Nuclear Engineering and Design 67 (1981) 245-258 245

North-Holland Publishing Company

EFFECT OF TEMPERATURE ON STEEL AND CONCRETE FOR PCRV's

U. SCHNEIDER, U. DIEDERICHS a n d C. E H M
Sonderforschungsbereich 148 "Brandverhalten yon Bauteilen", Technische Universitiit Braunschweig, Beethovenstr. 52,
D-3300 Braunschweig, Fed. Rep. Germany

Received 2 October 1981

The effect of temperature on the physical and mechanical properties of typical steels and concretes which are used in PCRV
design are presented. Special consideration has been given to the properties and phenomena concerning thermal behaviour, e.g.
thermal diffusivity and conductivity, specific heat, density, thermal expansion and decomposition effects. Data from the
literature and recent test results will be discussed. A temperature range from 20°C up to the melting points of the materials is
considered.

1. Introduction [2], the risk of a hypothetical core failure has been

estimated. During the course of research it became
Within the scope of deterministic and probabilistic obvious that certain failure conditions - e.g. loss of
safety studies for H T R s in the U S A [l] and in Germany offsite power - may cause an unrestricted core heatup

30.6 m

Refueling

E
CO

Note: No temperotures exceed 3600"C

Fig. 1. PCRV temperature distribution at 24 h for core heatup.

0029-5493/81/0000-0000/$02.75 © 1981 N o r t h - H o l l a n d
246 U. Schneider et al. / Ef]ect of temperature on .steel and concrete/or P( 'R ~"~

3500 - thermal conductivity X ( W / m K ) ,

_ er' g tive core _ - coefficient of thermal expansion a T (10 ~' K t),
(.J
e 3000
.c_ poiot
- modulus of elasticity E ( T ) / E ( 2 0 ° C ) ,
2500 - yield strength fo.2(T)//fo2(20°C),
- tensile strength folt(T)/f, at(20°C).
2000 _ l _ ..I co:. .................... The temperature for the reported data ranges from
Q.
I.........- ' ............. 20 to 1560°C. The survey of data from refs. [3,4] indi-
/ _ maximu . _

II
1500 cated a dependence of steel properties on the chemical
-~-'-- ----.--:gO'" concrete surface
/ - ~ I ...... \1 r composition and the heat treatment during manufactur-
1000
/~__....~'melting point [
ler /' concrete ing process. Therefore, a rough distinction has been
failure,..."" made between different steels in the discussion as there
500
are:
0 - liner steels,
20 40 60 80 100 120 140 160 - reinforcing steels,
time in h - prestressing steels,
low alloyed steels,
Fig. 2. Temperature transients for the core heatup. - heat resistant austenitic steels.

2.1.1. Transformations
whereby core temperatures exceeding 3000°C are to be During heating of steels crystalline transformations
expected. occur at certain temperatures. These transformations
Fig. I shows a PCRV and the temperature distribu- influence the physical properties. For understanding the
tion in it after 24 h core heatup. It is obvious that the typical changes of properties the main transition points
whole structure and the materials are highly stressed of pure iron are presented:
and attacked by extreme temperature conditions. (i) Ac~-transformation (perlite point): 721°C,
Fig. 2 shows the temperature transients for an unre- (ii) A~2-transformation (Curie point:
stricted core heatup of a gas-cooled high-temperature magnetic ~ nonmagnetic a-Fe): 768°C,
reactor. According to the physical model employed in (iii) A~3-transformation (a ~ y-Fe): 960°C,
the core an average heating rate of 6 ° C / m i n is ex- (iv) Aca-transformation (y --, 8-Fe): 1401°C,
pected. Concrete surface temperatures will increase at a (v) melting point: 1536°C.
rate of 0 . 5 ° C / m i n . The exposure temperatures are far Somewhat lower transformation temperatures are
beyond operating temperatures. valid for steels according to the concentration of carbon
Questions which arise from this situation, concern and other alloying constituents. The melting points of
- the performance of the whole PCVR, steels (1460-1530°C) are lower than those of pure iron
- the behaviour of single structural components. accordingly.
Both questions are closely related to the construction
materials itself and to its thermal behaviour. Therefore, 2.2. Physical properties of steel
it was necessary to establish the physical and mechani-
cal properties of the structural materials in the tempera- 2.2.1. Density
ture range from 20°C up to melting temperatures. The The density of steels, fig. 3, decreases with increasing
presented data belong to our own test experiences as temperatures. Up to the melting the density reduces
well as to general literature sources. 13%. Alloying constituents influence the density. The
density of pure iron and low carbon steels increases
during a - y transformation (cfc ~ cbc) and decreases
2. Steel during 1 ' ~ transformation. The density increase about
4% during melting.
2.1. General
2.2.2. Mean specific heat
The following physical and mechanical properties of For practical reasons the mean specific heat is more
steels are discussed: suitable than the true specific heat. Fig. 4 shows that the
- density p (Me/m3), mean specific heat (related to 20°C) of steels is heavily
- specific heat Cp (kJ kg i K i), temperature dependent in the range from 20 to 800°C.
 U. Schneider et al. / Effect of temperature on steel and concrete for PCR V's 247

8.4 position do not influence the specific heat. Only high

alloyed heat resistant steels indicate a somewhat differ-
% ent behaviour. Spontaneous changes of specific heat at
the A~3 and Ac4 conversions are indicated by dotted
~ 8.0 lines for pure iron and low carbon steels.
¢- / h e < a t resistQnt
>,

~h 2.2.3. Thermal conductivity and thermal diffusivity

c
Thermal conductivity, fig. 5, and diffusivity, fig. 6, of
7.6 different steels are simultaneously discussed because
liner steel ~
reinforcing steel -~,? ~1 both properties indicate similar temperature effects.
low olloyed steel ~ ~'-,. Thermal conductivity and diffusivity at 20°C decrease
with increasing alloying constituents. With low alloyed
72 steels conductivity and diffusivity decreases with in-
melting low ~ ~,~ creasing temperatures up to 800°C. With high alloyed
cQrbon steel ~,\\,
steels the opposite holds. The reason for this is a de-
crease in the order of crystal lattice with an increasing
6.8 part of alloying constituents. This influences the elec-
0 400 800 1200 1600
tronic part of the thermal conductivity. From 800°C to
temperoture in °C melting conductivity and diffusivity of all steels increase
slightly, independent of concentration of alloying con-
Fig. 3. Density of different steels. stituents.

70 \
At melting it increases spontaneously about 20%; this is
due to the melting enthalpy, which is necessary for
\
melting. Apparently, small changes in chemical corn-
\
\
60 \
090 \ .~...Iow carbon steel

rnelt[ng low \
corbon steel i~-
I 50
\
0.80 I--
liner steel I
I v

,¢. reinforcing steel I E ~'\, liner steel

I ~///////~ ~reinforcing steel
low alloyed steel I
40
070
.c_

,o. \. lJ
._o steel I /',_'~, I tt
ul
0.60
~ austenitic steel
s!stont
8 30

0.50 20

040
lkW/h -=3600kJ
I lO
L
0 400 800 1200 1600 0 400 800 1200 160(

temperature in oC ~- temperature in °C
Fig. 4. Mean specific heat of different steels. Fig. 5. Thermal conductivity of different steels.
 248 U. Schneider et al. / Effect of temperature on steel and concrete fin" P('R V's

20

ul
16 ' I
/ l o w carbon
L--- stee[
"7

o
v

N
E .c

12
.E " ~ \ \ line steel
"~__~'//~%~ ~ " r ei nrfor c i ng steel
.>
m
v,-

:6
-6 low a oyec x~X" ~ extrapolation / ~'
steel ' ~ _ \ _ _ / + ~ ,--:----"+"----I I. -..J
o~
~.../...,......:r'" l:~_

8
he/at resistant ~t~
oustenitic steel

0 400 800 1200 1600

~- temperature in *C o 4oo 800 1200 1600
Fig. 6. Thermal diffusivity of different steels. temperature in *C
Fig. 7. Coefficient of thermal expansion of different steels.

The conductivity changes at At3 (a 4 7 transforma-

tion: decrease of 7% for pure iron) and at melting 2. 3. Mechanical properties of steels
(increase of 4%). The diffusivity changes at At2 (Curie
point), where diffusivity reaches a minimum (the reason 2.3.1. General
is the maximum of true specific heat), and at melting The strength characteristics and mechanical proper-
(diffusivity increases 20%). ties of steels depend on many factors:
(a) amount and type of alloying constituents,
2.2.4. Coefficient of thermal expansion (b) heat treatment during manufacturing,
The coefficient of thermal expansion of different (c) retreatment in the cold state (cold twisting or
steels, fig. 7, shows two different regions: One region cold drawing).
covers the coefficient of thermal expansion of low al- Especially, distinctions are necessary between rein-
loyed steels, the other region for which the coefficient is forcing and prestressing steels. Structural steels and
about 30-40% higher matches high alloyed steels. For reinforcing steels behave similarly with respect to cer-
these steels the coefficient of thermal expansion in- tain properties due to their similar chemical composi-
creases nearly linear with temperature. There are no tions.
points of discontinuity, because these types of steel In this section the behaviour of reinforcing and
undergo no transformations. prestressing steels is presented in the temperature range
The steep decrease of the thermal expansion of the from 20 to 800°C. The information given here is neces-
low alloyed steels at about 800°C is due to the a ~ y sarily more general, because comprehensive investi-
transformation. At the end of 7 --" 8 dilatation thermal gations leading to more specific statements are lacking.
expansion increases. The expansion of melt is about Differences in the properties of similar types of steels
40% higher than the original value at ambient tempera- from country to country are possible, due to different
tures. test specimens and test procedures.
 U. Schneider et al. / Effect of temperature on steel and concrete for PCRV's 249

2.3.2. Modulus of elasticity lZ,O I i

1] hot rolled steel
Fig. 8 shows the modulus of elasticity of four Swedish BSt 2201340
reinforcing steels. The modulus decreases nearly linear O 120 =- 2) hot rolled steel
with increasing temperatures. Similar curves are availa- BSt 4201500
3) cold twisted steel
ble for structural steels, which indicate a somewhat ,,..o BSt 4201500
100 " ~ - 4) cold drawn steel
favourable behaviour. Most investigators indicate that A
..% it 42O1500
the modulus of elasticity is not dependent upon micro-
structure of the steel. The data of fig. 8, to a certain 80
extent, are representative for steels used in concrete
construction. 60

40
2.3.3. Yield strength
Fig. 9 shows the yield strength of German rein-
forcing steels. It should be noted that hot rolled steels 20
indicate higher strength than cold twisted and cold
drawn steels. At 500°C the mean value of yield strength 0
200 400 600 800 1000 1200
is about 50% of its original value.
-~- temperature in *C

2.3. 4. Tensile strength Fig. 9. Yield strength fo2(T) of different reinforcing steels in
Fig. 10 shows the tensile strength of 30 types of the heated state.
prestressing steels according to a Belgian investigation.
i i j
70% of all data lay in a comparatively small band of the 30 types of steels
whole area of scatter and almost all test results were wires : d'5 to 7ram
120
between the two extreme boundaries. Steels which are ? strands : ~ 5 to 12mm
tempered indicate higher values than stabilized and cold strength: 1500 to 2200
=:= 100 " ~ N/mrn2 at 20 °C.
drawn steels.
E
80

1.4 i
KS
• KS
i i
40, IS10mrn
4 0 , ~, 8 r a m
==

c
6O

,~0
~ ?0 % of all d~ata

1.2 • KS 60, ~ 8mm --

o KS /-,0SE, ~, 8 m m

1.0 20

g 0.8 .
0
o
0 200 ~6O 600 8OO 1000 1200
LU
0.6 "~,~,~/,~ m e a n curve
A
I.-
~-- temperature in °C
LU

0./. Fig. 10. Tensile strength fult(T) of different prestressing steels

in the heated state.
0.2

0
I 3. Concrete
0 2OO 1,00 6OO 8OO IOOO

3.1. General
temperature in °C

In this section it is endeavoured to give a summariz-

Fig. 8. Modulus of elasticity of different reinforcing steels. ing review of the state of knowledge concerning physical
250 U. Schneider et al. // Effect of temperature on steel and concrete fi)r P( 'R I/'~

properties of concrete at high temperatures. More par- 3.2. T h e r m a l reactions in concrete

ticularly, the range from 20°C up to melting of concrete
(at about 1350°C) will be examined. The melting pro- If concrete made with portland cement or with blast-
cess has been included in this treatment of the subject furnace slag cement is subjected to heat, a number of
because the advanced safety considerations relating to transformations and reactions of many different kinds
particular reactor configurations necessitate this ap- occur already with only a moderate rise in temperature.
proach [ 1,2]. Not surprisingly, our knowledge of material Apart from the crystal transformations occurring mainly
properties above 1000°C is very limited. in the aggregates, these phenomena comprise the so-
In the first part of this section the thermal stability called degradation reactions, these being reactions which
of concrete will be considered - mainly on the basis of bring about a progressive breakdown in the structure of
the present authors' own research work and results [5]. the concrete. They occur more particularly in the
It deals with crystal transformations, recrystallization hardened cement paste, but also in the aggregates. At
processes and degradation (dehydration) reactions due low temperatures these reactions mainly take the form
to the liberation of water and carbon dioxide in the of dehydration and water expulsion reactions. At higher
concrete. The heat effects occurring in connection with temperatures, decarbonation of carbonates will play a
the individual processes have been quantitatively de- prominent part if the concrete contains limestone ag-
termined as far as possible and are likewise discussed. gregates. Above 1300-1400°C concrete exists in the
With the aid of the thermodynamic data it is possible to form of a "melt". The reactions initiated during the
explain the temperature dependent behaviour of the heating of concrete can be studied by differential ther-
physical properties of concrete presented in the second mal analysis (DTA). Quantitatively, it is more particu-
part of this section. larly possible to study those reactions which are associ-

Table l
Transformation and decomposition reactions of quartzitic and limestone concrete

Temperature Transformation or Heat of reaction or Heat of reaction Reacted quantity

range decomposition reaction transformation or transformation (kg/m 3 concrete)
(°C) (kJ/kg) (MJ/m 3 concrete)

30- 120 Evaporation of physically bound Heat of evaporation of 290 130 kg water
water water: 2258
30- 300 Breakdown of gel, first stage Heat of hydration: 250 20 < 78 kg hardened
of dehydration cement paste
120- 600 Release of zeolitically bound Heat of evaporation Of 135 60 kg water
or chemisorbed water water: 2258
450- 550 Decomposition of portlandite 1000 <40 <40 kg
Ca(OH)2 ~ CaO + H20
570 Transformation of quartz Quartzitic: 5.86 8.8 1500 kg quartz
a ~ f l SiO2 Limestone: 5.90 1.2 200 kg quartz
600- 700 Decomposition of the Heat of hydration: 500 120 < 240 kg hardened
CSH-phases; formation cement paste
of fl-CzS
600- 900 Only limestone: Heat of decomposition: 1637 2360 1600 kg limestone:
decarbonation of the limestone CaCO 3-proportion
90%
from Melting of concrete, Melting heat: 500-1000 Quartzitic: 1575 2100 kg concrete
1100-1200 formation of vitreous Limestone: 1125 1500 kg concrete
substances
 251

ated with a conversion of mass, which manifests itself in Table 2

a distinct change in weight. Investigations relating to Heat of ablation of different concretes
this were carried out for concrete with the aid of a super
thermo-balance. The reactions occurring in the concrete, Type of Heat of Author
together with their temperature ranges, the heats of concrete ablation
(MJ/kg)
reaction or transformation, and the reacted quantities
(as based on the present authors's experience) are listed
Basaltic 6 -+3 Muir [6]
in table I. A limestone and a quartzitic mass concrete Limestone
made with blast-furnace slag cement were considered.
Basaltic 4.5 -+0.5 Chu [7]
Limestone
3.3. Melting and erosion of concrete
Quartzitic 2.2 Hildenbrand et al. [8]
Calcitic 1.5
3.3.1. General
Above 1200°C and up to 1300°C some components Quartzitic 2.4 Schneider et al. [5]
of the concrete begin to melt. During melting, some Calcitic 3.2
aggregates (e.g. igneous rocks such as basalt) show
degassing and expansion phenomena accompanied by
release of gases trapped within the rock at the time of
its formation. Above 1300-1400°C concrete exists in defined as the heat, which is dissipated per unit mass of
the form of a "melt", i.e., it has t h e n turned into a the material during a steady-state erosion process result-
liquid. ing in the removal of this unit mass. It is composed of
Little information on the actual process itself and the sensible heat to be provided and the heats of reac-
the phases that are formed in connection with it is to be tion and transformation.
found in the literature. F r o m such observations as have In table 2 heats of ablation for different concretes
been reported, however, it appears that the liquefaction are summarized. There are great differences, which re-
of the concrete commences with the melting of the sult from test conditions and the kind of determination.
matrix of hardened cement paste and that melting of
the aggregates takes place only after that [6-8]. The 3.3.3. Erosion rate
melting points of the aggregates vary greatly. At 1060°C The "erosion rate" describes the velocity at which
basalt is at the lower limit of all types of rock in this concrete subjected to high temperatures (1200°C and
respect, whereas quartzite does not melt below about
1700°C (cristobalite begins to produce a melt at 1713 ±
5°C). The latent heat to be supplied for the transition Table 3
from concrete from a "solid" to a "melt" is sometimes Erosion rates of different concretes
called, in the literature [6-8], the heat of melting. In
Type of Erosion Test Author
reality, however, a whole series of reactions take place concrete rate procedure
particularly in the range of temperatures in which the (mm/min)
melting is observed, and these reactions are themselves
in turn connected with heats of transformation, which Basaltic 20 Metallothermic Sutherland [9]
can sometimes be much greater than the actual heat of Limestone reaction
melting. The thermal process during the melting of Siliceous 40 Metallothermic Perinic
concrete has hitherto not been investigated at all from reaction et al. [10]
this point of view.
Basaltic 22 Molten pool Peehs
Limestone 35 Arc-heating et al. [8]
3.3.2. Heat of ablation
In connection with the accidental situation of nuclear Basaltic 12 Plasma jet Muir [6]
Limestone
reactor core melting it is of special interest to know how
much energy can be dissipated by the concrete of the Limestone 66 Immersed in Ehm et al. [1 I]
containment structure. The term "heat of ablation" is Siliceous 44 molten steel
applied in this context, being adopted fom aerospace Siliceous 25 Metallothermic Ehm et al. [I I]
industry, more particularly in connection with spacecraft reaction
re-entry into the atmosphere. The heat of ablation is
252 U. Schneider et al. / Effect of temperature on steel and concretej~)r P('R V'v

more) is decomposed by mechanical desintegration and - mean specific heat cp (kJ kg ] K i);
by melting processes. - thermal conductivity ~ ( W / i n K ) ;
In a case where concrete is subjected to the direct - thermal diffusivity a (m2/s);
action of molten metal, in addition to the phenomena - coefficient of thermal expansion aq- (10 ~ K ~).
discussed above there may occur substantial changes in The physical properties of concrete - apart from the
the melting temperatures and melting enthalpies, be- aggregates, cement and water-cement ratio - depend
cause at least some constituents of concrete react very on the manner of making and working the concrete, its
violently with molten metal. So the erosion rate can also hydration conditions and (in connection with these) the
be defined as the rate of penetration of molten metal m a n n e r of curing and the ambient conditions to which
into a concrete structure. it is exposed (temperature, humidity, etc.), as well as the
Table 3 shows results from investigations with con- geometric features of the concrete member or structural
crete and molten metals in the temperature range from component concerned. Of course, at the lower end of
1600°C (molten steel) to 2600°C (metallothermic reac- the temperature range, where the water in the concrete
tion). There are great differences in the data because of playes an important part, very large variations in the
different mixtures of concrete, different test conditions physical characteristic values must be expected.
(temperature and time of exposure) and interpretations.
3.4.2. Density
3. 4. Physical properties of concrete The density of various concretes, fig. 11, depends on
the density of their aggregates and their moisture con-
3.4.1. General tent in the temperature range between 20 an 150°C. The
This section deals with the principal physical proper- changes in density that occur within this range are very
ties of various normal-weigth concretes - made with greatly affected by the conditions of storage. At temper-
limestone, basalt or gravel aggregates. For the range atures from 150°C up to the start of limestone de-
from 20°C up to melting (at about 1400°C) the follow- carbonation (at 600°C), there is, in our experience, only
ing characteristics of the material have been studied and a very minor decrease in the density of concrete.
will be discussed here: Dependig on the composition of limestone and on
- density p ( M g / m 3 or g/cm3); the prevailing conditions of heating, decarbonation of

2.6

b~e"e" ~e"e"e"l~-~ I
2.4
' -'l i = I I ~1.0 I
~. - - - ---.4_ - ,"~..._ - ......... 4 [-"
I"-.. I I " '\ melt
2.2

~A~ A -- ---- ---- x . . . .

E 2.0
Ot --[ \ \ ~ \l
r
Schneider et aI. [5]
>, 1.8 i-~. ~ o limestone
\\ ~ x quartzitic
C
i-\ I/ .basaltic
-o 1.6
,,!-~ HiIdenbrond et ol. [81
II ~ limestone
I "4---..~i..L' x quOrtzitic
I 1.4
Harmathy [12]
• limestone
1.2
0 200 400 600 800 1000 1200 1400 1600 1600 2000
temperature ( ° C )
Fig. I 1. Density of different concretes.
 U. Schneider et al. / Effect of temperature o,1 steel and concrete for PCR V's 253

the limestone in the concrete commences between 600 3.4.3. Mean specific heat
and 900°C, finally resulting in a highly porous concrete Fig. 12 summarizes the mean specific heats of limes-
(density 1.4 g/m3). On further heating, the density at tone, siliceous, quartzitic and granitic concretes. The
first continues to decrease (expansion of specimens at measured values are, irrespective of the type of concrete,
virtually constant weight), but afterwards the density somewhat above 0.8 kJ kg ~ K - i at room temperature.
can be expected to increase a little (sintering process). The specific heat is increased by the presence of water
With the start of melting the density increases further, "(cp = 4.1868 kJ kg-1 K - l ) ; furthermore, the reactions
eventually attaining a density of 2.2-2.6 g / c m 2. occurring at high temperatures, which are almost invari-
In contrast with the calcitic concretes, those made ably endothermic in character cause an apparent in-
with siliceous aggregates show a somewhat steeper de- crease of specific heat.
cline in density in the range between room temperature
and 700°C. Here the initial rise in density is again 3.4.4. Thermal conductivity
determined by the expulsion of water, whereas at higher The measured values for the thermal conductivity of
temperatures it is more particularly the greater thermal concrete (see fig. 13) comprise a very wide-ranging set
expansion of quartzite that is responsible for the sharper of values over the whole temperature range envisaged.
decline in density. Basalt concrete shows a rather less The reasons for such variations are not fully under-
marked decrease in density because these aggregates stood. Important influencing factors are the aggregate,
have relatively low thermal expansion. the hardened cement paste, the pore volume (voids
Just below 1200°C the melting of concrete com- content) and pore distribution, and the water content of
mences, characterized initially by the formation of small the concrete. Besides, the test method and the interpre-
"island" areas of melt in the matrix of hardened cement tation of the test results will affect the individual values.
paste. For basalt concrete, incipient melting can be From considerations of transport phenomena the
expected to affect the density from a temperature of following temperature behaviour of the thermal conduc-
about 1200°C upwards, whereas the quartzitic coarse tivity can be predicted: At low temperatures and with
aggregates do not show any measurable melting until a moist concrete very high values for the thermal conduc-
temperature of 1400°C is reached. The density of the tivity exist, because, on the one hand, the conductivity
molten concrete at temperatures of 1400-1500°C is of the solid skeleton of the material is highest and
between 2.2 and 2.4 g / c m 3. because, furthermore, the heat transfer between the

1.8
Hiidenbrand et at. [8]
/ /: /.._ ..... I o o limestone
1.6 • quortzitic
/ fi Harmothy,Allen [ 13]
I i mestone
1.4 MTF , siliceous
_fYir h CoUet,Torver nier
,~ l i m e s t o n e
[ 14]
1.2
© siliceous
• Odeen [15]
t- 1.0 gronitic
u
Pogorzelski [16]
u
Q,, • (different c o n c r e t e s )
o.
in
0.8

I 0.6

0.4
0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200
temperature ( ° C )

Fig. 12. Mean specific heat of different concretes.

254 U. Schneider et al. / l-fleet of temperature on steel and concrete j o t PC'R l '~

i i

silicec~s ['WlZ =0.7

2.5 ~.] Harada [17] gravel (. WlZ =0.6

~ 1.5
~~~{~.
i C Crispino [181 , timestone
~] Harmathy{12] siliceous gravel

*'

e.
limestone
basalt (full scate
o Schneider[lCJ b a s a l t
(dried) DIN
test )

basalt ' 52612

r-
o
(moist)
u

E 1.O-- • ]Hundt {20l limestone(dried}

[ I"m estone ( moist )
c-
• Arg Nat Lab. in[6],limestone

l 0.5

0
0 400 800 1200 1600 2000

temperature ( °C )

Fig. 13. Thermal conductivity of normal concretes with different aggregates.

walls of the pores or voids is relatively good on account monotonically to around 0.35 x 10 6 m:/s at 1000°C.
of the water in them. At higher temperatures the "ap- Because of the considerable effect of moisture content
parent" thermal conductivity becomes somewhat better upon thermal diffusivity and density, the largest dif-
because the temperature gradient due to combined water ferences in the values of the diffusivity are to be ob-
and heat transport becomes less steep. A little below served in the temperature range below 200°C. At higher
100°C the thermal conductivity decreases again, be- temperatures the measured values (apart from those of
cause the water-filled pores are now partly emptied, so Chu) are relatively close together. The reason for the
that the heat transfer conditions between their walls are narrow range of scatter for the values of the thermal
less favourable. Up to 300-400°C there is a further diffusivity as compared with the thermal conductivity is
decrease in conductivity, the reason being that, for one perhaps that the diffusivity can be determined from
thing, the thermal conductivity of the solid skeleton unsteady-state heating tests, whereas in many methods
becomes lower (reduction of the average free path of the of determining the conductivity it is additionally neces-
phonons) and, furthermore, the heat transfer conditions sary to take account of cp and P (with their limited
in the pore space of the concrete are further worsened in accuracy) in the evaluation of the two values. For equal
consequence of drying. On the other hand, however, heating behaviour of different concretes, equal values
from temperatures of 300°C onwards, increased crack- for the thermal diffusivity may be obtained in some
ing develops in consequence of incompatibilities of the cases, whereas the corresponding values for the thermal
aggregates and hardened cement paste, as a result of conductivity are liable to differ greatly, depending on
which the transfer of heat between the pore walls de- how accurate the available values of cp and O are.
creases.
3.4.6. Coefficient of thermal expansion
3.4.5. Thermal diffusivity The differences in the expansion behaviour of con-
The values of the thermal diffusivity (see fig. 14) cretes are reflected in the linear expansion coefficients
display roughly the same trend as those of the thermal (see fig. 15). The values obtained for the limestone
conductivity. From relatively high values at room tem- concrete stored under water (Schneider et al. [5]) can
perature (0.6-1.1 X 10 6 m2/s) they decrease almost probably be regarded as representing the lower limit for
 U. Schneider et a L / Effect of temperature on steel and concrete for PCR V's 255

Hildenbrand etol [81

v limestone
A • quort zitic
Hormot hy,Allen 1131
E • limestone
'o 0.8 Chu 17]
v
x limestone
> Haroda [171
ul 0.6 o siliceous

"10
Schneider [5]
o siliceous
0.4
E
e-
0.2

0
0 200 400 600 800 1000 1200 I/-.00 1600 1800

temperature (°C)

Fig. 14. Thermal diffusivity of normal concretes with different aggregates.

the thermal expansion of concrete. In the lower part of The thermal expansion of quartzitic concretes shows
the temperature range the expansion coefficient de- the same trends as regards its temperature dependence.
pends on the moisture content of the concrete (compare However, it is about twice that of limestone concrete on
limestone concretes). account of the higher thermal expansion of quartzite:
It is to see that moist concrete specimens have a Moreover in consequence of the transformation that
considerably higher thermal expansion in the lower part quartz undergoes at 570°C (from the a- to the fl-form)
of the temperature range than dried ones. there occurs an additional steep increase in expansion
which is especially pronounced in the linear coefficient
of thermal expansion. Basalt concretes display a some-
28 i i what different type of behaviour because basalt has a
H i l d e n b r a n d e l a l . 181 J
26 " q*x)rlzitic f k finecrystalline structure compared to the coarsely crys-
talline structure of quartzite aggregates. This difference
is manifested more particularly in the absence of the
c zz o //\ \ "quartz jump", in the lower thermal expansion and in
o limestone, oI [ ~
the low melting point of basalt. From about 700°C

I/ \,, onwards, however, the siliceous concretes also exhibit

contraction, which initially is attributable to the con-

_ 1_/ '
ira.,,
//
,°
traction of the hardened cement paste, followed by
decomposition of the CSH phases.
~- zz
3.5. Mechanical properties of concrete
....... [~.--! i i
8 ill .../ [ 3.5.1. General
The discussion of the mechanical behaviour of con-
crete is based on three characteristic material properties
as there are:
- modulus of elasticity E ( N / m m 2, respectively %),
t e m p e r a t u r e (oC)
- transient creep ~ (1),
Fig. 15 Coefficient o f thermal expansion o f different concretes. - total strain during heating c (960).
256 U. Schneider et al. / Eflect of temperature on steel and com'rete.lor P('R l'~

3.5.2. Modulus of elasticity 8 /

symbol concrete /
As far as E ( T ) is concerned there are many results J/
available. Fig. 16 summarizes some tests with normal 7 light -weight _
/
;4- /
[ imestone
and light-weight concretes. The individual concrete / /
mixes were made with Portland cement and different 6 I basaltic / /
aggregates. For all concretes the modulus of elasticity
J
quortzitic L// /
5 siliceous /__./
decreases with increasing temperature. It turned out
that the modulus of elasticity being significantly in- g
fluenced by the type of aggregate. Whereas the original
strength and the type of cement have only little in-
g
i
4
ii/i I
1~7"

fluence. Test conditions have a great influence: a sus- 9- 3

tained stress during heating affects the modulus of
elasticity significantly. Stressed elasticities are always
higher than unstressed elasticities.
2 .......-
1
3.5.3. Creep and total strain Y ~./
The deformation characteristics of unsealed concrete o
specimens during heating have been studied in connec- o 100 200 300 400 500 600 700 800

tion with fire research. According to [21] the total -- - ---~.- temperature (°C)
deformation of a heated concrete specimen under con-
stant load can be described by Fig. 17. Nonsteady-state creep function ~(T) of different con-
cretes.
eto, = ( ` h ( T ) - a ( I + c p ( a , T ) ) / E ( T ) ,

where Cth(T ) is the thermal expansion, 6 the stress,

E ( T ) the modulus of elasticity, and ¢p(6, T) the non-
steady-state creep function. The ep-function can be de- tions of loaded concrete specimens being heated at a
termined, if (tot, E ( T ) and c lh(T) are being measured in constant heating rate of 2 K / m i n . Between the theoret-
separate tests. Fig. 17 shows the ¢p-function of different ical results calculated with the formulae mentioned
concretes [22]. A substantial increase of creep rates can above and the experimental results a good agreement is
be observed during temperature increase in the con- obtained up to 800°C. Transient creep values have been
crete. The type of aggregate influences the creep values. employed especially with calculations of PCVRs under
Fig. 18 shows typical results of the total deforma- extreme heating conditions.

100 rate oi heating up 2K rain-1

o.,R* hold t i m e 2 hours
>, 80
~.~.l,~ "'~.':£7- ~
:c,_ '
""-.C.7--
o¢, 60 •\ x
• "\ \
.~.i~_~.~ \

0
.symbol I concrete
quartzitic ~ "\.~"~---I",.
...... siliceous
E 20 .... [imestone

t 0
..........

0 100
basaltic
light-weight

200 300 400 500 600 700

I
800 g00 1000

temperoture (°C)
Fig. 16. Modulus of elasticity of different concretes at elevated temperatures.
 U. Schneider et al. / Effect of temperature on steel and concrete for PCR V's 257

}
[5] U. Schneider and U. Diederichs, Physical properties of
experiment / concrete from 20°C up to melting, Part l and 2, Be-
12 calculation - Z / tonwerk + Fertigteiltechnik, 3 + 4 ( 1981) pp. 141 - 149, 223
-230.
[6] J.F. Muir, Response of concrete exposed to a high heat
flux on surface, Research Paper SAND 77-1467, Sandia
Laboratories, Albuquerque NM, USA (1977).
[7] Tze Yao Chu: Radiant heat evolution of concrete - a
study of the erosion of concrete due to surface heating,
Research Paper SAND 77-0922, Sandia Laboratories, Al-
buquerque NM, USA (1978).
[8] G. Hildenbrand et al., Untersuchung der Wechselwirkung
yon Kernschmelze und Reaktorbeton, Abschlussbericht.
o F6rderungsvorhaben BMFT RS 154, KWU, E/'langen, Mai
1978; s.a.M. Peehs, A. Skokan and M. Reimann, Investi-
-8
gations in Germany of the Barrier Effect of Reactor
Concrete against Propagating Molten Corium in the Case
-12 of a Hypothetical Core Meltdown Accident of a LWR,
p: toad100% \
ENS/ANS Int. Top. Meeting on Nuclear Power Reactor
shor t ~fime streng?t ot20oC Safety, Brtissel, 16-9 October 1978, Vo/. 1.
[9] H.J. Sutherland, Acoustic measurement of the penetration
200 400 600 800
of a molten metallic pool into concrete, Nuc. Tech. 46
temperature (°C) (December 1979) ISO-355.
[10] D. Perinic et al., Betontiegelversuche mit Thermitschmel-
Fig. 18. Experimental and theoretical deformations of calcare- zen. Kernforschungszentrum Karlsruhe, KfK 2572, July
ous concrete being heated under load. 1979.
[11] C. Ehm et al., Verhalten von Beton unter Einwirkung
flfissiger Metalle, Report of Inst. ftir Baustoffe, Massivbau
und Brandschutz der Technischen Universit~tt Braun-
4. Conclusions schweig, 1981 (unpublished).
[121 T.Z. Harmathy, Thermal properties of concrete at elevated
temperatures, Research Paper No. 426 of the Division of
A comprehensive survey a n d research program o n
Building Research, Ottawa (March 1970).
high temperature properties of steel a n d concrete was [13] T.Z. Harmathy and L.W. Allen, Thermal properties of
conducted. Material d a t a between 20°C a n d the melting selected masonry unit concretes, Research Paper No. 562
p o i n t s were derived. W i t h increasing temperatures the of the Division of Building Research, Ottawa (April 1973).
reliability of d a t a decreases. In m a n y cases the actual [14] Y. Collet and E. Tavernier, Etude des propi+t+s du bEton
c o m p o s i t i o n of materials is a prevailing factor a n d must soumis a des temp6ratures +lev6es, Groupe de Travail,
be taken into account. Comportment du Mat6riau B6ton en Fonction de la
Temp6rature, Bruxelles, (November 1976).
[ 15] K. Odeen, Fire resistance of prestressed concrete double T
units, National Swedish Inst. for Mat. Testing, Stockholm,
References
1968.
[16] J.A. Pogorzelski, Thermal properties of some building
[I] W.]. Houghton et al., HTGR Accident Initiation and
materials, Report to RILEM Committee 44-PHT, Warsaw,
Progression Analysis (AIPA), Status Report, General 1980.
Atomic, GA-A 13 617, Vol. I-VIII, 1976/77.
[17] T. Harada et al., Strength, elasticity and thermal proper-
[2] W. KrOger et al., Sicherheitsstudie ftir Hochtemperatur-
ties of concrete subjected to elevated temperatures, ACI,
reaktoren unter deutschen Standortbedingungen, BMI RS
Seminar on Concrete for Nuclear Reactors, West,Berlin,
133, KfA Jiilich, 1981.
October 1970.
[3] F. Richter, Die wichtigsten physikalischen Eigenschaften
[18] E. Crispino, Studies on the technology of concretes under
yon 52 Eisenwerkstoffen, Stahleisen-Sonderberichte Heft 8
thermal conditions, ACI-SP-34, Concrete for Nuclear Re-
(Verlag Stahleisen mbH., DfisseldorL 1973).
actors, Vol. I, ACI, Detroit (1972).
[4] G.J. Gantvoort. The mechanical properties of reinforcing
[19] U. Schneider and K. Kordina, Bestimmung der
and prestressing steel during and after a fire, Institute
W~irmeleitf'ahigkeit des "Konstruktonsbetons 2.3" des
T N O for Building Materials and Building Structures, De-
Kernkraftwerkes SNR-Kalkar, Gutachtliche Stellung-
lft, Report No. B 1-79-55/62.1.3210 (September 1979).
nahme, Institut for Baustoffe, Massivbau und Brandschutz,
TU Braunschweig, 1976.
258 U. Schneider el a/. / I'ffect of temperature on s'teel and concrete/))r P( R I '~

[20] J. Hundt, W~.rme und Feuchtigkeitsleitung in Beton unter and relaxation of concrete at high temperature:,), [tabilita
Einwirkung eines TemperaturgefS.11es, DAfStB, Heft 256, tion, TU Braunschweig, 1979.
t975. [22] U. Schneider, Creep effects under transient temperature
[21 ] U. Schneider, Ein Beitrag zur Frage des Kriechens und der conditions. Prepints of symposium on Fundamental Re-
Relaxation yon Beton unter hohen Tempcraturen (Creep search on Creep and Shrinkage of Concrete. l.ausann¢,
September 1980.