CZECH TECHNICAL UNIVERSITY IN PRAGUE

Faculty of Civil Engineering

Doctoral Degree Program: CIVIL ENGINEERING

Branch of study: Building and Structural Engineering
Department of Concrete and Masonry Structures

Ing. Jakub Goringer

COUPLED DETERIORATION PROCESSES IN CONCRETE

-
EXPOSURE TO ACIDIC ENVIRONMENT, CYCLIC LOADING AND
IMPACT LOADING

KOMBINOVANA DEGRADACE BETONU

-
Vliv agresivnho prostred, cyklickeho zatzen a razoveho zatzen

DISERTATION THESIS

Supervisor: Doc. Ing. Marek Foglar, Ph.D.

Prague, August 2016

 Statement of originality

Statement of originality

I declare that I have accomplished this final thesis independently and that I have included all cited
literature and used sources of information. Neither the work as the whole nor any its substantial
part has been previously submitted to achieve a different or similar academic degree.

Prague 13 August 2016

Signed: ..
Jakub Gringer

iii
 Acknowledgements

Acknowledgements

I would like to first say a big thank you to my supervisor doc. Ing. Marek Foglar, Ph.D. for all the
support and encouragement he gave me during the last five years. Without his guidance and
constant feedback this PhD thesis would not have been written and the research behind it would
not have been possible.
Many thanks to all members of the Department of concrete and masonry structures for the kind
working atmosphere, their support, encouragement and many professional discussions.
My deep appreciation goes out to my colleagues Pavel Jiek, Radek Hjek and Martin Kov for
their help and assistance with the experimental program, their time to discuss the problems which
this thesis brought and discuss the problems of life, the universe and everything.
Also I must acknowledge the funding, without which this work would not have materialised.
Grants from the Ministry of the Interior of the Czech Republic (VG20132015114), from the Czech
Science Foundation (13-30441S) and Czech Technical University (SGS12/029/OHK1/1T/11,
SGS13/035/OHK1/1T/11, SGS14/031/OHK1/1T/11 and SGS15/033/OHK1/1T/11) were used in
part to support the PhD project.
Finally, I would like to thank my family and friends for their love, patience and support during
this work.

iv
 Abstract

Abstract
Durability of the structures is one of the most discussed issues of last decades. Modern high-
performance concretes enable to design extremely slender structures which are more vulnerable to
the deterioration effects due to their minimal dimensions. In case of the transport infrastructure at
least two deterioration mechanisms are involved; chemical - the material is exposed e.g. to acid
solution which reacts with concrete components and successively weakens composition of the
binder and mechanical the material is exposed to e.g. cyclic loading which leads to the formation
and further propagation of cracks in the material matrix.
The state of the art in this thesis provided information that the problems of the combined
deterioration processes are not so thoroughly described. Therefore this work is focused on the
problem of coupled deterioration processes in concrete, particularly acidic chemical deterioration,
cyclic loading and extreme type of loading impact.
The thesis presents an analytical tool for the evaluation of the chemical deterioration degree of
concrete elements based on the chemical kinetics. This model was further combined with the
assessment of the deflection increase due to the cyclic loading and in the estimation of the beam
energy dissipation during the impact loading.
The analytical model and its extensions were verified with the data obtained from the long-term
experimental program which involved many testing methods (Atomic Absorption Spectrometry,
Scanning electron microscopy, Energy-dispersive X-ray spectroscopy, Photon Doppler
velocimetry etc.).

Keywords
Concrete, deterioration, acid, aggressive environment, chemical kinetics, SEM, EDS, cyclic
loading, impact loading

v
 Abstrakt

Abstrakt
Trvanlivost stavebnch konstrukc je jednm z mnoha velmi diskutovanch tmat s ohledem na
souasnou filozofii trvale udritelnho rozvoje. Souasn vysokohodnotn betony umouj nvrh
vysoce thlch konstrukc, jejich ivotnost a pouitelnost je z dvodu minimalizovanch
rozmr v konenm dsledku vce ovlivnna degradanmi vlivy. V ppad staveb dopravn
infrastruktury se tchto pokozujcch vliv uplatn hned nkolik. Nejvznamnjmi jsou
chemick degradace vlivem agresivnho prosted, kdy je postupn rozruovna celkov integrita
materilu a mechanick degradace, napklad cyklick zaten, kter me zpsobovat tvorbu
a nsledn rozvoj trhlin.
Zmnn jevy jsou v odborn literatue jednotliv dobe popsny, bohuel informace o jejich
kombinovanm vlivu ji tak rozshl nejsou. Z tohoto dvodu je tato prce zamena na problm
kombinovan degradace betonu vlivem chemicky agresivnho prosted, cyklickho zaten
a nsledn i vlivu extrmnho typu zaten - rzem.
Disertan prce pedstavuje analytick nstroj pro vyhodnocen stupn degradace betonu za
vyuit kinetiky chemickch reakc. Tento model je dle kombinovn s nslednm posouzenm
vlivu nrstu deformac vlivem cyklickho zaten a odhadu mnostv energie disipovan
v prhybech v rmci zaten rzem.
Navren analytick model byl verifikovn na datech zskanch z dlouhodobho
experimentlnho programu, pro jeho vyhodnocen bylo pouito mnoho modernch metod,
napklad Atomov absorpn spektrometrie, skenovac elektronov mikroskopie, rentgenov
spektroskopie a fotonov Dopplerova velocimetrie.

Klov slova
Beton, degradace, agresivn prosted, chemick kinetika, SEM, EDS, cyklick zaten, rzov
zaten

vi
 Table of Contents

TABLE OF CONTENTS
1 Introduction .................................................................................................................. 1
2 A State of the art........................................................................................................... 4

2.1 Chemical deterioration of concrete ....................................................................... 4

2.2 Influence of cyclic loading on deflection development ........................................ 8
2.3 Impact loading .................................................................................................... 10

3 Coupled deterioration processes ............................................................................... 12

3.1 Mathematical description of concrete acidic deterioration ................................. 12

3.1.1 Basic chemical prerequisites ................................................................... 12

3.1.2 Adapting the reaction kinetics to determine the chemical deterioration . 13
3.1.3 Deterioration of the concrete specimens ................................................. 14

3.2 Extension of the strain development under cyclic loading calculation .............. 18

3.2.1 Extension of the deflection increase calculation process ........................ 19

3.2.2 Residual life-time assessment ................................................................. 22

3.3 Impact loading .................................................................................................... 24

3.3.1 Dynamic increase factor (DIF) ............................................................... 24

3.3.2 FEM model of the experiment ................................................................ 25
3.3.3 Energy balance of the system.................................................................. 28

4 Experimental verification .......................................................................................... 32

4.1 Experimental program ........................................................................................ 32

4.1.1 Sets of specimens .................................................................................... 32

4.1.2 Exposure class of the environments ........................................................ 33
4.1.3 Maintaining a stable pH value ................................................................ 35
4.1.4 Layout of the fatigue testing ................................................................... 37
4.1.5 The material model point ........................................................................ 38
4.1.6 Deflection measurements ........................................................................ 38
4.1.7 The impact loading.................................................................................. 39

4.2 Chemical deterioration ........................................................................................ 41

4.2.1 Time change of the pH of the aggressive solution .................................. 41

4.2.2 Integration method verification ............................................................... 42
4.2.3 Evaluating the rate constant .................................................................... 43
vii
 Table of Contents

4.2.4 Weighted average of the pH of the aggressive environment .................. 45

4.2.5 Verifying the rate equation ..................................................................... 45
4.2.6 Reduced mechanical properties of deteriorated concrete specimens ...... 46
4.2.7 Reduced cross-sectional area of specimen core ...................................... 48

4.3 SEM and EDS specimen analysis ....................................................................... 50

4.3.1 Sample selection and preparation ........................................................... 50

4.3.2 Observational method (SEM) ................................................................. 53
4.3.3 Analytical method (EDS) ........................................................................ 55

4.4 Cyclic loading ..................................................................................................... 59

4.4.1 Sensitivity analysis of the number of layers ........................................... 60

4.4.2 Fatigue testing of reference C sets....................................................... 60
4.4.3 Fatigue testing of deteriorated A sets .................................................. 63
4.4.4 Influence of the strain rate ...................................................................... 67
4.4.5 Evaluation of the remaining fatigue life-time ......................................... 70

4.5 Impact loading .................................................................................................... 72

4.5.1 Procedures for the measured data evaluation .......................................... 72

4.5.2 Fracture energy fitting in concrete material - MAT159.......................... 74
4.5.3 Experimental and FEM results of the impact testing .............................. 75
4.5.4 Energy estimation with deterioration assessment methods..................... 84

5 Discussion .................................................................................................................... 87

5.1 Chemical deterioration ........................................................................................ 87

5.2 Cyclic loading ..................................................................................................... 88
5.3 Impact loading .................................................................................................... 89

6 Conclusions ................................................................................................................. 90
7 References ................................................................................................................... 93
8 Publications ............................................................................................................... 100

8.1 WoS and SCOPUS indexed publications ......................................................... 100

8.2 Other conference and journal publications ....................................................... 100
8.3 Utility models ................................................................................................... 102

viii
 List of Figures

LIST OF FIGURES
Fig. 1 Principle of the proposed superposition of the damage components ................................ 2
Fig. 2 Bridge pier exposed to different deterioration processes .................................................. 4
Fig. 3 Generalized scheme of the acidic deterioration of concrete ............................................ 13
Fig. 4 Comparison of the pH distribution in direction of diff. according to Ficks law and .... 17
Fig. 5 Principle of the reduced cross-sectional area method ..................................................... 17
Fig. 6 Development of the secant modulus of elasticity of concrete under cyclic loading ....... 18
Fig. 7 Principle of partial integration of the compressive zone for n = 400,000 cycles ............ 21
Fig. 8 Amplitudes development and exponential parts development for Smax = 0. ................... 23
Fig. 9 Modulus of elasticity fatigue age development according to equation (3.2.5) ............... 23
Fig. 10 Strain rates for different loading velocities ................................................................... 24
Fig. 11 DIF curves for different mechanical properties according to Model Code 2010 .......... 25
Fig. 12 LS-DYNA model render ............................................................................................... 26
Fig. 13 External work done by applied force with force-deflection diagram ............................ 29
Fig. 14 Hourglass modes in the x-direction for a prismatic hexahedron ................................... 30
Fig. 15 Reinforcement scheme of the specimens ...................................................................... 33
Fig. 16 Set of specimens stored in the aggressive environment ................................................ 34
Fig. 17 Solution pH calculation based on the concentrations .................................................... 35
Fig. 18 Solution pH calculation based on the ionic strength ..................................................... 36
Fig. 19 The arrangement of the fatigue testing .......................................................................... 37
Fig. 20 Four-point bending arrangement and localization zone principle ................................. 38
Fig. 21 Steel frame and the scheme of experiment .................................................................... 39
Fig. 22 Experimental layout for the impact loading with the use of PDV ................................ 40
Fig. 23 Measured and approximated pH development in time ............................................... 41
Fig. 24 In time development of the integrated form of rate equation ........................................ 42
Fig. 25 Time development of compressive strength (left) and elastic modulus (right) ............. 46
Fig. 26 Calculation of the reduced compressive strength for small beam A8 ........................... 48
Fig. 27 Calculation of the reduced cross-sectional area for small beam A8 ............................. 50
Fig. 28 Zones of the specimen for SEM and EDS analysis, specimen cross-section, S-zone ... 51
Fig. 29 - Specimens prepared for the SEM ................................................................................... 52
Fig. 30 Selected locations of SEM and EDS analyses for all tested specimens ........................ 53
Fig. 31 SEM images of deteriorated layer, transition zone and sound concrete........................ 54
Fig. 32 Typical EDS spectra for the brown-coloured layer and sound concrete ....................... 55
Fig. 33 Atomic composition of the deteriorated concrete with depth dependence.................... 56

ix
 List of Figures

Fig. 34 Atomic percent of Ca2+ ions along the specimen depth ................................................ 57
Fig. 35 Atomic percent of Si4+ ions along the specimen depth (range as in Fig. 34) ................ 57
Fig. 36 Atomic percent of Al3+ ions along the specimen depth (range as in Fig. 34)................ 57
Fig. 37 Atomic percent of Cl- ions along the specimen depth (range as in Fig. 34).................. 58
Fig. 38 Atomic percent of Fe3+ ions along the specimen depth (range as in Fig. 34) ............... 58
Fig. 39 Comparison of measured and calculated deflections for set C1.................................... 60
Fig. 40 Comparison of measured and calculated deflections for set C2.................................... 61
Fig. 41 Comparison of measured and calculated deflections for set C3.................................... 61
Fig. 42 Cyclic beam C4 after collapse during the fatigue testing .............................................. 62
Fig. 43 Shear crack after first static loading of specimen C5 .................................................... 62
Fig. 44 Comparison of measured and calculated deflections for set A1 ................................... 63
Fig. 45 Comparison of measured and calculated deflections for set A2 ................................... 64
Fig. 46 Comparison of measured and calculated deflections for set A3 ................................... 64
Fig. 47 Comparison of measured and calculated deflections for set A4 ................................... 65
Fig. 48 Comparison of measured and calculated deflections for set A5 .................................. 65
Fig. 49 Comparison of measured and calculated deflections for set A6 ................................... 66
Fig. 50 Comparison of measured and calculated deflections for set A7 ................................... 66
Fig. 51 Comparison of measured and calculated deflections for set A8 ................................... 67
Fig. 52 Static and dynamic deflections with the calculated trend curves for specimen A3 ...... 69
Fig. 53 Static and dynamic deflections with the calculated trend curves for specimen A5 ...... 70
Fig. 54 Static and dynamic deflections with the calculated trend curves for specimen A7 ...... 70
Fig. 55 Statistical evaluated velocity of the burden impact testing of specimen C1 .............. 74
Fig. 56 Develop. of burden speed / deflection and force-deflection curve / area of spec. R1 ... 76
Fig. 57 Condition of the specimen R1 after impact loading ...................................................... 76
Fig. 58 Develop. of burden speed / deflection and force-deflection curve / area of spec. C1 ... 76
Fig. 59 Condition of the specimen C1 after impact loading ...................................................... 77
Fig. 60 Develop. of burden speed / deflection and force-deflection curve / area of spec. A1 ... 77
Fig. 61 Condition of the specimen A1 after impact loading ...................................................... 77
Fig. 62 Develop. of burden speed / deflection and force-deflection curve / area of spec. A2 ... 78
Fig. 63 Condition of the specimen A2 after impact loading...................................................... 78
Fig. 64 Longitudinal crack in specimen A2 due to the impact loading exp. vs. LS-DYNA .. 79
Fig. 65 Develop. of burden speed / deflection and force-deflection curve / area of spec. A3 ... 79
Fig. 66 Condition of the specimen A3 after impact loading...................................................... 80
Fig. 67 Develop. of burden speed / deflection and force-deflection curve / area of spec. A4 ... 80
Fig. 68 Condition of the specimen A4 after impact loading...................................................... 80
x
 List of Figures

Fig. 69 Develop. of burden speed / deflection and force-deflection curve / area of spec. A5 ... 81
Fig. 70 Condition of the specimen A5 after impact loading...................................................... 81
Fig. 71 Develop. of burden speed / deflection and force-deflection curve / area of spec. A6 ... 81
Fig. 72 Condition of the specimen A6 after impact loading...................................................... 82
Fig. 73 Develop. of burden speed / deflection and force-deflection curve / area of spec. A8 ... 82
Fig. 74 Condition of the specimen A8 after impact loading...................................................... 82
Fig. 75 Area under force-deflection curve according to reduced compressive strength ........... 85
Fig. 76 Area under force-deflection curve according to reduced cross-sectional area .............. 86
Fig. 77 Principle of the proposed approaches implemen. to different design states analysis .... 90

xi
 List of Tables

LIST OF TABLES
Tab. 1 DIF values for the impact and cyclic loading according to Model Code 2010 .............. 25
Tab. 2 Summary of the sets used in the experimental program ................................................. 32
Tab. 3 Geometry properties of the tested specimens ................................................................. 33
Tab. 4 Mix proportions of the concrete mixture ........................................................................ 33
Tab. 5 Exposure classes for a chemical attack according to EN 206 ........................................ 34
Tab. 6 A summary of calculated and lab. obtained conc. of Ca2+, Al3+, Fe3+, Si4+ ions ........... 37
Tab. 7 Specification of the fatigue testing set-up ...................................................................... 38
Tab. 8 Verification of the principle stated in equation (4.2.4)................................................... 44
Tab. 9 Excerpt from the pH measurement report ...................................................................... 45
Tab. 10 Equivalent pH values of the aggressive environment .................................................. 45
Tab. 11 A comparison of the calc. mass and the lab. obtained mass of dissolved Ca2+ ions .... 46
Tab. 12 Comparison of the calculated and measured reduced compressive strength values .... 47
Tab. 13 Comparison of the reference and calc. comp. strength of sound specimen cores ........ 49
Tab. 14 Specification of the SEM specimens ............................................................................ 51
Tab. 15 Calculated comp. strength and reduced cross-section dimensions for cyclic beams ... 59
Tab. 16 Comparison of static and dynamic deflection measurements ...................................... 68
Tab. 17 Age of specimens with regards to the deflection increase, compared with MC 10 ..... 71
Tab. 18 Fracture energies used in LS-DYNA model................................................................. 75
Tab. 19 Summary of the impact testing results ......................................................................... 83

xii
 List of Symbols

LIST OF SYMBOLS
Latin upper case letters
Acrack crack area [m2]
Areduced reduced cross-sectional area [m2]
Ai amplitude (i = 1 to 2) [-]
D diffusion coefficient [m2/s]
Ec,imp modulus of elasticity in impact loading [GPa]
Ec,init initial modulus of elasticity [GPa]
Ec,ni modulus of elasticity after ni cycles [GPa]
Edamp energy of damping system [J]
Ehg hourglass energy [J]
Eint internal energy [J]
Esi energy of sliding interface [J]
Ekin kinetic energy [J]
Eres residual modulus of elasticity [J]
Etotal total model energy [J]
Fc cyclic force [N]
Fburden impact force [N]
G strain energy release rate [J/m2]
Gf fracture energy [N/m]
Gf0 base value of fracture energy (aggregate size dependent) [N/m]
Gfc uniaxial compressive fracture energy [N/m]
Gft uniaxial tensile fracture energy [N/m]
Gfs pure shear fracture energy [N/m]
M() molecular mass of the substance in brackets [g/mol]
N limit number of applied cycles [-]
S surface area exposed to an aggressive environment [m2]
Smax maximum stress level [-]
V volume [m3]
Wext external work [J]
Wimp work done on deflections during the impact loading [J]
T temperature [C]
XCaO volume fraction of calcium oxide in the cement [-]

xiii
 List of Symbols

Latin lower case letters

a variable dependent on Smax (fatigue) [-]
a crack length (impact) [m]
b variable dependent on Smax (fatigue) [-]
b crack width (impact) [m]
c concentration in general [mol/dm3]
ci constants in fatigue damage function (i = 1 to 4) [-]
cX concentration of individual substances [mol/dm3]
cX,0 initial concentration of individual substance [mol/dm3]
cCa,0 initial calcium ion concentration in the concrete mixture [mol/dm3]
erf error function [-]
fc,imp,k characteristic compressive strength under high rates of loading [MPa]
fct,imp,k characteristic tensile strength under high rates of loading [MPa]
fc compressive strength of concrete [MPa]
fcm resistance of cement in 28 days (chemical deterioration) [MPa]
fcm mean value of compressive strength of concrete (impact) [MPa]
fctm mean value of tensile strength of concrete [MPa]
g gravitational acceleration [m/s2]
kc rate constant [(mol/dm3)(1-n)/s]
kc pseudo-rate constant [dm3/(mol.s)]
kf Fret coefficient (time dependent)
m mass in general [g]
mc mass of cement [g]
n reaction order (chemical deterioration) [-]
ni amount of an substance i (chemical deterioration) [mol]
ni number of applied loading cycles (fatigue) [-]
1/r curvature [1/m]
t time [s]
uburden burden displacements [m]
uburden burden velocity [m/s]
uburden burden acceleration [m/s2]
vimpact impact velocity [m/s]
vi the stoichiometric coefficient of a substance i (chemical equations) [-]
va air volume fraction in the mix design (Fret equation) [-]

xiv
 List of Symbols

vc cement volume fraction in the mix design (Fret equation) [-]

vw water volume fraction in the mix design (Fret equation) [-]
w deflection [m]
w/c water to cement ratio [-]
x location from the surface (depth) [m]
y0 offset [-]

Greek lower case letters

partial reaction order [-]
partial reaction order [-]
400k deflections after 400 thousand cycles [m]
beam beam deflections [m]
burden burden displacements [m]
init initial deflections [m]
c concrete compression strain rate [-]
ct concrete tensile strain rate [-]
c averaging factor of concrete stresses considering the stress gradient [-]
max maximal stress [MPa]
min minimal stress [MPa]
time [s]
density in general [kg/m3]
c density of cement [kg/m3]
CaO density of calcium oxide [kg/m3]
Fi fatigue damage function [-]
R residual life reduction parameter [-]

xv
 Introduction

1 INTRODUCTION
Concrete deterioration due to an aggressive environment has become a widely discussed issue,
mainly in the last 25 years. Exposition to an aggressive environment reduces the durability of
materials, which is an essential design requirement for a sustainable development. For most kinds
of materials, including concrete, durability is influenced by material characteristics (compressive
strength, elastic modulus, etc.) and by the matrix structure of the material. For concrete structures,
soundness of the cover layer is another durability criterion that ensures integrity of the designed
reinforced concrete cross-section.
Another problem of the deterioration due to the aggressive environment is that it does not act
separately but interacts with other deteriorative processes and loadings. Many common structures
located in the aggressive environment (de-icing salts for road maintenance, chemical plants) e.g.
brides, crane tracks and parking lots are often exposed to the cyclic loading or in extreme cases to
the impact loading.
It is commonly known fact that the cyclic loading has an influence on the increasing deflection
development during the structural lifetime. This phenomenon was widely investigated as
a separate issue but the researches were not particularly focused on the effect of the chemical
deterioration on this increasing development of the deflections.
This applies also for the impact loading which was widely researched for concrete structures with
the possibility of vehicle crashes or terrorist attacks. These structures are also affected by the
influence of the aggressive environment.
With regard to lack of the information about the coupled deterioration effects on the structures of
transport infrastructure and many other kinds of structures where the combined deteriorative effect
applies it may be very useful to provide a tool which can accurately describe the problem of
chemical deterioration and also will be usable in other types of structural analysis.
According to the following state of the art (section 2) the deterioration processes can be assessed
independently using the superposition principle. The link between different concrete deteriorations
should be sought in a mechanical properties reduction. With regard to the mentioned systems for
deterioration degree evaluation the reduction in compressive strength (aggressive environment)
and modulus of elasticity (cyclic loading) seems to be appropriate, thus the prerequisite for coupled
deterioration due to aggressive environment and fatigue caused by cyclic loading is to combine
reduction of compressive strength and modulus of elasticity of both damage components. The
principle of interaction (superposition) of both types of damage as presented in this thesis can be
seen in Fig. 1. The impact loading, as it is not the deterioration effect, but extreme type of loading

1
 Introduction

which usually leads to the element failure or at least a part of it, is applied to the already
deteriorated elements, thus it is introduced at the end of the diagram.

Fig. 1 Principle of the proposed superposition of the damage components

The structure of the thesis follows the system proposed with the diagram in Fig. 1. General
description of the problems with their theoretical background is provided in section 3 in the order
in which they are considered for evaluation: section 3.1) chemical deterioration of concrete;
section 3.2) influence of cyclic loading and section 3.3) impact loading.
The theoretical background is followed by the description of the long-term experimental program
together with the analysis of the obtained results are described in section 4.
Discussion in section 5 provides a brief summary of findings for each type of examined problems.
The evaluation of the goals of the thesis is made in Conclusions (section 6). Main goals of the
thesis are stated below.
Main goals of this thesis are following:
To propose an analytical model dealing with the acidic chemical deterioration of concrete
structures which will be applicable to other types of the structural analysis.
2
 Introduction

To extend previously developed fatigue damage function (previous work of doc. Foglar)
for the calculation of the deflection increase due to cyclic loading with deteriorative effects
of the aggressive environment.
To incorporate the analytical model of chemical deterioration into the estimation of the
concrete elements impact resistance together with the recommendation for the numerical
modelling of these coupled deterioration problems.

3
 A State of the art

2 A STATE OF THE ART

Many kinds of concrete structures may be exposed to several deterioration processes during their
lifetime, e.g. structures of transport infrastructure are exposed to aggressive environment (salt for
winter road maintenance, acid rains etc.) and cyclic loading (traffic loading) which together cause
long-term concrete deterioration. Consequently these structures may be exposed to the extreme
type of loading impact. The combination of loadings (chemical, mechanical) during the lifetime
of the bridge pier is shown in Fig. 2.

Fig. 2 Bridge pier exposed to different deterioration processes

To correctly describe the coupled deterioration problems a state of the art for each particular
deterioration process should be made. The detailed bibliography study should be proper
knowledge base for its further extension.

2.1 Chemical deterioration of concrete

Aggressive environments do not occur only in chemical plants, and ways of classifying them have
been widely discussed [1, 2]. For concrete structures, any environment can be considered as
aggressive if its pH is lower than 12.5, when the destabilization of Portlandite begins to occur
[1, 3]. Significant reduction of the lifetime of the structure takes place when the aggressiveness of
the environment is much higher (when there are much lower pH values).
Lauer [1] focused on the problem of classifying the chemical deterioration of concrete. Many
deteriorative reactions which may occurred in structural concrete are discussed (effect of pure

4
 A State of the art

water without few or no calcium salts, sulphate attack, alkali-aggregate reaction, effect of
seawater, acids). Result of the review is that the classification based only on rating system is
insufficient. The associated cause of the damage and the resulting physical condition of the
structure must be known.
In [2] Lauer made a review of the damage classification systems for concrete structures. Based on
the large dataset of survey data of existing structures it is recommended to use statistical evaluation
as used for e.g. concrete pavements in USA. The different national damage evaluation systems of
existing structures similar to the Czech standard SN 73 6221 [4] were introduced.
In addition to the classification problems there is no easy way to compare the deteriorative effects
of different acidic environments. This was pointed out in several articles [3, 5, 6].
Pavlk [3] focused on the depth of the corroded layer of cement pastes exposed to different acids.
The influence of calcium salts of reacting acids solubility was proven on the experimental data.
With higher solubility of the calcium salts, the transfer of the acidic environment through the
corroded layer is more rapid, so the corrosion rate can be assumed to be higher. The results also
showed, that there is strong influence of the acids dissociations constants, therefore pH values
cannot be used as a basis for comparing aggressiveness (when dissociations constant vary
significantly). Moreover, according to the results, even acid concentration values cannot be
correctly used as the basis for a comparison of weak and strong acids deterioration effects.
ivica and Bajza [5, 6] made a detailed review of the acid attack on cement based materials. In
first part [5] the principal of acid attack is described for different types of acids, mostly for the
organic acids. In the second part [6] different factors influencing the deterioration rate are
presented, e.g. solution concentration, action of the solution static / dynamic (further researched
in [7]), mix design etc.
The chemical deterioration of concrete can be divided into two groups: deterioration due to human
activity (acid rains, sewage systems, winter road maintenance) [712], and due to natural causes
(acid soils, sea water, and biodegradation caused by bacteriological activity) [5, 1316].
Okochi et al. [7] tested cylindrical specimens (100x50 mm) with two simulated acid rain solution
according to the chemical analysis of the fog-water and rainwater in Japan. Two different systems
of testing was used to simulate the summer and winter conditions. Furthermore different rain
intensities were investigated. No influence of the higher or lower temperature was found. Vice
versa the influence of the rate intensities was proven. If there is the same total rainfall, rain with
lower intensity has a stronger deteriorative effect than rain with higher intensity. This finding can
be explained by the longer contact time, and can be expected to occur only for short-time
phenomena such as rainfall. At the same time, it can be assumed that the higher flow rate of steadily

5
 A State of the art

flowing media (e.g. sewer systems) will disrupt the surface structure of the material and will lead
to faster deterioration [6, 14].
Kong and Orbison [8] prepared 140 concrete cylinders which were continuously immersed in
combined solution of sulfuric and nitric acid with different pH. The tested specimens were
removed from the solution each week, brushed and rinsed in tap water to simulate erosive effect
of rainfall. Reduction in weight and compressive strength increased with lower pH value.
Xie [9] tested cubes with 100 mm edge for different pH levels. Two different systems of aggressive
environment were compared; periodic soaking (5 days in aggressive solution and 1 day drying)
and spraying (as soon as the specimen is dry, the surface is sprayed again). The results showed,
that the two methods are comparable and the results are consistent. The corrosion rate was higher
for the soaking method, thus it was recommended for the accelerated test. The spraying method is
suitable for more realistic investigation, but with higher time consumption.
Kanazu et al. [10] tested small mortar plates (40x15x160 mm) with different w/c ratio. The
aggressive solution was mixed form different acids with different pH levels. The results showed,
that there is a linear relation between total rainfall and corroded layer thickness.
Fan et al. [11] tested 354 prism specimens made from ready-mixed commercial concrete. Four
groups were made (reference group in water (pH = 7.0) and three different pH groups (pH 1.5, 2.5
and 3.5). The quantitative evaluation of the experiment was performed. The specimens showed
a steady rate of mass loss when exposed to acid solutions. At the beginning of the experiment the
strength of the specimens slightly increased and then decreased gradually. The visual observation
showed that the surface colour gradually changed from grey to brown.
Chen et al. [12] made 40x40x160 mm specimens which were permanently immersed in aggressive
environment simulating the acid rains. The surface colour changed from grey to brown. The
compressive and flexural strength decreased with the time of exposure.
ivica and Bajza [13] focused on the influence of the organic compounds. The tests were
performed on the cubes with 20 mm edge and prisms 20x20x120 mm. Strong aggressiveness of
the acetic acid and even higher aggressiveness of the lactic acid had been shown. Vice versa the
solution of saccharose, glycerol, petroleum and mineral oils have shown as unaggressive.
Fernandes et al. [14] described a problem of severely damaged sewage system in Spain mechanical
after 9 years of usage. The extreme importance of correct classification was showed. The sewage
system was designed for XA2 aggressiveness but the research results proved aggressiveness
extremely higher than XA3. A petrographic analysis was made on the collected concrete
specimens together with the EDS testing.
Windt et al. [15] focused on bioleaching due to the microorganisms of Aspergillus niger fungi
which is commonly found in soil environment and colonize mortars. The bioleaching test was
6
 A State of the art

performed on cylinders from CEM I 52.5 N CP2. The influence of the organics acids secreted by
the fungal culture was modelled with HYTEC reactive transport model.
Bertron et al. [16] focused on the influence of manure on the concrete. Cylindrical specimens
(75x25 mm) were stored in aggressive solution of acids which are contained in the manure. The
main objective was to validate the use of solution at a pH of 4 for accelerated tests. The differences
in results for test with manure pH = 6 and pH = 4 were negligible, therefore the pH = 4 was
recommended for accelerated tests.
Another major part of cement material research is aimed at providing increased resistance against
the influence of aggressive environments.
Girardi et al. [17, 18] showed that the composition of the concrete material matrix can have
a significant effect on resistance to an acidic environment, e.g. partial substitution of Portlandite
with more stable calcium-aluminate hydrates can reduce the transport properties of the mixture,
and can thus increase its resistance to acidic environments. The significance of these substitute
materials decreases with increasing aggressiveness.
Chatveera and Lertwattanaruk [19] experimentally demonstrated that the total amount of substitute
material is another influencing factor. Results of the experiment showed that only 10% substitution
of Portland cement with black rice husk ash (0% - 50% substitutions were studied) significantly
improved the compressive strength and reduced the weight loss after exposure to an aggressive
environment.
Another influencing factor is the w/c ratio (or w/b in the case of blended cements). The
permeability of the concrete decreases with the w/c value, so the transfer of aggressive media
through the material matrix is slower. According to the recommendations stated in [6], concretes
with higher resistance to aggressive environments should have w/c lower than 0.45, and the w/c
value should not exceed 0.6. This recommendation corresponds with the European standard for
concrete EN 206 [20].
The chemical composition and the mechanical properties of the deteriorated specimens have been
tested experimentally and studied quantitatively in [7, 11, 12, 15, 18, 21]. Despite the extensive
research the only analytical model for the mechanical properties deterioration was found in [22].
Huang et al. [22] proposed a general analytical solution for mechanical properties on the basis of
experimental data. However, the concentrations of the acids used for the proposed analytical
equations were extremely high, pH (-0.15; -0.75). Although the agreement with the results was
good, the experimental data contained only one measurement after 24 hours of exposure. Thus it
can be concluded that the analytical formulas were not sufficiently verified.
The different systems for the evaluation of the deterioration phenomena was proposed, the system
using reaction kinetics seems suitable.
7
 A State of the art

Beddoe and Dorner [23, 24] presented a scheme for the numerical evaluation of the deterioration
based on the chemical reaction kinetics but no further extension of their work was found.
Miyamoto et al. [25] proposed theoretical base for the numerical modelling of the deterioration
system due to a combination of different acids.
The principle of chemical kinetics is applied in the analytical model proposed in this thesis.

2.2 Influence of cyclic loading on deflection development

The influence of fatigue, respectively cyclic loading, on changes in the material and their
descriptions have been studied by many authors [2639].
From this selected range, it is possible to distinguish between three different approaches to solving
the aforementioned problem.
First approach is based on usage of experimentally obtained stress-strain diagrams of concrete
subjected to cyclic loading [2628]. These approaches are easy to use in common practice or for
simple implementation into the FEM solvers. Their disadvantage is the inability to capture the
complex behaviour of concrete under cyclic loading.
Sinha et al. [26] divided the stress-strain diagram into three different curves; unloading curve,
reloading curve and envelope curve. From the experimental results all these curves were fitted to
provide a complex material model of concrete exposed to cyclic loading. This approach is further
used in [27, 28].
Martnez-Rueda and Elnashai [27] extended an existing concrete model to new, more general,
cyclic stress-strain model for both confined and unconfined concrete. The proposed model was
implemented to the non-linear finite element analysis program ADAPTIC. The verification was
made on reinforced concrete elements.
Bahn and Hsu [28] experimentally proposed stress-strain diagram for cyclically loaded concrete
in compression. The concrete cylinders (150x75 mm) were loaded in uniaxial compression in
different time setups. Monotonic loading to obtain complete monotonic stress-strain curves; cycles
to envelope curve where the specimens is unloaded to zero load level and consequently reloaded
up to the yield stress; cycles to common point where the specimens is unloaded to zero load level
and reloaded to the common point (intersection of the reloading and unloading curve); cycles with
random loading mix of different loadings.
More sophisticated approaches are based on continuum damage mechanics. From this group, one
type are the models based on usage of bounding surfaces [2931]. The bounding surface models
are defined in e.g. [36]. The second type of mentioned approaches is damage and damage-plasticity
models [3235].

8
 A State of the art

Voyiadjis and Abu-Lebdeh [29, 30] presented a rate-independent, plasticity-type bounding surface
model for the multiaxial monotonic and cyclic behavior of concrete. The used bounding surface is
a function of maximal strains in compression with assumption that the surface is reduced with
higher strains. The deflection development is calculated based on the plastic modulus which is
a function of maximal strains and of the distance from the present stress point to the bounding
surface along the deviatoric stress direction. As the bounding surface modified Ottosen failure
criterion [40] was used. The material model was verified on the dataset from different research
programs.
Yang et al. [31] developed complex mathematical model including bounding surface for plain
concrete. Its main advantage is that the concrete compressive strength fc is the only necessary input
material constant. As in case of [29, 30] the material model can predict the concrete behaviour
under uniaxial, biaxial and triaxial compression. Model was fitted and compared with available
experimental data.
Papa and Taliercio [33] proposed a theoretical model for the description of the static and fatigue
behaviour of concrete under uniaxial and triaxial loading. Firstly, the parameters for the static
behaviour are described together with the unilateral damage assumptions. Consequently the
extension for the behaviour during the cyclic loading is made. As a result a model based on damage
theory with use of tensorial damage in conjunction with a cyclic damage evolution law was
proposed. The model was verified with available experimental data.
Alliche [34] developed a 3D model for the concrete behaviour exposed to fatigue loading. The
model is based on the damage mechanics using the thermodynamics potential for cementitious
materials. A second-order tensor is introduced for the anisotropic behaviour of the micro-cracked
materials. The damage dissipation due to the loadings under the yield surface is incorporated. The
crack development is forced by tensile stresses. The verification was performed with available
experimental data.
Grassl and Rempling [35] proposed a three-dimensional damage-plasticity interface model. In the
model the ratio of permanent and reversible inelastic displacement can be controlled. As an input
eight model parameters are needed (modulus of elasticity, Poissons ratio, compressive strength,
shear strength, tensile strength and uniaxial tension / compression fracture energies). Author of the
thesis believes that the pure shear strength and fracture energies may be difficult to obtain
experimentally.
At last but not least, it is appropriate to mention the analytical models based on experiments. The
influence of changes in the material caused by cyclic loading on the deflections of concrete
samples, or structures was developed by Holmen [37]. This approach was further extended by
Foglar [38].
9
 A State of the art

Holmen [37] developed a function for strain development dependent on the duration of cyclic
loading and maximal stress. In total 462 cylindrical specimens (250x100 mm) were tested
statically and dynamically in compression. A total of 140 specimens were tested under constant
amplitude and 180 specimens were tested under various variable load histories. The results showed
that the maximum longitudinal strain development consists of two components; time-dependent
and stress-dependent. The first one is known as cyclic creep.
The cyclic creep was studied in [39] where the model based on Paris law for fatigue growth is used
in analogy to the steel structures. The proposed model is based on the fracture mechanics and was
verified with long-term measurements of the bridges deflections.
The extension of [37] proposed in [38] is further used in this thesis for the assessment of the cyclic
loading influence on deflection development.

2.3 Impact loading

Impact as a type of extreme loading is widely studied since the early 1980s. The main reason for
the investigations is concrete sensitiveness to the strain rates. For these testing a drop-weight
mechanism is usually used [4152].
The research is mainly focused on the improvement of the fracture energy of concrete exposed to
high strain rates by addition of the fibres into the concrete mixture [42, 43, 45].
Mindess and Bentur [41] tested three differently reinforced concrete specimens and reference plain
concrete specimens. First type of specimens was reinforced with 1.25% of 25 mm long glass fibres,
second type was reinforced with 1.25% of 50 mm long steel fibres and one type was reinforced
with conventional two 10 mm steel bars. The testing was performed on drop-weight machine with
345 kg hammer from drop height 500 mm. The testing was recorded with high-speed camera at
500 fps. The results showed that the cracking process under impact loading is similar to static
loading. On these results the consequent research program was based [4244]
The following experimental program of Mindess, Bentur et al. [4244] was focused on the testing
of beam specimens (1400x100x125) with different setup. The concrete specimens were made as
normal strength concrete (compressive strength around 40 MPa) and high strength concrete
(compressive strength around 80 MPa). The specimens were reinforced with conventional steel
bars together with 0.5% of fibrillated polypropylene fibres. In [43] the specimens had notch to
evaluate fracture energy and critical stress intensity factor.
The results from [42] showed that the polypropylene fibres had only small effect during the static
loading, but a much bigger influence on the ductility in impact. This was even more significant for
the high strength concrete, therefore the addition of the fibres can be recommended to compensate
its more brittle behaviour during the impact.
10
 A State of the art

The results from [43] showed that the dynamic value of fracture toughness (critical stress intensity
factor) is much larger than the static value. The high strength concrete required less fracture energy
during the impact loading than the normal strength concrete which reflects its more brittle
behaviour. Moreover the high strength concrete specimens was able to carry higher maximum
loads for un-notched specimens, but lower maximum loads for notched specimens. The fracture
toughness as well as the fracture energy was increased with use of the polypropylene fibres.
Another extension of the research in [41] was made by Wang et al. [45]. Nine different types of
fibre reinforced concrete with different percentage of fibres were produced and tested in three-
point bending with span length 305 mm. The increase in fracture energy of the specimens was
about 22% for 0.5% of polypropylene fibres and about 41% for 0.5% steel fibres. It appeared from
the results that there is a critical volume of fibres at which the fracture energy of the concrete
jumped to a much higher level. Beyond this point, there was only a small increase in the fracture
energy with further increases in the volume of fibres.
Rieder and Mindess [46] developed a new technique for the impact testing of confined concrete
specimens. Due to experimental layout different prestress were chosen for the confinement. The
used specimens were cubes with 100 mm edge and compressive strength of 85 MPa. The
preliminary results of the extensive testing (over 2000 specimens) showed that the variations in
the impact energy and in the maximum impact load due to the confinement were small. The
consequent splitting tension tests showed that the damage depends on the level of confinement.
Bischoff and Perry [47] tested plain concrete specimens in uniaxial compression. Concrete
cylinders (250x100 mm) from two concrete mixes with 28-day design cube strengths of 30 and 50
MPa were prepared. According to the results the compressive strength increased by 50-60% during
the impact loading, axial strains increase between 14% and 36% and energy absorption increase
between 64% and 118%.
The influence of the strain rate value on the development of mechanical damage in concrete was
described in [48].
Sukontasukkul et al. [48] tested concrete cylinders (200x100 mm) to evaluate the influence of
strain rate on the development of mechanical damage in concrete during the impact loading. Two
different methods based on the damage mechanics were used for the evaluation (strain rate
variation and deterioration of modulus of elasticity). The results showed that the deterioration of
modulus of elasticity method is more realistic especially after the peak value of loading.
The high topicality of the impact loading can be supported by effort to develop standardized tests
[49, 50] or with a development of complex FEM models for concrete under impact loading
[51, 52].

11
 Coupled deterioration processes

3 COUPLED DETERIORATION PROCESSES

3.1 Mathematical description of concrete acidic deterioration
3.1.1 Basic chemical prerequisites
For most cases of material deterioration, it can be assumed that the main factors influencing the
degree of deterioration are the duration of the process and its aggressiveness (rate). For chemical
deterioration, the rate is based on the chemical reaction rate, which can be expressed by chemical
kinetics. In order to achieve a chemical reaction, particles of the reactants must come into close
contact. Thus the chemical reaction rate is influenced by factors which increase the probability of
contact between molecules: an increasing number of particles per unit volume (an increase in
concentration), increasing intensity of molecular motion (an increase in temperature), and the area
of the contact surfaces (reactions of gases and liquids are significantly faster than reactions of
solids). These factors are applied in a differential rate equation, which can be used to determine
the reaction rate. The rate equation can be written in the form stated in equation (3.1.1).

dni
= kc (T ) cA cB (3.1.1)
iVd

The partial orders of reaction and are not generally identical to the stoichiometric coefficients
(except in the case of an elementary reaction). The order of the reaction n = + and the rate
constant kc have to be determined experimentally. Methods used in chemical kinetics usually
provide the concentration in dependence on time, which can be further used to determine the
reaction rate. The basic prerequisite for the measurement is to obtain the development of the
concentration, or other variables, from which the change in concentrations can be determined (e.g.
a pH measurement).
Several methods can be used to determine the reaction order and the rate constant: the integration
method, the half-life method, differential methods, for details see [53]. An integration method was
used for the purposes of the experimental program described in this thesis. The integration method
compares the dependence of the concentration in time with data from the integrated forms of rate
equations. With the use of the graphical solution, the left sides of the integrated form of rate
equations (3.1.2) are plotted against time. The correct solution exhibits linear progression.

ln c A c A,0 for 1st order

(3.1.2)
(1 c A 1 c A,0 ) for 2
nd
order

12
 Coupled deterioration processes

3.1.2 Adapting the reaction kinetics to determine the chemical deterioration of

concrete
As described above, chemical kinetics can also be used for determining the chemical deterioration
of concrete. The rate of chemical deterioration of concrete depends mainly on the matrix of the
material and the aggressiveness of the environment. The aggressiveness of an acidic environment
in relation to concrete structures is primarily defined by the pH value, which is a function of the
concentration of H+ ions, see equation (3.1.3).

pH = log ( cH ) (3.1.3)

For the experimental program (described in section 4.1), hydrochloric acid (HCl) was used as
a deteriorative substance. In this case, it can be assumed (in a simplified way) that the deteriorative
reaction that takes place when an aggressive environment comes into contact with concrete will
correspond with the neutralization according to equation (3.1.4). Thus hydrochloric acid reacts
with calcium hydroxide (Ca(OH)2) to form calcium chloride (CaCl2) and water. Generally, the
ion-molecule reaction occurs when dissolved H+ ions react with solid Ca(OH)2, which
disintegrates into calcium ions Ca2+ and water (for the scheme of the expected chemical processes
see Fig. 3).

Fig. 3 Generalized scheme of the acidic deterioration of concrete [23]

A reaction of the acid with compounds of iron, aluminium and silicon takes place to a negligible
extent, as described in [3]. This assumption was experimentally verified by SEM and EDS analysis
(section 4.3) and also by a chemical analysis of the aggressive solution applied to the concrete
specimens, Tab. 6.

2 HCl + Ca ( OH )2 CaCl2 + 2 H 2O (3.1.4)

Using the equation of neutralization (3.1.4), the general form of the rate equation (3.1.1) is
transformed into the specific rate equation describing the above-mentioned neutralization equation
(3.1.5). This yields an equation with three unknowns that cannot be easily solved.
13
 Coupled deterioration processes

dcCa
=
kc cCa cH (3.1.5)
d

If the pH (H+ ion concentration) is kept constant (a buffer solution), or is in such excess to the
concentration of the second reactant that their concentration remains practically unchanged, the
rate of the reaction in (3.1.4) can be described by a first order rate equation (3.1.6). This solution
uses a pseudo-rate constant, which also includes the effect of the concentration of the H+ ions
( kc = kcH ). With this simplification, the second order reaction is converted to a pseudo-first order
reaction that can be solved effectively by conventional methods.

dcCa
A products kccCa
=
(3.1.6)
d

Equation (3.1.6) provides the rate of loss of concentration of the calcium ions. Integrating this
differential equation gives the concentration value at a specific time dependent on the pseudo-rate
constant, the initial calcium ion concentration and time. The integrated form follows:

=cCa cCa ,0 exp ( kc ) (3.1.7)

The initial value of the concentration of the calcium ions cCa,0 is different for each mix design. In
order to determine it accurately, the composition of the concrete has to be known, particularly the
specific type of cement and its volume. For commonly used types of concrete, the initial value for
the concentration of the calcium ions ranges from 3.5 to 5.0 mol/dm3 of concrete.
Generally, the rate constant is constant for a particular chemical reaction. Due to the diffusion
phenomenon which takes place in the concrete deterioration rate, thus in the decrease in the
reaction rate in time, the rate constant should be a function of time. In the experimental program
described below, the rate constant was determined as a function of time, which already includes
the effect of the decrease in the reaction rate. With this simplification, the transport of the
aggressive media through the concrete material matrix can be neglected. In order to keep correct
chemical terms, from now on the designation rate constant will be used, although it is a function
of time. The rate constant must be determined experimentally using chemical or physical
measurements.

3.1.3 Deterioration of the concrete specimens

With the rate constant and the solved rate equation for the deterioration of concrete due to chemical
reactions, we can obtain the actual rate of dissolution of the calcium ions in time. These quantities
are insufficient for quantifying the degree of deterioration of the concrete elements. To overcome

14
 Coupled deterioration processes

these difficulties, it is necessary to link the chemical equations that are obtained with methods for
quantifying the material properties.
For this purpose two methods to determine the deterioration degree of concrete exposed to
aggressive environment were proposed: 1) determination of the reduced compressive strength and
2) evaluation of the time variable cross-sectional area of sound specimen core (analogy with fire
damage assessment in [54]).
Both approaches are described below. The connecting link between these two methods is the
solution of the differential equation (3.1.5) in its simplified form as stated in (3.1.7).

3.1.3.1 Reduced compressive strength

Compressive strength, which is closely associated with other material properties and is one of the
main characteristics for the concrete structure design, was chosen as an essential mechanical
property. Standardized approaches use the compressive strength to derive other material
properties, e.g. modulus of elasticity, fatigue resistance, fracture energy etc.
The effect of the composition on the properties of the concrete has been widely described in the
literature [5557]. Four main components of cement may be considered: tricalcium silicate (alite)
3CaOSiO2 (C3S); dicalcium silicate (belite) 2CaOSiO2 (C2S); tricalcium aluminate
3CaOAl2O3 (C3A), and tetracalcium aluminoferrite 4CaOAl2O3Fe2O3 (C4AF). The first two of
these components have a major influence on the final compressive strength of concrete [55, 58],
and their volume in the cement can be determined using Bogues equations (C3S approximately
65%, C2S approximately 10%). During hydration of the cement, these two compounds form
primarily a C-S-H gel (1.7CaOSiO24H2O the formula is only approximate, because the structure
of the hydration product is highly variable, particularly in the context of water [57] and calcium
hydroxide (Ca(OH)2). The C-S-H gel together with calcium hydroxide is about 75% of the volume
of the cementitious binder. The amount of calcium oxide (CaO) entering the hydration process
greatly exceeds other oxides that enter the process, so it can be assumed that the total amount of
calcium oxide in the cement has the greatest influence on the final compressive strength. On the
basis of the quantity of calcium oxide in the total volume of the concrete, we can determine the
initial concentration of calcium ions cCa,0 and subsequently, by calculating the change in
concentration according to equation (3.1.7), we can determine the reduced compressive strength.
The Fret function [59] for calculating the compressive strength on the basis of the mix design was
chosen for this purpose. The Fret function is used in the form according to equation (3.1.8).

2
vc
f ck ( ) = k f ( ) f cm (3.1.8)
vc + vw + va

15
 Coupled deterioration processes

The Fret coefficient must be determined experimentally for each mixture. For the purposes of the
experimental program described below, the tests were performed on reference cube and beam
specimens.
To include the concrete deterioration rate into the Fret equation, the calculation of the Ca2+ ions
concentration in time (cCa) according to equation (3.1.7) can be used. The change in the
concentration of the calcium ions will be reflected in the volume of cement, according to equation
(3.1.9).

mc cCa M ( CaO )
=
vc = (3.1.9)
c X CaO CaO

Substituting (3.1.7) into (3.1.9) yields the relationship for the compressive strength as a function
of time and the pH value of the aggressive environment (3.1.10). The time-dependent increase in
compressive strength is taken into account with the Fret coefficient.

vc exp ( kc ( pH ) )
2

f ck ( , pH ) = k f ( ) f cm (3.1.10)
vc exp ( kc ( pH ) ) + vw + va

The 28-day strength Fret coefficient can be used for the calculation in accordance with
standardized approaches. It can be assumed that due to the time development of the compressive
strength this simplification will be on the safe side.

3.1.3.2 Reduced cross-sectional area of specimen core

This method is based on the research described in [60] where it is stated that the transport of H+
ions may be considered as steady-state diffusion due to slow increase of the corroded layer
thickness. Two boundary conditions are proposed on the surface of the concrete specimen the
concentration of H+ ions is equal to the concentration of aggressive solution; in contact of the
corroded and uncorroded layer the pH value is due to neutralization around 7 (cH+ = 10-7 mol/dm3).
The gradient dcH+/dx is nearly constant. This was experimentally verified for 0.5 M nitric acid
(HNO3) [60].
From the viewpoint of the conventional diffusion it is possible to describe the transport process of
H+ ions with the Ficks second law of diffusion (3.1.11).

c 2c
=D 2 (3.1.11)
t x

16
 Coupled deterioration processes

Considering constant diffusion coefficient, constant boundary conditions and constant initial
conditions (3.1.12), the problem (3.1.11) has simple mathematical solution (3.1.13) [61]. For its
simple applicability it is widely used in concrete diffusion problems [6264].

c (=
0, t ) c0 c (
= ,t) 0 c ( x=
, 0) 0 (3.1.12)

x
c ( x=
, t ) c0 1 erf (3.1.13)
2 Dt

The comparison of the pH distribution according to equation (3.1.13) and Pavlk [60] for time
t = 365 days and initial concentration c0 = 10-2 mol/dm3 (pH = 2) can be seen in Fig. 4.

Fig. 4 Comparison of the pH distribution in direction of diffusion according to Ficks law and [60]

The comparison shows disagreement in the mentioned assumptions. The suitability of the
approach from [60] will be described in the results of the experimental program.
When the pH distribution through the corroded layer according to Pavlk [60] proves to be valid,
there is a possible presumption that for the deterioration degree evaluation, the time variable
surface exposed to aggressive environment can be used. At the basis of the kinetics of chemical
reaction the amount of dissolved Ca2+ ions can be determined. For the known mix design, it is
then possible to identify the thickness of the layer with almost none Ca2+ ions. This layer may be
eliminated from further cross-sectional assessment. Proposed principle can be seen in Fig. 5.

Fig. 5 Principle of the reduced cross-sectional area method

17
 Coupled deterioration processes

A long-term experimental program was designed and executed in order to verify the proposed
assumptions, together with the proposed form of the rate equation, the time/pH dependent
compressive strength calculation and reduced cross-sectional area method. Laboratory tests, SEM
and EDS analysis methods were used for the verifications.

3.2 Extension of the strain development under cyclic loading

calculation
It is well-known fact that the concrete structures deflections increases under the cyclic loading, the
phenomenon is called cyclic creep [39]. The strain development is similar to the crack growth rate
of steel under cyclic loading. The secant modulus of elasticity under cyclic loading reflects the
strain development, see Fig. 6.
As described in [37] the phenomenon of deflection increase consists of two components:
1) increase due to the internal mechanical damage of the concrete matrix caused by the repeated
loading and 2) time-dependent strain increase which is independent of the concrete endurance and
can be equated with an ordinary creep.

Fig. 6 Development of the secant modulus of elasticity of concrete under cyclic loading [37]

Based on the research from [37] Foglar, in his dissertation thesis [38], proposed an analytical
model dealing with the deflection increase based on the maximum stress level Smax. The proposed
mathematical function, fatigue damage function, was experimentally verified on the prestressed
concrete specimen and is defined in equation (3.2.1). For the detailed description of the function
see [38].

c
Smax 3

n c4 ni
Fi = 1 a i
+ b exp 1 c2 (3.2.1)
c1 N N

18
 Coupled deterioration processes

The aims of this thesis section can be divided into three parts: 1) extend the calculation process of
the deflection increase due to the cyclic loading using the fatigue damage function and verify the
extension with the experimental program; 2) investigate the influence of the concrete acidic
deterioration on the deflection increase due to the cyclic loading and 3) propose an analytical
solution for the remaining lifetime calculation of the structures exposed to cyclic loading
(independent on the ratio of the number of load cycles that the structural element has already
resisted n to the total number of load cycles that the structural element is able to resist at the
particular load level N).

3.2.1 Extension of the deflection increase calculation process

The fatigue damage function (3.2.1) is designed as a reduction factor of the initial secant modulus
of elasticity (see equation (3.2.2)).

Ec=
, ni Fi Ec ,init (3.2.2)

For the most simple calculation of the deflection increase the reduced modulus of elasticity Ec,ni
can be used for the whole analysed specimen. This corresponds to the calibration experiments in
[37] where the stress distribution over the specimen cross-section (probe cylinder) was constant.
This simplification can be used for the specimens or structures exposed to the uniaxial loading
with good agreement, but is not valid for specimens subjected to bending, where each part of the
cross-section is exposed to a different compressive or tensile stress.
For the specimens exposed to bending the calculation using the reduced modulus of elasticity Ec,ni
for whole specimen may lead to significantly conservative values. It can be assumed that most of
the structures subjected to cyclic loading are exposed to bending.
To consider the variously stressed fibres in the structural element subjected to bending the
compressive zone is divided over the height of the cross-section. The analysed cross-section
decomposes into an ideal cross-section of layers with a different modulus of elasticity depending
on the stresses that it is subjected to. This approach should reflect the differences between the
fatigue behaviour of the compressive cyclic loading of the cylinders, on which the fatigue damage
function is based, and specimens subjected to cyclic bending.

3.2.1.1 Calculation of the deflections

For an evaluation of the deflections, the partial integration over the compressive zone height was
used to reflect the real stress distribution. The method is based on the average stress in each layer,
and not on the maximal stress in the top fibres, as e.g. in [65].

19
 Coupled deterioration processes

From the definition of the fatigue damage function, the modulus of elasticity decreases to zero at
the limit number of applied cycles N (this corresponds to the behaviour of the probe cylinders). In
the case of cylinders, the specimen fails as a whole element after N cycles; but in the case of
specimens subjected to bending, the bond between differently stressed neighbouring parts will
influence the behaviour of the specimen. When one layer reaches the limit number of load cycles
corresponding to the applied Smax, there is still a bond with the neighbouring part. When one of the
layers deteriorates, the other layers take over its role. Due to this stress redistribution, some
residual value of the modulus of elasticity of the extremely damaged part can be assumed. The
position of the neutral axis changes with the deterioration of the layers, and thus changes the
maximum stress applied to the layer.
To include the facts mentioned above, the constant multiplier of the modulus of elasticity is used
when the fatigue damage function reaches the third phase according to Fig. 6. This constant
multiplier is taken as an average value of the fatigue damage function at the start of the third phase
of the strain development, where the function takes values between approximately 0.65 and 0.
Thus the modulus of elasticity value is 0.33 (Eres = 0.33Ec,init) for Smax 0.2. For low maximum
stresses, it can be assumed that the influence of fatigue is lower. For that reason, the upper limit
value of the fatigue damage function ( = 0.65) was chosen for Smax < 0.2 (Eres = 0.65Ec,init). These
instant changes in the modulus of elasticity lead to instant increases in the calculated deflections.
The accuracy of the method is dependent on the fineness of the compressive zone division. With
more layers used for calculation the average stress in each layer will better reflect the real stress
distribution over the height of compressive zone. This should also lead to the reduction of the
instant increases in the calculated deflections as mentioned before. The sensitivity analysis should
be performed to determine the adequate number or height of the layers to ensure sufficient
accuracy at reasonable computational time. The principle of the partial integration together with
the influence of the number of layers used for calculation is illustrated in Fig. 7.
In order to calculate the deflections, the ideal moment of inertia is needed. An evaluation of this
cross-sectional characteristic is carried out in 4 steps:
1) From the fatigue analysis, the maximum stress levels Smax is obtained for each
partial height (layer) within each specimen.
2) The fatigue damage function is evaluated depending on the maximum stress levels
Smax.
3) The decreased modulus of elasticity is calculated with the fatigue damage function
for each layer.
4) The ideal moment of inertia is calculated on the basis of the decreased modulus of
elasticity of each layer.
20
 Coupled deterioration processes

Fig. 7 Principle of partial integration of the compressive zone for n = 400,000 cycles

3.2.1.2 Limit number of applied cycles N usage and discussion of standards

To evaluate the increase in deflection with the fatigue damage function (3.2.1) a limit number of
applied cycles N has to be known for the specific Smax. This value can be calculated by procedures
given e.g. in EN 1992-1-1 [65] or in the Model Code [66].
EN 1992-1-1 [65] states that the fatigue should be evaluated in the most stressed fibres. Model
Code 2010 [66] partially reflects the stress distribution in compression zone with c coefficient.
Calculations show that design values of the limit number of applied cycles N according to these
standards differs significantly (equations in (3.2.3) for Model Code 2010 and equation (3.2.4) for
EN 1992-2 [67]).

0.45 + 1.8 S min

Y=
1 + 1.8 S min 0.3 S min
2

8
=
log N1 ( Smax 1) (3.2.3)
Y 1
8 ln (10 ) S S
log N 2 =
8+ (Y Smin ) log max min
Y 1 Y S min

(a) if log N1 8, then log N = log N1

(b) if log N1 > 8, then log N = log N2

1 S max
log N = 14 (3.2.4)
1 min max

21
 Coupled deterioration processes

Experimental testing of reinforced concrete specimens (described further) did not show any visible
damage after 400 thousand cycles. This may correspond with the value of N calculated according
to Model Code 2010 [66] N = ~107, whereas for the calculation according the Eurocode [67] the
limit number of applied cycles N = 1900 (both values were calculated using material safety factor
M = 1.0). Based on these findings the Model Code 2010 [66] is used in the proposed extension.

3.2.1.3 Influence of the acidic deterioration of concrete

Two methods for the concrete deterioration assessment were proposed in sections 3.1.3. To verify
their suitability for use in this type of the structural analysis both methods will be integrated
(separately) to the deflection increase calculation.
For the verification of the proposed extension of the deflection increase calculation the reinforced
concrete specimens were prepared as will be described further in the text (section 4.1.1). To
evaluate the influence of the acidic concrete deterioration the same specimens were used.

3.2.2 Residual life-time assessment

The residual life-time is often discussed topic in case of the reconstruction of bridges and
engineering structures. Focus of this section is to propose an analytical model which could predict
the remaining life-time of the structural elements according to the results of the loading test of the
existing structure. The evaluation of the damage caused by the cyclic loading is based on the
deflection increase development and is independent on the limit number of applied cycles N which
as a calculated value is standard-dependent and can vary significantly as described in section
3.2.1.2. The approach is designed according to results of the extensive experimental program
described in [37].
In case of bridges, in Czech Republic, the initial loading test is quite usual, thus the starting data
for the assessment are often available. The results of loading test during the lifetime can show
increase in deflection, thus decrease in secant modulus of elasticity. On the basis of this reduction
the remaining lifetime can be predicted.
To obtain the secant modulus of elasticity after cyclic loading, the initial modulus of elasticity
should be multiplied by a reduction parameter R. The reduction parameter should take into
account the different development of deflections under different values of Smax. With regard to the
shape of the modulus of elasticity stated in [37] the function for reduction parameter was sought
for in the form of two phase exponential decay function, see equation (3.2.5).

Ec,sec= Ec ,init R= Ec ,init ( y0 + A1e x t1 + A2 e x t2 ) (3.2.5)

22
 Coupled deterioration processes

As a safe threshold of the fatigue failure 80% of limit number of applied cycles N was chosen due
to subsequent unstable increase in deflections after 85-90% of N. The increase in deflection at this
limit value is between 45-65% depending on the value of Smax.
The offset, amplitudes and time parameters were designed as constants or variables dependent on
Smax value. Their definition can be seen in equation (3.2.6).

=y0 0.6= =
t1 158 t2 5.6
(3.2.6)
A1 =
0.283S max + 0.004 A2 =
0.287 S max + 0.378

The amplitudes dependency on the Smax, thus the percentage influence of exponential parts on the
shape of the proposed function (equation (3.2.5) substituted with (3.2.6)) can be seen in Fig. 8.

Fig. 8 Amplitudes development (left) and exponential parts development for Smax = 0.5 (right)

The coefficient of determination R2 values are higher than 0.90 (except of the extremely high value
Smax = 0,95 where the R2 = 0.65), thus the function exhibits good agreement with the fitted data
from [37], comparison can be seen in Fig. 9.

Fig. 9 Modulus of elasticity fatigue age development according to equation (3.2.5)

23
 Coupled deterioration processes

3.3 Impact loading

Unlike the already mentioned deterioration phenomena the impact loading will be evaluated
separately, but for the reduced material characteristics due to aggressive environment and cyclic
loading. This approach is based on the assumption, that the impact loading is not a deteriorative
effect, but extreme type of loading which usually leads to the element failure or at least a part of
it (concrete spalling, reinforcement yielding etc.). The impact performance of a deteriorated
specimen is therefore sought for.

3.3.1 Dynamic increase factor (DIF)

Concrete is a strain rate sensitive material which results in an increase of the instantaneous
mechanical properties with higher strain rate, thus with higher loading speed. This phenomenon is
called dynamic increase factor (DIF), i.e. the ratio of the dynamic to quasi-static (strain rate 10-6)
material properties (usually compressive strength, tensile strength and modulus of elasticity).
Classification of the different kinds of loading according to the strain rates, as in Model Code 2010
[66], can be seen in Fig. 10.

Fig. 10 Strain rates for different loading velocities [66]

The DIF value is dependent on the strain rate thus the DIF curves can be developed on the basis
of the experiments. The DIF curves were published e.g. in [68, 69], standardized in Model Code
1990 [70] and further modified in Model Code 2010 [66]. For the purposes of the experimental
program the DIF values for impact loading were determined according to the [66]; equation (3.3.1)
for the dynamic compressive strength; equation (3.3.2) for the dynamic tensile strength and
equation (3.3.3) for the dynamic modulus of elasticity.

f c ,imp ,k f cm = ( c c 0 )
0.014
(3.3.1)

f ct ,imp ,k f ctm = ( ct ct 0 )
0.018
(3.3.2)

Ec ,imp Eci = ( c c 0 )
0.026
(3.3.3)

24
 Coupled deterioration processes

The DIF values were evaluated for the impact loading based on the strain rates which were
calculated from the change of the specimens deflections in time; the scheme of the calculation is
represented by the equation (3.3.4).

w ( t ) 1 r ( t ) (3.3.4)

The evaluated strain rates with corresponding values of DIF are listed in Tab. 1. The DIF values
for the cyclic loading are listed for the comparison as well.

Tab. 1 DIF values for the impact and cyclic loading according to Model Code 2010 [66]

Loading Strain rate fc,imp,k / fcm fct,imp,k / fctm Ec,imp / Eci

type -1
[s ] [-] [-] [-]
Impact 2.89e+00 1.174 1.307 1.348
Cyclic 1.17e-02 1.087 1.184 1.168

The DIF curves for mechanical properties according to equations (3.3.1) (3.3.3) with emphasized
values for the impact and cyclic loading can be seen in Fig. 11.

Fig. 11 DIF curves for different mechanical properties according to Model Code 2010 [66]

3.3.2 FEM model of the experiment

Due to the limited possibilities of the real-time measurements of the impact experiments the FEM
model in LS-DYNA software was designed. Based on the results of the experiment a model
adjustment will be made especially in material properties.
The FEM model should be a carbon copy of the proposed experiment as described further in the
section 4.1.7. The geometry of the model consist of 5 different part: concrete of the beam, beam
reinforcement, two supports and a burden. The description of each part is given below. Render of

25
 Coupled deterioration processes

the model with the mesh depicted can be seen in Fig. 12. The concrete elements are slightly
transparent in order to see the beam reinforcement.

Fig. 12 LS-DYNA model render

The model was designed using two types of elements: solid elements and beam elements. The
solid elements (concrete, supports and burden) are primarily 8-node hexahedron with one-point
integration (constant stress solid element) in case of burden 6-node pentahedrons are used as well.
The Hughes-Liu beams [71, 72] are used for the rebar beam elements.

3.3.2.1 Beam
The beam is consisting from two different parts: concrete elements and reinforcement. The
concrete elements are cubes with sides of 10 mm to reduce the time demands of the calculation.
To determine if the accuracy of the proposed element size is sufficient model with cubes with sides
of 5 mm was made. Results from both models were compared with a negligible differences. To
reduce the time demands of the calculation the bigger elements were used for the analysis.
Continuous Surface Cap (CSC) Model MAT159 was used for concrete elements. The material
model parameters will be described further in the text.
The reinforcement bars were modelled as a beam elements which nodes were connected with the
concrete elements nodes by the Lagrange in solid function which is often recommended for the
rebar modelling [73, 74]. As a material model piecewise linear plasticity was selected to define
the bilinear elasto-plastic properties of the rebar as specified in Eurocode 2 [65].

3.3.2.2 Supports
The supports was modelled with a solid elements with equal size as the concrete elements (10 mm)
to ensure simplicity of the contact between beam and supports. In the experiment two steel bars
20 mm were used. The circular cross-section of the supports increased the model time demands,

26
 Coupled deterioration processes

thus the comparison of the results obtained with rectangular and circular cross-section was made.
The evaluated differences were negligible. The elastic material model was used.

3.3.2.3 Burden
The burden was modelled using combination of the 8-node hexahedrons and 6-node pentahedrons
to obtain sufficiently accurate geometry. The burden geometry was partially simplified. The
manipulation bar was omitted. To maintain the burden weight an elastic material model for steel
with modified density was used.

3.3.2.4 Boundary and initial conditions

The correct boundary and initial conditions are essential to ensure the proper behaviour of the
model.
During the impact loading the steel bars were fixed to the base with an adhesive tape for easier
manipulation. This adhesive tape was not damaged after the impact, thus there was no movement
or rotation. Based on this fact the displacements and rotations of the supporting steel bars were
constrained.
The contacts between the beam specimen and supports or burden were generated automatically by
LS-DYNA function Automatic surface to surface. These contacts are active only in compression.
As described before the bond between reinforcement and concrete elements was modelled with
the Lagrange in solid function.
The gravitational acceleration was incorporated into the model together with the initial velocity of
the burden. The initial velocity and the burden distance to the beam was calculated according to
impact speed vimpact = 7.096 m/s which was obtained from the experiment.

3.3.2.5 Loads
For partial inclusion of the cyclic loading influence on the plastic strains in the beam a force was
applied on the beam in the four-point bending layout corresponding to the experiment described
in section 4.1.4. The time gap between the force load and impact was long enough to ensure that
the beam settles. No additional loads were used.

3.3.2.6 Concrete material model MAT159-CSCM

For the concrete material model MAT159-CSCM, elasto-plastic damage model with rate effects,
was used. Its comprehensive description is provided in [73]. This model was selected from the
several concrete material models in LS-DYNA due to its simple adjustments. The simplified
version, if selected, allows to calculate all the necessary material characteristics on the basis of the
compressive strength. The full version of the material model allows to change all the calculated
characteristics arbitrarily.
27
 Coupled deterioration processes

The material characteristics were calculated for the concrete grade C25/30 which was used for the
experimental program. Consequently the material model fracture energy can be modified for all
the tested specimens to obtain agreement between results of the model and measured data from
the experiments.
The material model includes three different fracture energies (in uniaxial stress Gfc, in uniaxial
tension Gft and in pure shear stress Gfs). Default values for the tension and shear fracture energy
are set to a hundredth of the compressive fracture energy which is calculated according to the
formula stated in Model Code 1990 [70], equation (3.3.5). For the model evaluation an updated
formula from the Model Code 2010 [66] which is no longer dependent on the aggregate size,
equation (3.3.6), will be used for an initial value of compressive fracture energy.

G f = G f 0 ( f cm f cm 0 )
0.7
(3.3.5)

G=
f 73 f cm0.18 (3.3.6)

Influence of each fracture energy will be tested on the models. The influence of the concrete
deterioration due to aggressive environment on its fracture energy will be evaluated on the basis
of the model fitting.

3.3.3 Energy balance of the system

According to the law of the conservation of energy the total energy of the described system should
remain constant during the whole analysis if there is no additional external work. Thus the total
energy is described by the equilibrium from equation (3.3.7) which should be valid during the
whole analysis.

Etotal = Ekin + Eint + Esi + Edamp + Ehg = Ekin

0
+ Eint
0
+ Wext (3.3.7)

Each energy components will be discussed to explain their effect on the overall results of the LS-
DYNA modelling except the dampness energy which is negligible in the experiment, thus it will
be excluded from the energy balance.
The time course of the impact experiment can be divided into three phases: 1) before the impact;
2) during the impact and 3) after the impact.
Before the impact the total energy mainly consists of the kinetic energy of the falling burden (this
corresponds with the experiment layout). Due to the model boundary conditions the external work
has to be considered in the total energy calculation. The external work has to be divided to the two
components: a) work done by applied forces and b) work done by the gravity boundary. As
described in previous section, for partial inclusion of the cyclic loading effect into the model the

28
 Coupled deterioration processes

force is applied to the beam before the impact. The external work of this force can be calculated
according to the equation (3.3.8).

res ,beam

=
Wext , F Fc ( t ) d beam (3.3.8)
0,beam

The strains caused by this work are reflected in the increase of the beam internal energy Eint. If
there is a plastic strains due to the applied force, part of the beam internal energy remains even
after unloading (energy dissipated in the plastic deformation). The development of the external
work done by applied forces together with the force-deflection curve can be seen in the results of
a sample LS-DYNA analysis in Fig. 13.

Fig. 13 External work done by applied force with force-deflection diagram

The work done by the gravity boundary is in accordance with the Newtons second law of motion,
thus the work done by the acceleration on the burden is equal to the change in the burden kinetic
energy, equation (3.3.9).

mg burden =
Wext , g = Ekin (3.3.9)

At the end of the first phase, i.e. an instant before the impact the total energy is equivalent to the
relationship from equation (3.3.10). The influence of the sliding interface energy Esi is negligible.

=
Etotal ,1 Ekin ,burden + Eint,beam +=
Esi (E 0
kin,burden + Wext , g ) + Wext , F (3.3.10)

During the impact the burden kinetic energy transfers with its decreasing velocity to the strains
(beam internal energy, strain energy dissipation of the burden is negligible), hourglass energy and
sliding interface energy. The second phase, impact, ends when the burden velocity vburden = 0, thus

29
 Coupled deterioration processes

the kinetic energy of the whole system is equal zero as well. The total energy at the end of the
second phase is equivalent to the relationship from equation (3.3.11). The work done by the gravity
boundary during the second phase is small due to the small burden displacement and can be
neglected. During the impact part of the energy is transformed to the beam kinetic energy which
at the end of the phase is transformed to the beam internal energy.

Etotal=
,2 Eint ,beam + Esi + E=
hg Etotal ,1 (3.3.11)

The hourglass energy is associated with the mesh instability of under-integrated elements (solid,
shell) which is known as hourglassing. It is a special case of a zero-energy modes that can lead to
singularity of the assembled stiffness matrix [75]. To prevent the mesh instability hourglass forces
reacting against the formations of the hourglass modes of the elements are incorporated into the
LS-DYNA analysis. The energy dissipated by these hourglass forces is therefore reported as the
hourglass energy. The hourglass modes problem can be entirely eliminated with use of the fully-
integrated elements but at large additional computing time cost and possible shear-lock problem.
The x-components of the hourglass modes of an eight-node element with one integration point are
shown in Fig. 14; the 8 other hourglass modes can be imagined by changing the x-axis to the y-
axis and then the z-axis [75].

Fig. 14 Hourglass modes in the x-direction for a prismatic hexahedron [75]

The problem with hourglassing was localized on the contact between the burden and beam
elements where significant deformations occur. Part of the hourglass energy (elastic) disappears
in the third phase when the burden lose contact with the beam. The hourglass energy was up to 5%
in maximum.
The sliding interface (contact) energy is the energy lost during the contact due to the contact
searching and recalculation of the elements penetration to the forces. Negative contact energy (the

30
 Coupled deterioration processes

impacted element gains more energy than the impactor had) may be caused by undetected initial
penetrations and indicates poorly designed mesh. No negative contact energy was observed during
the models analysis. The positive contact energy (the impacted element gains less energy than the
impactor had) is usually due to the penetration recalculation (if there is no friction defined). This
phenomenon can be eliminated with finer mesh but at higher computing time cost. Considering
that for all models the same mesh were used it can be assumed that the contact energy will be
similar for all the models and therefore it should not affect the evaluation.
During the third phase (the unloading phase) the burden acquires negative velocity, thus gains part
of the system energy in a form of the kinetic energy. This kinetic energy is mainly transformed
from the internal energy dissipated in the elastic deformations of the concrete (around 30%) and
beam reinforcement (around 70%). Small part of the energy is from the hourglass and contact
energy. The total energy decreases in time due to the negative external work caused by the
gravitational boundary condition. The energy equilibrium for the third phase is stated in equation
(3.3.12). The beam kinetic energy is negligible.

Etotal ,3= Ekin ,burden + Eint,beam + Esi + Ehg= Etotal ,2 + Wext ,g (3.3.12)

Due to the specification of the material model used for concrete some of the elements erode during
the calculation (when the strains exceed specified value). Energy dissipated in these elements is
therefore tracked and reported as eroded energy. This eroded energy influences mainly beam
internal energy.
The energy balance, more specifically the beam internal energy will be examined in more detail
to determine the influence of the cyclic loading and aggressive environment on the concrete energy
dissipation.

31
 Experimental verification

4 EXPERIMENTAL VERIFICATION
4.1 Experimental program
The experimental program was designed for possible evaluation of all three types of deterioration.
More precisely to evaluate chemical deterioration due to aggressive environment, mechanical
damage due to the cyclic loading (deteriorated and sound concrete specimens) and for the
subsequent evaluation of the influence of combined deterioration on the performance of concrete
specimens exposed to extreme loading (impact loading).

4.1.1 Sets of specimens

Fourteen sets of samples were made within the experimental program; 6 reference sets were stored
in a dry environment (5 exposed to cyclic loading, 1 as reference sound specimen), and 8 sets were
stored in an aggressive environment. Each set consists of one cube with sides of 150 mm, three
beams (100x100x400 mm) and one large reinforced concrete beam (300x150x1300 mm). The age
of the specimens at testing together with the length of exposure to the aggressive environment are
stated in Tab. 2. The reference set is marked R, the sets exposed to cyclic loading are marked
C, and the sets stored in the aggressive environment are marked A.

Tab. 2 Summary of the sets used in the experimental program

Set R1 C1 C2 C3 C4 C5 A1 A2 A3 A4 A5 A6 A7 A8
Age [day] 436 28 63 182 364 728 35 42 77 105 133 224 406 777
Acid [day] - - - - - - 7 14 34 62 90 181 363 734

Three beams were used to determine the elastic modulus according to ISO 1920-10 [76], the tensile
strength in bending and the compressive strength. The cube was used as the reference compressive
strength for the modulus of elasticity test, according to the requirements stated in ISO 1920-
10 [76]. This reference value was extremely important for the deteriorated concrete samples,
where the compressive strength cannot be estimated as precisely as for the sound specimens.
Beams 1300 mm in length were included to evaluate the influence of the cyclic loading and
subsequent impact loading. For this purposes the specimens were designed as over-reinforced
(616 grade B500 reinforcing steel), thus the failure by compressive-zone crushing should occur
and the fatigue failure of concrete can be assumed. The scheme of the dimensions and the specimen
reinforcement can be seen in Fig. 15.

32
 Experimental verification

Fig. 15 Reinforcement scheme of the specimens

Additionally the reinforced concrete specimens allowed to verify the effect of different ratios
between the area of the surface of the element exposed to an aggressive environment S and the
element volume V. The effect of the S/V ratio is explained in section 4.2.3. The geometric
properties of each tested specimen are summarized in Tab. 3.

Tab. 3 Geometry properties of the tested specimens

Specimen H B L S V S/V
[m] [m] [m] [m2] [dm3] [m-1]
Cube 0.15 0.15 0.15 0.135 3.375 40.0
Small beam 0.10 0.10 0.40 0.180 4.000 45.0
Beam 0.15 0.30 1.30 1.260 58.50 21.5

All sets were made of concrete class grade C25/30 with the mix design according to Tab. 4. Fine
aggregate consisted mainly of quartz and granitoid. Coarse aggregate was mix of spilite and tuffite.

Tab. 4 Mix proportions of the concrete mixture

Constituent kg/m3
CEM II 32.5 320
Sand (0-4) 836
Fine aggregate (4-8) 495
Coarse aggregate (8-16) 443
Water 185
w/c ratio [-] 0.58

4.1.2 Exposure class of the environments

The classification to the exposure classes should allow not only to determine the type of the
damage but also to evaluate its effect on the serviceability and to indicate the need for repair and
maintenance. The problems with the proper classification are described e.g. in [1, 2]. Today, the
set of European standards, namely EN 206 [20] should be used to classify the exposure class.

33
 Experimental verification

As was mentioned above, two types of sets were made; a) sets stored in a dry interior environment,
in accordance with EN 206 [20], defined as the XC1 exposure class, and b) sets stored in the
chemically aggressive environment. The aggressive environment was designed as a diluted
solution of 35% hydrochloric acid (HCl) with pH = 2. This corresponds with an H+ ion
concentration of cH = 0.01 mol/dm3. This set-up cannot be classified in accordance with the
European standards, which define a minimum of pH = 4 (cH = 10-4 mol/dm3) for the extreme
chemical aggressive environment XA3, see Tab. 4.

Tab. 5 Exposure classes for a chemical attack according to EN 206 [20]

Chemical properties XA1 XA2 XA3

Aggressiveness Weak Medium Strong
pH 5.5-6.5 4.5-5.5 4.0-4.5
H+ ion concentration 10-6.5-10-5.5 10-5.5-10-4.5 10-4.5-10-4.0
max. w/c 0.55 0.50 0.45
min. concrete grade C25/30 C30/37 C35/45
min. cem. vol. [kg/m3] 300 320 360
Water permeability [mm] 50 35 20

A significantly higher concentration of H+ ions was intentionally chosen to speed up the

deteriorative effects of the aggressive environment on the concrete specimens. For the general
concept, it should be noted that orange juice, black coffee and beer are examples of commonly
used products with a pH value of about 4. A set of specimens stored in the aggressive environment
is shown in Fig. 16.

Fig. 16 Set of specimens stored in the aggressive environment

34
 Experimental verification

4.1.3 Maintaining a stable pH value

During the neutralization, reaction of dissolved H+ ions with the concrete matrix, the pH value of
the aggressive environment increases constantly. To maintain a stable pH value, the aggressive
agent (HCl) was regularly replenished according to the pH measurements. The pH measurements
were performed with the VWR pH10 pen portable pH meter. The measured values were stored
together with the measured temperature for subsequent evaluation of the rate constant. The
required amount of acid, together with the amount of dissolved Ca2+ ions, were calculated
according to the ionic strength. This provides higher accuracy than a calculation dependent only
on the H+ ion concentration. The pH value was also measured after the aggressive agent had been
replenished, thus ensuring that the time change in the pH values can be monitored.
Comparison of the calculation methods for the first HCl addition can be seen in Fig. 17 and Fig.
18. The method based on the concentrations (Fig. 17) calculates the needed volume of the HCl on
the basis of the desired pH value, vice versa for the method based on the ionic strength (Fig. 18)
the pH value is dependent on the exact solution composition, thus the desired pH value is obtained
iteratively with the change in the volume of the added HCl. For the comparison the needed amount
of hydrochloric acid from the concentration calculation was entered into the ionic strength
calculation (green-tinged values). The difference in desired and calculated pH value is equal to
0.05 which corresponds to about 10% error in H+ ions concentration. The error of the calculation
based on the H+ ions concentration increases with the increasing content of dissolved Ca2+ ions in
the solution.

CONCENTRATION AND WEIGHT CALCULATION FROM SOLUTION pH (via concentration)

solution description: solution of hydrochloric acid + Ca(OH)2
initial pH: pHinit = 7.00 - final pH: pHfin = 2.00 -
neutralisation
initial concentration: c(OH-)= 0.000E+00 mol/dm3 acid type: monoprotic
concentration for neutralisation: c(H3O+)= 0.000E+00 mol/dm3 base type: diprotic
3
amount of added substance for neut.: n(H3O+)= 0.000E+00 mole for dm
change of solution acidity
final concentration: c(H3O+)= 1.000E-02 mol/dm3
amount needed: n(OH-)= 1.000E-02 mole for dm3
sum for dm3
total amount needed: n(H3O+)= 1.000E-02 mole for dm3 Molar mass: M(H3O+)=36.458 g/mol
3
total mass needed: m(H3O+)= 3.646E-01 g/dm
Total sum for solution
Volume of solution: V= 100.00 dm3 admixture concentration: 35%
total mass dissolved: mTotal = 36.46 g admixture density: adm= 1.18 g/cm3
admixture dissolved: madm= 104.17 g admixture volume: Vadm= 88.28 cm3
Fig. 17 Solution pH calculation based on the concentrations

35
 Experimental verification

CONCENTRATION AND WEIGHT CALCULATION FROM SOLUTION pH (via Ionic strength)

solution properties
Water volume: Vw = 99.912 dm3 Total vol. of solution: V= 100.000 dm3
admixture #1 volume: Vadm,1 = 0.00 cm3 admixture #2 volume: Vadm,2 = 88.28 cm3
admixture #1 concentr.: 100.00% admixture #2 concentr.: 35.00%
pure adm. #1 volume: Vp,adm,1 = 0.00 cm3 pure adm. #2 volume: Vp,adm,2 = 30.90 cm3
Molar mass of adm. #1 M adm,1 = 74.10 g/mol Molar mass of adm. #2 M adm,2 = 36.46 g/mol
density of adm. #1 adm,1 = 2.24 g/cm 3
density of adm. #2 adm,2 = 1.18 g/cm3
amount of adm. #1 n adm,1 = 0.00 mol amount of adm. #2 n adm,2 = 1.00 mol

ionic strength calculation (via molarity)

no. of ions concentration charge no. ionic strength
# admixture ion
[-] [mol/dm3 ] [-] [mol/dm3 ]
Ca2+ 1 2 0.00E+00
1 Ca(OH)2 0.00E+00
OH- 2 -1 0.00E+00
H+ 1 1 5.00E-03
2 HCl 1.00E-02
Cl- 1 -1 5.00E-03
Total ionic strength of solution 1.00E-02
negative ion type: OH- positive ion type: H+

activity coefficient: = 0.90 - ion concentration: c(H+)= 1.000E-02 mol/dm3

3
activity: a(H+)= 8.99E-03 mol/dm solution pH: pH= 2.05 -
Fig. 18 Solution pH calculation based on the ionic strength

With the ionic strength calculation, the amount of dissolved calcium ions can be determined
iteratively on the basis of the total amount of added hydrochloric acid and the measured pH value.
The calculated amount of dissolved calcium ions can subsequently be used for verifying the
proposed rate equation, and also for the proposed rate constant. All these calculations neglect the
influence of possible dissolution of other chemical elements, e.g. ferrites (Fe3+), aluminates (Al3+),
or silicates (Si4+).
To verify the correctness of not including these elements, samples of the aggressive environment
solution were chemically analysed to obtain the concentration of the considered ions. The content
of chlorides was determined using the Mohr method for argentometric determination of chlorides
according to ISO 9297 [77]. Atomic Absorption Spectrometry (AAS) according to ISO 8288 [78]
was used for the content of calcium and other elements. A comparison of the ionic strength based
calculation and the laboratory obtained values is shown in Tab. 6.

36
 Experimental verification

Tab. 6 A summary of calculated and laboratory obtained concentrations of Ca2+, Al3+, Fe3+, Si4+ ions

Set Added HCl Calculated Ca2+ Laboratory Ca2+ Laboratory Al3+ /

concentration concentration Fe3+ / Si4+ conc.
[cm3] [mg/dm3] [mg/dm3] [mg/dm3]
A1 1240 1384 1240 14.3 / 5.8 / 29.4
A2 1720 2215 2430 22.3 / 6.3 / 24.0
A3 1085 1588 1540 5.5 / 0.52 / 20.9
A4 1940 3959 3900 -
A5 3110 4375 4380 -
A6 4645 8685 8970 -
A7 6760 14177 13500 -

4.1.4 Layout of the fatigue testing

The arrangement of the cyclic loading is the four-point bending with span length 1000 mm and
overhangs of the length of 150 mm. This testing layout has been chosen for its several advantages
which are later discussed in the section 4.1.5. Experiments were conducted in the Experimental
Centre of Faculty of Civil Engineering, Czech Technical University in Prague. The arrangement
of the fatigue testing can be seen in Fig. 19.

Fig. 19 The arrangement of the fatigue testing

Parameters of the fatigue testing: loading frequency, maximum / minimum cyclic force and its
eccentricity, can be seen in Tab. 7. Due to high time demands of the experiments the expected
maximum number of applied cycles for each specimen is around 400 thousand cycles.

37
 Experimental verification

The deteriorated specimens were removed from the acidic environment three days before the
fatigue testing. The drying-out of the specimens had to be performed to prevent damage of the
testing equipment by the wet acidic surface of the specimen.

Tab. 7 Specification of the fatigue testing set-up

Loading Cyclic force

Eccentricity
frequency [kN]
[Hz] min max [m]
5 5 100 0.3

4.1.5 The material model point

The two basic types of testing layout were compared. For the evaluation of the crack propagation
and determination of the fracture energy or fracture toughness (material properties that are
commonly investigated in relation to fatigue of materials), the three-point bending layout with
notch is usually chosen.
From the perspective of the structural analysis, bending cracks are crucial for acceleration of the
deteriorative effects of the aggressive environment. Material point in the presented experiment is
therefore a zone subjected to pure bending (decomposed to compression / tension) without the
influence of shear. According to this approach, the four-point bending layout corresponds to the
mentioned assumptions. The localization zone between the loading forces is exposed to pure
bending. Due to the absence of a notch it can be assumed that the concentration of maximal normal
stresses occurs at the point of maximal material deterioration. This corresponds to the behaviour
of common structures and therefore it is used in the experimental program. The above described
principle is illustrated in Fig. 20.

Fig. 20 Four-point bending arrangement and localization zone principle

4.1.6 Deflection measurements

Two types of deflection measurements were performed during the fatigue testing. The first type -
static deflection measurements - takes place after every hour of cyclic loading (circa 18000 load
cycles). The second type - dynamic deflection measurement - was carried out during the fatigue
38
 Experimental verification

testing immediately after the static deflection measurements. In order to obtain the exact deflection
in the middle of the span, the settlement of the supports was also measured.
The deflections of the reinforced concrete specimens were measured by the linear variable
displacement transducer (LVDT).
Because of the time restrictions of the laboratories, additional measurements will be executed. One
after the end of the fatigue testing (3 pm), one before another testing on the following day (7 am).

4.1.7 The impact loading

Consequently after the exposure to the cyclic loading or to the combination of aggressive
environment and cyclic loading (according to the Tab. 2) the reinforced concrete beams were tested
with extreme loading conditions - impact loading.
Before the impact loading all the specimens were exposed to another cyclic loading for one hour.
This should remove the influence of self-healing effect of concrete during the different time
periods between the cyclic loading and impact loading. Specimen C2 failed during the additional
cyclic loading.
For the impact loading, a braced steel frame (Fig. 21) carrying drop weight was built in the
Experimental Centre of the Czech Technical University. Due to laboratory build-in crane, height
of the frame was limited to 4.1 m. Horizontal base of the construction is made of steel frame
anchored to the floor of the laboratory. Vertical stability is provided by bracing diagonals. The
burden is considered as a set of different weights to enable maximal variety of the drop-weight.

Fig. 21 Steel frame and the scheme of experiment

39
 Experimental verification

The tested concrete specimen is located in the biaxial centre of the horizontal frame. The sample
is supported on two parallel sides by reinforced steel girders which are anchored to the floor. To
ensure similar layout as in case of the cyclic loading two reinforcing bars were attached to the steel
girders. This ensured the span length of 1000 mm and simple supported beam behaviour. The
difference between the cyclic loading and impact loading is in the force position (four-point
bending for cyclic loading and three-point bending for impact loading).
The steel burden used during the experiment consisted of a central pole and a main cylindrical
hammer, which supports steel discs of different weights. The required weight was obtained by
combination of the steel discs. On the lower surface of the steel cylinder, a hemispherical tip
(R=20 mm) which concentrates the applied impact loading into one exact point is placed. The
burden used for the testing weights 75 kg.
The impact testing was supplemented with the Photon Doppler Velocimetry (PDV) which
measured the instantaneous velocity of the bottom surface of the tested specimen. These
measurements enable to evaluate instantaneous deflection and acceleration of the beam with high
precision. The impact zone (contact zone between burden and concrete specimen) was recorded
with the high-speed camera with 2500 fps (frames per second). The testing layout with the PDV
laser located below the specimen (in covering black-box) can be seen in Fig. 22.

Fig. 22 Experimental layout for the impact loading with the use of PDV

The wooden lath with 50 mm black and white stripes and the measuring points in Fig. 22 were
used as a scale for the evaluation of the data recorded with the high-speed camera.

40
 Experimental verification

4.2 Chemical deterioration

The experimental program dealt not only with the deterioration rate, described by the chemical
kinetics, but also with the change in the mechanical properties due to the deterioration. A dataset
was obtained for each set of specimens. Each dataset consists of compressive strength (cube and
small beam), modulus of elasticity (small beams) and, for the specimens stored in the aggressive
environment, the time protocol of the pH values and temperatures. The ambient temperature was
almost constant in time (ranging between 15C and 16C), and its influence on the deterioration
rate was therefore neglected.

4.2.1 Time change of the pH of the aggressive solution

The change in time of the pH was regularly measured to determine the kinetics of the chemical
reaction. An advantage of this pH time recording is that we could capture the change in the reaction
system from reaction-controlled to diffusion-controlled. This is caused by the transport of the
H+ ions through the increasing thickness of the corroded layer [60]. This change in the
deterioration system is accompanied by a decrease in the reaction rate, and therefore by a decrease
in the change in pH in time. Under normal laboratory conditions, when the reactants are added
separately (as liquids or as a powder), the change in pH in time will be constant.
The time dependence of the change in pH value is shown in Fig. 23. With the procedure described
here, the diffusion problem of the corroded layer with regard to its incremental increase in
thickness and the solubility of the emerging acid salts [3, 5, 6, 60] is already included in the
measured values.

Fig. 23 Measured and approximated pH development in time

For the purposes of subsequent calculations, the change in pH per day (pH()) was approximated
by the function in the following form:
41
 Experimental verification

23
pH ( ) = (4.2.1)
+ 17

Due to the dependence of the pH on the concentration of the H+ ions, the obtained function is valid
only for the particular case of stable pH 2. It can be assumed that for lower pH values the change
in the H+ ion concentration per day will be higher (the deterioration rate will be higher), but the
pH change per day will be smaller. This phenomenon results from the logarithmic nature of the
definition of pH in equation (3.1.3). This data is sufficient for determining the pseudo-rate
constant, because of subsequent implementation of the particular pH value. Integrity for other pH
values is therefore ensured.

4.2.2 Integration method verification

The rate constant and the system of the chemical reaction was simplified to the first order system.
If the proposed simplification is correct and can be used in the experimental program should be
verified on the obtained data set. As described in section 3.1.1 the change in Ca2+ ions
concentration should exhibit linear progression when the corresponding equation from (3.1.2) is
used. The collected pH data were converted to the Ca2+ ions concentration loss and used in
verification of the proposed first order system. The linear progression of the integrated form of the
rate equation can be seen in Fig. 24. Based on this verification the pseudo-rate constant was
evaluated for the first order system as described in section 3.1.2.

Fig. 24 In time development of the integrated form of rate equation (measured points, linear trend)

42
 Experimental verification

4.2.3 Evaluating the rate constant

The rate constant was determined from the function pH(), i.e. from the change in H+ ion
concentration in the aggressive solution. On the principle of an elementary chemical equation
(4.2.2), it is clear that the change in the concentration of calcium ions is one half of the
concentration of H+ ions. This applies for the neutralization equation (3.1.4).

vHCl cH = vCa(OH ) cCa (4.2.2)

2

According to these assumptions, the rate constant can be determined with the use of equation
(4.2.1) and the total amount of dissolved Ca2+ ions in time.
It is necessary to take into account the variable pH value. The initial pH value pH = pH0 changes
within the evaluation time by the value of pH(), according to equation (4.2.1), together with the
appropriate change in concentration according to equation (3.1.3). In the subsequent interval, the
initial pH is modified back to pH0, and there is another change by the value of pH(). The
evaluation interval was set to one day with respect to the current design lifetime of structures (50-
100 years) and with respect to the continuous system of the deterioration process.
As was mentioned in section 3.1, it can be assumed that due to the diffusion problems and due to
the transport of the aggressive media through the material matrix, the rate constant will be
a function of time. This phenomenon was confirmed by the decreasing change of pH in time,
Fig. 23. The function of the rate constant for neutralization (3.1.4) was obtained from the
experimental measurements, as follows.

1 0.15 dm3
=kc ( ) + mol day (4.2.3)
4.76 + 0.12 0.75

This rate constant has to be multiplied by the concentration of the H+ ions to obtain the pseudo-
rate constant, which can subsequently be substituted into function (3.1.7).
Calculated values of the calcium ion concentration in time are generally valid only for the
described experimental sets, and particularly for a specific volume and surface area of the
specimen. Conversely, the rate constant should be inherently dependent only on the type of
reaction.
The calculation of the change in calcium ion concentration for different specimens and structural
shapes was generalized on the basis of the S/V ratio (surface exposed to the aggressive environment
/ element volume). This approach is based on the assumption that for various kinds of elements of
various shapes and volumes, the total number of dissolved Ca2+ ions in time should be the same,

43
 Experimental verification

if the exposed surface is equal. While the area of the exposed surface is constant, the total number
of dissolved calcium ions is also constant, irrespective of the S/V ratio.
The calcium ion concentration calculation according to equation (3.1.7) is therefore modified for
an area of 1 m2, and is subsequently modified by the S/V value to obtain appropriate values for
a particular specimen or structural element. A modified form of equation (3.1.7) is stated in
equation (4.2.4).

S
cCa cCa ,0 exp ( kc ) 0.0382
= (4.2.4)
V

This method was verified numerically on the basis of the rate constant modified for different S/V
ratios. The initial calcium ion concentration cCa,0 was modified according to the change in volume.
The exposed surfaces were kept constant, so the total amount of dissolved calcium ions was also
constant. It is clear that with different values of cCa,0 it is necessary to change the development of
the exponential part in equation (3.1.7). This can only be done by changing the rate constants.
With these assumptions, the rate constants for different geometric shapes with different S/V ratios
were determined according to equation (3.1.2). The formulas dependent on S/V were extremely
comprehensive, and complicated the calculation significantly. For these reasons, the approach was
used only as a further verification of equation (4.2.4). A comparison of these two methods is shown
in Tab. 8. The cylinder, sphere and different size deck elements were used only in the verifying
calculations for their different S/V ratios (preserving constant surface area). These type of elements
were not tested experimentally.

Tab. 8 Verification of the principle stated in equation (4.2.4)

Specimen shape S/V S Number of Ca2+ Number of Ca2+ Agreement

(size [m]) ions eq. (4.2.4) ions verification
[-] [m2] [mol] [mol] [%]
cylinder (0.6x1.0) 6.494 89.331 95.981 93.1
sphere (D=0.785) 7.643 89.364 91.271 97.9
deck (H:B=1x2.5) 10.130 89.312 87.987 101.5
cube (0.57) 10.563 89.355 87.795 101.8
deck (1:5) 14.884 89.321 87.345 102.3
constant
deck (1:7.5) 19.918 89.303 92.041 97.0
deck (1:10) 25.000 89.367 88.654 100.8
exp. set 26.193 89.270 89.270 100.1
deck (1:15) 35.262 89.367 90.804 98.4
deck (1:20) 45.652 89.183 91.200 97.8

44
 Experimental verification

4.2.4 Weighted average of the pH of the aggressive environment

The calculation of the calcium ion concentration in time is dependent on the specific pH of the
aggressive environment. This also applies for both the rate constant and the pseudo-rate constant
derived in the previous section. The pH of the aggressive environment was measured and
maintained about 3 times per week (for a excerpt from the measurement report, see Tab. 9). It
cannot be asserted that the aggressive solution had stable pH = 2.

Tab. 9 Excerpt from the pH measurement report

Date Time Temp Initial pH HCl added Modified pH

[C] [-] [cm3] [-]
2013-12-04 15:50 15.5 2.45 85 2.00
2013-12-06 16:30 15.8 2.61 95 2.01
2013-12-09 8:50 15.7 2.68 110 1.98
2013-12-11 17:00 15.7 2.55 95 2.01
2013-12-14 15:00 15.6 2.70 115 1.99

For this reason, we proceeded to determine the substitute values of pH as a weighted average. For
each two consecutive measured intervals, the average concentration of the H+ ions is calculated
on the basis of the initial and final pH values (see Tab. 9). The calculated concentration of the H+
ions is subsequently multiplied by the length of the time interval between two consequent
replenishments of the aggressive agent. To obtain the weighted average, the total sum of the
multiplied values is divided by the total time of the measurements. This yields the average value
of the H+ ion concentration, which can be used for calculating the equivalent pH value. With this
procedure, even longer intervals between pH measurements can be taken into account. The
equivalent pH values for each set are listed in Tab. 10.

Tab. 10 Equivalent pH values of the aggressive environment

A1 A2 A3 A4 A5 A7 A8
2.113 2.124 2.461 2.392 2.289 2.178 2.238

4.2.5 Verifying the rate equation

To verify the proposed rate equation (3.1.6), the total number of dissolved calcium ions was
obtained in the laboratory. The experimentally obtained values were compared with the
calculations according to equation (3.1.7), with the use of the determined rate constant (4.2.3) and
the equivalent value of pH (described in section 4.2.4). The measured and calculated values are
compared in Tab. 11. Set A8 was not laboratory tested, but on the basis of the compliance of the

45
 Experimental verification

values obtained by calculation and in the laboratory, see Tab. 6, it can be expected that the
calculations using the total amount of added HCl and the ionic strength of the solution are
sufficiently precise.
The comparison showed good agreement between the calculated amount and the laboratory-
obtained amount of dissolved calcium ions. The maximum error of about 15% may be due to the
large fluctuations of the pH values. This applies especially at the early measurement times, where
the calculation of the equivalent pH value may be inaccurate. In a quasi-stable environment, the
equivalent pH value will stabilize, and this will reduce the error.

Tab. 11 A comparison of the calculated mass and the laboratory-obtained mass of dissolved Ca2+ ions

Set Time in the Calc. mass of Lab. obtained Agreement

acidic solution dissolved Ca2+ mass of Ca2+
[day] [g] [g] [%]
A1 7 126.827 121.160 104.7
A2 14 219.922 237.430 92.6
A3 34 208.192 246.197 84.6
A4 62 380.306 435.510 87.3
A5 90 618.026 678.114 91.1
A6 181 1020.243 1033.495 98.7
A7 363 1699.172 1516.923 112.0
A8 742 1971.837 1859.8061) 106.0
1) estimated mass calculated using a known amount of added HCl (ionic strength method)

4.2.6 Reduced mechanical properties of deteriorated concrete specimens

The compressive strength and the modulus of elasticity were determined within the experimental
program. The results of the material properties for the reference specimen sets and the deteriorated
specimen sets are shown in Fig. 25.

Fig. 25 Time development of compressive strength (left) and elastic modulus (right)
46
 Experimental verification

The exponential trend of the reduction of the material properties can be seen in the trend lines of
the deteriorated specimens (A sets). This trend corresponds with the change in pH (Fig. 23), and
therefore with the deterioration rate.
The reference values for the non-deteriorated material were determined from the sets marked C.
The Fret coefficient, which is necessary for calculating the reduced compressive strength
according to equation (3.1.10), was evaluated from the compressive strength tests. The Fret
coefficient is time-dependent and for the mix design used here, it was found in the form shown in
equation (4.2.5).

k f (=
) 8.8 0.977( 40) (4.2.5)

The compressive strength was tested on the cube and on the small beams after the modulus of
elasticity tests. In order to calculate the reduced compressive strength, it is necessary to determine
the amount of calcium ions dissolved for a particular specimen. This requirement arises from the
nature of the Fret equation, which indicates the volume fraction of the cement. The volume faction
of the cement is reduced by means of the rate equation. The ratio of the amount of released calcium
ions to the total mass differs for each element, depending on the S/V ratio. If the reduction of the
compressive strength is calculated for the total amount of dissolved Ca2+ ions and for the total
volume of one set, the S/V ratio will not correspond to the specimens tested for compressive
strength (cube and small beams), and the results will be underestimated (on the dangerous side).
The modified rate equation, as stated in (4.2.4) can be used for calculating the amount of dissolved
Ca2+ ions from the small beams (100x100x400 mm). For a sample calculation of the reduced
compressive strength of a beam specimen from the set A8 see Fig. 26.
A comparison of the calculated and experimentally-obtained compressive strength values is
presented in Tab. 12.

Tab. 12 Comparison of the calculated and measured reduced compressive strength values

Set Age Fret Reference fc Deteriorated Deteriorated Agreement

coefficient kf calculated fc measured fc calculated
[day] [-] [MPa] [MPa] [MPa] [%]
A1 35 7.677 30.6 25.2 29.7 118.1
A2 42 7.845 31.2 28.3 29.8 105.4
A3 77 8.377 33.4 31.6 31.9 101.1
A4 105 8.580 34.2 31.6 31.5 99.8
A5 133 8.685 34.6 34.0 30.3 89.1
A6 226 8.787 35.0 29.4 27.8 94.7
Table continue on next page
47
 Experimental verification

Set Age Fret Reference fc Deteriorated Deteriorated Agreement

coefficient kf calculated fc measured fc calculated
[day] [-] [MPa] [MPa] [MPa] [%]
A7 406 8.800 35.0 23.7 23.2 97.7
A8 777 8.800 35.0 22.1 21.1 95.5

The proposed approach for evaluating the compressive strength of the deteriorated specimens
showed good agreement with the experimentally obtained values. A significant difference can be
seen at an early age of the concrete (set A1), where the measured deteriorated compressive strength
was extremely low, despite the short exposure to the aggressive environment (7 days). The
measured compressive strength is doubtable and may be caused by poor compaction of beam
specimens in set A1. Reverse fluctuation can be seen for set A5, where the measured values are
almost identical with the reference values.

REDUCED COMPRESSIVE STRENGTH CALCULATION

mix design
cement density: CEM= 3050.0 kg/m3 calcium oxide density: CaO = 3316.0 kg/m3
cement content: mc= 320.0 kg/m3 amount of CaO in cement XCaO = 61.0 %
3
water content: mw = 185.0 kg/m cement properties based on its type
3
cement volume fraction: v c= 0.105 m air volume fraction: v a= 0.010 m3
water volume fraction: vw= 0.185 m3 initial Ca2+ concentration: cCa.0 = 3.784 mol/dm3
calculation of the Ca2+ ions concentration
specimen type: beam surface: S= 0.180 m2
surface/volume ratio: S/V= 45.000 m-1 volume: V= 4.000 dm3
initial Ca 2+
amount: n Ca= 15.138 mol time in acid: = 734.000 day
3
rate constant: kc= 0.046 dm /(mo weighted pH value: pH= 2.238 -
3 +
pseudo-rate constant: kc'= 2.68E-04 dm /(mo weighted H conc.: cH+= 0.006 mol/dm3
time Ca2+ concentration: cCa= 2.624 mol/dm3 reduced cem. vol. fract.: v c,red = 0.073 m3
strength calculation
age of concrete: 0 = 777.000 day Fret coefficient: kf= 8.800 day
cement resistance: fcm= 32.500 MPa Reduced comp. strength: fck= 21.107 MPa
Fig. 26 Calculation of the reduced compressive strength for small beam A8

4.2.7 Reduced cross-sectional area of specimen core

The verification of the proposed method is based on the assumption that the corroded layer does
not interact with the sound specimen core during the loading, thus only the reduced cross-section
is exposed to the stress.
The experimental methods described in the following section 4.3 provided the results which
confirmed the assumptions of the steady-state diffusion as described in 3.1.3.2. Thus the method
seems applicable for the structural assessment.

48
 Experimental verification

During the compressive strength tests the applied force is recorded and consequently the maximum
stress (compressive strength) is calculated based on the cross-sectional area of the specimen. In
case of the reference cube or beam specimens the cross-sectional areas are Acube = 0.0225 m2 and
Abeam = 0.01 m2 (dimensional deviations are neglected). For the calculation of the compressive
strength of the deteriorated specimens the same areas were used.
To verify the reduced cross-sectional area method the recorded forces from the compressive
strength testing should be recalculated using the reduced specimen area based on the total amount
of Ca2+ ions depleted. The obtained compressive strengths of the sound core should be similar to
the reference compressive strength of non-deteriorated specimens.
The reduced cross-sectional area is calculated using the results from Tab. 11 in section 4.2.5. and
the known mix-design (Tab. 4). For a sample calculation of the reduced cross-sectional area of
a beam specimen from the set A8 see Fig. 27.
A comparison of the reference compressive strengths and calculated strengths of sound cores of
the deteriorated beam specimens is presented in Tab. 13.

Tab. 13 Comparison of the reference and calculated compressive strength of sound specimen cores

Set Age fc,ref fc,det hbeam Fc,core fc,core

calculated acid - tests reduced calculated / fc,ref
[day] [MPa] [MPa] [mm] [MPa] [%]
A1 35 30.6 25.2 99.0 25.7 84.0
A2 42 31.2 28.3 98.3 29.3 93.9
A3 77 33.4 31.6 98.4 32.6 97.7
A4 105 34.2 31.6 97.0 33.6 98.1
A5 133 34.6 34.0 95.1 37.6 108.6
A6 226 35.0 29.4 91.8 34.9 99.6
A7 406 35.0 23.7 86.0 32.1 91.6
A8 777 35.0 22.1 83.5 31.7 90.5

The proposed approach for evaluating the reduced cross-sectional area of the deteriorated
specimens showed good agreement with the reference values. Significant errors can be observed
for sets A1 and A5. The explanation of these deviations is described in the previous section and
applies for this method as well.
The proposed approach for evaluating the reduced cross-sectional area of the deteriorated
specimens showed good agreement with the reference values. Significant errors can be observed
for sets A1 and A5. The explanation of these deviations is described in the previous section and
applies for this method as well.

49
 Experimental verification

REDUCED CROSS-SECTIONAL AREA CALCULATION

mix design
cement density: CEM= 3050.0 kg/m3 calcium oxide density: CaO = 3316.0 kg/m3
cement content: mc= 320.0 kg/m3 amount of CaO in cement XCaO = 61.0 %
2+ 3
initial Ca concentration: cCa.0 = 3.784 mol/dm cement properties based on its type
2+
calculation of the Ca ions concentration
specimen type: beam surface: S= 0.180 m2
cross-sectional area: A= 0.010 m2 volume: V= 4.000 dm3
initial Ca2+ amount: n Ca= 15.138 mol dissolved Ca2+ weight: mCa= 183.427 g
2+ 2+
initial Ca weight: mCa= 606.692 g dissolved / initial Ca d/i= 30.234 %
2+ 2
area without Ca ions: A w/o = 3023.393 mm corroded layer thickness: x= 8.237 mm
x = b 4 + h 4 + 1 4 b + 2bh + h 4 Aw/o
2 2

reduced height: h red = 83.526 mm reduced width: b red = 83.526 mm

reduced x-sectional area: Ared= 6976.607 mm2
core strength calculation
measured comp. strength: fck,mea= 22.100 MPa testing force: Fmeas.= 221.000 kN
Core comp. strength: fck,core= 31.677 MPa

Fig. 27 Calculation of the reduced cross-sectional area for small beam A8

4.3 SEM and EDS specimen analysis

The scanning electron microscope (SEM) and Energy-dispersive X-ray spectroscopy (EDS) were
used to verify certain assumptions used in the proposed deterioration model, e.g. the partially
neglected diffusion problem. The combination of an observational method (SEM) and an
analytical method (EDS) provides a better understanding of the deterioration phenomenon in
particular zones of deteriorated concrete.

4.3.1 Sample selection and preparation

The beam specimen (100x100x400 mm) from set A8 was selected for the SEM and EDS analyses.
The specimen was stored for two years in an acidic environment, see Tab. 2. After the modulus of
elasticity had been tested, the specimen was split into two pieces during the tensile bending
strength test, and one of the halves was selected for SEM testing.
To obtain a comprehensive set of data, the specimen was cut into several zones, which included
deteriorated parts as well as non-deteriorated (reference) parts, see Fig. 28.

50
 Experimental verification

Fig. 28 Zones of the specimen for SEM and EDS analysis (left) specimen cross-section, S-zone (right)

The S-zone represents the splitting surface of the specimen during the testing of tensile strength in
bending. The centre of the zone is in the middle of the element thus the acquired specimen should
not be influenced by the chemical deterioration. The R-zone is specified as the base of the beam,
thus the specimens should be the most deteriorated. The C-zone is partially equivalent to the R-
zone thus only one deteriorated specimen was prepared. The D-zone, about 30 mm thick stripe
through the middle of the specimen, was selected for the preparation of the thin-section specimens
(described below), which may include deteriorated layer, sound part and transition layer in-
between. In the macro-scale, the extremely deteriorated layer of concrete (brown-coloured) is
between 5 and 7 mm in thickness, see the specimen cross-section photo, Fig. 28.
Two different types of specimens were prepared for SEM and EDS analysis; a) polished-section
specimens (maximum diameter 25 mm) and b) thin-section specimens (25x40 mm). For a detailed
specification of the specimens, see Tab. 14.

Tab. 14 Specification of the SEM specimens

Zone No. Type of Deteriorated Note

specimen / reference
[-] [section] [MPa]
1 polished deteriorated transition zone
R
2 polished deteriorated surface view
1 polished deteriorated brown zone
S
2 polished reference centre part
1 thin deteriorated transition zone
D 2 thin deteriorated brown zone
3 thin deteriorated transition zone
C 1 polished deteriorated brown zone
F 1 polished deteriorated dust

51
 Experimental verification

Each specimen, with the exception of R2 (deteriorated surface view), S2 (reference specimen from
the centre of the beam) and F1 (dust mixed with sand from the brown-coloured layer) had
a deteriorated part as well as the sound part. Incoherence of the brown-coloured layer caused
problems when preparing the specimens, so some specimens have only the transition zone (the
brown layer fell off). Specimens R2 and F1 were prepared for the eventuality of indeterminate
results on the epoxy-coated deteriorated layer. Prepared specimens with carbon-coating are shown
in Fig. 29 (thin sections are turned upside down to show their structure).

Fig. 29 - Specimens prepared for the SEM

Before SEM and EDS analyses each specimen was magnified in the scanning electron microscope
to define suitable locations of image and atomic analysis. Selected locations can be seen in Fig.
30. Locations for SEM imaging and point or area EDS analysis are marked with red spots, line
EDS analysis are marked with blue lines. Map was created to determine the spatial distribution of
chemical elements for small zone of specimen D1. The mapping process is highly time-consuming,
thus only one attempt was made. The mapping zone is marked with green rectangle.

52
 Experimental verification

Fig. 30 Selected locations of SEM and EDS analyses for all tested specimens

4.3.2 Observational method (SEM)

The TESCAN VEGA 3 XMU scanning electron microscope belonging to the Institute of Geology
of the Czech Academy of Sciences was used to determine changes in the material matrix due to
chemical deterioration.
Image scanning was performed in high vacuum mode (cca 0.05 Pa) with 15-20 kV accelerating
voltage and detection of back-scattered electrons. Magnification between 1000x and 4000x was
selected for the evaluation, depending on the area that was being examined.
Three typical areas of interest were chosen; a) a brown-coloured layer with extreme deterioration,
b) the transition layer between the brown-coloured layer and the grey concrete, and c) the sound
concrete core. Three images from each typical area are shown in Fig. 31.

53
 Experimental verification

Fig. 31 SEM images of deteriorated layer (upper row), transition zone (middle row) and sound concrete
(bottom row)

The SEM images show extreme changes in the concrete matrix. The brown-coloured deteriorated
layer consists mainly of very fine gravel and sand, which is glued rather than bonded together with
residual cement paste. It can be assumed that the content of calcium ions will be very low.
The transition zone is much more compact than the deteriorated layer. The large number of visible
voids in upper part of the SEM images (dark parts) indicate a small amount of calcium ions. The
transition zone is hardly detectable, and was evaluated as the end of the brown-coloured layer. The
transition zone is between 100 and 500 m in thickness, as was confirmed by EDS analysis.
The sound core of the concrete specimens is continuously filled with cement paste. The small
visible failures may have been caused by mechanical testing and by specimen preparation (sharp-
edged cracks), or by early age shrinkage (smooth-edged cracks).

54
 Experimental verification

The cement paste in the SEM analysis can be observed as continuously spread surface around the
aggregate particles. This is due to the resolution of the SEM. In order to examine the cement paste
as single grains, as in [79], it would be necessary to use a different type of diode, e.g. a Schottky
diode.

4.3.3 Analytical method (EDS)

EDS was used to determine the composition of the material matrix (sound and deteriorated). For
this purpose, the TESCAN VEGA 3 XMU was equipped with the BRUKER QUANTAX EDS.
Different types of analysis were used for the evaluation; a) point evaluation for initial identification
of the cement paste and aggregates, b) area evaluation to determine the average composition of
a typical zone (as mentioned above), and c) line evaluation to obtain a continuous composition
curve.
Typical average EDS spectra obtained by the area evaluation for the deteriorated brown-coloured
layer and for sound concrete are shown in Fig. 32.

Fig. 32 Typical EDS spectra for the brown-coloured layer (top) and sound concrete (bottom)

55
 Experimental verification

The comparison of the EDS spectra corresponds with the assumption, stated in section 4.1.3 and
[3], that the reaction of the acid with compounds of iron, aluminium and silicon is negligible. In
the whole data set, no considerable variability in the amount of these elements was found.
The discrete evaluation in particular points and areas showed that the trend of the content of
selected ions (Ca2+, Si4+, Al3+, Fe3+ and Cl-) was dependent on the depth of the specimen. The
significant difference in the content of Ca2+ ions at the end of the deterioration layer corresponds
with the assumption that the transition zone between the deteriorated layer and sound concrete is
very thin. This phenomenon is shown in Fig. 33. All other compounds do not change significantly.

Fig. 33 Atomic composition of the deteriorated concrete with depth dependence (discrete points evaluation)

The reason for the change in colour of the deteriorated layer was searched. No significant increase
of the Fe3+ ions content in deteriorated layer was observed, which partially corresponds with [8].
The change in colour may be caused by the change in atomic percent ratio between Ca2+ and Fe3+
ions resulting in the bigger influence of ferric ions. This phenomenon is accompanied by the
decrease in Ca2+ ions and consequent disruption of the cement paste, which cover the fine
aggregate, thus the brown colour of the fine aggregate (sand) will reveal. There was no increased
content of Fe3+ ions around the fine aggregate particles which corresponds with its composition
(quartz). The coarse aggregate did not show any surface damage, therefore any possible content
of Fe3+ ions in spilite should not influence the obtained results.
For more accurate results, the discrete evaluation was supplemented by a continuous line analysis.
These analyses were performed at least twice in the designated area to provide a reliable dataset.
A continuous dataset was obtained for several specimens (S1, D2, D3 and C1), with comparable
results. Specimen D2 contains all three typical areas, and was therefore selected for the
presentation of the results. The development of the atomic content of Ca2+ (in relation to other
investigated elements Si4+, Al3+, Cl- and Fe3+) along the depth of the element is shown in Fig. 34.

56
 Experimental verification

Fig. 34 Atomic percent of Ca2+ ions along the specimen depth (evaluated lines specified on the SEM image)

The discontinuity of the dataset is caused by the presence of big aggregates, which were excluded
from the evaluation. The development of the remaining elements (Si4+, Al3+, Cl- and Fe3+) can be
seen in Fig. 35 Fig. 38.

Fig. 35 Atomic percent of Si4+ ions along the specimen depth (range as in Fig. 34)

Fig. 36 Atomic percent of Al3+ ions along the specimen depth (range as in Fig. 34)

57
 Experimental verification

Fig. 37 Atomic percent of Cl- ions along the specimen depth (range as in Fig. 34)

Fig. 38 Atomic percent of Fe3+ ions along the specimen depth (range as in Fig. 34)

In Fig. 35 Fig. 38 the ions (Si4+, Al3+, Cl- and Fe3+) exhibit slightly higher values in the
deteriorated layer than in sound concrete. This is caused by the evaluation using the atomic percent,
which sum has to be 100%, thus if the Ca2+ ions are missing, the atomic percent content of each
remaining ions should be higher. In mass percent the increase is maximally up to 1.5% for all the
mentioned ions.
Due to the specimen transport which preceded the thin-section preparation part of the deteriorated
layer (around 2 mm) detached from the specimen. The reduced thickness was measured with
Vernier caliper. The EDS analysis proved that the brown-coloured deteriorated layer was 4 to
5 mm in thickness which corresponds with the caliper measurements. The transition zone exhibits
a steep increase in the amount of Ca2+ ions in the material matrix, and is difficult to distinguish,
similarly as in the visual observation of the SEM images. On the basis of these findings, it can be
stated that the diffusion through the sound concrete is negligible, and that the chemical reaction in
the transition zone is mainly reaction-controlled. The diffusion problem has an effect only on the
transport of H+ ions through the deteriorated layer.

58
 Experimental verification

4.4 Cyclic loading

As described in section 4.1.1, in total 13 specimen sets were cast for the fatigue testing: 5 sets of
a non-deteriorated dry specimen and 8 sets of deteriorated specimen which was taken out of the
acidic environment three days before the fatigue testing.
For the verification of the proposed extension of the fatigue damage function (section 3.2.1) and
the influence of the concrete deterioration on the deflection increase due to the cyclic loading
(section 3.2.1.3) the deflection measurements were recorded regularly.
In case of the specimens stored in aggressive environment both methods for the assessment of the
deteriorated specimens (sections 3.1.3.1 and 3.1.3.2) were used to compare their suitability for the
deflection increase calculation as it was proposed.
Based on the selected method for the deteriorated specimens some additional properties had to be
calculated before the calculation: a) compressive strength of the cyclic beam which differs from
the directly measured values on the small beams as described in section 4.2.6; b) reference
compressive strength on the basis of mix-design and Fret coefficient (equation (4.2.5)) and
reduced cross-sectional dimensions based on the total amount of depleted Ca2+ ions as described
in section 4.2.7. These additional properties are listed in Tab. 15.

Tab. 15 Calculated compressive strength and reduced cross-section dimensions for cyclic beams

Set fc,ref fc,cyclic hbcyclic bcyclic

calculated calculated reduced reduced
[MPa] [MPa] [mm] [MPa]
A1 30.6 30.2 149.1 299.1
A2 31.2 30.5 148.4 298.4
A3 33.4 32.7 148.5 298.5
A4 34.2 32.9 147.2 297.2
A5 34.6 32.5 145.4 295.4
A6 35.0 31.5 142.4 292.4
A7 35.0 29.3 137.2 287.2
A8 35.0 28.2 135.0 285.0

In addition to verify the extended fatigue damage function method on deteriorated and non-
deteriorated specimens the influence of the loading speed and the fatigue age of the specimens was
examined for the cyclic loading tests. This phenomenon is further described in section 3.3.1.

59
 Experimental verification

4.4.1 Sensitivity analysis of the number of layers

For the calculation using the partial integration over height the compressive zone of the tested
specimens was divided into 20 layers (in the case of 300x150x1300 specimens, each layer has
around 3 mm in thickness). This division was based on a sensitivity analysis.
Using coarse division, up to 5 layers, the small average stresses in the wide integration layers leads
to results on the unsafe side (in contrast with the standardized approaches). The calculated
deflections did not show good agreement with the measured values, and were distinctly lower than
the measured values.
With a finer division, up to 15 layers, the average stresses in the layers are higher (and correspond
more to the real stresses). Thus some of the layers may exceed the limit number of applied cycles,
which leads to a significant reduction in the modulus of elasticity (see below). The division is still
coarse, and the height of the layer represents a considerable part of the cross-section. This
significantly influences the ideal moment of inertia of the cross section, which may lead to very
conservative results.
Dividing the compressed zone of the ideal cross-section into 20 layers proved to be the optimal
solution for these experimental settings. Finer division into more than 20 layers is not necessary,
as the difference in the results is negligible in comparison with the possible increase in
computational time requirements.

4.4.2 Fatigue testing of reference C sets

For the first set C1 (age 28 days) the evaluated mechanical properties of the specimens are
following: modulus of elasticity Ec,init = 29.7 GPa, compressive strength fc = 29.1 MPa, flexural
strength fctm = 3.9 MPa and density = 2257 kg/m3. Comparison of the measured values of static
deflections and calculated deflections according to the extended method for the fatigue damage
function as described in section 3.2.1 can be seen in Fig. 39. Specimen behaviour shows good
agreement with the calculated data except for the two first measurements which were erroneous.

Fig. 39 Comparison of measured and calculated deflections for set C1

60
 Experimental verification

The second set C2 was stored in dry condition and it was tested at the age of 60 days. The
determined mechanical properties of the specimens are following: modulus of elasticity
Ec,init = 30.0 GPa, compressive strength fc = 34.0 MPa, flexural strength fctm = 4.0 MPa and density
= 2206 kg/m3. During the fatigue testing, the crack in the direction of principal tensile stress
appeared. This crack influenced the deflection measurements, as can be seen in Fig. 40, between
110 thousand cycles and 230 thousand cycles. The increase in the deflection was about 0.4 mm.
This increase was included into the calculated deflection with an instant increase. With this
inclusion, the specimen behaviour corresponds to the fatigue damage function, especially in the
post crack domain.

Fig. 40 Comparison of measured and calculated deflections for set C2

The third specimen C3 was stored in dry condition and it was tested at the age of 182 days. The
evaluated mechanical properties of the specimens are following: modulus of elasticity
Ec,init = 30.1 GPa, compressive strength fc = 34.2 MPa, flexural strength fctm = 4.2 MPa and density
= 2221 kg/m3. The comparison between measured and calculated deflection can be seen in Fig.
41. The measured deflections show gradually decreasing rise which asymptote is probably a little
lower than for the calculated values.

Fig. 41 Comparison of measured and calculated deflections for set C3

The fourth specimen set C4 was stored in dry condition and was tested in the age of 364 days. The
specimen subjected to cyclic loading collapsed after a few thousand cycles. The results for the

61
 Experimental verification

effect of the fatigue were excluded from the evaluation. The specimen collapsed in direction of the
maximum principal stress (tensile), due to shear. The collapsed specimen can be seen in Fig. 42.

Fig. 42 Cyclic beam C4 after collapse during the fatigue testing

The fifth specimen set C5 was stored in dry condition and it was tested at the age of 728 days. The
specimen collapsed in direction of maximum principal stress similarly to the specimen C4. In case
of specimen C5 wide shear crack appeared already during the first static loading, Fig. 43.

Fig. 43 Shear crack after first static loading of specimen C5

The comparison between the calculated and measured values for the reference specimens suffers
from the lack of data due to the collapse of specimens C4 and C5 together with the small amount
of measured data-points in case of specimen C1. Still the agreement is satisfying and will be further
discussed and analysed for the deteriorated specimens A1 A8.
The cause of the specimens collapse should be further discussed. The deflection increase due to
cyclic loading diminishes with higher age of the specimens. The differences are negligible for the
tested specimens. This complies with the concrete time-dependent behaviour when the material
properties (especially modulus of elasticity and compressive strength) increases with time. With

62
 Experimental verification

the increasing age the specimens became brittle due to the dry conditioning. This corresponds e.g.
with the behaviour of the plastics which become brittle due to the age and exposure to the weather.
Considering the fact that the deteriorated specimens did not collapsed (as can be seen in next
section) the cause of the failure can be attributed to the drying-out process of the specimens. That
this phenomenon had such significant influence on the specimens behaviour may indicate that it
was a problem of the badly used mix-design (Tab. 4).

4.4.3 Fatigue testing of deteriorated A sets

The first evaluated set of specimens stored in aggressive solution of hydrochloric acid (A1) was
tested after 7 days of storing (age of 35 days). The evaluated mechanical properties of the
specimens are following: modulus of elasticity Ec,init = 30.3 GPa, measured compressive strength
on small beams fc,beam = 25.2 MPa, calculated compressive strength for cyclic beam fc,cyclic = 30.2
MPa, flexural strength fctm = 3.2 MPa and density = 2311 kg/m3. Similarly as in case of set C2,
the crack in the direction of the principal stress occurred during the fatigue testing. The increase
in the deflection was about 0.5 mm. The measured values are decreasing at the end of the testing
which may be caused due to the measurement error. Comparison of the measured and calculated
deflections can be seen in Fig. 44.

Fig. 44 Comparison of measured and calculated deflections for set A1

The second set of deteriorated specimens A2 was tested after 14 days in aggressive solution (age
of 42 days). The evaluated mechanical properties of the specimens are following: modulus of
elasticity Ec,init = 31.7 GPa, measured compressive strength on small beams fc = 28.3 MPa,
calculated compressive strength for cyclic beam fc,cyclic = 30.5 MPa, flexural strength fctm = 4.8 MPa
and density = 2311 kg/m3. Similarly as in cases of set C2 and A1 a crack in the direction of the
principal stress occurred during the fatigue testing. The increase in the deflection due to the crack
development cannot be properly quantified due to the lack of data at the beginning of the testing.
The shear crack influence is therefore not considered in the calculation, thus the calculated values

63
 Experimental verification

are underestimated but the behaviour pattern shows good agreement with the measured data.
Comparison of the measured and calculated deflections can be seen in Fig. 45.

Fig. 45 Comparison of measured and calculated deflections for set A2

The third set of deteriorated specimens A3 was tested after 34 days in aggressive solution (age of
77 days). The evaluated mechanical properties of the specimens are following: modulus of
elasticity Ec,init = 32.4 GPa, measured compressive strength on small beams fc = 31.6 MPa,
calculated compressive strength for cyclic beam fc,cyclic = 32.7 MPa, flexural strength fctm = 3.5
MPa and density = 2262 kg/m3. Comparison of the measured and calculated values can be seen
in Fig. 46. The measurements and calculations show good agreement.

Fig. 46 Comparison of measured and calculated deflections for set A3

The specimen set A4 was tested after 62 days in aggressive solution (age of 105 days). The
evaluated mechanical properties of the specimens are following: modulus of elasticity
Ec,init = 31.8 GPa, measured compressive strength on small beams fc = 31.6 MPa, calculated
compressive strength for cyclic beam fc,cyclic = 32.9 MPa, flexural strength fctm = 4.6 MPa and
density = 2246 kg/m3. The comparison of the measured deflections and the calculated deflections
shows overestimation on side of the fatigue damage function. Still the behaviour pattern
corresponds. Possible explanation for this behaviour may be due to the higher modulus of elasticity
of cyclic specimens than that evaluated from small beams. This can be supported by Fig. 47,
particularly on the first measured and calculated deflections (no. of cycles ni = 0), which should be

64
 Experimental verification

similar since there is no influence of fatigue. Comparison of the measured and calculated
deflections can be seen in Fig. 47.

Fig. 47 Comparison of measured and calculated deflections for set A4

The specimen set A5 was tested after 90 days in aggressive solution (age of 133 days). The
evaluated mechanical properties of the specimens are following: modulus of elasticity
Ec,init = 34.1 GPa, measured compressive strength on small beams fc = 34.0 MPa, calculated
compressive strength for cyclic beam fc,cyclic = 32.5 MPa, flexural strength fctm = 3.7 MPa and
density = 2300 kg/m3. Comparison of the first measured and calculated deflections (no. of cycles
ni = 0) shows good agreement. The calculated increase in the deflection due to the cyclic loading
is almost identical to the calculated values. Comparison of the measured and calculated deflections
can be seen in Fig. 48.

Fig. 48 Comparison of measured and calculated deflections for set A5

The sixth set of deteriorated specimens A6 was tested after 181 days in aggressive solution (age
of 224 days). The evaluated mechanical properties of the specimens are following: modulus of
elasticity Ec,init = 31.9 GPa, measured compressive strength on small beams fc = 29.4 MPa,
calculated compressive strength for cyclic beam fc,cyclic = 31.5 MPa, flexural strength fctm = 3.9 MPa
and density = 2238 kg/m3. The overestimation of the calculated deflection between 0-150
thousand cycles is possibly caused by the combination of two factors: 1) different initial modulus
of the cyclic beam and 2) usage of a completely cracked cross-section in the calculation of
65
 Experimental verification

deflections. The specimen showed progressive crack development over the cycles. Comparison of
the measured and calculated deflections is on Fig. 49.

Fig. 49 Comparison of measured and calculated deflections for set A6

The seventh set of deteriorated specimens A7 was tested after 363 days in aggressive solution (age
of 406 days). The evaluated mechanical properties of the specimens are following: modulus of
elasticity Ec,init = 31.0 GPa, measured compressive strength on small beams fc = 28.0 MPa,
calculated compressive strength for cyclic beam fc,cyclic = 29.3 MPa, flexural strength
fctm = 3.4 MPa and density = 2245 kg/m3. The measured and calculated data show good
agreement. Comparison of the measured and calculated deflections can be seen in Fig. 50.

Fig. 50 Comparison of measured and calculated deflections for set A7

The specimen set A8 was tested after 734 days in aggressive solution (age of 777 days). The
evaluated mechanical properties of the specimens are following: modulus of elasticity
Ec,init = 28.5 GPa, measured compressive strength on small beams fc = 22.1 MPa, calculated
compressive strength for cyclic beam fc,cyclic = 28.2 MPa, flexural strength fctm = 3.0 MPa and
density = 2192 kg/m3. The deflection increase shows good agreement with the calculated values.
Comparison of the measured and calculated deflections can be seen in Fig. 51.

66
 Experimental verification

Fig. 51 Comparison of measured and calculated deflections for set A8

As can be seen in Fig. 44 Fig. 51 the deflections calculated using the reduced cross-sectional
area are getting extremely overestimated with the higher deterioration degree. With reduced
dimensions the cross-sectional moment of inertia is reduced, thus the calculated stresses in
evaluated layers are higher causing its faster deterioration of modulus of elasticity. This is
multiplied due to the deflections dependency on the ideal moment of inertia which is dependent
on the modulus of elasticity of each layer as well. The reduction of the moment of inertia is
between 2.2% for set A1 and 31.8% for specimen A8 with comparison to the reference sets C.
That the reduced cross-section area method is not valid for the deflection increase calculation or
for the stresses evaluation (service limit states SLS) does not mean that it is not suitable for the
ultimate limit states (ULS) assessment. With the assumption that during the structural element
failure the materials are at their yield stress (in simplified way) it can be expected that the corroded
layer (with almost none concentration of Ca2+ ions) does not have sufficient cohesion with the
element sound core and does not contributes to the element mechanical resistance. Thus the author
suppose that the reduced cross-sectional area method can be used as an analogy to the concrete
fire design in [54].

4.4.4 Influence of the strain rate

The cyclic loading has a frequency of 5 Hz as described in section 4.1.4. With higher loading
speed, respectively with the higher strain rates, the increase in the mechanical properties can be
observed. This phenomenon is commonly referred in the scientific literature as dynamic increase
factor (DIF) as described in section 3.3.1. With the used cyclic loading speed the DIF can be
assumed. For this reason two types of the deflection measurements (static and dynamic) were
conducted as described in section 4.1.6. With this data set the influence of the strain rates can be
evaluated. The measured deflections for all the tested specimens are listed in Tab. 16.
The increase in instantaneous material characteristics with higher strain rate should result in the
reduction of the dynamic deflections (in comparison with the static deflections). This largely
corresponds with the measured data in Tab. 16.

67
 Experimental verification

Tab. 16 Comparison of static and dynamic deflection measurements

68
 Experimental verification

Numerical comparison shows that the maximum deflection reduction is up to 10% (specimen C1).
The common extreme for all the specimens is around 7% (the dynamic deflection is 93% of the
static deflection). The average reduction is about 2% (the average dynamic deflection
measurement is 98% of the static deflection measurement).
To verify if the approach using the fatigue damage function can be applied to the dynamic
deflections calculation the influence of the DIF was incorporated into the calculation using the
change of the modulus of elasticity. The compressive strength increase was omitted due to the
assumption that the cyclic deterioration of the concrete layers is caused by the microstructural
damage, thus it is influenced by the extreme stresses at the tips of the cracks and micro-cracks
where the slight increase of the compressive strength does not take effect. This damage
corresponds to the time-dependent strain increase as described in [37].
For each numerical layer of the cross-section the strain rate was evaluated as described in section
3.3.1. Subsequently the modulus of elasticity of each layer was multiplied with DIF value
determined on the basis of the strain rates according to the equation (3.3.3). With modified values
of modulus of elasticity the ideal moment of inertia is calculated.
For the verification, values calculated for the static deflection were compared with the values
calculated for the dynamic deflection (with the modification described above). With regard to the
fact that the strain rates should be equal for all the tested specimens the ratio between the calculated
static and dynamic deflections should not vary. Therefore the verification was made graphically
for the specimens A3, A5 and A7 which showed stable agreement with the calculated data in
previous section. The result can be seen in Fig. 52 Fig. 54.

Fig. 52 Static and dynamic deflections with the calculated trend curves for specimen A3

69
 Experimental verification

Fig. 53 Static and dynamic deflections with the calculated trend curves for specimen A5

Fig. 54 Static and dynamic deflections with the calculated trend curves for specimen A7

As can be seen in Fig. 52 Fig. 54 there is good agreement between the calculated dynamic
deflections and the measured values. The error is up to 10% in maximum. With the proposed
calculation method the difference between the evaluated values of static deflection and dynamic
deflection slightly increases with number of applied cycles (ratio is around 97.1% at the ni = 0 and
around 96.8% when ni = 450 000 cycles). This phenomenon cannot be observed in the measured
values. For specimens C1, A1, A3, A4, A7 and A8 the ratio is higher with increasing ni and vice
versa for the specimens C2, C3, A2, A5, A6.
There is no significant trend which can be attributed to the deterioration due to aggressive
environment. The increase in mechanical properties appears to be the same for the applied strain
rates.

4.4.5 Evaluation of the remaining fatigue life-time

The remaining fatigue life-time was calculated for all tested specimens according to the
equation (3.2.5). Due to the high time demands of the cyclic loading tests, the specimens cannot
be tested until their fatigue failure, thus there is no measured data to modify the used constants and
variables from (3.2.6) which were fitted only on the basis of the large data-set obtained from [37].
The verification of the proposed approach is therefore made as a comparison with Model Code
2010 [66]. For this purpose the predicted fatigue age of the specimens was calculated on the basis
of the deflection increase (approach to verify, described in section 3.2.2) and on the basis of the

70
 Experimental verification

ratio between number of applied cycles ni (from the fatigue testing) and limit number of applied
cycles N (calculated according to [66], equation (3.2.3)). Results of both methods are listed in Tab.
17.

Tab. 17 Age of specimens with regards to the deflection increase, compared with Model Code 2010

Set fc,cyclic init 400k init / 400k *Einit Lifetime MC2010

[MPa] [mm] [mm] [-] [-] [%] [%]
C1 29.1 1.3801) 1.430 1.036 0.965 0.4 37.2
C2 34.0 1.343 1.4562) 1.084 0.922 1.6 1.9
C3 34.2 1.091 1.317 1.207 0.828 5.7 1.7
A1 30.2 1.211 1.3422) 1.108 0.902 2.5 17.9
A2 30.6 1.238 1.808 1.460 0.685 77.2 13.9
A3 32.7 1.088 1.438 1.322 0.757 15.5 3.9
A4 32.9 0.985 1.278 1.297 0.771 12.1 3.4
A5 32.5 0.925 1.106 1.196 0.836 5.5 4.2
A6 31.5 0.949 1.368 1.442 0.694 87.2 7.5
A7 29.3 1.193 1.409 1.181 0.847 5.4 32.0
A8 28.2 1.183 1.586 1.341 0.746 35.3 72.7
1) calculated deflection due to the erroneous first measurements
2) the shear crack influence is neglected in the final deflections

As can be seen the calculated data differs quite significantly for all the tested specimens. Extreme
calculated values based on the deflection measurements occur for the specimen A2 and A6. This
is due to the higher deflection increase caused by the cracks development. For these two specimens
the crack contribution to deflections cannot be quantified properly from the measured data, thus
the fatigue age of the specimens was calculated including the crack influence. Where the effect of
the shear cracks could be removed the fatigue age was calculated without its influence (specimens
C2 and A2).
On the values calculated according to the Model Code 2010 [66] significant influence of the
specimens compressive strength can be observed. This phenomenon occurs even in the calculated
values on the basis of equation (3.2.5) but in lower extent (specimen A8).
The values calculated according to the proposed approach and to the standard are not in great
agreement, but they are comparable. The main advantages of the proposed approach are its
independence on the number of applied cycles ni which is usually hard to determine for the
structures of transport infrastructure, respectively its independence on the limit number of applied
cycles N which is usually standardized value and varies significantly for different standards.

71
 Experimental verification

The proposed approach or its modification can be used for the preliminary estimation of the fatigue
age or remaining life-time of the structures.

4.5 Impact loading

As mentioned in section 4.1.7 the specimens were exposed to the additional cyclic loading before
the impact testing. The specimen C2 failed during this testing.
The impact loading was carried out on two dates. The first round (specimens R1, C3 and A7) was
performed on 5 November 2014 together with the impact loading tests for another experimental
program with slightly different setup (10000 fps without PDV). The second round (specimens C1,
A1, A2, A3, A4, A5, A6 and A8) was performed on 22 October 2015 with the setup described in
section 4.1.7.
Problems with insufficient accuracy of the camera recordings occurred during the evaluation of
the first round (described further). In addition, due to a trigger error the impact test was not
recorded for the specimen A7. Therefore the first round of testing was not evaluated (except of the
specimen R1 which is the only reference specimen). For the second round the testing setup was
modified to remove the identified problems.

4.5.1 Procedures for the measured data evaluation

For the evaluation of the specimen resistance to the impact loading the energy dissipated in the
deflection of the beams was chosen. This energy can be determined as a work of the burden on the
beam deflections, thus as the area under the force-displacement curve in analogy to the equation
(3.3.8), see equation (4.5.1).

res ,beam

=Wimp Fburden ( t ) d beam (4.5.1)

0,beam

The deflections of the specimen can be evaluated with two different methods. When only the
camera recording is available the development of the specimen deflections can be obtained from
the sequence of consecutive frames involving duration of the impact loading. When the Photon
Doppler Velocimetry is used during the impact loading the specimen deflections can be easily
calculated from the PDV velocity results as an integral. Firstly, the signal recorded from the laser
measurements needs to be converted with Fast Fourier Transformation to obtain the velocities.
The main problem of the described testing setup is the absence of some additional measurements,
in particular, acting force or burden velocity or at least reactions in supports. While these data are
not available, it is possible to use the camera records as a substitute. Therefore the video sequence
from the impact loading was used to determine the development of the loading force.
72
 Experimental verification

The time-development of the acting force has to be determined to substitute in equation (4.5.1).
The force evaluation is derived from the Newtons second law of motion (equation (4.5.2)).

dv ( t )
Fburden ( t ) = mburden = mburden a ( t ) = mburdenuburden ( t ) (4.5.2)
dt

At any time t the velocity (equation (4.5.3)) and acceleration (equation (4.5.4)) can be calculated
from the burden position uburden on at least two consequent frames.

duburden ( t )
uburden ( t ) = (4.5.3)
dt
duburden ( t ) d 2uburden ( t )
uburden ( t ) =
= (4.5.4)
dt dt 2

The accuracy of these evaluations are dependent on the frames resolution which decreases with
the higher frame rate (fps). Vice versa with bigger time gap between the consecutive frames the
error of the visual method can increase as well. In the experiment 2500 fps were used, as described
in section 4.1.7, as an optimum between the frames resolution and the frame rate. This is based on
the problems during the evaluation of the first round of experiment (10000 fps were used) where
one pixel correspond to approximately one millimetre.
Another problem occurred during the evaluation of the camera recordings. As the process is
a visual method the human factor is presented, thus some error can be expected. This error was
not visible in the deflection evaluation, where low values are summed into relatively high
deflection but when the numerical derivative process was executed the scatter in the results due to
the error in the visual analysis was clear. To remove this error two methods were tested.
The first method was smoothing of the signal obtained from the visual analysis. The smoothing
was carried out with the method for signal processing, the Savitzky-Golay filter with 16 points of
window and 4th polynomial order [80]. The second method was based on statistics. The visual
analysis was performed at least ten times with different observational points. Consequently the
maximum, minimum and average values can be acquired.
Both methods had similar results for the average. The second (statistical) method presented
additional interval between relative minimum and maximum values. Based on these values the
resistance of the beam specimens to impact loading can be stated. The sample of the data set
obtained by the statistical method can be seen in Fig. 55.

73
 Experimental verification

Fig. 55 Statistical evaluated velocity of the burden impact testing of specimen C1

With regard to the possible errors in the visual evaluation the exact beam deflection measurements
from the PDV system were taken as a basis for the FEM modelling of the impact loading. The
model in LS-DYNA software, described in section 3.3.2, was fitted according to these beam
deflections through the change in the concrete material fracture energy.

4.5.2 Fracture energy fitting in concrete material - MAT159

The fitting of the model was based on the change in the fracture energy. The material model
MAT159 consists of three different types of fracture energies as described in section 3.3.2.6 (in
uniaxial stress Gfc, in uniaxial tension Gft and in pure shear stress Gfs).
Primarily, the model was modified only by changing the compressive fracture energy Gfc. Only
when the agreement between the results from the LS-DYNA and PDV was not sufficient the
modification of other fracture energies were made. Otherwise the default value of tensile and shear
fracture energy (Gft = Gfs = 0.01Gfc) was used.
For the first fitting the non-deteriorated set C1 with data from PDV was selected. The initial
compressive fracture energy was considered according to the Model Code 2010 [66], equation
(3.3.6). The calculated deflection from this settings were slightly overestimated in comparison
with the PDV dataset. For this reason the increased compressive strength due to the strain rates, as
described in section 3.3.1, was used for the fracture energy calculation. Therefore substituting
equation (3.3.1) into equation (3.3.6) yields increased fracture energy, see equation (4.5.5).

0.18
G fc ,imp =73 f cm ( c c 0 ) =73 f c0.18
0.014
(4.5.5)
,imp , k

The influence of each fracture energy on the deflection results was investigated on simplified
model in LS-DYNA. The change in compressive fracture energy Gfc influences mainly the

74
 Experimental verification

duration of the impact. This is due to the crushing of the contact area between beam and burden.
With higher fracture energy the crushing is reduced, therefore the burden rebounds faster. Change
in other two fracture energies (tensile fracture energy Gft and shear fracture energy Gfs)
significantly interacts. When one of the energy is increased disproportionately the element failure
occurs in the opposite mode (when shear energy is increased the element fails in tension). The
deflection development is significantly influenced by both the energies. Based on these facts the
fracture energies were modified for each specimen set. Values of the fracture energies as used in
LS-DYNA model are listed in Tab. 18.
Tab. 18 Fracture energies used in LS-DYNA model

Set fc,cyclic Gfc Gfc Gfc,imp Gft Gft Gfs Gfs

MC90 MC10 / Gfc,imp / Gfc,imp
[MPa] [Nm-1] [Nm-1] [Nm-1] [%] [Nm-1] [%] [Nm-1]
R1 35.0 83.328 143.684 147.894 0.98 1.449 0.98 1.449
C1 35.0 83.328 143.684 147.894 1.00 1.479 1.00 1.479
A1 30.2 76.606 140.610 144.729 0.70 1.013 0.65 0.941
A2 30.5 77.147 140.865 144.992 0.85 1.232 0.80 1.160
A3 32.7 80.076 142.221 146.388 0.83 1.215 0.78 1.142
A4 32.9 80.390 142.364 146.535 0.82 1.202 0.80 1.172
A5 32.5 79.865 142.125 146.288 1.03 1.507 1.00 1.463
A6 31.5 78.530 141.510 145.655 0.90 1.311 0.80 1.165
A8 28.2 73.824 139.279 143.359 1.00 1.434 1.00 1.434

4.5.3 Experimental and FEM results of the impact testing

For each specimen the results of FEM analysis were compared with the measured data from the
experiment. Differences between results were up to 10%, in most cases nearly exact match was
accomplished. Results for all the specimens are described below.
Results for the specimen R1, not exposed to cyclic loading or aggressive environment, are in Fig.
56. After the impact the steep decrease in velocity can be observed. The absorbed energy is
reflected in a deflection increase. The impact lasts about 7 ms (when the velocity begins to gain
negative values). The specimen was not completely damaged as can be seen in Fig. 57, thus the
weight rebounded (velocities lower than 0) and the specimen deflection decreased. The specimen
failed in shear similarly to the fatigue testing.

75
 Experimental verification

Fig. 56 Development of burden speed / deflection (left) and force-deflection curve / area (right) of spec. R1

Fig. 57 Condition of the specimen R1 after impact loading

Results for the specimen C1, exposed to cyclic loading with no initial shear crack, are in Fig. 58.
Duration of the impact was about 5 ms. The specimen failed in shear similarly to the fatigue testing
(see Fig. 59).

Fig. 58 Development of burden speed / deflection (left) and force-deflection curve / area (right) of spec. C1

76
 Experimental verification

Fig. 59 Condition of the specimen C1 after impact loading

Results for the specimen A1, chemically deteriorated specimen exposed to cyclic loading (shear
cracks appeared during the cyclic loading), are in Fig. 60. The decrease in velocity after the impact
is slower than for the specimen C1. Shear cracks from the fatigue testing were opening, thus the
energy was absorbed at slower rate, but with result of higher deflections. The impact lasts about
11 ms. For condition of the specimen A1 after the impact testing see Fig. 61.

Fig. 60 Development of burden speed / deflection (left) and force-deflection curve / area (right) of spec. A1

Fig. 61 Condition of the specimen A1 after impact loading

Results for the specimen A2, chemically deteriorated specimen exposed to cyclic loading, are in
Fig. 62. Duration of the impact was about 9 ms. The specimen cracked close to the right support
(see Fig. 63). The comparison of the camera records and PDV data showed, that the results from
77
 Experimental verification

laser measurements were incorrect probably because larger part of concrete fell from the bottom
surface. This problem was observed only with this specimen. The beam deflections was therefore
obtained by visual analysis of the camera recording.

Fig. 62 Development of burden speed / deflection (left) and force-deflection curve / area (right) of spec. A2

Fig. 63 Condition of the specimen A2 after impact loading

After the impact loading a longitudinal crack was found in the specimen A2. This crack
corresponds with the results from the LS-DYNA model. Visual comparison between the FEM
results and actual condition of the specimen can be seen in Fig. 64. This longitudinal crack can be
caused by the transversal bending moment as a result of the high-speed loading, when the short-
term stiffness of the specimen influences the support system and the beam parts with reinforcement
can be considered as flexible supports. Due to the symmetrical reinforcement the central part of
the beam (without a reinforcement bar) is the weakest part in the transversal direction.

78
 Experimental verification

Fig. 64 Longitudinal crack in specimen A2 due to the impact loading experiment vs. LS-DYNA

Results for the specimen A3, chemically deteriorated specimen exposed to cyclic loading, are in
Fig. 65. Duration of the impact was about 8.5 ms. The specimen cracked only on the right side
(see Fig. 66).

Fig. 65 Development of burden speed / deflection (left) and force-deflection curve / area (right) of spec. A3

79
 Experimental verification

Fig. 66 Condition of the specimen A3 after impact loading

Results for the specimen A4, chemically deteriorated specimen exposed to cyclic loading, are in
Fig. 67. Duration of the impact was about 9.5 ms. The specimen cracked only on the left side.
Extreme surface deterioration due to the aggressive environment can be seen (Fig. 66).

Fig. 67 Development of burden speed / deflection (left) and force-deflection curve / area (right) of spec. A4

Fig. 68 Condition of the specimen A4 after impact loading

Results for the specimen A5, chemically deteriorated specimen exposed to cyclic loading, are in
Fig. 69. Duration of the impact was about 5.5 ms. The specimen cracked symmetrically on both
sides, which resulted in significantly smaller deflections (the influence of the reinforcement was
higher). Extreme surface deterioration due to the aggressive environment can be seen (Fig. 70).

80
 Experimental verification

Fig. 69 Development of burden speed / deflection (left) and force-deflection curve / area (right) of spec. A5

Fig. 70 Condition of the specimen A5 after impact loading

Results for the specimen A6, chemically deteriorated specimen exposed to cyclic loading, are in
Fig. 71. Duration of the impact was about 8 ms. The specimen cracked only on the right side,
which resulted in higher deflection, similarly as in case of specimens A2, A3 and A4. Extreme
surface deterioration due to the aggressive environment can be seen (Fig. 72).

Fig. 71 Development of burden speed / deflection (left) and force-deflection curve / area (right) of spec. A6

81
 Experimental verification

Fig. 72 Condition of the specimen A6 after impact loading

Results for the specimen A8, chemically deteriorated specimen exposed to cyclic loading, are in
Fig. 73. Duration of the impact was about 7 ms. The specimen cracked symmetrically on both
sides without significant damage close to supports. This resulted in relatively small deflections.
Extreme surface deterioration due to the aggressive environment can be seen (Fig. 74).

Fig. 73 Development of burden speed / deflection (left) and force-deflection curve / area (right) of spec. A8

Fig. 74 Condition of the specimen A8 after impact loading

Summary of the results from LS-DYNA model together with the results from the PDV and camera
recordings can be seen in Tab. 19 The statistically obtained maximum force (average, minimum
and maximum) is stated together with the maximal force from the FEM model. The total energy
of the system Etotal may be considered as constant. The beam internal energy as stated in LS-DYNA
82
 Experimental verification

can be divided into two parts: a) work done on the deflection Wimp according to equation (4.5.1)
and b) supplementary damage (crushing of concrete, cracking of the upper surface after the impact
due to the rebound of the specimen etc.). The work done on the deflection obtained from the FEM
analysis and from the experimental data evaluation are listed together. The damage energy is listed
only for the FEM analysis, because it is not possible to evaluate this property from the experimental
data.

Tab. 19 Summary of the impact testing results

Set Fmax Fmax,av Fmax,min Fmax,max Etotal Eint,beam Wimp Wimp Eint,damage
FEM FEM exp. FEM
[kN] [kN] [kN] [kN] [J] [J] [J] [J] [J]
R1 153.5 392.9 156.5 430.6 1893 1843 1153 1181 824
C1 174.5 250.5 159.4 380.6 1818 1081 1114 858
A1 134.7 308.8 206.1 488.1 1789 1174 1079 615
A2 144.7 210.0 52.3 298.2 1836 1006 1074 830
A3 147.0 300.1 241.8 411.0 1845 1023 1059 812
1923
A4 149.0 251.8 103.6 370.6 1825 1106 1135 720
A5 154.9 328.8 173.7 369.5 1809 776 1039 1033
A6 152.5 388.2 290.7 441.1 1818 1038 1052 780
A8 151.0 241.8 179.4 402.6 1839 872 965 967

The results listed in Tab. 19 and Fig. 56 - Fig. 74 showed that the visual method for the data
evaluation is suitable only for certain calculations. For example the maximum force is almost
constant from the viewpoint of FEM model, which may be considered as correct. The maximum
force is dependent on the stiffness ratio of the two colliding bodies. The burden stiffness is
constant. The short-term stiffness of the beam (cca for 1 ms before the maximum force is reached)
will be nearly identical for all the specimens without the possible influence of the deterioration.
A lack of precision in the force evaluation by visual method can be observed even from the
significant scatter of the calculated values.
The visual evaluation can be used in properties where the small increments in consequent frames
are summed into relatively high values and the insufficient precision is partially eliminated. This
applies for deflections, velocities or work on deflections.
The short discussion about the energies listed in the Tab. 19 should be made. The total energy for
specimen R1 is lower than for other specimens. This is due to the absence of the force loading
which is a partial replacement for the damage caused by the cyclic loading, see section 3.3.2.5.
The energy dissipated from the work done by these forces was removed from the Wimp. The internal
energy of the beam is therefore only about 50 J lower than the total energy (in average). This
83
 Experimental verification

corresponds with the kinetic energy of the burden after the rebound (maximum velocity
around 1 m/s).
The main criterion for the comparison of the concrete elements behaviour during the impact testing
was work done on the specimen deflections. A trend is that with higher deterioration degree the
dissipated energy is lower. This can be attributed to the deteriorated material matrix. The energy
needed to open or form the cracks in the deteriorated specimens is lower. In case of the reference
specimen R1 (dry condition, without cyclic loading) the dissipated energy is 1153 J. In case of the
specimen C1 (dry condition, cyclically loaded) the dissipated energy is only 1081 J. Previously
formed bending cracks (during the cyclic loading) with already concentrated stresses at the tips
are elongated with lower energy addition. In case of specimen R1 new cracks has to be formed,
thus the energy release is higher. The chemical deterioration is similar to the cyclic loading
damage. The softer material matrix can be disintegrated with the addition of lower energy. This
will be further discussed with the methods for the deterioration assessment in the next section.

4.5.4 Energy estimation with deterioration assessment methods

In the sections 3.1 and 4.2 two different methods for the assessment of elements or structures
exposed to acidic chemical deterioration were described and verified. Their possible use in the
estimation of the work done on the displacement of the tested specimens will be discussed further
in text.
For each method a different approach should be chosen according to their main principle:
a) reduced compressive strength and b) reduced cross-sectional area of the specimen. Both these
principles were already discussed in previous sections. As an input data the work done on the
deflections from the Tab. 19 will be used.
Both modifications of the proposed assessment methods are based on the strain energy release rate
which in its simplified form can be written as in equation (4.5.6).

W W
G=
=
(4.5.6)
ba Acrack

When the strain energy release rate (4.5.6) is equal or greater than the critical value, the fracture
energy Gf, the crack will grow.
The first method, the reduced compressive strength method, is based on these assumptions. The
fracture energy is a property of the material and can be evaluated on the basis of standards as
described in section 3.3.2.6. Furthermore, if the specimens exposed to the impact loading are of
the same width b and exhibit identical or similar damage it can be assumed that the area of the
crack Acrack will be the same or at least similar for all the tested specimens. Thus for the

84
 Experimental verification

approximation of the energy dissipated in the deflections of the specimens the crack area can be
determined as a constant. Therefore substituting (4.5.1) into the energy release rate (4.5.6) yields
the equation (4.5.7).

G f ( f cm ,imp ) Acrack = Wimp (4.5.7)

The fracture energy can be evaluated according to Model Code 2010 [66] based on the compressive
strength of the specimens. The work done on deflections Wimp can be obtained from the LS-DYNA
analysis or from the experimental measurements. The crack area Acrack is sought for. With regard
to the fact that the tested specimens are highly reinforced the Acrack value is only imaginary. For
the best fit of the values from Tab. 19 the crack area Acrack = 7.53 m2. Substituting Acrack and (4.5.5)
into the equation (4.5.7) the formulation (4.5.8) is obtained.

0.18
,imp , k= 549.69 f cm ( c c 0 )
549.69 f c0.18
0.014
= Wimp (4.5.8)

The comparison of the values obtained from the LS-DYNA analysis and values calculated
according to the equation (4.5.8) using the compressive strength for the beams (Tab. 15) can be
seen in Fig. 75.

Fig. 75 Area under force-deflection curve according to reduced compressive strength

The second method, the cross-sectional area reduction method, is also based on the strain energy
release rate. The basic assumption, as stated before, is that the non-deteriorated concrete core
should have the same material properties across all the samples. This implies that the fracture
energy is also the same and the equation (4.5.6) can be rewritten as follows:

G f Areduced = Wimp (4.5.9)

85
 Experimental verification

It is obvious that the value of reduced cross-sectional area Areduced is not the crack area but its
substitute which incorporates the deterioration degree of the specimen. As with the imaginary
Acrack in the previous method the fracture energy Gf is an imaginary value. With regard to the fact
that all the specimens were exposed to cyclic loading (except specimen R1) the C1 specimen was
selected as the reference value, therefore Gf = 1081 J. Substituting Gf into equation (4.5.9) yields
formula for the cross-sectional area reduction method, see equation (4.5.10).

1081 breduced hreduced =

Wimp (4.5.10)

The comparison of the values obtained from the LS-DYNA analysis and values calculated
according to the equation (4.5.10) using the reduced cross-sectional dimensions for the beams
(Tab. 15) can be seen in Fig. 76.

Fig. 76 Area under force-deflection curve according to reduced cross-sectional area

Both methods follow the trend that with higher deterioration degree the work done on the
deflections decreases. The method of the reduced cross-sectional area exhibits better fit than the
method of reduced compressive strength. The main advantage of these methods is in the reduction
of the calculation time which is needed for the specimens fitting. Only one model of non-
deteriorated concrete specimens can be analysed in LS-DYNA or other explicit dynamics software
and the energy balance for the deteriorated specimens will be estimated on the basis of these
methods and obtained results from the FEM model.

86
 Discussion

5 DISCUSSION
5.1 Chemical deterioration
Results of the experiment enabled to fit the proposed model of the acidic chemical deterioration
of concrete. Both proposed methods for the assessment of the deterioration degree of concrete
elements were verified with good agreement between calculated and experimentally obtained
values.
The diffusion problem of the chemical deterioration in fluids can be described as a steady-state.
The corroded layer is uniform all around the specimen surfaces. The thickness of the corroded
layer increases very slowly even when the aggressive environment is extremely severe (pH = 2.0).
This is partially due to the stable corroded layer which acts like a diffusion barrier, therefore it can
be assumed that when the elements is exposed to weather conditions or mechanical abrasion the
gradual removal of this corroded layer may cause faster deterioration rate. On the other hand the
environment chosen for the experiment is 100 times more aggressive than the standardized XA3
value in EN 206 [20], therefore it can be assumed that even when the weather influence will apply
the results obtained with the proposed methods are on the safe side.
The laboratory results, and also the SEM and EDS analysis, have verified the assumption that, in
the case of acidic chemical deterioration, the calcium ions are the most dissolved elements.
Dissolution of other strongly represented ions (Fe3+, Al3+, Si4+ etc.) occurs to a considerably less
extent. On this basis, the assumption that the amount of dissolved calcium ions (Ca2+) plays the
key role in the deterioration process of the material characteristics was made.
The reason for brown colour of the corroded layer is another highly discussed topic. General
consensus is that the orange-brown colour is due to the increased presence of the ferric ions. There
is only small increase in the mass of the Fe3+ ions in the corroded layer (up to 1.5%), but in the
viewpoint of atomic percentage the increase is higher due to the lack of the calcium ions. Therefore
the slightly red or orange shade of the deteriorated layer is probably the result of the Fe3+ ions
content. But with regard to the SEM and EDS analysis results the brown colour is due to the
decrease in Ca2+ ions which cover the fine aggregate, thus the colour of the aggregate (sand) will
reveal.
The proposed methods were compared with several laboratory techniques (SEM, EDS, AAS) with
good agreement, thus their suitability was verified.

87
 Discussion

5.2 Cyclic loading

The experimental part focused on the cyclic loading tests should verify the proposed extension of
the fatigue damage function [38] together with the possible use of the methods for the deterioration
degree assessment.
The extension of the method for the deflection increase calculation of the specimens exposed to
the cyclic loading showed good agreement with the experimentally measured data. The approach
based on the partial integration over the height of the beam compressive zone incorporates the
stress distribution in the cross-section, therefore reduces the conservativeness of the standardized
approaches. For optimum accuracy of the calculation the division of the compressive zone to 20
layers is recommended. With finer division the time cost of the calculation increases with
negligible improvement in the results. On the other hand the coarse division may results in the
underestimated deflection calculation.
The implementation of both methods for deterioration degree assessment into the deflection
development calculation showed, that the reduced cross-sectional area is not appropriate for this
type of evaluation. With higher deterioration degree the reduced cross-section significantly
influences the stress calculation in evaluated layers. This affects the calculation of the modulus of
elasticity deterioration and consequently the deflections which are dependent on the ideal moment
of inertia, thus on the modulus of elasticity. The reduced cross-section area method is therefore
not recommended for the evaluation of the service limit states SLS. For the ultimate limit states
(ULS) where the materials are considered at their yield stress the corroded layer will not apply in
the load bearing capacity. Thus it can be assumed that this method can be used as an analogy to
the concrete fire design as described in EN 1992-1-2 [54].
The method of reduced compressive strength showed good agreement with the experimentally
obtained data.
The influence of the chemical deterioration on the measured deflection was up to 20%, after 400
thousand cycles, when comparing specimens stored in dry condition (C sets) and specimens
stored in aggressive environment (A sets).
The strain rate influence on the deflections was up to 7% when comparing dynamic and static
deflections. The average values were lower but when the structure is analysed on the basis of the
dynamic deflections their potential decrease due to loading speed should be included.

88
 Discussion

5.3 Impact loading

The impact tests were performed on the reinforced concrete beam specimens to evaluate the
influence of cyclic loading and chemical deterioration on their impact resistance.
The experimental setup with use of high-speed camera has proven insufficient for the correct
evaluation of the impact tests. The visual analysis of the camera recordings is highly dependent on
the resolution of the frames which decreases with the frame rate. Vice versa the time gap between
consecutive frames should be adequate to the loading speed. An optimum between the frame rate
and its resolution should be evaluated for each specific camera.
To obtain better measurements it can be recommended to extend the measurements of at least one
other method. Very accurate and simple method for the additional measurements is Photon
Doppler Velocimetry (PDV) which can provide velocity of the beam deflections, therefore its
acceleration and deflection. Problem may be the purchase price.
Based on the beam deflection results obtained with PDV a numerical model in LS-DYNA (an
explicit dynamics FEM software) was made. The method of changing the mechanical properties
of concrete on the basis of the material model fracture energy proved to be suitable. The numerical
modelling of reinforced concrete specimens showed good agreement with the data evaluated from
the camera recordings.
The influence of the previous damage caused by the cyclic loading, aggressive environment or
their combination resulted in the lower energy dissipated in the beam deflections. This can be
attributed to the cracks formation during the cyclic loading together with the concentrated stress
at the tips of the micro-cracks in material matrix. Influence of the aggressive environment is mainly
in the reduction of the effective cross-sectional area which is reflected in good agreement of the
proposed method for energy estimation (based on the reduced cross-sectional area method) with
values from the experiment and numerical modelling.
The second proposed method of the energy estimation (based on the reduced compressive strength)
seems less suitable for this type of evaluation.
The differences in the energy dissipated in the deflection (work done on deflection) between the
reference specimen R1 and specimen exposed to cyclic loading C1 is up to 7%. In case of the
deteriorated specimens the difference is higher, up to 20% when comparing to specimen C1 and
up to 25% when comparing to specimen R1.
The strain rate effect cannot be properly quantified, but it was incorporated in the fracture energy
entered to the LS-DYNA model.

89
 Conclusions

6 CONCLUSIONS
The main objective of this thesis was to propose an analytical model dealing with the acidic
chemical deterioration of concrete structures which can be further extended for the use in
assessment of the deflection increase due to the cyclic loading and for use in the estimation of the
impact resistance of concrete elements.
The thesis includes comprehensive state of the art of all three topics covered in the work; chemical
deterioration of concrete, cyclic loading and impact loading.
The goals stated in the introduction (section 1) can be summarized as follows:
Two different analytical approaches for the evaluation of the deterioration degree of the
concrete elements were proposed and successfully verified with the experimental program.
All used assumption were proved with laboratory testing, SEM and EDS.
The analytical system for the evaluation of the deflection increase due to the cyclic loading
was extended and the obtained results were experimentally verified. An approach for the
estimation of the remaining fatigue life-time independent on the total number of applied
cycles N and number of already applied cycles n was proposed.
The approach for possible estimation of the energy dissipated during the impact loading in
dependency on the deterioration degree was formulated. The recommendations for the
FEM modelling of the impact loading were proposed.
Both proposed methods for the assessment of deteriorated specimens were implemented in
the consequent systems of cyclic loading and impact loading analysis. Suitability of
particular method for different limit states was obtained by comparing the calculated and
measured data. Principle of the proposed methods implementation can be seen in Fig. 77.

The detailed conclusions are provided below for each research part of the thesis.

Fig. 77 Principle of the proposed approaches implementation to different design states analysis

The first part of this thesis, chemical deterioration of concrete, described the fundamentals of the
chemical reaction kinetics that are needed for the proposed analytical model of chemical
90
 Conclusions

deterioration. The model is based on the philosophy of the simplified chemical kinetics of second
order reaction with use of the pseudo-first order reaction which can be solved effectively by
conventional methods. The proposed analytical model enables to calculate the total loss of calcium
ions Ca2+ from the concrete material matrix based on the pH value of the aggressive environment,
duration of the exposure and surface to volume ratio.
Results from the model are further used in two proposed methods of deterioration degree
assessment. The first method is based on the reduction of the compressive strength and the second
method is based on the reduction of the cross-sectional area.
The experimental verification of the proposed analytical model together with both methods for the
concrete deterioration assessment is presented. The calculated values of the calcium ions loss were
compared with the laboratory measurements with very good agreement. All stated assumptions
were further verified using scanning electron microscopy and Energy-dispersive X-ray
spectroscopy.
Both presented methods for deterioration degree assessment showed good agreement with results
from the experimental testing.
The second part of this thesis, cyclic loading, proposed an extension for the calculation of
deflection increase due to the cyclic loading (fatigue damage function) from [38]. To consider the
stress distribution in concrete specimens exposed to cyclic bending the method of partial
integration over the height of the specimen compressive zone was used.
The experimental verification proved the suitability of the proposed extension for the deflection
increase evaluation. Further, the extension allows to easily implement both calculation methods
for the deterioration degree assessment. Comprehensive dataset covering the results of the non-
deteriorated together with the deteriorated specimens was obtained from the experimental
program. The proposed extension showed good agreement with the experimental data for the
specimens stored in dry environment. The comparison with the dataset of deteriorated specimens
have showed inaptitude of the reduced cross-sectional method for this type of calculations. The
method is not suitable for the service limit state assessment, still it can be used for the ultimate
limit states evaluation.
The reduced compressive strength method showed very good agreement with the measured
dataset.
The third part of this thesis, impact loading, described influence of the strain rate effects on
concrete together with the recommendations for the numerical modelling of the impact testing.
The energy balance of the impact test had been widely described.
The experimental part sought for the influence of the previous damage due to the cyclic loading,
chemical deterioration or their combination on the impact resistance of the concrete specimens. In
91
 Conclusions

total 9 specimens of three different damage type were tested; 1) reference specimen, 2) specimens
stored in dry environment and exposed to cyclic loading and 3) specimens stored in aggressive
environment with subsequent cyclic loading.
Numerical modelling in LS-DYNA software was based on the measurements of the impact testing
with high-speed camera and Photon Doppler velocimetry. The influence of the change in fracture
energy was investigated during the model fitting.
The evaluation of the specimens behaviour during the impact testing was based on the value of the
energy dissipated in the beam deflections. The values obtained from the numerical models were
comparable with the values calculated using the camera recordings.
Both methods for the deterioration degree assessment were extended for the estimation of the
energy dissipation during the impact testing based on the strain energy release rate. For this
purpose the method of reduced cross-sectional area showed good agreement with measured data.
This fact verifies the assumption that the method is more suitable for the ultimate limit states,
which impact loading definitely is.
In conclusion, it has to be emphasized that the analytical model of the chemical deterioration
should be generally valid for the acidic environment only with small adjustments for the specific
chemical reaction. The advantage of the approach presented here is its generality, due to its
applicability to the various types of widely-used structures just by using a specific S/V ratio and at
least an approximate estimation of the aggressiveness of an environment with the equivalent pH
value. In addition the proposed methods for the deterioration degree assessment proved their
suitability for different purposes and can be easily used in common praxis.

92
 References

7 REFERENCES
[1] LAUER, K. R. Classification of concrete damage caused by chemical attack. Materials and
Structures [online]. May 1990. Vol. 23, no. 3, p. 223229. DOI 10.1007/BF02473022.
[2] LAUER, K. R. State-of-the art report: the use of damage classification systems for concrete
structures. Materials and Structures [online]. 1991. Vol. 24, no. 4, p. 265267.
DOI 10.1007/BF02472081.
[3] PAVLK, V. Corrosion of hardened cement paste by acetic and nitric acids part I:
Calculation of corrosion depth. Cement and Concrete Research [online]. January 1994.
Vol. 24, no. 3, p. 551562. DOI 10.1016/0008-8846(94)90144-9.
[4] SN 73 6221. Inspection of road bridges. 2011. Prague : Czech office for standards,
metrology and testing.
[5] ZIVICA, Vladimr and BAJZA, Adolf. Acidic attack of cement based materials a review
Part 1. Principle of acidic attack. Construction and Building Materials [online]. December
2001. Vol. 15, no. 8, p. 331340. DOI 10.1016/S0950-0618(01)00012-5.
[6] ZIVICA, Vladimr and BAJZA, Adolf. Acidic attack of cement-based materials - A review
Part 2. Factors of rate of acidic attack and protective measures. Construction and Building
Materials. 2002. Vol. 16, no. 4, p. 215222. DOI 10.1016/S0950-0618(02)00011-9.
[7] OKOCHI, Hiroshi, KAMEDA, Hideki, HASEGAWA, Shin-ichi, SAITO, Nobuhiko,
KUBOTA, Ken and IGAWA, Manabu. Deterioration of concrete structures by acid
deposition an assessment of the role of rainwater on deterioration by laboratory and field
exposure experiments using mortar specimens. Atmospheric Environment [online]. January
2000. Vol. 34, no. 18, p. 29372945. DOI 10.1016/S1352-2310(99)00523-3.
[8] KONG, Hendrik L. and ORBISON, James G. Concrete Deterioration Due to Acid
Precipitation. ACI Materials Journal [online]. 1987. Vol. 84, no. 2, p. 110116.
DOI 10.14359/1825.
[9] XIE, Shaodong, QI, Li and ZHOU, Ding. Investigation of the effects of acid rain on the
deterioration of cement concrete using accelerated tests established in laboratory.
Atmospheric Environment [online]. September 2004. Vol. 38, no. 27, p. 44574466.
DOI 10.1016/j.atmosenv.2004.05.017.
[10] KANAZU, Tsutomu, MATSUMURA, Takuro, NISHIUCHI, Tatsuo and YAMAMOTO,
Takeshi. Effect of simulated acid rain on deterioration of concrete. Water, Air, and Soil
Pollution [online]. 2001. Vol. 130, no. 1/4, p. 14811486.
DOI 10.1023/A:1013925316797.
[11] FAN, Y.F., HU, Z.Q., ZHANG, Y.Z. and LIU, J.L. Deterioration of compressive property

93
 References

of concrete under simulated acid rain environment. Construction and Building Materials
[online]. October 2010. Vol. 24, no. 10, p. 19751983.
DOI 10.1016/j.conbuildmat.2010.04.002.
[12] CHEN, Meng-Cheng, WANG, Kai and XIE, Li. Deterioration mechanism of cementitious
materials under acid rain attack. Engineering Failure Analysis [online]. January 2013.
Vol. 27, p. 272285. DOI 10.1016/j.engfailanal.2012.08.007.
[13] IVICA, Vladimr. Deterioration of cement-based materials due to the action of organic
compounds. Construction and Building Materials [online]. November 2006. Vol. 20, no. 9,
p. 634641. DOI 10.1016/j.conbuildmat.2005.02.011.
[14] FERNANDES, I., PERICO, M., HAGELIA, P., NORONHA, F., RIBEIRO, M. a. and
MAIA, J. Identification of acid attack on concrete of a sewage system. Materials and
Structures [online]. 11 March 2012. Vol. 45, no. 3, p. 337350. DOI 10.1617/s11527-011-
9769-y.
[15] DE WINDT, Laurent and DEVILLERS, Philippe. Modeling the degradation of Portland
cement pastes by biogenic organic acids. Cement and Concrete Research [online]. August
2010. Vol. 40, no. 8, p. 11651174. DOI 10.1016/j.cemconres.2010.03.005.
[16] BERTRON, a., DUCHESNE, J. and ESCADEILLAS, G. Accelerated tests of hardened
cement pastes alteration by organic acids: analysis of the pH effect. Cement and Concrete
Research [online]. January 2005. Vol. 35, no. 1, p. 155166.
DOI 10.1016/j.cemconres.2004.09.009.
[17] GIRARDI, F, VAONA, W and MAGGIO, R Di. Resistance of different types of concretes
to cyclic sulfuric acid and sodium sulfate attack. Cement and Concrete Composites [online].
2010. Vol. 32, no. 8, p. 595602. DOI 10.1016/j.cemconcomp.2010.07.002.
[18] GIRARDI, F. and MAGGIO, R. D. Resistance of concrete mixtures to cyclic sulfuric acid
exposure and mixed sulfates: Effect of the type of aggregate. Cement and Concrete
Composites [online]. 2011. Vol. 33, no. 2, p. 276285.
DOI 10.1016/j.cemconcomp.2010.10.015.
[19] CHATVEERA, B. and LERTWATTANARUK, P. Evaluation of nitric and acetic acid
resistance of cement mortars containing high-volume black rice husk ash. Journal of
Environmental Management [online]. 2014. Vol. 133, p. 365373.
DOI 10.1016/j.jenvman.2013.12.010.
[20] EN 206. Concrete, Specification, performance, production and conformity. 2013. Brussels :
European Committee for Standardization.
[21] FAN, Y. F., LUAN, H. Y. and ZHANG, S. Y. Multiscale Investigation of Porosity
Characteristic in Concrete Exposed to Acidic Environment. International Journal of
94
 References

Corrosion [online]. 2012. Vol. 2012, p. 19. DOI 10.1155/2012/923010.

[22] HUANG, Pengfei, BAO, Yiwang and YAO, Yan. Influence of HCl corrosion on the
mechanical properties of concrete. Cement and Concrete Research [online]. March 2005.
Vol. 35, no. 3, p. 584589. DOI 10.1016/j.cemconres.2004.06.027.
[23] BEDDOE, Robin E. and DORNER, Horst W. Modelling acid attack on concrete: Part I.
The essential mechanisms. Cement and Concrete Research [online]. 2005. Vol. 35, no. 12,
p. 23332339. DOI 10.1016/j.cemconres.2005.04.002.
[24] DORNER, H W and BEDDOE, R E. Prognosis Of Concrete Corrosion Due To Acid Attack.
9Dbmc-2002. 2002. P. 19.
[25] MIYAMOTO, Shintaro, MINAGAWA, Hiroshi and HISADA, Makoto. Deterioration rate
of hardened cement caused by high concentrated mixed acid attack. Construction and
Building Materials [online]. 2013. Vol. 67, p. 4754.
DOI 10.1016/j.conbuildmat.2013.11.008.
[26] SINHA, B. P, GERSTLE, Kurt H and TULIN, Leonard G. Stress-Strain Relations for
Concrete Under Cyclic Loading. ACI Journal Proceedings [online]. 1964. Vol. 61, no. 2,
p. 195212. DOI 10.14359/7775.
[27] MARTNEZ-RUEDA, J. Enrique and ELNASHAI, A S. Confined concrete model under
cyclic load. Materials and Structures [online]. April 1997. Vol. 30, no. 3, p. 139147.
DOI 10.1007/BF02486385.
[28] BAHN, B Y and HSU, C T. Stress-Strain Behavior of Concrete under Cyclic Loading. ACI
Materials Journal [online]. 1998. Vol. 95, no. 2, p. 178193. DOI 10.14359/363.
[29] VOYIADJIS, George Z. and ABU-LEBDEH, Taher M. Plasticity model for concrete using
the bounding surface concept. International Journal of Plasticity [online]. January 1994.
Vol. 10, no. 1, p. 121. DOI 10.1016/0749-6419(94)90051-5.
[30] VOYIADJIS, George Z. and ABU LEB
Using Bounding Surface Concept. Journal of Engineering Mechanics [online]. September
1993. Vol. 119, no. 9, p. 18651885. DOI 10.1061/(ASCE)0733-9399(1993)119:9(1865).
[31] YANG, Bing-Lin, DAFALIAS, Yannis F. and HERRMANN, Leonard R. A Bounding
Surface Plasticity Model for Concrete. Journal of Engineering Mechanics [online]. March
1985. Vol. 111, no. 3, p. 359380. DOI 10.1061/(ASCE)0733-9399(1985)111:3(359).
[32] MAEKAWA, KOICHI, HAJIME OKAMURA, and Amorn Pimanmas. Non-linear
mechanics of reinforced concrete. New York : Spon Press, 2003.
[33] PAPA, Enrico and TALIERCIO, Alberto. Anisotropic damage model for the multiaxial
static and fatigue behaviour of plain concrete. Engineering Fracture Mechanics [online].
September 1996. Vol. 55, no. 2, p. 163179. DOI 10.1016/0013-7944(96)00004-5.
95
 References

[34] ALLICHE, Abdenour. Damage model for fatigue loading of concrete. International Journal
of Fatigue [online]. September 2004. Vol. 26, no. 9, p. 915921.
DOI 10.1016/j.ijfatigue.2004.02.006.
[35] GRASSL, Peter and REMPLING, Rasmus. A damage-plasticity interface approach to the
meso-scale modelling of concrete subjected to cyclic compressive loading. Engineering
Fracture Mechanics [online]. November 2008. Vol. 75, no. 16, p. 48044818.
DOI 10.1016/j.engfracmech.2008.06.005.
[36] CHEN, Wai-Fah and HAN, Da-Jian. Plasticity for structural engineers. New York :
Springer-Verlag, 1988. ISBN 978-1-4612-8380-5.
[37] HOLMEN, Jan Ove. Fatigue of Concrete by Constant and Variable Amplitude Loading.
University of Trondheim, 1979.
[38] FOGLAR, Marek. Strain development in concrete under cyclic loading. Czech Technical
University in Prague, 2008.
[39] BAZANT, Zdenek P. and HUBLER, Mija H. Theory of cyclic creep of concrete based on
Paris law for fatigue growth of subcritical microcracks. Journal of the Mechanics and
Physics of Solids [online]. February 2014. Vol. 63, no. 1, p. 187200.
DOI 10.1016/j.jmps.2013.09.010.
[40] OTTOSEN, N S. A Failure Criterion for Concrete. American Society of Civil Engineers.
Engineering Mechanics Division. Journal. 1977. Vol. 103, no. 4, p. 527535.
[41] MINDESS, Sidney and BENTUR, Arnon. A preliminary study of the fracture of concrete
beams under impact loading, using high speed photography. Cement and Concrete
Research [online]. May 1985. Vol. 15, no. 3, p. 474484. DOI 10.1016/0008-
8846(85)90121-8.
[42] MINDESS, S, BANTHIA, N and BENTUR, A. The response of reinforced concrete beams
with a fibre concrete matrix to impact loading. International Journal of Cement Composites
and Lightweight Concrete [online]. August 1986. Vol. 8, no. 3, p. 165170.
DOI 10.1016/0262-5075(86)90037-0.
[43] MINDESS, Sidney, BANTHIA, Nemkumar and YAN, Cheng. The fracture toughness of
concrete under impact loading. Cement and Concrete Research [online]. March 1987.
Vol. 17, no. 2, p. 231241. DOI 10.1016/0008-8846(87)90106-2.
[44] BANTHIA, N., MINDESS, S, BENTUR, A and PIGEON, M. Impact testing of concrete
using a drop-weight impact machine. Experimental Mechanics [online]. March 1989.
Vol. 29, no. 1, p. 6369. DOI 10.1007/BF02327783.
[45] WANG, Nianzhi, MINDESS, Sidney and KO, Keith. Fibre reinforced concrete beams under
impact loading. Cement and Concrete Research [online]. March 1996. Vol. 26, no. 3,
96
 References

p. 363376. DOI 10.1016/S0008-8846(96)85024-1.

[46] RIEDER, K. -A. and MINDESS, Sidney. New test method to evaluate the impact behaviour
of biaxially confined concrete. Materials and Structures [online]. December 1998. Vol. 31,
no. 10, p. 669675. DOI 10.1007/BF02480443.
[47] BISCHOFF, Peter H. and PERRY, Simon H. Impact Behavior of Plain Concrete Loaded in
Uniaxial Compression. Journal of Engineering Mechanics [online]. June 1995. Vol. 121,
no. 6, p. 685693. DOI 10.1061/(ASCE)0733-9399(1995)121:6(685).
[48] SUKONTASUKKUL, Piti, NIMITYONGSKUL, Pichai and MINDESS, Sidney. Effect of
loading rate on damage of concrete. Cement and Concrete Research [online]. November
2004. Vol. 34, no. 11, p. 21272134. DOI 10.1016/j.cemconres.2004.03.022.
[49] BADR, Atef and ASHOUR, Ashraf F. Modified ACI Drop-Weight Impact Test for
Concrete. ACI Materials Journal [online]. 2005. Vol. 102, no. 4, p. 249256.
DOI 10.14359/14618.
[50] SOLEIMANI, S.M. and BANTHIA, N. A Novel Drop Weight Impact Setup for Testing
Reinforced Concrete Beams. Experimental Techniques [online]. May 2014. Vol. 38, no. 3,
p. 7279. DOI 10.1111/j.1747-1567.2012.00810.x.
[51] ABBAS, H, GUPTA, N.K. and ALAM, M. Nonlinear response of concrete beams and
plates under impact loading. International Journal of Impact Engineering [online].
September 2004. Vol. 30, no. 89, p. 10391053. DOI 10.1016/j.ijimpeng.2004.06.011.
[52] GEORGIN, J.F. and REYNOUARD, J.M. Modeling of structures subjected to impact:
concrete behaviour under high strain rate. Cement and Concrete Composites [online].
January 2003. Vol. 25, no. 1, p. 131143. DOI 10.1016/S0958-9465(01)00060-9.
[53] WRIGHT, Margaret Robson. An Introduction to Chemical Kinetics. Chichester : John
Wiley & Sons, Ltd., 2004. ISBN 0-470-09058-8.
[54] EN 1992-1-2. Eurocode 2: Design of concrete structures - Part 1-2: General rules -
Structural fire design. 2004. Brussels : European Committee for Standardization.
[55] NEVILLE, Adam M. Properties Of Concrete [online]. 4th and fi. Harlow : Longman
Group, 1995. ISBN 0-582-23070-5.
[56] SCRIVENER, Karen L. Calcium aluminate cements. In : NEWMAN, John and CHOO, Ban
Seng (eds.), Advanced Concrete Technology - 1: Constituent Materials. Oxford : Elsevier
Butterworth-Heinemann, 2003. p. 6393. ISBN 0-7506-5103-2.
[57] MINDESS, Sidney, YOUNG, J. Francis and DARWIN, David. Concrete. 2003. ISBN 0-
13-064632-6.
[58] BOGUE, Robert Herman. The chemistry of Portland cement. 2nd ed. New York : Reinhold
Publishing Corp., 1955.
97
 References

[59] FRET, R. Sur la compacit des mortiers hydrauliques. Annales des Ponts et Chausses.
1892. Vol. 4, no. 7, p. 5196.
[60] PAVLK, V. Corrosion of hardened cement paste by acetic and nitric acids part II:
Formation and chemical composition of the corrosion products layer. Cement and Concrete
Research [online]. 1994. Vol. 24, no. 8, p. 14951508. DOI 10.1016/0008-8846(94)90164-
3.
[61] JOHN, Crank. The mathematics of diffusion. 2nd ed. Oxford : Oxford University Press,
1975. ISBN 978-0-19-853411-2.
[62] VU, Kim Anh T and STEWART, Mark G. Structural reliability of concrete bridges
including improved chloride-induced corrosion models. Structural Safety [online]. January
2000. Vol. 22, no. 4, p. 313333. DOI 10.1016/S0167-4730(00)00018-7.
[63] SAASSOUH, Bassem and LOUNIS, Zoubir. Probabilistic modeling of chloride-induced
corrosion in concrete structures using first- and second-order reliability methods. Cement
and Concrete Composites [online]. October 2012. Vol. 34, no. 9, p. 10821093.
DOI 10.1016/j.cemconcomp.2012.05.001.
[64] STEWART, Mark G. and ROSOWSKY, David V. Time-dependent reliability of
deteriorating reinforced concrete bridge decks. Structural Safety [online]. January 1998.
Vol. 20, no. 1, p. 91109. DOI 10.1016/S0167-4730(97)00021-0.
[65] EN 1992-1-1. Eurocode 2: Design of concrete structures - Part 1-1: General rules and rules
for buildings. 2004. Brussels : European Committee for Standardization.
[66] CEB-FIP. fib Model Code for Concrete Structures 2010. Lausanne : fdration
internationale du bton, 2013. ISBN 978-3-433-03061-5.
[67] EN 1992-2. Eurocode 2: Design of concrete structures - Part 2: Concrete bridges - Design
and detailing rules. 2005. Brussels : European Committee for Standardization.
[68] ROSS, L Javier Malvar and C Allen. Review of Strain Rate Effects for Concrete in Tension.
ACI Materials Journal [online]. 1998. Vol. 95, no. 6. DOI 10.14359/418.
[69] TEDESCO, J W and ROSS, C A. Strain-Rate-Dependent Constitutive Equations for
Concrete. Journal of Pressure Vessel Technology [online]. 1 November 1998. Vol. 120,
no. 4, p. 398. DOI 10.1115/1.2842350. 10.1115/1.2842350
[70] CEB-FIP. CEB-FIP model code 1990. Comit euro-international du bton, 1993. ISBN 0-
7277-1696-4.
[71] HUGHES, Thomas J.R. and LIU, Wing Kam. Nonlinear finite element analysis of shells:
Part I. three-dimensional shells. Computer Methods in Applied Mechanics and Engineering
[online]. June 1981. Vol. 26, no. 3, p. 331362. DOI 10.1016/0045-7825(81)90121-3.
[72] HUGHES, Thomas J.R. and LIU, Wing Kam. Nonlinear finite element analysis of shells-
98
 References

part II. two-dimensional shells. Computer Methods in Applied Mechanics and Engineering
[online]. 1981. Vol. 27, no. 2, p. 167181. DOI 10.1016/0045-7825(81)90148-1.
[73] MURRAY, Yvonne. Users Manual for LS-DYNA Concrete Material Model 159. Federal
Highway Administration, 2007.
[74] MOUTOUSSAMY, L, HERVE, G and BARBIER, F. Qualification of * Constrained _
Lagrange _ In _ Solid command for steel / concrete interface modeling. 8th European LS-
DYNA Conference. 2011.
[75] BELYTSCHKO, Ted, ONG, Jame Shau-Jen, WING KAM LIU and KENNEDY, James M.
Hourglass control in linear and nonlinear problems. Computer Methods in Applied
Mechanics and Engineering [online]. May 1984. Vol. 43, no. 3, p. 251276.
DOI 10.1016/0045-7825(84)90067-7.
[76] ISO 1920-10:2010. Testing of concrete - Part 10: Determination of static modulus of
elasticity in compression. 2010. International Organization for Standardization.
[77] ISO 9297:1989. Water quality - Determination of chloride - Silver nitrate titration with
chromate indicator (Mohrs method). 1989. International Organization for Standardization.
[78] ISO 8288:1986. Water quality - Determination of cobalt, nickel, copper, zinc, cadmium and
lead - Flame atomic absorption spectrometric methods. 1986. International Organization
for Standardization.
[79] FAN, Ying Fang and LUAN, Hai Yang. Pore structure in concrete exposed to acid deposit.
Construction and Building Materials [online]. 2013. Vol. 49, p. 407416.
DOI 10.1016/j.conbuildmat.2013.08.075.
[80] SAVITZKY, Abraham. and GOLAY, M. J. E. Smoothing and Differentiation of Data by
Simplified Least Squares Procedures. Analytical Chemistry [online]. 1964. Vol. 36, no. 8,
p. 16271639. DOI 10.1021/ac60214a047.

99
 Publications

8 PUBLICATIONS
8.1 WoS and SCOPUS indexed publications
[P1] Foglar, M. - Gringer, J.: Influence of the Structural Arrangement of Bridges on the Noise
Induced by Traffic. Engineering Structures. 2013, vol. 56, p. 642-655. ISSN 0141-0296.
[P2] Gringer, J. - Foglar, M.: Analysis of the performance of the concrete specimens
subjected to cyclic loading and aggressive environment. In Bridge Maintenance, Safety,
Management and Life Extension 2014. Leiden: CRC Press/Balkema, 2014, p. 269-276.
ISBN 978-1-138-00103-9.
[P3] Gringer, J. - Foglar, M.: Experimental and Theoretical Study of the Performance of
Reinforced Concrete Specimen Subjected to Cyclic Loading and Aggressive
Environment. In 11th International Fatigue Congress. Zurich: Trans Tech Publications,
2014, p. 494-499. ISSN 1022-6680. ISBN 9783038350088.
[P4] Gringer, J. - Foglar, M.: Concrete deterioration due to aggressive environment and cyclic
loading - theoretical study. In 20th International Conference Engineering Mechanics 2014.
Brno: Brno University of Technology, 2014, p. 196-199. ISSN 1805-8248. ISBN 978-
80-214-4871-1.
[P5] Gringer, J. - Foglar, M. - Jiek, P.: Impact peformance of specimens subjected to
fatigue loading - experimental investigation. In 11th International Conference on the
Mechanical and Physical Behaviour of Materials under Dynamic Loading. Les Ulis: EDP
Sciences, 2015, ISSN 2101-6275. ISBN 978-2-7598-1817-4.

8.2 Other conference and journal publications

[P6] Foglar, M. - Gringer, J.: Influence of Bridge Structural Arrangement on the Noise
Induced by Traffic. Silnice eleznice. 2012, ro. 7, . 4, s. 73-84. ISSN 1801-822X. (in
Czech).
[P7] Gringer, J. - Foglar, M.: Review of Approaches to Solve Deterioration Processes in
Concrete due to Cyclic Loading and Aggressive Environment. In 19th Czech Concrete
Days 2012 - Proceedings. Praha: esk betonsk spolenost SSI, 2012, p. 241-244.
[P8] Gringer, J. - Foglar, M.: Review of Approaches to Solve Deterioration Processes in
Concrete due to Cyclic Loading and Aggressive Environment. In 19. betonsk dny 2012
- sbornk pspvk. Praha: esk betonsk spolenost SSI, 2012, s. 414-417. ISBN
978-80-87158-32-6. (in Czech).
[P9] Gringer, J. - Foglar, M.: Modeling of Deterioration Processes of Concrete Structures due
to Cyclic Loading and Aggressive Environment. In Proceedings of the 3rd Conference
100
 Publications

Nano and Macro Mechanics NMM 2012. Praha: Czech Technical University in Prague,
Faculty of Civil Engineering, 2012, p. 35-42. ISBN 978-80-01-05097-2.
[P10] Gringer, J.: Influence of Bridge Structural Arrangement on the Noise Induced by Traffic.
In Workshop doktorand katedry betonovch a zdnch konstrukc. Praha: Czech
Technical University in Prague, Faculty of Civil Engineering, 2012, art. no. 7, ISBN 978-
80-01-05037-8. (in Czech).
[P11] Gringer, J. - Foglar, M.: Investigation of the Combined Effect of Fatigue and Aggressive
Environment on Concrete. In Concrete under Severe Conditions. Cachan: RILEM
Publications, 2013, p. 1065-1075. ISBN 978-2-35158-124-7.
[P12] Gringer, J. - Foglar, M.: Performance of specimens subjected to cyclic loading and
aggressive environment. In CCC 2013 - Concrete Structures in Urban Areas. Wroclaw:
Dolnoslaskie wydawnictwo edukacyjne, 2013, p. 300-303. ISBN 978-83-7125-229-7.
[P13] Gringer, J. - Foglar, M.: Influence of the long term exposure to aggressive environment
on the fatigue performance of concrete specimens. In Engineering Mechanics 2013.
Praha: stav termomechaniky AV R, 2013, p. 130-135. ISSN 1805-8256. ISBN 978-
80-87012-47-5.
[P14] Gringer, J. - Foglar, M.: The Influence of the Bridge Structural Arrangement on the
Distribution of the Noise Impact Induced by Traffic. In Engineering Mechanics 2013.
Praha: stav termomechaniky AV R, 2013, p. 107-114. ISSN 1805-8256. ISBN 978-
80-87012-47-5.
[P15] Gringer, J. - Foglar, M.: Experimental Investigation of Concrete Specimens under
Severe Environment. In Proceedings of the 4th Conference Nano & Macro Mechanics.
Praha: Czech Technical University in Prague, 2013, p. 35-42. ISBN 978-80-01-05332-4.
[P16] Gringer, J.: Fatigue Properties of Concrete Specimens Exposed to Aggressive
Environment. In PhD Workshop. Praha: Czech Technical University in Prague, Faculty
of Civil Engineering, 2013, art. no. 4, ISBN 978-80-01-05236-5. (in Czech).
[P17] Gringer, J.: Evaluation of increase in the deflections of concrete specimens due to cyclic
loading. In PhD Workshop - Proceedings. Praha: Czech Technical University in Prague,
Faculty of Civil Engineering, 2014, art. no. 5, ISBN 978-80-01-05471-0. (in Czech).
[P18] Gringer, J.: Impact testing of deteriorated concrete specimens. In PhD Workshop 2015
- Full Versions. Praha: Czech Technical University in Prague, Faculty of Civil
Engineering, 2015, art. no. 6, ISBN 978-80-01-05722-3.
[P19] Foglar, M. - Gringer, J.: Influence of cyclic loading on the deflection development of
concrete specimens. Stavebn obzor. 2015, vol. 2015, no. 4, art. no. 24, ISSN 1805-2576.

101
 Publications

8.3 Utility models

[P20] Foglar, M. - Gringer, J.: Steel assembly bearing. Utility model, Industrial Property
Office, 27973. 2015-03-17. (in Czech).
[P21] Foglar, M. - Gringer, J.: Reinforced concrete frame with optimized static action. Utility
model, Industrial Property Office, 27974. 2015-03-17. (in Czech).

102