with pozzolans. This increases the capacity for plastic deformation with the advantage of increased tensile strain. It is mainly governed by economic aspects and by the extent to which the blended material is able to contribute to strength development. and the reduced risk for AlkaliAggregate Reaction (AAR). The disadvantage of slower early strength development is not essential for dam concrete. the main benefits from blending is a reduced heat of hydration. The amount of blending proportions differs largely. 2. Another advantage comes from the increased paste content when using pozzolans or other fine-grained additions. Tab. 2. fly ash or slag and other mineral admixtures are very common for dam concrete because of their generally economic and performance advantages.4] 2. Apart from economy. however more pronounced if intergrinding the natural pozzolans with Portland cement clinker and not mixing the two compounds at the site.9 – Compressive strength development for different conditions of curing [2. As submitted for ICOLD review.9 lists examples of dams with use of blended cement.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) Fig. This is. particularly as this strength is recuperated at later ages. improved workability due to the finer gradation.

lower strengths are recorded than in the case of standard test. The breakdown of concrete consists in the formation of cracks. Furthermore using blended cements reduces the heat of hydration. a higher strength is recorded.5 to 30 30 30 fly ash fly ash pozzolans pozzolans pozzolans fly ash fly ash fly ash pozzolans fly ash calcinated diatom.11 demonstrates this phenomenon impressively [2. 2. Under the low rates of loading.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) Tab. The process of cracking during loading explains this behaviour. march 2008 Section 2-26 . which does not exceed 70 to 80 % of the short-term strength. the load is applied more rapidly.42] shows an example of a test series resulting in the choice of 30% fly ash for this particular dam concrete. Fig. clay fly ash fly ash fly ash fly ash fly ash pozzolans pozzolans fly ash fly ash fly ash pozzolans pozzolans Blended cements need more testing (a) because of the properties of pozzolans may vary widely and (b) because the quantity of blended materials has to be optimized with respect to strength and durability.10 [2. or their extension without arrest (unstable crack propagation at high stress levels). The well-known Rüsch-diagram in Fig. either their extension with arrest (stable cracking at low stress levels). creep occurs. however only at a degree of loading.43]. the compressive strength is determined at loading in the range of minutes. 2. 2. on the other hand. As submitted for ICOLD review. As this range of minute increases.3. 2. It flattens the stress-strain curve. If.6 Effect of sustained loading In a standard test.9 – Use of blended cements in selective dams Concrete volume 10 3 m 3 ZillergrŸndl/Austria Katse/Lesotho Luminei/Italy Cancano II/Italy Vajont/Italy Tagokura/Japan Hitotsuse/Japan Itaipu/Brazil Francisco Morazan/Honduras Hungry Horse/USA Monticello/USA Yellowtail/USA Longyangxia/China Ankang/China Ertan/China Puylaurent/France Pieve di Cadore/Italy Valle di Lei/Switzerland Baserca/Spain Llauset/Spain Las Portas/Spain Alcantara/Spain Cedillo/Spain arch arch arch gravity arch gravity arch buttress arch arch-gravity arch-gravity arch-gravity gravity gravity arch-gravity arch arch-gravity arch arch arch arch buttress gravity 1'400 100 520 265 1'950 560 12'000 1'510 2'350 248 1'340 1'587 3'250 4'100 85 377 862 238 211 641 950 800 Content of cementitious material for interior concrete kg/m3 170 180 270 240 250 140 215 130 150-180 165 170 177 160 130-155 174 250 200 230 200 200 200 250 225 Dam/Country Type of dam Type of blended material Blending % 33 30 to 50 25 37 35 25 30 23 25 55 25 28 30 30 to 55 30 40 25 35 to 37 25 25 22.

10 – Trial tests with variable contents of fly ash [2.42] Fig.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) Fig. march 2008 Section 2-27 .11 – Influence of sustained loading on strain according to Rüsch [2.43]. 2. As submitted for ICOLD review. 2.

it is not common to test dam concrete under biaxial or even triaxial conditions and uniaxial test results are used as strength parameters. they are applied to a structure pre-loaded due to self-weight.3. For this kind of loading there are little laboratory results available. These cycles are extremely slow as compared to common laboratory loading with rates of 0. Because of obvious reasons. Other examples are given in [2. 2. march 2008 Section 2-28 .12 – Example for biaxial strength envelope [2.4). Interesting.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) In dams.46]. the results of uniaxial testing enter as corner point in the chosen strength envelope.45] and [2. Such envelopes exist from the literature.44]. 2. In other words. This is an acceptable approach as long as biaxial strength envelopes are used for safety evaluations of compressive stress domains in the dam.7 Multiaxial compressive stress domain in a dam Generally concrete in dams is stressed multiaxially. Moreover. Fig.15 to 0.12 is an example [2. loading cycles are mainly depending on variation of the reservoir level and on annual temperature cycles. in this respect. Fig.44] As submitted for ICOLD review. 2.4 MPa/s. It would be utterly conservative to use uniaxial compressive strength values alone for safety evaluations. is the observation of increased ultimate tensile strain at loading rates which are common for dams (see Section 2.

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 2. Tensile strength is commonly and basically tested in 3 ways: direct (pure) tension.47]).1 Tensile loading in the dam Which type of tensile strength shall be employed for simulating the real conditions in the dam? Is it direct tension. As submitted for ICOLD review. is it splitting strength or is it the result from wedge splitting testing? The choice should relate to the anticipated mode of failure and is therefore not unique. both under static and seismic loading. The following sections try to outline common practices and trends to account for tensile strength. that provides most valuable information about concrete loading under tensile stress (see Appendix A). In spite of the significance of tension. it is not common to test and evaluate it as consistently as compression.13 – Compressive vs flexural strength [2.13 (from [2. It is common to relate tensile to compressive strength as shown by the example of Fig.4. 2.464 to obtain coefficients for MPa. It has been argued that the modulus of rupture gives values that can be directly used 1 multiply the coefficients of the βBZ-formulas by 0. the wedge splitting test. flexural tension (also termed modulus of rupture) and splitting tension (Brazilian test). Fig.47] 1 2.4 TENSILE STRENGTH A critical and important factor in the safety of concrete dams is the tensile strength. is it the modulus of rupture. march 2008 Section 2-29 . Recently a test based on fracture mechanics has been introduced. 2.

In order to relate the flexural to the direct tensile strength. Tab.3].g. on a localized level (conditions at the crack tip).2.4. which allows to respect the ratio between sample size and MSA. 150x300mm Sample size recommended for dam concrete 200x200x600mm or 300x300x900mm Drilled core φ = 75 to 100mm At selective location as dictated by design or monitoring demands Comments ft fr Only to be tested in specialized laboratories fsp Tensile testing show high scatter as initial stresses might have already impaired the concrete's tensile strength potential even before the external load starts to act. it is suggested that typical cracking due to static loading. Fig. with splitting test results as a redundant type of test. rather follows the laboratory condition of direct tension. a more specific consideration for parts of the dam under tensile stress is warranted and attention should be paid to the following: • • The ability to relate any value of tensile strength to the corresponding mode of cracking in the given part of the dam The type of analysis being conducted for which the tensile strength is to be specified (see Section 2.4.1. The As submitted for ICOLD review.2 Type of tensile testing Notwithstanding the above. at the upstream toes of gravity and arch dams. The full mix of the flexural test will exhibit a more elastic strength value (lower modulus of elasticity) because of accounting for aggregate interlock. 2. However. tensile testing at site may be restricted to evaluating the flexural strength.10 lists some suggestions for tensile strength testing.10 . Tab. 2. march 2008 Section 2-30 . 2.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) for matching the results of finite element analysis [2. On the other hand.Static tensile strength testing Symbol Type of tensile strength Direct tension Flexural strength (modulus of rupture) Splitting strength Standard (example) RILEM CPC 27 ASTM 78 BS 1881 DIN 1048 ASTM 496 BS 1881 DIN 1048 Sample size of standard test variable 152x152x508mm 150x150x750mm 150x150x700mm Variable Variable Cyl.48]. Such testing needs wet-screening of the large fractions down to a grain size. Initial stresses are those from drying shrinkage (causing a tensile stress) and from relaxation due to viscoelastic creep (relieving shrinkage stresses). From the above. splitting tests are the easiest to handle and a large number of these simple test would also provide a somewhat more reliable value of this random variable.14 [2. the latter should be either taken from the literature or tested in specialised laboratories.8) 2. e. preferably on 300x300x900 mm prisms. see Section 2.

It also allows gaining conclusions from observing the broken halves of the specimen. As submitted for ICOLD review. Fig. This is of particular interest for judging the consequences of tensile cracking in dams. Drying out between removal (drilling the core) and testing explains the differences of up to 50% lower tensile strength for cores.3 Tensile strength ratios Similar to what was said for the relation between laboratory and field compressive (core) strength. 2. strain softening curve) instead of only a single strength value. J. This is beautifully demonstrated by wedge splitting tests and can be a valuable input for nonlinear dynamic analysis (damage parameters). They are also the reason for high scatter of tensile in-situ testing. 2.49]. there are several factors for tension which need to be interpreted wisely before drawing conclusions.14 – Influence of shrinkage and creep on tensile strength [2. Raphael reports that large differences in tensile strength between laboratory and field specimens stems from its curing history [2. In view of the above arguments and uncertainties it is recommended that careful thought be given before finally selecting tensile strength criteria.4. march 2008 Section 2-31 . This phenomenon is further exposed in Appendix A.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) superposition of these effects determine weather or not tensile stresses will result in cracking.48] Among the types of testing tensile strength. Post-peak tensile strain not necessarily leads to disintegration of the concrete and thus not to failure. the wedge splitting test renders an enhanced level of information as it gives a strain history result (fracture energy.

as reported in [2.4 Size effect The size effect is made evident through comparison of geometrically similar structural elements of variable sizes.49] Formula f1 = (2. flexural and splitting strength.76 2/3 Other relations (all in kg/cm2) Source Swiss standard SIA 162 Emosson dam A.53]. Some of these results are given in Tab. Here it should only be mentioned that the size effect depends on the type of test and on the rate of loading. such as they are induced by reservoir or temperature fluctuations. The difference is basic.fc f2 = 0. Such argumentation is questionable. This phenomenon defeats the occasionally brought forward argument that a higher content of cementitious material is alleviating cracking.2 average 1.fc 2/3 f4 = 0. Results from fracture mechanics indicate that it cannot be discarded when comparing laboratory tensile strength with the one expected in the structure [2. 0. It is therefore suggested to use a so-called apparent tensile strength for safety evaluations of tensile zones in the dam [2.Khan et al.e.48 0.16].fc2/3 1) somewhat too low because of including tests by Gonnerman and Shuman (1928) Multiplier of c for transformation kg/cm2 to MPa: 0.fc min. 2.5 Tensile strain capacity in dams It is actually the tensile strain capacity (rather than the tensile stress). Tab.11 . third-point load 2 prisms 10cm high and several lengths as above 10x10x40cm 15x15x53cm cyl. 2. and also different for static and dynamic loading. 1. 2. Appendix A oulines this phenomenon.49] 0. where a much higher tensile strain capacity due to slow cyclic reservoir loading is reported as compared to laboratory test results (Fig.54].Ãfc f2 = 2.Conversion Tensile to Compressive Strength No. Interesting in this respect is the observation on Austrian dams. 15x30 and 15x25cm.A.51]. As submitted for ICOLD review. Research in this field indicates that slow rates of loading. can increase the strain capacity by a factor of 1.46] c in fi=c.5 0.15). I 1 Flexural: center point load Flexural.52] [2.7. 9x15x60cm 0. which is of interest for the control of cracking in dams. cyl. [2.1 with respect to laboratory loading rates [2.44] = 10kg/cm 1 MPa 2 Type of strength test Type of specimens Conversion to compressive strength fc after Hellmann [2. i. 2.86.5 to 3.64.681) Kupfer-Gerstle [2.fc 2/3 2/3 [kg/cm2] max. because only strain (unlike stress) depends on the rate of tensile loading.4.86 0. prisms 5x20x20.521) 0.464 in 2/3exp.0). it is different for direct. march 2008 Section 2-32 .28 1.0.11.formulas fc is 28 days 20cm cube strength throughout all relations 2.95.50] Raphael [2.98 f4 = 0.Ãfc f2 = 0.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) A multitude of test series have compared tensile with compressive strength (at the same types of specimen) resulting in conversion factors.361) 0.1 to 2.07 0.4. 3 4 Splitting (Brazilian) Pure tension 6"x12" (15x30cm) cyl.7 Raphael [2.

if stable cracking is the consequence of tensile stresses. Even with particular efforts during construction to prepare concrete and foundation contact surfaces for bonding. If tensile stresses are vertical then excess tension may cause opening of lift joints with their reduced tensile strain capacity.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) Fig.15 – Tensile strain capacity at slow loading rate [2. tensile strengths across bonded lift surfaces and foundation contacts should be expected to be at least 10 to 20 % less than the corresponding intact tensile material strength [2. and in particular the typical requirements and methods for joint treatment. march 2008 Section 2-33 . As submitted for ICOLD review. however too conservative for transient earthquake and flood loads. will be covered in more detail in the APPENDIX B (“The Physical Properties of Construction Joints in Concrete Dams). Some designers cautiously assume no tensile joint strength at all (zero tension approach) and consider a condition as being safe. This seems to be a reasonable approach for static loading.53].16] 2.6 Tensile strength at joints The limiting tensile concrete strength is that which exists at lift joints and across the contact surfaces between the dam and its rock foundation. This aspect.4. 2.

analysis results.16 are practically inexistant in dams. Such criteria comprise a no. 2. 2. nonlinear constitutive stress/strain models for high tensile stress loading are more realistically approaching reality. idealized failure types can be deduced [2. 2. Considering the fact that concrete is an elastic-plastic material. 2.17 demonstrates how the fracture energy (Gf) decreases within the elastic range and that at. Gf starts to increase again prior to complete material disintegration. which model the conditions in dams more closely than uniaxial wedge splitting results [2. As submitted for ICOLD review. transition between 1 and 3 (cases B and C) 3. There exist also biaxial test results from wedge splitting. From phenomenological observations of crack patterns on samples exposed to biaxial loading. or better limited. cases A and D of Fig.12). the scatter of biaxial compression-tension test series is too large as to draw a somewhat reliable biaxial envelope line in this sector and to use it for defining the FS as AC/AB (as in Fig. Designers use other safety criteria as soon as tension is involved. These types depend on the principal stress ratio at the ultimate state of stresses (point at the failure envelope in Fig.7 Tensile strength criteria Similarly to what was said for multiaxial compressive strength considerations. march 2008 Section 2-34 . Other than locally. They may be termed: 1.57]. cleavage or tensile splitting orthogonal to the tensile stress (case A) 2. The relevance for dam concrete is a fringe resistance of a cracked concrete mass for post-peak loading. uniaxial tensile strength values can be used to define the corners of a biaxial strength envelope in the tension-compression sectors (Fig. the onset of microcracking. tension analyses.4.12).55] [2. However. Fig. This unsteady development ensues from restraints of expansion of the fracture process zone for the descending Gf branch and from aggregate interlock for the rising branch.56]. shear failure or splitting parallel to tensile stress (case D) From FE-results it may therefore be assessed which type of tensile strength is most appropriate to be introduced in the biaxial failure envelope for a particular part of the dam. The concept of fracture energy (Gf) is further introduced in Appendix A.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 2. which demonstrates a margin against tensile cracking not considered with a purely elastic model. 2.16) and can therefore be assessed from Finite Element (FE) .

2.Development of fracture energy under biaxial loading[2.16 – Typical failure pattern and corresponding Mohr circle representation [2.57] 1 1 Gf0 is the fracture energy under corresponding uniaxial loading As submitted for ICOLD review.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) Fig.56] Fig. 2.17 . march 2008 Section 2-35 .

8 Use of tensile strength in mathematical models The development of mathematical models calls for performing laboratory testing of strength parameters to be used in constitutive models.18). To be consistent with a linear analysis.18 .4. In this case it has to be considered that a uniaxial tension test will give a nonlinear stress-strain relationship. This is particularly relevant for tensile strength parameters. 2.1 Linear elastic analysis In many cases a linear-elastic stress analysis will be a most adequate representation of reality. flexural tension tests or wedge splitting tests.58] As submitted for ICOLD review. different to what is assumed in the model. The same is true for a flexural test: the load deflection response will be non linear (Fig.Apparent tensile strength (in a linear analysis) compared to tensile test results [2. the stress difference between point A (linear analysis) and B (test result) has to be considered by conversion factors for the stress ratios AB.2 and 1. march 2008 Section 2-36 . Factors between 1.4.8. 2.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 2.4 can be expected from tensile test series for mass concrete of dams. These factors can be obtained from pure tension. 2. Fig.

Standard diameters are 200mm. Here the validation of constitutive strength modelling via experimental data is more cumbersome and delicate. march 2008 Section 2-37 . The sequence of actions comprises: 1 Sample extraction from the dam.61].60]. which best match.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 2. ¾ material degradation due to a-priori strain histories.2 Non linear analysis To conduct a nonlinear analysis.8.59] or ACI [2. 2+3 Fig. 5 Example of a bilinear strain softening relation to be used in a plasticity-based model. and the forcecrack opening diagram as the result of such a test. the present state-of-the-art includes the ability to formulate: ¾ plastic straining under compressive and tensile states of stress. as an example. 4 The post-peak response of the material has to be transferred into a constitutive model by FE-discretization.19 shows a wedge splitting test on cores.62].19 shows the example for validating tensile strain softening by means of wedge splitting tests.59]. 2. It is suggested to follow curing conditions of test standards such as RILEM[2. the lower branch by overcoming friction and tortuosity of aggregate interlock. this test is commonly performed also on beams or cubes [2. ¾ tensile strain softening. ¾ influence of creep. Equally. Curing immediately after core extraction can influence the result. Fig. It is obtained by numerical compliance to experimental data. the experimentally found softening curve. and it generally needs mesh adaptations[2. 2. Shown on 4 are stress contours obtained from FE-analysis. This scatter can be large given the influence of paste-large aggregate heterogeneity on the fracture mechanism. This means that the computed and measured fracture energy (Gf) during testing (surface under the post-peak softening curve) must comply. As submitted for ICOLD review. ¾ viscoplastic bi.or triaxial stress-strain solutions with cycles to reach conversion at time-dependent plastic strain conditions (for seismic loading).4. The kink indicates that energy is consumed by two different fracture mechanisms: the upper branch of the curve is due to microcracking. This needs consideration of the size effect relating the in-situ strength to the laboratorydetermined strength [2. Several cores should be extracted to average the scatter of strain softening curves.

indicating two types of crack propagation: Gf1 strain softening due to microcracking. A. . march 2008 Section 2-38 . 4: FE-discretization of concrete test sample. here a quarter-cube of a wedge splitting test with some stress contours in % of uniaxial compressive strength.Strain softening: tensile strength testing for FE-validation. Fig. and Gf2 continued cracking due to debonding of aggregate interlock [2. 5: bilinear softening diagram which best fits the measured softening curve for input into a nonlinear FE-stress analysis.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) Fig. As submitted for ICOLD review. here shown on a core. 1: core extraction. 3: strain softening curve from wedge splitting test.4). with FH the tensile force. 2.17] (see also APPENDIX A.19 . 2: wedge splitting testing.

The common terminology relates a curved envelope to the Mohr rupture diagram (constructed by the uniaxial compression and tension circles) whereby the term Coulomb relates to a linear/bilinear failure criterion 1 As submitted for ICOLD review. As may be expected. since shearing strength is represented by the point where the failure envelope intersects the vertical shear axis in a σ-τ(normal stress-shear) diagram.A + tanφ.g. roughness or joint grouting. such shear tests demonstrate a wide range of joint quality values between low-shear strength (core separation) and full material strength (invisible joints). Shear strength depends on the adhesion between the substrate material and the grout (in the case of grouted block joints). In addition to the Mohr-Coulomb failure envelope. normal force. whereby the higher flatter branch is for shear resistance of the joint surface (lower φ and some c= f [total block joint surface to shearkeyed surface]). of joint The Mohr-Coulomb relationship is an appropriate presentation for shear resistance. cracks) and to the rock-concrete contact. total area times a coefficient assessing the efficiency curing. no cohesion). say τc=0. and the cohesive strength of the lift joint surfaces. thermal cracking due to lacking curing) or if block joints are taking no or little grout. The lower. c can conservatively be put to zero. For intact concrete a curved Mohr rupture diagram can be assumed enveloping the two uniaxial Mohr circles for compressive and tensile strength.5 SHEAR STRENGTH The significance of shear strength in dam concrete is practically limited to discontinuities in the concrete mass (joints. tanφ N = friction coefficient. This implies that the steeper branch being the shear resistance of concrete. c= cohesion of intact lift surface or contraction (block) joint surface accounting for shear key. A= intact area.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 2.N where T= shear force. steeper branch in the σ−τ diagram is for shear resistance of the plain concrete surface within the keys (high φ. Quality of bonding at joint surfaces is generally checked by shear tests from cores drilled through joints. a bilinear shear resistance criterion is meaningful for assessing the shear resistance in shear-keyed block joints. with τ being the intercept of the failure envelope.g. fc can be assumed. If doubts about effective joint treatment (e.15. e. Shear resistance of the grout material (in block joints or at the concrete-rock contact) depends on the water-cement ratio and thus on the compressive strength fc. fc and the flatter branch a function of the key geometry.2. In absence of test results a shear value of τ=0. if constructed. march 2008 Section 2-39 . as mentioned above. Shear strength is usually assumed to follow the Mohr-Coulomb1 relationship T=c.

2. Fig.20 shows the result of a simple shear test along a discontinuity. Test results can be evaluated to render several resistance parameters.Shear results from core testing Such shear test results can then be used for modelling shear resistance envelopes in a σ−τ diagram according to the anticipated shear resistance parameters.63] and summarised in APPENDIX B. 2. 2.21 . As submitted for ICOLD review.21). march 2008 Section 2-40 . say.Reduction of test data into a Coulomb σ − τ diagram. and ¾ degree of separation from the inspection of the shear surface.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) A fine survey of joint treatment in dams and their benefits are found in [2. providing: ¾ a peak shear value ¾ a residual shear value (from a second test). Fig. ¾ displacement at peak shear value.20 . 2. of block joints with questionable grouting success or of "cold" construction joints (Fig. Fig.

48 As submitted for ICOLD review.68]: testing for evaluating blasting impacts: upper value for 20 MPa and lower value for 50MPa concrete Santa Anita Dam as reported by J.66] Multiaxial loading from Ref.60 1. [2.32 Flexural strength 1. [2. A caveat needs to be added using these strength ratios. Tab.25 1. row) Crystal Springs Dam as reported by J. Bureau's of Reclamation test series. [2.15 1. 2 concrete types: 25 (1.65] Uniaxial loading from Ref.12 1.00 1. Ratios from experimental tests. This means that dynamic strength is less reliable than the static strength.17 2.6 DYNAMIC STRENGTH This section concentrates on strength at loading rates which are related to strong earthquakes. [ 2. Corresponding results from several test series under static and dynamic loading are summarised as static/dynamic strength ratios in Tab. *)for fc'=30MPa 1. Ref.50 1.12. Raphael [2.48 1. row) and 36 MPa strength Crystal Springs Dam ( J. in Ref. at strain rates of 10-3 to 10-2 /s.73 Splitting strength (Brazilian) 1. 2. CEP has a similar proposal.67] for saturated dam concrete Big Tujunga dam.12. march 2008 Section 2-41 .33 1.66] Auburn Dam (Bureau of Reclamation) Ref.00 1./stat.50 1. Ratio av. cores drilled from the dam.[2.00 Comments min.15 Compressive strength 1.12 – Examples for Static/Dynamic Strength Ratios Parameter max. Raphael [2. This has a bearing on the stipulated dynamic strength for safety analyses.27 1.20 1. 2.18 1.32 Big Tujunga dam.51] with reference to size effects Bureau's extensive tests on Auburn dam (1977/78) Crystal Springs Dam as reported by J.60 1. What is not shown explicitly in this Table is the larger scatter of dynamic strength values as compared to static strength testing.31 1.70] suggests a ratio r = 14.57 1.g.69] with tests on dry (1.21 1. e.28 1. should therefore be cautiously chosen (e.49] Big Tujunga Dam as reported by J. rep.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 2. cores drilled from the dam. lower bounds be taken) when using them for safety assessment.0*) 1.49] Ref. Young's modulus Dyn.64] 1.16 1. row) and saturated Specimen (2. like those in Tab.83 1. 2.20 1. 36 MPa strength (2. Raphael [2.06 1.40 1.46 1.07 1.42 1.row) Zervreila Dam (Switzerland): cores drilled from the dam Reinhard [2. Raphael) 32 MPa concrete Ref.49] Ref.49] Tests on Swiss dams.13 1. Raphael [2.38 1.fc' -0. [2. 2 concrete types: 25 (1.29 1.23 1.[2.g.00 1.48 1. row) .57 respecting the dependence of r on the compressive strength.28 Pure tension 1.28 1.[2.50 1./single value 1.

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) The US Bureau of Reclamation performed an extensive test series for dynamic properties from cores of existing dams [2. by tuning the choice of static/dynamic strength ratios with the ability of the constitutive model to consider the above mentioned structural safety margin. This is as expected.89±0.07±0. This well documented publication concludes that the predominant parameter influencing the ratio is static compressive strength (and not mix proportions.66].73 0.45 1.71].9 for strain rates -3 -2 between 10 to 10 /s.69 0. Tab. 2. e.78 0. the other is the large-scale structural response. given the higher ductility of low strength concrete. Evidently.44±0. march 2008 . This is the most complete data series available for dynamic/static strength ratios for dams.6.17 0.).98 The values are from 103 specimens of 10 dams constructed between 1916 and 1995.1 Dynamic Compressive Strength A number of test results exist about the dynamic-to-static strength ratio: • A state-of-the-art paper is [2.73 Min. ratio 0. ratio 1.20 0. Not considered in laboratory-obtained strength ratios is the dam’s strength margin during strong shaking due to the concrete’s ability of plastic straining (strain softening) before failure becomes imminent. Here one is faced with two phenomena: one is the pure material response under dynamic loading (obtained from small samples).7 0. Main results are summarized in Table Tab.58 1.12 1. This means that the designer has to weight the compatibility of these phenomena on a safety outcome. 2.g.29 1.15 Max. The test data reflect a high variability due to the highly variable mix proportions and due to different degrees of saturation of the specimens. with high strength concrete tending towards the lower ratios of the above range and low strength concrete towards the higher ratios.13 – Dynamic Concrete Properties from Dam Cores Parameter Compression Young’s Modulus Failure strain Poisson ratio Splitting strength Ratio dynamic/static (mean ± coefficient of variation 1.09±0. 2.93±0. Section 2-42 As submitted for ICOLD review.3 and 0.13. one has to link these test results to what might happen in the dam. w/c ratio type of aggregate etc.1 1. Again the scatter of evaluating some 30 test series is high and varies between a ratio of 1.69 1.

5. fc with ft fc dynamic tensile strength from pure tension tests compressive cylinder strength 2/3 [MPa] The comparison with Tab.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) • In 1975 tests were performed with the concrete of Big Tujunga dam in California on cores drilled from different quality concrete [2. Again.8 38. Reference [2. both for static and dynamic load.11 shows that this formula suggests a 50% increase of dynamic against static strength.Ratio between dynamic and static strength Static strength (MPa) 24. Another As submitted for ICOLD review.: ft = 1.9 40. march 2008 Section 2-43 . A strain rate of 0. 2.6 1. such high ratios should be used with caution and in combination with the mathematical model to be employed for stress analysis.8 34.14 .28 0.07.7 22.98 0.4 37. fc2/3 [kg/cm2] ft = 0.2 Dynamic Tensile Strength Commonly the 2/3-power law is used for the relation between tensile and compressive strength. Tab.14.11 Ratio Given this information the a caution approach suggests using a ratio close to unity for any analytical consideration unless test are performed for the particular dam.99 1.72]. 2.16 1.7 Dynamic strength (MPa) 28.6. The results from 6 dynamic tests are as shown in the following Tab.9 26.04s was applied corresponding to the 1/4 of the assumed earthquake cycle.8 36.3 22. 2.2 36. 2.27 1.49] uses the same power law for assessing dynamic tensile strength. ft. as mentioned earlier.8 46.

2. is shown in Fig.Tensile (static/seismic) vs.73].35 for dynamic loading and for common dam concrete (fc§20 MPa) is suggested.22 [2. see Fig. the dynamic tensile strength is also an ambiguous material parameter as soon as it is used as input in linear stress analysis. expressed by the formula ft = 3. 2. The linear high strength was termed apparent tensile strength [2. both for static and dynamic tensile strength.58]. with maximum aggregate size up to 25 mm. The strain-rate effect on the tensile strength for different concretes. The increase of the tensile strength with increasing relative humidity was observed [2. If a linear stress-strain relation is used on a material. As submitted for ICOLD review. then the analysis will predict failure at a higher stress than the test value (peak of the stress-strain softening curve).22. Studies at Delft University indicate a decreasing static/dynamic strength ratio with stronger. In [2.70]. 2.1 [MPa] Fig. compressive strength [2. typically from ~ 10-2 s-1 of earthquakes to ~ 102 s-1 of impacts. more brittle concrete.3] a ratio of 1.74]. The range of strain rates might be very large. march 2008 Section 2-44 . The difference between apparent (linear) and nonlinear peak at ultimate strain can be considerable. 0.70] Besides the above mentioned post-earthquake strength margin. which actually behaves nonlinearly at high stresses. fc for earthquake loading.23 [2.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) factor influencing the static/dynamic strength ratio is compressive strength.2.

static tensile strength 4 5 mm (cube 60 mm side) 10 mm (cube 60 mm side) 10 mm (cube 200 mm side) 25 mm (cube 200 mm side) 3 2 1 0 10 -7 10 -5 0.3 Dynamic Shear Strength No increase for dynamic as compared to static shear parameters is generally assumed. dynamic tensile strength / max.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 5 max.6. From an engineering aspect the friction coefficient should be larger for short term loading which leads occasionally to accept an increase of 10%.73] 2. 2.23 – Dynamic/static tensile strength at different strain rates [2.001 0.1 10 STRAIN RATE [ s -1 ] Fig. march 2008 Section 2-45 . As submitted for ICOLD review.

Lausanne 1985. San Francisco 1988. Q. Dungar. 1999. [2. Bouilder. Chapter 5.7 REFERENCES [2. Vol.9] Paolina. R. ACI Publication. "How to avoid thermal cracking of mass concrete". J.. p. Part 3. "Testing practice affects concrete's perceived quality".. Vienna. Nov.1R-96. eds. Cement and Concrete Research.J. 1991. I. Baldovin. 1998. August 2000. [2. “ACI Building Code Requirements for Structural Concrete”.4] US Bureau of Reclamation: "Concrete Manual" . Berra.10] Barlett. 35-39. [2. 16.. Appendino.J. Morris. "Contrôle des bétons durcis in situ . Wilk W. [2.65.8] American Concrete Institute. Tschegg E.A.. M. 2001.18] American Concrete Institute. Q.4. 1. ACI Material Journal.. [2. Vol. ICOLD Congress Proc. [2. As submitted for ICOLD review. 207. F. Morreu.1R. R. 04. MacGregor. et al. June 1991. "Does Core Size Affect Strength Testing?".Etat de l'art".6] LERM (Laboratoire d'etudes et de recherches sur les materiaux).77.15. of the International Conferente on Dam Fracture.17] Kreuzer H. American Concrete Institute (ACI). “In Place methods to Estimate Concrete Strength”. “Fracture Enery of Concrete under Biaxial Loading”.5] Soares de Pinho. Q. 1994.2] Experience from Swiss dams: Bulletin Technique de la Suisse Romande. 228.02 [2.7] Neville. [2. 28. Proc. J. 15th ICOLD Congress on Large Dams. ACI 318R-95. Sep. Carabelli. Aug.15] Akers. [2. No.E. Not Easy to Interpret".14] American Concrete Institute.. ACI Manual of Concrete Practice. LERM rapport d'etude n° 01. 1981 [2. [2. Aug. E.62..13] EN 12504: "Essais pour béton dans les structures (Testing of concrete in structures)". 1998 [2. D. Section 4. n° 93-101.43. J. 59-68. 2. [2. A.. Vielmo. Vol. Annual Book of ASTM Standards.01.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) 2.57. 447-457. G. "Control of Mass Concrete for Dams: Full. "Core Tests: Easy to Perform. [2..3] Raphael J.mix and Wet-sceened Concrete Tests". Concrete International. M. July 2001. “Mass Concrete”. New York. [2. 42-45. M. R. march 2008 Section 2-46 . Posta. V... M.1] Vilardell. Norme européenne. 1956.. [2.11] Ahmed A...8th ed.. R. Concrete International. Vol. 91. “Deterioration problems for concrete and masonry dams in Italy”. Colerado. Paris. Arles. Saouma. "The Nature of Mass Concrete in Dams". “Estimation of the modulus of elasticity for dam concrete”.. R. Wiley. 17th ICOLD Congress.12] ASTM C 42-99: "Standard Test Method for Obtaining and Testing Drilled Cores and Sawed Beams of Concrete".27.16] Widmann. "Effect of Core-Length-to-Diameter Ratio on Concrete Core Strength". April 1990. [2. R.5529. Concrete International. ACI SP 55-6.001. 339-348. H. Bianchini.

Rackwitz.24] ISO 2859 . 4954.29] Lombardi G. 1990. “Recommended Practice for Evaluation of Strength Test Results of Concrete”. Collected from several publications. 1951. "Distributions a double borne logarithmique". edited to the 20th ICOLD congress. [2. USA. Heft 206. [2. "Better concrete for dams: recent experiences and trends in California". 1.T.34] Popovics S. H. August 1996 [2..22] ANIDEL. R.33] American Concrete Institute. Bury K.39] Huber.32] Swiss Committee on Dams. “Statistische Analyse der Betonfestigkeit” (Statistical Analysis of Concrete Strength). 214. Ernst. [2. Journal of ACI. 1979.28] Lambotte H. 1979. Vol. “Variables in Concrete Aggregates and Portalnd Cement Paste with influence on Strength of Concrete”. 512 – 529.1-91 (Reapproved 2002). XP P 18-305.. [2.. Grindelwald April 1993.. 2. Gillespie. H.37] Gordon. E. Morreu H.C. R. “Concrete Information”. [2. vol.. et al. As submitted for ICOLD review.A. September – October 1990. Civil Engineering. [2.. [2.11. 16th ICOLD Congress. "The effect of distribution bounds on dam safety evaluation". 3. (1969).Le dighe di ritenuta negli impianti idroelettrici italiani. [2.38] Leemann.5.35] AFNOR Association Francaise de Normalisation: "Béton prêt à l'emploi". Vol. “Maximum size of aggregates affects strength of concrete”.27] Comité Euro-International du Beton (CEB): CEB-FIP Model Code. Manual of Concrete Practice.. C.. [2.H. 1963 [2. Heft 5. Berlin. 24.21] Kreuzer. 10th ICOLD Congress on Large Dams. march 2008 Section 2-47 .. [2. “Concrete of Swiss dams”. et al. [2.3. L. Vol. Skokie III.39.-Revue No.Tests for departure from the normal distribution. [2. et al. ACI 211. “Die Kölnbreinsperre – Neue Wege in der Technik des Massenbetons” [Kölnbrein Arch Dam – Recent Developments in Mass Concrete Technology]. Beton & Stahlbeton. “Analysis of the concrete strength vs water/cement ratio relationship”. 1429 [2. [2.31] Rüsch. ACI Publication. Proc. 1987. Sept. September 2000. Heavyweight and Mass Concrete.. W. 1988. Q. June 1999... San Francisco. 532-536.36] Higginson.H. Beijing.S. W. “Gebrochene Zuschlagstoffe” [Crushed Aggregates]. Nr.Sampling procedures for inspection by attributes.Sampling procedures for inspection by variables.. [2. [2. 1963.25] ISO 3951 .20] Portland Cement Association (PCA). [2.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) [2. Montreal 1970. “Standard Practice for Selecting Proportions for Normal. ACI Material Journal. Nov. A. Deutscher Ausschuss für Stalbeton.23] American Concrete Institute.. [2. H. International Workshop on Dam Safety Evaluation.Statistical interpretation of data -.W.26] ISO 5479 .C. Aug.19] Tuthill. "Controle de la qualité du béton chantier et en usine". Schweizer Ingenieur und Architekt (SIA).30] Kreuzer H.

49] Raphael J. 22-24 May 1984.. n° Sept. ASCE. Issue 3.. [2. July 1997.. [2. 1989. P. “Der Einfluss der Deformationseigenschaften des Betons auf den Spannungszustand (the influence of deformation properties on the state of stress in concrete)”. 1991. Vol.44] Kupfer.53] The National Academy of Sciences.43] Rüsch. 1998. 1973. Febr. ACI Concrete International. 1992.. 3. [2. [2. G. Teeni. [2.. As submitted for ICOLD review. 4-6 October 1972.B. “Neue Wege in der Technik des Massenbetons” [New developments in mass concrete technology]. [2... [2. Ferrara. Esevier Applied Science. [2. 308 -92. Gerstle. n° ASCE. Proc. Journal ACI. 158165. Conf..G.. 487-493. Vol.. [2..al. 77. Paper 9917. 1977.56] Kreuzer. III..40] American Concrete Institute. “Reliability analysis of the Mohr failure criterion”. Newman. ACI Material Journal Sep.47] Hellmann. “Prediction of tensile strain capacity of mass concrete”. K. on concrete under multiaxial conditions. J. [2. Aug.. “Give it a week: 7 day initial curing”.41] Haque. “Apparent tensile strength for arch dam design: a review for rate. K. Jg.57] Tschegg. Z. R. “ Experimental research on deformation and failure of concrete under triaxial loads”. medium and high strength concrete at early ages”. H. T.3. SIKA Information. H. n°5. “Standard Practice for Curing Concrete”. of the Southampton 1969 Civil Engineering Materials Conference. of the Conference on Fracture mechanics of Concrete Structures. 1959 [2. “Behaviour of Concrete under Biaxial Stress” Journal of Engineering mechanics Div. et. FramCos 1.48] Metha.. Bazant ed. London.. 89. Cannes.N. size and strength dependency” Dam Engineering. ed.-Oct. 2-69. [2. No. M.55] Newman.. H. 1992. ACI Manual of Concrete Practice. 99. Dec. H. march 2008 Section 2-48 .ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) [2.C. Kreuzer.. Bury.M. P. Aug. J. R.. 1971. ACI Concrete International. “Fracture in concrete under biaxial loading – numerical evaluation of wedge splitting test results”.. “Durability – Critical Issue for the future”.B. Gettu. R. “Beziehung zwischen Zug-und Druckfestigkeit des Betons”. 68-70. Performance. 75. 455-460./Oct. “Earthquake Engineering for Concrete Dams: Design. 1969. USA. H. Heft 9. 1978.46] Berra. P. “Tensile Strength of Concrete”. Breckenridge 1992....45] Bertacchi. et al. “Tensile strength of low. 853866. Commission on Engineering and Technical Systems. Vol.. [2. 5. Schweizerische Bauzeitung. Mc Donald. [2. M. Bellotti. Sep. Toulouse.51] Bazant. H. [2.42] Huber. 115.52] Dungar.54] Liu. “Failures theories and design criteria for plain concrete”. Proc. Vol. March. Fatticcioni. [2. “Rate Effect and Load Relaxation in Static Fracture of Concrete” ACI Material Journal. M. 456-468..H. n° EM4. K. 1998. and Research Needs”. Journal of Engineering Mechanics. Vol.. “Triaxial behaviour of concretes of different weights” Int.K. 1996.. Beton. E.50] Khan A. [2. A. ACI Journal. Colloque Rilem.B. 2. April 1984.

[2.34 . Richter & Ass.January-February 2001 . 1991.64] Tarbox G.67] Hatano T. Labibes. 13 ICOLD Proc. edt. 1995.12. march 2008 Section 2-49 .. Plesha. C. See also comment to this reference in [2. No.3] [2. Albertini. 9.73] Berra. Perry.M..65] Brühwiler E.74] Berra.3R-97. As submitted for ICOLD review. et al. [2. Arch Dam Final Design. R. “Fracture Mechanics of Dam Concrete Subjected to Quasi-static and Seismic Loading Conditions”. P. [2. He. May 1981 (unpublished).62] Gunn. New Dehli 1979. “Rate and size effect in concrete fracture: implications for dams”. of International Conference. Static and Dynamic Stress Analysis.D.63] Pacelli. "Aseismic design criteria for arch dams in Japan". 1999. Water Power & Dam Construction. Albertini. et al.. Issue 3.-Febr. "Seismic Analysis of Concrete Dams". R. Part 4. et al. 1989.L.. r.Gettu. Dolen T. Magazine of Concrete Research. M. 1991.70] Discussion by H. Dam Fracture. Giangrasso.P. “Fracture mechanics test methods of concrete”. RILEM Materials and Structures. S.W..68] Harris. American Concrete Institute. RILEM Materials and Structures. Cadoni.H. Ph.. Carpenter L.60] ACI 446. [2.66] Bischoff. carpinteri. Vol. [2. In Ref.H.).1.P. [2. ICOLD [2. Giangrasso. 94-98. 1978. 739... 24.W.61] Bazant. "Compressive behavior of concrete at high strain rates". Q. [2. Hydropower & Dams.P. M..69] Saucier K.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 2 (Strength properties) [2. “Finite Element analysis of fracture in concrete structures”. Ecole Polytechnique Federale de Lausanne. Manual of Concrete Practice.. 21 – 26. Nov. Thesis no.pp. 97. Proc. 2001.E. 2000 – pp. [2.72] Lindvall. Istanbul 1967. Cadoni. STP [2.71] Bureau of Reclamation (Mohorovic.51 R. "Treatment and performance of construction joints in concrete dams". ACI Materials Journal... Harris D. [2. "El Cajon Hydroelectric Powerplant. S.59] RILEM Report 5.W..58] Dungar. [2. D. "Dynamic Properties of Mass Concrete". [2. ” High strain-rate tensile concrete behaviour ".3]. W.n°235 .. Vol. Reinhard and others included in ACI Jan. " Strain-rate effect on the tensile behaviour of concrete at different relative humidity levels ". 1988. Shah and A. 1993.A. 52 – n° – Oct. R. Congress Proc. Q. Vol. Boulder CO.. S. Congress [2. Chapman & Hall. May-June 2000. Sept. "Dynamic Properties of Mass Concrete Obtained from Dam Cores". [2.35 R.. Los Angeles 1975 (unpublished). "Final Report for Investigations of the Big Tujunga Dam". 5 365-370. z. 163-178. Rowlands. A. 1998.. [2. “Non-linear design and safety analysis of ach dams using damage mechanics”. ASTM 654. M.3. “Dynamic Properties of Mass Concrete Obtained from Dam Cores” Dam safety Office. Labibes.

..........7 REFERENCES . 13 3........................18 3..............................................................................1 3.............2........................2..........2 STATIC MODULUS OF ELASTICITY................... 8 Transition zone....................................................4.........................2........................1 GENERAL ...3...........................6 Aggregates..............................................2..........................3 Empirical expressions through correlation with the compressive strength ......2 Prediction based on elastic models ................................................3... 11 3........... 8 Test conditions .......4................................................4 POISSON’S RATIO....................................................2.................4 Estimation of the elastic modulus . 14 3.....................5 3...3.......2.......................5 SIGNIFICANCE OF ELASTIC PROPERTIES ON DAM BEHAVIOUR ........................1 Modulus of Elasticity under Seismic Loading ..................................1 Typical behaviour and definitions .............................2............5 3.... march 2008 Section 3-1 .2......................................................2 Internal mechanisms affecting the behaviour .................................................................... 1 3..................................................................................................................4 3.........................3..........................................2................3 3....... 12 3..19 3......................6 USE OF ELASTIC PROPERTIES IN MATHEMATICAL MODELS FOR DAM STRUCTURAL ANALYSIS.................................................................3.. 9 Age ...4...................3...................................3 Factors affecting the elastic modulus ............4 3.28 As submitted for ICOLD review........................... 10 Damage conditions ............................................2..........................2....3 DYNAMIC MODULUS OF ELASTICITY ................................................................2...17 3..........................2 3.24 3....................23 3......................................2 3..3....................12 3.........................................ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) 3 ELASTIC PROPERTIES 3 ELASTIC PROPERTIES........................... 6 Cement paste matrix .............6 3...4 3......2...2 Dynamic Ultrasonic Modulus of Elasticity ........................................................................................1 Experimental determination in laboratory ...........................17 3.......3..............

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) 3. 3.g. among others.4] and Chen [3.[3.6]) behaviour includes the effects of fluid pressure in the pore spaces of an elastic-plastic material. Concrete has a non-linear stress-strain curve. it is useful to present a brief summary of other constitutive models which have been used to determine the stressstrain response of concrete structures. “visco-elastic”. Fauchet et al. In non-linear elastic models. Concrete is variously characterised theoretically as “elastic”.Generalised stress-strain curve for concrete [3. “Poro-plastic” (see for example. Elastic properties provide a measure of the stiffness of a material. the elastic properties may vary depending on the level of stress or other controlling parameters. “elasticplastic”.1 .1 GENERAL The elastic characteristics of a material define how the material will deform when various loads are applied. Fig.3]) behaviour includes treatment of time dependent deformation such as creep.1] Prior to discussing the elastic properties of concrete. “Elastic-plastic” behaviour (see for example. 3.2]) behaviour by definition has the property that the strains appear and disappear immediately on the application and removal of stress. At higher stress levels the time dependent behaviour of concrete is non linear. 3.2 shows a rheological equivalent of the above mentioned stress-strain response. Fig. the typical stress transfer between the damping (viscosity) and As submitted for ICOLD review. It shows. [3. Bangesh [3. e. “Elastic” (see for example [3. Viscoelastic rheological material models are applied to concrete when the stresses are relatively low which is similar to the range of applicability of linear elastic material models. “poro-plastic” or otherwise depending on the behavioural model assumed. “Viscoelastic” (for example Neville et al. Creep of concrete is discussed further in Section 4. Elastic behaviour does not imply a linear stress-strain relationship.1). however to a certain degree it may be considered an elastic material (Fig. march 2008 Section 3-2 .5]) includes permanent deformations and a plastic flow rule subsequent to “yielding”.1] and Timoshenko and Goodier [3.

it is necessary to select further idealisations in terms of homogeneity and isotropy. with time.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) the friction element. vertical block joints or weak horizontal lift joints are marked discontinuities too. with “smeared cracks” and strain-softening or as discrete cracks depending on the level of detail available and behaviour desired. march 2008 Section 3-3 . Moreover. In some instances. As submitted for ICOLD review. In addition. Fig. 3. stress will be completely transferred from the first to the latter resulting in irreversible (plastic) strain. Homogeneity is frequently assumed at a macro level but of course at the particle level the grains and aggregate form a very inhomogeneous mixture. These discontinuities may be modelled as equivalent continua as “no tension” models. Plastic strain continues until the accumulated strain is enough to activate resisting stress. This kind of process can then be simulated in a non linear stress-strain relation for Finite Element-modelling. Isotropy is usually assumed in conventional mass concrete. inhomogeneity may develop with time due to deterioration mechanisms varying spatially.2 . The classic and simplest constitutive material model is one of linear elasticity in an isotropic homogeneous material in which the behaviour can be characterised by two parameters. In structures where distinctly different concrete mixes were used. then selected volumes may be assumed to be homogeneous within themselves but may vary from volume to volume. In fact it can be shown that strength parameters (compressive strength. A material is isotropic when its elastic properties are identical in all directions. the modulus of elasticity (E) and the Poisson’s ratio (ν). Young’s modulus) are only insignificantly (if ever) different when comparing their values parallel or perpendicular to the direction of concrete placement.Rheological equivalent of typical stress-strain response in concrete [3.7] The above models consider the concrete as “continua” but in fact internal cracking due to shrinkage or other source of tension may introduce discontinuities. hereby.

in a viscoplastic model) or for short term loading. The tangent modulus is of little practical significance since it applies to very small stress-strain conditions. Concrete in dams will exhibit time dependent straining (creep) and the notion of an elastic modulus does not strictly apply. Since the curve is non-linear.g. 3. and. 3. As submitted for ICOLD review. is fully justified. the chord modulus (or an incremental modulus) given by the slope of the line between two points.4). the tangent modulus given by the slope of the line drawn tangent to the stress-strain curve at any point. march 2008 Section 3-4 . 2. with respect to using elastic moduli in finite element analyses (FEA). in a linear-elastic FEA the artifice to choose a secant modulus. in which the strain is measured between stipulated load percentages (see 3. using a modulus of elasticity and Poisson’s ratio to include certain elastic and inelastic behaviours exhibited by concrete.2 STATIC MODULUS OF ELASTICITY 3. including creep effects.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) The behaviour of concrete is usually characterised differently for static and dynamic loading. It is the modulus defined in most standards. This will depend on the rate of loading and the idealisation. or stress level.2.8]). for which the modulus is regarded being higher than long term loading. for example. such as blast loading. Here we discuss the use of the two parameter model. the tangent modulus is a choice when non-linearities are taken as rate-dependent (e. However.2. In some cases. the secant modulus given by the slope of the line drawn from the origin to the point on the curve at the stress level.3 (from Mehta [3. However. in others the distinction is related to modelling assumptions and strain magnitude. the rate effects are significant in themselves. These are: 1. various definitions of elastic moduli are possible as illustrated. on the curve. the use of secant modulus for static loading versus tangent modulus for dynamic loading of small amplitude. 3.1 Typical behaviour and definitions The behaviour of concrete in compression is shown diagrammatically in Fig.

3 . mix design and curing. i. the micro-cracks increase in length.2 Internal mechanisms affecting the behaviour The non-linear stress-strain behaviour has been explained in terms of the progression of micro-cracking of concrete through four stages [3. the concrete would contain micro-cracks in the transition zone between the matrix and the coarse aggregate. Subsequently the micro-cracks become unstable and non-linearity is stronger and eventually spontaneous crack growth and large strains occur which could eventually lead to failure. As submitted for ICOLD review. although they may still be stable.2.e. march 2008 Section 3-5 . the micro-cracks will be stable and the modulus will be approximately constant.A typical concrete stress-strain curve and various definitions of elastic moduli 3. before the application of external load. Then as the strain progresses. Up to a certain level depending on the placing. chemical shrinkage (autogenous volume change) and thermal cracks. essentially linear.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) Tangent C Initial Tangent B Unit Stress D Chord A Secant Unit Strain Fig.8]. and non-linear behaviour becomes apparent. 3. In this Section. The number and extent would depend on the mix proportion. Initially. the discussion mainly pertains to linear elastic behaviour which occurs at low level of strain and when micro-cracking is small. curing history and strength of the material and consist of drying shrinkage.

Concrete made with natural aggregate in a soft paste conforms more closely to the lower bound. 3.10].2.3. the series model.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) 3. A dense aggregate has a high elastic modulus and this also produces a concrete with a higher modulus. loading rate application). i. facing concrete) the greater is the porosity importance of the aggregate. like for example sandstone and limestone. The higher is the concrete grade (e. As submitted for ICOLD review.g.1 Aggregates The relationship between the concrete static modulus and the aggregate used for the concrete can be obtained through the volumetric contents of both cement paste and aggregates. 3.4) [3. Ea. march 2008 Section 3-6 . In particular this relationship can be expressed with the two-phase composite models. age and damage conditions. which combine elements with parallel (Voigt model) and series (Reuss model) phases (Fig.2.e. Other aggregate properties that influence the elastic modulus of concrete are the maximum size and size distribution. These last through the influence of micro-cracking in the transition zone between aggregate and cement paste (see 3. (2) each phase behaves linearly in the linear elastic regime. Ea is considerably higher than Em and the true value for EC should lie between the theoretical lower limit (series model) and the spherical model. For good quality normal weight aggregates.9]. are characterised by EG between 10 and 60 GPa. (3) there is no interaction between the aggregate particles. and their relative static moduli. These assumptions are not all always valid. temperature. the matrix and the coarse aggregate. the surface texture and mineralogical composition. Aggregates from low porosity rock such as basalt and granite have Ea greater than 60 – 70 GPa while aggregates from more porous rocks. test conditions (humidity. the shape. 3.2. aggregate – cement paste interface (transition zone).9] or that one consisting of spherical particles in a continuous matrix [3.4) [3. This last can be expressed as: (1 − Va ) E m + (1 + Va ) E a ]E m (1 + Va ) E m + (1 − Va ) E a Ec = [ These models depend on the assumptions that (1) the concrete is a three-dimesional combination of the two homogeneous and isotropic phases. However other more realistic models for concrete have been suggested as those by Hirsch and Counto (Fig. In fact it directly affects the aggregate stiffness.3 Factors affecting the elastic modulus These are aggregates.3). The parallel system is the upper bound for elastic properties of interest while the series system provides the lower bound. and (4) that there is a perfect bond between aggregate and the matrix. The porosity of the aggregate is another aggregate parameter able to influence the concrete elastic modulus.3. cement paste matrix. such as those used in dams.

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) Model Voigt model Equation Assumes ε is the same in aggregate and matrix E C = Vm ⋅ E m + Va ⋅ E a Reuss model V V 1 = m + a EC E m E a Assumes σ is the same in aggregate and matrix Hirsch’s model 1-X X §V V · 1 1 =X⋅ + (1 − X ) ⋅ ¨ m + a ¸ ¨E ¸ EC Vm ⋅ E m + Va ⋅ E a © m Ea ¹ Counto’s model Va 1 1 − Va = + EC Em 1 − Va ⋅ E m + Va ⋅ E a ( ) Matrix Aggregate EC Em Ea Vm Va X Elasticity modulus of concrete [MPa] Elasticity modulus of matrix [MPa] Elasticity modulus of aggregate [MPa] Volumetric fraction of matrix in concrete Volumetric fraction of aggregates in concrete Arbitrary parameter for combination of parallel and series composition within Hirsch’s model Fig.Concrete two-phase composite models and computation of elastic modulus [3.4 . march 2008 Section 3-7 .9] As submitted for ICOLD review. 3.

3. The reason of this is the behaviour of the cement paste – aggregate interface or transition zone.8] As submitted for ICOLD review.2 Cement paste matrix Similarly the capillary porosity of the cement paste matrix controls the concrete elastic modulus. in the considered concrete stress range. aggregate and concrete [3. It should be noted that these values are similar to the elastic moduli for low quality high porosity aggregates. The porosity of the cement paste matrix is affected by the water/cementitious ratio.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) 3. the concrete curve lies between those of the two components but while these two last ones are substantially linear elastic.5 schematically shows the typical stress – strain curve for a concrete and its main components (cement matrix or paste and aggregate). march 2008 Section 3-8 . The modulus of the cement paste matrix is generally in the range of 10 GPa to 30 GPa. air content and mineral admixtures.3. Fig.3 Transition zone Fig.5 – Typical stress-strain behaviour of cement paste. 3.2. 3.2. As for all composite materials. the concrete curve is not.3.

11].3]. and the moisture doesn't decreases. There are several authors that have experimented this phenomena. march 2008 Section 3-9 .5) and an elastic modulus EC can be measured. width and numbers and the higher the stress level the lower the stability of the microcracks system. Above this range. whose amount is related to degree of consolidation and degree of hydration and type of curing. to maintain their in situ moisture content during specimen preparation and transportation to the laboratory (for example wrapping them with plastics and moist towels). E (saturated) = E (dry) + 3 GPa.8] [3. when taking cores from concrete dams.5. The influence of temperature on elastic modulus is not direct. the transition zone microcracks begin to increase in length.4 Test conditions The test conditions also affect the value of measured elastic modulus. on the contrary. 3. However. between the normal range of temperatures (0ºC-50ºC) the effect of the temperature on the modulus is less important.2. Tests conducted under dry conditions indicate approximately 15 to 20 per cent lower elastic modulus than corresponding samples tested in the wet conditions. 3. the modulus has lost 60% of the initial value. The dry samples are adversely affected by drying which increases the amount of microcracking in the transition zone. higher than in wet conditions. due to the wedge effect produced by the presence of water inside the concrete porosity. In the most cases that have been analysed.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) The transition zone contains void spaces. and at 150ºC the 60% of the initial value can be lost. In case of saturation of the concrete. depending on the saturation degree [3. Gorisse [3. Above 50 – 60 % of the ultimate load cracks begin to form also into the cement matrix and then very high strain are finally developed up to failure [3. 3. micro-cracks and calcium hydroxide crystals. such as the humidity degree and the sealing degree. and reaching 400ºC. This is reflected by a clear non linearity in the concrete curve of Fig.3. They already exist in this interface zone even before the application of external loads but generally they remain stable up to about 1/3 of the ultimate load.9]. For this reason it is important. The modulus of elasticity varies also with changes in temperature and a tendency of decreasing values with increasing temperature conditions have been found [3. It is worth to note an apparent inconsistency in that the compressive strength under dry conditions is. This is approximately the range where stress-strain curve remains linear (Fig. a dry concrete has a reduction of elastic modulus that varies with the temperature in quasi-linear way when the value exceeds 50ºC.12] has established the following relationship that quantifies the value of the E modulus. As submitted for ICOLD review. the value modulus can fall quickly. because there are other aspects that must be considered at the same time.

3.11 in Section 2). The average increase within this period is 122% with a low coefficient of variation of 7. such cases of no modulus increase on core samples may be related to eventual micro-cracking on sample from coring.6). that still can continue even after a long time. An increase between 105% and 138% within the period of 28 days and one year is reported. This is though to depend on the density improvement of the transition zone. lower than in static conditions.2. because of the slow chemical beneficial interactions between cement paste and aggregate. on the contrary. f(τ) = cube strength of concrete at time τ and fcu = cube strength of concrete at 28 days.3 on the dynamic modulus of elasticity and Figure 2. march 2008 Section 3-10 . 3. also due to the significant creep effects. Fig. Raphael in his “The nature of mass concrete in dams” [3. For example a commonly accepted expression developed by Geen and Swanson [3. under instantaneous loading only a little strain can occur prior to failure and the modulus of elasticity is higher than in static conditions [3.9]. Investigations on dam concrete show a considerable increase in moduli with age. Research has also shown that concrete modulus of elasticity tends to increase with time at a faster rate compared to the compressive strength [3.6« »¾ ¬ fcu ¼ ¿ ¯ where E = Young’s modulus at 28 days. 3.15] reports results for ratios “modulus under sustained load after one year to the instantaneous modulus after 28 days”. thus lowering the elastic modulus further [3. the elastic modulus is. In fact creep strains would be superimposed. For very slow loading rates. As submitted for ICOLD review.13] is as follows: ­ ª f (τ ) º ½ E(τ ) = E ®0.5 Age The concrete modulus varies with time and numerous investigators have presented results of tests and equations to describe this behaviour.6 also indicates some cases where the modulus of elasticity increases during the period of cement hydration (say until an age of 90 days) and is constant thereafter. 3. The effect of an increase of the modulus of elasticity with time even beyond an age of 90 days is particularly pronounced when pozzolanic materials are used within the cementitious content of the concrete mix.5% (Fig.4 + 0. ACI 207 (“Mass Concrete”) [3.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) The rate of loading application is another important factor that influence the concrete elasticity modulus.8] (see also the Section 4 on creep properties). In fact.14] shows elastic properties (modulus of elasticity and Poisson ratio) for 13 dams.8] (see also the paragraph 3. Because all laboratory samples show significant increase with age.

2. damaged concrete can mobilize margins of strength due to aggregate interlock of a micro-cracked concrete.3. From Fig.7 can be observed how the elastic parameter can evolve due to the effects. one has to be cautious in not relating an E-modulus reduction to a corresponding reduction in strength. fatigue. More information on this aspect can be found in Appendices A and C (Fracture Energy Application for Dams. 3.Increase of modulus of elasticity with age [3.). and Physical properties of concrete subjected to expansion phenomena as AAR in dams). A reduction of this parameter. its elasticity modulus becomes.6 . for example of micro-crackings. However. 3. 3. chemical or physical deterioration etc. as the damage proceeds. of course. can be observed through the changing of slope of loading and unloading cycles. both in tension and compression. As fracture mechanics testing shows.7 – Elastic modulus evolution due to damage increase As submitted for ICOLD review.15] 3. lower compared to that in no damage conditions. Stress E E’ Strain Fig. march 2008 Section 3-11 .ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) Fig. Such strength margins are of particular interest for judging post-earthquake stabilities in dams.6 Damage conditions If concrete is damaged or plasticized (because of high loads.

ASTM C 469 uses 6 by 12 inch (15 by 30 cm) cylinders.18] provides the static Young’s Modulus and Poisson’s Ratio using a compressive test procedure. For drilled core specimens.1. estimating the elastic modulus directly from the concrete compressive strength (paragraph 3. this is not always easy: in fact the maximum size aggregate usually ranges from 80 to 200 mm and large specimens. It stipulates the use of chord modulus with a lower point at strain of 50 µε and the upper point corresponding to 40 per cent of the compressive strength at the time of loading. The preferred gage length is typically about one half the height of the specimen.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) 3. less complicated and time-consuming compared to the experimental determination. removing aggregates larger than 40 mm. should be required [3. The sizes of these samples should be at least three times the maximum size of the aggregate used in concrete. accurately placing the test specimen in the centre of the loading platens of the testing machine and monitoring the strains values provided by the gauges.17] in order to predict the modulus of the real size aggregate concrete from the experimental results obtained form the wet-screened concrete specimens (paragraph 3.2. A lot of empirical relationships have been proposed and recommended by national and international standards. the ASTM C469 [3. If these strains eccessively differ by more than 20% from their mean value.2. It is recommended to recentre the test specimen and repeat the test. As submitted for ICOLD review. and using standard cylinders of 15x30 cm (2. Due to the practical difficulties in performing such tests.1 Experimental determination in laboratory Dam concrete is designed for a compressive stress state. for example cylinders of 45 x90 cm. The lower and upper chord modulus points are chosen to avoid seating effects and also the modulus is obtained in the “elastic” range of the stress-strain curve.2.4. Furthermore empirical approaches can be used to estimate the elastic modulus.3).4. For example.2).2. ASTM C469 specifies that the gage shall not be less than three times the maximum size of aggregate or more than 2/3 the height of the specimen.2.2.5 are used.1 in Section 2).2.4. Before loading the specimen it is important to follow a specific centering sequence.16]. only diamond drilled cores with length-to-diameter ratios of greater than 1. The strains along the axis of compression should be measured with two or more gage lines such that eccentric loading and non-uniform response can be monitored. The gage length for strain measurement is an important consideration. march 2008 Section 3-12 .1). therefore the static modulus of elasticity is typically measured with concrete in compression. A method has been recently proposed [3.4 Estimation of the elastic modulus The elastic modulus of a concrete is conventionally measured using standardised tests directly based on concrete samples subjected to uniaxial loading (paragraph 3. 3.3. However. However this procedure can result in a incorrect estimation of the elasticity values. from an experimental point of view.4. dam concrete is usually wetscreened. It is based on the application of the simple elastic models presented in the paragraph 3.

The use of a stiff testing machine will result in gradually softening behaviour in the post-peak response stages of the test.1.2. As submitted for ICOLD review.2 Prediction based on elastic models A valuable proposal for predicting the elasticity modulus of a real dam concrete from the experimental results of tests carried out on standard wet-screened concrete specimens (e. σb is the initial stress (for ASTM C469: stress corresponding to a strain εb of 50 µε) εa is the mean strain under the upper loading stress: εb is the mean strain under the initial loading stress (for ASTM C469: 50 µε).3. 3.4 fc). is also effective in order to stabilise the concrete behaviour and reduce the creep effect that can be induced by the strain-time diagram used for the determination of the elastic modulus. it was found that this aim can be easily obtained by simply using the wet-screened concrete data. along with estimations of the aggregate modulus. with loading and unloading cycles. The static modulus of elasticity in compression. An interesting aspect of the stress strain curve obtained from such a test is that the sample fails suddenly shortly after the maximum load is obtained. march 2008 Section 3-13 .4. It is less complicated and time-consuming than testing on large size specimens. The modulus of elasticity in tension and flexure is typically taken to be equal the value obtained from compression tests.120 mm) as the aggregate phase. On the basis of an experimental research on dam concrete. in MPa (for ASTM C469: σa = 0.17]. Such a response is related to the properties of the testing machine rather than the behaviour of the tested concrete specimen. the elastic moduli calculated for the real size concrete were quite close to those experimented on large size specimens (45x45x90 prisms with aggregate 0-120 mm) of real concrete.2.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) This preloading procedure. is given by the formula ∆σ σ a − σ b = ∆ε εa − εb where σa is the upper loading stress. Data from tension and flexure tests indicate that this is a reasonable approach. in a multiphase concrete model such as the Hirsch model presented in paragraph 3. cylinders of 15x30 cm) is presented in [3. in MPa. Ec.g. By taking the matrix phase as the wet-screened concrete (15x30 cm cylinders with aggregate size up to 40 mm) and the gravel above 40 mm (40 .

British Standard BS 8110 and Eurocode 2 (Design of concrete structures) are here presented.17].6% 2. Eurocode 2 (Design of concrete structures) [3.0 37.20]: E=9.7 41.37 fck 1/2 This formula applies for normal density concrete with unit weight of 2.8% Error 7 3.1 – Comparison between the predicted moduli of the real concrete (Hirsch model) and the experimental ones (at different ages and for two aggregate moduli) Age (days) Em (GPa) Wet-screened concrete 24.3% 7.4 42.6 38.3 43. 3.5 35.5% 2. Among the relationships for NSC the equations of American Concrete Institute (ACI Building Code 318).7 40.2 32.2 41.21]: E=9. where the fck is the characteristic compressive strength (in MPa) from 15x30 cm cylinders and E is the average modulus of elasticity in compression (in GPa) at the same age (28 days).5% -10.1.0% -3.3 Ec (GPa) Real concrete experimented 30.300 kg/m3 2. fck 1/3 3. In this case the calculated error is less than 10%.19]: E=4. The British Code of Practise (CP 110.3 -0. Different national building codes propose various formulas for normal strength concrete (NSC) and high strength concrete (HSC).3 Empirical expressions through correlation with the compressive strength The concrete modulus of elasticity is usually and widely expressed as a function of concrete uniaxial compressive strength.4. 3.5 * (fck +8) 1/3 As submitted for ICOLD review. ACI Building Code (ACI Committee 318) [3.8 34.2 43. Part 1) [3. 1.1.2 Ea (GPa) Aggregate 50 65 28 90 180 50 65 50 65 50 65 Ec (GPa) Real concrete predicted 30.2. More detailed information on this procedure can be found in [3.2 38.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) A synthesis of the comparison between the experimented and calculated results is reported in Tab.0% -4.4% 10. march 2008 Section 3-14 .1 37. Tab. with two different values of aggregate moduli (50 and 65 GPa).

9) 36 (5.000) 27 (4.2) 6 Tab.500) 60 (9. if fck is replaced by the effective strength at the considered time. Tab.6) 28 (4.000) 30 (4.000) 67 (10. Tab.2.3 shows the elastic moduli calculated according to the Eurocode 2 for the different concrete grades. 3. 3.2 .5) 34 (4. 3.7) 6 British Code of Practice (CP 110) fck MPa (psi) 20 (3.21] Concrete grade C 12/15 C 16/20 C 20/25 C 25/30 C 30/37 C 35/45 C 40/50 C 45/55 C 50/60 E (GPa) 26 27.4) 35 (5.5 35 36 37 As submitted for ICOLD review.000) E GPa (x 10 psi) 21 (3.500) 40 (6. Typical estimates of static modulus of elasticity obtained from the ACI and British codes is given in Tab. from C12/15 to C50/60 (cylinder strength / cube strength in MPa). march 2008 Section 3-15 .1) 39 (5. compared to their compression strength. Information on concrete mix design and time of testing are also provided.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) According to the Eurocode 2.5 29 30. Tab.8] ACI Building Code 318 f’ck MPa (psi) 20 (3.1) 31 (4. Common to all equations is a high coefficient of variation.Typical estimates of static modulus versus compressive strength [3. 3.000) 40 (6. the elastic modulus can be calculated even at a later age.6) 30 (4.4 shows modulus of elasticity values for concretes from German and Swiss dams.000) 50 (7.000) E GPa (x 10 psi) 25 (3.3 – Elastic moduli for different concrete grade according to Eurocode 2 [3.5 32 33.000) 53 (8.1) 25 (3. 3.

Water ca.7 15.3 3) 365 44 [3.23] Mauvoisin Switzerland.8 35. NMSA 63mm. 1991 90 29.8-15.4 . 1961 90 16. strength C20 on 200x200x200mm cubes Cube compr. C 140kg/m3.7 Cylinder compr.1 31. NMSA 100mm. Water ca. C 3 140kg/m Static Elastic Modulus Age [d] 4 7 28 90 17. strength C20 on [3. C FA 120kg/m3. 50kg/m3.7 [3.5 365 30.7 365 32. Remark Leibis Lichte CVC Gravity Germany.23] Grand Dixence CVC Gravity Switzerland.0 Emosson CVC Arch Switzerland.5 29.6 365 13. C 3 140kg/m . NMSA 125mm. 1956 90 32. 110kg/m3 90 25.23] CYL(16x32) d=160xh=320mm cylinder Cylinder compr. NMSA 150mm. 40kg/m3. flexural strength.0 27.9 [3. 1959 Granite.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) Tab.2-37. C FA 150kg/m3. strength on [3. NMSA 150mm.23] 200x200x200mm cubes Cube compr.8 34. C 140kg/m3. strength C30 on 300x300x300mm cubes Zeuzier CVC Arch Switzerland. cube compr. C 160kg/m3.7 21.9 As submitted for ICOLD review. Water ca.3 30. NMSA 100mm.9 27.2 Leibis Lichte CVC Gravity Germany.3 26.5 [3. Water ca. 130kg/m3 Gneiss. Water ca. modulus on 200x200x800 mm beams Cube compr. 2004 Quality control programme.23] CYL(30x45) d=300xh=450mm cylinder Cube compr.4 24.3 365 24. 130kg/m3 Limestone.7 30. modulus on 200x200x800 mm beams Quality control programme. NMSA 100mm. 105kg/m3 Gneiss. NMSA 120mm. strength C30 on 300x300x300mm cubes 4 7 28 90 20. 1974 90 27.Compressive strength vs elastic modulus from German and Swiss dams Dam Type Cube Compres. strength C20 on 200x200x200mm cubes Cube compr.23] CYL(30x45) d=300xh=450mm cylinder Moiry CVC Arch Switzerland. 2004 Albigna 90 25.22] 300x300x300mm cubes. 135kg/m3 Gneiss.5 Cylinder compr. C 140kg/m3.6 32. Water ca. 120kg/m3 Gneiss.6 30. split tensile strength. 115kg/m3 Granite. 1957 90 30. strength on [3.0-28. strength. Water 103kg/m3 7 28 90 180 360 3 Greywacke.23] Mauvoisin CVC ArchGravity CVC ArchGravity (Height ening) Switzerland. Water 110kg/m3 7 28 90 180 360 CVC Gravity Switzerland. C 3 140kg/m . 1958 Gneiss.0 E [GPa] Ref.9 16. Strength Country Mix Design Age [d] 3 Greywacke. 3.5 365 38. cube compr.22] 300x300x300mm cubes.23] Zervreila CVC Arch Switzerland. 1957 90 23. NMSA 120mm. flexural strength. march 2008 Section 3-16 . density on [3. density on [3.2 Project fc [MPa] 10. split tensile strength.3 / 18. strength. C 140kg/m3. NMSA 125mm.2 10. Water ca. strength on [3.0 / 32.

row) and 36 MPa strength (2. 3. av. Such a note of caution certainly can also be applied to the ratio between seismic and static moduli.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) 3. march 2008 Section 3-17 . are summarised in Tab.28 1.Results from investigations for seismic moduli of elasticity [3. • As submitted for ICOLD review. Bureau's of Reclamation test series Big Tujunga dam.5. Tab.18] states that the modulus of elasticity values obtained under rapid load application (dynamic or seismic rates) are usually higher than moduli in the static conditions.57 1. 3. In the dynamic condition.38 0. row) Bureau's of Reclamation test series Crystal Springs Dam Tests on Swiss dams extensive testing for Zervreila Dam (Switzerland) Reference For a better evaluation of the results of such investigations. The dynamic modulus can be considered as approximately equal to the initial tangent modulus while the static modulus is equal to the cord modulus (Fig. 105 that “… indiscriminate application of increased [tensile] strength criteria in analysis should be treated with caution as incipient cracks in most dams exist due to static thermal loading…”.48 1. ASTM C469 [3.10]. the observed high scatter comes mainly from difficulties to test specimens under short rates of load application and from the influence of the individual granular texture and size distribution in the generally small test sections. The ICOLD Bulletin 52 [3.24] Dynamic/static Ratio max. cores drilled from the dam.89 1.3).25] mentions on p.5 .66 min. the following considerations can be drawn: • In general.3 DYNAMIC MODULUS OF ELASTICITY An important characteristic of concrete is its behaviour under short term loading such as experienced during earthquakes (e.g.32 0.15 1.1 Modulus of Elasticity under Seismic Loading Several investigations were carried out to study values and relations between static and seismic moduli. 3. specifically concerning dam concrete.42 1.04 1. 3. 2 concrete types: 25 MPa (1.18 1. the strain behaviour of concrete is not influenced by the micro-cracking and creep. strain rates of 10-3 to 10-2 /s)./single value 1. Some of them.10 1.25 1.3. as it happens in the case of static modulus in presence of applied stresses [3.

for example ASTM C 597 [3. which may broadly be characterised in Finite Element Analysis by a lowering of the equivalent elastic modulus (together with an increase of damping). In the ultrasonic pulse velocity method an ultrasonic pulse is emitted by a transducer coupled to the structure surface. A receiving device detects the ultrasonic pulse and the transit time is determined. [3. typically around a value of 25 per cent. This will result in hysteresis effects.89 with a coefficient of variation of 17%”. Ed can be expressed as follows: E d = ρ *V 2 (1 + ν ) * (1 − 2ν ) (1 − ν ) where: ρ = concrete density V = pulse velocity ν = Poisson coefficient Two ultra-sonic test methods are currently used to determine the concrete modulus in the high frequency range i.8]. A conservative approach would therefore tend to use lower bounds when applying corresponding reference data or test results. “Consequently the average dynamic modulus did not tend to increase as the dynamic strength increased”. The cross As submitted for ICOLD review. Another characteristic for concrete exposed to reversed straining. 3. as it is common during high intensity earthquake loading. This means that the reliability of assessing values for the modulus of elasticity under seismic loading is lower than that for static loading. Other authors report figures as high as 40% [3. All this demonstrates that the ratio between seismic and static modulus of elasticity is burdened with a high scatter.3.2 Dynamic Ultrasonic Modulus of Elasticity The dynamic ultrasonic modulus of elasticity of concrete.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) • • Nagayama et al. The velocity is used to compute the dynamic modulus for high frequency excitation. The range is dependent on the transmitter energy with a range of 3 m for low powered systems and 10 m for high powered systems.27] comment about test series with cores from 10 US dams: “the average dynamic to static modulus of elasticity was 0. report in [3. The determination of the dynamic modulus of elasticity Ed is based on the studies on the sound wave propagation and is related to the propagation velocity of longitudinal stress wave pulses through concrete.26 that the modulus of elasticity is not affected by loading rate. march 2008 Section 3-18 . The path length is known and the ultrasonic pulse velocity is computed. ultrasonic pulse velocity method and the cross hole sonic logging method.e.. is that irreversible plastic straining will occur together with reversible elastic deformations. Harris et al. as determined according to the national standards. is greater than the static modulus.28].

Also.The Poisson coefficient as ratio between lateral and axial strain [3. non-destructive sonic methods can also be used also for compressive strength evaluation and diagnostic investigations on concrete in existing dams.11]. 3. Poisson’s ratio increases rapidly due to the level of cracking within the section.15 to 0. 3. 3.1] Under high loads. In fact the resonant frequency of a structure is directly related to its dynamic modulus and hence its mechanical integrity.21 (and generally from 0.11 to 0.30].20) based on strain measurements for normal concrete [3.8).6): As submitted for ICOLD review. However this response is non linear and indicative of near-failure conditions. for example the ASTM C215 [3.4 POISSON’S RATIO Poisson’s ratio is the ratio between lateral strain accompanying an applied axial strain (Fig. 3. Poisson’s ratio must be determined from strains which are within the elastic range of the stress-strain curve. Under static conditions. the Poisson’s ratio varies in the range of 0. and torsional frequencies of concrete cylinders and prisms. Another possible test method for determining the dynamic modulus of elasticity is that based on measurements of the fundamental transverse.8 . march 2008 Section 3-19 . for static stresses below 40 percent of the compressive strength. As already stated in Section 2 (Strength Properties). Poisson’s ratio is related to the elastic modulus E as follows (Tab. National Standards cover this testing method as. This method is useful if a particular zone in the structure is to be investigated. longitudinal. Fig.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) hole sonic method is a down-hole version of the previous method.29]. In the German Specification DIN 4227 [3.

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) Tab.24. 3. As submitted for ICOLD review.20.07 with coefficient of variation of 31%.000 0. Again this demonstrates the high scatter inherent in dynamic test methods. A dynamic determination of Poisson’s ratio yields an average value of 0. therefore this also indicates that Poisson’s ratio will vary from 0.17-0.7 shows modulus of elasticity values for concretes from US and Brasilian dams. However the tests carried out by Harris [3. 3. Tab.15 to 0. It is therefore not recommended increasing the Poisson ratio for dynamic loading from 0.000 0. Poisson’s ratio is higher for dynamically loaded concrete.15-0.25 The elastic modulus of concrete in mass concrete is typically in the range of 24 to 30 GPa.18 30.20 35. Information on concrete mix design and time of testing are also provided.15/0.20-0.000 0.27] from 16 dam core samples showed an average ratio dynamic/static Poisson of 1.24 which is significantly different than the static value.Typical estimates of compressive strength versus Poisson’s ratio E (MPa) ν 24.20 to 0. together with their Poisson ratio.6 . march 2008 Section 3-20 .

22 0. P 3 53kg/m .0 34. Water 3 88kg/m Limestone.0 30.0 37.0 37. NMSA 150mm.0 41.20 0. Andesite. P 3 56kg/m .0 180 28 90 180 90 180 365 28 90 180 365 0.24 0. Age [d] 28 90 365 28 90 365 28 Ratio [-] 0. Tuff. C 3 225kg/m .31] Morrow Point CVC Arch USA. NMSA 150mm. march 2008 Section 3-21 . NMSA 225mm. C 3 221kg/m .27 0.0 42.Elastic Modulus vs Poisson’s ratio from US and Brazilian dams (part I) Dam Type Static Elastic Modulus Age [d] Limestone & Granite. C 3 117kg/m . Water 3 83kg/m Limestone. Water 3 134kg/m Limestone. C 3 224kg/m .7 .0 47.26 0.0 90 0.17 0. Water 3 130kg/m Basalt.31] 180 28 90 180 90 180 365 28 90 180 365 46. C 3 111kg/m .22 0.20 0.31] Grand Coulee CVC Gravity USA.0 43. 3.20 0.25 0. C 3 111kg/m . 1962 [3.20 Flaming Gorge CVC Arch Gravity USA. 1942 [3.0 37.0 32.23 0. NMSA 150mm.19 0.31] Glen Canyon CVC Arch Gravity USA.31] Yellowtail CVC Arch Gravity USA. Water 3 82kg/m Andesite. Sandstone. NMSA 114mm. Sandstone. Basalt. 1965 [3.0 42.21 0.15 [3.0 30. Water 3 93kg/m 28 90 365 28 90 365 28 E [GPa] 38.31] As submitted for ICOLD review. NMSA 150mm.15 Project Country Mix Design Hoover [3.0 32. Chert.0 Poisson Ratio Ref.0 32.0 43.0 24. 1967 [3.18 0. P 3 50kg/m .13 0.23 0.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) Tab. 1963 90 43.

Water 3 79kg/m 28 22.31] 90 27.0 28 0.0 45.14 [3.21 0. C 3 128kg/m .16 Itaipu Hollow Brazil / Gravity Paraquay.0 37. P 3 32kg/m .22 0. 3. C 3 88kg/m . Basalt. 1973 Basaltic Gravel.21 [3.0 43.0 365 Ilha Solteira CVC Gravity Brazil. 1995 Granite. NMSA 150mm. Water 3 97kg/m 90 26.0 28 0. 1972 Granite Gneiss.0 43. NMSA 150mm. Water m3 85kg/ 28 90 180 365 28 90 365 38.0 28 0. C 3 125kg/m .Elastic Modulus vs Poisson’s ratio from US and Brazilian dams (part II) Dam Type Static Elastic Modulus Age [d] E [GPa] Poisson Ratio Ref.20 0.0 90 0.31] 365 26. Buttress 1982 Basalt.18 Dworshak CVC Gravity USA. Water 3 82kg/m 28 35.18 0.31] 90 41.0 90 [3. P 3 27kg/m . NMSA 150mm.21 0.31] 90 28. P 3 13kg/m . 3 C 82kg/m . Water 3 82kg/m 28 19. NMSA 150mm. NMSA 100mm. Age [d] Ratio [-] Project Country Mix Design Lower Granite CVC Gravity USA.19 [3.0 90 0.0 90 0.15 [3. march 2008 Section 3-22 .ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) Tab. NMSA 150mm. P 3 42kg/m .0 31.0 43.31] Theodore Rossevelt Modificatio n CVC Arch Gravity USA.31] As submitted for ICOLD review.20 Libby CVC Gravity USA. P m3 29kg/ .8 . C 3 108kg/m . P 3 29kg/m . 1974 Quartzite Gravel.0 28 90 180 365 28 90 365 0. Water 3 85kg/m [3. C 3 86kg/m . 1972 Quartzite Gravel.20 0.

5 SIGNIFICANCE OF ELASTIC PROPERTIES ON DAM BEHAVIOUR The modulus of elasticity is an important concrete parameter. For these structures.18]. As already discussed in paragraph 3.17]. in order to avoid extreme situations that are not expected.2. From a series of laboratory modulus data at different ages (up to 365 days) from more than 50 Brasilian dam concretes [3. problems could arise in transferring direct loads to the foundation rock or in reacting to imposed deformations like thermal loads. over a period of time with constant load. except for the testing time that can attain two years or more. and could be even faster than for compressive strength (Section 2). Sometimes this last one is generally accounted for by determining a “sustained” modulus of elasticity to be used in static analyses [3. This fact could also have a detrimental effect in highly hyper-static structures such as dams and especially arch dams that have a complex relation with their foundation. it was found that. the sustained modulus of elasticity should be evaluated from the concrete specimens with the same testing procedure prescribed for the instantaneous elastic modulus [3. it is recommended to consider the possible increase of the concrete elastic modulus already in the design stage. compared to those predicted in the design phase. the average ratio between the elasticity modulus at 365 days and at 90 days is about 1. As submitted for ICOLD review. march 2008 Section 3-23 . Thus. the elasticity modulus has an appreciable increasing trend with time. Single values can even attain ratios in excess of 1.3. especially when their design is based on elasticity considerations [3. Different stress levels and distributions. in fact. These sustained moduli are further discussed in Section 4 (paragraph 4. under constant sustained loads.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) 3. This problem should be taken into account also in the monitoring and interpretation of the dam behaviour [3. The subsequent development of the deformation.5 modelling creep in structural analysis). vital for structural analysis and for evaluating the strain distributions and deformations in concrete dams. particularly in pozzolanic cement concrete.2. in the presence of mineral admixtures.5.5. can appear and cracks could develop.32]. is the result of creep.33]. From the experimental point of view.34]. The concrete deformation that occurs immediately after the load applications depends on the “instantaneous” elastic modulus.

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) 3.4. • • In this respect the following references to other Sections of the Bulletin are in order: ¾ reference to Section 2 .8 Use of tensile strength in mathematical models linear elastic analysis and non linear analysis ¾ reference to Section 4 . Tangent moduli (O-A) are required for modelling when non-linearity is taken as a rate-dependent (e.Definitions of apparent moduli of elasticity used in mathematical models • Static correlations: medium-term creep strain is often included in the secant modulus (O-D). 3.5 Modelling creep in structural analysis (concept of sustained modulus of elasticity) ¾ reference to Appendix C on AAR . 3.C.9 illustrates that different loading rates and situations result in different “apparent elasticities”.6 USE OF ELASTIC PROPERTIES IN MATHEMATICAL MODELS FOR DAM STRUCTURAL ANALYSIS The degree of refinement of the mathematical model used for static and earthquake loading influences the choice of the modulus of elasticity. Most important is the effect of rate of loading as it both influences strength and elasticity.C. The shown relations refer to mathematical modelling as follows: Fig. Modulus of elasticity.2.3. march 2008 Section 3-24 .9 . As submitted for ICOLD review.36]. but for some models creep can be explicitly calculated with respect to the initial tangent elasticity [3.4.g. differences exist between static (O-C) and dynamic (O-B) laboratory tests due to strain rate dependency. visco-plastic) phenomenon. These differences can be modelled by a visco-plastic model [3.35].4. Fig.

as suggested by the Bureau of Reclamation [3. derived from the best fitting of experimental behaviours of sample population. As an example. march 2008 Section 3-25 . As submitted for ICOLD review. deterioration. characterised by accumulated non recoverable strains. The zone above this level represents a transition zone to non-linear behaviour.10 typical experimental compressive-stress-strain curves for a Bureau of Reclamation dam are reported.37]. 3. data obtained from standard elastic material tests are still important because they can be used. in several conditions such as seismic loading. However. to provide a comprehensive stress-strain diagram. This is particular important if damage prone conditions are to be expected for existing dams. loading beyond capacity. is that below the stress corresponding to 40% of the average ultimate strength. together with a comprehensive “failure state diagram”. It allows a rational evaluation on the need for either linear or non-linear modelling. In this respect a reference to Appendix A “Fracture energy: application for dams” is suggested (A. able to represent different failure states of the concrete. Reference [3. the structural dam behaviour is better modelled with non-linear analysis. with an elastic pattern.18].ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) Even if linear elastic analysis is the standard method for dam structural analysis.37]. Both linear and non-linear laboratory tests are recommended in this zone.11). based on loading and unloading cycles (Fig. Post failure tests and fracture energy special testing are required for the cracked zone.3 – Determination of fracture parameters).38] uses the principles of damage mechanics to analyse non-linear static or dynamic behaviour. in Fig. Some supplemental tests to represent the non-linear behaviour is suggested by [3. as defined by the ASTM C469 [3. The working stress zone. 3.

37] As submitted for ICOLD review.10 . 3.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) Fig. march 2008 Section 3-26 .Compressive-stress-strain curves for the Deadwood dam (Bureau of Reclamation dam) and the obtained “failure state diagram” [3.

11 – Loading and unloading cycles testing of concrete cores to evaluate non linear pre-failure parameters [3. march 2008 Section 3-27 . 3.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) Fig.37]. As submitted for ICOLD review.

IV. G. [3.cement. Montréal.. and Related Topics. CP 110. S.10] The University of Memphis. 207.14] American Concrete Institute. Prentice-Hall. Cortright.P. [3. “Poroplastic Analysis of Concrete Dams and their Foundations”.. and Tardieu. McGrow-Hill Book Company. 1996.5] Chen.F. Carrere. Aguado. 1989. 1993.7 REFERENCES [3. pp. London. "Standard Test Method for Static Modulus of Elasticity and Poisson's Ratio of Concrete in Compression”. “Concrete: Structure. Prentice Hall. [3.2] Timoshenko. [3. 181-197. London. Agullo. S. Coussy.7] Franklin Dusseault "Rock Engineering" (1991). Ed. L. O. Michigan.3] Neville. 1982. Dilger.17] Vilardell. Dam Engineering. Upper Saddle River.R. Gettu.1] Cement Association of Canada . march 2008 Section 3-28 .html.ca/cement... 1991. “Properties of Concrete”. Specifications. Englewood Cliffs. 1986. 1998. A.H. M. McGrow-Hill Book Company. [3.Y. S.6] Fauchet. “Design of concrete structures“.13] Geen. [3. [3..4] Bangash.nsf/ [3. Young. J. 93-101.edu/1101/notes/concrete/everything_about_concrete/11__compositee. S. “Concrete and Concrete Structures: Numerical Modelling and Applications”. New York. 2003. and Brooks..1963. 321.J.F. Second edition. 28.1R-96. [3. C. W.: Concrete. B. [3. Inc. “Ensayos y control de hormigones”. ETA [3. J. R. Issue 3. Technical Report AFWC-TR-72.. ACI SP 55-6.. http://www.21] Eurocode 2. 1983.H. “Creep of Plain and Structural concrete”.P. and Goodier. New Jersey. D. Vol. Proceedings of 10th ICOLD International Congress. B.J.. n° 1. M.memphis.. Cement and Concrete Research. [3. W. [3.N..11] Neville. L. Longmann Group Limited.. Detroit. Elsevier Applied Science.S. Vol. 1934. [3. ACI Manual of Concrete Practice Part 3: Use of concrete in Buildings – Design. Part 1 [3.19] ACI 318-95 Building code requirements for structural concrete. [3. [3.M...9] Mindess.. Darwin. and Swanson. ACI Publication. “Mass Concrete”.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) 3. K. J. New York. A. Construction Press. [3.20] British Code of Practice. Properties and Materials”. "The Nature of Mass Concrete in Dams". 1970. J. [3.J.16] Tuthill. p. 2002. As submitted for ICOLD review.ce.. ”Theory of Elasticity”. “Plasticity in Concrete”.12] Gorisse. A. Department of Civil Engineering. Volume II.. ISBN 0-13064632-6. “Static Constitutive Model for Concrete in compression”. pp.http://www. Sir Isaac Pitmann & Sons LTD. [3. “Estimation of the modulus of elasticity for dam concrete”. F. A.. Sarkaria...8] Metha.15] Raphael J.18] ASTM C469-02e1. 1973.

3. R. P. American Concrete Institute ACI. D.9. Rotterdam. “Numerical Models in Geomechanical Engineering Practice”.28] ASTM C597-02. May-June 2000.A. C. R.23] Kreuzer. Issue 2.35] Dungar. 2001. "Failure states from concrete stressstrain data for use in analysis”. T. 2001. 1998... [3. “The effect of cyclic creep on the ageing of arch dams”. 2005”. personal comunication. [3.32] Bureau of Reclamation “Design criteria for concrete arch and gravity dams”. Concrete Data Basis for Bulletin "Concrete of Swiss Dams". Vienna.37] Harris. Wiley. New York. " Standard Test Method for Pulse Velocity Through Concrete”. M.W. Dolen T. Longitudinal. A... Beton-Informationen 2/3 2004.31] ACI 207. Engineering Monograph n°19. prestressed lightweight concrete structural members”. [3. 2007.12.29] ASTM C215-02 “Standard Test Method for Fundamental Transverse.36] Dungar. 1986. nd As submitted for ICOLD review. [3. 1985. and Torsional Frequencies of Concrete Specimens”. R. [3..22] Wagner.25] ICOLD Bulletin n° Earthquake analysis for dams. [3. W.24] US Bureau of Reclamation: "Concrete Manual" .A Balkema. Studer. [3..34] Filho J.30] German Specification DIN 4227..-P.P.. J. E.26] “Experimental study on tensile strength of concret for dams”. Vol. personal comunication. 2004..M.1R-05 Mass Concrete. Deutscher Beton-Verein. a water resource technical publication. [3.8th ed.. Hydropower & Dams.33] Buil Sanz. [3. Proc. Dolen. J. [3.27] Harris. H. 1999. Q.2007. [3. Congress. Mohorovic.E. [3. ICOLD. Proc.65... J. 2002. 1981. march 2008 Section 3-29 . D. ACI Materials Journal. [3. Technical Memorandum of PWRI. [3. [3.38] Gunn R. “Non-linear design and safety analysis of arch dams using damage mechanics”.. partially published. J. J. N . 1991. of 2 USJapanese Workshop on Advanced Research on Earthquake Engineering for Dams. M.. 17. "Dynamic Properties of Mass o Concrete Obtained from Dam Cores". “ Prestressed concrete. M.. Massenbetone für die Gewichtsstaumauer Talsperre Leibis/Lichte. Email Communication with M. 23. [3.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 3 (Elastic properties) [3. Conrad. 97. 52. Swiss Committee on Dams. Mohoroovic C.

.................1 GENERAL ...................................................2............................................3 4.............2 Cement ......................................................2.....................................................3 MEASUREMENT OF CREEP.6 Ambient humidity .......................................2..5 MODELLING CREEP IN STRUCTURAL ANALYSIS ...............2..............2.................6 4..............................6 4.....................................................................................4 Age at loading...................1 Aggregate .....................................................2.................................................................................................................................................8 Specimen size .......................5 4... 1 4..................................................................................3 4...............................4 4.......................2 4.....................6 4..................................................7 Temperature .......2...............3 Strength and mixture proportions ................................................... march 2008 Section 4-1 ................................4 4......................2..6 REFERENCES .2 FACTORS INFLUENCING CREEP..2...................................5 Level of stress .............14 As submitted for ICOLD review.............................................5 4......9 Creep under different states of stress................................5 4......................................................ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 4 (Creep properties) 4 CREEP PROPERTIES 4 CREEP PROPERTIES.........................11 4..............4 SIGNIFICANCE OF CREEP PROPERTIES ON DAM BEHAVIOUR .................................................................9 4..........6 4.......

The movement of water into and out of the gel in response to changes in ambient humidity produces the well-know shrinking and swelling behaviour of concrete.1 GENERAL Creep is time-dependent deformation due to sustained load [4. based on certain observed mechanisms of deformation.1.1]. have been proposed. the elastic instantaneous strain ∈inst is recovered. 4. In creep. If the load were maintained. Although several hypotheses of creep. and some creep recovery is seen to occur. If the load is removed. the instantaneous and creep deformations develop again as show [4. Fig. it is generally accepted that concrete creep is a paste property.4].e.3] [4. i.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 4 (Creep properties) 4. as show by the dashed curve. a rheological phenomenon associated with the gel-like structure of cement paste [4.2] This particular concrete was loaded at age 28 days with resulting instantaneous strain ∈inst. If the concrete is reloaded at some later date.Typical concrete creep curve [4. march 2008 Section 4-2 . The load was then maintained for 230 days.2]. 4. The nature of the concrete behaviour can be shown schematically in Fig. during which time creep is seen to have increased the deformation to ∈creep which is almost 3 times its instantaneous value. gel water movement is caused by changes in applied pressure instead of differential hygrometric conditions between the concrete and its environment.1 . the deformation would follow the solid curve. As submitted for ICOLD review.

Creep of concrete. and degree of hydration [4.2) As submitted for ICOLD review.1) where cp is creep of neat cement paste of the same quality as used in concrete and α is defined as follows: α= 3(1. are related by [4.7]: cp 1 log = α log c 1− g (4. compaction.2 FACTORS INFLUENCING CREEP Creep is influenced by properties of aggregate and cement. It is partly reversible and partly irreversible 4. This concept offers explanations for the linearity of creep strain over a wide range of stress. the reduction in strain rate with time. curing conditions.3]. and the volumetric content of aggregate. properties of aggregate.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 4 (Creep properties) This concept is supported by the similar manner in which creep and shrinkage are affected by such factors as water-cement ratio. Creep deformation can also be explained partly in terms of visco-elastic deformation of cement paste and to the gradual transfer of load from the cement paste to the aggregate. and by strength. This section briefly summarises the current state of knowledge regarding the major factors that influence the creep of concrete.1 Aggregate Since the cement paste is subject to creep and the aggregate generally is not.2. level of stress. and conditions of storage of the concrete. mixture proportions. size and ambient temperature conditions. age of loading. and E = Modulus of elasticity of the neat cement paste E Ea (4. In basic creep no exchange of water from the concrete specimen to the environment is possible and only permanent and irreversible strain occur.µ ) 1+ µ + 2(1− 2µa) where µa = Poisson’s ratio of aggregate µ = Poisson’s ratio of the neat cement paste Ea = Modulus of elasticity of aggregate. c. Drying is the other creep that develops in concrete. and the sensitivity of creep to temperature [4. 4. Creep has been classified as “basic creep” and “drying creep” [4. march 2008 Section 4-3 . It is not influenced by humidity. Most of the creep studies that have been conducted have been for the purpose of determining the effect for one or more of these variables on creep.6]. the effect of aggregate is to reduce the effective creep of concrete.5]. mixture proportions. g.

4. but approximate. granite. at the time of loading.2.2). which influences creep. and II. the higher the modulus. Therefore. The influence of the cement type on creep is less significant for dams than for common civil structures due to the late hydrostatic loading (early creep only due to self weight). Fineness of cement affects the strength development at early ages and thus influences creep.2 Cement Creep is generally not influenced by the type and composition of cement if the concrete.3 (Tab.2. for a constant applied stress at the same early age. 4. Mass concrete for dams commonly contains large size (e. Type IV the lowest early strength.9] [4. is proportional to the amount of paste [4. Tests indicated that the influence of the type of rock of which the aggregate is composed has an effect on creep primarily through the modulus of elasticity of the rock [4.. the greater the restraint offered by the aggregate to the creep of concrete. V and IV. This finding has made it possible to develop significant data for mass concrete from small specimens. is fully hydrated. 150-mm) aggregate.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 4 (Creep properties) As evident from Eq. I.7]. having identical water-cementitious material ratio and air contents in the mortar phase but different amounts and size of a given coarse aggregate. 4.8]. Concrete made with sandstone aggregate exhibited creep more than twice as much as concrete made with limestone aggregate. Creep as a function of gain of strength is also valid for all kinds of blended cements.10]. Since the usual portland cements differ from one another primarily in the rate of hydration (gain of strength). Examples of creep data in dam concretes characterised by different types of aggregates are shown in 4. type III has the highest. creep increases in order for ASTM Types III. The relationship between creep and cement paste content is of particular interest in the field of mass concrete for dams. measure of the water-cementitious material ratio and degree of hydration of the cement in concrete. Fabrication and testing of concrete specimens containing such large aggregate are expensive. quartzite. As submitted for ICOLD review. (4. Typical rocks used as aggregates in the order of increasing creep are limestone. march 2008 Section 4-4 . and sandstone. creep is inversely proportional to the strength of concrete. For paste contents normally used in mass concrete. it can also be stated that creep increases with an increase in water-cementitious material ratio of the concrete mixture. the modulus of elasticity of aggregate affects the creep.e. the creep of sealed specimens. The amount of gypsum in the cement may affect creep in a manner similar to its influence on shrinkage [4.g.3 Strength and mixture proportions Strength is a convenient. I. basalt. This means that finer cements need more gypsum. For a constant cement paste content and the same applied stress.1).

4. the magnitude of creep will be independent of the RH [4.6 Ambient humidity Creep increases with a decrease in the ambient relative humidity (RH) as shown in Fig. 4. Fig.12].2 [4.4 Age at loading Age of concrete at loading is a factor in creep in so far as age influences the degree of hydration and the development of strength. 4. and there exists a stress-strength ratio above which creep resulting from sustained load causes failure [4. creep increases with an increase in stress at an increasing rate.40.11]. 4.8 to 0.9.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 4 (Creep properties) 4.2.Creep of concrete at different relative humidities [4. If concrete has reached moisture equilibrium prior to loading. Strictly speaking. At later ages when the degree of hydration remains substantially constant.2 .2. the rate of creep becomes independent of the age at loading. the ambient RH only affects creep if drying takes place while the specimen is under load.6].12] As submitted for ICOLD review. Above the limit of proportionality. This stress-strength ratio is in the range of 0.5 Level of stress It is generally accepted that there is a nearly linear relation between creep and the applied stress up to stress-strength ratios of approximately 0.2. march 2008 Section 4-5 .35 to 0.

69:0.84:0.13].17].9] that creep of sealed specimens is independent of specimen size. it is evident that there must be a size effect associated with the moisture gradients within the specimen. Test results indicate that the rate of creep increases approximately 3.2. This is advantageous against the development of tensile stresses at the upstream toe of dams. Creep was less under multi-axial compression than under a uniaxial compression of the same magnitude in the given direction [4.5 times when temperature increases from 21C to 70 C [4.15]. triaxial. march 2008 Section 4-6 .ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 4 (Creep properties) 4.16]. The rate of creep increases with increasing temperature. It is clearly not possible to reproduce these conditions exactly. Creep in saturated concrete is larger than in dry concrete. 4. can be maintained though loaded springs or hydraulic systems. Both cylinder and prism creep specimen As submitted for ICOLD review.2. This is similar to the behaviour of concrete at alternating humidity.00:0. biaxial. creep tests should be carried out in conditions appropriate to those existing in the structure.14]. 4. particularly in the case of the temperature. respectively [4.3 MEASUREMENT OF CREEP The method normally used to measure creep is based on the application of a constant load to concrete cylinders or prisms placed in a loading frame. The strains can be measured periodically by means of internal strain gages or by demountable mechanical extension gages applied to the external faces of the prism.7 Temperature The effect of concrete temperature on creep is of importance for mass concrete in dams where later temperatures can be significantly higher than the concrete placement temperatures due to heat of hydration.4]. As regards to the humidity conditions for mass concrete in dams. Laboratory investigations have shown that when the temperature alternates creep is higher than at either extreme of the alternating temperatures [4. 4.9 Creep under different states of stress Most creep testing has been concerned with specimens subjected to compression.45 for uniaxial. which will vary. it is generally considered appropriate to carry out the test in a sealed specimen to prevent moisture loss from the concrete (basic creep). It was reported that shrinkage and creep were dependent only on the ratio of surface to volume but independent of the shape of the specimen [4. The constant load. which may be applied for several months or even years. For unsealed specimens exposed to drying environment. and hydrostatic loading.8 Specimen size It has been demonstrated [4. Loads may be applied at different ages and can be decreased or removed when desired to investigate creep recovery. McDonald reported that the average values of compressive creep differed by a factor of 1.2. Since the amount of creep is influenced by humidity and temperature. For creep under tension there are only conflicting results reported [4.

19]. E = instantaneous elastic modulus at the age when it is first loaded. Other valuable creep test results on dam concrete can be found in references [4. Some creep data from Brazilian concrete dams are reported in reference [4. Such data can be represented pictorially by a creep surface such as that shown in Fig. in Fig.5]. which are kept in the same environment as the creep test specimens but which remain unloaded. or neoprene or butyl rubber encasements. Seasonal reservoir storage schemes are typically subjected to cycling load variation depending on the year period.18]. Tab. F(k) = creep rate. strains resulting purely from shrinkage are measured directly. Many sets of data show an approximately straight line over a considerable period of time. These accumulated effects are to be taken into consideration.23] and in reference [4. 4. The usual approach is to determine the shrinkage by making a parallel set of concrete cylinders or prisms. For a large concrete dam project. temperature effects also have a cycling variation. polyethylene plastic bags.20] As submitted for ICOLD review. 4.21] and. when interpreting laboratory test results for creep in real dam behaviour [4. and t = time after loading. the shrinkage strain.1 Creep values for Brasilian mass concrete [4. It is common to plot creep test results on a semi-logarithmic graph in which the linear axis represents creep strain and the logarithmic axis represent time. In addition the cycling nature of the hydrostatic load.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 4 (Creep properties) subjected to constant loads have been enclosed in sheer copper casing with soldered joints. A device for in-situ measurement of creep and its back analysis by the LNEC (Portugal) is reported in [4.19] [4. it is a generally necessary to conduct creep tests on sealed specimens in the laboratory at enough ages of loading so that a complete knowledge of rheological behaviour is available. in the case of unsealed specimens.1.24]. This has led to the development of numerous logarithmic equations for creep. While it is not intended that a theoretical logarithmic law should be inferred from the equation. the slope of the least squares line is a convenient parameter for comparing the creep characteristics of different concretes. for Swiss Dams. 4.22] [4. 4.20] and briefly summarised in Tab. in days (duration of loading). calculated as the slope of a straight line representing the creep curve on the semi-log plot. In this way. march 2008 Section 4-7 . Creep and shrinkage are difficult to separate in measurements of deformation during testing of unsealed specimens because they are inter-related. The strain measurements obtained in creep tests must be processed to take account of the immediate elastic strain due to loading and. at least at near the upstream and downstrem dam faces. An example of a typical creep equation is shown below: ∈= 1 + F( K ) ln(t + 1) E where: ∈ = total strain per unit stress.3 [4.4 [4.

0 3.5 6.5 - 1/E -6 F(k) 6.7 5.5 8.2 6.4 6.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 4 (Creep properties) Age at time of loading (Days) Schist Quartzite Basalt Greywacke 1/E -6 F(k) 18.1 - 1/E -6 F(k) 9.3 13. 4.0 3.8 10*E /MPa 2 7 28 90 365 23.9 3.5 3.5 - 10*E /MPa 6.7 9.3 22.5] As submitted for ICOLD review.9 10*E /MPa 12.5 4.3 4.3 .8 5.1 6.13 5.2 4.7 - 10*E /MPa 6.Typical Creep surface [4.4 1/E -6 F(k) 9.1 3. march 2008 Section 4-8 .8 10.0 8.2 9.3 7.6 8.1 Fig.

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 4 (Creep properties)

Fig. 4.4 – Comparison of results of creep tests on two different types of dam concrete (Swiss dam and Shasta dam) [4.22] [4.23]

4.4 SIGNIFICANCE OF CREEP PROPERTIES ON DAM BEHAVIOUR Due to the special conditions of concrete in dams, as for example the high temperature in the body mass and the high sustained load (self-weight) at early ages, the cycling hydrostatic loading, the low cement content, with higher strain at equal stress, etc., creep in dams is usually larger than in other civil structures and needs a careful estimation. In particular, information on creep properties is of relevance to an understanding of the mechanism leading to the prediction of potential thermal cracking of mass concrete when undergoing thermal cycles due to the heat of hydration and ambient conditions. Therefore, the most extensive use of creep data has been in the thermal stress analysis for concrete dams. The effect of creep strain of a dam concrete, particularly due to the cycling loading conditions, together with the changes in the mechanical material properties, plays also As submitted for ICOLD review, march 2008 Section 4-9

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 4 (Creep properties) an important role in the ageing of concrete dams [4.22]. In fact, the creep effect does not stabilise rather quickly for thick structures like dams and creep is very close to the so called “basic creep” for which no evidence of complete stabilisation is obtained after several decades of loading [4.25]. For example creep tests carried out on 27 year old concrete samples of the Zervreila arch dam, in Switzerland, have shown that creep deformation can form an important contribution to the total deformation process for many years [4.26]. This is clearly shown in Fig. 4.5 where the radial displacement at the top of the crown cantilever of the arch dam was chosen as a measure for the correlation of the irreversible creep strain with age. In this case the effect of creep deformation has been assessed taking into account the creep recovery, which occurs when the structures is unloaded (seasonal variation of the loads) [4.22].

Fig. 4.5 – Normalised radial deformation of a Swiss arch dam, assumed as a measure of the irreversible creep strain over the years [4.22]

As submitted for ICOLD review, march 2008

Section 4-10

ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 4 (Creep properties) Information on creep properties is also of particular interest to evaluate the stress behaviour and the structural safety of a mass concrete dam when subjected to the expansion phenomena due to Alkali-Aggregate Reaction (AAR). Literature and field 1 data seem to indicate that AAR expansion tends to increase the rate in concrete [4.27] .

4.5 MODELLING CREEP IN STRUCTURAL ANALYSIS One method of expressing the effect of creep in structural finite elements analysis is the sustained modulus of elasticity in which the stress is divided by the total deformation (i.e., instantaneous deformation plus creep deformation) for the time under the load (see Fig. 4.1). Typical instantaneous and sustained moduli of elasticity of mass concrete are given in Tab. 4.2. It can be seen that the values for sustained moduli of elasticity are approximately one-half those for the instantaneous moduli when loads are applied at early ages and slightly higher percentage of the instantaneous moduli when loading age is 90 days or greater. The sustained modulus of elasticity has been used, for example, in simplified linear thermal stress analyses to account for the creep effects. It allows a quick approximate determination of thermal stresses analyses to account for the creep effects. It allows a quick approximate determination of thermal stresses resulting from heat of hydration of cement and ambient conditions. Tab. 4.2 - Typical instantaneous and sustained modulus of elasticity values (GPa) for mass concrete [4.28] Age at time of loading (Days) E 2 7 28 90 365 12 16 24 28 34 Basalt E’ 5.7 7.6 12 17 21 Andesite & Slate E 9.7 14 24 30 32 E’ 3.7 6.9 12 19 24 Sandstone E 19 29 31 36 39 E’ 10 13 18 22 25 Sandstone & Quartzite E 9.7 15 25 29 32 E’ 4.3 6.5 12 18 21

REMARKS (All concrete mass mixed, wet screened over 37.5 mm sieve): E= instantaneous modulus of elasticity at time of loading, GPa E’= sustained modulus of elasticity after 365 days under load, GPa

More information on the physical properties of concrete subject to expansion phenomena are reported in Appendix C.

1

As submitted for ICOLD review, march 2008

Section 4-11

03 41.30].032 7. the long term creep strain of concrete: in fact this last can play.33] [4. caused by an uniaxial constant stress acting since age “τ”. march 2008 Section 4-12 .34] [4.33] has also been extended to consider the cyclic creep caused by the cycling loading conditions of dams [4. 4.τ ) = (1 + Φ (τ −m + α )(T − τ ) n * ( K1 + K 2 (1 − e f )) E T −τ K1 is the fraction of elastic strain which is instantaneously recovered on unloading K1 + K2 = 1. particularly in presence of cycling loading conditions [4.058 0.510 17. m and α are materials parameters. by CEB (Comité Euro-international du Beton) [4. The principle of superposition was involved and the effect of unloading was simply added to that of loading.31] and by other Authors [4.35] can be reported. Examples of the values of such parameters for some dams are shown in Tab.3 – Values of the parameters of the creep compliance function for some dams [4.267 2.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 4 (Creep properties) The experimentally generated creep surface (Fig. The following correction function was then included in the creep compliance function.114 0.01 0. taking into account a Tf relaxation time constant. at age “T”.451 0. together with the changes in the mechanical material properties. Even more elaborated models have been proposed in the technical literature.56 43.3) can be also used to formulate the creep equation required in thermal and structural stresses in the mass concrete of dams.10 0. Tab.32] [4. 4. while Φ. Among the models.701 n 0. As submitted for ICOLD review. For example the so called “double power” creep model of Bazant [4.46 This model [4.3222 0.τ ) = where E and n are the Elasticity modulus and a power constant respectively.353 0.26 24. 1 (1 + Φ(τ −m + α )(T − τ ) n E J (T .099 α Φ 0. those proposed by ACI (American Concrete Institute) Committee 209 [4. in a general and comprehensive way.33] describes a creep compliance function J(T. an important role in the ageing of concrete dams [4. 1 T J (T .22].τ) which represent the total strain (elastic + creep) per unit stress.29].814 13.33] E (GPa Canyon Ferry Ross Dam Dworshak Dam Shasta Dam 85.52 m 0.3. able to predict. 4.22] [4.22].115 0.00 0.

Even lower ratio than those presented in Tab. 4.38]. it was possible to obtain a suitable correlation between the calculated cyclic total strain and the normalised chosen function (deformation of the crown cantilever crest of the dam).36] [4. is shown in the following Fig. at Zervreila arch dam [4. As submitted for ICOLD review. not consistent with the real stress behaviour at the dam site[4. 4. 4.6 .22].ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 4 (Creep properties) Based on this function. in order to evaluate the stress within the mass concrete structure.37].6. including creep effects. usually adopted. Redistribution effects caused by creep are to be taken into consideration in the evaluating criteria. Fig.Calculated normalised strain due to seasonal variation of loading. to be compared with the experimental data of the previous Fig.2 are reported in the literature. In the case of Mactaquac. in the structural analysis of the Zervreila arch dam [4. if the instantaneous modulus of elasticity is used in the mathematical models. again. for example. Alternative to models with more or less complicated creep laws is the simple approach. based on the use of a sustained elastic modulus lower than the instantaneous modulus of elasticity. very high stress states come up. This calculated normalised strain due to seasonal variation of loading. march 2008 Section 4-13 .5.22] In the case of expansion phenomena due to Alkali-Aggregate Reaction (AAR). including creep effects. 4. a sustained modulus of elasticity of about one third of the instantaneous modulus was used for the evaluation of the long term loading effects (concrete grows loads) [4.

146. Detroit.E... [4. [4.6] Ali I.. third edition. A. “Mechanisms of Creep in Concrete. Z. R. USA.. E. “Design of Concrete Structures” McGraw-Hill Book Company. and Hillsdorf..H.... p. and Best. ”Investigation of Creep in Concrete – Creep of Mass Concrete. Philadelphia. 1943. V.M. H. 1963. and Davis. Vol. p.12] Troxell. “Long-time Creep and Shrinkage Tests of Plain and Reinforced Concrete. Michigan.” Symposium on Mass Concrete. U. USA. [4. American Concrete Institute.35-57. Michigan..ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 4 (Creep properties) 4.11] Neville. ”Cement and Concrete Terminology”. [4. 1964. No.. “Studies of Creep in Mass Concrete. Kordina. Vol 58. ACI Manual of Concrete Practice. 1567-1579. 1964. [4. R.” Deutscher Ausschuss fur Srahlbeton.” Significance of Tests and Properties and Concrete and Concrete-Marking Materials. Army Engineer Waterways Experiment Station. E.3] Smerda. and Meyers. American Society for Testing and materials.” RILEM Bulletin (Paris). p. and Mattock. H.” Miscellaneous Paper No. NY. J. A. Vol 62...4] Neville. “Properties of Concrete”.” Proceedings.” American Concrete Institute Monograph No. 1971... Detroit. “Creep and Shrinkage of Concrete Elements and Structures”.E. G.1-68. As submitted for ICOLD review... Detroit. 1990. 1965. Vol 63. A. Michigan. 1979. 267-290. ASTM STP 169C. B. ACI Journal. C. American Concrete Institute. M. A. “Der Einfludd des mineralogischen Charakters der Zuschlage auf das Kriechen von Beton. ”Creep of Concrete: Influencing Factors and Prediction. 16. ACI SP-6. K. p.W. [4. New Series No. Michigan. [4. R. [4. 192-201. p.M. A. 6.C.5] Philleo..” Proceedings. 1988. And Kesler. 47-57. Detroit. G. 257-283. USA. [4.. 21. 1963.F. ACI 116R-90. Vicksburg. K. USA.9] McCoy. and Nilson. Part 1.E. H. 6-132. [4. ACI SP-9. New York.” Symposium on Creep of Concrete.15] Hansen. pp.M.. American Concrete Institute. Raphael.2] Winter. 134 pp. p.. D.. [4. 1983. American Concrete Institute.” Symposium on Creep of Concrete.. published by Wiley. Report 3. “Hardened Concrete: Physical and Mechanical Aspects. “Creep of Concrete with and without Ice in the System. American Concrete Institute. 260 pp. A. American Society for Testing and Materials. Bulletin No. 1966. 1-33. 1958. Pirtz. published by Elsevier.10] Polivka. p. 1994. “Creep of Concrete at Elevated Temperatures. PA.S.8] Rusch. ACI SP-9. R. [4. USA. and Adams. [4. ACI Journal. 22 pp. T.” Proceedings. 12. Mississippi.M. 1966: “The Influence of Size and Shape of Member on the Shrinkage and Creep of Concrete. Michigan. USA. p. E. Detroit.13] Nasser. Fristek. Developments in Civil Engineering. march 2008 Section 4-14 . 1962. “Elastic Properties and Creep. 1943.7] Neville. H. and Neville. C.. [4.1] American Concrete Institute.6 REFERENCES [4.. 1101-1120.14] Johansen. USA.

p. Neville. ”A Hypothesis on Mechanism of Creep of Concrete with Reference to Multi-axial Compression”.. [4.. and Ghali. [4. [4.21] Gunn. American Concrete Institute.17] McDonald. Bulletin d’information n° 199.. Q. [4. [4.A .22] Dungar. 1993. [4.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 4 (Creep properties) [4. J. C. [4. Conference on Alkali-Aggregate Reaction in Concrete. “ Concrete of Swiss Dams – Experiences and Syntesis”. Proc.56.29] Gunn. International Journal on Hydropower and Dams. ACI Manual of Concrete Practice. Balkema 1990.27] Blight. “Evaluation of the time dependent behaviour of concrete”.F.20] Guerra. “Cooling and Insulating Systems for Mass Concrete. Technical report C-75-4. “The effect of cyclic creep on the ageing of arch dams”. Q. J. July 1953.M. 1. “Arch Dams”.FURNAS CENTRAIS ELETRICAS SA. ICOLD. 1993. Hydopower & Dams. 73-79.M. Cape Town. Bussoney. Congress.32] Brooks..M. Issue six. V. “Observed and computer-simulated movements in the 157 m high Zervreila Arch Dam (2-creep and plasticity of mass concrete)”. R.. A. et al. R. R. A. [4. USA.J. R. As submitted for ICOLD review.. “Creep analysis of mass concrete dams”. Detroit.18] Ramos. Congress. R.65..4R-93.9. 1996. [4.Fontoura. R. Rimmer.E.. Proceedings.S.16] Gopalakrishnan. 1991. ACI Journal. 1996.. “Predicting long-term creep and shrinkage from shortterm tests”. Vienna.. 1975. 1990. Bittencourt R.. Denver.S. and Pacelli.31] CEB.65. E. C.. Soares de Pinho J. [4. [4. 30. Congress. 462-472.Part 3.M.A . march 2008 Section 4-15 . pp. Bossoney. ACI Manual of concrete practice . Proc. “Ageing of Dams – Swiss Experience”. No. Vicksburg. edited for the ICOLD Congress in Beijing. Shrinkage.19] Raphael. Schuttle. J.. “Creep analysis of Mass Concrete Dams”. 29-35. Magazine of Concrete Research.. U. “A 10 year study of creep properties of concrete”. Q.28] American Concrete Institute. and temperature effects in concrete structures”. J.. 17. 1981. 306 pp. “The effects of AlcaliAggregate Reaction of Reinforced Concrete Structures made with Witwatersrand Quartzite Aggregate”. ACI 209R-92.T. “ The Nature of Mass Concrete in Dams”.24] Swiss Committee on Dams. Issue six.10.. ”Time-Dependent Deformation of Concrete Under Multiaxial Stress Conditions”. R. Vienna. Mciver. American Concrete Institute. A.. J. 73-79. 1991. “Delayed effects observed in concrete dams”... Lausanne.. Vol 67.23] USBR Concrete Laboratory Report n° SP-38. “Mass Concrete Properties for Large Hydropower Constructions”. “Prediction of Creep.38. W. [4. Army Engineer Waterways Experiment Station. Neville.” ACI 207. 17. 1970.25] Sinninger. n°103. [4. J.. [4.K. South Africa.. K.. G.M. 1976. Mississippi. W.E. SP 55. [4.. Proc. ICOLD 15.26] Dungar R.M. ICOLD. M. R.R. June 1978.30] ACI. 1985. [4. Part 1. 2000. Michigan.

. Vicksburg.I. 1994. r. final report.35] Gardner. D. R. “Creep and shrinkage prediction model for analyses and design of concrete structures – Model B3”. Mississippi. As submitted for ICOLD review. [4. Z. and Hammons.J.B..P. 68.G.98. 28. [4.P. 24.. V. 1991. Army Engineer Waterways Experiment Station.A. S. N. “Mactaquac Generating Station Intake and Spillway Remedial Measures”. 2004. march 2008 Section 4-16 .34] Bazant.37] ICOLD European Club. G... vol. “Design provisions for drying shrinkage and creep of normal-strength concrete”. ACI Materials Journal. U. Charlwood.. Mech. Curtis. Technical Report SL 91-7. Q. R. 1985. M. Durban..33] Bazant. 111. Working Group on Ageing on Concrete Dams.J. Materials and Structrures. [4. n°1.. “ Ageing on Concrete Dams”.ICOLD Bulletin: The Physical Properties of Hardened Conventional Concrete in Dams Section 4 (Creep properties) [4.38] Thompson. [4. Proceedings 18th ICOLD Congress.36] Garner. Vol. March-April 2001. 1995. pp159167.D. “Triple Power law for Concrete Creep” ASCE Journal of Eng. Steele. [4. Z. “Development and Implementation of a TimeDependent Material Model for Concrete”.S. M. Lockman.R.

Sign up to vote on this title
UsefulNot useful