Professional Documents
Culture Documents
Paolo Morabito 1
ISBN 91 - 89580 18 4
2001:18-4
SE
CS PA I
Improved Production of Advanced Concrete Structures
THERMAL PROPERTIES OF CONCRETE Variations with the temperature and during the hydration phase
Author Address
Paolo Morabito
morabito.paolo@enel.it
Task/Subtask no:
Brite EuRam Proposal No. BE96-3843 IPACS - IMPROVED PRODUCTION OF ADVANCED CONCRETE STRUCTURES
Betongindustri AB Cementa AB Selmer ASA Technical University of Delft ENEL Technical University of Lule NCC AB Skanska Teknik AB Technical University of Braunschweig Ismes Norwegian Public Roads Directorate Elkem AS Norcem AS Technical University of Trondheim 31 May 2001 31 May 2001
Project funded by the European Community under the Industrial & Materials Technologies Programme (Brite-EuRam III)
Main tasks and investigations in IPACS and output from the project:
The Expert System synthesises the results from the project into a robust engineering tool for planning and control of the production of concrete structures. It contains modules of varying simplicity, which can be used in all the phases of a construction project from pre-design to maintenance Project Partners: See earlier page Project Co-ordinator: Dr Mats Emborg Betongindustri AB (Heidelberger Zement North Europe) (SE) Dr Hans-Erik Gram/Mr Mats berg Cementa AB (Heidelberger Zement North Europe) (SE) Disclaimer
The author/authors and producer of this report have used their best effort in preparing this report. These efforts include the development, research and testing of the theories and programs to determine their effectiveness. The author/authors and producer make no warranty of any kind, expressed or implied, with regard to these programs or documentation contained in this report. The author/authors and publisher shall not be liable in any event for incidental or consequential damage in connection with, or arising out of, the furnishing, performance, or use of these programs. Editorial/production supervision: Cover design: Prepress material: Printed and published by Prof. Lennart Elfgren Hans Hedlund By report authors Lule University of Technology, Department of Civil and Mining Engineering, Division of Structural Engineering SE-971 87 Lule, Sweden
Table of Content
1 2 3 4 Introduction ......................................................................................................................... 7 The two-linear-parallel-probe method................................................................................. 7 Composition of the tested concretes.................................................................................. 10 Testing programme............................................................................................................ 12 4.1 4.2 5 Tests performed during the hardening stage. .............................................................. 12 Tests performed in hardened concrete samples under temperature variation............. 12
Test results against the temperature variation ................................................................... 12 5.1 5.2 5.3 5.4 Influence of temperature on thermal conductivity...................................................... 13 Influence of temperature on thermal diffusivity. ........................................................ 13 Influence of temperature on specific heat. .................................................................. 14 Modelling of the thermal properties against the temperature variation ...................... 15 Thermal conductivity ........................................................................................... 15 Specific heat ......................................................................................................... 16 Thermal diffusivity............................................................................................... 16
Test results during the hydration stage .............................................................................. 17 6.1 6.2 Experimental tests on a pure cement paste sample ..................................................... 17 Experimental tests in concrete samples ...................................................................... 18 Thermal conductivity ........................................................................................... 19 Thermal diffusivity............................................................................................... 20 Specific heat ......................................................................................................... 20
Abstract
An experimental study on the thermal properties of hardening concrete is presented in this report. It takes into account the results of measurements of thermal conductivity and thermal diffusivity and the results of specific heat calculated from the knowledge of the above measured parameters. The experimental testing method is described as well: based upon the linear heat source theory, it requires the use of two special probes to be inserted into the sample. The experimental programme was forwarded to test concrete mixtures with different kinds of cement and aggregate. The measurements were carried out from the pouring time of cylindrical samples and were ended up when hardened conditions were achieved. The unavoidable temperature variations during the hydration have required the knowledge of the influence of the temperature on the thermal properties. This was studied as well by performing experimental measurements on the same samples of concrete under different levels of temperature. The whole results have been modelled by empirical relationships. They describe both the variation of the thermal properties against the maturity age and the variation of the thermal properties against the temperature in a range going from about 0C up to 100C.
D c
T te TLPP GHP
1 Introduction
Within the task # 2 of the IPACS project, an experimental research program was carried out with the aim to determine the variations of the thermal properties of concrete with the temperature and during the hydration phase. The tests were performed by an innovative transient measuring technique based upon the linear heat source theory. The method, the so-called Two Linear and Parallel Probe method (TLPP) allows to measure simultaneously the coefficients of thermal conductivity and thermal diffusivity. The specific heat is determined according to the following relationship: c= D being: c = specific heat, [kJ/(kgC)]; = thermal conductivity, [W/(mC)]; D = thermal diffusivity, [cm2/s]; = bulk density, [kg/m3]. The experimental research program has mainly taken into account limestone concretes mixed with two different cement types, but additional measurements carried out during the hydration of a pure cement paste sample and in a hardened sample of gravel concrete, used for the construction of a sluice gate in Italy, are presented. The measurements performed during the hydration started right after the pouring of cylindrical samples, 16 cm in diameter and 32 cm in height, and were stopped when no significant change in the thermal properties was detected. A complete set of tests took about up to 200 hours. The effects of the temperature variations are analysed by performing TLPP tests in hardened samples placed inside a controlled climatic chamber by which a temperature variation from about 0C up to 100C is applied to the specimen.
(1)
2 dT (t ,r ) Q = exp r dt 4t 4Dt
(2)
where: Q = the heating power per unit length of the source, [W/cm]; r = the radial distance of the point from the heat source [cm]; t = the time elapsed from the start of heating [s]. The rate of the temperature rise is plotted in Figure 1 against the time; the graph shows that eq.(2) increases until a peak value M is reached and then decreases going to zero at infinite time.
Time
Figure 1 Rate of temperature rise against the time according to the transient linear heat source theory.
It is possible to demonstrate that the thermal conductivity and diffusivity are in relationship with the peak value M and with the corresponding time tM according to the following equations: = Q 4exp(1) M t M
2 D= r 4t M
(3)
The experimental set-up adopted to perform the test is given in Figure 2. Two thermal probes, 4 mm in diameter and 300 mm in height, are inserted in a parallel way into cylindrical samples having a diameter of 160 mm and a height of 320 mm. One probe is used as heating probe and is equipped with an electrical heating wire over the entire length; the other probe is the temperature probe, it is usually spaced 2025 mm from the heating probe and is equipped with a thermistor to measure the temperature.
Thermistor Sample r
The rate of the temperature rise, measured by the temperature probe, is computed from the start of supply the heating probe and is fitted by eq. 2; the coefficients M and tM are thus determined by a last mean square procedure. A typical example of test carried out in a concrete sample is given Figure 3
0.003
Experimental data Best fit curve
dT/dt [ C/s ]
0.002
0.001
0
0 200 400 600 800
Time [ s ]
The TLPP test method is particularly suitable to be applied in damp and porous solids, like concrete. The main features of the method are: little thermal gradients applied to the sample (less than 0.5 C/cm); very short duration of the test; 9
capability to perform tests during the hardening phase of concrete; suitability to test incoherent solids; use on site. The reliability of the method has been verified through comparative measurements carried out by means of the standard guarded hot plate (GHP) method (ASTM C177-63) in samples of reference materials and in fully dried concrete samples. The results of the comparative tests are given in Figure 4
2.5
1.5
0.5
Pyrex sample PTFE sample Dry concrete 1 sample Dry concrete 2 sample
0 0.5 1 1.5 2 2.5 3
Figure 4 Comparison between the standard guarded hot plate (GHP) method and the TLPP method.
10
Glow loss [%] no solving rest [%] CO2 Mn3O4 S [%] [%] [%]
11
4 Testing programme
4.1 Tests performed during the hardening stage.
For each of the two concretes a cylindrical sample, having a diameter of 160 mm and a height of 320 mm, was prepared. Right after the mixing procedure the fresh concrete was cast in the moulds, compacted and then subjected to a set of thermal conductivity and diffusivity measurements by the TLPP method. The set of measurements in each sample started after about 0.5 hours from the beginning of the mixing procedure. To avoid loss of free water during the tests, which in turn causes variation of the thermal properties due to the variation of moisture content, the samples were properly sealed. Additionally, the thermal properties of concrete will depend on the temperature of the sample under test. Due to the development of heat of hydration, it is almost impossible to keep constant the samples temperature. So, empirical relationships between thermal properties and temperature were determined on the same samples after that their have reached complete hydration; such relationships were used to reduce to the reference temperature of 20C the thermal conductivity and diffusivity measurements performed under variable temperature. 4.2 Tests performed in hardened concrete samples under temperature variation.
The samples were placed inside a climatic cell that allows controlling the temperature and the humidity. The temperature was varied from about 0C up to 100C at steps of 10 C. In correspondence of each step the samples were allowed to reach the temperature of the cell and this was checked by the temperature probe inserted into the specimen and then the TLPP test was performed.
12
With such an additional test it is possible to take into account both the effects of different cement types as well different types of aggregate. 5.1 Influence of temperature on thermal conductivity.
The results of the thermal conductivity measurements against the temperature are plotted in Figure 5. They put in evidence that: the thermal conductivity decreases with the increase of the concrete temperature; the relationship does not seem to depend on the cement type; the decrease in thermal conductivity is more pronounced in limestone aggregate than in -1 -1 natural gravel (0.15%C against 0.06%C ).
3
2.9
2.8
2.7
2.6 Limestone concrete - Cem.32.5 IIIA Limestone concrete - Cem.42.5 IIA-L Gravel concrete 2.4 0 20 40 60 80 100
2.5
Temperature [C]
Figure 5 Thermal conductivity against the temperature variation for the three tested concretes. 5.2 Influence of temperature on thermal diffusivity.
The thermal diffusivity measurements are plotted in Figure 6 and lead to the same comments made for conductivity: the thermal diffusivity decreases with the increase of the concrete temperature; the relationship does not depend on the cement type; the decrease in thermal diffusivity is more pronounced in limestone aggregate than in -1 -1 natural gravel (0.27%C against 0.12%C ). It is also to point out that the variations in thermal diffusivity are more pronounced than those in conductivity.
13
1.5
1.4
1.3
1.2
1.1
Limestone concrete - Cem.32.5 IIIA Limestone concrete - Cem.42.5 IIA-L Gravel concrete 0 20 40 60 80 100
0.9
Temperature [C]
Figure 6 - Thermal diffusivity against the temperature variation for the three tested concretes. 5.3 Influence of temperature on specific heat.
The specific heat was calculated according to eq. 1 from the experimental data of thermal conductivity and diffusivity. The results are plotted in Figure 7. They put in evidence that: the specific heat increases with the increase of the concrete temperature; the relationship seems to be independent on the cement type; the variations of specific heat are only a bit more pronounced in limestone aggregate than in natural gravel.
1.10 1.05 Limestone concrete - Cem.32.5 IIIA Limestone concrete - Cem.42.5 IIA-L Gravel concrete
Temperature [C]
Figure 7 Specific heat against the temperature variation for the three tested concretes. 14
5.4
The experimental results seem to put in evidence that thermal conductivity and specific heat could be conveniently described by a linear relationship with the temperature. For both of these parameters the following relationship has been adopted:
XT Xo = (T To ) Xo
(4)
being XT the thermal conductivity or the specific heat at the generic temperature T, the slope of the relationship to be determined from the experimental data, To a reference temperature and Xo the thermal property at the reference temperature. The reference temperature has been assumed to be equal to 20C so the coefficient in eq. 4 represents the relative variation of the thermal property against the unit variation of temperature from the reference temperature of 20C. The variation of thermal conductivity with the temperature is thus described by the coefficient whereas the variation of the specific heat is described by the c coefficient. As the measurements of thermal diffusivity exhibit a slight non-linear trend, its variations are calculated by eq. 1 from the knowledge of and c. 5.4.1 Thermal conductivity
In Figure 8 the experimental results of thermal conductivity are plotted according to the eq. 4
0.1
0.05
-0.05
-0.1
-0.15
Limestone concrete - Cem.32.5 IIIA Limestone concrete - Cem.42.5 IIA-L Gravel concrete -20 0 20 40 60 80
-0.2
15
The coefficient depends on the aggregate type and from a best fit on the experimental data the following values can be assumed: = 0.0006 [C ] for gravel aggregate; = 0.0015 [C ] for limestone aggregate. 5.4.2 Specific heat
-1 -1
In Figure 9 the variations of specific heat are plotted according to eq. 4. The corresponding c coefficients are: c = 0.0007 [C ] for gravel aggregate; c = 0.0016 [C ] for limestone aggregate.
0.15
-1 -1
0.1
0.05
-0.05
Limestone concrete - Cem.32.5 IIIA Limestone concrete - Cem.42.5 IIA-L Gravel concrete
-0.1 -20 0 20 40 60 80
Figure 9 Relative variation of specific heat against the temperature variation. 5.4.3 Thermal diffusivity
The relative variation of thermal diffusivity can be predicted from the following relationship, derived from eq. 1: D T D 20 ( c )(T 20) = D 20 1+ c (T 20 )
(5)
A comparison between the measured and predicted results is given in Figure 10.
16
0.1
0.05 0 -0.05 -0.1 Predicted -0.15 -0.2 -0.25 -20 0 20 40 60 80 Limestone concrete - Cem.32.5 IIIA Limestone concrete - Cem.42.5 IIA-L Gravel concrete
The gradual transition of the cement paste from plastic to hardened material gives rise to a variation of the thermal properties in young concrete. Preliminary tests were performed on a pure cement paste sample of Portland cement with a water/cement ratio of 0.4. The total temperature variation of the sample during the run time of the tests was of only about 3.5 C (see Figure 11) so no temperature correction was carried out on the measured results of conductivity and thermal diffusivity.
27
26
Temperature [C]
25
24
23
22
21 0 5 10 15 20 25
Time [h]
The measurements result of thermal conductivity, diffusivity and specific heat are plotted in Figure 12 as ratio between the actual measurement X(t) and the corresponding one in the hardened condition XH. They put in evidence that: the thermal properties seem to reach a plateau level after about 20 hours; both thermal conductivity and thermal diffusivity increase during the hydration phase of cement whereas the specific heat exhibits a slight decrease; thermal conductivity and thermal diffusivity reach a plateau value after an increase of about 9% and 13%, respectively; as a consequence of that, the specific heat, calculated according to eq. 4, decreases of only about 3%. Water exhibit, from one hand, a lower thermal conductivity and diffusivity than the main solid minerals and, from the other hand, a greater specific heat than solid minerals. As a result, the variations of thermal properties in a hydrating cement paste are caused by the gradual transformation of the free water in bound water and by the consequent increase of the solid/fluid ratio.
1.1
X(t) / XH
1.0
0.9
H = 1.013 [W/(mC)]
D H = 0.00340 [cm2/s] c H = 1.6 [kJ/(kgC)]
Thermal conductivity Thermal diffusivity Specific heat
0.8
0.7 0 5 10 15 20 25
Age [h]
Figure 12 Change in thermal properties during the hardening phase of a pure Portland cement paste sample. 6.2 Experimental tests in concrete samples
As previously mentioned, such tests were carried out in limestone concretes mixed with two different cement types. The tests, being performed during the hydration of the cement, are subjected to unavoidable temperature variations. Such variations were measured as well and the thermal conductivity, diffusivity and specific heat measurements were corrected for the temperature changes according to the corresponding relationships described in chapter 5. It is assumed that the correlations between thermal properties and temperature, determined in hardened concretes, can be applied at any age of concrete. As the aggregate is by far the main constituents of a concrete mixing and the experimental results have demonstrated that the variation of the thermal properties with the temperature depends on the nature of aggregate, such an assumption is likely to be realistic. 18
The experimental results during the hydration stage are thus reduced to the reference temperature of 20C and are plotted against the equivalent age. The latter has been calculated taking account of the temperature variations during the hydration stage and of the experimental relationships of energy activation (Morabito, 2000) for the two tested cements, given by: EA = 45.14 + 0.99 (20 T) EA = 58.80 + 1.43 (20 T)
35
for cement type 32.5 IIIA for cement type 42.5 IIA-L
(6) (7)
25
20 0 24 48 72 96 120 144
Age [h]
Figure 13 Temperature variation during the hydration stage of the tested concretes. 6.2.1 Thermal conductivity
The ratio /20 between thermal conductivity at early age and the conductivity of the hardened concrete at the reference temperature of 20C is plotted in Figure 14 against the equivalent age for the two tested concretes.
1. 1
(te) / 20
1. 0
0. 9
0. 8 Cem ent t ype 32. II 5 IA Cem ent t ype 42. IA 5 I L 0. 7 0 24 48 72 96 120 144 168 192
Figure 14 Variation of thermal conductivity against the equivalent age of hardening concretes. 19
From a best fit on the experimental data, the conductivity variation can be described by the following relationship: (t e ) = 0.92 + 0.08 e t e 20
8.2
(8)
being 20 equal to 2.75 W/(mC) for both types of concrete. 6.2.2 Thermal diffusivity
In a similar way as for conductivity, the ratio D/D20 between the thermal diffusivity at early age and the diffusivity of the hardened concrete at 20C is plotted in Figure 15 against the equivalent age. The rise of thermal diffusivity at early age can be described by the following equation: D(t e ) = 0.93 + 0.07 e t e D 20
2
6 .6
(9)
with D20 equal to 0.0126 cm /s for the concrete with cement type 32.5 IIIA and 0.0129 cm /s for the other tested concrete.
1. 10 D ( e) /D 20 t 1. 00
0. 90
0. 80 Cem ent t ype 32. II 5 IA Cem ent t ype 42. IA 5 I L 0. 70 0 24 48 72 96 120 144 168 192
Figure 15 - Thermal diffusivity against the equivalent age of hardening concretes. 6.2.3 Specific heat
The specific heat has been calculated according to eq. 1 and is plotted in Figure 16.
20
The results put in evidence only a very slight decrease of this parameter during the hydration stage so, for all the practical purpose, a constant value, equal to 0.21 kcal/(kgC) for the two tested concretes, can be assumed.
1.1
c(te) / c20
1.0
0.9
0.8 Cement type 32.5 IIIA Cement type 42.5 IIA-L 0.7 0 24 48 72 96 120 144 168 192
7 Remarks
Some relationships dealing with the variations of thermal properties of concrete, namely thermal conductivity, thermal diffusivity and specific heat, are proposed in this paper. They have been obtained from best-fit analyses on experimental measurements carried out in concrete samples. The tests have been performed by the Two-Linear-Parallel-Probe method, a transient method particularly suitable to measure thermal conductivity and diffusivity in concrete samples. The tests have taken into account the variations of the thermal properties at early ages, caused from the gradual transition from a plastic to a lytic material during the cement hydration, and the variations of the thermal properties with the temperature. The temperature effects, examined in a limestone concrete and in a gravel concrete, have put in evidence that the thermal conductivity and diffusivity decrease as far as the temperature increases whereas the specific heat increases with the temperature rise. The variations, which are more pronounced in the limestone concrete, can be conveniently described by the following general relationship: X T X 20 = (T 20 ) X 20
21
where is a temperature coefficient that depends on the aggregate type and must be experimentally determined. To have an idea of the order of magnitude and the range of variation of , in Table IV the values of this coefficient for two tested concretes are summarised. Table IV Temperature coefficients determined in the two tested concretes. Limestone concrete Gravel concrete Temperature coefficient [C-1] Thermal conductivity, Specific heat, c Thermal diffusivity, D= c -0.0015 0.0016 -0.0031 -0.0006 0.0007 -0.0013
In Table IV, it has been assumed that the thermal diffusivity changes linearly with the temperature and a D coefficient equal to c has been adopted. This rises from eq. 5 where the second term of the denominator can be neglected in respect to 1 for low temperature variations. At last, the temperature coefficients do not seem to be affected from the cement type and this is in agreement with the main role played from the aggregate. The test results carried out during the hydration stage put in evidence a gradual increase of thermal conductivity and diffusivity up to a plateau value corresponding to that one measured in the hardened samples whilst the specific heat can be assumed to be constant during the hydration process. The gradual transformation of the free water in bound water gives rise to a lytic material. Water conductivity and diffusivity are lower than conductivity and diffusivity of a lytic material typically 0.6 W/m/C against 2.9 W/m/C for thermal conductivity and 0.00142 2 2 cm /s against 0.015 cm /s for thermal diffusivity, respectively and this explains the increase of conductivity and diffusivity in hardening concretes. Additionally, a better process of heat conduction is to be ascribed to a solid than to a fluid and composite material because in the latter the unavoidable thermal contact resistances between the different components can obstruct the process of heat conduction. On the other hand, the specific heat of water is greater than the specific heat of a lytic material 4.186 kJ/kg/C against 0.75 kJ/kg/C - so a decrease in specific heat would be expected during the hardening stage. The experimental results have put in evidence a decrease in the pure cement paste of the order of only 3% but negligible variations are observed in the concrete samples. Such a different behaviour is to be ascribed to the role of the aggregate which, from one hand, it is not affected from physical/chemical changes during the hydration stage and, from the other hand, its content in the concrete mixes is of the order of 80% against 12% of cement content. At last, the increases in conductivity and diffusivity in hardening pure cement paste samples have been of about 9% and 13%, respectively. The corresponding increases in concrete samples during the hydration are of the order of 78%. The effects of the aggregate are still evident but to a lower extent than in the specific heat.
22
8 References
Carslaw, H. S. and Jaeger, J. C. Conduction of Heat in Solids. 3rd edition, Clarendon Press, Oxford 1959. Morabito, P. Measurements of the thermal properties of different concretes. 11th ECTP. Pion Limited, London 1989. Lanciani, A. et al. Measurements of the thermophysical properties of structural materials in laboratory and in situ: methods and instrumentation. High Temperature-High Pressure, vol. 21, pp. 391-400, 1989. Lanciani, A. et al. The two-linear-parallel-probe method: a review. 12th ECTP. Pion Limited, London 1993. A.S.T.M. Thermal conductivity of materials by means of a guarded hot plate. ASTM Specification C177-63, 1963. Morabito, P. Determination of the apparent activation energy by adiabatic tests on concrete samples. IPACS Report, Task # 2, Sub-task # 2.2, 2000. Morabito, P. Field test in Italy: sluice gate on the Brembo river. IPACS Report, Task # 5, Sub-task # 5.1, 2000.
23