Lifetime-Oriented Structural Design Concepts- P7 ppsx

For this purpose the coefficients ofthermal expansion α T -value of concrete constituents can become important.However, under normal conditions, in practice, differences in the thermalexpan

Number of cycles

Measuring point 2 Measuring point 1

microc-age model based on the small scale yielding approach of linear elastic fracture mechanics So the propagation of microcracking can be described with stress intensity factors of the cracks near tip field embedded in an isotropic material

with the properties of the macroscopic scale

Related cycle ratio

Fig 3.21 Degradation process of relevant concrete properties due to flexural

load-ings [866]

Fig 3.22 Stiffness reduction by high cycle fatigue

Brittle Damage by Microcracks

As in this context only damaging processes caused by microcracks, whichare triggered by elastic stresses, are regarded no macroscopic plasticity has to

be considered Imagine such a member with growing microcracks undergoing

a process, in which it is deformed by a total deformation, a certain part of thiswill be elastically recoverable, and another part can be induced by damage.When these loads are released, the member will have, in contrast to plastic-ity, not any remaining permanent deformation Nevertheless, the state of themember could have changed; its elastic stiffness could have been reduced bythe growth of microcracks For a process which involves no further damag-ing, the total deformation is an elastic one, but starting from a state with

Fig 3.23 Model for brittle damage by microcrack growth

changed elastic properties The underlying micromechanics for a continuumpoint and the corresponding macro-stresses and strains are sketched in Fig-ure 3.22 Starting with an unstressed member, containing a crack of length

2a and the resulting average stiffness E (the stiffness of the matrix remains

unchanged by crack growth), up to a certain load the crack will not grow inlength but only open its width Beyond this threshold the crack length willincrease and the average stiffness decreases When the member is unloadedagain the crack will close and no further growth occurs For the same stress agreater strain will result, due to the reduced stiffness In the stress free statethere is - as already mentioned - no permanent deformation Only the stiffness

remains on a lower level than in the initial stage The free energy W stored at the end of the process and the energy dissipated by crack growth Wcrack arealso sketched in Figure 3.22 This means the process of stiffness degradationcan be modelled by finding a correct representation for the energy dissipated

by crack growth This will be the basis for the continuum damage model forhigh-cycle fatigue of metals presented in Section 3.3.1.2.2.2

3.1.2 Non-mechanical Loading

3.1.2.1 Thermal Loading

3.1.2.1.1 Degradation of Concrete Due to Thermal Incompatibility of Its

Components

If the thermal behaviour and the thermal properties of the various concreteconstituents are quite different from each another, in cases of significant

temperature changes, incompatibilities in the deformations of the differentmaterials cause internal stresses between the aggregates and the cement paste,wich further can result in internal cracking For this purpose the coefficients of

thermal expansion (α T -value) of concrete constituents can become important.However, under normal conditions, in practice, differences in the thermalexpansion coefficient are not necessarily deleterious when the temperature

two relevant α T-values (aggregates, cement paste) differ seriously (much more

than 5.5 · 10 −6 K −1) from each another the durability of concrete concerning

freezing and thawing may be affected [567]

3.1.2.1.2 Stresses Due to Thermal Loading

Much more important for microcracking and degradation processes in concretestructures are restraint stresses, caused by restraining of thermal deformations

( T = α T · Δt) Such restraint can be external as well as internal.

In most cases temperature profiles over a cross-section are not constant orlinear, but more or less stochastic and non-linear (Figure 3.24) Thus, the re-sulting stresses can be divided into longitudinal, warping and internal stresses.For longitudinal stresses with constant magnitude an external restraint andconstant temperature changes over the cross-section of concrete are respon-sible Such an external restraint is caused in practice e.g by bond to a stifffoundation or an already hardened concrete member A linear distributed tem-perature gradient results in warping stresses, since the bending deformationsusually are restrained already by the deadload or also by external restraint.Internal stresses are formed by the restraint of non-linear thermal deforma-tions In this case the restraint is internal, as the cross-section cannot deformunevenly (Bernoulli-hypothesis)

In context with restraint stresses due to restrained thermal deformationsespecially in thicker concrete members often already the load-case ”heat ofhydration” becomes relevant

Fig 3.24 Stresses in a concrete slab at one-sided, non-linear cooling from the top

[145]

3.1.2.1.3 Temperature and Stress Development in Concrete at the Early Age

Due to Heat of Hydration

When concrete has been placed, initially the temperature remains unchanged,because the hydration process is still in its rest period (stage I) (Figure 3.25)

A few hours after starting of hydration also a moderate temperature raise can

be observed, however (also in case of restraint) without developing significantcompressive stresses At this stage II the concrete has not yet set and is there-fore still plastically deformable Along with further hydration the stiffness ofthe concrete increases and may lead – if the deformations are restrained – tocompressive stresses (stage III) The concrete temperature at the beginning

of this third stage is called the first zero-stress temperature (1 T z) However,also in this stage the relaxation of the young concrete is still high, so that

in spite of a significant temperature rise only small compressive stresses areraised In the consequence of this high relaxation at the end of stage III themaximum of the compressive stresses is obtained in general some time beforethe temperature maximum After exceeding the temperature maximum theremaining compressive stresses decrease rapidly (stage IV) Only a few degrees

below the temperature maximum the second zero-stress temperature (2 T z) isobtained Already starting from this point tensile stresses are caused during

T air = T fresh concrete

Fig 3.25 Temperature and stress development during the first hydration phase in

restrained concrete elements [763, 145, 466]

further cooling (stage V) When the not yet significantly developed tensilestrength is exceeded in this cooling period, at an age of only a few days, first

cracks will be formed at the so-called cracking-temperature (T crack) [145].Especially in mass concrete structures the internal restraint and thus theresulting internal stresses can become a dominant cause for thermal cracking

If the heat of hydration is not controlled and large temperature differencesbetween the inner core and the surface are raised, internal stresses with tension

at the surface develop in the concrete member Thus, a surface map-cracking

in the surface-zone can occur, whereby the crack-width usually is very small

It is evident, that the described cracking also at such thermal loadings doesn’tdevelop suddenly Furthermore it has to be considered, that also in such cases acomplex micro-cracking is preceding the macro-cracking formation Thus also

by this way degradation processes can take place, even if the tensile strength

is not exceeded, i.e when the ambient temperature is achieved before cracks could be formed In this case the concrete structure remains on a hightensile stress level and micro cracks (with resulting degradation) develop

macro-3.1.2.2 Thermo-Hygral Loading

3.1.2.2.1 Hygral Behaviour of Hardened Cement Paste

Authored by Max J Setzer and Christian Duckheim

Due to its nano- and microporous structure hardened cement paste acts strongly with its environmental humidity This gain or loss of water has

inter-a deep impinter-act on durinter-ability inter-and minter-ateriinter-al properties below 0◦C (e.g frost) as

well as above 0◦C (e.g creep and shrinkage) [633] Even if further research is

required, freeze-thaw-resistance of concrete structures and the correspondingmechanisms have been investigated extensively in the last years and can be ex-plained well today In contrast, despite numerous different analyses creep- andshrinkage-mechanisms are only fragmentarily understood up to now Amongstothers, this fact can be attributed to the manifold parameters which influenceexperimental results (such as sample composition and shape or the measur-ing setup and procedure) but most of all to the complex colloidal structureformed by nano-sized CSH-particles, where only complicated ascertainablesurface interactions play a decisive role Drying shrinkage and swelling as a

basic hygric property of hardened cement paste (w/c = 0, 35; 0, 40; 0, 50 and

0, 60) has been investigated over the complete humidity range by means of

a newly developed laser supported measuring principle This new techniqueallows the speedy, precise measurement of the pure material characteristic

of several filigree samples with an accuracy of about 20 nm Further mainlynovel methods have been applied for examining sorption behaviour as well asinner volume and density change Measurement data have been analysed with

Fig 3.26 Hygric strains vs relative humidity

Fig 3.27 Hygric strains vs relative humidity & vs water content

respect of the prevailing mechanisms on the nano- and microscale as surfaceenergy, disjoining pressure and capillary tension

In Figure 3.26 hygric strains of four samples with different w/c-ratios

during first de- and adsorption are illustrated Figure 3.27 shows the relation

between the measured deformations (w/c = 0, 40) and water content of

-9 -8 -7 -6 -5 -4

Fig 3.28 Hygric strains vs surface free energy change For further details

(calcu-lation of surface energy and deformations due to capillary tension) see [239]

the structure including a second desorption-adsorption-cycle The total

deformation, which grows with increasing w/c-ratio, lies in between 7 mm/m and 9 mm/m Examining the results, in the range from 0 % r h to 100 %

found A close connection between water content of the structure and studiedproperties is demonstrated with only a marginal hysteresis between dryingand wetting as well as the influence of capillary condensation It could beproved that in the lower humidity range shrinkage and swelling are indeedproportional to changes in the surface free energy indeed (Figures 3.28

to 3.31) However, an energy reduction during adsorption does not lead

to an expansion as assumed up to now (Munich Model), but rather to acontraction of csh-particles (Figure 3.31), while the pore volume increasessimultaneously and vice versa during desorption Solid density which is

attributed to the dispersive component of disjoining pressure which prevails

in the lower humidity range, whereas in the range of condensation repulsivcomponents (electrostatic and structural component) and capillary tensiondominate the processes in hardened cement paste Consequently here adistinct linear relation-ship exists between hygric strains and water content(Figure 3.27) Irreversible strains have to be attributed merely to firstdrying

-5 -4 -3 -2 -1

0

relative humidity (%)

w/c = 0,35 (calculated) w/c = 0,35

w/c = 0,40 (calculated) w/c = 0,40

w/c = 0,50 (calculated) w/c = 0,50

w/c = 0,60 (calculated) w/c = 0,60

Fig 3.29 Hygric strains vs surface free energy change & comparison between

mea-sured hygric strains and hygric strains calculated by capillary tension For furtherdetails (calculation of surface energy and deformations due to capillary tension) see[239]

Fig 3.31 Solid density vs relative humidity

The presented findings and additional results are merged in a schematicdiagram (Figure 3.32) which describes the change of various hygric proper-ties qualitatively and illustrates the effects of the two different mechanisms(disjoining pressure and capillary tension) on the system of hardened cementpaste during first desorption and adsorption Elaborate explanations and fur-ther details can be found in [239]

3.1.2.2.2 Influence of Cracks on the Moisture Transport

Cracks, irrespective of their origin, have a considerable influence on themoisture permeability of cementitious materials As a consequence, the trans-port of aggressive substances is promoted and the degradation process isfurther accelerated The significant influence of fracture on the transport prop-erties of porous materials was first recognized in the context of the coupledmechanical and hydraulic behavior of fractured rock masses Experiments byZoback & Byerlee [874] indicate an increase of the permeability of granitecaused by microcracking Particularly for materials with very low moisturepermeabilities, such as granite and shale, flow through the connected porespace was found to be insignificant compared to flow through fracture zones.The role of cracks on the transport properties of cement-based materials hasbeen investigated in e.g [92, 155, 309, 41, 310], see Breysse and G´erard [153]for a state-of-the-art survey It has been shown, that the problems of moisturetransfer change the scale, in fact that the permeability is increased by severalorders of magnitude, when cracking is considered

Fig 3.32 Schematic diagram of hygric mechanisms and properties of hardened

cement paste

3.1.2.2.3 Freeze Thaw

Authored by Max J Setzer and Ivanka Bevanda

Deterioration under freeze thaw attack has been discussed in literatureunder various aspects The most recent development deals with the aspects

of fracture mechanics But for understanding the damage mechanism of frostand frost deicing salt attack, it is important to understand the following:

(1) freeze thaw cycles (with and without de-icing salt) acted as a micro pump and (2) a distinction between the forgoing transport process and the

following damage process and final degradation is essential The unusualfreezing behaviour of the pore solution in cement paste i.e the special poresystem is responsible that water and solution will be sucked up during freezethaw cycles [723],[725] This phenomena: (1) frost suction and (2) followingdeterioration after reaching the critical degree of saturation explained bySetzers modul of the micro-ice-lens [725],[727],[728],[730] This model, based

on the fact that within the concrete matrix water, vapour and ice cancoexist in a wide temperature range, as concrete deviates from macroscopicbehaviour due to its nanostructure This leads to shrinkage of the gel during

I: Cooling

pressure due to triple phase condition expansion / contraction

liquid water flow internal vapor transport

ice matrix

bulk- water Vapor

gel-II: Heating

approx 150 nm approx 150 mm

'p T

x

0 x

Freezing zone

External heat Compression zone

only water - vapor

Melting zone

0 0

External water

External heat

Expansion zone 0

external water

Triple phase

Fig 3.33 Comparison of macroscopic and microscopic situation of the

micro-ice-lens model during the heating and cooling phase [731] Because of a pressure ence between the unfrozen gel water and ice in larger pores, water transport occurs,when freezing starts Water from the micro pores is transported to existing ice incapillary pores Simultaneously shrinkage of smaller pores can be registered Thewater, i.e ice, content of the macro pores increases During thawing the gel tries

differ-to expand be again sucking water, which is available not from the still frozen ice

in larger pores but from external sources Independently with the 9% expansion ofvolume of ice, frost damage occurs in completely filled pores

freezing and to transport process within the pore system during ing (see Figure 3.33)

melt-Powers[644] and Fagerlund [270] particularly describe models of retardedice formation which leads to hydraulic pressure Powers and Helmunt werethe first who discussed the problems arising from transport phenomena, os-motic and hydraulic pressure [644],[645],[366] Fagerlund [269] refined thisand stated the distinction between the critical degree and when damage started.Some models discussed the submicroscopic stress which is generated by sur-face interaction and by curved surfaces Everentt and Haynes [266] describedthis phenomena as an ice crystal which is successively penetrating a micro pore

pro-cess Litvan [503] explained the transport from unfrozen pore water to icedue to diffusion The impact of surface forces on frost damage was taken into

account by the thermodymamic model developed by Setzer [723] One group ofdamage models was explained by macroscopically generated stress Podval-

to temperature and salt concentration gradients, R¨osliand Harnik [689] thetemperature and stress gradient due to sudden melting of ice by de-icing agents.Besides this, physical models a chemical model for the damage mechanism un-der frost de-icing salt attack is described in [509],[508] Ludwig has shown apreferred formation of ettringite under low temperatures

It should be noted that it was found in the SFB 398/ TP A11 thatsmall amounts of dissolved ions increase the surface scaling dramatically (seeSubsection 2.4.2, [120]) This can neither be explained by macroscopic orsemi-macroscopic physics - concentrations are much too small - nor by purechemical effects - the phenomena reach a maximum between 0.2 mol/l and 0.5mol/l Similar to chromatographic effects during transport both must proba-bly be attributed to superimposed effects of surface physics

3.1.2.3 Chemical Loading

3.1.2.3.1 Microstructure of Cementitious Materials

Concrete is a nano-porous multi-component system composed of aggregatesand cement matrix The cement matrix is consisting of a heterogeneous sys-tem of non-hydrated cement, hydration products, pores and pore solutions

It should be noted, that during the hydration process of cement both largecalcium hydroxide crystals and CSH-gel are generated simultaneously result-ing in a pore structure characterised by a pore size distribution ranging fromcapillary pores to gel pores (Figure 3.34) Different mechanical, physical andchemical processes which may considerably affect the durablity of the materialare caused and controlled by the pore size distribution, the fluid saturation,the mechanically and chemically induced changes in the porosity and thechemical composition of the pore fluid and the matrix

Durability of concrete is highly affected by the transport of moistureand ionic (corrosive) species eventually leading to damage processes caused

by chemically expansive reactions as well as by dissolution of load bearingconstituents

The pore size distribution of hardened cement paste covers a large spectrum

of pores extending over 7 orders of magnitude, see Figure 3.34 The smallestpores are smaller than one nanometre and the sizes of voids due to non-completedcompaction might exceed some millimetres Furthermore, at least two differentkinds of pores have to be distinguished: The gel pores resulting from the cementhydration within these CSH-products and the larger capillary pores in thecement-paste between the original cement particles The different pores also may

Fig 3.34 Volume fractions of constituents of hardened cement paste as a function

of the water cement ratio [448]

be distinguished according to their behaviour and/or to the necessary time to tain a specific capillary pore system [724] This yields a classification into coarsepores, capillary pores, meso- and micro-gel pores, see Table 3.1

ob-Table 3.1 Classification of pore sizes in concrete according to [724]

freez-able, highly mobile, small illary rise

cap-

Meso-capillary

< 30μm sucking within minutes,

refill-able within weeks

free macroscopic water, able, considerable capillary risewithin a few days

freez-

Micro-capillary

strong capillary attraction, butincreasing internal frictionMeso-

gel

behaviour to surface physics;

filled by condensation at 50 % to

98 % rel humidity

pre-structured, condensed ter, evaporation below 50 % rel-ative humidity, not freezable be-yond−23 ◦C

Table 3.2 Influences on the degree of chemical attack

Acid attack increases with

- increase in acid concentration and decrease in pH-value

- constant and fast renewal of acidic solution at the concrete/liquid interface

- higher temperatures

- higher pressure

Environmentally induced deterioration of cementitious materials and sequently the lifetime of concrete structures are to a large extent controlled bytransport processes within the pore system In particular, the accumulation ofenvironmentally induced deterioration processes such as dissolution processes(e.g calcium leaching), chemical expansive reactions (e.g sulfate attack) orthe transport of chlorides, which interacts with damage caused by time variantexternal loading, may limit the durability of concrete and reinforced concretestructures

con-Main concrete constituents subjected to aggressive substances also may bechemically damaged by calcium leaching which is controlled by the dissolutionand de-calcification of different cement phases and the diffusion of dissolvedspecies through the pore system

The physical and chemical processes strongly interact with mechanical formations and degradations of concrete structures, such as microcracks andhence the increase of the pore spaces due to the additional mechanical load.Considerable progress was achieved in material-oriented research on envi-ronmentally induced degradation mechanisms, which led to a better under-standing of the microstructural mechanisms and to analysis tools to simulatethe relevant processes In Subsections 3.1.2.3.2, 3.1.2.3.3, 3.1.2.2.2 the mainexperimental findings associated with long-term degradation processes caused

de-by environmental loading and their interactions with external loading are marized

sum-3.1.2.3.2 Dissolution

When concrete structures remain in continuous contact with acidic fluids,exchangeable salt solutions or softened water with a low content of alka-line earth ions (e.g Ca++) chemical dissolution processes lead to a contin-uous deterioration of the material The degree of such dissolutions depends

on chemical conditions of the fluid as well as on environmental conditions(Table 3.2)

It should be noted, that the chemical dissolution strongly can interact withmechanically induced micro- and macrocracks caused, e.g by external loading(see Figure 3.35) This may considerably affect the long term serviceablilty andthe integrity of concrete structures Cooling towers, containments for nuclear

or other waste disposal, cement-bound coatings of drinking water reservoirs,grouted anchors and tunnel linings are examples for structures and structural