Professional Documents
Culture Documents
Alain Millard
Pierre Pimienta Editors
Modelling of
Concrete Behaviour
at High Temperature
State-of-the-Art Report of the RILEM
Technical Committee 227-HPB
RILEM State-of-the-Art Reports
RILEM STATE-OF-THE-ART REPORTS
Volume 30
Editors
Modelling of Concrete
Behaviour at High
Temperature
State-of-the-Art Report of the RILEM
Technical Committee 227-HPB
123
Editors
Alain Millard Pierre Pimienta
Service d’Études Mécaniques et Thermiques Centre scientifique et technique du bâtiment
Commissariat à l’énergie atomique et aux (CSTB), Safety, Structures and Fire
énergies alternatives (CEA) Université Paris-Est
Saclay, France Marne la Vallée, France
© RILEM 2019
No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any
means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written
permission from the Publisher, with the exception of any material supplied specifically for the purpose of
being entered and executed on a computer system, for exclusive use by the purchaser of the work.
Permission for use must always be obtained from the owner of the copyright: RILEM.
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Dedication and Acknowledgements
To the memory of Pr. Ulrich Schneider
This state-of-the-art report is dedicated to the
memory of Pr. Ulrich Schneider. Ulrich was
the chairman of the RILEM 227-HPB
Technical Committee. He mapped out our
path and shared with us his impressive, deep
and extensive scientific expertise and his
human value. He has left to the scientific
community an immense and rich heritage.
Foreword
This state-of-the-art report represents a valuable step forward with respect to the
available literature. Till now, other STARs were limited to a range of temperatures
not so far from the standard ones, up to 80 °C, or considered mainly ordinary
concrete with only some short descriptions of the behaviour of high-performance
concrete. Furthermore, the existing reports or monographs do not cover the most
recent progress in the field of computational modelling. In the last 20 years, the
computer processing power has increased dramatically, enabling engineers to for-
mulate and implement ever more sophisticated mathematical models, taking into
account increasingly complex aspects of the physical and chemical behaviour of
cementitious materials.
The analysis of concrete behaviour when the material is exposed to high tem-
perature is a crucial problem in several fields ranging from civil to nuclear engi-
neering and it is of paramount importance with regard to spalling. Spalling, which
can appear in various forms including explosive spalling, is especially dangerous
for concrete structures because it can lead to a generalized collapse. Taking into
account the technical and scientific progress in recent years allowing for a better
understanding of the topic, we can state that classical numerical models should not
be used for the simulation of the behaviour of concrete with respect to spalling.
In this state-of-the-art report, the mathematical/numerical models are classified
into two major groups:
Engineering models, which are based on the prescriptions of the Eurocode and
take into account essentially the thermo-mechanical aspects of the behaviour of the
material at high temperature;
Advanced models, referring to computational ones considering concrete as a
heterogeneous material with one or more fluid phases flowing in its pores.
For the first class of models, the report emphasizes their easy use for a large
range of structural design. At the same time, their limits are clearly identified and
described, especially as far as spalling is concerned.
In the second class of models, a plethora of approaches is nowadays available in
the literature based on different assumptions and taking into account different
physicochemical processes. This makes their description rather difficult, so the
vii
viii Foreword
authors of the STAR have chosen a top-bottom approach. First, a very complete
model based on multiphase porous media mechanics (i.e. MPMM), likely the most
sophisticated one currently available, is presented and discussed. Then, a set of
progressive simplifications are introduced to make this type of approach easier to
use in practical applications.
The structure of the STAR, the approach chosen for the description, and the
clarity of the presentation make the report a must for all those who want to approach
the rather complicated subject of heated concrete.
All the authors are extremely grateful to Katarzyna Mróz from Cracow University
of Technology and Colin Davie from Newcastle University for their very important
and efficient involvement in the formatting and review of the document.
ix
RILEM Technical Committee 227-HPB
(Physical Properties and Behaviour
of High-Performance Concrete at High
Temperature)
Chairmen
Secretary
Members
xi
xii RILEM Technical Committee 227-HPB …
The following list is presenting the global offer of RILEM Publications, sorted by
series. Each publication is available in printed version and/or in online version.
xiii
xiv RILEM Publications
Report 19: Considerations for Use in Managing the Aging of Nuclear Power
Plant Concrete Structures (ISBN: 2-912143-07-1); Ed. D. J. Naus
Report 20: Engineering and Transport Properties of the Interfacial Transition
Zone in Cementitious Composites (ISBN: 2-912143-08-X); Eds. M. G. Alexander,
G. Arliguie, G. Ballivy, A. Bentur and J. Marchand
Report 21: Durability of Building Sealants (ISBN: 2-912143-12-8); Ed. A. T. Wolf
Report 22: Sustainable Raw Materials—Construction and Demolition Waste
(ISBN: 2-912143-17-9); Eds. C. F. Hendriks and H. S. Pietersen
Report 23: Self-Compacting Concrete state-of-the-art report (ISBN:
2-912143-23-3); Eds. Å. Skarendahl and Ö. Petersson
Report 24: Workability and Rheology of Fresh Concrete: Compendium of Tests
(ISBN: 2-912143-32-2); Eds. P. J. M. Bartos, M. Sonebi and A. K. Tamimi
Report 25: Early Age Cracking in Cementitious Systems (ISBN:
2-912143-33-0); Ed. A. Bentur
Report 26: Towards Sustainable Roofing (Joint Committee CIB/RILEM) (CD
07) (e-ISBN 978-2-912143-65-5); Eds. Thomas W. Hutchinson and Keith Roberts
Report 27: Condition Assessment of Roofs (Joint Committee CIB/RILEM) (CD
08) (e-ISBN 978-2-912143-66-2); Ed. CIB W 83/RILEM TC166-RMS
Report 28: Final report of RILEM TC 167-COM ‘Characterisation of Old
Mortars with Respect to Their Repair (ISBN: 978-2-912143-56-3); Eds. C. Groot,
G. Ashall and J. Hughes
Report 29: Pavement Performance Prediction and Evaluation (PPPE):
Interlaboratory Tests (e-ISBN: 2-912143-68-3); Eds. M. Partl and H. Piber
Report 30: Final Report of RILEM TC 198-URM ‘Use of Recycled Materials’
(ISBN: 2-912143-82-9; e-ISBN: 2-912143-69-1); Eds. Ch. F. Hendriks,
G. M. T. Janssen and E. Vázquez
Report 31: Final Report of RILEM TC 185-ATC ‘Advanced testing of
cement-based materials during setting and hardening’ (ISBN: 2-912143-81-0;
e-ISBN: 2-912143-70-5); Eds. H. W. Reinhardt and C. U. Grosse
Report 32: Probabilistic Assessment of Existing Structures. A JCSS publication
(ISBN 2-912143-24-1); Ed. D. Diamantidis
Report 33: State-of-the-Art Report of RILEM Technical Committee TC
184-IFE ‘Industrial Floors’ (ISBN 2-35158-006-0); Ed. P. Seidler
Report 34: Report of RILEM Technical Committee TC 147-FMB ‘Fracture
mechanics applications to anchorage and bond’ Tension of Reinforced Concrete
Prisms—Round Robin Analysis and Tests on Bond (e-ISBN 2-912143-91-8); Eds.
L. Elfgren and K. Noghabai
Report 35: Final Report of RILEM Technical Committee TC 188-CSC ‘Casting
of Self Compacting Concrete’ (ISBN 2-35158-001-X; e-ISBN: 2-912143-98-5);
Eds. Å. Skarendahl and P. Billberg
RILEM Publications xxiii
1 Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Alain Millard and Pierre Pimienta
2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Alain Millard
3 Engineering Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Sven Huismann, Matthias Zeiml, Manfred Korzen and Alain Millard
4 Advanced Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Fekri Meftah, Francesco Pesavento, Colin Davie, Stefano Dal Pont,
Matthias Zeiml, Manfred Korzen and Alain Millard
5 Constitutive Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Fekri Meftah, Colin Davie, Stefano Dal Pont and Alain Millard
6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Alain Millard and Pierre Pimienta
xxv
Contributors
xxvii
Chapter 1
Scope
Abstract This book presents the work done by the RILEM Technical Committee
227-HPB (Physical properties and behaviour of High-Performance Concrete at high
temperature). It contains the latest research results on the modelling of concrete
behaviour at high temperature. Some monographs on the subject have been pub-
lished already but generally they do not cover the whole range of possibilities which
are encountered in the literature as well as in practice. Moreover, there has been a
rapidly increasing development of computational models during the last twenty
years, which deserves attention. Therefore, it is the aim of this report to compile and
present most of the tools that are proposed in the literature and are nowadays
available for practice in some commercial computational packages. This chapter
presents the main literature produced during the last decades on the behaviour at
high temperature of concrete. This literature mainly focus on 2 main fields of
application: the structural response and safety analysis of buildings and structures
under fire and the nuclear industry and in particular the nuclear reactor containment
buildings generally made of pre-stressed concrete. A special attention is paid to
high performance concrete and their specific properties.
This report presents part of the work done by the RILEM Technical Committee
227-HPB (Physical properties and behaviour of High-Performance Concrete at
high temperature). This Committee was created in 2007 and was adjourned in 2014.
A. Millard (&)
Service d’Études Mécaniques et Thermiques, Commissariat à l’Énergie Atomique
et aux Énergies Alternatives (CEA), Saclay, France
e-mail: alain.millard@cea.fr
P. Pimienta
Centre Scientifique et Technique du Bâtiment (CSTB), Université Paris-Est,
Marne la Vallée, France
e-mail: pierre.pimienta@cstb.fr
© RILEM 2019 1
A. Millard and P. Pimienta (eds.), Modelling of Concrete Behaviour
at High Temperature, RILEM State-of-the-Art Reports 30,
https://doi.org/10.1007/978-3-030-11995-9_1
2 A. Millard and P. Pimienta
This committee assured the continuity of the work that had been done by the
RILEM Technical Committee 200-HTC (Mechanical concrete properties at high
temperatures-modelling and applications) and previous ones on ordinary concretes.
The main objectives of the TC HPB were to write the State-of-the-Art on two
areas of research. The first, which is published in this book, focuses on the mod-
elling of the behaviour of concrete at high temperatures. The second is published in
a second book and gathers and analyses the latest experimental research results on
the behaviour of high-performance concretes (HPC) at high-temperature (Pimienta
et al. 2018).
Because of their actual significant use and their specific properties at high
temperature, special attention will be focus to HPC through the document.
The field of application of such tools is constantly enlarging. First, the tem-
peratures that are considered range from 20 °C up to the melting point of concrete,
around 1350 °C. These temperatures can be obtained either in service or in fault
conditions, depending on the nature and serviceability of the structure. Of course,
one main field of application is related to the structural response and safety analysis
of buildings and structures under fire. In particular, recent accidental fires in railway
as well as roadway tunnels have highlighted the need to improve the safety of the
structures under fire, as well as their design, in particular by improving the concrete
performances under high temperature. This trend concerns not only concrete
structures but, more generally, all types of cementitious-like materials (such as
mortar, plaster, etc.) which are routinely used in buildings, and for which the same
type of approaches can be contemplated.
Another main field of applications concerns the nuclear industry. A first
important issue is related to the safety analysis of nuclear reactor containment
buildings generally made of pre-stressed concrete. Although the temperature
increase up to about 210 °C in the case of a Loss of Coolant Accident (LOCA) in a
Pressurized Water Reactor (PWR) is already a problem, advanced reactors such as
liquid metal cooled reactors or high temperature gas cooled reactors are more
critical, with temperatures reaching as high as 1150 °C (Schimmelpfennig and
Altes 1982). In the case of a severe accident, several scenarios that can lead to high
temperatures are studied, such as the spillage of melted core on to the containment
basemat, or the thermal shock due to liquid metal on concrete walls. Another
important issue is related to the intermediate or long-term surface storage of
high-level nuclear wastes. In the case of a failure of the cooling system, tempera-
tures up to 250 °C might be encountered.
HPC is now often used because of its improved strength and durability compared
to OPC. These increased performances are mainly obtained by means of reduced
water-cement ratios and use of special additives such as silica-fume. As a conse-
quence, the porosity and the permeability are reduced, thus improving the resistance
of concrete to the ingress of chloride ions or carbon dioxide, which can in the
long-term lead to the active corrosion of the reinforcement.
A further consequence is that the behaviour of HPC under high temperatures
may be rather different from that of OPC, especially in the case of rapid heating, as
can be encountered during a fire. In particular, spalling may occur, leading to the
detachment of concrete pieces or layers from the surface, and to the subsequent
exposure of the remaining structure to the fire, thus accelerating the degradation
process and finally the collapse of the structure.
The needs for an improved design of HPC structures subjected to high tem-
peratures, as well as their performance and safety assessments, have boosted the
development of analytical and numerical tools during the last three decades. In
particular, in view of its inherent complexity as outlined in the next sections, a
satisfactory explanation of spalling requires a clear understanding not only of the
fundamental mechanisms but also of their orders of magnitude. This can only be
done by resorting to experimental tests complemented by modelling, because of the
non-homogeneous response of the specimens that behave like true structures.
4 A. Millard and P. Pimienta
Moreover, the cost and time required by experimental testing can be significantly
reduced by appropriate modelling investigations, which allow a better under-
standing of what is happening, what the dominating physical mechanisms are and
the most influential parameters.
In the next section, a description of the main phenomena and processes taking
place in concrete structures exposed to high temperature, is given for a better
understanding of the difficulties faced in the formulation of models that can be
considered as really predictive of the structure behaviour.
References
Bažant, Z.P., Kaplan, M.F.: Concrete at High Temperatures: Material Properties and Mathematical
Models. Longman, Harlow (1996)
CEB-FIP: Fire design of concrete structures – materials, structures and modelling, State-of-the-art
report, Bulletin 38. International Federation for Structural Concrete (fib) (2007)
CEB-FIP: Constitutive modelling of high strength/high performance concrete, State-of-the-art
report, Bulletin 42. International Federation for Structural Concrete (fib) (2008)
Pimienta P., Jansson McNamee R. and Mindeguia J-C. (Eds.) “Physical Properties and Behaviour
of High-Performance Concrete at High Temperature - State-of-the-Art Report of the RILEM
Technical Committee 227-HPB” Book published by Springer (2018)
RILEM Committee 44-PHT: Properties of materials at high temperature – Concrete. U. Schneider
(eds), Dept. of Civil Engineering, Kassel University (1985)
RILEM Committee HTC: Behaviour of ordinary concrete at high temperatures, U. Schneider and
J. Howarth (eds), Institute of Building Materials, Vienna University of Technology (2002)
RILEM TC 129-MHT: Test methods for mechanical properties of concrete at high temperatures:
Part 3: Compressive strength for service and accident conditions, Materials and Structures 28
(7), 410–414 (1995)
RILEM TC 129-MHT: Test methods for mechanical properties of concrete at high temperatures,
Recommendations Part 6: Thermal strain. Materials and Structures, Supplement March 1997,
17–21 (1997)
RILEM TC 129-MHT: Test methods for mechanical properties of concrete at high temperatures,
Recommendations, Part 7: Transient creep for service and accident condition. Materials and
Structures 31, 290–295 (1998)
RILEM TC 129-MHT: Test methods for mechanical properties of concrete at high temperatures,
Part 4: Tensile Strength for service and accident conditions. Mater. Struct. 33(228), 219–223
(2000a)
RILEM TC 129-MHT: Test methods for mechanical properties of concrete at high temperatures,
Recommendations, Part 8: Steady-state creep and creep recovery for service and accident
conditions. Materials and Structures 33, 6–13 (2000b)
RILEM TC 129-MHT: Test methods for mechanical properties of concrete at high temperatures,
Recommendations, Part 9: Shrinkage for service and accident conditions.Materials and
Structures 33, 224–228 (2000c)
RILEM TC 129-MHT: Test methods for mechanical properties of concrete at high temperatures:
Part 5: Modulus of elasticity for service and accident conditions, Materials and Structures 37
(2) 139–144 (2004)
RILEM TC 200-HTC: Test methods for mechanical properties of concrete at high temperatures,
Recommendations, Part 10: Restraint stress. Materials and Structures 38, 913–919 (2005)
1 Scope 5
RILEM TC 200-HTC: Test methods for mechanical properties of concrete at high temperatures,
Recommendation, Part 1: Introduction-General presentation. Materials and Structures 40(9),
841–853 (2007a)
RILEM TC 200-HTC: Mechanical concrete properties at high temperatures - modelling and
applications, Part 2: Stress-strain relation, Materials and Structures 40(9) 855–864 (2007b)
RILEM TC 200-HTC: Mechanical concrete properties at high temperatures-modelling and
applications, Part 11: Relaxation. Materials and Structures 40, 449–458 (2007c)
Schimmelpfennig, K., Altes, J.: Special aspects on the behaviour of PCRV under extremely high
core temperature loading. Nuclear Engineering and Design. 75, 291–302 (1982)
Schneider U.: Kinetic of strength reducing reactions of concrete under high temperatures. PhD
Dissertation (1973)
Schneider U.: A contribution to creep and relaxation of concrete at elevated temperatures,
Habilitation Thesis (1979)
Schneider U.: Behaviour of concrete at high temperatures. Deutscher Ausschuss für Stahlbeton,
Heft 337 (1982)
Schneider, U.: Modelling of concrete behaviour at high temperatures. In: Anchor, R.D, Malhotra,
H.L. and Purkiss, J.A. (eds.) Design of structures against fire, pp. 53–69. Elsevier Applied
Science Publishers, London (1986)
Schneider, U.: Concrete at high temperatures – A general review. Fire Safety Journal 13(1), 55–68
(1988)
Schneider, U., Horvath, J., Behaviour of ordinary concrete at high temperature, distributed by,
Institute of Building Materials, Vienna University of Technology, Austria, (2003)
Chapter 2
Introduction
Alain Millard
A. Millard (&)
Service d’Études Mécaniques et Thermiques, Commissariat à l’Énergie
Atomique et aux Énergies Alternatives (CEA), Saclay, France
e-mail: alain.millard@cea.fr
© RILEM 2019 7
A. Millard and P. Pimienta (eds.), Modelling of Concrete Behaviour
at High Temperature, RILEM State-of-the-Art Reports 30,
https://doi.org/10.1007/978-3-030-11995-9_2
8 A. Millard
expansion, which is particularly important above the temperature of about 160 °C,
(Gawin et al. 2002a). These effects cause a significant decrease of the gas perme-
ability, because the space available for the gas is decreased.
Increasing temperature causes the material expansion which in part is due to
concrete dehydration (products of the thermal dissolution of concrete components
have greater volume than their initial volume), in part due to the material cracking
and progressive cracks opening (Schneider 1988), and finally due to ‘normal’
thermal expansion of the material skeleton. The concrete cracking during heating is
caused by an incompatibility of thermal expansion of the aggregate and the cement
paste, resulting in high traction stresses and development of local micro-cracks.
Due to these cracks and chemical transformations of concrete (generally called
dehydration), the concrete strength properties degrade gradually (Gawin et al.
2003, 2005).
10 A. Millard
2.2 Modelling
Concerning design purposes, codes and standards generally apply to classical sit-
uations, i.e. typical structural members, such as beams, columns and plates. In case
of non-conventional projects, it is necessary to turn to specific modelling tools to
demonstrate, for example, the fire resistance of a building. Moreover, modelling
tools can nowadays be contemplated as possible tools to optimize the HPC for-
mulation itself, in order to improve its behaviour under high temperature.
Obviously, there are a variety of modelling approaches, which on one hand
reflect the complexity of the phenomena, and on the other hand correspond to the
expected level of accuracy in the description of these phenomena. The most
straightforward approach is based on the so-called prescriptive methods. It consists
in assessing the load bearing capacity of a given structure (which can be either a
structural member or a substructure) subjected to high temperatures, by applying
codes and standards, such as Eurocode 2 (EN 1992-1-2 2004) for structural fire
design. For example, in Eurocode 2, geometrical tabulated data are available to
allow a design which guarantees the resistance of the structure under high tem-
perature for a given time.
Although very simple, this approach can be too severe and does not cover all the
structural configurations (heating scenarios, geometries, loadings, support condi-
tions, etc.), which can be encountered in practice. Therefore, there is an increasing
trend to have recourse to performance-based approaches in order to optimize
2 Introduction 11
structure design. It consists in studying the response of part of the structure or the
entire structure during the whole real heating process. When fire safety engineering
approach is applied, the cooling phase is also taken into account. For this purpose,
computer simulations are generally required, which can be more or less sophisti-
cated, the differences coming from the scale and the level of description of the
physico-chemical mechanisms. Nevertheless, whatever the level of accuracy or
complexity, a thermal analysis is required to determine the evolution of the tem-
perature distribution in the structure.
A common practice consists in adopting a thermo-mechanical framework,
together with phenomenological models. In this context, the simplest model is
based on a simplified limit state analysis, known as the “isotherm method”. In this
method, a temperature threshold is defined (around 400 °C for HPC), above which
the concrete is supposed to be totally damaged, while below, the concrete is not
affected by the temperature increase. Even though this simple method gives good
results for example in case of normalized fires, it may give non-conservative results
in case of slow heating rates and does not take into account indirect fire actions.
Moreover, it cannot be used for the cooling phases since it does not capture the
irreversible changes of the concrete strength.
Most of the routinely made calculations resort to thermo-mechanical analysis
where, in general, the temperature evolution is coupled to the mechanical analysis
through the thermal strains and the variation of the mechanical properties (Young’s
modulus, thermal expansion coefficient, compressive strength, etc.), whereas the
mechanical evolution has no significant influence on the temperature. Therefore, the
transient thermal and mechanical analysis are performed sequentially. Despite the
fact that major phenomena like moisture transfer and chemical reactions, are not
explicitly modeled, the thermo-mechanical approach can predict rather accurate
results provided that ad-hoc phenomenological equations are established to account
indirectly for their effect. For example, the variation of the concrete specific heat
with respect to the temperature accounts for different hydro-chemical phenomena
such as the vaporization of the free water around 100 °C, the decomposition of
calcium carbonate above 400 °C, etc. A great effort must be put on the constitutive
models in order to be able to describe the various strains which develop during
heating, such as drying shrinkage, transient creep and transitional thermal creep
strains, etc. These strain components can be carefully identified from experimental
measurements, following for example the recommendations for test methods issued
by the RILEM Committee 129-MHT (see references in Chap. 1).
However, one main underlying assumption in these models is to consider the
concrete as behaving like a sealed or an unsealed specimen, with respect to moisture
transfer. It is obvious that, in reality, the situation is much more complex and
generally lies between these two extreme cases, depending on the time-scale con-
sidered. Moreover, these models cannot account for the changes in pore pressure
due to heating, and in particular to the vapour pressure increase which can partly
explain spalling.
12 A. Millard
References
Dal Pont S., Ehrlacher A.: Numerical and experimental analysis of chemical dehydration, heat and
mass transfer in a concrete hollow cylinder submitted to high temperatures. International
Journal of Heat and Mass transfer 1(47):135–147 (2004)
England, G.L., Khoylou, N.: Moisture flow in concrete under steady state non-uniform temperature
states: experimental observations and theoretical modelling. Nuclear Engineering Design 156,
83–107 (1995)
EN 1992-1-2, Eurocode 2 Design of concrete structures, Part 1.2 General rules – Structural fire
design, (2004)
Gawin, D., Alonso, C., Andrade, C., Majorana, C.E., Pesavento, F.: Effect of damage on
permeability and hygro-thermal behaviour of HPCs at elevated temperatures, Part 1.
Experimental results. Computers and Concrete 2(3), 189–202 (2005)
Gawin, D., Pesavento, F., Schrefler, B.A.: Modelling of hygro-thermal behaviour and damage of
concrete at temperature above critical point of water. International Journal of Numerical and
Analytical Methods in Geomechanics 26(6), 537–562 (2002a)
Gawin, D., Pesavento, F., Schrefler, B.A.: Modelling of thermo-chemical and mechanical damage
of concrete at high temperature. Computer Methods in Applied Mechanics and Engineering
192, 1731–1771 (2003)
Gawin, D., Pesavento, F., Schrefler, B.A.: Simulation of damage–permeability coupling in
hygro-thermo-mechanical analysis of concrete at high temperature. Communications in
Numerical Methods in Engineering 18(2), 113–119 (2002b)
Gawin, D., Pesavento, F., Schrefler, B.A.: Towards prediction of the thermal spalling risk through
a multi-phase porous media model of concrete. Computer Methods in Applied Mechanics and
Engineering 195(41–43), 5707-5729 (2006)
Kalifa, P., Menneteau, F.D., Quenard, D.: Spalling and pore pressure in HPC at high temperatures.
Cement and Concrete Research 30, 1915–1927 (2000)
2 Introduction 13
Abstract This chapter presents the most commonly used approach to analyse the
thermo-mechanical behaviour of concrete structures subjected to high temperatures
as in the case of fire loading. Prescriptions of the Eurocode are detailed for the
thermal as well as the mechanical analysis. Finally, recommendations from two
national (Austrian and German) guidelines give some improvements for under-
ground infrastructure.
© RILEM 2019 15
A. Millard and P. Pimienta (eds.), Modelling of Concrete Behaviour
at High Temperature, RILEM State-of-the-Art Reports 30,
https://doi.org/10.1007/978-3-030-11995-9_3
16 S. Huismann et al.
3.2 Eurocode
where, Ed,fi is the design effect of actions for the fire situation, determined in
accordance with Eurocode 1 (EN 1991-1-2, 2010), including effects of thermal
expansions and deformations, and Rd,t,fi is the corresponding design resistance in
the fire situation.
In general, the fire design of concrete structures or structural components is
performed in two consecutive steps. First, a thermal analysis gives access to the
temperature distribution within the member. Within the subsequent mechanical
analysis, the previously-determined temperature field serves as input for the sim-
ulation of the mechanical behaviour of the structural member and/or structure.
Fig. 3.1 Thermal material properties of concrete according to Eurocode 2 (EN 1992-1-2, 2004):
a thermal conductivity and b specific heat for different values of the moisture content by weight
For fire loading following the standard temperature-time fire curve according to
EN 1991-1-2, 2010, the corresponding temperature field can be determined directly
from Eurocode 2 where pre-calculated temperature distributions are available for
typical types of members (e.g. slabs, beams, columns) and selected time durations
(i.e., 30, 60, 90, 120, 180, 240 min).
Within the design process, the following design methods can be used in order to
satisfy condition (3.1):
18 S. Huismann et al.
3.5 Spalling
Fig. 3.2 Stress-strain relationship for concrete under compression at elevated temperatures (EN
1992-1-2, 2004)
3 Engineering Modelling 21
Fig. 3.3 Stress-strain relationship for concrete under compression at elevated temperatures (EN
1992-1-2, 2004)
specified in the national annex to Eurocode 2 for each country. The recommended
value is 3%. For moisture contents greater than k, a more accurate assessment of
moisture content, type of aggregate, permeability of concrete and heating rate
should be considered. From a scientific point of view, this simplistic recommen-
dation is to be questioned, since it is commonly accepted that spalling is caused by
the interaction of numerous thermo-hydro-chemo-mechanical processes with the
moisture content being only one of numerous influencing parameters.
Since most of the recommendations of the Eurocode are valid only for fire-loading
scenarios up to the standard temperature-time fire, national guidelines were estab-
lished for more severe fire-loading scenarios. Such fire-loading scenarios are to be
considered, e.g. in tunnels and other underground structures.
Austrian ÖBV-guideline “Improved structural fire protection with concrete
for underground infrastructure” (ÖBV 2013)
This Austrian guideline was established in order to specify the details and typical
areas of application of numerical models with different levels of complexity within
the model category defined in the Eurocode 2 as advanced calculation methods.
In this guideline it is suggested that determination of the thermal loading and the
resulting temperature distributions within the structural members should either be
based on a thermal analysis according to Eurocode 2 or on empirical temperature
distributions determined from fire experiments (Kusterle et al. 2005) with
22 S. Huismann et al.
fire-loading conditions typical for tunnel fires (see Fig. 3.4 and (ÖBV 2013) for
details as well as the numerical values corresponding to the displayed curves).
For the subsequent mechanical analysis following the recommendations given in
Eurocode 2, (ÖBV 2013) distinguishes between linear and nonlinear model
assumptions:
– linear model assumptions:
In order to be used in standard linear-elastic beam-spring models, the nonlinear
temperature distribution resulting from the thermal analysis is transformed into a
so-called equivalent temperature (see Kusterle et al. 2005; Ring et al. 2013 for
details). The internal forces caused by the governing load combination for the
design case fire are determined considering linear-elastic material behaviour of
both concrete and reinforcing steel.
– nonlinear model assumptions:
In the case that nonlinear model assumptions are applied, the nonlinear tem-
perature distributions are incorporated directly into the mechanical analysis
model. The internal forces are determined considering elasto-plastic material
behaviour and the mechanical material parameters according to Eurocode 2 (see
Sect. 3.5 for details).
Based on the above statements of linear and nonlinear model assumptions,
recommendations are given on when linear or nonlinear model assumptions are to
be applied. Based on numerical simulations of fire experiments and benchmark
examples (see Ring 2012; Ring et al. 2013a, b for details), the main influencing
parameters determining the choice of the model assumptions are the duration of the
Fig. 3.4 Empirical temperature distributions for structural fire design according to (ÖBV 2013)
(curves are valid for PP-fibre reinforced concrete with a moisture content comparable to (Kusterle
et al. 2004) and a dry density between 2000 and 2600 kg/m3; curves cover the following
fire-loading scenarios: RWS, HCinc, HC, EUREKA, ISO 834)
3 Engineering Modelling 23
fire (i.e. the required fire resistance) and the complexity of the structure (e.g.
rectangular, frame-like structures, arched or circular cross sections).
German DB-guideline 853 “Design, construction and maintenance of railway
tunnels” (DB-guideline 853 2011)
This German guideline deals with, among numerous other topics, the structural
design of tubbing tunnels and recommends a specific type of advanced calculation
method (as defined in Eurocode 2). Hereby, the special model feature of tubbing
tunnels, namely nonlinear hinges representing the joints between tubbing segments,
is incorporated into the model by the following design steps:
– determination of time-dependent internal forces due to thermal restraint:
The structural model for determining the internal forces due to thermal restraint
represents a tubbing tunnel without nonlinear hinges representing the joints. The
overall stiffness of the numerical model is determined at the time instant where
the restraint forces reach their maximum values.
– determination of the restraint internal forces for a unit temperature loading:
With the same structural model as in the previous step, the internal forces due to
thermal restraint are determined with a unit temperature loading (i.e. a fixed
value for the components of the so-called equivalent temperature).
– determination of unit temperatures for the given structural model:
The unit temperatures for the structural model under consideration are deter-
mined by the quotient of the internal stress resultants of the two
above-mentioned analysis steps.
– fire simulation of final tunnel structure:
With the temperature values determined in the previous analysis step, the final
structural model (i.e. tubbing tunnel including the nonlinear hinges representing
the joints between tubbing segments) is analysed. Hereby, the previously
determined stiffness is considered.
The reason for recommending this rather laborious analysis scheme is that
common engineering software tools can consider nonlinear material behaviour
together with nonlinear temperature distributions (that typically develop in concrete
with its low heat diffusivity), but are not able to consider these features together
with nonlinear hinges at nodes, which are necessary to take into account all relevant
nonlinearities present in common tubbing tunnels. Nonlinear hinges are currently
incorporated only in standard beam-spring models which in turn cannot incorporate
nonlinear temperature distributions. Since software tools will evolve in the future,
eventually becoming capable of incorporating all above-mentioned nonlinearities in
one numerical model, the described analysis scheme will become obsolete and a
direct non-linear analysis (in the sense of an advanced calculation method
according to Eurocode 2) will be possible.
24 S. Huismann et al.
References
Franssen, J.-M.: Plastic analysis of concrete structures subjected to fire. Fire design of concrete
structures: What now? What next?, Milan, 133–145 (2004)
ÖBV-guideline “Improved structural fire protection with concrete for underground infrastructure”
(“Erhöhter baulicher Brandschutz mit Beton für unterirdische Verkehrsbauwerke”). Austrian
Society for Construction Technology, in German (2013)
Kusterle, W., Lindlbauer, W., Hampejs, G., Heel, A., Donauer, P.-F., Zeiml, M., Brunnsteiner, W.,
Dietze, R., Hermann, W., Viechtbauer, H., Schreiner, M., Vierthaler, R., Stadlober, H., Winter,
H., Lemmerer, J., Kammeringer, E.: Brandbeständigkeit von Faser-, Stahl- und Spannbeton
[Fire resistance of fiber-reinforced, reinforced, and prestressed concrete]. Tech. Rep. 544,
Bundesministerium für Verkehr, Innovation und Technologie, Vienna, in German (2004)
Ring, T., Zeiml, M., Lackner, R.: Underground concrete frame structures subjected to fire loading:
Part II – re-analysis of large-scale fire tests. Engineering Structures, in print (2013a)
Ring, T., Wikete, C., Kari, H., Zeiml, M., Lackner, R.: Der Einfluss des Rechen- und
Materialmodells auf die Strukturantwort bei der Simulation von Tunnel unter Brandbelastung
[The influence of numerical and material model on the structural response of tunnels subjected
to fire loading], Bauingenieur 88, p. 35–44, in German (2013b)
Ring, T.: Experimental characterization and modelling of concrete at high temperatures – structural
safety assessment of different tunnel cross-sections subjected to fire loading. PhD thesis,
Vienna University of Technology, Vienna, Austria (2012)
DB-guideline 853 “Design, construction and maintenance of railway tunnels” (“Eisenbahntunnel
planen, bauen und instandhalten”). Deutsche Bahn AG, in German (2011)
Chapter 4
Advanced Modelling
F. Meftah (&)
Institut National des Sciences Appliquées, Rennes, France
e-mail: fekri.meftah@insa-rennes.fr
F. Pesavento
Dipartimento di Ingegneria Civile, Edile e Ambientale, Universita degli
Studi di Padova, Padua, Italy
e-mail: pesa@dicea.unipd.it
C. Davie
School of Engineering, Newcastle University, Newcastle upon Tyne, UK
e-mail: colin.davie@newcastle.ac.uk
S. D. Pont
Laboratoire 3SR, Université Grenoble-Alpes, Grenoble, France
e-mail: stefano.dalpont@3sr-grenoble.fr
M. Zeiml
Material Technology, University of Innsbruck, Innsbruck, Austria
e-mail: matthias.z@gmx.at
M. Korzen
Bundesanstanlt for Materialforschung und prüfung (BAM), Berlin, Germany
e-mail: manfred.korzen@bam.de
A. Millard
Service d’Études Mécaniques et Thermiques, Commissariat à l’Énergie
Atomique et aux Énergies Alternatives (CEA), Saclay, France
e-mail: alain.millard@cea.fr
© RILEM 2019 27
A. Millard and P. Pimienta (eds.), Modelling of Concrete Behaviour
at High Temperature, RILEM State-of-the-Art Reports 30,
https://doi.org/10.1007/978-3-030-11995-9_4
28 F. Meftah et al.
4.1 Introduction
The advanced models can be classified into two main categories: models essentially
focused on the thermo-mechanical behaviour of concrete, and the models consid-
ering also the fluid phases (one or more) inside the pores of the material.
Firstly, we will discuss about the models considering the fluids inside the pore
network of concrete. Then, we will illustrate the more commonly used, purely
thermo-mechanical models (Sect. 4.6).
The first class of models aims at the simulation of the overall thermo-hygral and
mechanical behaviour of the material, considering sometimes also some chemical
aspects (e.g. the reactions of hydration and dehydration). To analyse hygro-thermal
phenomena in porous media, two different approaches are currently used: phe-
nomenological and mechanistic ones. In phenomenological approaches (Bažant and
Thonguthai 1978, 1979; England and Khoylou 1995; Abdel-Rahman and Ahmed
1996) moisture and heat transport are described by diffusive type differential
equations with temperature- and moisture content-dependent coefficients. The
model coefficients are usually determined by inverse problem solution, starting
from experimental tests results, in order to obtain the best agreement between
theoretical prediction and experimental evidence. A characterisitc of such models is
that they are very accurate for interpolation, but rather poor for extrapolation of the
known experimental results. Moreover, various physical phenomena are lumped
together and important processes, such as phase changes, especially at high tem-
peratures where they play a relevant role, are neglected. Mechanistic models
(Bažant and Kaplan 1996; Consolazio et al. 1998; Ulm et al. 1999a, b; Gawin et al.
1999, 2002a) are more complicated from a mathematical point of view and, con-
trary to phenomenological ones, their coefficients have a clear physical meaning.
This kind of model is often obtained from microscopic balance equations written
for each constituent of the medium, which are then averaged in the space applying
special averaging operators.
Several mathematical and numerical models, usually based on extensive labo-
ratory tests, have been developed for the analysis of heat and mass transfer in
concrete at high temperature, e.g. Gawin et al. (2003, 2004, 2006), Tenchev et al.
(2001, 2005), Ichikawa and England (2004), Davie et al. (2006), Chung et al.
(2006), Dwaikat and Kodur (2009). However, one should remember that every
interpretation possesses all limitations and simplifications that are assumed in the
mathematical model on which the simulations are based. Thus, for this purpose one
should use models considering possibly the whole complexity and mutual inter-
actions of the analysed physical processes. Hence, in the case of concrete at high
temperature, a proper choice may be models based on mechanics of multiphase
porous media, taking into account chemical reactions (dehydration), phase changes,
cracking and thermo-chemical material degradation, as well as their mutual cou-
plings and influence on the hygral, thermal, chemical and mechanical properties of
concrete. These physical aspects of the behaviour of concrete subjected to high
temperature are described below.
4 Advanced Modelling 29
and the gas pressure pg = pv + pa (pv being the pressure of water vapour and pa the
pressure of the dry air) is called the capillary pressure pc and is a function of the
liquid water saturation Sl:
pc ðSl Þ ¼ pg pl ð4:1Þ
Assuming the contact angle c between the liquid phase and the solid matrix to be
zero, the capillary pressure of water pc can be related to the pore radius r with the
Laplace equation:
2rðT Þ
pc ¼ ¼ 2rðT Þv ð4:2Þ
r
where r(T) is the surface tension of water which depends upon temperature, and v
is the curvature of the capillary meniscus. Any change in the curvature of meniscus
will change the equilibrium between liquid and vapour phases. A relationship
between the liquid water and the vapour can be obtained by means of Kelvin’s
equation considering that liquid is incompressible and the vapour is a perfect gas.
Water can exist in the hydrated cement paste in many forms:
• capillary and physically adsorbed water; their loss is mainly responsible for the
shrinkage of the material while drying
• interlayer water; it can be lost only during strong drying, which leads to con-
siderably shrinkage of the C-S-H structure
All these types of water can evaporate at a temperature of 105 °C (at atmo-
spheric pressure and slow rate of heating). Another type of water, which is
non-evaporable, is chemically bound water. It is considered to be an integral part of
the structure of various cement hydration products and it is released when the
4 Advanced Modelling 31
x- interlayer water
o - adsorbed water
A - inter particles links
B - CSH layers
Fig. 4.3 Probable structure of hydrated silicates (Feldman and Sereda 1968)
used in modelling, giving reasonable results from a physical point of view and in
part confirmed by the available experimental results.
Concrete is here considered to be a multiphase medium where the voids of the
solid skeleton could be filled with various combinations of liquid- and gas-phases
(Fig. 4.2). In the specific case the fluids filling the pore space are the moist air
(mixture of dry air and vapour), capillary water and physically adsorbed water. The
chemically bound water is considered to be part of the solid skeleton until it is
released on heating.
Below the critical temperature of water Tcr, the liquid phase consists of physi-
cally adsorbed water, which is present in the whole range of moisture content, and
capillary water, which appears when the degree of water saturation Sl exceeds the
upper limit of the hygroscopic region, Sssp (i.e. below Sssp there is only physically
adsorbed water). Above the temperature Tcr the liquid phase consists of the
adsorbed water only. In the whole temperature range the gas phase is a mixture of
dry air and water vapour.
Thanks to the mechanics of porous media and the schematisation of the material
described above, it is possible to analyse three different classes of phenomena
which characterise the behaviour of concrete at high temperature:
• Hygral phenomena
• Thermal phenomena
• Mechanical phenomena
Fig. 4.4 Mass transport mechanism in a non-saturated porous medium subjected to heating (Gens
and Olivella 2001)
Concerning thermal phenomena, it can be stated that in most cases the main
mechanism for heat transport is heat conduction. Heat conduction responds to
gradients of temperature T. However, additional heat transfer will also be accom-
plished by advection due to the movement of the three phases: solid, liquid and gas.
The latent heat inherent to phase changes may also have significant thermal effects
(Gens and Olivella 2001) (Fig. 4.5).
The evolution of the temperature distribution in any structure is governed by the
thermal properties of the material, particularly heat capacity and thermal conduc-
tivity. In the case of concrete, it is difficult to determine these properties because of
the numerous phenomena that occur simultaneously within the microstructure of
concrete and cannot be separated easily. These phenomena are affected in particular
34 F. Meftah et al.
Fig. 4.5 Heat transport mechanism in a porous medium (Gens and Olivella 2001)
by the evolution of the porosity, by the moisture content, by the type and amount of
aggregate, by changes in the chemical composition and by the latent heat con-
sumption generated by certain chemical phenomena. Because of these effects, a
unique relationship cannot rigorously describe the dependence of concrete prop-
erties on temperature (Harmathy 1965, 1970).
When mechanically loaded and simultaneously heated, the overall measured strain
of concrete is assumed as an additive combination of different components. These
components can be conventionally classified into three families according to the
origin of the driving mechanisms:
– Mechanical strains that occur due to an applied mechanical load only. Elastic
strain, cracking strain and basic creep strain are the main mechanical
components
4 Advanced Modelling 35
The general approach to heat and mass transfer processes in a partially saturated
open porous medium is to start from a set of balance equations governing the time
evolution of mass and heat of the solid matrix and the fluids filling the porous
network, taking into account the exchange between the phases and with the sur-
rounding medium. These balance equations are supplemented with an appropriate
set of constitutive relationships, which permit a reduction in the number of inde-
pendent state variables that control the physical process under investigation
(Hassanizadeh and Gray 1979a, b, 1980; Lewis and Schrefler 1998; Pesavento
2000; Schrefler 2002). In the following, the full set of balance equations will be
presented. The governing equations of the model are given in terms of the chosen
state variables: the capillary pressure pc the gas pressure pg and the temperature
T. This choice is of particular importance: the chosen quantities must describe a
well-posed initial-boundary value problem, should guarantee a good numerical
performance of the solution algorithm, and should make their experimental iden-
tification simple.
A detailed discussion about the choice of state (i.e. primary) variables can be
found in Sect. 4.3.
The balance equations can be obtained by using the procedure of space averaging of
the microscopic balance equations, written for the individual constituents of the
medium (i.e. the governing equations at local level). The theoretical framework is
based on the works of Hassanizadeh and Gray (1979a, b, 1980), Gray and Schrefler
(2007) and Lewis and Schrefler (1998). A detailed description of the procedure can
be found also in (Pesavento 2000, Gawin et al. (2003).
36 F. Meftah et al.
For the sake of brevity, only the form of the conservation equations at macro-
scopic level is given (Gawin et al. 2003; Dal Pont et al. 2001). Hence, mass balance
equations read, for the solid matrix:
@ms
þ r ðms vs Þ ¼ m_ dehyd ð4:3Þ
@t
@ml
þ r ðml vl Þ ¼ m_ vap m_ dehyd ð4:4Þ
@t
@mv
þ r ðmv vv Þ ¼ m_ vap ð4:5Þ
@t
@ma
þ r ð m a va Þ ¼ 0 ð4:6Þ
@t
where mp is the mass per unit volume of porous medium of each constituent:
ms ¼ ð1 uÞqs ; ml ¼ ql Sl u
ð4:7Þ
mv ¼ qv ð1 Sl Þu; ma ¼ qa ð1 Sl Þu
D E
mdehyd meq
dehyd
þ
m_ dehyd ¼ ð4:8Þ
sdehyd
where the symbol hi þ stands for the positive value of the argument (otherwise the
value is set to zero), sdehyd is the characteristic time and meq
dehyd ðT Þ is the dehydration
mass at equilibrium, that is, the mass loss when the temperature rate T_ is slow
enough; i.e. when the time interval of variation of temperature is great enough with
regard to sdehyd . Alternatively, a time impendent evolution law may be adopted
(Gawin et al. 2003) by considering that mass loss attains systematically its equi-
librium value irrespective of the temperature rate:
@meq
dehyd _
m_ dehyd ¼ T þ ð4:9Þ
@T
@u m_ dehyd
¼ ð4:10Þ
@t qs
where the solid phase density qs is assumed here to be constant, otherwise the
variation of qs with time should appear in Eq. (4.10). It is worth noting that
Eq. (4.10) is strictly local since the effect of the deformation of the solid skeleton on
the porosity has been neglected in the adopted context of small-strain deformations.
In this case, the variation of porosity of the medium is only due to the temperature
induced dehydration: porosity increases with a rate that is proportional to dehy-
dration rate. A more general presentation including deformability of the solid
skeleton is discussed in Sect. 4.2.3.
Moreover, fluid velocities vp in conservation equations (4.4)–(4.6) are split into
relative components in order to describe mass transport within the porous network,
by both permeation and diffusion phenomena due to pressure and concentration
gradients, respectively. They read:
vl ¼ vs þ vls vls
ð4:11Þ
vp ¼ vs þ vgs þ vpg vgs þ vpg with p ¼ v; a
where vs (which can be neglected) is the velocity of the solid phase, vp−s is the
velocity of the liquid water (p = l) and gas mixture (p = g) with respect to the solid
38 F. Meftah et al.
phase and vp−g is the velocity of the vapour (p = v) and dry air (p = a) with respect
to the gas mixture (moist air).
Having assumed here that all phases of the material are locally in thermody-
namic equilibrium, i.e. their temperatures are the same, Tp = T (p = s, w, g), with
regard to the energy conservation equation, the space averaging procedure gives
(Gawin et al. 2003):
@T
qCp þ ml Cl vls þ mg Cg vgs rT þ r q ¼ rHdehyd m_ dehyd Þ DHvap m_ vap
@t
ð4:12Þ
with the heat capacity of the whole porous medium given by:
X
qCp ¼ mp Cp with p ¼ s; l; v; a ð4:13Þ
p
Mp
qp ¼ pp ð4:14Þ
RT
where pp is the pressure, Mp the molar mass and R the universal gas constant.
Furthermore, the pressure and density of the gas mixture can be related to the partial
pressures and densities of the constituents by:
4 Advanced Modelling 39
qg ¼ qv þ qa ; pg ¼ pv þ pa ð4:15Þ
which gives:
pv
M g ¼ M a þ ðM v M a Þ ð4:16Þ
pg
Liquid–Vapour equilibrium
By assuming that the evaporation process occurs without energy dissipation, that is,
liquid and vapour water have equal free enthalpies, one can derive the generalized
Clausius–Clapeyron equation, which is a relationship between liquid and vapour
pressure:
Mv
pv ¼ pvs exp pg pc pvs ð4:17Þ
ql RT
pc ¼ pg pl ð4:18Þ
has been introduced for the porous medium and pvs is the saturation vapour pres-
sure, which only depends on temperature.
Mass fluxes
According to the velocity decomposition (Eq. 4.11), the mass fluxes can be made
explicit:
ql krl
J ls ¼ ml vls ¼ K rpl ð4:19Þ
ll
qv krg Mv Ma pv
J vs ¼ mv vgs þ mv vvg ¼ K rpg Dqg r ð4:20Þ
lg Mg2 pg
qa krg Mv Ma pa
J as ¼ ma vgs þ ma vag ¼ K rpg Dqg r ð4:21Þ
lg Mg2 pg
In the above equations, where Darcy’s and Fick’s laws are introduced, K is the
intrinsic permeability (assumed isotropic), krp is the relative permeability, lp is the
dynamic viscosity, D is the diffusivity and Mp is the molar mass. For the gas
mixture, the Darcy’s part of the mass flux J ps with p = v, a is controlled
by the
barycentric velocity of the gas vps while the Fick’s part J pg ¼ mp vpg , which is
controlled by the concentration gradient, gives the diffusion of each component
within the gas mixture. Furthermore, the concentration gradient can be written as:
40 F. Meftah et al.
pv 1 qv pv qv pa
r ¼ rpp rpc ¼ r ð4:22Þ
pg pg ql pg ql pg
Thus, the vapour and dry air fluxes can be recast in the following generic form:
qp krg M v M a q v pv Mv Ma qv
J ps ¼ K rpg D rpg D rpc ð4:23Þ
lg RTMg ql pg RTMg ql
where the upper signs correspond to vapour case (p = v) while the lower signs give
the dry air flux (p = a). This form depends directly on the state variables pg and pc.
Conductive heat flux
In the energy balance equation (4.12), the heat conduction process in the porous
medium can be described by Fourier’s law which relates the temperature to the heat
flux as follows:
q ¼ kðSl ; T Þ rT ð4:24Þ
Sl ¼ Sl ð pc ; T Þ ð4:25Þ
(2003), in which the equation of vapour mass conservation is summed up with the
liquid water mass balance equation.
This has the effect of reducing the set of unknowns further as the evaporation
source term m_ vap is eliminated, thus obtaining the balance equation for total water
mass mw:
@mw
þ r J ws ¼ m_ dehyd ð4:26Þ
@t
with:
mw ¼ ml þ mv ð4:27Þ
In this way, at the end of the mathematical developments, we have a model with
two mass balance equations: one for the dry air and another one for the water
species (liquid plus water vapour).
The final form of the governing equations of heat and mass transfer will be
described in the next subsection, taking into account the set of primary variables
chosen at the beginning of the formulation.
Some additional discussions about the choice of the state variables and about
some key points in modelling concrete exposed to high temperature can be found in
Sect. 4.3.
where C_ dehyd is a normalized dehydration rate that may be related to m_ dehyd and to
the apparent solid density (Gawin et al. 2003).
The Mass balance equation of gaseous and liquid water considers the diffusive
(L.28.6) and advective flows of water vapour (L.28.7) and water (L.28.8), the mass
sources related to phase changes of vapour (evaporation-condensation, physical
adsorption–desorption) (the sum of those mass sources equals to zero) and dehy-
dration (R.28.1), and the changes of porosity caused by variation of: temperature
(L.28.3), dehydration process (L.28.10), variation of skeleton density due to
dehydration (L.28.9) and deformations of the skeleton (L.28.2), as well as the
variations of: water saturation degree (L.28.1) and the densities of vapour (L.28.4)
and liquid water (L.28.5). This gives the following equation (Gawin et al. 2003):
with bslg the thermal expansion coefficient of the multiphase porous medium,
defined by:
bslg ¼ bs ð1 uÞ ql Sl þ qv Sg þ ubl ql Sl ð4:31Þ
where q is the conductive thermal flux and the enthalpies of vaporisation and
dehydration are:
DHvap ¼ Hv Hl
ð4:33Þ
DHdehyd ¼ Hl Hls
where bsl is the thermal expansion coefficient of the solid-liquid phase, defined as:
bsl ¼ ql Sl ½ð1 uÞbs þ ubl ð4:35Þ
Equation (4.34) is obtained from the water mass conservation equation, con-
sidering the advective flow of water (R.32.4), the mass sources related to dehy-
dration (R.32.7) and the changes of porosity caused by: variation of temperature
(R.32.3), dehydration process (R.32.6), variation of skeleton density due to dehy-
dration (R.32.5) and deformations of the skeleton (R.32.2), as well as the variations
of water saturation degree (R.32.1).
The hygro-thermal problem consists of determining the temperature field T and the
pressure fields pg and pc, satisfying the conservation equations (4.29), (4.30) and
(4.32) within the domain X with the following boundary conditions at the domain
boundary R:
pg ¼ pg on Rp ð4:36Þ
pc ¼ pc on Rp ð4:37Þ
44 F. Meftah et al.
T ¼ T on RT ð4:38Þ
J vs n ¼ qv hg qv q1
v on Rp ð4:39Þ
J as n ¼ qa hg qa q1 p
on R ð4:40Þ
a
J ls n ¼ ql p
on R ð4:41Þ
q DHvap J ls n ¼ qT hT ðT T1 Þ rB T 4 T1
4 T
on R ð4:42Þ
where R is part of the boundary at which the pressures and the temperature are
known (Dirichlet type boundary conditions), while R is the complementary part
(with unit outward normal n) at which the mass fluxes of fluids and heat flux are
imposed. The terms qp (p = T, a, v, l) are the imposed fluxes, while q1 p with p = v,
a, are the densities of vapour and dry air, respectively. T∞ is the temperature in the
far field surrounding gas, while hg, and hT, are the convective mass and energy,
exchange coefficients, respectively. Finally is the emissivity and rB is the Stefan–
Boltzmann constant.
Note that the convective term on the right-hand-side of Eq. (4.42) corresponds to
Newton’s law of cooling and describes the conditions occurring in most practical
situations at the interface between a porous medium and the surrounding fluid
(moist air in this case). The second left-hand-side term gives the energy exchange at
the external surface due to the vaporisation process.
Particular attention has to be given to the mass convective boundary conditions
(4.39) and (4.40) if the staggered approach discussed below is used for time inte-
gration. The mass densities qp (p = v, a) have to be expressed in terms of the
retained variables of the problem (pc, pg, T). However, the relationships relating the
densities to the state variables are strongly nonlinear, which does not permit a
straightforward algebraic form of these boundary conditions suitable for the dis-
cretised problem to be obtained. Indeed, the variables at the boundary R p should be
factorised such that additional terms, associated to these boundary conditions, arise
in the operators of the discretized formulation in order to make them well condi-
tioned. In order to overcome this difficulty, it is proposed here to linearize, for a
given time stage t, the relationships relating the densities to the state variables in the
neighbourhood of a reference time stage tref:
@qp @q @qp
qp qref
p þ pg pref þ p pc pref þ T T ref ð4:43Þ
@pg g
@pc c
@T
where (pc, pg, T)ref are the values of the variables at tref. The choice of the reference
time will depend on the adopted iterative algorithm presented in the following.
It is worth noting that this peculiar problem does not exist in the case where the
monolithic approach algorithm is adopted for the solution of the numerical model
(Gawin et al. 2003, 2006). In such a case all the variables related to the calculation
4 Advanced Modelling 45
of the boundary conditions are evaluated at the same time instant and at the same
iteration.
state variable selected in the model is capillary pressure (Gawin et al. 2002b), which
was shown to be a thermodynamic potential of the physically adsorbed water and,
with an appropriate interpretation, can be also used to describe water at pressures
higher than atmospheric. The capillary pressure has been shown to assure good
numerical performance of the computer code (Gawin et al. 2002b, 2003, 2006) and
is very convenient for analysis of the stress state in concrete, because there is a clear
relationship between pressures and stresses (Gray et al. 2009).
Therefore, as already shown, the chosen primary state variables of the presented
model are the volume averaged values of: gas pressure, pg, capillary pressure, pc,
temperature, T, and displacement vector of the solid matrix, u.
For temperatures lower than the critical point of water, T < Tcr, and for capillary
saturation range, Sl > Sssp(T) (Sssp is not only the upper limit of the hygroscopic
moisture range as mentioned earlier, but also the lower limit of the capillary one),
the capillary pressure is defined according to Eq. (4.18).
For all other situations, and in particular for T
Tcr where condition Sl < Sssp
is always fulfilled (there is no capillary water in the pores), the capillary pressure is
substituted—only formally—by the water potential Wc, defined as:
RT pv
Wc ¼ ln vs ð4:44Þ
Mw f
where Mw is the molar mass of water, R the universal gas constant and f vs the
fugacity of water vapour in thermodynamic equilibrium with a saturated film of
physically adsorbed water (Gawin et al. 2002b). For physically adsorbed water at
lower temperatures (S < Sssp and T < Tcr) the fugacity f vs should be substituted in
the definition of the potential Wc, by the saturated vapour pressure pv. Having in
mind Kelvin’s equation, valid for the equilibrium state of capillary water with water
vapour above the curved interface (meniscus):
pv pc M w
ln ¼ ð4:45Þ
pvs ql RT
We can note that in the situations where (4.44) is valid, the capillary pressure
may be treated formally as the water potential multiplied by the density of the liquid
water, ql , according to the relation:
pc ¼ Wc ql ð4:46Þ
Above this temperature, the liquid water and gaseous water phase (water vapour)
cannot be distinguished and only the gas phase exists. As a result, there are no
phase changes of the pore water (condensation–evaporation) and capillary pressure
has no physical meaning.
Hence, after reaching, in a region of moist porous material, temperatures higher
than the critical point of water, we deal with a kind of Stefan’s problem, where two
regions, with the temperatures below and above the critical point of water, are
separated by the moving interface boundary.
In Gawin et al. (2002b) a method, which formally allows direct tracking of the
boundary position in the space to be avoided, is presented. It consists of giving
formally different physical meaning to the capillary pressure (as discussed above)
and using it still for description of the hygrometric state of concrete in the zone
where the temperature exceeds the critical point of water. Then, a special
‘switching’ procedure is applied for a finite element where this temperature is
encountered, from the below- to the above- critical temperature description of the
medium, still using the governing equations of the same form, but with different
physical meaning.
In the following, some numerical procedures that are adopted for the discretisation
and the solution of the thermo-hydraulic model are presented.
The set of non-linear partial differential equations that controls the processes of
mass transport (see 4.2.3) is usually discretised in the space domain by means of
two methods, Finite Elements or Finite Volumes. The latter has the advantage of
being flux conservative but the former is usually adopted in the literature and will
be retained for this presentation.
According to the FEM procedure, the governing equations are re-written using
the weighted residual method, and the standard Galerkin procedure is usually
adopted. Time discretisation is achieved by means of the standard Finite Difference
Method. A non-symmetric, non-linear system is finally obtained and linearization,
usually by means of the Newton-Raphson method, is required (Zienkiewicz and
Taylor 2000).
The final set of discretised, linearized governing equations is solved as a
monolithic system or, alternatively, by means of a staggered coupled procedure (see
Gawin et al. 2003 and Meftah and Dal Pont 2010 or Meftah et al. 2012).
Starting from the previous initial-boundary value problem, the variational formu-
lation can be obtained as below, respectively, for total water, dry air and energy:
48 F. Meftah et al.
Z Z Z
@mw
p dX rp J ws dX ¼ p ðql þ
q v h g q v q1
v ÞdR
@t
X X Rp
Z
p m_ dehydr dX ð4:47Þ
X
Z Z Z
@ma
p dX rp J as dX ¼ p ðqa hg qa q1
a ÞdR ð4:48Þ
@t
X X Rp
Z Z
@T @ml
T qCp þ ml Cl vls þ mg Cg vgs rT dX T Hvap dX
@t @t
X X
Z Z
þ T rHvap ml vls dX þ rT Hvap ml vls dX
ZX Z X
ð4:49Þ
rT qdX T Hvap þ Hdehydr m_ dehydr dX
X X
Z
¼ T ðqT hT ðT T1 Þ er T 4 T1
4
ÞdR
Rp
where p* and T* are weighting functions that respectively vanish at the boundaries
Rp and RT and in which essential boundary conditions are used.
The standard Galerkin discretisation, replacing the weighting functions with the
shape functions, is adopted. For the state variables we can write:
pg ¼ N p pg ; pc ¼ N p pc ; T ¼ Nt T ð4:50Þ
After some manipulations, the weak formulation previously stated can be written
in the following form:
@Y
Cij ðY Þ þ K ij ðY ÞY ¼ f i ðY Þ ð4:51Þ
@t
where:
4 Advanced Modelling 49
0 1 0 1 0 1
Cgg Cgc Cgt K gg K gc K gt fg
Cij ¼ @ 0 Ccc Cct A; K ij ¼ @ K cg K cc K ct A; f i ¼ @ f c A ð4:52Þ
0 Ctc Ctt K tg K tc K tt ft
For the sake of brevity, the matrix C and K are not given (see e.g. Gawin et al.
2003; Meftah and Dal Pont 2012 for further details).
~ cc Dpn þ 1 þ K cc pn ¼ f n þ # K
K ~ cg Dpn þ 1 þ K cg pn K
~ cT DT n þ 1 þ K cT T n ð4:53Þ
c c c g g
~ gg Dpn þ 1 þ K gg pn ¼ f n þ # K
K ~ gc Dpn þ 1 þ K gc pn K
~ gT DT n þ 1 þ K gT T n ð4:54Þ
g g g c c
~ TT DT n þ 1 þ K TT T n ¼ f n þ # K
K ~ Tc Dpn þ 1 þ K Tc pn K
~ Tg Dpn þ 1 þ K Tg pn ð4:55Þ
T c c g g
with:
and:
~ ¼ C þ #K
K ð4:57Þ
Dt
where Dt is the time step, the index n makes reference to the time station tn and tn þ #
to the time station tn þ #Dtn .
A monolithic approach is often adopted for the solution of the linearized system. In
this approach all the equations are solved at the same time and all the state variables
are updated simultaneously during the iterative procedure. This approach guaran-
tees the optimal convergence of the system especially when dealing with strong
non-linearities (Pesavento 2000; Gawin et al. 2003).
Alternatively, the system can be solved using a staggered procedure that permits
the equations to be solved sequentially (Dal Pont et al. 2007, Meftah and Dal Pont
2010). A specific iterative algorithm to account for the interactions between all of
50 F. Meftah et al.
the transfer processes to conserve their full coupling is required. Staggered pro-
cedures show a superior flexibility compared to monolithic ones if successive fields
have to be introduced in the problem. Moreover, an appropriate partitioning reduces
the size of the discretized problem to be solved at each time step. For the sake of
clarity, some details are provided on this latter procedure.
The set of nonlinear equations (4.53)–(4.55) is solved using a staggered iterative
scheme with two nested levels of iterations. The first level (local iteration k) con-
cerns the convergence process for each of the three equations when solved for one
variable while the other variables are kept constant. The second level (global
iteration j) concerns the convergence of the interaction among the three equations
when considering simultaneously the three updated variables. Moreover, a
quasi-Newton iterative algorithm is adopted; each equation is solved for the total
increment Dxn þ 1 (with Dxc ¼ Dpc , Dxg ¼ Dpg ; Dxt ¼ DT) by keeping fixed the
linearization point at the previous converged step (at time stage tn ) and by cumu-
lating the residuals during the iterations. Accordingly, the set of Eqs. (4.53)–(4.55)
can be put in the following generic form:
Hence, after some manipulations, the final form of the set of Eqs. (4.53)–(4.55)
to be solved at each local iteration k + 1 is given by:
with:
n
Y na þ #;j þ 1;k ¼ f na þ # þ K ~ K ~ n þ #;j þ 1;k Dxn þ 1;j þ 1;k K n þ #;j þ 1;k xn
aa aa a aa a
K ~ n þ #;j þ 1;k
Dx n þ 1;j
K n þ #;j þ 1;k n
x ð4:62Þ
ab b ab b
~ n þ #;j þ 1;k Dxn þ 1;j K n þ #;j þ 1;k xn
K ac c ac c
In this section, we describe the Bažant’s model as the most representative one in the
class of the phenomenological models.
Bažant’s model has been, probably, the most popular model for analysing the
behaviour of concrete at high temperature between the 1980s and 1990s. Bažant
and Thonguthai (1978, 1979), on the basis of Luikov’s theory, proposed field
equations for the coupled heat and moisture transfer in concrete at high tempera-
tures. Making use of experimental data, they simplified the expression of the mass
flux (a function only of the gradient of the pressure) and presented some
semi-empirical relationships for the desorption isotherms, for the dependence of
permeability on temperature, for the change of porosity because of dehydration, and
so on. Here, only a brief description will be given and the reader is referred to
Bažant and Thonguthai (1978, 1979), and Bažant and Kaplan (1996) for further
details.
This model is based on a typical phenomenological approach and considers a
single moisture flux, i.e. a single fluid phase is present, which includes both liquid
water and moist air. Thus, the flux of mass in concrete J consists of a part due to the
gradient of moisture concentration, i.e. is a diffusive flux (Fick’s law), and a part
due to the gradient of temperature, i.e. convective flux (Soret flux). Similarly, the
heat flux q, consists of a flux due to the gradient of temperature, i.e. conductive flux
(Fourier’s law) as well as the flux due to the gradient of moisture concentration, i.e.
convective flux (Dufour flux). We can write (Bažant and Kaplan 1996):
where mw is the water-content (Eq. 4.27), i.e., the total mass of free water (not
chemically bound) per m3 of concrete at a certain temperature.
Note that generally the diffusion coefficients Dij (i, j = w, T) in Eq. (4.63) are not
symmetric and that the relationship is written assuming that the mass and heat
fluxes are not affected by solid (concrete) deformability.
From Eq. (4.63)1, Bažant gives the following relationship for moisture flux by
neglecting the convective flux due to thermal gradient (Soret effect):
J ¼ Dh rh ð4:64Þ
Using Eq. (4.64) and Kelvin’s equation (4.45), the moisture flux in Eq. (4.64)
can be alternatively expressed in terms of gradient of capillary pressure:
Likewise, it is possible to write down the relationship for heat flux, neglecting
the convective part of the flow (Dufour):
q ¼ brT ð4:67Þ
@w @wd
¼ r J ð4:68Þ
@t @t
@T @w
qC Ca Cw JrT ¼ r q ð4:69Þ
@t @t
in which wd is the mass of chemically bound (hydrate) water that has been released
into the pores as a result of dehydration of cement paste (i.e. part of the solid phase)
caused by heating, q and C are the density and specific heat of concrete including
chemically bound water, but excluding free water, and finally, Ca and Cw are the
heat of sorption of free water and the isobaric heat capacity of bulk (liquid) water,
respectively. The term Cw JrT represents the rate of heat supply due to heat
convection by moving water. Usually this term is neglected in Bažant’s model, but
in rapid heating, such as during a fire or a nuclear accident, it can be of importance.
The most important shortcoming of this model is that water vapour and liquid
water content are treated as one variable (moisture). The formulation of the des-
orption isotherms proposed by Bažant is analytically derived from the thermody-
namics of water at high temperature, but it is empirically modified to match the
experimental data. In such a way the equilibrium between the liquid phase and its
vapour phase is no longer assured and it is possible to obtain values of relative
humidity greater than 100%, which is physically meaningless. Moreover, part of the
description of the material microstructure, i.e. the porosity, is automatically
encompassed in the isotherms definition.
Furthermore, in the original formulation of the model the mechanical behaviour
is not considered, so it is limited to the heat and mass transfer simulation.
54 F. Meftah et al.
In this approach, only the heat transfer process is considered, with the temperature
field as the unknown. It is determined by solving the linearized energy balance
equation, which can be put in the following form when neglecting mechanical
interaction (which has negligible contribution when high temperatures are consid-
ered) and local source terms:
@T
qCp þr q ¼ 0 ð4:70Þ
@t
where q is the material density, Cp is the specific heat and q is the heat flux.
It is worth noting that, in a simplified thermo-mechanical approach, the rate of
energy release or energy consumption due to phase changes (as evaporation,
condensation and dehydration) at any material point depends only on the temper-
ature at this point. In other words, the supply (or loss) of water to this point by the
surrounding medium cannot be considered during the heating process.
Thus, phase changes are taken into account by adopting suitable phenomeno-
logical evolution laws of the specific heat capacity as functions of temperature. For
instance, an increase in specific heat capacity can be considered in the range of
temperature where water evaporation is assumed to intensively occur.
This simplification is a limitation in the thermo-mechanical approach since the
heat capacity of the material point is constrained by the adopted
temperature-dependent evolution law disregarding the transient hygral state of the
point.
Another limitation in the simplified thermo-mechanical approach is that
advective heat transfer (which may affect significantly the temperature field due to
mass fluxes of fluids) is ignored.
4 Advanced Modelling 55
The heat flux is classically given by Fourier’s constitutive equation for heat
transfer:
where kðT Þ is the thermal conductivity. Assuming that the thermal conductivity
depends on the temperature only restricts the model in a way that there is no
possibility of accounting for the effect of the transient hygral state on the heat
transfer mechanism.
The heat transfer problem consists of determining the temperature field T ðx; tÞ
satisfying Eq. (4.70) (supplemented by constitutive Eq. (4.71)) within a given
domain X. The boundary conditions can be of various types:
on a part @XT of the domain boundary:
– prescribed temperatures T
T ¼ T 8 x 2 @XT ð4:72Þ
n q ¼ q 8 x 2 @XT ð4:73Þ
where n is the outward normal to the boundary and q is the heat flux across the
boundary, accounting for conduction, convection and radiation where appropriate:
q ¼ hT ðT T1 Þ þ rB T 4 T1
4
ð4:74Þ
For completeness, the momentum balance equations are recalled here while limiting
the analysis to quasi-static mechanical loadings. Introducing the Cauchy stress
56 F. Meftah et al.
second-order tensor rðx; tÞ at time stage t and spatial position x of any material
point of the studied domain X, the momentum balance equations can be written, in
the quasi-static case, i.e. neglecting the inertia effects, as:
x 2 @Xu : u ¼ u ð4:76Þ
x 2 @Xr : r n ¼ f ð4:77Þ
where n denotes the outward normal vector to the boundary @Xr and stands for the
scalar product.
e ¼ ee þ ef þ eth þ ec ð4:78Þ
Note that the identification of the component emf requires analysis of the
behaviour of the material at the meso-scale, i.e., the scale at which concrete is seen
as a set of granular inclusions embedded in a cementitious matrix.
4 Advanced Modelling 57
With regard to the shrinkage strain es two main driving mechanisms can be
discriminated:
– Drying shrinkage which involves the mechanical action of the capillary pressure
which increases due the loss of water by desiccation. This component appears to
be less significant than the second mechanism since the liquid saturation of
heated concrete quickly drops to zero in addition to the counteracting effect of
dilatation of initially available capillary water.
– Dehydration shrinkage which is due to a reduction of the material volume
induced by the departure of the initially chemically bounded water. The pro-
gressive reorganization, at the microstructure scale, of the formed anhydrous
phases lead to this shrinkage component, the magnitude of which appears to be
more important than the previous one.
Furthermore, the transient creep strain ec in Eq. (4.78) encompasses both the
basic creep strain ebc and the transient thermal creep strain etc :
From experimental tests, LITS is obtained by removing the free thermal strain eth
from the strain measured during a test at which the specimen is firstly loaded and
then heated while the load is maintained.
It is worth noting that the identification of strain components of heated concrete
according to Eqs. (4.78)–(4.81) is not straightforward since it requires a proper
quantification of the contribution of cracking processes. Cracking strain needs to be
evaluated by ad hoc cracking models since it cannot be measured experimentally.
58 F. Meftah et al.
_
e_ th ¼ ath ðT ÞTd ð4:82Þ
where d is the second-order unit tensor, T_ is the temperature rate and T_ is the
tangent temperature-dependent thermal expansion coefficient.
Constitutive equation (4.82) is the usual adopted approach for modelling thermal
expansion. However, the thermal expansion coefficient is identified experimentally
from the standard free thermal expansion test during which mass and heat transfer
occur within the specimen with some characteristic times that depend on the
specimen size and the transfer properties (conductivity, permeability, dehydration
kinetics…) of the material.
Since the identified thermal expansion coefficient depends on the test conditions
it cannot be considered as an objective constitutive (material) parameter. To
overcome the inherent drawback in the standard experimental identification pro-
cedure, the free thermal strain eth should be split according to Eq. (4.79) with an
appropriate constitutive law for each strain. Then, experiment and computation
should be used in conjunction to identify all involved physical parameters.
Effective thermal strain
Similarly to the approach of apparent thermal strain, an effective thermal expansion
coefficient at ðT Þ may be considered for expressing the effective thermal strain rate e_ t :
_
e_ t ¼ at ðT ÞTd ð4:83Þ
4 Advanced Modelling 59
e_ s ¼ as hm_ l i d ð4:84Þ
where as is the shrinkage coefficient, m_ l is the mass loss of water that comes from
mass and heat transfer modelling (cf. Sect. 4.2) and the symbol hi stands for the
negative part, that is, only a water loss (negative m_ l ) will lead to shrinkage.
An alternative constitutive equation for the drying shrinkage component comes
from making the two driving mechanisms previously described (cf. Sect. 4.7.2)
explicit in the constitutive law. Distinction is made between shrinkage induced by
drying and shrinkage induced by dehydration. The former is assumed to be the
deformation of the solid skeleton that is induced mechanically by capillary effects:
capillary forces give rise to an increase of suction within the pore space leading to a
contraction of the porous medium. The second mechanism is rather related to a
volume decrease of the solid phase (skeleton) due to loss of chemically bound water
by dehydration. This is conventionally assumed to start when the temperature
exceeds a threshold of about 105–115 °C.
Hence, the shrinkage strain tensor in this alternative approach
reads (Benboudjema et al. 2007; Sabeur et al. 2008; Sabeur and Meftah 2008):
While an explicit constitutive law is needed for strain rate tensor of dehydration
shrinkage e_ hyds , the drying (capillarity) induced shrinkage may be implicitly
embedded in the total strain by the use of a poro-mechanical constitutive behaviour
(cf. section on effective stress). Indeed, the capillary pressure (Eq. 4.18) as a part of
pore pressure will generate an effective stress r0 (Eq. 4.89) that leads to a defor-
mation of the porous medium. For instance, a time dependent constitutive law
(creep) may be adopted for the porous medium leading to a drying shrinkage strain
that is the time dependent response of the medium due to capillary pressure (suc-
tion) acting on the creeping skeleton (Benboudjema et al. 2007; Sabeur et al. 2008;
Sabeur and Meftah 2008).
60 F. Meftah et al.
Transient Creep
During first-time heating of a concrete specimen under sustained load, a significant
increase in strains associated with the increase in temperature occurs. This phe-
nomenon cannot be solely explained by the thermo-mechanical degradation of the
elastic properties nor by the occurrence of micro-cracking. It has rather been linked
to an additional strain component, which takes place in a very short period of time.
This strain component, called transient thermo-mechanical interaction strain or
transient creep strain, is considered to be a function of the temperature as well as of
the stress. Even though there is no consensus on the physical mechanisms driving
this phenomenon, a federating twofold assumption is made: whereby between 100
and 200 °C drying of the cement matrix is the most important factor (drying creep),
while for higher temperatures a change of the chemical structure of the hydrates is
primarily assumed to be responsible.
Based on experimental investigation, Anderberg and Thelandersson (1976)
proposed a uni-axial constitutive law for transient creep strain rate that has been
extended to multi-axial behaviour by Thelandersson (1987):
_
e_ tc ¼ TQ:r ð4:86Þ
where Q is the forth order tensor of transient creep compliance given by:
ath btc 1
Qijkl ¼ ð1 þ ctc Þ dik djl þ dil djk ctc dij dkl ð4:87Þ
fc 2
in which fc is the compressive strength at ambient temperature and ðbtc ; ctc Þ are
model parameters identified from transient creep tests: btc comes from a uni-axial
test while ctc refers to the “Poisson’s” ratio of transient creep.
The constitutive relationship (4.86) has been successfully incorporated into a
thermo-elasto-plastic modelling approach for heated concrete by Khennane and
Baker (1992) and by Heinfling (1998). The approach was also extended to a
damage-based framework by Nechenech et al. (2002):
_ r
e_ tc ¼ TQ:~ ð4:88Þ
r ¼ r0 þ aps d ð4:89Þ
where r0 is the effective (Bishop’s) stress tensor, a is the Biot coefficient, ps is the
pore pressure transferred to the solid skeleton and d is the identity tensor (Gray et al.
2009).
Pore pressure definition
Different expressions for the pore pressure that transfers to the solid concrete
material phase as stress can be found in the literature. A first definition is (Tenchev
et al. 2001):
ps ¼ pg ð4:90Þ
where pl is the liquid water pressure, patm is the atmospheric pressure and Sbl is the
saturation threshold beyond which the liquid phase present in the pore cannot exert
a significant pressure on the solid skeleton. Based on the same key idea (Gawin
et al. 2003), an alternative expression to the previous reads:
(
pg patm for Sl Sbl
ps ¼ Sl Sbl ð4:92Þ
1Sbl
pl þ 1Sl
1Sbl
pg patm for Sl [ Sbl
Finally, the pore pressure can be related to gas pressure and capillary pressure as
follows:
ps ¼ pg v l pc ð4:93Þ
where vl is the solid surface fraction in contact with the wetting film (Al Najim et al.
2003; Benboujema et al. 2007; Sabeur and Meftah 2008; Gray et al. 2009).
Stress-strain relationship
Typically when modelling concrete exposed to elevated temperatures, it is con-
sidered most appropriate and convenient to assume elastic behaviour of the solid
component and account for non-linear behaviour via a damage formulation (e.g.
Gawin et al. 1999, 2004; Tenchev and Purnell 2005; Davie et al. 2010; de Sa and
Benboudjema 2011). That said, some authors have recently started to combine
damage and plasticity models in order to more comprehensively capture observed
behaviour, especially in multi-axial stress conditions (e.g. Nechnech et al. 2002; Al
Najim et al. 2003; Gernay et al. 2013).
62 F. Meftah et al.
The use of a damage formulation has the added advantage that, for thermo-
hydro-chemo-mechanical models, the change in some other material properties
(particularly for example permeability (see Sect. 5.1.5) can be directly associated
with the damage formulation (e.g. Gawin et al. 2002b; Al Najim et al. 2003; Davie
et al. 2010).
In the general case, the damage formulation is based on the concept of the
~ in the sense of damage mechanics (Kachanov 1986), which may
effective stress r
be related to the effective stress r0 in the sense of poro-mechanics as:
r0 ¼ ð1 xÞ~
r ¼ ð1 xÞC0 :ee ð4:94Þ
Whereby the damage parameter, x.,ontrols the reduction of the initial elasticity
tensor C0 due to the effective loss of load carrying cross-section as micro-cracks
develop in the solid skeleton of the concrete.
Classic damage models consider the effects of mechanical stress in the devel-
opment of damage, but when considering concrete at elevated temperatures it is
necessary to consider also the fact that damage can be caused by thermal degra-
dation of the material (i.e. the breakdown of the solid and the development of
micro-cracks due to an increase in temperature, even without load).
This thermal damage xT is typically accounted for through the degradation of
the elastic modulus (as described in Sect. 5.3.1.3) (e.g. Baker and Stabler 1999;
Nechnech et al. 2002; Al Najim et al. 2003; Gawin et al. 2003; Davie et al. 2010)
and is treated multiplicatively with the mechanical damage xM such that:
However, some authors account for thermal effects through the temperature
dependent evolution of the damage and/or yield surfaces (e.g. Gernay, Millard et al.
2013).
The definition of the mechanical damage can take many forms and the most
appropriate form is dependent on the required application. Generally, thermal
dependence of the parameters of the model is required in order to capture the
change in mechanical behaviour at elevated temperatures. Examples of various
approaches to thermally dependent mechanical damage formulations can be found
in Gawin et al. (1999, 2003), Nechnech et al. (2002), Al Najim et al. (2003),
Luccioni et al. (2003), Tenchev and Purnell (2005), Davie et al. (2010), de Sa and
Benboudjema (2011), Gernay et al. (2013).
References
Abdel-Rahman, A.K., Ahmed, G.N.: Computational heat and mass transport in concrete walls
exposed to fire, Numerical Heat Transfer, Part A: Applications. 29(4), 373–395 (1996)
4 Advanced Modelling 63
Al Najim, A., Meftah, F., et al.: A Non-Saturated Porous Medium Approach for the Modelling of
Concrete Behaviour Submitted to High Temperatures. Computational Modelling of Concrete
Structures, Proceedings of the EURO-C Conference, Austria, Balkema (2003)
Anderberg Y. and Thelandersson S. (1973), Stress and deformation characteristics of concrete at
high temperature, Lund Institute of technology (Sweden): Division of Structural Mechanics
and Concrete Construction, 84p. Internal rep. no Alba15/04-01
Anderberg, Y., Thelandersson, S.: Stress and deformation characteristics of concrete at high
temperature:2. Bulletin 54, Lund Institute of technology, Lund. (1976)
Baker, G., Stabler, J.: Computational modelling of thermally induced fracture in concrete. Proc.
Euro-C, 530–545 (1999)
Baroghel-Bouny, V., et al.: Characterization and Identification of Equilibrium and Transfer
Moisture Properties for Ordinary and High-Performance Cementitious Materials. Cement and
Concrete Research. 29, 1225–1238 (1999)
Bažant, Z.P., Kaplan, M.F.: Concrete at High Temperatures: Material Properties and Mathematical
Models. Longman, Harlow (1996)
Bažant, Z.P., Thonguthai, W.: Pore pressure and drying of concrete at high temperature. J. Engng.
Mech. Div. ASCE. 104, 1059–1079 (1978)
Bažant, Z.P., Thonguthai, W.: Pore pressure in heated concrete walls: theoretical prediction. Mag.
of Concr. Res. 31(107), 67–76 (1979)
Benboudjema F. (2002), Modélisation des déformations différées du béton sous sollicitations
biaxials. Application aux enceintes de confinement de bâtiments réacteurs des centrals
nucléaires, Thèse de doctorat, UMLV, France, 258 p (In French)
Benboudjema, F., Meftah, F., Torrenti, J.M.: A viscoelastic approach for the assessment of the
drying shrinkage behaviour of concrete, Materials and Structures, 40(2), 163–174 (2007).
Chung, J.H., Consolazio, G.R., McVay, M.C.: Finite element stress analysis of a reinforced
high-strength concrete column in severe fires. Computers and Structures. 84(21), 1338–1352
(2006)
Consolazio, G.R., McVay, M.C., Rish, J.W.: Measurement and prediction of pore pressures in
saturated cement mortar subjected to radiant heating. ACI Mater. J. 95(5), 526–536 (1998)
Dal Pont, S., Ehrlacher A.: Numerical and experimental analysis of chemical dehydration, heat and
mass transfer in a concrete hollow cylinder submitted to high temperatures. International
Journal of Heat and Mass Transfer. 1(47), 135–147 (2004)
Dal Pont, S., Durand, S., Schrefler, B.A. A multiphase thermo-hydro-mechanical model for
concrete at high temperatures - finite element implementation and validation under LOCA load.
Nuclear Eng. and Design, 237/20: 2137–2150 (2007)
Davie, C.T., Pearce, C.J., Bicanic, N.: Coupled Heat and Moisture Transport in Concrete at
Elevated Temperatures-Effects of Capillary Pressure and Adsorbed Water. Numerical Heat
Transfer, Part A. 49(8), 733–763 (2006)
Davie, C.T., Pearce, C.J., et al.: A Fully Generalised, Coupled, Multi-Phase,
Hygro-Thermo-Mechanical Model for Concrete, Materials and Structures 43 (Sup. 1) (2010)
de Sa, C., Benboudjema, F.: Modeling of concrete nonlinear mechanical behavior at high
temperatures with different damage-based approaches, Materials and Structures 44(8): 1411 –
1429 (2011)
Dwaikat, M.B., Kodur, V.K.R.: Hydrothermal model for predicting fire-induced spalling in
concrete structural systems. Fire Safety Journal. 44, 425–434 (2009)
England, G.L., Khoylou, N.: Moisture flow in concrete under steady state non-uniform
temperature states: experimental observations and theoretical modeling. Nucl. Eng. Des.
156, 83–107 (1995)
Feldman, R.F., Sereda, P.J. :The model for hydrated Portland cement as de-duced from
sorption-length change and mechanical properties. Materials and Con-struction, 1, 509–520
(1968)
Feraille A.: Le rôle de l’eau dans le comportement à haute température des bétons. Dissertation,
ENPC, Marne la Vallée (2000)
64 F. Meftah et al.
Gawin, D., Majorana, C.E., Schrefler, B.A.: Numerical analysis of hygro-thermic behaviour and
damage of concrete at high temperature. Mechanics of Cohesive-Frictional Materials. 4, 37–74
(1999)
Gawin, D., Pesavento, F., Schrefler, B.A.: Simulation of damage–permeability coupling in
hygro-thermo-mechanical analysis of concrete at high temperature. Communications in
Numerical Methods in Engineering. 18(2), 113–119 (2002a)
Gawin, D., Pesavento, F., Schrefler, B.A.: Modelling of hygro-thermal behaviour and damage of
concrete at temperature above critical point of water. International Journal of Numerical and
Analytical Methods in Geomechanics. 26(6), 537–562 (2002b)
Gawin, D., Pesavento, F., Schrefler, B.A.: Modelling of thermo-chemical and mechanical damage
of concrete at high temperature. Computer Methods in Applied Mechanics and Engineering.
192, 1731–1771 (2003)
Gawin, D., Pesavento, F., Schrefler, B.A.: Modelling of deformations of high strength concrete at
elevated temperatures. Materials and Structures. 37(268), 218–236 (2004)
Gawin, D., Pesavento, F., Schrefler, B.A.: Towards prediction of the thermal spalling risk through
a multi-phase porous media model of concrete. Computer Methods in Applied Mechanics and
Engineering. 195(41–43), 5707–5729 (2006)
Gens, A. and Olivella, S.: THM phenomena in saturated and unsaturated porous media, RFGC.
Environmental Geomechanics. 5(6), 693–717 (2001)
Gernay, T., Millard, A., et al.: A multiaxial constitutive model for concrete in the fire situation:
Theoretical formulation, International Journal of Solids and Structures 50(22–23): 3659-3673
(2013)
Gray, W.G., Schrefler, B.A.: Analysis of the solid phase stress tensor in multiphase porous media.
Int. J. Numer. Anal. Meth. Geomech., 31(4), 541–581 (2007)
Gray, W.G., Schrefler, B.A., Pesavento, F.: The solid phase stress tensor in porous media
mechanics and the Hill-Mandel condition. J. Mech Phys Solids. 57, 539–554 (2009)
Harmathy, T.Z.: Effect of moisture on the fire endurance of building elements. In ASTM special
technical publication 385. Philadelphia (1965)
Harmathy, T.Z.,. Thermal properties of concrete at elevated temperatures. J. Mech. JMLSA. 5(1),
47–74 (1970)
Hassanizadeh, S.M., Gray W.G.: General Conservation Equations for Multi-Phase Systems: 1.
Averaging Procedure. Adv. Water Resources. 2, 131–144 (1979a)
Hassanizadeh, S.M., Gray, W.G.: General Conservation Equations for Multi-Phase Systems: 2.
Mass, Momenta, Energy and Entropy Equations. Adv. Water Resources. 2, 191–203 (1979b)
Hassanizadeh, S.M., Gray, W.G.: General Conservation Equations for Multi-Phase Systems: 3.
Constitutive Theory for Porous Media Flow. Adv. Water Resources, 3, 25–40 (1980)
Heinfling G. (1998), Contribution à la modélisation numérique du comportement du béton et des
structures en béton armé sous sollicitations thermomécaniques à hautes températures, Thèse de
doctorat, INSA de Lyon, 227 p (In French)
Ichikawa, Y, England, G.L.: Prediction of moisture migration and pore pressure build-up in
concrete at high temperatures. Nucl. Eng. Des. 228(1–3), 245-259 (2004)
Kachanov, L. M.: Introduction to continum damage mechanics, Martinus Nijhoff Publishers,
Dordrecht, The Netherlands (1986)
Khennane, A., Baker, G.: Thermo-plasticity models for concrete under varying temperature and
biaxial stress. Proc. Royal Soc. Lond. A. 439, 59–80 (1992)
Khoury G.A. (2003), Creep & Shrinkage, Course of Heat on Concrete, International Center for
Mechanical Sciences (CISM), 9-13 June 2003, Udine, Italy
Khoury G. A. ., Grainger B. N. and Sullivan, P. I. E. (1985), Grainger B. N. and Sullivan P. J. E.,
Transient thermal strain of concrete: literature review, conditions within specimen and
behaviour of individual constituents. Magazine of Concrete Research, 37 (132), 131–144
Khoury G.A, Majorana C.E, Pesavento F. and Schrefler B.A. (2002), Modelling of heated
concrete, Magazine of concrete Research, 2002, 54, No.2, April, 77–101
Lewis, R.W., Schrefler, B.A.: The Finite Element Method in the Static and Dynamic Deformation
and Consolidation of Porous Media. second ed. John Wiley & Sons, Chichester (1998)
4 Advanced Modelling 65
Meftah, F., Dal Pont, S.: Staggered finite volume modeling of transport phenomena in porous
materials with convective boundary conditions. Transport Porous Media. 82(2), 275–298
(2010)
Meftah, F., Dal Pont, S., Schrefler, B.A.: A three-dimensional staggered finite element approach
for random parametric modeling of thermo-hygral coupled phenomena in porous media. Int.
J. Num and Anal. Meth. Geomech. 36(5), 574–596 (2012)
Mehta PK, Monteiro PJM (2014) Concrete: Structure, Properties, and Materials (4th Ed).
Prentice-Hall, Englewood Cliffs, NJ, USA
Nechnech, W., Meftah, F., et al.: An elasto-plastic damage model for plain concrete subjected to
high temperatures, Engineering Structures 24(5): 597–611 (2002)
Pasquero D. (2004), Contribution à l'étude de la déshydratation dans les pâtes de ciment soumises
à haute température, Thèse de doctorat, ENPC, France, 191 p. (In French)
Pesavento, F.: Non-linear modeling of concrete as multiphase porous material in high temperature
conditions. Dissertation, University of Padova (2000)
Sabeur H, Meftah F. Dehydration creep of concrete at high temperatures. Materials and Structures
41(1), 17–30 (2008)
Sabeur H, Meftah F, Colina H, Plateret G. Correlation between transient creep of concrete and its
dehydration. Magazine of Concrete Research, 60(3):157–163 (2008)
Schneider U. (1988), Concrete at high temperatures: A general review, Fire safety Journal, vol. 13,
p 55–68
Schrefler, B.A.: Mechanics and Thermodynamics of Saturated-Unsaturated Porous Materials and
Quantitative Solutions. Applied Mechanics Review. 55(4): 351–388 (2002)
Tenchev, R., Purnell, P.: An application of a damage constitutive model to concrete at high
temperature and prediction of spalling. Int. J. Solids Struct. 42(26), 6550–6565 (2005)
Tenchev, R.T., Li, L.Y., Purkiss, J.A.: Finite element analysis of coupled heat and moisture
transfer in concrete subjected to fire. Num. Heat Transfer, Part A: Appl. 39(7), 685–710 (2001)
Thelandresson, S.: Modeling of combined thermal and mechanical action in concrete. J. Engrg.
Mech., ASCE, 113(6), 893–906 (1987)
Thelandersson S., Martensson A. and Dahlblom O. (1988), Tension softening and cracking in
drying concrete, Materials and Structures, 21, p. 416–424
Torrenti J.M., Granger L., Diruy M. and Genin P. (1997), Modélisation du retrait du béton en
ambiance variable, Revue Française de Génie Civil, 1 (4), p. 687–698 (In French)
Ulm, F.-J., Coussy, O., Bažant, Z.P.: The “Chunnel” fire. I. Chemoplastic softening in rapidly
heated concrete. J. Eng. Mech. ASCE 125(3), 272–282 (1999a)
Ulm, F.-J., Acker, P., Levy, M.: The “Chunnel” fire. II. Analysis of concrete damage. J. Eng.
Mech. ASCE 125(3), 283–289 (1999b)
van Genuchten, M.T.: A closed-form equation for predicting the hydraulic conductivity of
unsaturated soils. Soil Sci. Soc. Am. J. 44(5), 892–898 (1980)
Zienkiewicz, O.C., Taylor, R.L.: The Finite Element Method, vol. 1: The Basis,
Butterworth-Heinemann, Oxford (2000)
Chapter 5
Constitutive Parameters
Fekri Meftah, Colin Davie, Stefano Dal Pont and Alain Millard
The loss of stability of solid phases in concrete, paste and aggregates, when
exposed to high temperatures, affects its pore structure. The physical and chemical
changes occurring in solid phases induce changes in pore size distribution and total
porosity. Noumowe (1995) has given the variation of porosity with different levels
of temperatures for an ordinary concrete and a high performance concrete heated up
to 600 °C. The total pore volume of a heated concrete increases non-linearly with
the increase of temperature. On heating up to 300 °C, a relatively small increase in
pore volume is observed in comparison to the weight losses. This was considered to
F. Meftah (&)
Institut National des Sciences Appliquées, Rennes, France
e-mail: fekri.meftah@insa-rennes.fr
C. Davie
School of Engineering, Newcastle University, Newcastle upon Tyne, UK
e-mail: colin.davie@newcastle.ac.uk
S. D. Pont
Laboratoire 3SR, Université Grenoble-Alpes, Grenoble, France
e-mail: stefano.dalpont@3sr-grenoble.fr
A. Millard
Service d’Études Mécaniques et Thermiques, Commissariat à l’Énergie Atomique
et aux Énergies Alternatives (CEA), Saclay, France
e-mail: alain.millard@cea.fr
© RILEM 2019 67
A. Millard and P. Pimienta (eds.), Modelling of Concrete Behaviour
at High Temperature, RILEM State-of-the-Art Reports 30,
https://doi.org/10.1007/978-3-030-11995-9_5
68 F. Meftah et al.
be due to structural changes resulting from the desiccation and dehydration pro-
cesses occurring mainly in pores of radii less than 40 Å (Andrade et al. 2003).
Heating up to 600 °C will increase the total pore volume. This increase is higher
than would be expected in comparison with weight losses (Fig. 5.1). On the other
hand Bažant and Thonguthai (1978, 1979) observed an upward jump in concrete
permeability of two orders of magnitude. This was related to the fact that the pore
volume available for capillary water must increase significantly when temperature
and pore pressure increase. Then, it may therefore be suspected that either an
expansion of pores has occurred by break down of partition walls, or micro-cracks
are formed (Andrade et al. 2003).
Many authors have investigated micro-cracking (Dougill 1968; Blundell et al.
1976). They have indicated that the differential strain between the aggregate (which
expands) and the cement paste (which shrinks) at different temperatures will ini-
tially induced a small compressive stress in the paste. As the temperature increases,
the compressive stress is reduced and changes to a larger tensile stress. The changes
of porosity with the increase of temperature were measured for several types of
concrete by Schneider and Herbst (1989). Their results showed that the dependence
of porosity on temperature can be approximated for concrete by a linear
relationship:
Fig. 5.2 Porosity from experimental data (Schneider and Herbst 1989) and approximations using
Eq. 5.1
uT ¼ u0 þ Au ðT T0 Þ ð5:1Þ
Another relationship for the evolution of porosity due to the dehydration pro-
cesses has been proposed by Feraille (2000) as follows:
where u0 is the initial porosity and mdehyd is the mass of water which is to be
experimentally determined from the dehydration processes, as will be seen in detail
in the following section. It should be noted also that the above equation takes into
account the Le Chatelier contraction.
70 F. Meftah et al.
The combined reaction in cement paste and aggregate in concrete, which are ini-
tiated when concrete is subjected to elevated temperature, have been examined by
many authors (Harmathy and Allen 1973; Schneider 1982) with the aid of differ-
ential thermal analysis investigations. Dehydration (i.e. release of chemically bound
water) from the C-S-H becomes significant above 105 °C, while the dehydration of
the calcium hydroxide Ca(OH)2 takes place at about 500 °C producing CaO and
H2O. Both will produce a reduction in the mass of the cement paste of the solid
skeleton. However, the strength of concrete is more related to the C-S-H than the Ca
(OH)2.
The amount of water that is released upon heating due to dehydration into the
pores of concrete depends on the amount of hydrated water mhydr (consumed by the
hydration process) before heating. Therefore, it is a function of the degree of
hydration. The degree of hydration before heating depends on the age of concrete as
well as its temperature and humidity history. The degree of hydration may con-
veniently be referred to the equivalent hydration period (maturity) te. It represents
the period of hydration, in water and at 25 °C, needed to give the same degree of
hydration as the actual time period gives at variable RH and T. By fitting the results
of Powers and Brownyard (1948) at 25 °C, it has been found (Bažant and Kaplan
1996) that:
te
mhyd ðte Þ 0:21c ð5:3Þ
te þ se
where the characteristic period se ¼ 23 days for a concrete with w/c = 0.43.
The amount of water that is released by the dehydration mdehydr (kg per m3 of
concrete), when reaching various temperatures is experimentally obtained by
weight loss measurements on heated specimens and can be expressed according to
Bažant and Kaplan (1996) as:
mdehyd ðT Þ ¼ m105
hydr f ðT Þ
C
ð5:4Þ
105 C
in which mhydr is the hydrated water content at 105 °C and f ðT Þ is a function of
dehydration. Gawin et al. (2001) have calculated the dehydrated water content
mdehyd ðT Þ, when temperature increases, by using a step function showing a sharp
change between 200 and 500 °C. The following law can represent this kind of
behaviour:
where c is the cement content, faging is the aging degree of the concrete (between
0 and 1), fstechio is stechiometric factor and f ðT Þ is a function of dehydration (Gawin
et al. 2001).
5 Constitutive Parameters 71
Based on an experimental work on cement paste, Feraille (2000) has noted the
evolution of dehydration and observed thatit is not an instantaneous process but
needs some time to take place. The author has proposed the following formula,
which takes into account the asymptotic evolution of dehydration m_ dehyd through
the characteristic time sdehyd of mass loss:
D E
mdehyd meq
dehyd
þ
m_ dehyd ¼ ð5:6Þ
sdehyd
where meqdehyd is the water mass at the equilibrium measured during the
thermo-gravimetric tests. It is expressed as:
7:5 105 C T 105 2 105 C
meq
dehyd ¼ mhydr 1 exp H ðT 105Þ þ mhydr
100 200 100
T 400 1:5 105 C T 540
1 exp H ðT 400Þ þ m 1 exp
10 100 hydr 5
H ðT 540Þ
ð5:7Þ
105 C
where H is the Heaviside function and mhydr ¼ c 1 þ 0:2faging in (kg/m3).
Equation 5.7 is used to calculate the dehydrated water content mdehyd ðT Þ, and the
evolution of dehydration as follows:
@ @T
m_ dehyd ¼ mdehyd ðT Þ ð5:8Þ
@T @t
It is assumed here that no dissipation occurs during phase change. This in turn
requires the assumption that the liquid water (both capillary and adsorbed) and the
vapour remain in a local thermodynamic equilibrium. Therefore, this local equi-
librium gives rise to a set of restrictions in the equilibrium configuration that must
be satisfied at all times. However, it can be considered that the specific Gibbs
functions of the two phases, considered at the same temperature, are equal at all
times.
In the work of Dal Pont (2004) the following steps are followed to establish a
relationship between the pure liquid water and vapour phases; starting with the
consideration that the enthalpies of both liquid water gl ðpl ; T Þ and vapour gv ðpv ; T Þ
are functions of temperature T and pressure p, the partial derivatives of these two
functions with respect to their pressures can be written as follows:
72 F. Meftah et al.
@ 1
gl ð pl ; T Þ ¼ ð5:9Þ
@pl ql ðpl ; T Þ
@ 1
gv ð pv ; T Þ ¼ ð5:10Þ
@pv q v ð pv ; T Þ
and considering that the liquid water is incompressible (ql is independent of pl ) and
the vapour is an ideal gas qv ¼ pRT
v Ml
. After integrating the two equations above, we
obtain:
pl
gl ð pl ; T Þ ¼ þ C1 ðT Þ ð5:11Þ
q l ðT Þ
RT
gv ð pv ; T Þ ¼ lnðpv Þ þ C2 ðT Þ ð5:12Þ
Ml
pl RT
þ lnðpv Þ þ C1 ðT Þ þ C2 ðT Þ ¼ 0 ð5:13Þ
q l ðT Þ Ml
Now if we consider the case in which three phases (i.e. liquid water, vapour and dry
air) can coexist at the same time, like in the case of a partially saturated concrete,
then the condition of equilibrium over the separation surface must change so that
(pv ¼ pvs and pl ¼ pvs þ pa ) and we can get the following relationship:
ql ðT ÞRT pv
pl ¼ pvs ðT Þ þ pa þ ln ð5:15Þ
Ml pvs ðT Þ
If we know that at 20 °C the saturation vapour pressure is pvs ¼ 2350 Pa and the
dry air pressure is pa ¼ 99;000 Pa, then pg ¼ patm ¼ 101;325 Pa. This can be a
5 Constitutive Parameters 73
particular case of (5.15) which can give the known relationship of Kelvin as
follows:
ql ðT ÞRT pv
pl ¼ patm þ ln ð5:16Þ
Ml pvs ðT Þ
The saturation pressure of pure vapour pvs ðT Þ, which depends only upon temper-
ature, could be obtained from empirical correlation. Tabulated results (Raznjevic
1970), which link the water vapour saturation pressure pvs with temperature T could
be approximated by the following formula (Monteith and Unsworth 1990):
7:5T
pvs ðT Þ ¼ 610:7 10T þ 237:3 ð5:17Þ
It should be noted that this equation is taken until the critical point of water
647.15 K, after which it is not possible to distinguish between the liquid and
vapour. Another equation can be taken for pvs after the critical temperatures fol-
lowing the L-function given in Ju and Zhang (1998) for the superheated steam:
" 2 #
T T
pvs ðT Þ ¼ pvs ð647:15Þ L0 þ L1 þ L2 ð5:18Þ
647:15 647:15
where q0l ¼ 999:84 kg m3 is the water density at the reference temperature T0
and pressure p0l , bl ¼ m1l @m
@T pl is volume thermal expansion coefficient of the
l
74 F. Meftah et al.
@ml
water, al ¼ m1l @pl T is the isothermal compression modulus of water
9 1
(bl ¼ 4:3 10 Pa ), and ml ¼ q1 is the water specific volume. The coefficient bl
l
Gas Properties
Ideal Gases
The following expressions of density of gases can be found in the literature:
Mp
qp ¼ pp ð5:21Þ
RT
where, pp is the density of gas phase, R is the ideal gas constant, T is the tem-
perature, Mp is the molar mass of gas and pp is the partial pressure of the gas phase
(e.g. Gawin et al. 1999; Tenchev et al. 2001; Davie et al. 2006)
Mv
qv ¼ pv þ Dqv ð5:22Þ
RT
where, qv is the density of the vapour phase, Mv is the molar mass of vapour, pv is
the partial pressure of the vapour phase and Dpv is a correction factor to bring the
ideal gas behaviour in line with measured values of vapour density with temper-
ature (Gawin et al. 2002).
5 Constitutive Parameters 75
pg ¼ pv þ pa ð5:23Þ
qg ¼ qv þ qa ð5:24Þ
where pg is the pressure of the combined gas phase, pv is the partial vapour pressure
and pa is the partial pressure of dry air. pg is the density of the combined gas phase,
pv is the density of the vapour phase and pa is the density of the dry air (Gawin et al.
1999; Tenchev et al. 2001; Davie et al. 2006).
The sorption and desorption isotherms are useful for calculation of pores water
vapour pressures. These isotherms are the equilibrium curves of pore relative
vapour pressure versus specific water content at constant temperatures. In another
words, relative humidity RH is controlled while the moisture content within the
sample is measured. Since the degree of saturation Sl is fundamental for realistic
modelling of thermo-hygral behaviour of the specific concrete, it must be deter-
mined during sorption tests at several temperatures (Bažant and Thonguthai 1978).
However, isotherms are only available for moderate temperature. Theoretically,
it should be possible to determine these isotherms also above 100 °C, but this
would require measurements at very high ambient pressure. Such data are not
available (Bažant and Kaplan 1996).
At elevated temperature Bažant and Thonguthai (1978) assumed that a local
thermodynamic equilibrium always exists between the phases of pore water (vapour
and liquid) within a very small element of concrete. An approximate, but rational,
constitutive relation, which relates pore pressure pv, water content w and temper-
ature T is presented as follows:
w h
ws0 im1 pv
¼ RH with RH ¼ ð5:25Þ
c c pvs ðT Þ
2
T0 T þ 10
mðT Þ ¼ 1:04 with T 0 ¼ ð5:26Þ
22:34 þ T 0 T0 þ 10
ðw=cÞRH;T
Sl ðRH; T Þ ¼ ð5:27Þ
ðw=cÞRH0 ;T0
where ðw=cÞRH0 ;T0 is calculated at the initial state. According to this equation it can
be seen that at a higher temperature the saturated state is reached with lower water
content. The degree of liquid saturation Sl can also be related to the capillary
pressure pc. Since a direct experimental determination is difficult, an indirect
determination of the Sl − pc relationship by using the desorption isotherm is gen-
erally employed. Applying Kelvin’s law, which relates the relative humidity to the
capillary pressure, yields an expression for the missing capillary pressure curve and
the degree of saturation Sl ¼ Sl ðpc ; T Þ as follows:
1=mðT Þ
1 ws0 Ml
Sl ð pc ; T Þ ¼ exp pc ð5:28Þ
ðw=cÞRH0 ;T0 c ql RT
The following equation between capillary pressure, temperature and saturation can
be alternatively developed based on the work of Bendar (2002):
rðT Þ
S l ð pc ; T Þ ¼ Sl ð pc Þ ð5:29Þ
rð T 0 Þ
where r(T) is the surface tension of water which depends upon temperature, and
can be given by Le Neindre (1993) as follows:
T n
rðT Þ ¼ r0 1 ð5:30Þ
Tcr
where r0 = 0.155 [N/m2] and n = 1.26. Based upon the pore network model of
Mualem (1976), a relationship between the capillary pressure and the liquid satu-
ration is proposed in van Genuchten (1980) in the form:
" 1=ð1AÞ #A
j pc j
Sl ðpc Þ ¼ 1 þ ð5:31Þ
B
the evolution of Sl with capillary pressure pc using (5.31). We can relate the
capillary pressure pc to the saturation degree Sl by the following relationship Baggio
et al. (1993):
11=b
pc ðSl Þ ¼ a Sb
l 1 ð5:32Þ
where a and b are material parameters identified for each studied case.
5.1.5 Permeability
Darcy’s Law
Pressure driven fluid flow can be written as:
K Kp
vi ¼ rpp ð5:33Þ
li
78 F. Meftah et al.
where vp is the velocity of the fluid p, K is the intrinsic permeability of the concrete,
Kp is the relative permeability of the phase, lp is the dynamic viscosity of the phase
and pp is the driving pressure. pp is typically the pressure of the phase pl or pg but
may be calculated from capillary pressure and may include pressures due to gravity
(e.g. Gawin et al. 1999; Tenchev et al. 2001; Davie et al. 2006; Mounajed and
Obeid 2004; Baroghel-Bouny et al. 1999; Chung and Consolazio 2005).
In a partially saturated porous medium like concrete the liquid and gas transport
will be governed by the absolute permeability of the material, which can be defined as:
K krp
K0 ¼ ð5:34Þ
lp
Fig. 5.4 The intrinsic permeability versus temperature for an ordinary concrete (NC) and for high
performance concrete (HPC) (Kalifa et al. 1998)
cracking and, therefore, a sharp increase in permeability (Gérard et al. 1996; Wang
et al. 1997; Torrenti et al. 1999).
An experimental study was conducted by Picandet et al. (2001) to characterize
the effect of external load-induced cracking on the permeability of concrete after
unloading. Compared to the undamaged sample, a uniaxial compressive load at
90% of the ultimate strength can increase the axial permeability by about one order
of magnitude after unloading. This increase in permeability is directly related to the
maximum applied strain during loading and is due to the formation of a connected
network of micro-cracks, which does not close down completely after the samples
are unloaded (see Fig. 5.5). An expression has been proposed by the author in order
to reproduce the experimental results as follows:
Fig. 5.5 Relationship between increase in permeability and damage value from ‘Grindosonic’
evaluation and the curve fitted according to (5.35) (Picandet et al. 2001)
K ¼ K0 ku ðuÞ þ Kd ð5:37Þ
The relationship between permeability and pore structure of hardening cement paste
at early ages was investigated experimentally by Nyame and Illston (1981).
According to this work a relationship between permeability and the pore structure
can be expressed as follows Meschke and Grasberger (2003):
6ðu u0 Þ
ku ðuÞ ¼ 10d with d ¼ ð5:38Þ
0:3 0:4u0
w3h Wc2
Kd ¼ with wh ¼ for wc wh ð5:39Þ
12lc R2:5
where Sir is the irreducible saturation (sometimes assumed as Sir = 0), and Al is a
constant, with the value from the range 〈1, 3〉.
When the relative humidity RH reaches a value higher than 75%, a rapid increase
of the capillary water flux is observed (Bažant and Najjar 1972). Such behaviour
can be described by the following relationship (Gawin et al. 1999).
" #1
1 RH Bl
krl ¼ 1 þ SA l ð5:42Þ
0:25
where Al, Bl are constants with values in the range 〈1, 3〉. It should be noted that the
term in square brackets has a form originally proposed in Bažant and Najjar (1972).
According to Gawin et al. (1999) this equation has good numerical properties and
allows the use of the irreducible saturation concept to be avoide; a concept that
creates serious theoretical and numerical problems (Couture et al. 1996).
82 F. Meftah et al.
Moreover it can also be described by the formula given by Couture et al. (1996) and
Nasrallah and Perre (1988):
Ag
S
krg ¼ 1 for S\Scr ð5:44Þ
Scr
where Scr is the critical saturation value, above which there is no gas flow in the
medium, Ag is a constant, which usually has value from the range 〈1, 3〉.
Viscosity of Fluids
The dynamic viscosity of liquid water µl [Pa s] depends strongly upon temperature
and can be evaluated, with sufficient accuracy, in a wide temperature range, using
the approximate formula (Thomas and Sansom 1995):
The dynamic viscosity of moist air µg [Pa s], which depends upon the temperature
and the ratio of the vapour and gas pressure, using the data from Mason and
Monchik (1965), can be approximated using the following formula (Forsyth and
Simpson 1991):
a 0:608
p
lg ¼ lv þ ðla lv Þ ð5:46Þ
pg
where the ratio pa/pg is the mole fraction of dry air in the gas phase, and lv is the
water vapour dynamic viscosity:
lv ¼ lv0 þ av ðT T0 Þ ð5:47Þ
with lv0 ¼ 8:85 106 ½Pa s, av ¼ 3:53 108 Pa s K1 . Further, the dry air
dynamic viscosity la is given by:
la ¼ la0 þ aa ðT T0 Þ þ ba ðT T0 Þ2 ð5:48Þ
with la0 ¼ 17:17 106 ½Pa s, aa ¼ 4:73 108 Pa s K1 , ba ¼ 2:22 1011
Pa s K2 . Another formula exists in the literature (Pezzani 1988), which is simpler
since it is function of temperature only:
5 Constitutive Parameters 83
5.1.6 Diffusivity
where Dv0 ¼ 2:58 105 ½m2 s1 is the diffusion coefficient of vapour species in
the air at the reference temperature T0 ¼ 273:15 K and pressure pg0 ¼ 101;325 Pa
(Forsyth and Simpson 1991). Av is a constant, and for the value Av ¼ 1:667 there is
a good correlation with the experimental data concerning vapour diffusion at dif-
ferent temperatures (Mason and Monchik 1965).
The pore space in concrete has a very complex inner structure, that influences the
vapour diffusion process. Therefore, and because of the lack of experimental
results, the simplest way of taking it into account is the introduction of a tortuosity
factor, as it has been suggested by Bažant and Najjar (1972) and Perre (1987),
which considers the tortuous nature of the pathway in the porous media. The
tortuosity factor is usually assumed to be constant in the whole range of moisture
content although (Millington 1959) showed that it can be approximated by the
following formula:
For concrete the range is typically s ffi 0:40:6, used by several authors (e.g. Daian
1989).
Several authors (e.g. Perre 1987) have defined the effective diffusion coefficient,
taking into account the reduction of space offered to the diffusion of gaseous
constituents and the tortuosity, as:
Deff ðSl Þ ¼ fs ðu; Sl ÞDva T; pg ð5:53Þ
where ug ¼ ð1 Sl Þu.
This structure factor fs could be also approached through the expression derived
by Millington (1959) for variably saturated porous media and used by several
authors (e.g. Adenekan et al. 1993; Sleep and Sykes 1993; Mainguy et al. 2001),
2
fs ðu; Sl Þ ¼ u4=3
g ð1 Sl Þ ¼ u
4=3
ð1 Sl Þ10=3 ð5:54Þ
During diffusion of vapour in a porous material with very narrow pores (e.g. in
concrete), the number of collisions of water molecules with the solid matrix is
non-negligible compared with the number of collisions with air molecules. This is
known as the Knudsen effect, but it is generally ignored.
For an ordinary concrete thermal conductivity will decrease when the temperature
rises. Thermal conductivity as a function of high temperature is very difficult to
measure due to the influence of many parameters such as porosity, moisture con-
tent, type and amount of aggregates. The reduction in thermal conductivity due to
the temperature is marked for a concrete with siliceous aggregates, weak for cal-
careous aggregates, and very weak for lightweight aggregates [Fig. 5.6 (Collet
1977)].
The thermal conductivity, k, of siliceous concrete presented in Anderberg (2003)
and used in Swedish fire engineering design is shown in Fig. 5.7. In Eurocode 2
(2004) and Eurocode 4 (2004) a compromise has been made in order to permit the
thermal conductivity to be chosen between an upper and a lower limit as shown in
Fig. 5.7. These limits, for normal weight concrete, are described by Eq. 5.55 in the
range of temperature 20 C
T
1200 C:
where ql is the liquid water density, u is the porosity, ql is the density of liquid
water, S is the saturation with liquid water, qs is the density of the solid phase, and
kd(T) is thermal conductivity of the dry material, whose temperature dependence
can be expressed by means of the following relationship (Harmathy 1970):
kd ¼ kd0 ½1 Ak ðT T0 Þ ð5:57Þ
kd0 being the dry conductivity at the reference temperature T0 and Ak being a
coefficient which takes into account the effect of the thermal damage on the dry
thermal conductivity. An example of the thermal conductivity for concrete using
(5.56), with the following data (Table 5.1) is presented in Fig. 5.8.
5 Constitutive Parameters 87
Heat capacity is the amount of heat per unit mass required to change the temper-
ature of the material by one degree. At a constant pressure p, the heat capacity, Cp,
may be expressed as follows (Harmathy 1970; Harmathy and Allen 1973):
@H
Cp ¼ ð5:58Þ
@Tp
@H @H @n p þ DHp @n
Cp ¼ þ ¼C ð5:59Þ
@T p; n @n p; T @T @T
The enthalpy of evaporation, also called the latent heat of evaporation, depends on
the temperature and may be approximated by the Watson formula (Forsyth and
Simpson 1991):
chemically combined water in the C-S-H, all the water in the Ca(OH)2, and also
some water not held by chemical bonds. It takes 1670 J/kg to establish a bond of
non-evaporable water in the C-S-H while the energy of the water of crystallisation
of Ca(OH)2 is 3560 J/kg (Khoury and Majorana 2003). The enthalpy of dehydra-
tion is supposed constant by Tenchev et al. (2001) and Davie et al. (2006).
5.2.4 Density
In order to assess the volumetric heat capacity qCp which is normally used in the
energy conservation equation, taking into account the effect of latent heat,
Eq. (5.61) and Table (5.2) can be used. In a partially saturated concrete the volu-
metric heat capacity qCp [J/(m3 K)] can be expressed as a combination of the
volumetric capacities of its constituents, as well as the effect of the evaporation and
dehydration processes (Gawin et al. 1999).
@T @T
qCp ¼ qCp eff þ ðDH m_ Þvap þ ðDH m_ Þdehydr ð5:61Þ
@t @t
where (qCp)eff is the effective volumetric heat capacity, DHvap and DHdehydr are the
enthalpies of evaporation and dehydration respectively and m_ vap and m_ dehydr are the
rates of evaporation and dehydration respectively. The effective volumetric heat
capacity in Eq. (5.61) reads:
Table 5.2 Variation of density with temperature influenced by water loss (Anderberg 2003)
q ¼ q20 C For 20 °C
T
115 °C
q ¼ q20 C (1–0.02(T − 115)/85) For 115 °C
T
200 °C
q ¼ q20 C (0.98–0.03(T − 200)/200) For 200 °C
T
400 °C
q ¼ q20 C (0.95–0.07(T − 400)/800) For 400 °C
T
1200 °C
90 F. Meftah et al.
qCp eff
¼ ð1 uÞqs Cps þ u Sql Cpl þ ð1 Sl Þ qa Cpa þ qv Cpv ð5:62Þ
where the thermal capacities Cpi [J/(kg K)] depend upon the temperature of the
material and qi is a density with i ¼ s; l; v; a:S:
The dependence of temperature on the volumetric heat capacity for the solid
skeleton may be approximated by a linear relationship (Harmathy and Allen 1973):
where ð1 u0 Þqs0 Cps0 is the volumetric heat capacity of the solid skeleton at the
reference temperature T0, and Ac is a coefficient which considers the effect of
temperature on the dry volumetric heat capacity. The thermal capacity for the
gaseous constituents, in the case of a perfect gas and at a constant pressure, would
be chosen independent from temperature changes for both air and vapour. Their
heat capacities will be Cpa ¼ 1003:5 and Cpv ¼ 1880:0 (Feraille 2000). Based on
the results for the thermal capacity of the liquid water given by Raznjevic (1970),
an approximated formula has been given by Feraille (2000) as follows:
T 273 2
Cpl ¼ 4180 þ 300 ð5:64Þ
T 715
The solid phase skeleton of concrete generally undergoes degradation upon heating
such that its mechanical properties change (reduce) with increasing temperature.
When considering numerically the mechanical performance of concrete exposed to
elevated temperatures it is necessary to take into account these evolutions.
Primarily, the changes in strength and stiffness need to be considered but other
properties such as fracture energy may also be of interest.
While the strength of concrete may be shown to increase slightly in the early stages
of heating it is generally accepted that there is an overall loss of tensile strength and
various simplified relationships have been employed for numerical models in the
literature.
5 Constitutive Parameters 91
For example (Gernay et al. 2013) employ the function presented in Eurocode 2
(see also Sect. 4.2.2)
where,
Tenchev and Purnell (2005) present an exponential curve, for a specific concrete,
that predicts a much slower reduction in strength with increasing temperature:
ft ¼ 6:5 14 exp 430T 0:9 ð5:68Þ
and (Nechnech et al. 2002) present a relationship derived from experimental work,
which reduces more steeply even than the Eurocode curve. Luccioni et al. (2003)
also present relationships derived from experimental work and these lie in close
proximity to the curve employed by Davie et al. (2012).
These functions are all presented in Fig. 5.10.
ft
1
0.9
Normalised Tensile Strength (-)
0.8
0.7 EC2
0.6 Davie et al.
0.5 Nielsen et al.
0.4 Nechnech et al.
0.3 Luccioni et al. 1
0.2 Luccioni et al. 2
0.1 Tenchev and Prunell
0
0 200 400 600 800 1000 1200
Temperature (C)
experimental data, but reducing more steeply than the Eurocode curves. Luccioni
et al. (2003) present two experimentally derived curves that both suggest signifi-
cantly lower strengths than other authors in the early stages of heating but meet the
upper curves at higher temperatures.
Davie et al. (2012) do not explicitly employ compressive strength in their work
but implicitly the compressive strength follows the same evolution as the tensile
strength (5.67).
These functions are all presented in Fig. 5.11.
fc
1.0
0.9
Normalised Compressive Strength ( -)
0.8
0.7 EC2 (Siliceous)
0.6 EC2 (Calcareous)
0.5 Davie et al.
0.4 Nechnech et al.
0.3 Gawin et al.
0.2 Luccioni et al. 1
0.1 Luccioni et al. 2
0.0
0 200 400 600 800 1000 1200
Temperature (C)
The elastic modulus of concrete is also known to generally decrease with increasing
temperature due to the breakdown of the solid skeleton of the concrete and the
development of micro-cracking. Again, several authors have presented relationships
to describe this although, again, a numerical description is rarely found.
Davie et al. (2010) use a quadratically reducing curve:
2
T T0
EðT Þ ¼ E0 1 0:1 ð5:70Þ
100
while (Tenchev and Purnell 2005) proposed an exponential curve, for a specific
concrete:
32:3
E¼ ð5:71Þ
1 þ expð0:0068T 1:5Þ1:13
and (Gawin et al. 2003; Luccioni et al. 2003) and (Nechnech et al. 2002) employ
functions derived from experimental data. When plotted (see Fig. 5.12) these
functions are all broadly similar.
0.9
0.8
Normalised Elasc Modulus (-)
0.7
Davie et al.
0.6
Gawin et al.
0.5 Nechnech et al. 1
Nechnech et al. 2
0.4
Luccioni et al. 1
0.3
Luccioni et al. 2
0.2 Tenchev and Prunell
0.1
0
0 200 400 600 800 1000 1200
Temperature (C)
The variation of Poisson’s ratio with temperature is much less often considered and
most numerical work has considered it to remain constant with increasing tem-
perature (e.g. Gawin et al. 1999; Nechnech et al. 2002; Davie et al. 2010).
However (Gernay et al. 2013) employed the following function derived from the
work of Luccioni et al. (2003)
T
mðT Þ ¼ m20 0:2 þ 0:8 TTvv20 for T
Tv
ð5:72Þ
mðT Þ ¼ 0:2m20 for T [ Tv
This function is plotted in Fig. 5.13, along with experimental data from Luccioni
et al. (2003).
The dependence of fracture energy release rate in temperature is not often con-
sidered explicitly when modelling concrete at elevated temperatures. However
(Davie et al. 2010) do employ a function as below:
0 TT0 2
Gf ¼ G0f 1 þ 0:39 TT
100 0:07 100 for 0
T
500 C ð5:73Þ
and (Luccioni et al. 2003) do report experimentally derived values in their work.
These are plotted in Fig. 5.14. Interestingly the function presented by Davie et al.
1
0.9
Normalised Poisson's rao (-)
0.8
0.7
0.6
0.5 Luccioni et al. 1
0.2
0.1
0
0 200 400 600 800 1000 1200
Temperature (C)
Gf
1.8
1.6
Normalised Fracture Energy ( -)
1.4
1.2
1
Davie et al.
0.8
Luccioni et al. 1
0.6 Luccioni et al. 2
0.4
0.2
0
0 100 200 300 400 500 600 700 800
Temperature (C)
(2010) indicates that the fracture energy initially increases with temperature before
decreasing, while the work of Luccioni et al. (2003) shows the fracture energy
simply to decrease with increasing temperature.
References
Gérard B., et al. (1996) Cracking and permeability of concrete under tension, Mater. Struct. 29,
pp. 141–151.
Gernay, T., Millard, A., & Franssen, J.-M. (2013). A multiaxial constitutive model for concrete in
the fire situation: Theoretical formulation. International Journal of Solids and Structures, 50
(22–23), 3659–3673.
Harmathy T.Z. (1970) Thermal properties of concrete at elevated temperatures. ASTM Journal of
Materials, 5(1):47–74.
Harmathy T.Z & Allen L.W. (1973) Thermal properties of selected masonry unit concretes,
Journal of American Concrete Institute, vol. 70, no 2, pp. 132–142.
Ju J.W. & Zhang Y. (1998) Axisymmetric thermo-mechanical constitutive and damage modelling
for air field concrete pavement under transient high temperature, Mechanics of Materials, 29,
307–323.
Kalifa P., Tsimbrovska M. & Baroghel-Bouny V. (1998) High performance concrete at elevated
temperature – An extensive experimental investigation on thermal, hydral and microstructure
properties. Proc. of Int. Symp. On high-performance and reactive powder concrete. Aug.
16–20, Sherbrooke Canada.
Khoury G. & Majorana C. (2003) Course on effect of heat on concrete, International Center for
Mechanical Sciences (CISM), 9–13 June 2003, Udine, Italy.
Le Neindre B. (1993) Tensions superficielles des composés inorganiques et des mélanges, volume
K476 of Traite Constantes Physico-Chimiques. Techniques de l’Ingenieur.
Luccioni B M, Figueroa M I, Danesi R F. (2003) Thermo-mechanic model for concrete exposed to
elevated temperatures. Eng Struct, 25 (6): 729–742.
Luckner L., van Genuchten M.T. & Nielsen D.R. (1989) A consistent set of parametric models for
the two-phase-flow of immiscible fluids in the subsurface, Water Res., 25(10), 2225–2245.
Mainguy M., Coussy O. & Baroghel-Bouny V. (2001) Role of air pressure in drying of weakly
permeable materials, J. Eng. Mech., 127(6), 582–592.
Mason E.A. and Monchik L. (1965) Survey of the equation of state and transport properties of
moist gases, Humidity Moisture Measurement Control Science, 3, 257–272.
Meschke G. & Grasberger S. (2003) Numerical modelling of coupled hygro-mechanical
degradation of cementitious materials, J. Eng. Mech., 129(4), 383–392.
Millington R.J. (1959) Gas diffusion in porous media, Science, 130, 100–102.
Monteith J.L. & Unsworth M.H. (1990) Principles of environmental physics. Edward Arnold,
London.
Mounajed G. & Obeid W. (2004) A new coupling F.E. model for the simulation of thermal-
hydro-mechanical behaviour of concretes at high temperatures, Mater. Struct. 37, 422–432.
Mualem Y. (1976) A new model for predicting the hydraulic conductivity of unsaturated porous
media. Water Resour. Res., 12, 513–522.
Nasrallah S. B. & Perre P. (1988) Detailed study of a model of heat and mass transfer during
convective drying of porous media, Int. J. Heat Mass Transfer, 31(5), 957–967.
Neville A.M. (1973) Properties of Concretes. A Halsted Press Book, J. Wiley & Sons, New York,
N.Y.
Nechnech W., Meftah F., Reynouard J. M. 2002) An elasto-plastic damage model for plain
concrete subjected to high temperatures, Engineering Structures, 24, pp. 597–611.
Nielsen, C. V., C. J. Pearce, et al. (2002). Theoretical model of high temperature effects on uniaxial
concrete member under elastic restraint. Magazine of Concrete Research 54(4): 239–249.
Noumowé A. (1995) Effet de hautes températures (20–600 °C) sur le béton - Cas particulier du
béton à hautes performances. Thèse de Génie Civil: Institut National des Sciences Appliquées
de Lyon et Univ. Lyon I, 1995.
Nyame B.K. & Illston J.M. (1981) Relationships between permeability and pore structure of
hardened cement paste, Magazine of Concrete Research, 33(116):139–146.
Ohigishi S., Miyasaka S. & Chida J. (1972) On properties of magnetite and serpentine concrete at
elevated temperatures for nuclear reactor shielding, In International Seminar on Concrete for
Nuclear Reactors, AC1 Special Publication No. 34, Vol. 3, Paper PS34–57, American Concrete
Institute, Detroit, pp. 1243–53.
98 F. Meftah et al.
Perre P. (1987) Measurements of softwoods’ permeability to air: importance upon the drying
model, Int. Comm. Heat Mass Transfer, 14, 519–529.
Pezzani P. (1988) Propriétés thermodynamiques de l’eau (K585), Techniques de l’ingénieur, traité
constantes phisico-chimiques.
Picandet V., Khelidj A. & Bastian G. (2001) Effect of axial compressive damage on gas
permeability of ordinary and high-performance concrete, Cement and Concrete Research, 31,
1525–1532.
Powers T.C. & Brownyard T.L. (1948) Studies of the physical properties of hardened cement
paste. Research Department Bulletin No. 22. Portland Cement Association, Chicago.
Rastiello G., Boulay C., Dal Pont S., Tailhan J., Rossi P., Real-time water permeability evolution
of a localized crack in concrete under loading. Cement and Concrete Research, Vol.56/2,
pp. 20–26: 2014.
Raznjevic K. (1970) Tables et diagrammes thermodynamiques, Editions Eyrolles.
Savage B.M. & Janssen D.J. (1997) Soil physics principles validated for use in predicting
unsaturated moisture movement in Portland cement concrete, ACI Mat J., 94(1), 63–70.
Schneider U. (1982) Behaviour of concrete at high temperatures. Paris: RILEM, 72p. Report to
Committee no 44-PHT.
Schneider U. (1988) Concrete at high temperatures: a general review, Fire safety J. 13, 55–68.
Schneider U. & Herbst H.J. (1989) Permeabilitaet und Porositaet von Beton bei hohen
Temperaturen (in German), Deutscher Ausschuss Stahlbeton, 403, 23–52.
Sleep B.E. & Sykes J.F. (1993) Compositional simulation of groundwater contamination by
organic compounds, 1. Model development and verification, Water Resour. Res., 29(6), 1697–
1708.
Tenchev R.T., Li L.Y. & Purkiss J.A. (2001) Finite element analysis of coupled heat and moisture
transfer in concrete subjected to fire, Numer. Heat Transfer, 2001, 39(7), 685–710.
Tenchev R.T. & Purnell P. (2005) An application of a damage constitutive model to concrete at
high temperature and prediction of spalling, Int. J. Solids Struct. 42(26), 6550–6565.
Thomas H.R. & Sansom M.R. (1995) Fully coupled analysis of heat, moisture and air transfer in
unsaturated soil, J. Eng. Mech., 121(3), 392–405.
Torrenti J.-M. et al. (1999) La dégradation des bétons, Hermès (Eds.), Paris.
Van Genuchten M. Th. (1980) A closed-form equation for predicting the hydraulic conductivity of
unsaturated soils, Soil Science Society of America, 44, 892–898.
Wang K., Jansen D.C. & Shah S.P. (1997) Permeability study of cracked concrete, Cem. Concr.
Res. 27 3, pp. 381–393.
Chapter 6
Conclusion
A. Millard (&)
Service d’Études Mécaniques et Thermiques, Commissariat à l’Énergie
Atomique et aux Énergies Alternatives (CEA), Saclay, France
e-mail: alain.millard@cea.fr
P. Pimienta
Centre Scientifique et Technique du Bâtiment (CSTB),
Université Paris-Est, Marne La Vallée, France
e-mail: pierre.pimienta@cstb.fr
© RILEM 2019 99
A. Millard and P. Pimienta (eds.), Modelling of Concrete Behaviour
at High Temperature, RILEM State-of-the-Art Reports 30,
https://doi.org/10.1007/978-3-030-11995-9_6
100 A. Millard and P. Pimienta