Recommendations for preventing disorders due to Delayed ...media.lcpc.fr/ext/pdf/sources/gtrsi-e.pdf · ISSN 1151-1516 Réf : GTRSI/E techniques et méthodes des laboratoires des

ISSN 1151-1516

Réf : GTRSI/E

techniques et méthodesdes laboratoires des ponts et chaussées

Guide technique

Recommendationsfor preventing disorders due

to Delayed Ettringite Formation

The recommendations presented in this document aim at limiting the risk of disorders occurring dueto an internal sulfatic reaction. The latter is caused by the formation of delayed ettringite in acementitious material and occurs in particular because of an important heating of the concreteintervened several hours or several days after its casting. It causes an expansion of concrete whichgenerates in its turn a cracking of the structures. This reaction can be encountered with two types ofconcrete: the heat treated concretes and the concretes cast in place in elements known as critical. These recommendations are concerned with civil engineering structures and buildings comprisingelements of important size that are in contact with water or subjected to a humid environment. Theyfix the level of prevention to be reached according to the category of the structure (or to the part ofstructure) and to the exposure conditions. For each of the four levels of prevention selected,associated precautions are applied and associated checks are carried out. They also presentprovisions related to the design and dimensioning of the structures, the formulation and themanufacture of the concrete as well as to its pouring.

Rec

omm

enda

tions

for p

reve

ntin

g di

sord

ers

due

to D

elay

ed E

ttrin

gite

For

mat

ion

Guide technique

May 2009

English Translation of a Technical Guidance published by the LCPC in August 2007 and entitled :

Recommandations pour la prévention des désordres dus à la réaction sulfatique interne

Recommendations for preventing disorders

due to Delayed Ettringite Formation

Laboratoire Central des Ponts et Chaussées58, bd Lefebvre, F 75732 Paris Cedex 15

This document has been written by a group of experts placed under the supervision ofa technical committee :

Members of the technical committee :

President of the committee : M. Thierry Kretz (SETRA).

MM. Abdelkim Ammouche (LERM), Claude Bois (CGPC), François Cussigh (FNTP), Michel Delort (ATILH), LoïcDivet (LCPC), Gilles Escadeillas (LMDC Toulouse), Hervé Foucard (Ville de Paris), Bruno Godart (LCPC), BrunoHuvelin (SNBPE), Yann Jaffré (SETRA), Alain Jeanpierre (EDF), Jean-Marc Morin (SNCF), Christian Parthenay(SNBPE), Alexandre Pavoine (LCPC), Patrick Rougeau (CERIB).

Members of the group of experts :

Moderators of the group : MM. Bruno Godart and Loïc Divet (LCPC).

MM. Jean-Luc Clément (LCPC), François Cussigh (FNTP), Gilles Escadeillas (LMDC Toulouse), Patrick Guiraud(CIMBETON), Lofti Hasni (CEBTP), Sylvie Lecrux (CTG ITALCEMENTI), Lionel Linger (Vinci Construction GrandsProjets), Alexandre Pavoine (LCPC), Patrick Rougeau (CERIB), Angélique Vichot (ATILH), Vincent Waller (CTGITALCEMENTI).

Ce document est propriété du Laboratoire central des ponts et chaussées et ne peut être reproduit, même partiellement, sans l’autorisation de son Directeur général

(ou de ses représentants autorisés)

© 2009 - LCPCISSN 1151-1516

N° DOI/Crossref 10.3829/gt-gtrsi/E-fr

3

Introduction......................................................................................................... 5

Chapter 1. General remarks on the DelayedEttringite Formation (DEF) ............................................................ 71. The DEF phenomenon.................................................................................................... 92. The main parameters associated with the delayed ettringite formation ........................ 103. The current standardized environment.......................................................................... 104. Influence of cement additions........................................................................................ 11

Chapter 2. Determination of the required level of prevention.................................................................................................. 131. Categories of the given structure or structural part ....................................................... 152. Exposure classes relative to DEF ................................................................................. 163. Levels of prevention ...................................................................................................... 16

Chapter 3. Precautions adopted based on level of prevention.......................................................................... 191. Level of prevention : As................................................................................................. 222. Level of prevention : Bs................................................................................................ 223. Level of prevention : Cs ................................................................................................ 224. Level of prevention : Ds ................................................................................................ 23

Sommaire

4

Chapter 4. Provisions related to design andcalculation of structures, concrete mix designand fabrication, as well as placement ............................ 251. General remarks ........................................................................................................... 272. Provisions related to the design and calculation of structures ...................................... 27

2.1. Avoiding extended contact with water .............................................................. 272.2. Limiting the temperature rise in concrete......................................................... 28

3. Provisions related to concrete mix design..................................................................... 284. Provisions related to concrete fabrication and transport ............................................... 295. Provisions related to concrete placement ..................................................................... 30

5.1. Jobsite organisation ......................................................................................... 305.2. Concrete cooling .............................................................................................. 305.3. Choice of formwork .......................................................................................... 31

6. Provisions specific to prefabrication.............................................................................. 31

Appendices ..................................................................................................... 35Appendix I : Guidelines on the delayed ettringite formation (DEF) ................................ 37

Appendix II : Evaluation of structures affected by DEF and presentation of disorders ... 39

Appendix III : Review of the exothermic reaction of concretes ....................................... 43

Appendix IV : Estimation of temperatures reached in structures planned for construction.......................................................................................... 47

Appendix V : Performance-based testing ....................................................................... 55

References...................................................................................................... 57

Bibliography ................................................................................................... 59

5

RECOMMENDATIONS FOR PREVENTING DISORDERS DUE TO DELAYED ETTRINGITE FORMATION

Introduction

T he Delayed Ettringite Formation (DEF) is a cause of disorders capable of damagingconcrete structures quite severely. While only a few cases have been catalogued inFrance, they are still significant enough to justify taking certain precautions for new

construction. The recommendations presented in this document are intended to minimisethe risk of inducing disorders related to this reaction.

These recommendations pertain to engineering structures and buildings containing large-sized structural elements that are in contact with water or exposed to a humid environment.They serve to target the desired level of prevention on the basis of both the category ofstructure (or part of structure) and the environmental conditions affecting the structure.Each of these levels of prevention is associated with precautions to be applied as well asverifications to be performed. These recommendations also set forth a number of buildingprocedures to be implemented.

The first cases of delayed ettringite formation appeared outside of France beginning in1987 in certain precast components that had been subjected to a heat treatmentinappropriate to the type of concrete mix design and exposure (e.g. railroad ties). Both themix designs and heat treatment cycles were thus adapted so as to avoid the occurrence ofnew disorders ; since that time, such adaptations have been formalised within a standardrelative to prefabricated products (NF EN 13369). Disorders caused by the delayedettringite formation were first observed in France in 1997 on bridges whose concrete hadbeen cast in place; more massive structural elements were primarily involved in suchobservations (e.g. piers, crossbeams on piers or abutments, tower bases), either in contactwith water or exposed to a high moisture.

According to our current state of knowledge, only a few research results have actuallybeen published on the topic of treating structures affected by the delayed ettringiteformation and moreover no treatment solution has been proposed to halt the progressionof this reaction within structures. Given this situation, a preventive approach provesessential in seeking to avoid the occurrence and development of this reaction in structuresto be build.

These recommendations serve to supplement the conventional state-of-the-art guidelinesthat address both concrete material quality and durability within structures ; morespecifically, these guidelines are complementary to Standard NF EN 206-1, precast

6

concrete product standards and the entire series of NF EN 1992 standards. They offeradapted solutions in the event of contradictions raised with some of the currently-applicable rules and regulations ; such is especially the case with recommendationsrelative to hardened concrete frost resistance prescribing the use of high contents of typeCEM I cement in order to resist spalling, this condition being able to be detrimental asregards durability in the presence of delayed ettringite formation. In addition, this set ofrecommendations reinforces some of the basic notions that, on occasion, tend to beoverlooked or even forgotten in the building process :

for structures cast in place :

• Current trends favouring increasingly compressed construction schedules, theattainment of high concrete strength at the early age in order to accelerate formworkrotation or the early introduction of prestressing forces, particularly in the case ofmassive elements, must not come at the expense of the durability of structures builtunder these conditions (see Appendix 3) ;

• The principle must be respected to allow selecting a concrete mix design welladapted to the structural part within its specific environment. This principle extends toadopting a multicriteria optimisation approach when choosing both the cement andthe concrete mix, in avoiding for example the use of highly-exothermic CEM Icements at a heavy content within a massive structural element ;

• The casting of massive elements must be avoided during the hottest periods of thesummer season when no specific conditions have been adopted to limit excessiveconcrete heating.

for structures composed of precast concrete elements :

• For these elements, it would be necessary to avoid all heat treatment cycles thatinvolve excessive maximum temperatures in conjunction with the excessive constanttemperature plateau for concrete mix designs that prove sensitive to DEF.

The set of recommendations contained herein combine a performance-based approachwith specification of means, this latter being mandatory when performance-orientedspecifications cannot be applied.

Lastly, these recommendations are provisional, reflective of the current state of knowledge.In the case where their application leads to especially restrictive provisions (e.g. level ofprevention Ds), it would be highly advised to solicit the services of an expert or requestinput from a specialised laboratory with experience and practice in studying this type ofreaction.

7

Chapter 1General remarks

on the Delayed Ettringite Formation (DEF)

1. The DEF phenomenonThis internal sulphate reaction may be characterised by the delayed formation of ettringite within acementitious material several months or even years after the cement has set and without thecontribution of any external sulphate. The term delayed (see Appendix 1) used in this contextindicates that the ettringite was not able to form during cement hydration, due to a significantheating of the concrete taking place a few hours or days after casting.

A distinction must be drawn between the delayed ettringite formation (DEF) phenomenon and theexternal sulphate reaction (ESR) phenomenon, which was first discovered in 1887 and detected byCandlot subsequent to observations recorded on mortars along Paris's fortifications when incontact with gypsum water. In the case of ESR, sulphate sources may be added by soils and de-icing salts or be conveyed by groundwater, infiltration water, seawater or water effluent fromindustrial sites. Sulphates penetrate via the concrete capillary network and can cause so-called"secondary" ettringite formation (see Appendix 1), which in turn could induce expansionphenomena and hence concrete deterioration. As regards ESR therefore, deterioration startsgradually from the surface and heads towards the core of the concrete element.

In some instances, when concrete undergoes heating at an early age, the delayed ettringiteformation (DEF) phenomenon may appear without necessarily any sulphate input from an externalsource. The international term ascribed to this condition is Delayed Ettringite Formation (or DEF forshort). Such a phenomenon can arise for two types of concrete : heat-treated concretes, and cast-in-place concretes in critical structural elements, i.e. concrete elements whose heat release isonly marginally discharged to the exterior, which leads to a sizeable rise in concretetemperature (see Appendix 3).

A large majority of internally-generated sulphate ions stem from the cement and, under certainconditions, can be dissolved in the concrete pore solution. The sulphate reaction involves theseions present in the pore solution along with the cement aluminates, potentially leading to theformation of ettringite capable of causing expansion inside the hardened concrete. The internalsulphate swelling phenomenon can be detected by the occurrence at the concrete surface ofmultidirectional cracking with a relatively wide mesh spacing of between 10 and 30 cm.

No specific symptoms indicate this delayed ettringite formation since other structural pathologiesalso display the same symptoms, particularly the alkali-aggregate reaction phenomenon.

Up until now, it has been observed that concrete elements affected by this pathology :• are critical elements, i.e. concrete elements whose heat release is only marginally discharged tothe outside, leading to a significant concrete temperature rise (Appendix 3) ;

9


Chapter 1 - General remarks on the Delayed Ettringite Formation (DEF)

10

• were cast during summertime and exposed to a temperature estimated at the core to be above80 °C during the concrete hardening period ;• were subjected to a moist environment for a number of years.

These elements might also be precast elements that had undergone a thermal treatment at veryhigh heat.

2. The main parameters associated with the delayedettringite formationThe causes, physicochemical mechanisms and kinetics of the reaction that gives rise to the internalsulphate swelling phenomenon, as well as the impact of the various parameters affecting thesulphate reaction, are not yet accurately known and continue to be the topic of widespreadresearch. It appears however that a combination of several parameters is essential to initiating andextending the DEF, and this observation would most likely explain the low number of structurescurrently identified as experiencing DEF. The principal parameters herein are water, temperatureand its duration of application, the sulphate and aluminate contents of cement, and lastly the alkalicontent of concrete :

• Water and humidity : One longstanding observation, both in the laboratory and on actualstructures, involves the fundamental role played by water as the reaction develops. Water is areactive medium essential to producing the reaction ; it is as much involved in the transfer processas in the actual formation of reaction products. DEF primarily affects the parts of structures either incontact with water (submerged zone, tidal zone) or subjected to water inflow (waterproofingdefects, absence of drainage, etc.), or perhaps exposed to a high moisture rate.

• Temperature and its duration of application : The maximum temperature reached as well asthe amount of time high temperature is maintained both influence the risk of delayed ettringiteformation. Laboratory work has shown that if temperature were to exceed 65 °C and if other keyparameters were present, a situation would be encountered in which DEF is typically able todevelop. Intense heating of the concrete during setting and hardening is a necessary precondition,yet on its own remains insufficient.

• Sulphate and aluminate contents of the cement : Sulphates and aluminates are directlyinvolved in the reactive mechanism that serves to form ettringite, which is composed of a hydratedcalcium trisulfoaluminate. Consequently, DEF can only arise if the cement used contains a highenough quantity of both tricalcic aluminates (3CaO Al2O3 or C3A) and sulphates (SO3).

• Alkali content of the concrete : The role of this content on ettringite solubility is welldocumented. Ettringite is more highly soluble at higher alkali rates. As a result of ettringite solubilityvariation with temperature, a strong interaction exists between these two parameters during theDEF process. With all other parameters being the same, a drop in alkali content serves to increasethe critical temperature value.

3. The current standardized environmentStandard NF EN 206-1 Concrete - Part 1 : Specifications, performance, production and conformityestablishes for concrete a set of compositional prescriptions, which depend on the level ofexposure the structure or part there of will sustain during its use period. These prescriptions areintended to guarantee structural durability with respect to certain physicochemical aggression, inparticular with respect to the external sulphate reaction. Yet this standard does not provide any

11


response in terms of DEF prevention, and the classes of exposure it defines are not at all adaptedto the inclusion of this risk.

On the other hand, draft Standard prEN 13670-1, entitled Construction of concrete structures,offers several recommendations specific to concrete heating during the setting period. It actuallyindicates that the concrete temperature peak within an element must not exceed 70 °C, unlessdata prove that with the particular combination of materials used, greater temperatures would exertno significant adverse effect on concrete service performance.

In the area of prefabrication, Standard NF EN 13369 Common rules for precast productsacknowledges DEF and provides a set of environment-based recommendations (i.e. dry vs. wet)that focus on :• either the maximum temperature of heat treatment,• or a concrete durability experiment,• or threshold contents of both cement sulphate and concrete alkali.

The delayed ettringite formation is also taken into account in a number of standards specific toprefabricated products such as NF EN 13230-1 : Railway applications - Track - Concrete sleepersand bearers.

4. Influence of cement and additionsIn the case of cast-in-place concretes for the fabrication of critical structural elements, it ispreferable to use ordinary cements with low hydration heat, as set forth in Standard NF EN 197-1,Amendment A1 (LH rating). Another possibility would be to use cements whose hydration heat, asdetermined according to Standard NF EN 196-9 and verified at a frequency identical to that appliedfor LH cements, does not surpass a threshold value to be established in the specifications. Thislatter point pertains in particular to the family of CEM I cements, which in the vast majority of casesare unable to satisfy the characteristic value of 270 J/g at 41 hours, as stipulated in StandardNF EN 197-1/A1.

Possibilities also exist to use mineral additions as a substitution for the CEM I type of cement, inorder to decrease the concrete exothermic reaction, or to use CEM II composite Portland cements,CEM III blast furnace slag cements, CEM IV pozzolanic cements, CEM V composite cements, orCSS super-sulphated cements (see Standard NF P15-313). Moreover, when introducing blastfurnace slag or fly ash, concrete strength with respect to DEF is expected to improve, especiallyowing to both a relative decrease in the quantity of aluminates stemming from clinker andmodifications to hydrate type and texture. The use of additions as a substitute for cement alsohelps reduce the quantity of sulphates in the concrete.

13

Chapter 2Determination of the required level

of prevention

The purpose of these recommendations is to provide precautionary measures for the placementand design of a concrete, so as to better contain the risks capable of arising due to delayedettringite formation (DEF) throughout the lifetime of the structure.

The approach adopted consists of identifying those parts of structures most at risk of developingdisorders due to the DEF reaction ; such parts tend to be considered as critical elements(according to the interpretation offered in Appendix 3), as well as precast concrete productssubjected to heat treatment. To proceed with this approach, cross-referencing is performedbetween the category describing the structure (or part thereof), as differentiated by the acceptablelevel of risk, and environmental actions affecting the structure (or part thereof) over the life cycle ofthe structure.

This cross-referencing step serves to establish for these structural parts a level of prevention that inturn determines the set of precautionary measures to be implemented. Such measures depend to alarge extent on the threshold of the maximum temperature reached at the core of the targetedstructural elements during concrete hardening and on the choice of a satisfactory concrete mixdesign.

1. Category of the given structure or structural partThe structures (or parts of structures) are grouped into 3 categories representing the acceptablelevel of risk relative to the delayed ettringite formation for a given structure (or part). The choice ofstructural category falls under the responsibility of the project owner ; this decision depends on :the type of structure, its intended purpose, the consequences of disorders on the desired level ofsafety, and lastly on its future maintenance.

15


Catégory Examples of structures or structural elements

Categorie I Structures made of concrete rated in a strength class of less than C 16/20(small or acceptable Non load-bearing building componentsconsequences) Easily-replaced elements

Temporary structuresThe majority precast non-structural products

Categorie III Buildings housing nuclear power plant reactors and cooling towers(unacceptable or nearly Damsunacceptable consequences) Tunnels

Exceptional bridges and viaductsMonuments and landmark buildingsRailroad ties

Categorie II The load-bearing components of most buildings and engineering structures(rather severe consequences) (including standard bridges)

The majority of prefabricated structural products (including pressurised pipes)

TABLE I - EXAMPLES OF STRUCTURES OR STRUCTURAL ELEMENTS CLASSIFIED BY CATEGORY

Category I applies to structures (or parts) for which the consequences of a potentialdisorder would remain small or acceptable. The majority of precast concrete productsfall into this category, with the notable exception of prefabricated structural elementsand products intended for use in more aggressive environments (acoustic screens,bridge cornices, some types of drainage pipes, etc.).

Category II comprises structures (or parts) for which the consequences of a potential disorderwould be rather severe. The load-bearing elements of most buildings and engineering structures(including standard bridges) are placed in this category, as are prefabricated structuralcomponents.

Category III represents structures (or parts) for which the consequences of a potential disorderwould be totally or near totally unacceptable. Typically included herein are exceptional structuresthat may necessitate a complete absence of disorders for safety or aesthetic reasons, or as a resultof the infeasibility of their repair or replacement.

2. Exposure classes relative to DEFSince Standard NF EN-206-1 does not define a class of exposure specifically adapted to thedelayed ettringite formation, three complementary classes XH1, XH2 and XH3 will be introducedinto the present document to address this case. These additional classes acknowledge the fact thatthe presence of water or high ambient relative humidity constitutes a factor essential to developingthe delayed ettringite formation. The inflows of alkalis and sulphates from the surroundingenvironment also serve to exacerbate disorders, yet they get considered as part of a surfacedegradation process and are governed by preventive procedures addressed elsewhere (e.g. inStandard NF EN 206-1).

These three exposure classes XH1, XH2 and XH3 are defined in accordance with the indicationslisted in Table 2, which also displays for information purposes a number of examples of structuralparts classified within their appropriate ambient environments.

Exposure classes XH1 through XH3 serve to supplement the 18 classes already established inStandard NF EN 206-1 and must be indicated in the Specific Technical Clauses (of the Contract)for each structural part. No apparent direct correlation can possibly exist between the 3 classesXH1-XH3 and the 18 exposure classes set forth in Standard NF EN-206-1 ; it can still beconsidered however that structural parts classified in XC4 fall into class XH2.

For a given structure (or part), the specifications listed in the present document must beincorporated as a complement to those imposed by Standard NF EN 206-1.

3. Levels of preventionFour levels of prevention have been established, as designated respectively by the letter codes As,Bs, Cs and Ds. A level of prevention is determined on the basis of both the structure category andXH exposure class applicable to the particular part of the structure. The level of prevention can beidentified by considering the entire structure, yet it is advised to assess each structural partindividually in order to derive a specifically adapted level of prevention. Responsibility for selectingthese various levels of prevention lies with the project owner, for whom Table 3 provides someguidance.

For purposes of illustration, let's take a bridge rated in category II : both the piles and foundationfootings qualify for a level of prevention Cs, whereas the piers and deck require a Bs level ; as forthe crossbeams on piers and abutments, the choice of level of prevention will be made by

Chapter 2 - Determination of the required level or prevention

16

17


Exposure Environmental Informational examples to illustrateclass description the choice of exposure class

designation

XH1 Dry or moderate humidity Part of a concrete structure localed insidebuildings where the ambient air humidity rateis either low or mediumPart of a concrete structure located outsideand protected from rain

XH2 Alternation of humidity and Part of a concrete structure located insidedryness, high rate of humidity buildings where the ambient air humidity rate

is highPart of a concrete structure unprotected by acoating and exposed to inclement weather,without any water stagnation at the surfacePart of a concrete structure unprotected by a coating and frequantly subjected to condensation

XH3 In long-lasting contact with water : Part of a concrete structure permanentlystate of permanent immersion, submerged in waterwater stagnation at the surface, Maritime structural elementstidal zone A large number of foundations

Part of a concrete structure regularly exposedto water projections

TABLE II - EXPOSURE CLASSES FOR A STRUCTURAL PART WITH RESPECT TO DEF

Exposure class ofthe structural part XH1 XH2 XH3

Structure category

Category I As As As

Category II As Bs Cs

Category III As Cs Ds

TABLEAU III - CHOICE OF LEVEL OF PREVENTION

evaluating the conditions adopted to ensure water discharge on these structural parts : the level ofprevention will thus be set at Bs or Cs depending on the risk of water stagnation.

19

Chapter 3Precautions adopted based

on level of prevention

Each of the four levels of prevention As, Bs, Cs and Ds corresponds with a type of precaution tobe implemented. The prevention principle basically relies on limiting concrete heating, ascharacterised by the maximum temperature Tmax capable of being reached inside the structureand, if applicable, by the length of time a high temperature can be maintained.

To enable estimating the maximum temperature capable of being reached at the element core,Appendix 4 presents a simple method for this estimation by only utilising a minimal number of basicdata, such as element thickness (over its smallest dimension) and a few data regarding concretecomposition (contents, 2-day and 28-day compressive strength of the cement, hydration heat of thecement at 41 hours, etc.). A more accurate method for estimating maximum temperature consistsof running a finite element computation code and including the heat released by concrete during aspecific test.

In order to avoid any excessive and uncontrolled concrete heat release, it would be necessary toimplement all possible resources (choice of mix design and concrete components, choice ofconcreting duration, cooling of fresh concrete, adapted construction procedures, etc.) in order tolower concrete temperature during placement and then over the first few days thereafter.

In all cases, a concrete mix design must meet the specifications contained in the standards andrecommendations currently in effect. However, it may be extremely difficult, or even impossible, toreconcile the requirements of means imposed due to durability considerations with the present setof recommendations, especially as regards the type of binder and its minimum content. As anexample, the « G + S » specification from LCPC's « frost » recommendations or the XA3 class fromStandard NF EN 206-1 stipulates a minimum concentration of 385 kg/m3 (for a value of Dmax equalto 20 mm) and significantly limits the possibility of incorporating additions or using compositecements, which could lead to considerable heating in the case of massive structural elements.Since the overlap of specifications of means does not always prove relevant, it is necessary toconduct a special study at the project design stage in order to lay out realistic requirements to beprescribed in the specifications. Such a study might lead to: adopting a performance-basedapproach geared to the targeted exposure class and based on a series of recognised tests,modifying building procedures, or revising the project altogether.

As for considerations of frost and de-icing salts, these performance-oriented tests consist of P18-424 or P18-425 (for « severe » or « moderate » frost, respectively) and/or the spalling test inP18-420. The reader's attention is drawn to the time periods necessary to produce suchjustifications (at least 3 months).

21


Chapter 3 - Précautions adopted based on level of prevention

22

1. Level of prevention : As For this level of prevention, the risk relative to the delayed ettringite formation must be taken intoaccount by means of the following precaution :

The temperature Tmax capable of being reached within the structure must remain less than 85°C.

In the case of controlled heat treatment*: it is authorised to exceed temperature Tmax = 85 °C by arise of up to 90 °C, provided that the duration during which the temperature remains above 85 °C islimited to 4 hours.

2. Level of prevention : Bs For this level of prevention, the risk with respect to the delayed ettringite formation must be takeninto account by adopting one of the two following precautions, denoted 1) and 2), respectively :

1) The maximum temperature reached within the concrete must remain below 75 °C.

2) If the maximum temperature obtained within the concrete is unable to remain below 75°C,then it must never exceed 85 °C and at least one of the six following conditions must be satisfied :

• The heat treatment is controlled*, the duration of concrete temperature staying above 75 °Cmust not exceed 4 hours, and the equivalent active alkali content of the concrete must remainless than 3 kg/m3. (The term «temperature holding duration» is defined as the period duringwhich the temperature surpasses 75 °C) ;• Use of a cement compliant with Standard NF P15-319 (ES) with, in the case of use of CEM Iand CEM II/A, a limitation set at 3 kg/m3 of the equivalent active alkali content of theconcrete ; • Use of cements that do not comply with Standard NF P 15-319 (ES) that are of the typeCEM II/B-V, CEM II/B-S, CEM II/B-Q, CEM II/B-M (S-V), CEM III/A or CEM V ; moreover, theSO3 content of these cements must not exceed 3 %, and the clinker introduced into cementfabrication must not contain more than 8 % of C3A ;• In combination with CEM I, use of fly ash compliant with Standard NF EN 450-1, groundblast-furnace slag compliant with Standard NF EN 15167-1, or even calcinated naturalpozzolana (the corresponding French standard is currently under preparation). The proportionof additions must attain at least 20 %, provided standard requirements have been met (thisapplies in particular to Standard NF EN 206-1). The CEM I employed must meet the followingrequirements : C3A (as a ratio of cement) ≤ 8 % and SO3 ≤ 3 % ; • Verification of concrete durability with respect to DEF by relying upon performance testingand by satisfying a number of decision-making criteria ;• For prefabricated elements, the intended concrete/heating pair is identical or analogous to aconcrete/heating pair that can be associated with at least 5 satisfactory references of usecovering different conditions of application**. This analogy needs to be justified byappropriate documentation and then approved by an independent laboratory with expertcredentials in DEF.

3. Level of prevention : Cs For this level of prevention, the risk with respect to the delayed ettringite formation must be takeninto account by adopting one of the two following precautions, denoted 1) and 2), respectively :

1) The maximum temperature reached within the concrete must remain below 70 °C.

2) If the maximum temperature obtained within the concrete is unable to remain below 70 °C,then it must never exceed 80 °C and at least one of the six following conditions must be satisfied :

• The heat treatment is controlled*, the duration of concrete temperature staying above 70°Cmust not exceed 4 hours, and the equivalent active alkali content of the concrete must remainless than 3 kg/m3. (The term « temperature holding duration » is defined as the period duringwhich the temperature surpasses 70°C) ;• Use of a cement compliant with Standard NF P15-319 (ES) with, in the case of use of CEM Iand CEM II/A, a limitation set at 3 kg/m3 of the equivalent active alkali content of theconcrete ; • Use of cements that do not comply with Standard NF P 15-319 (ES) that are of the typeCEM II/B-V, CEM II/B-S, CEM II/B-Q, CEM II/B-M (S-V), CEM III/A or CEM V ; moreover, theSO3 content of these cements must not exceed 3 %, and the clinker introduced into cementfabrication must not contain more than 8 % of C3A.• In combination with CEM I, use of fly ash compliant with Standard NF EN 450-1, groundblast-furnace slag compliant with Standard NF EN 15167-1, or even calcinated naturalpozzolana (the corresponding French standard is currently under preparation). The proportionof additions must attain at least 20 %, provided standard requirements have been met (thisapplies in particular to Standard NF EN 206-1). The CEM I employed must meet the followingrequirements : C3A (as a ratio of cement) ≤ 8 % and SO3 ≤ 3 % ;• Verification of concrete durability with respect to DEF by relying upon performance testingand by satisfying a number of decision-making criteria ;• For prefabricated elements, the intended concrete/heating pair is identical or analogous to aconcrete/heating pair that can be associated with at least 5 satisfactory references of usecovering different conditions of application**. This analogy needs to be justified byappropriate documentation and then approved by an independent laboratory with expertcredentials in DEF.

4. Level of prevention : DsFor this level of prevention, the risk with respect to the delayed ettringite formation must be takeninto account by adopting one of the two following precautions, denoted 1) and 2), with the firstbeing recommended as the priority precaution:

1) The maximum temperature reached within the concrete must remain below 65°C.2) If the maximum temperature obtained within the concrete is unable to remain below 65°C,

then it must never exceed 75°C and the two following conditions must be satisfied :• Use of a cement compliant with Standard NF P15-319 (ES) with, in the case of use of CEM Iand CEM II/A, a limitation set at 3 kg/m3 of the equivalent active alkali content of theconcrete ; • Validation of the concrete mix design by an independent laboratory with expert credentials inDEF.

23


* The controlled heat treatment can be performed either in a prefabrication plant or on adequate project site facilities.** A satisfactory reference of use corresponds to the use of a concrete/heating pair for building a structure exposed toconditions that promote DEF development (i.e. environment XH2 or XH3) over a significant period of time (at least 10years), during which absolutely no DEF-related disorder appears.Two concrete/heating pairs are considered to be analogous whenever the concrete mix designs closely resemble oneanother and especially when the conditions listed below have been met :

• heating of the project mix design does not surpass heating of the reference design ;• C3A and SO3 contents in the project design cement do not exceed the C3A and SO3 contents of the referencecement ;• the alkali contents of the two concretes do not differ by more than 10 % ;• aggregates used in the two concretes are derived from the same origin ;• the component mix contents do not differ by more than 10 %.

25

Chapter 4Provisions related to design and calculation of structures,

concrete mix design and fabrication,

as well as placement

1. General remarks The detailed conditions set forth in the present chapter are intended to :

• avoid extended contact of the critical part with water throughout the period of use of the structure,• limit the maximum temperature reached inside the concrete contained in critical parts,• control the heat treatments applied to prefabricated elements.

2. Provisions related to the design and calculation of structures

2.1. Avoiding extended contact with waterThe structure must be designed so as to avoid, to the greatest extent possible, water accumulationand stagnation zones as well as preferential flow paths due to runoff ; this effort necessitatespredicting the slope profiles and shapes that allow for rapid water drainage.

It is also possible to implement the protocol required to avoid water and humidity penetration withinconcrete structures either :• by ensuring waterproofing of the critical part,• or by ensuring waterproofing of the structural components that serve to protect the critical partand by implementing the appropriate drainage measures.Such is particularly the case with bridge decks where a waterproofing membrane* is required inaddition to planning for water drainage systems that are efficient and maintained on a regular basis.The application of a well-adapted waterproofing system (in the form of a membrane) could make itpossible to classify the structure or structural part in the XH1 category, yet keep in mind that thedurability of this waterproofing system imposes a regular replacement schedule.

27


* For the horizontal parts of road traffic-bearing bridge decks, the reference document is Fascicle 67 of the GeneralTechnical Clauses, Part I, to be complemented by the Technical Guideline issued by the SETRA Road EngineeringAgency. Concerning the technical procedures to be implemented, the STER 81 Guide, also published by SETRA, and itstwo most recent updates would constitute the applicable documentation.

Chapter 4 - Provisions related to design and calculation of structures, concrete mix design and fabrication, as well as placement

28

Among the other surfacing/coating materials capable of limiting humidity and/or water penetration,the most widespread are protective coatings (paints, thin coatings, impregnation techniques, etc.)(See the LCPC Guide entitled Concrete protection). Applying paint constitutes a solution with justextremely limited efficiency in counteracting the effects of delayed ettringite formation (DEF) and isthus not recommended. The placement of a thicker (up to several millimetres) concrete protectivecoating offers one means of protection, provided that the systems employed are sufficiently sealed(to water vapour as well). This type of coating however only maintains its efficiency over a limitedlife cycle (on the order of ten years), which necessitates a number of coating replacements duringthe duration of structural use, which in turn directs the choice of protection towards the kinds ofprevention solutions presented in Chapter 3.

Applying a concrete protective coating thus enables avoiding reaction by means of minimisingwater penetration into the structure. Such a solution may be employed to accompany a morereliable prevention solution, while not definitively ensuring DEF prevention.

Moreover, within the scope of structural monitoring efforts, it is necessary to inspect the partsdeemed to be critical so as to detect potential cracks that may appear and allow water to penetrateinto the concrete. It is thus essential to treat these cracks with a sealant; various techniques areavailable, including bridging and injection).

For the other parts, the reader is advised to consult the LCPC Guide on concrete protection (citedabove) as well as a SETRA memorandum (CTOA N° 25): Do not mistake a bridge deck surfaceseal for concrete protection.

In the case of an underground critical part, the surrounding embankments must be equipped withdrainage systems capable of channelling runoff water.

Note : The application of cladding can accompany a more reliable solution, yet does not in and ofitself constitute an adequate means of prevention.

2.2. Limiting the temperature rise in concrete

It is recommended to avoid critical structural parts by jointly optimising element material anddesign. Thus, the use of High Performance Concrete (HPC) could enable producing slenderstructures with less sensitivity to the DEF risk. From a general standpoint, it is advised to designstructures that integrate hollow parts or, whenever possible, voided parts.

As an example, the order of magnitude associated with the potential impact from using a hollowelement on the maximum temperature reached is as follows :• for the addition of an internal formwork to a pier (erection of a hollow pier instead of a solid pier,yielding an apparent thickness of 0.5 m instead of 3 m, with a concrete containing 350 kg/m3 ofCEM I 52.5 N), the drops of the maximum temperature equals approximately 15 °C.

3. Provisions related to concrete mix designThe choice of type of cement and eventual addition by the prescriber must incorporate theinfluence of these materials on concrete heating.

The selected binder must in fact be the least exothermic possible while remaining compatible bothwith specifications tied to exposure classes and with the early-age concrete strength requirements(it is preferable to define early-age strength specifications with as few constraints as possible, andeven to define no specification at all for potentially difficult concreting, yet such a decision mustnaturally be integrated into the overall project schedule). Similarly, binder concentration may beminimised while still meeting the set of workability, strength and durability requirements. The use ofcomposite cements and/or incorporation of additions offer adapted solutions for minimisingconcrete exothermic reactions.

Similarly, binder concentration may be minimised while still meeting the set of workability, strengthand durability requirements. The use of composite cements and/or incorporation of additions offeradapted solutions for minimising concrete exothermic reactions.

For purposes of illustration, the replacement of a concrete with CEM I 52.5 N cement by onecontaining CEM III 42.5 N cement in the mix design (on an element with a 1-m apparent thickness)is reflected by a drop in maximum temperature of approximately 15°C.

4. Provisions related to concrete fabrication and transport Before proceeding with an examination of the provisions potentially adopted to limit concreteheating temperature, it would be worthwhile to review a key parameter involved in this type ofheating: concrete heat capacity.The concrete heat capacity is defined as the quantity of heat needed to be input to a unit mass ofconcrete to raise its temperature by 1 °C. This capacity depends on the actual heat capacity ofeach of the individual concrete components. The following table provides an order of magnitude ofheat capacities experienced when producing one cubic metre of concrete :

29


Component Mass Mass heat capacity Component heatcapacity

(kg) (kJ/kg*K) ∗ (kJ/K)

Coarse limestone aggregate (dry) 1,050 0.84 882

Siliceous limestone sand (dry) 750 0.80 600

Cement 400 0.75 300

Total water 200 4.18 836

Total 2,618

∗ According to Loïc Divet - ERLPC OA 40.

TABLE IV - HEAT CAPACITY OF VARIOUS CONCRETE COMPONENTS

This table shows that due to their presence in large quantities, aggregates (coarse + fine) exertconsiderable influence on the mix in terms of heat capacity ; this finding signifies that a change inaggregate temperature will cause the most significant temperature change at the level of the entiremix. The above table also demonstrates that mixing water plays a major role in the ultimate mixtemperature and that substituting this water by ice could trigger a sizeable decrease in heat contentof the mix (in which case, it becomes necessary to include the heat of ice melting, which equals334 kJ/kg). Concrete temperature upon setting must be controlled and can be lowered by means ofseveral methods :• use of cold or refrigerated mixing water,• aggregate cooling (spraying water on the coarse aggregates),• protection of aggregate stocks from exposure to the sun,• substitution of ice for a portion of mixing water.The first two methods are relatively simple to employ even though they both require specialcustomised equipment that is not typically available at ready-mixed concrete plants. The use of iceis more complex and necessitates a specific large-scale installation, which in particular demandsextended mixing time to guarantee complete melting.


30

The technique of injecting liquid nitrogen into the concrete mixing plant or mobile mixer may proveadvantageous; it does not however enjoy widespread use due to its relatively high cost andtechnical sophistication.As an example, the orders of magnitude for the potential impact of the various parameters citedabove on the maximum temperature reached are :• with cold water mixing (temperature of 4 °C instead of 20 °C), cooling provides an impact ofapproximately 3 °C ;• for the spraying of coarse aggregates using cold water (yielding a 10 °C temperature drop),cooling also amounts to approximately 3 °C.Note : It is necessary to account for cement temperature during its onsite delivery. A recently-produced cement leaving the plant can in fact reach a high temperature (above 50 °C in someinstances). As an example, a 10 °C rise in cement engenders a 1 °C temperature increase inconcrete.

The impacts of both material transport and mixer truck delays also need to be considered, for thepurpose of minimising wait times. Every effort must obviously be made as well to limit the amountof time mixer trucks are parked without shade protection.

5. Provisions related to concrete placement

5.1. Jobsite organisation

Ambient temperature conditions are difficult to control ; more specifically, it is generally impossibleto select the concreting season given the jobsite's overall scheduling constraints. It would still bepreferable however to choose a time of day that helps minimise the temperature of fresh concrete(i.e. end of the day or at night).

For example, the order of magnitude associated with the potential impact of this parameter on themaximum temperature reached would, for night-time concreting (during the summer season), bereflected by a temperature decrease of approximately 5 °C.

It is highly recommended, on large-sized elements, not to attempt saving on prescribed internalformwork so as to retain just those sections that actually prove vital to structural strength (e.g.bridge piers). It is also advised to avoid creating, strictly for the purpose of facilitating erection, solidparts even though the design calls for hollow parts.

For large-sized elements, it is possible to anticipate separating the concreting into several phasesso as to promote heat exchange. This division is only efficient if a long enough period (at least aweek) is maintained between successive castings. It is nonetheless advised to remain withinacceptable time limits that enable : preserving monolithic structural behaviour, positioning theconcrete cast joints within appropriate zones from a mechanical perspective, and lastly respectingrules governing the appropriate realisation of concrete cast joints.

5.2. Concrete cooling

It is also possible, as a complementary approach, to cool the concrete after setting by means ofincorporating coils into the concrete, with such coils constituting a cooling circuit through which coolwater is circulated. This method must undergo a cooling system design in order to avoid thedevelopment of thermal gradients inside the concrete mass, especially in the vicinity of tubes, giventhat these gradients can produce radial or tangential cracks. Furthermore, the method is only trulyefficient when the concrete displays a moderate or weak exothermic reaction. In the case of ahighly exothermic composition, which is not optimal from the standpoint of thermal problemsspecific to the element being concreted, the (temperature-activated) heat release is much fasterthan the rate at which calories are removed by the cooling circuit. Lastly, cooling circuit installation

31


Pre-treatment

t1

Gm Gr

TP

TO

t2 t3 t4

Theoretical cycleActual cycle

T (˚C)

CoolingPlateau phaseTemperaturerise

FIGURE 1 - General shape of a cycle.

interferes with reinforcement work and execution times get extended. Cooling in the concrete massmust be undertaken only as a last resort (let's also point out the ultimate necessity of plugging thepipes with a cement grout).

5.3. Choice of formworkFor medium-sized parts, formworks that favour heat exchanges may serve to limit the maximumtemperature reached within the concrete.

As an example, a 40-cm shear wall concreted in a wooden formwork with a C40/50 class concreteand a 400 kg/m3 content of CEM I 52.5 R could lead to a 40 °C temperature rise. When using ametal formwork, this increase would be reduced to 35 °C.

6. Provisions specific to prefabricationThe need to decrease the duration during which the means of production are immobilised whileincreasing the number of daily production runs could lead to accelerating, via external heat input,the various chemical reactions involved in the concrete hardening process.

Concrete hardening is being accelerated in order to introduce into the concrete sufficientmechanical strength to perform, depending on the specific case, the steps of demoulding, handling,the release of prestressing forces or even appearance treatments.

As a general rule, both the heat treatment and means used for its application must be determinedby incorporating element geometry and dimensions, concrete composition, concrete plasticity andfabrication conditions, such that the steps of element demoulding, prestressing, lifting or transportmay be performed upon completion of treatment. Moreover, treatment protocols must be examinedby including the ambient temperature and relative humidity conditions associated with productionand storage, so as to avoid any thermal shock as well as the occurrence of cracks or surfacedefects detrimental to either concrete durability or structural element appearance.

In its most general form, a thermal cycle comprises four phases (Fig. 1), with each defined by aduration-temperature pair or by a rate :• pre-treatment phase,• temperature rise phase,• constant temperature plateau phase,• cooling phase.

Preliminary tests are carried out to optimise each of these phases.


32

Pre-treatment phase

The pre-treatment phase is intended to introduce into the concrete a sufficient amount of cohesionso that it is able to absorb the internal forces generated due to the thermal dilatation of itscomponents, in particular water and air, during a temperature rise. The pre-treatment period mustbe extended as the rate of temperature rise during the subsequent phase quickens and concretesetting slows (Fig. 2).

Temperature rise phase

The rate of temperature increase must be such that the forces caused by dilatation developedwithin the element can be absorbed at any time by the concrete, which gradually becomes stiffer.For discussion purposes, Figure 3 shows the order of magnitude for the maximum rate oftemperature increase Gm (as measured in °C/hr) vs. the maximum steaming radius Remax (in cm). Definition of Remax :In considering the set of shortest distances separating each point of the concrete from the heatedface, the value of Remax corresponds to the greatest of these distances (Fig. 4).

CementCEM I 52.5 R

CementCEM I 52.5 N

10 20 30 40

1

2

3

4

5

30˚C

30˚C

20˚C

Duration of pre-treatment (hrs)

Speed of temperature riseduring the second phase (˚C/h)

Temperatureat the end of mixing 15˚C

FIGURE 2 - Influence of the type of cement and rate of temperature

rise on pre-treatment phase duration.

Gmmax (˚C/h)

Remax (cm)0 5 10 15 20 25 30 35 40

0

10

20

30

40FIGURE 3 - Values of the thermal

gradient Gmmax vs. maximumsteaming radius Remax.

Constant temperature plateau phase

Both the duration and temperature of this phase, during which the concrete hardening process(already initiated during the previous phase) continues its course, depend on :• maturity acquired by the concrete upon completion of the temperature rise phase,• number of daily production runs,• desired level of strength.

The duration of this plateau depends on the actual plateau temperature; it typically lies between 1and 3 hours for 85 °C, and between 4 and 12 hours for 65 °C. As regards the pertinent standards,NF EN 13369 stipulates maximum concrete temperature as a function of the environment to whichthe precast element will be exposed.

Over the course of this phase, special attention is focused on applying the necessary precautionsto avoid concrete drying so as to ensure that hydration is completed as much as possible.

It must also be ensured that the temperatures between the various points on the large-sizedelements or among the various elements subjected to the same treatment remain close in valueand constant in order to obtain identical strength levels and avoid the detrimental consequences ofdifferential dilatations.

Cooling phase

Cooling must also proceed in a homogeneous manner. Disorders can in fact be more easilyascribed to the temperature differences that exist between the various points of an element than tothe actual cooling rate itself. The surface cooling rate is higher than the rate at the element core.Cracking risks arise whenever the temperature difference between element core and surfaceexceeds 15 °C.

Heat treatment example

Figure 5 depicts a thermal cycle example applied to a self-compacting concrete mix design. Themaximum temperature reached at the concrete core equals 68 °C for the first cycle. The

33


Remax

Remax = a (a>b)

10 cm

20 cm

40 cm

(a)

(b)

θ

θ

Remax

FIGURE 4 - Examples serving to define the maximum steamingradius Remax.


34

mechanical compressive strength values obtained on 11 × 22 cylindrical specimens after 18 hoursand 28 days are 39 MPa and 59.5 MPa, respectively.

0

20

40

60

80

0 5 10 15

T (˚C)

Time (hrs)

Recommendedsteam temperature

Specimen core

FIGURE 5 - Exampleof a thermal cycle.

35

AppendicesI. Some precisions on the Delayed Ettringite Formation

II. Evaluation of structures affected by DEF and presentation of disorders

III. Review of the exothermic reaction of concretes

IV. Estimation of temperatures reached in planned structures

V. Performance-based testing

37


Appendix ISome precisions on the Delayed Ettringite Formation

The phenomenon of Delayed Ettringite Formation (DEF) actually stems from the delayed formationof a mineral called ettringite, which displays the chemical formula 3CaO.Al2O3.3CaSO4.32H2O.This time-delayed formation of ettringite is capable of causing an expansion, as revealed by theappearance at the concrete surface of multidirectional cracking with a relatively wide grid pattern.Yet ettringite is not systematically detrimental to concrete given that it is a standard product ofcement hydration. For this reason, we will provide below some background information on thevarious types of ettringite encountered in concrete.

The various forms of ettringite (primary, secondary and delayed)At present, several terms are found in the literature to distinguish the various modes and formationtimetables for ettringite in concrete. Three types of ettringite capable of coexisting in a givenconcrete sample can be identified. The names assigned are those proposed by the FrenchAssociation of Civil Engineering (AFGC), as forwarded by its working group « GranDuBé » (Frenchacronym for Magnitudes associated with Concrete Durability) [1] :• primary formation of ettringite, which does not cause any expansion ;• secondary formation of ettringite, which might cause expansion ;• delayed formation of ettringite, subsequent to a temperature rise imposed on the concrete atan early age, which in turn can also cause expansion.

Primary ettringite formation corresponds to a product yielded by cement hydration taking placedue to reaction between the setting/hardening regulator (calcium sulphate) and the calciumaluminates. Ettringite is encountered in the form of needle-shaped (or acicular) crystals, which donot cause swelling since they assemble prior to concrete hardening within unoccupied spaces ofthe material. They even serve a beneficial purpose by contributing to cement paste cohesion at anearly age by means of decreasing porosity while simultaneously increasing mechanical strength ofthe mix. This phenomenon is especially predominant during the setting of supersulfated cementsand sulphoaluminate cements.

Secondary ettringite formation occurs once the concrete has already hardened and is the resultof water movement inside the concrete (dissolution / precipitation phenomena) as well as externalexternal sulphate-based contributions (soils, water, etc.) or internal contributions (use ofaggregates containing sulphates, mixing water). In this latter case, compliance with standardsallows avoiding such a reaction. For the dissolution / precipitation phenomena, ettringite crystallisesin a needle-shaped (acicular) pattern within the unoccupied space of concrete and typically doesnot display expansive behaviour. On the other hand, this secondary-stage formation subsequent toan external contribution of sulphates is capable of generating swelling. As opposed to the non-expansive facies of the material, this pathological ettringite crystallises into a massive andcompressed arrangement.

Delayed ettringite formation pertains solely to concretes that had been exposed during the earlyage to heating in excess of 65 °C. Beyond this temperature, the primary ettringite does not formduring cement hydration reactions and/or becomes decomposed. The source of sulphate ionswould thus be internal given that these ions are generated due to the absence or decomposition ofprimary ettringite. After returning to ambient temperature and in the presence of humidity, ettringiteis able to form or reform, at which point it is capable of generating swelling pressures under certainconditions.

Appendix

38

The physicochemical mechanisms associated with DEFAn internal sulphate reaction therefore arises by the delayed formation of ettringite within acementitious material, following setting and without any contribution from external sulphates. Themechanism may be schematically decomposed into several sequences, which feature the creationof a source for potentially-remobilised sulphates, the delayed precipitation of ettringite andexpansion as revealed by concrete cracking.

Constitution of a sulphate sourceA considerable rise in temperature modifies cement hydration reactions. As a matter of fact, thesulphates inputted by the setting/hardening regulator cannot be entirely mobilised in order to formprimary ettringite. Moreover, primary ettringite solubility increases with temperature, which isreflected by a higher sulphate ion concentration in the concrete pore solution. A significant quantityof these ions is also trapped by physical adsorption at the surface of some cement hydrationproducts. This phenomenon is indeed reversible, thus offering a reserve of sulphate for subsequentettringite formation.

Ettringite precipitationAfter the hardening of concrete and a return to ambient temperature, eventually in conjunction withleaching of alkali contained in the concrete pore solution, ettringite precipitation can arise atreactive sites that include aluminates. Concentration conditions may then become dominant, whichin turn would lead to a highly unstable local chemical system along with the formation, within anenclosed space, of an ettringite often qualified as poorly-crystallised. This formation process canlocally produce high pressures and cause swelling.

ExpansionThe expansive or non-expansive nature of ettringite depends on the initial chemical composition,particularly on the type of cement (contents of aluminates and alkali, quantity of potentially-formedPortlandite) and the quantity of sulphates capable of being mobilised. The precise mechanism bywhich ettringite formation is able to generate pressures inside concrete is still not unanimouslyacknowledged. Two principal mechanisms, which are loosely associated, have been proposed toexplain the swelling induced by ettringite formation :• swelling correlated with the crystallisation pressures inherent in ettringite crystal growth,• swelling correlated with the osmotic pressures due to colloidal ettringite expansion.In reality, it is likely that both of these mechanisms play a role simultaneously and cannot actuallybe dissociated from one another.

39


Evaluation of structures affected by DEFThe first known cases in which DEF has been considered as the main source of degradationpertain to precast concrete railroad ties subjected to heat treatment. Several countries haveexperienced this phenomenon, with references available in particular for: Finland in 1987 [2],Germany in 1989 [3], the former Czechoslovakia in 1991 [4], Australia in 1992 [5], South Africa in1992 [6], the United States in 1995 [7], and more recently Sweden in 2004 [8]. The first disordersstarted occurring after a number of years in operation, typically within 10 years and in all casesthese ties had been exposed to humidity: the same ties placed in tunnels in fact show no signs ofalteration. In certain cases, DEF is associated with other concrete degradation mechanisms likefreeze-thaw or the alkali-silica reaction.

DEF has also been identified in other prefabricated concrete elements. More specifically, thisreaction has been spotted on a staircase wall of a parking lot in the United States [9], on post-tensioned beams and gutters in Great Britain [10], in the stands of a stadium in the United States[11], and inside asbestos cement roofing in Italy [12]. In most of these prefabricated elements,disorders were observed less than 10 years after construction.

DEF has also been discovered in massive, cast-in-place concrete components. Theseobservations relate in particular to foundations of electric transmission line towers in the UnitedStates [9,13] and Italy [12]. Such disorders began appearing between 3 and 8 years after concretepouring. A major structural evaluation campaign, undertaken in Great Britain, revealed 23 caseswhere parts of bridges had been affected by DEF : foundations, abutments, joists, wing walls [10].These concretes were most often cast during summertime and contained high contents of cement(between 420 and 550 kg/m3), as well as high equivalent alkali contents (> 4.0 kg/m3). Suchstructural elements tend to be quite thick (at least 60 cm). The maximum temperature reachedwithin the concrete components was estimated at close to 80 °C, and the disorders beganappearing at between 8 and 20 years following construction.

In France, the discovery of DEF and its signs of degradation is rather recent, with the initial casesbeing identified during the 1990's. This phenomenon was primarily observed in massive concretebridge components cast in place during summer heat waves [14,15]. The cases of unhealthystructures, inventoried in France by the Laboratoire Central des Ponts et Chaussées (LCPC), arefor the time being quite limited in number (around twenty), yet most concern Category II structures.The concretes in this category have for the most part been attained by the DEF while not by thealkali-silica reaction. From a general standpoint, the manifestation of disorders is visible on astructure between 5 and 10 years after construction. Furthermore, these disorders never affect theentire structure but primarily the massive parts exposed to humidity or water inflow. A considerableheating of concrete has been suggested (on the order of 80 °C) as a result of specimen geometry,casting period (summertime) and a high cement proportion.

Appendix IIEvaluation of structures affected by DEF

and presentation of disorders

Appendix

40

A more detailed expert appraisal was conducted on the basis of eight structures for the purpose ofsimultaneously identifying the parameters present and a priori essential to DEF development [15].These parameters have been classified into four groups :• temperature-related-parameters ;• cement-related parameters ;• concrete-related parameters ;• environment-related parameters.The corresponding date have been collated in Table V.

Bridge A Bridge B Bridge C Bridge D Bridge E Bridge F Bridge G Bridge H

Year built 1955 1967 1980 1988 1990 1982 1988 1989

Targeted part of the structure Crossbeam Pier Crossbeam Pier Pier Crossbeam Pier base Crossbeam

Temperature-related parameters • T max (°C) > 80 > 80 > 80 > 75 > 80 > 70 > 75 > 75• Concreting period August Anknown August/ July/ August/ July/ July/ July/

September August September August August August

Cement-related parameters• Type of cement CPA CPAL CPA 55R CPJ 55 CPA 55R CPA CPA CPA 55R

(10 % (10 % lime-slag) stone filler)

• SO3 (% mass) 2.5 2.7 2.6 2.5 2.8 3.2 2.2 3.5

• C3A (% mass) 11.2 9.6 9.8 7.0 8.2 11.0 7.1 10.1

Concrete-related parameters• Cement content (kg/m3) 430 430 400 380 410 350 385 400• W/C ratio 0,50 0,50 0,47 0,54 0,46 0,49 0,48 0,50• Type of aggregates Siliceous Siliceous Sand-lime Siliceous Siliceous Siliceous- Siliceous Siliceous-

calcareous calcareous• Alkali content (Na2O équivalent en kg/m3)* 2.0 4.3 4.0 4.1 2.3 3.0 3.9 4.6

Environment-related parameter• Humidity Water- Lack of Condensation Tidal zone Tidal zone Exposed to Exposed to Lack of

profing drainage Wetting- inclement inclement drainageproblem drying weather weather

alternation

TABLE V - COMPARATIVE STUDY OF THE VARIOUS DEF DETERMINANT FACTORS ENCOUNTERED ON THE GROUP OF BRIDGES EVALUATED

* Na2O equivalent = Na2O + 0,658 K2O.Note : The data on bridges «A» through «E» have been extracted from reference [14], while bridges F through H were taken asanonymous sources.

41


FIGURE 6 - Extremity of railroad tiedisplaying a network of cracks causedby DEF (outside of France).

FIGURE 8 - Localised multidirectionalcracking on the vertical face, at theextremities of a crossbeam on the bridgeshown in Figure 2.

FIGURE 7 - General view of a bridge witha crossbeam affected by DEF (France).

Appendix

42

From study findings, it would appear that the DEF phenomenon arises whenever the followingconditions are applicable :• considerable concrete heating, near a temperature of 80 °C, resulting from several factors(massively-sized parts, highly-exothermic cement, a strong cement concentration and concretingtaking place during the summer) ;• the use of cements whose alkali content exceeds 0.6 % Na2O equivalent, with a SO3 content onthe order of 2.6 % and a C3A content value between 7 % and 11 % ;• the presence of wetting / drying cycles or relatively humid conditions ;• aggregates created for the most part from siliceous or silicate rocks.

It still needs to be pointed out however that we presently lack knowledge on the actual status of allFrench structures experiencing this pathology, given that any cataloguing effort proves verychallenging for various reasons :• a set of symptoms similar to the alkali-silica reaction ;• an alkali-silica reaction in some instances associated with DEF ;• difficulty in diagnosing DEF ;• structures under judicial investigation requiring strict confidentiality.

Présentation of disordersDEF is manifested by the development of expansion, which in turn produces major degradations atthe structural level. Visual symptoms are similar to those caused by an alkali-silica reaction. Mapcracking is frequently observed on the facings of structures exhibiting DEF, and this cracking isoften considered to be disorderly to the extent that it is composed of cracks with erratic patterns.

Cracking can also assume the form of a network of widely-spaced multidirectional cracks andsometimes shows preferred orientations, depending on the distribution of reinforcements. The sizeof crack openings varies from several tenths of millimetres to a few millimetres and increases asthe reaction advances. These cracks often get exacerbated by humidity and a whitish exudate issometimes observed at the level of individual cracks.

43


Concrete setting and hardening is accompanied by a heat release due to the exothermal nature ofhydration reactions. Depending on the concrete mix design and especially the type and content ofbinder, heat quantities can actually vary widely, as can the rate of heat release and the sensitivity ofthis rate to temperature.

For purposes of illustration, Figure 9 will now provide the quantities of heat released by eightFrench manufactured cements of different types, all measured on standardised mortars conservedunder quasi-adiabatic conditions within a thermally-insulated « Langavant » bottle :

Appendix IIIReview of the exothermic reaction of concretes

0

50

100

150

200

250

300

350

400

450

-10 10 30 50

Heat release (J/g)

Actual age (hrs)

CEM I 52.5 R

CEM I 52.5 N

CEM I 52.5 N PM-ES

CEM II/A-LL 42.5 R

CEM II/A-LL 42.5 N

CEM III/A 42.5 N

CEM III/B 32.5 N

CEM V/A (S-V) 32.5 N

FIGURE 9 - Examples of heat release curves for various types of cement.

A study conducted by Committee TC 51 of the CEN in 1995 on heat releases under semi-adiabaticconditions for a dozen European cements revealed that the quantities of heat released after 41hours (Q 41) varied between 210 and 320 J/g. The ratios of heat released at 41 hrs (Q41) to thatreleased at 72 hrs (Q 72) varied between 0.85 and 0.95 depending on type of cement.

In general terms, hydration reaction rates are heavily influenced by temperature (e.g. anacceleration by a factor ranging from 2 to 4 is observed when temperature equals 40 °C instead of20 °C); moreover, the sensitivity of concrete mix design is described by a parameter referred to asthe activation energy (Fig. 10).

Appendix

44

This sensitivity (and thus this energy) drops when the binder shows greater reactivity during theearly age, as illustrated in the table 6, which offers a number of sample values for the activationcoefficient (E/R, which stands for activation energy divided by the perfect gas constant - keep inmind that this value can still vary significantly from one concrete to the next for a given type ofcement), corresponding to different types of cement :

0

10

20

30

40

50

60

0 5 10 15 20 25 30

Rate of heat release (J/g/hr)

Time (hrs)

20 ˚C

40 ˚C

FIGURE 10 - Rate of heat release for a CEM I 52.5 R cement vs. temperature.

Cement CEM I 52.5 R CEM I 42.5 R CEM I 42.5 N CEM II 32.5 R CEM II 32.5 N CEM III/C 32.5 N

E/R (°K) 3,540 3,970 4,150 4,810 5,530 6,700

TABLE VI - SAMPLE VALUES OF THE ACTIVATION COEFFICIENT FOR VARIOUS TYPES OF CEMENT

The amount of temperature rise within a concrete element depends not only on the exothermalreaction of the material, but also on element geometry, the initial material temperature and thermallosses. It is impossible to establish a precise limit regarding element thickness (in most cases,thermal losses occur along a preferred direction, and the element dimension along this directionyields the « thickness » as denoted herein). Based on this « thickness », the term « massiveelement »may be employed, and this raises inherent fears over an excessive concrete temperatureincrease, since the race between heat release and heat losses involves at the same time thematerial, the element geometry and the boundary conditions. A base plate measuring 1.5 m thickfor a C30/37 concrete with a content of 370 kg/m3 of CEM III/A 42.5 N cement could display a29 °C temperature rise, whereas a 60-cm shell concreted using wooden formwork with a C40/50containing 400 kg/m3 of CEM I 52.5 R would lead to a rise of 45 °C. The notion of massive elementthus proves irrelevant in preventing risks associated with the Delayed Ettringite Formation.

The notion of « critical » element will therefore be used in reference to a concrete elementwhose heat release will only be very slightly discharged to the outside and will lead to asizeable concrete temperature rise.

It should be noted that temperature is not uniform within the concrete and that relatively sharpgradients (depending on insulation conditions provided by the formwork) are present around theperiphery. For this reason, the maximum temperature being targeted for the delayed ettringiteformation is the one actually reached at the element core.

45


In order to evaluate the temperature rise within a concrete element and determine whether theconditions of a « critical » element have been met, a number of tools are provided in Appendix 4 ofthis report.It also needs to be pointed out that the temperature rise within a concrete part can cause damageat several levels :• cracking due to impeded thermal shrinkage,• alteration to long-term concrete mechanical properties,• risk of delayed ettringite formation.A substantial temperature rise is in fact often correlated not only with high gradients between theconcrete core and skin, which are capable of causing cracks during initial surface cooling (thesecracks are less serious since subsequent cooling at the concrete core tends to keep them closed),but also with temperature gradients relative to the previously-cast adjacent elements that have hadsufficient time to cool. In this latter case, it is observed a through-cracking pattern which isregularly-distributed (with the distribution pattern being even more regular at higher densities ofreinforcement bars), and is occurring in the concrete cast joints. This cracking may serve tofacilitate water penetration into the material and promote delayed ettringite formations. Gradientsmay also develop in zones featuring an abrupt change in concrete element cross-section, and thisin turn gives rise to damaging through-cracking.

Furthermore, as mentioned above, hydration reactions may be accelerated as a result of an inflowof heat, yet it is well known that high maturation temperatures modify the nature of hydrates formedand diminish concrete characteristics over the long term, in comparison with the same concretethat had not been steam-cured, particularly in terms of compressive strength. As was the case forthe impact on short-term strength values, long-term strength impact depends to a great extent onthe actual concrete mix design. As an example, a 70 °C heat treatment lasting 9 hours (with a10 °C/hr temperature rise) can induce a 10 % compressive strength loss measured at 700 days,whereas a 90°C treatment for 6 hours (at the same rate of temperature increase) can yield a dropof 20 % (ref. : Mamillan).

It goes without saying that by optimising the set of heat treatment parameters, durable concretescan still be obtained despite the decrease in mechanical strength.

47


1. PurposeThis appendix is intended to propose a simplified methodology for assessing, prior to breakingground on a building project, whether some elements need to be considered as critical with respectto the risks of delayed ettringite formation (in correlation with the risk of excessive temperature atthe core of cast elements), in light of the concrete design principles adopted as part of the contracttechnical specifications.

The present appendix thus offers an estimation of the maximum temperature at the element core,for which only the thickness (at its smallest dimension) and a smattering of basic data on concretecomposition are actually known.

The precision of this method remains limited due to the small number of parameters introduced (i.e.those known or easily obtained during the preliminary phase). It should be used as a warning tool ;if its conclusion is that the element is critical, then a more in-depth study must be carried out or theparameters must be modified.

2. Data required to estimate the maximum temperature reachedThe simplified method can be run once the following parameters have been determined:• cement content of the concrete mix, C (en kg/m3),• content of mineral additions, A (kg/m3),• mass density of the concrete, Mν (kg/m3),• content of efficient water in the concrete, Eeff (kg/m3),• 2-day compressive strength of the cement, Rc2 (MPa), according to Standard NF EN 196-1, • 28-day compressive strength of the cement, Rc28 (MPa), according to Standard NF EN 196-1, • cement hydration heat at 41 hours, Q41 (kJ/kg), according to Standard NF EN 196-9,• element thickness, EP (m) - only if greater than 0.25 m (below this threshold value, the element isnot critical with respect to the risks of delayed ettringite formation).The cement-related data are generally available from technical specifications.

Element thickness is established from the smallest dimension (in the preferred direction, forthermal losses).

3. Calculation steps The sequence of calculation steps is presented in the flowchart below (Fig. 11) and then describedin greater detail in the following discussion :

Appendix IVEstimation of temperatures reached in planned structures

Appendix

48

3.1. Estimation of the heat release at infinity for the selected cementSince the technical files on cements only indicated heat release values at 41 hours, this first stepconsists of estimating the maximum heat released over the long term by the cement, Qm (in kJ/kg),by applying the formula :

Qm = Q41 × ratio_Qm/Q41

where the value of ratio_Qm/Q41 is given by Abacus n° 1 (Fig. 12), based on the Rc2 ratio.Rc28

Step n˚ 1

Estimation of the heat release at infinity for the selected cement:determination of Qm (Q41, Rc2/Rc28)

Step n˚ 2

Incorporation of mineral additions:Determination of the heat equivalent binder

LEch (type and quantity of mineral additions, EP, C)

Step n˚ 3

Incorporation of the impact of Eeff/LEch ratioon temperature rise:

Determination of a corrective term α (Eeff/LEch)

Step n˚ 4

Estimation of the temperature risein the absence of thermal losses:

determination of ΔTadia (Qm, LEch, Cth, Mv, α)

Step n˚ 5

Incorporation of thermal losses:Determination of a reduction coefficient R (Q41, EP) to assign to ΔTadia

in order to estimate temperature rise of the element ΔT

The estimated temperature value at the core of the element equalsthe sum of concrete temperature at the time of casting plus ΔT

FIGURE 11 - Successive steps for the calculation of the maximum temperature reached in structures to be build.

49


1

1.1

1.2

1.3

1.4

1.5

1.6

0.2 0.3 0.4 0.5 0.6

Qm/Q41

Rc2/Rc28

Qm/Q41 Ratio for the evaluation of Qm

ABACUS n° 1

3.2. Incorporation of mineral additionsConcrete mix additions contribute to the concrete heat release. To incorporate these components,a heat equivalent binder is introduced, denoted LEch (in kg/m3), by means of the followingformula :

LEch = C + K' Awhere : A is the addition content

K' is the weighting coefficient of additions derived from Abacus n° 2 (Fig. 13).

For both limestone and siliceous additions, the coefficient K' equals 0.

FIGURE 12 - Abacus n° 1 for the estimation of the maximumheat released over the long termby the cement (Qm).

FIGURE 13 - Abacus n° 2 for the estimation of the weightingcoefficient of additions (K').

0

0,2

0,4

0,6

0,8

1K'

0 1 2 3 4 5 6 7Element thickness EP (m)

Fly ash

Silica fume (K' = 1) Slag

Value of coefficient K' for additions.as input into the"heat equivalent binder" calculation

ABACUS n° 2

3.3. Incorporation of the impact from the Eeff/LEch ratioThe temperature rise stemming from heat released by the binder also depends on the Eeff/LEchratio, which affects the maximum long-term rate of hydration: as the value of this ratio decreases,hydration remains further from completion and less heat gets released. This correlation is takeninto account via a corrective term α (in °C) that can be derived from Abacus n° 3 (Fig. 14) :

Appendix

50

3.4. Estimation of the temperature rise in the absence of thermal lossesAt this stage, it is possible to evaluate the temperature rise ΔTadia (en°C) under adiabaticconditions (i.e. perfect insulation) based on the formula :

ΔTadia = (Qm × LEch)/(Cth × Mv) + α

where Cth is the thermal capacity of the concrete, set equal to 1 kJ/(kg .°C).

3.5. Incorporation of thermal losses Thermal losses heavily depend on the level of cement response and element thickness (as the rateof heat production is competing with dissipation rate). Using Abacus n° 4 (Fig. 15), the reductioncoefficient R (lying between 0 and 1) can be obtained ; this result allows taking thermal losses intoaccount, with cement response being expressed via the Q41 parameter :

-6

-4

-2

0

2

4

6

8

0.35 0.450.40 0.550.50 Eeff/LEch

α (˚C)

Correction of the temperature riseas a function of the ratio Eeff/LEch

ABACUS n° 3

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

200 300 400

Reduction coefficient R

Q41 (kJ/kg)

EP = 0.25 m

EP = 0.5 m

EP = 0.75 m

EP = 1 m

EP = 2 m

EP = 3 m

EP = 4 mABACUS n° 4

FIGURE 14 - Abacus n° 3 for theestimation of the corrective

coefficient α as a function ofEeff/LEch.

FIGURE 15 - Abacus n° 4 for theestimation of the reduction

coefficient R as a function ofthermal losses.

51


Should the thickness EP be greater than or equal to 5 m, the value R = 1 is assigned.

The R value serves to estimate the temperature rise ΔT (°C) using the formula :

ΔT = R × ΔTadia

If this addition to the value of fresh concrete temperature at the time of concreting exceeds thespecified limit with respect to the selected level of prevention (see Chapter 3), the element isconsidered to be critical and then only a more detailed study would justify the acceptability ofheating from the standpoint of delayed ettringite formation risks.

4. Application examples

4.1. Example n° 1 : Concrete without any additions (element thickness : 1.00 m)Data• Cement content in the concrete mix C = 350 kg/m3

• Content of mineral additions A = 0 kg/m3

• Mass density of the concrete Mv = 2,400 kg/m3

• Efficient water content of the concrete mix Eeff = 175 kg/m3

• 2-day compressive strength of the cement Rc2 = 27 MPa• 28-day compressive strength of the cement Rc2 = 68 MPa• Hydration heat of the cement at 41 hours Q41 = 306 kJ/kg• Element thickness EP = 1.00 m

Step n° 1 : Estimation of the heat release at infinity for the selected cementFrom the data presented above, the ratio Rc2/Rc28 = 0.4. The use of Abacus no. 1 enablesdetermining the ratio Qm/Q41 = 1.25The maximum heat released over the long term is found to be: Qm = Q41 x 1.25 = 383 kJ/kg.

Step n° 2 : Incorporation of mineral additionsThe concrete mix design contains no mineral additions, thus A = 0 kg/m3. In this case, theequivalent binder LEch = C.

Step n° 3 : Incorporation of the impact from ratio Eeff/LEchIn this example, Eeff/LEch = Eeff/C = 0.5, which corresponds (i.e. in reading from Abacus n° 3) tothe corrective term α = 2 °C.

Step n° 4 : Estimation of temperature rise in the absence of thermal lossesΔTadia = (Qm × LEch)/(Cth × Mv) + α = (383 × 350)/(1 × 2400) + 2 = 58 °C.

Step n° 5 : Incorporation of thermal lossesThe use of Abacus n° 4, with a cement hydration heat Q41 = 306 kJ/kg and a thickness of theelement to be cast EP = 1.00 m, leads to the reduction coefficient R = 0.85.The temperature rise of the element can thus be estimated by the relation :

ΔT = R × ΔTadia = 0,85 × 58 = 49 °C.

By assuming a value of fresh concrete temperature at the time of concreting equal to 20 °C, thetemperature at the element core can be estimated at 69 °C.

Appendix

52

4.2. Example n° 2 : Concrete with fly ash additions (element thickness : 3.00 m)Data• Cement content in the concrete mix C = 320 kg/m3

• Content of mineral additions A = 80 kg/m3

• Mass density of the concrete Mv = 2400 kg/m3

• Efficient water content of the concrete mix Eeff = 175 kg/m3

• 2-day compressive strength of the cement Rc2 = 27 MPa• 28-day compressive strength of the cement Rc28 = 68 MPa• Hydration heat of the cement at 41hours Q41 = 306 kJ/kg• Element thickness EP = 3 m.

Step n° 1 : Estimation of the heat release at infinity for the selected cementFrom the data presented above, the ratio Rc2/Rc28 = 0.4. The use of Abacus no. 1 enablesdetermining the ratio Qm/Q41 = 1.25

The maximum heat released over the long term is found to Qm/Q41 = 1,25.

The maximum heat released over the long term is found to be Qm = Q41 × 1,25 = 383 kJ/kg.

Step n° 2 : Incorporation of mineral additionsThe concrete mix design contains fly ash additions, yielding A = 80 kg/m3. In this case, theequivalent binder LEch = C + K' × A. The coefficient K' = 0.55 is read from Abacus n° 2 for an EP =3.00 m and silica-alumina fly ash.

LEch = C + K' × A = 320 + 0,55 x 80 = 364 kg/m3.

Step n° 3 : Incorporation of the impact from the ratio Eeff/LEchDans cet exemple, Eeff/LEch = 175/364 = 0,48 qui correspond, à partir de l'abaque n° 3, au termecorrectif α = 1,2 °C.

Step n° 4 : Step n° 4: Estimation of temperature rise in the absence of thermal lossesΔTadia = (Qm × LEch)/(Cth × Mv) + α = (383 × 364)/(1 × 2400) + 1,2 = 59 °C.

Step n° 5 : Prise en compte des déperditions thermiquesThe use of Abacus n° 4, with a cement hydration heat Q41 = 306 kJ/kg and a thickness of theelement to be cast EP = 3.00 m, leads to the reduction coefficient R = 0.95.

The temperature rise of the element can then be estimated by the relation :

ΔT = R × ΔTadia = 0,95 × 59 = 56 °C.

With a value of fresh concrete temperature at the time of concreting equal to 20 °C, thetemperature at the element core can then be estimated at 76 °C.

5. Method precisionThis method has been developed based on a series of 19 actual structural cases with knownvalues of temperature rise, through either in situ measurements or calculation by means of finiteelement simulation. An adjustment has then been introduced to yield results with a built-in marginof safety (Fig. 16).

53


6. Calculation of heating using a finite element modelA more precise method for estimating the maximum temperature reached within the concretespecimen consists of using a finite element computation code and including the heat released bythe specimen based on heat release measurements vs. time (performed as part of the calorimetrictest). This heat release can actually be deduced from measurements performed on thestandardised mortar (the so-called « cement » test), or derived either from equivalent concretemortar (ECM) or directly on the concrete.

Running a finite element calculation requires inputting a certain number of data, including :• geometry of the structure,• casting protocol (number and duration of castings, wait time between castings, etc.),• initial thermal conditions (concrete temperature, temperature of the outside environment, etc.),• physical characteristics relative to material heat conduction (volumetric calorific capacity andconductivity of the concrete, soil, steel, etc.),• boundary conditions at the periphery (imposed temperature, and eventually the imposed flow),• exchange coefficients on the model contour (surface exposed to air, metal formwork vs. woodenformwork, etc.),• data from a calorimetric test that entails recording over time the temperature curve of a concretesample representative of the structural concrete (sample temperature history, history oftemperature outside the calorimeter, coefficients of calibration specific to calorimeter losses,volumetric calorific capacity of the sample, and the Arrhenius constant Ea/R).

0

10

20

30

40

50

60

70

80

0 20 40 60 80

Estimated heating (˚C)

Measured or calculated heating (˚C)

Y = X

Y = X + 5

Y = X - 5

Actual cases

FIGURE 16 - Precision of the estimaion model (based on 19 actual structural cases).

55


General remarksA performance test focusing on concrete specimens has been developed in partnership betweenthe French Technical Association of the Hydraulic Binder Industry (ATILH), the Concrete IndustryStudy and Research Centre (CERIB) and the LCPC. This test has been published by LCPC underthe heading :

LPC Test Method n° 66 : Reactivity of a concrete mix design with respect to the delayedettringite formation - Performance testing

This operating procedure was refined for the purpose of proposing a reliable alternative method tothe tests described in the literature, whose representativeness is now challenged due to both theexcessive temperatures imposed on the material and the selected specimen dimensions.

Since the production of an initial publication as draft LPC test method n° 59 (May, 2003),subsequent tests have been conducted by CERIB and the LPC network of Ponts et Chausséeslaboratories on concretes representative of massive structural elements or on products generatedby the concrete prefabrication industry with known in situ behaviour, in the presence of the delayedettringite formation. These studies have made it possible to validate test pertinence anddifferentiation.

Reproducibility of the test method was then examined during a campaign of cross-referencing testsperformed within the scope of the work program undertaken by the GranDuBé group affiliated withthe AFGC association. The level of reproducibility depends on concrete expansion and can reachas high as 60% of the average swelling value when a 0.2 % expansion has been measured.Despite this level of dispersion, which may be attributed to the swelling phenomenon, themeasurements recorded have led all laboratories involved to the same conclusion regarding the« non-reactive » or « potentially reactive » nature of the concrete/heating pairs studied.

Interpretation of the cross-referenced test results, as a complement to the set of studies completedover the past six-plus years by the Public Works Ministry's Scientific and Technical network, hasenabled establishing the decision-making criteria described below.

Testing principleThe test consists of characterising the swelling risk for a concrete with respect to DEF. Theconcrete is defined by both its mix design and heating exposure during the early age.• concrete fabrication,• heat treatment to simulate concrete heating,• drying and wetting cycles,• definitive immersion in water, and longitudinal deformation monitoring.The minimum duration of this test calls for 12 months of immersion, and this period may beextended to 15 months should significant expansion be measured.

Appendix VPerformance-based testing

Appendix

56

Results interpretationThe test report is to contain, at the very least, the following elements :• the data listed in the test sheets appended to the test method ;• record of temperatures inside the climate-controlled chamber during thermal treatment and, ifpossible, the temperature recordings inside the concrete ;• a graph showing the plots of the expansion of each specimen as well as the average expansion ofall three specimens vs. immersion time ;• results interpretation relative to the decision-making criteria presented below. To confirm thesource of swelling, these measurements may be accompanied by a scanning electron microscopeevaluation whenever concrete swelling has been detected.The « concrete mix design / heating » pair is considered to be suitable for use provided that one ofthe two following criteria (1 or 2), pertaining to the swelling threshold and swelling curve slope, isbeing met :

Criterion 1The average longitudinal deformation of 3 specimens is less than 0.04 % and no individual valueexceeds 0.06 % by the end of the 12-month period,

ANDthe monthly variation in average longitudinal deformation for the 3 specimens measured as of the3rd month remains below 0.004 %.

Creterion 2The individual longitudinal deformation of the 3 specimens lies between 0.04 % and 0.07 % uponexpiration of the 12-month period, in which case it becomes necessary to extend the test throughthe 15th month,

ANDthe monthly variation in average longitudinal deformation for the 3 specimens measured as of the12th month is less than 0.004 %, and the cumulative variation between the 12th and 15th monthsremains below 0.006 %.

57


Norme NF EN 196-1 Méthodes d'essai des ciments - Partie 1 : Détermination des résistances mécaniques.

Norme NF EN 196-9 Méthodes d'essai des ciments - Partie 9 : Chaleur d'hydratation - Méthode semi-adiabatique.

Norme NF EN 197-1 Ciment - Partie 1 : Composition, spécifications et critères de conformité des ciments courants.

Norme NF EN 197-1/A1 Amendement A1 à la norme NF EN 197-1.

Norme NF EN 206-1 Béton - Partie 1 : Spécifications, performances, production et conformité.

Norme NF EN 450-1 Cendres volantes pour béton - Partie 1 : Définition, spécifications et critères de conformité.

Norme NF EN 1992-1-1 Eurocode 2 - Calcul des structures en béton - Partie 1-1 : Règles générales et règles pour les bâtiments.

Norme NF EN 1992-1-1/NA Eurocode 2 : Calcul des structures en béton - Partie 1-1 : Règles générales et règles pour les bâtiments - Annexe Nationale à la NF EN 1992-1-1: 2005 - Règles générales et règles pour les bâtiments.

Norme NF EN 1992-1-2 Eurocode 2 : Calcul des structures en béton - Partie 1-2 : Règles générales - Calcul du comportement au feu.

Norme NF EN 1992-2 Eurocode 2 - Calcul des structures en béton - Partie 2 : Ponts en béton - Calculet dispositions constructives.

Norme NF EN 1992-2/NA Eurocode 2 - Calcul des structures en béton - Partie 2 : Ponts en béton - Calculet dispositions constructives - Annexe nationale à la NF EN 1992-2 : 2006 - Ponts en béton - Calcul et dispositions constructives.

Norme NF EN 1992-3 Eurocode 2 - Calcul des structures en béton - Partie 3 : Silos et réservoirs.

Norme EN 13230-1 Applications ferroviaires - Voie - Traverses et supports en béton - Partie 1 : Prescriptions générales.

Norme NF EN 13369 Règles communes pour les produits préfabriqués en béton.

Norme NF EN 13369/A1 Amendement A1 à la norme NF EN 13369.

Norme NF EN 15167-1 Laitier granulé de haut fourneau moulu pour utilisation dans le béton, mortier etcoulis - Partie 1 : Définitions, exigences et critères de conformité.

Norme NF P15-313 Liants hydrauliques - Ciment sursulfaté - Composition, spécification et critères de conformité.

REFERENCES

References

58

Norme NF P15-319 Liants hydrauliques - Ciments pour travaux en eaux à haute teneur en sulfates.

Norme XP P18-420 Béton - Essai d'écaillage des surfaces de béton durci exposées au gel en présence d'une solution saline.

Norme P18-424 Bétons - Essai de gel sur béton durci - Gel dans l'eau - Dégel dans l'eau.

Norme P18-425 Bétons - Essai de gel sur béton durci - Gel dans l'air - Dégel dans l'eau.

Norme NF P 18-508 Additions pour béton hydraulique - Additions calcaires - Spécifications et critères de conformité.

Norme NF P 18-509 Additions pour béton hydraulique - Additions siliceuses - Spécifications et critères de conformité.

Norme XP ENV 13670-1 Exécution des ouvrages en béton - Partie 1 : Tronc commun et document d'application nationale.

Méthode d’essai n° 66 - LCPC, Essai performantiel (à paraître).

Cahier des clauses techniques générales - Fascicule 67 - Titre I - Étanchéité des ponts routes. Support en béton de ciment - Numéro spécial 85-32 bis.

Protection des bétons par application de produits à la surface du parement - LCPC ; SETRA, 2002 - (Guide technique) - Réf. F0231.

Recommandations pour la durabilité des bétons durcis soumis au gel - LCPC, 2003 - (Guide technique) - Réf. RECDUR.

STER 81 : Surfaçage, étanchéité et couches de roulement des tabliers d'ouvrages d'art - SETRA, 1981 - (Guide technique) - Réf. F8210.

STER 81 : Mise à jour n° 2 - Réfection des étanchéités et des couches de roulement des tabliers d'ouvrages d'art. Réparations localisées - SETRA ; LCPC, 2001 - (Guide technique) - Réf. F0112.

Ne pas confondre étanchéité de surface de tablier et protection du béton - Note d'Information - Ouvrages d'Art - Série(OA) n° 25 - SETRA, août 2004 - Réf. 0422w.

59


[1] G. ARLIGUIE, H HORNAIN, GranDuBé, Mesures des grandeurs associées à la durabilité des bétons,Presses de l'ENPC, 2007.

[2] P. TEPPONEN, B. ERIKSSON, Damages en concrete railway sleepers en Finland, Nordic Concrete Research,n° 6, pp. 199-209, 1987.

[3] D. HEINZ, U. LUDWIG, I. RÜDIGER, Delayed ettringite formation in heat treated mortars and concretes,Concrete Precasting Plant and technology, vol. 11, pp. 56-61, 1989.

[4] L. VITOUVA, Concrete Sleepers in CSD tracks, International symposium on precast concrete sleepers,Madrid, pp. 253-264, 1991.

[5] A. SHAYAN, G.-W. QUICK, Microscopic features of cracked and uncracked concrete railway sleepers, ACIMaterials, Vol. 89, n° 4, pp. 348-361, 1992.

[6] R.-E. OBERHOLSTER, H. MAREE, J.-H.-B. BRAND, Cracked prestressed concrete railway sleepers : alcali-silica reaction or delayed ettringite formation, Proceedings of the 9th International conference on alkali-silicareaction in concrete, London, pp. 739-749, 1992.

[7] R.-C. MIELENZ, S. MARUSIN, W.-G. HIME, Z.T. ZUGOVIC, Prestressed concrete railway tie distress : alkali-silica reaction or delayed ettringite formation, Concrete International, Vol. 17, n° 12, pp. 62-68, 1995.

[8] S. SAHU, N. THAULOW, Delayed ettringite formation in swedish concrete railroad ties, Cement and ConcreteResearch, Vol. 34, pp. 1675-1681, 2004.

[9] W.-G. HIME, Delayed ettringite formation - a concern for precast concrete ?, PCI Journal, Vol. 41, pp. 26-30, 1996.

[10] D.-W. HOBBS, Cracking of concrete attributed to delayed ettringite formation, Proceedings of the eleventhannual BCA/concrete society conference on higher education and the concrete industry, UMIST, Manchester,paper 6, pp. 51-60, 2001.

[11] M.-O. OZOL, W. Strand, Delayed ettringite formation at Brewer Stadium Boone, North Carolina, Cementconcrete and aggregates, pp.24-34, 2000.

[12] M. COLLEPARDI, Damage by delayed ettringite formation, Concrete international, pp. 69-74, 1999.

[13] L. LAWRENCE, J.-J. MYERS, R.-L. CARRASQUILLO, Premature concrete deterioration in Texas department oftransportation precast elements, chapter Ettringite - The sometimes host of destruction, American concreteinstitute international, Vol. SP 177-10, Farmington Hills, MI, USA, B. Erlin edition, pp. 140-157, 1999.

[14] L. DIVET, F. GUERRIER, G. LE MESTRE, Existe-t-il un risque d'attaque sulfatique endogène dans les piècesen béton de grande masse : le cas du pont d'Ondes (Haute-Garonne), Bulletin des Laboratoires des Ponts etChaussées, n° 213, pp. 59-72, 1998.

[15] L. DIVET, Les réactions sulfatiques internes au béton : contribution à l'étude des mécanismes de laformation différée de l'ettringite, Études et recherches des Laboratoires des Ponts et Chaussées, OA n° 40,227p, 2001.

BIBLIOGRAPHY

Document publié par le LCPC sous le numéro C1502546

Conception et réalisation LCPC-DISTC, Marie-Christine Pautré

Infographie LCPC-DITSC, Philippe Caquelard

En couverture : - Support of the Allonne cable-stayed bridge (bridge 17) on the Beauvais circle road (RN 31) (Photo LCPC).

Page 7 : - Acicular crystals of ettringite and massive ettringite (Photos LCPC).

Page 13 : - Footings of Normandy bridge (Photos LCPC).

- Construction of the foundations of a pier of the Normandy bridge, in the Seine river (Photos de la Chambre de Commerce et d’Industrie du Havre).

Page 19 : - Millau viaduct. Pier 1, concreting n°24, prepairing the concreting (Photo Eiffage Construction).

Page 25 : - Concreting of a girder of the Aquitaine bridge and cooling circuit with water circulation (Photos Vinci Construction).

ISSN 1151-1516

Réf : GTRSI-E

techniques et méthodesdes laboratoires des ponts et chaussées

Guide technique

Recommendationsfor preventing disorders due

to Delayed Ettringite Formation

The recommendations presented in this document aim at limiting the risk of disorders occurring dueto an internal sulfatic reaction. The latter is caused by the formation of delayed ettringite in acementitious material and occurs in particular because of an important heating of the concreteintervened several hours or several days after its casting. It causes an expansion of concrete whichgenerates in its turn a cracking of the structures. This reaction can be encountered with two types ofconcrete: the heat treated concretes and the concretes cast in place in elements known as critical. These recommendations are concerned with civil engineering structures and buildings comprisingelements of important size that are in contact with water or subjected to a humid environment. Theyfix the level of prevention to be reached according to the category of the structure (or to the part ofstructure) and to the exposure conditions. For each of the four levels of prevention selected,associated precautions are applied and associated checks are carried out. They also presentprovisions related to the design and dimensioning of the structures, the formulation and themanufacture of the concrete as well as to its pouring.

Rec

omm

enda

tions

for p

reve

ntin

g di

sord

ers

due

to D

elay

ed E

ttrin

gite

For

mat

ion

Documents

Recommendations for preventing disorders due to Delayed ...media.lcpc.fr/ext/pdf/sources/gtrsi-e.pdf · ISSN 1151-1516 Réf : GTRSI/E techniques et méthodes des laboratoires des