SP 092 PDF

SP92-l
Experience of Fires
in Concrete Structures
by A. K. Tovey and R. N. Crook
Synopsis: This paper outlines the procedures adopted in obtaining

information on fire-damaged concrete structures since 1975,
Details are given on the information received from questionnaires
and a summary of the building and construction types, damage and
repairs are tabula ted, The general conclusion is that concrete
structures behave well under fire conditions with the majority of
cases being repairable.
Keywords: buildings; concrete construction; damage; fire resistance;

fires; performance; prestressed concrete; reinforced concrete; surveys
l
2 Experience of Fires
A. K. Tovey is Principal Structural Engineer in the Building

Group, Cement and Concrete Association, England. Member of the
Institution of Structural Engineers and Chartered Institute of
Arbitrators. Involved with various British Standards committees
and the Concrete Society Fire Resistance Committee.
R. N. Crook, at the time of writing the paper, was with the Cement
and Concrete Association, England. He is now Technical Advisory
Engineer for ARC Concrete Ltd, Conbloc Division. He is a member
of the Institution of Civil Engineers and obtained his PhD on fire
resistance of reinforced concrete from the University of Aston.
INTRODUCTION
A small number of major fires in concrete structures have

been 1vell investigated and reported in various publications ( 1-5).
These indicate the good fire resistance performance of both
reinforced and prestressed concrete structures but, in order to
gain a more comprehensive picture, the Fire Resistance Committee
of The Concrete Society, England, decided to investigate a large
number and variety of fire-damaged concrete structures within the
UK.
An initial questionnaire, as given in Appendix A, was

prepared and distributed to several hundred UK consultants in
order to obtain contact with as many people as possible having
experience of fire-damaged concrete structures.
Having obtained the contact, the next stage was to obtain

detailed information on the structures concerned and, in order to
be consistent, a comprehensive record summary sheet as given in
Appendix B was developed. This is also to be issued as a
questionnaire in future cases of fires.
RF~CORDING OF DATA
The record summary sheet is divided into six sections. The

first part is concerned with the identification of the building
and the contact point for information. Section 2 requires a
description of the building, particularly its design and
construction, in order to have a clear idea of the type of
structure before the fire and its potential resistance to fire.
Of particular interest is the question of structural continuity.
If this was present, it could have had a significant influence on
the structure's behaviour pattern and details or an estimate of
this are requested. Section 3 aims to establish the type of fire
Tovey and Crook 3
that occurred. It may be that either the fire brigade or

subsequent examination by an expert established the severity of
the fire. Such information provides a base against which the
actual behaviour of the construction can be judged. Section 4
asks primarily for the visual record of damage by spalling and its
extent. Spalling occurs in most fires but it is necessary to
establish the circumstances under which it may have serious
consequences. Section 5 asks for data on structural damage to the
parts affected by the fire and how the extent of the damage was
established. This is followed by information in Section 6 on the
repair work that had to be undertaken to put the building back
into commission. A question is asked in the case of
reconstruction as to whether the decision was based on the nature
of the damage or because of other practical reasons.
BUILDINGS EXAMINED
As a result of this investigation, detailed information has

been gathered on the performance, assessment and repair of over
100 fire-damaged concrete structures. The information was
obtained from consultants and from the Cement and Concrete
Association (England) whose staff have visited and advised on the
reinstatement of many fires in concrete structures within the UK.
The types of building examined, the type of construction together
with the degree of damage and decision in respect to repair is
summarized in Appendix C. From this, it will be seen that the
structures covered include dwellings, offices, warehouses,
factories and car parks of both single and multi-storey
construction. Some fires under bridge structures were examined
but only one is listed in Appendix c. The forms of construction
examined included flat, trough and waffle floors and associated
beams and columns of in situ and precast construction in both
reinforced and prestressed concrete.
Examining the items under damage and repair shows that:
(1) most of the structures were repaired and many of those which
were not could have been but were demolished for reasons
other than the damage sustained;
(2) much spalling occurred but most structures still performed

satisfactorily even those where supplementary reinforcement
might now be required under current UK provisions;
(3) almost without exception the structures performed well

during and after the fire.
It is hoped that these investigations will provide essential

information on the behaviour of concrete structures in fire to
enable recommendations given in the Codes and other relevant
documents to be updated. The intention is to prepare a
publication giving detailed findings of the investigation. The
information gathered contains details of tests on components and

methods used for repair. To date, the reports show the
performance of concrete structures in fire is generally excellent.
An example of the information that is being obtained may be

illustrated by reference to Figure 1. This shows an example of
damage to the main structural beams to an in situ reinforced
concrete warehouse built around 1930. The fire occurred in 1983
and there was noticeable separation of the floor slab and buckling
of the main reinforcing bars over the central section of the beam.
The omission of stirrups over the central section has clearly
demonstrated a detailing deficiency. Such details would not
generally be permitted under our current UK Codes but, despite
this, the structure adequately withstood the fire. It is clear
that examination of such cases can identify structural
deficiencies and hence, when appropriate, be used to modify the
Code provisions.
Another case which illustrates a detailing deficiency is

shown in Figure 2. In this case, the structure consisted of a
continuous reinforced concrete waffle floor slab that was detailed
to cater for the appropriate negative and positive bending moments
along the span. In accordance with the bending moment envelope,
the bottom reinforcement can be stopped short of the spans. This
is adequate for strength condi tiona but the effect of the fire
caused the unsupported bottom bars to buckle away from the support
zone, The Joint Report ( 1) recommends that 50% of the bottom
reinforcement should continue to the support and that 20% of the
top bars should be continued across the span. Again, the
structure performed adequately in the fire but had all
reinforcement be taken to the support then the element would
ultimately have performed better and there would probably have
been less spalling and damage.
The building fires investigated varied from minor fires of

short duration through to major fires reaching, and at times
exceeding, the designed fire period. The equivalent fire period
was determined, where possible, from examination of debris, tests
on the elements and consideration of the fire load.
The evidence to date shows all structures to have behaved

well, even those having less than satisfactory details. Improving
detailing can significantly improve the performance under fire and
thus, if these are incorporated into the Codes, it should be
possible to relax some of the Code provisions in respect to
notional fire resistance requirements and yet still maintain a
sufficient factor of safety in the event of a fire.
INVESTIGATION OF BEHAVIOUR OF STRUCTURE
The working group responsible for the review is seeking to

obtain data on the overall performance of the structures and, in
particular, on items which have prompted some concern in the past.
Tovey and Crook 5
One such area being investigated is the overall effect of

expansion and contraction on the complete structure. Although
expansion of a hot suspended floor of a building can cause bending
and shear to be developed in supporting columns and other members,
significant problems arising from such are indeed rare within the
UK. In fact, the data already collected has not yielded a single
instance where significant damage has been noted in a concrete
frame remote from the fire. The only collapses recorded were
those resulting from instability of other structural materfals.
Another area of concern in the UK is the loss of strength or

restraint on cooling which according to the Joint Report (1)
"could lead to rapid collapse". Some loss of strength of concrete
on cooling is not in dispute but the essential point is that the
beneficial restraint due to the effects of expansion obtained when
hot is additional to the strength normally expected during a fire.
If such benefit is lost on cooling then, according to the severity
of the fire, the member still has the appropriate proportion of
strength due to its normal methods of carrying load at normal
temperature. Once again, no immediate evidence has been obtained
which shows adverse or unacceptable behaviour on cooling.
Nevertheless, the data is to be re-examined to identify those
fires where restraint may have had a major influence on the
structural behaviour.
The evidence so far indicates that the major reserve in

strength comes from continuity reinforcement at supports rather
than from expansive restraint during the fire. The former has the
benefit in that the strength is also retained on cooling.
ACKNOWLEDGEMENT
The authors thank The Concrete Society and the Chairman of

the Fire Resistance Committee, Mr D.F. Lyon, for permission to
publish the results of the survey.
REFERENCES
1. The Institution of Structural Engineers., Fire resistance of

concrete structures, Report of a joint committee of the
Institution of Structural Engineers and The Concrete Society,
London, Institution of Structural Engineers, August 1975, 59 pp.
2. Forrest, J.C.M., An international review of the fire

resistance of lightweight concrete, The Concrete Society, 1980,
The International Journal of Lightweight Concrete, Vol.2, No.2,
ppo81-94o
3. Malhotra, H.L., Some noteworthy fires in concrete structures,
London, FIP Congress, 1978, pp.86-98.
4. Malhotra, H.L., Some recent experiences of fires in concrete

buildings, London, FIP Congress, 1982, pp.165-173.
5· Cembureau., Concrete for fire resistant construction, Wexham

Springs, Cement and Concrete Association, 1979, 71 pp.
APPENDIX A
INITIAL QUESTIONNAIRE
1. Have you had experience of a fire-damaged structure

since 19757 Yes/No
2. What type of structure was involved:

a) Shed type
b) Multi-storey
c) Other
3. What was the type of construction:

a) Reinforced concrete
b) Precast concrete
c) Other
4. What type of concrete was damaged:

a) Dense
b) Lightweight
5. How much of the structure was damaged:

a) All
b) Part
6. Was the structure:

a) Demolished
b) Repaired
7. Please give the name and address of a contact for

further details:
Name
Firm
Address
Tel. No,
8. Are there any other comments you wish to add.

Tovey and Crook 7
APPENDIX B
FIRES IN CONCRETE STRUCTURES
Summary Record No, ....
1. REFERENCE DATA
1.1 Name of the building:

1.2 Address:
1.3 Contact:
1 .4 Firm:
1.5 Telephone No:
2. BUILDING DESCRIPTION
2.1 Type of occupancy:

2.2 Contents:
2.3 Plan size:
2.4 Height I No. of storeys
2.5 Type of construction (reinforced, prestressed, in situ, etc):
2.6 Brief description:
2.7 Aggregate type:
2.8 Cover to reinforcement (design I actual):
2.9 Supplementary requirement:
2.10 Detailing for continuity:
2.11 Structural design by and construction date:
2.12 Fire resistance requirement:
3. FIRE DETAILS
3.1 Date of fire:

3.2 Fire duration:
3.3 Fire brigade:
3.4 Estimation of fire severity (e.g. BS 476 equivalence):
3.5 Examination of debris for fire damage assessment:
4. FIRE EFFECTS
4.1 Extent of fire spread (No. of storeys, floor areas involved):

4.2 Extent of structure affected:
4.3 Extent of spalling (localized or extensive):
4.4 Reinforcement exposed (locally or extensively):
4.5 Spalling damage to floors:
beams:
columns:
walls:
5. STRUCTURAL DAHAGE ASSESSHENT
5.1 Name of assessor:

5.2 Brief description of damage floors:
beams:
columns:
walls:
5.3 Hethod of damage assessment:
5.4 Residual deformation:
6. REPAIR
6.1 Was the damaged part repaired or replaced?

6.2 Hethod of repair:
6.3 If replaced, the reason:
7. GENERAL
7.1 Has there been a previous fire in the building?

7.2 Is a report or photographs of the incident available?
7.3 Can all or some of the information given be published?
Tovey and Crook 9
APPENDIX C
CONCRETE SOCIETY - FIRE RESISTANCE COMMITTEE
Working Group A
SUMMARY OF INFORMATION ON
FIRES IN CONCRETE STRUCTURES
Ref Type of Building Type of Construction Damage Repaired ?

No.
Office block In situ hollow pot Localized spalling Yes - reconcreted
Supermarket Prestressed double Spalling of beams Yes - gunite/

tee beams. reconcreted
In situ frame
Office block In situ waffle Spalling to ground Yes - gunite

floor soffit
Office block In situ RC Spalling to soffit Yes - reconcreted

one floor
5 Information
not received *
Department Store In situ RC Spalling to soffit Yes - bonded mortar
one floor
Office block In situ RC Yes
8 College Precast Localized to Yes

Prestressed service shaft steel sub-frame
9 Bus Garage Steel frame Severe No - demolished

+ concrete
10 Office block Steel frame Spalling of beams Yes - reconcreted

+ concrete
11 Furniture shop In situ RC Spalling to ground Load test

and first floor Yes - epoxy mortar
21 Hulti-storey RC and PS Local damage Yes
22 Office block Steel frame Spalling of beams Yes - gunite

+ concrete
23 t~arehouse In situ + Spalling of beams No - demolished

prestressed pot floo
24 Office block In situ RC frame 5 floors damaged Yes/no

hollow pot floor Part demolished
repair by gunite
lO Experience of Fires
APPENDIX C (continued)
Ref
Type of Building Type of Construction Damage Repaired ?
No.
25 Hare house Precast portal Damage to single No - demolished

unit Replaced with
PC units
26 Hospital In situ frame Basement Yes - gunite
27 Tunnel In situ RC Spalling to Yos - major gunite

entire soffit - minor epoxy
28
29 Office block In situ Spalling to soffit Yes/no

25 mm hard soot demo 1 ished bad areas
41 l.ibrary In situ Limited Not required
Office block In situ Basement soffit Yes/no

some demolished
gunited areas
columns jacketed
44 Power Station Steel frame Thermal expansion Yes

+ concrete Distortion of
steelwork
Spalling to concrete
45 Oil jetty Concrete on Severe damage to No - deo1ol ished

steel piles superstructure and replace
Deck burnt through superstructure
(hydrocarbon fire)
46 Office block In situ Spalling of beams Yes - bonded mortar
47 Steel frame
48 Of fica block In situ \o/holc building to Yes/no

some degree some members replaced
Fire burnt for others repaired
one week
61 Warehouse In situ Considerable -

62 School In situ + Precast First floor soffit Yes
structural steel
stiHeners
63 flarket Steel frame + Steel - bad damage Yes/no

precast/prestressed cone - minimal steal replaced
units concrete - gunited
Tovey and Crook ll
Ref Type of Building Type of Construction Damage Repaired 1
No,
64 Foundry In situ 100 mm spalling to Yes - gunite

basement floor
(molten steel)
65 Power Station RC and steel frame Yes -

PC roof
66 Portal frame RC Yes - epoxy mortar
67 Hospital In situ trough Spalling to soffit No - damaged part

demolished
suspect original cone
81 Hotel Nasonry + precast Spalling to soffit Yes/no

Cracking of masonry part demolished
walls mainly repaired
82 Nulti-storey In situ Yes
83 ~lulti-storcy In situ Yes
84 Nulti-storey In situ + precast Yes
85 Information
not received *
86 Factory Yes
87 Various In situ, precast Yes

and steel
88 Hulti-storey In situ Yes
89 Hulti-storey In situ Yes,
90 Nulti-storey In situ Yes
101 Workshop Precast portal Shearing of Yes - steel brackets

scarf joints added
102 Information
not received *
103 School In situ, precast Expansion of steel Yes
prestressed beams bracing punched
through concrete
box sections
104 School Precast panels Damage to purlins Yes - external plates

prestressed purlins glued to purlins
105 Information
not received
Ref
Type of Building Type of Construction Damage Repaired ?
No.
106 Boiler House Precast frame, Damage to roof Yes

omn-ia floors, beams
aac roof slab
107 School In situ + timber Roof collapsed No - demolished

roof although cols sound
108 Information
not received *
109 School In situ walls Steel roof collapsed. Yes
Limited spatting
110 Information
not received *
E1 Shop Steel beams + Spalling of plaster Replastered

precast slabs + two slabs Replaced two units
E2 College In situ + Precast General spalling Yes - gunite
E3 Factory In situ + Precast Slight spalling No - precast beam

replaced
E4 Sports Centre Steel frame + Frame collapsed Yes/no - frame

in situ floor some spalling to demolished
concrete concrete repaired
C1 Basement car In situ frame Spalling to soffit Yes - bonded mortar

Park to high waffle slab
rise flats
C2 Flats In situ + Spalling to soffit Yes

Prestressed
C3 Bridge In situ Spalling to beam Yes - bonded mortar
C4 Flats In situ Yes
C5 Flats Timber frame on Structural frame No - structure

concrete raft destroyed. Spall ing demolished
to floor raft Yes - cement mortar
C6 Garage In situ Spalling to soffit Yes
C7 ~1u l t i-s torey Concrete encased Spalling to soffit Yes - gunite

Brewery steel frame
in situ floor
CB Flats In situ Spall ing to soffit Yes - gunite

Tovey and Crook 13
Ref Type of Building Type of Construction Damage Repaired ?

No.
C9 Warehouse Waffle slab on Spalling to soffit Yes/no - demolished

RC columns waffle,cols repaired
C10 Warehouse Precast portal frame Collapsed No
C11 Flats In situ hollow pot Plaster spall Yes - replace broken
pots
C12 School In situ frame Spalling to soffit Yes - gunite
C13 Coach station Concrete encased Spalling to column Yes - gunite

steel frame + soffit
C14 Warehouse Portal Spalling to frame Yes - hand applied

epoxy
C15 Car park Waffle Spalling to soffit Unknown - gunite

suggested
C16 Office block In situ + Spalling to cols Yes - gunite

woodwool slab burnt woodwool
C17 Barn p,;rtal Spall ing Yes - hand applied

mortar
C18 Car park In situ Spalling to soffit Yes - gunite
C19 Flats In situ + Severe spalling Yes - gunite

omnia floor
C20 Brewery In situ Spalling to soffit Yes - gunite
C21 Retaining wall In situ Spalling to bottom Yes - gunite

of wall
C22 Office Spanform General spalling Recollllllend repair
C23 Flats In situ Spalling to soffit Yes - gunite
C24 Supermarket In situ frame Spalling to soffit Yes - hand applied

some expansion mortar
C25 Car park In situ RC Limited spalling Yes - hand applied

mortar
C26 Squash court Precast lightweight Slight crazing Yes

concrete walls timber gallery
destroyed
C27 Warehouse In situ Spalling to ground Yes - gunite

floor soffit
* Replies to Appendix 8 questionnaire not yet received from consultants concerned.

14· Experience of Fires
Slab cracked Crack

/beam
Separation of slab
Crack
: : : """:
"' /
1 I I 1
_·-1-- -l--LJ. l L//~
~
'---'
Bars buckled Links revoalod nt ends

significant spalling none at mid -span
+
0'
••
Scpa<aUon
at fillet
Section
Fig. 1-Effect of unrestrained bars and slab
Bottom reinforcement
stopped off before suppon
Oars deflected away from

slab with increased spalling
Fig. 2-Effect of non-support of bars at end of waffle slab

SP 92-2
Fires During Nuclear

Power Plant Construction
by M. Diaz-Llanos, V. Sanchez Velasco,
and I. Cerezo Preysler
Synopsis: In some cases, it is necessary to evaluate and document

structures which were subjected to fires during construction. Due
to the lack of official regulations on the subject, the analyses
and documentation on these incidences, when reported to the regula
tory authorities, implied unconventional activities, both for the-
utility and the A/E.
The paper describes several fires affecting nuclear power
plant concrete structures. They were all caused by inadvertent
human actions. The presence of highly combustible auxiliary cons-
truction materials contributed to their propagation.
It is recommended that the cost (purchase and installation)
of these auxiliary materials be evaluated against the use of alter
nate (noncombustible)ones before making any decision, sufficiently
in advance to avoid costly and time-consuming changes that may
affect a usually tight schedule.
As a result of visual inspection, followed by "in situ" and
laboratory tests and engineering analyses of the affected elements,
some elements had to be demolished and reconstructed while others
were found to be acceptable.
Keywords: acceptability; bond (concrete to reinforcement); compres-

sive strength; concrete construction; evaluation; fires; fire tests;
impact hammer tests; modulus of elasticity; nuclear power plants;
tensile stress; tests; ultrasonic tests
15
16 Nuclear Power Plants
M. Diaz-Llanos responsible for the Analysis and Design of Nuclear

Power Plants during the last 15 years. Now is Supervisor Engineer
for BWR at Empresarios Agrupados S.A. Madrid, Spain.
Vicente Sanchez Velasco responsible for the Analysis & Design of
Power Plant during the last 20 years. Now is Supervisor Engineer
for PWR at Empresarios Agrupados S.A. Madrid, Spain.
I. Cerezo Preysler responsible for Analysis and Design of Power
Plants during the last 12 years. Now is Squad leader Engineer at
Empresarios Agrupados S.A. Madrid, Spain.
INTRODUCTION
Two types of fires which occurred during the construction of
Nuclear Power Plant (BWR and PWR) during 1983 and 1984 are descri-
bed in this paper.
The fires occuring in the BWR type Nuclear Power Plant (Fig.
1) corresponded to the ignition of the vertical formwork materials
between walls (Fig. 2), due to incandescent particles, coming from
welds or oxyacetylene cuts, falling on them. The fire was produced
in the junctions between buildings when the second wall which
forms the junction was being prepared for concreting, The fires
caused a local elevation of the temperature in the concrete of the
wall already built and in the reinforcement steel bars of the wall
that was going to be concreted. The aim of the analysis was to
ascertain if the materials of the affected areas were still suita-
ble from the engineering and design viewpoint or if they should be
demolished and replaced.
The fire which occurred in the PWR Nuclear Power Plant in
1983 affected the Electrical Building (Fig. 5), where the form-
works between the 7th and 8th floors burned. As a result of the
visual inspection, the analysis was aimed at determining the areas
to be demolished.
ANALYSIS OF THE BWR FIRES BETWEEN WALLS
The analysis of the fires which occurred between building
walls consisted in the following activities:
Description of the Incident
F-1 At 16 h 25m the fire was detected (Fig. 1) and use was made
of the powder fire extinguishers. At about 16 h 45 m water
hoses were used to extinguish the fire, which was accomplished
at 18 hours. The cause was attributed to the oxyacetylene
cutting of reinforcement steel.
F-2 Water was directed from the roof of the Auxiliary Building
(Fig. 1) at 21 h 49 m. Due to various impediments in the
water supply the fire was not extinguished until 23 h 40 m.
Water continued to be hosed until 2 o'clock the next morning.
The cause of this fire was not clearly identified, although
it was assumed due to welding operations.
Diaz-Llanos, Velasco, and Preysler 17
Visual Inspection
Visual inspection of the fires indicated the same conclusions
in both cases, although of greater dimensions in the second. ~lmg
of the concrete up to 15 mm deep was observed, cracks were not seen
Only superficial concrete spalling resulted in the most
affected areas. As the specified minimum protective covering of 40
mm had been reduced locally to still admissible values, the aim of
the analysis was to substantiate that the constructed work subject
to this incident still complies with the project requirements.
Drilled Sample Test
Table 1 summarizes the results of the drilled concrete and
reinforcement steel samples tested from the affected and unaffected
areas. Discrepancies in the strength properties of the materials
were not observed. The safety margins between the tested and
required materials strengths were sufficient after the two fires.
In Situ Tests
The results of in situ concrete tests with the sclerometer
are included in table 2 (SCHMIDT hammer, type N). A large number of
measurements were taken to suitably cover all the disturbed area.
The results confirmed a local superficial concrete spalling,because
the test results changed substantially with repeated hammer strikes.
In figure 3, the statistical correlation between sclerometer
readings and ultimate concrete strength is given. This study used a
sample of 290 concrete test cylinders, and each sclerometric measu-
rement is the average of six readings. The test specimens were 3, ~
28 and 45 days old.
The statistical interpretation indicates that for a minimum
sclerometric value of 32 corresponds to an average ultimate concrete
strength of 25 N/mm2 with a 90% level of confidence. The specified
concrete compressive strength was 25 N/rnm 2 •
The propagation speed measurements, done by PUNDIT device,are
summarized in table 2. The test samples which broke with ultimate
concrete strengths between 32.7 and 44 N/mm2 had a direct propaga-
tion speed between 3.48 and 3.86 Km/s. Comparing these results with
3.42 to 3.78 Km/s in situ propagation speed (Table 1) is a procedu-
re for qualifying the ultimate concrete strength in the fire
exposed structure.
Simulated Fire
Five reinforced concrete test samples were prepared to carry
out the bond test according to EH-80 (8). A brick wall was cons-
tructed with four of these test samples embedded, subjecting one
of the faces to a simulated fire equivalent to the one under study.
A thermocouple was placed to measure the temperatures reached at a
depth of 15 mm from the face exposed to the fire. The test sample
not subjected to the fire was used as a reference.
Water spray was applied to the simulated fire, as in the live

fires, but not in a continous manner in order to not extinguish it.
Not all the structure that was exposed to the live fired had the
same water spray treatment. The bond test results are considered
representative, because the heat and temperature reached in the
contact between concrete and reinforcement were comparable in the
specimen and in the real structure.
The results of the simulated fire were:
Result of thermocouple reading
Fire duration: 26 min
Temperature at the start: 10.92 C
Maximum temperature: 101.42 C after 19 min
Recovery
after 29 min: 93.52 C
after 2 h 49 min: 41.62 C
after 19 h 13 min:10.42 C
The results of the bond tests are summarized in table 3 and
are acceptable according to code EH-80 (8),
CONCLUSIONS
It is concluded that the fire temperatures and heat in the
affected concrete areas did not reduce the overall and local
structural capacity of the reinforced concrete walls.
It is worth mentioning that, in view of the test results
obtained and of the pertinent visual inspection, superficial damage
only occured in the concrete in isolate areas and at depths limited
to 15 mm. This fact is not considered preoccupying from the struc-
tural viewpoint as a minimum covering of 40 mm is provided for the
rebars. Furthermore, the bond tests carried out to verify if there
was resistance capacity loss in the concrete covering the rebars
has made clear the valid behavior of the former.
The estimation of maximum temperatures reached in the concre-
te during fires is based on:
a) The maximum reading obtained in the thermocouple located at 15
mm from the face exposed to the fire in the simulated test was
101.42 c.
b) From the compared results of the modulus of elasticity that in
dicate a decrease to the order of 10%, and in view of graph A~
25 of ACI-216-12-81 (11), it is estimated that the average tern
perature reached in the concrete is around 1002C, This same va
1ue can also be reached following the indications of PeterSmith
(9) or of P.V. Sullivan and Meller P. Poucher (10).
c) From the concrete color change, which was detected in the
extracted test samples and which affected a depth of not more
than 10 mm, it is gathered that the temperature reached could
be around 3002 C, although it may even be assumed that said
coloring could have been caused by the effects of the chemical
components of the wood panels used as part of the formwork on
burning, due to which it could even be considered that this
temperature of 3002 C had not been reached.
d) In the ultrasonic tests in fire F-2 the samples most affected

by low speeds correspond to the hollow spaces with 3432 and
3332 azimuth, with a relative average speed of 0.86. Using
this datum in figure 4 it can be assumed that the concrete
temperature did not exceed 1702 C.
It is concluded that the maximum concrete temperature due to
fires did not exceed 3002 C. There is some evidence that the tempe
rature may not have exceeded 1702 C. -
Concrete test
No estimate of the concrete temperature during the fire can
be reached from the concrete test sample data. However, the obser-
ved concrete compressive strength from these tests exceeds the
specified design minimum.
Steel test
The tensile yield and ultimate strength for reinforcement
bars exposed to fire were tested and their results were above the
specified minima. Some steel test showed reduced strain values
at rupture, or a loss of ductility.
ANALYSIS OF THE FIRE IN A PWR
The fire affected the Electrical Building during the cons-
truction of a PWH Nuclear Power Plant in 1983. This building was
to be completely constructed (slabs, walls and coiumns) of reinfo£
ced concrete. The concrete used rolling siliceous aggregate, with
a water cement ratio of less than 0.5 and 375 Kg/m3 of cement. It
consists of 8 floors (two of which are basements) and a roof on
which there are two penthouses. When the fire started the 7th floor
was complete, as well as the walls and columns between the 7th and
8th floors. Likewise, the formwork and reinforcement for the 8th
floor slab were in place.
The fire started during the early morning from a small fire
used by the laborers to keep warm. The forms of the floors were
prefabricated aluminum frames with phenolic linings, The wall and
column forms were agglomerated panels supported by wooden beams
glued with resin. The nearness of the small fire to the forms and
their combustibility facilitated propagation of the fire. The fire
lasted approximately two hours, The absence of workers because of a
change of shifts delayes in its detection. The fire, affected
a~roximately a quarter of the surface of the 7th floor, that is 400
m , Practically all the forms installed in this area burned.
An inspection was made to determine the extent and the magn!
tude of the fire damage.
The fire affected approximately a quarter of the 7th floor, and
an area between the 7th and 8th floors. The forms for the 8th
floor with the reinforcement placed had completely burned. This
reinforcement was badly distorted, smoked, and showed slag and
fly ash deposits. The interior wall and columns were greatly
affected by the fire. The concrete protective covering detached
and exposed the reinforcement.
The exterior walls, on the inside faces, showed similar damage.

Their external faces were slightly damaged.
The 7th floor slab fire damage was limited to two areas of
spalling which exposed the reinforcement and another area with
only superficial concrete damage.
Melted parts of the aluminum formwork were found, which
indicated that concrete temperatures of over 600Q C were reached.
Fly ash and vitreous element deposits were found on the
reinforcement for the 8th floor. Their subsequent bond to the
concrete could not be trusted. Furthermore, the fire caused defor
mations in the reinforcement, in accordance with the codes and
standards, made them not suitable for reuse.
However, mechanical tests of these rebars were carried out to
determine the elastic limit, obtaining acceptable results in all
cases except for the rebars of diameter less than 16 mm.
To quantify the effects of the fire on the reinforcement,
tensile strengths tests were carried out on the 16 mm diameter bars
taken from the 7th flcor slab. The strength and ductility proper-
ties were acceptable according to the test results.
It was decided to replace all the 8th floor slab reinforce-
ment, and accept the reinforcement in the walls and columns (all
greater than 16 mm). All other bars that remained in the concrete
after the fire were also be considered acceptable.
It was observed that the wall and column concrete with
considerable fire exposure between the 7th and 8th floors was in a
precarious state, completely cracked, and decomposed. The repair
would have meant chipping off the concrete suface to reach sound
concrete, and repairing the walls with epoxy resin. This meant
considerable cost and difficulty, with uncertainties about the
structural behavior of complete repairs. The state of the construe
tion permitted easy demolition and, therefore, this type of solu--
tion was not justified. Consequently, the problem was limited to
defining the limits of the areas to be demolished.
The areas to be demolished was based on the determinations
of the concrete compressive strength, the elastic modulus and on
the bond between reinforcement and concrete.
The first two characteristics were obtained extracting con-
crete test samples. These were laboratory tested and the results
compared with the specified values.
In situ reinforcement bond tests were carried out, trying to
reproduce as close as possible the Spanish standard EH-80 for the
bond test. Such testing was performed on probably acceptable ele-
ments (basically walls between the 7th and 8th floors).
For the bond tests, wall reinforcement bars were chosen (see
Fig. 6), and cut or isolated at a certain distance from the upper
surface of the wall (60 em).
The concrete was chipped off the isolated bar, except in a
length 10 times the bar diameter. A load was applied to the bar by
a hydraulic jack placed on top of a pair of beams resting against

the upper surface of the wall and against an auxiliary metallic
support. Load phases were applied measuring the bond stresses
corresponding to displacements of 0.01 mm, 0.1 mm and 1.0 mm. The
results were compared to the requirements of standard EH-80 (8)
bond tests.
Furthermore, six tests were performed on reinforcement welds
(tensile and bending) used for those bars without sufficient length
for overlapping. These results were compared to UNE 36088 (5).
The affected areas were determined to be demolished. Basica-
lly, the greater part of the walls, all the columns between the 7th
and 8th floors of the area affected by the fire and the two ~rete
spalling areas on the 7th floor slab.
EXPERIENCE SUMMARY
The steps followed for the qualification of the structures
subjected to fires were:
Visual Inspection
Provides general idea of the damage.
Establish documentation test.
Prepare a precise description of the fire: Geometry and
affected areas; Duration; Burned materials.
In Situ Test
Concrete and reinforcement samples.
Sclerometric measures for the superficial damage.
Propagation velocity measures, if possible through
thickness.
Bond test.
Laboratory Test
Samples taken from affected areas shall be tested and
compared with quality control test results.
Quantitative analysis of sclerometric measures with
ultimate concrete strenght.
Propagation velocity measures as a qualitative interpre-
tation of the ultimate concrete strength.
Simulated fire for bond test.
LIST OF TABLES
1. Vertical construction joint fires. Summary on sample
test
2. Vertical construction joint fires. Summary on in situ
test
3. Bond test according to EH-80
LIST OF FIGURES
1. BWR Plot Plan zones affected by fires
2. Typical section between walls with forms
3. Correlation between ultimate concrete strength and
sclerometer measures in laboratory samples
4. Variation of relative propagation velocity with
temperature
5. PWR Plot Plan and zone affected by fire
6. "In situ" Bond Test
LIST OF REFERENCES
1. "Fabricaci6n y conservaci6n de probetas de hormig6n" UNE
7240, Instituto Nacional de Racionalizaci6n y Normaliza-
ci6n, Madrid, Spain, 1971. (Fabrication and Conservation
of Concrete Samples).
2. "Extracci6n y conservaci6n de probetas testigo de hormi-
g6n", UNE 7241, Instituto Nacional de Racionalizaci6n y
Normalizaci6n, Madrid, Spain, 1971. (Extraction and
Conservation of Concrete Samples).
3. "Ensayo de rotura por compresi6n de probetas de hormig6n','
UNE 7242, Instituto Nacional de Racionalizaci6n y Norma-
lizaci6n, Madrid, Spain, 1971. (Test for Ultimate Compr~
ssive Concrete Strength).
4. "Static Modulus of Elasticity and Poisson's Ratio of
Concrete in Compression", ASTM C-469, American Society
for Testing and Materials, Philadelphia, P a. , USA, 1965.
5. "Barras corrugadas para hornig6n armado", UNE 36088, Ins
tituto Nacional de Racionalizaci6n y Normalizaci6n, Ma=
drid, Spain, 1972. (Steel Quality for ReinforcementBars).
6. "Toma y preparaci6n probetas de acero", UNE 7282, Insti-
tuto Nacional de Racionalizaci6n y Normalizaci6n, Madrid,
Spain, 1972. (Extraction and Preparation of Steel Sarples).
7. "Deformed and Plain Billet-Steel Bars for Concrete
Reinforcement", ASTM A-615-76, American Society forTes-
ting and Materials, Philadelphia, Pa. USA, 1976.
8. "Instrucci6n EH-80 para el proyecto y la ejecuci6n de
obras de hormig6n en masa o armado", Comisi6n Permanente
del Hormig6n, MOPU, 1980. (Bond test).
9. Peter Smith, "Resistance to High Temperatures", Signifi-

cance of Test and Properties of Concrete and Concrete-
Making Materials. ASTM Special Technical Publication
1698, Pages 388 to 419. 1619 Race Street, Philadelphia,
Pa. 19103, 1978.
10. P.J. Sullivan and Meller P. Poucher, "The Influence of
Temperature on Physical Properties of Concrete and Mortar
in the Range 202 C to 4002 C", ACI-SP25, American Crncrete
Institute, Detroit, Mich. (Nov., 1968).
11. "Guide for Determining the Fire Endurance of Concrete
Elements", ACI-216R-81, American Concrete Institute,
Detroit, Mich., 1977.
12. LOGOTHETIS, L. and others. The Influence of High Tempera
tures on Calibration of Non-Destructive Testing of Con--
crete. Materiaux et Constructions, Vol. 14, N2 79 (1981).
t'.,j
TABLE 1 ,f:.
VERTICAL CONSTRUCTION JOINT FIRES SUMMARY ON SAMPLES TEST

2:
AFFECTED ZONES
=
NUMBER
F-1
6
F-2
18
NON AFFECTED ZONES
-
n
('!)
~
'"'!
MAXIMUM 51.22 60.84 ~

~
CONCRETE SPECIFIED LABORATORY TEST=25
STRENGTH MINIMUM 37.89 29.23 TEST SAMPLE = 42.55 '"'!

(Reference1,2,3 ~ 8)
MEAN 45.44 43.80 :sl
~
STD. DEVIATION 4. 56 9. 38 =
......
CJl
CONCRETE (I l
MODULUS OF NUMBER 1 7 MEAN LABORATORY TEST AT 360 DAYS

ELASTICITY
23501
!TANGENT) MEAN 22987 23600
(Reference I ,2 E 4) TEST SAM Pl. E : 25212
NUMBER OF TEST 7 54 SPECIFIED VALUES
MEAN YIELD STRESS 516 490 420

REINFORCING
BARS
MEAN ULTIMATE STRESS 687 688 550
(Reference 5,6 t 7)
MEAN ULTIMATE STRAIN(%) 18 19 18
NOTES:ALL UNITS ARE N/mm2 , EXCEPT ULTIMATE STRAIN IN °/0

(I) 60 mm OIAMETER SAMPLES, LENGTH BETWEEN 86 AND 130 nom
TABLE 2
VERTICAL CONSTRUCTION JOINT FIRES SUMMARY IN SITU TEST
AFFECTED ZONES
NON AFFECTED
F- I F-2 ZONES
P.(l) NP. (2) P. NP. NP.
NUMBER 134 171 196 196 119

v
MAXIMUN 55 54 55 54 51 6i"
N
I I
SCLEROMETER (I) MINIMUM 36 20 21 20 3B
I t:
~
MEAN 44.50 34.20 42.18 37.05 42.47 ::s
0
Cfl
STD. DEVIATION 4.84 605 5.67 6.51 2. 90 ~
~
NUMBER 9
- -
~
Cfl
~
0
PROPAGATION MAXIMUM
- 3.78 - ~
VELOCITY ~
( Km/s)
MINIMUM - 3.42 - ::s
0....
MEt 'II
- 3.50 - ~
~
ctl
-
-
STD. DEVIATION 0.17
- '<
Cfl
I
ctl
~
P. PREPARED SURFACE
NP. NON PREPARED SURFACE
(I) HAMMER REBOUND 1:\J
CJ1
TABLE 3
EH- 80 BOND TEST
BOND STRESS
FOR 0=32 mm Ibm= 4. 16 N /mm2

EH-BO REQUIRED VALUES
FOR IJ= 32mm tbu = 6. 92 N/mm 2
TEST RESULTS
LEFT INDICATOR RIGHT INDICATOR
NON AFFECT ED BY FIRE
SAMPLE Ibm= 7. 79 N /mm2 Ibm= 9.03 N/mm2
1b u = 12.13 N/nim 2 tbu = 13.76 N /mm 2
LEFT INDICA TOR RIGHT INDICATOR

SUMMARY OF RESULTS ,
--·
MEAN VALUE OF FOUR
Ibm= B.3B N/mm 2 Ibm= B.36 N/mm2
SAMPLES SUBJECTED TO
SIMULATED FIRE
lbu =13.70 N/mm2 lbu = 14.03 N/mm 2
4No- POLISHING BLD.
D~
D FIRE F-1
TURBINE BLD.
FIRE
I
S?.
Qj
~
I
Dl
Qj
=
0
00
~
AUXILIARY BLO.
~
..........
):ll
00
C":l
0
~
):ll
=
0...
0 10 20 30 40 m. ~
'"<
00
..........
,..,
(b
1'-.:)
Fig. 1--BWR plot plan--affected zones by fires -l
800 BOO
1 fTt 1
REBARS
·,'.J,I
. u.··.
··o-.J·
.\ .· .
." >o".':J
WOOD FORM
CONCRETE WALL
(ALREADY POURED l
WALL TO BE
CONCRETED
EXPANDED
POL YURETANE
PLASTIC SHEET
PANEL
DIMENSIONS IN mm . BOARD
Fig. 2-Typical section between walls w:ith forms

UPPER 90% CONFIDENCE

LIMIT
40 /
/
/
G ,4;
C\1 G G i_l>"'
/
E
E
'z G (J /
/
J: c_, (, //
I- /
, /s
(,
C)
z
w 30
/
,·,
a:: \.•
I-
(/)
w /
/
1--
w
a:: /
u /
z / l:') /
0 /
0
/
u /
/ e
G
w /
G
/
I- LOWER 90% CONFIDENCE LIMIT
< 2
:::!!
1--
...J
:::l
I
>-
20 30 40
X- SCLEROMETER MEASURES
CORRELATION COEFICIENT =0. 43

HIGHLY SIGNIFICANT FOR 1% LEVEL OF SIGNIFICANCE
(SNEDECOR STATISTICAL METHODS, 1940)
Fig, 3-Correlation between ultimate concrete strength and

sclerometer measures in laboratory samples
1,00
--- r----... .............

0::>0
"'" ..........
"-.. ~
~-
0
0 100 200 300 400 500 600 700 800
TEMPERATURE °C
BY LOGOTHETIS ( 12)
1(,: INITIAL PULSE VELOCITY
VT: PULSE VELOCITY AFTER HEATING
Fig. 4-Ratc Vt as a temperature function

v~
•NO·
REACT01 SLOG.
_ __._~~~~~- - - - ----1---+-
TURBINE BLDG. 0
""
~
I
t:
~
:::::
c(fJ
AUXILIARY BLDG.
-
~
~
(fJ
~
0
~
~
AFFECTED :::::
ZONE
0 10 20 30 40 m. -
0...
--:'"'
(t
'<
g:_
(t
ELECTRICAL BLDGS. '"':
Fig. 5--PWR plot plan--affected zone by fire w

........
~
,__,
fJ 25 REBAR
...
I I THREADED SLEEVE
I
,
yr-
8
II
~
II
, II
HOLLOW JACK
~ II
I I
~ I I
~ II
II STEEL BEAM ( 2 lPN -140 SHAPE)
~
II
1:1
~
lf
I l> 150 mm. : I I I
I
I
'I
I I
I CONCRETE II VERTICAL
I WALL II SUPPORT
I
I FLEXIMETER
I
I
I I
II
I II
II
I il
II ,J
"
BOO
Fig. 6-In situ bond test

SP 92-3
Old Flames Speak Well of Young Concrete

by Peter Smith
Syno2sis: A major three span highway bridge in Southern 0ntario

was-1nvolved in a construction fire while it was being widened
25 years ago. Concrete, ranging in age from 6 days to 20 years,
in an abutment, the adjacent arch rings and spandrel columns was
severely damaged. The structure is now revisited to assess the
long term performance of the remedial works and to consider
whether the investigation and repairs might have taken a
different course if today's testing, analytical and repair
techniques had been available then. The strengthening of the
damaged arch springings and the shotcrete restored concrete
sections exposed to a harsh environment have stood the test of
time. The review of investigative and repair methods identifies
advances made in recent years, and justifies long term
confidence in older procedures which are still in use.
Keywords: arch bridges; concrete construction; damage; fire resis-

tance; fires; nondestructive tests; performance; repairs; shotcrete
33
34 Young Concrete
Peter Smith is Director, Research and Development Branch,

Ontario Ministry of Transportation and Communications. He was
involved in the original investigation and repair of the fire
damaged concrete in 1959, and made the appraisal 25 years later
which is now reported. He is a Fellow and Past President of ACI.
INTRODUCTION
On March 24, 1959 a construction fire, fueled by straw
insulated wooden formwork and timber falsework, seriously
damaged the reinforced concrete arches and abutment of a highway
bridge that was being widened, Fig. 1 shows the scene on the
morning after.
The existing bridge carried a major freeway over the Bronte
Creek, some 40 km West of Toronto, Ontario, therefore the first
imperative was to determine if this 20 year old structure could
be reopened safely to traffic. Once this had been accomplished,
attention turned to investigating the extent and structural
significance of the damage with a view to remedial work.
Naturally, the young concrete in the new arch and abutment
which, depending on the location was only 6, 19 or 40 days old,
was of special concern.
The first surprise in the investigation was the lack of
published information on the effect of fire on young concrete.
A knowledge base had to be assembled from information on mature
concrete exposed to fires and fire tests and interpreted with
cons i derab 1e judgement. Next, investigative methods had to be
selected and used to probe the condition and properties of the
concrete and steel that remained, apparently intact, in order to
analyze what could be salvaged by restoration of the section or
which elements needed strengthening or replacing. Lastly,
specific remedial works had to be devised, specified and
executed which would ensure a durable structure with a long term
performance in no way inferior as a result of the fire.
All these steps are described in detail in a paper
presented at the time (1). The purpose of this paper is to
revisit the bridge to determine if the diagnosis made, and the
remedial work undertaken over 25 years ago have stood the test
of time. In light of this experience, advancements in knowledge
of the effects of elevated temperatures on materials, in investi-
gative techhiques and in repair methods are reviewed to see
whether any of the advancements would warrant a different course
of action if a fire of similar consequence occurred today.
Smith 35
LONG TERM PERFORMANCE PF THE REPAIRED CONCRETE

Nature of the Repairs and Subsequent 5xposure
The investigation determined that the fire damage was more
serious than evident from the obvious surface cracking and
spalling of the concrete and the deformation of exposed
reinforcing steel alone. Concrete, at least to the depth of the
main reinforcing steel and especially on the vertical faces of
the arches, was found to be unsound. Fortunately, the main steel
and the core of concrete in both the old and new arch rings had
been affected little by the fire. Figs. 2 and 3 are indicative
of the extent of the damage and the concrete section requiring
restoration in the 20 year old arch and the 19 day old arch ring
respectively. These arches were the structural members of
greatest concern as to their integrity and the serviceability of
repairs. Ref. 1 contains detailed plots of the test results and
condition surveys which determined the extent of repairs
required and upon which the decision to proceed with repairs was
made.
Except at the arch springings, where very deep spalling
occurred, the 20 year old and the 6 and 40 day old concrete in
the more massive abutment was only superficially damaged since
formwork, which was still in place, did not fully burn. Formwork
also afforded some protection to the underside of the new arch.
As detailed in Fig. 15 of Ref. 1, the arch springings were
repaired by encasernent in a single concrete pedestal.
After removal of loose and suspect concrete, both the young
and old concrete sections remaining were rebuilt with shotcrete.
Following final cleanup by sandblasting, a small gauge wire mesh
was attached to the existing concrete by ramsets. The shotcrete
was then formed against edge strips and lightly trowelled to the
intended section and 25 mm minimum cover over the reinforcing
steel. The shotcrete was then wet cured for 7 days. As shown in
Figs. 4 and 5, the work was carried out with care for quality
and appearance.
The 19 day old concrete was about 3 months old when
repaired and it has been continuously in an open, south-facing
exposure, including many cycles of freezing and thawing, ever
since. Parts of the substructure have been subjected also to
brine runoff from deicing salt applied to the highway above each
winter. As a result of adding the new arch, the repaired 20 year
old concrete became one of the inner arches that were protected
to some extent from the weather. The common strengthening of the
old and new arch springings seen in Fig. 5 was later backfilled
and protected from the elements.
36 Young Concrete
Recent Maintenance Work

No appreciable maintenance was required to the bridge until
routine inspections during the late 1970s showed general
deterioration to the deck slabs, supporting beams, bearing seat
areas and parts of the arch rings and spandrel columns. Damage
was much more severe in the, by then, 40 year old original
bridge than in the new widenings, and appeared similar to that
in comparable reinforced concrete bridges exposed to the effects
of freezing and thawing and deicing chemicals for many years.
The upper surface of the deck was repaired, waterproofed and
repaved in 1982. The work on the structure below the deck
surface (where the fire had occurred) was undertaken in the
following year. After removing deteriorated concrete and
cleaning the reinforcing steel, latex-modified shotcrete or
mortar was applied to restore the section and provide at least
25 mm cover to the steel. In addition, significant cracks were
either epoxy grouted or, if subject to movement, sealed.
Surveys of Performance
The condition of the original repairs to the fire damaged
concrete at the time of the 1977-78 inspections may be judged
from Fig. 6. The outer arch ring and column stubs in the fore-
ground, which contained the 19 day old concrete and were at the
heart of the fire, were found not to require further repairs. An
area of the old inner arch and a column (showing between the
second and third columns from the right of the foreground arch),
required repair. This concrete, located down the arch from Core
No. 20 in Fig. 13 of Ref. 1, was the only significant area of
repaired fire damaged concrete scheduled to be repaired again.
This localized damage seemed entirely due to brine leakage
channeled from above. Fig. 7 shows the 1983 repair work in
progress on this part of the old arch. Some areas of concrete
which had been exposed to fire, but not damaged to the point of
repair, did require repair in 1983. An example is the right hand
column of the old inner arch shown in Fig. 6. But since similar
deterioration was prevalent in other old spans that had been
well clear of the fire, delayed effects of the fire are unlikely
to have contributed to the corrosion spalling.
Late in 1984 a further visual inspection, with soundings,
was made of the bridge substructure affected by the fire. No
apparent or latent defects were found. The arch in the fore-
ground of Fig. 8 contained the concrete that was only 19 days
old at the time of the fire. The general integrity of the
concrete arch and the condition of the 25 year old shotcrete
repairs appear to be excellent. A similar observation applies to
the repaired concrete in the exposed abutment which was 6 days
old at the time of the fire. The column in the background of
Fig. 8 is one of those in which the concrete was 20 years old at
the time of the fire but which had been repaired recently
because of corrosion spalling. Unlike the original shotcrete
Smith 37
repairs, the new latex-modified shotcrete, here and elsewhere in

the structure, displays significant surface shrinkage cracking.
The strengthening pedestal to the arch springings could not
be examined because it is buried. No structural cracking or
settlement has been observed in the arches or abutment, so it
seems reasonable to deduce that this remedial measure has also
stood the test of time.
REVIEW OF ASSESSMENT AND REPAIR TECHNIQUES USED
The remaining question is: If the construction fire had
happened today, would anything have been done differently in
light of experience and new knowledge?
The first admission is that the references cited in the
original paper {1), though still useful, need updating.
References 2 through 16 provide more recent sources on material
properties, testing, structural behaviour and fire damage assess-
ment. Ref. 4, 8 and 16 in particular contain bibliographies, or
lead to other references, that may be helpful in a particular
situation.
Comparison of Testing.and Investigating Methods
The investigation of the construction fire damage to the
Bronte Creek bridge in 1959 used a then state-of-the-art panoply
of inspection and testing techniques. These were:
1. Visual inspection to identify and locate spalls, cracks,
exposed reinforcing steel, fused materials and other
evidence of the duration and temperature of the fire and the
severity of damage.
2. Petrographic examination of the concrete, from both loose
and core samples, to identify colour change, dehydration,
microcracking, loss of bond to aggregate or steel,
autogenous healing, etc.
3. Physical testing of concrete cores for residual compressive
strength and of cut samples of reinforcing steel for tensile
strength.
4. Non-destructive testing using ultra-sonic pulse velocity
measurements and Schmidt Rebound Hammer readings.
Fortunately, both fire damaged and unaffected concrete of
the same class and age were available in different parts of the
structure so that direct comparisons were possible. This greatly
increased the confidence in the data, especially that from
measurements of pulse velocity.
38 Young Concrete
In the intervening years many of the procedures outlined

above have been refined in light of experience and new knowledge
(2), (3), (4), (5), (6). Improved testing equipment has made
site work a lot easier; for example, ultra-sonic testers have
been miniaturized and now have easy-to-read digital displays.
Confidence in interpreting data has increased; for example, that
from the two non-destructive tests used (7). In some instances
alternative devices are available; for example, the Windsor
Probe in place of the Schmidt Hammer. But proven new methods
that might have influenced or radically changed the course and
outcome of the 1959 investigation are few.
Some newer, non-destructive tests (8), (9), have potential
for assessing the effects of fire on reinforced concrete. These
tests include; impulse radar and infrared thermography (to
detect discontinuities at right angles to the detector),
acoustic emissions (to detect deformation and stress under
changing load) and radiographic or radiometry methods (to give
an internal picture or measure density). However, all of these
require elaborate equipment, site access may be a problem, and
interpretation of the data is not an established routine. rther
new tests, such as determination of indentation microhardness or
plucking of steel tendons have been used in assessing damage to
prestressed concrete structures only (3), (4), (6).
Twenty-five years ago, the duration and intensity effects
of a fire had to be judged by such minimum temperature
indicators as colour change and quartz spalls in the concrete or
by identifying fused materials. Since data on the residual
properties of concrete and steel after heating to various
elevated temperatures (and quenching) were available; the amount
of core drilling, pulse-velocity work and laboratory testing
might have been reduced, if more reliable means had been
available to establish the temperature regime to which the
concrete, in various locations, had been subjected. Subse-
quently, Harmathy (10) proposed a method based on thermogravi-
metric and dilatometric testing of samples, secured within a day
or two of the fire, to determine a temperature history inwardly
into the concrete. Another approach is the residual thermo-
luminescence of quartz method reported later by Smith and
Placido (11), which they state gives an estimate of thermal
exposure (time/temperature) that can be correlated with residual
concrete compressive strength. 0nly a small sample is required
that can be extracted easily from loose concrete or a core that
is suspect.
Material and Structural Considerations
The striking resistance of immature concrete to fire
damage displayed at the Bronte Creek bridge has been amply
confirmed by other incidents; one of which involved concrete in
a building that was less than 2 days old (12). Yet the paucity
of knowledge as to why this should be so, in face of low
Smith 39
strength and high moisture content in the concrete. still

remains. But then. alternative opinions still remain to be
resolved on the mechanism of spalling. which is such an obvious
feature of fire damaged concrete of any age (13). As already
referenced. significant information. has been published over the
last quarter century about the behaviour of mature concrete and
structural elements and assemblies exposed to high temperature.
including the beneficial effects of continuity and restraint
from cooler parts of the structures. With care. much of this
information can be extended to young concrete.
The structural elements affected by the fire at Bronte
Creek were essentially compression members and hence the
possibility of loss of bond strength between concrete and steel
was not considered as seriously as might have been in tension
members. Furthermore in 1959. unless dehydration at the inter-
face was evident. it was not considered that significant
reduction in bond was 1ike ly ( 1). More recent work ( 14). ( 15)
tends to show this judgement was naive. and that possible
reduction in bond efficiency should always be questioned after
a fire (15). The adequacy of bond with epoxy coated reinforcing
bars. used on constructions that will be exposed to deicing
chemicals. has raised concern in some minds. Fire could be an
added detriment. At the time of writing (June 1985). full scale
fire tests. which should resolve this doubt. are about to start
at the Portland Cement Association laboratories.
The task of analyzing the load carrying capacity of a fire
damaged structure. has become easier in recent years. In
addition to a better understanding of the effects of elevated
temperatures on reinforced and prestressed concrete. improved
structural analysis techniques and the calculating power of the
computer are now on command. Notwithstanding. proof load tests
still may be called for in some critical situations to settle
ultimate doubts.
RevJ!~- of Repair Techniques
The need to repair damage to bridges due to reinforcing
steel corrosion and to rehabilitate other concrete structures.
has led to many new or modified materials or techniques in
recent years. For example. the recent repairs to the Bronte
Creek bridge used latex-modified shotcrete and mortar. The
attributes of latex-modified shotcrete are reviewed in Ref. 17.
Other modifications to shotcrete to improve strength and
adhesion and to reduce shrinkage and rebound. such as the
addition of steel fibres and silica fume. are possible but they
have not yet been widely used. Notwithstanding these develop-
nents. the performance of the less expensive. conventional shot-
crete originally used to repair the fire damage. could hardly be
bettered.
40 Young Concrete
The importance of removing all unsound concrete and

achieving a clean substrate to aid good bond cannot be over-
emphasized. For these purposes, high pressure water jet cutting
in replacement of jack hammering and sandblasting is a promising
development.
The secrets of successful concrete restoration work have
not changed over the years. It is desirable to match the
properties of the new work as closely as possible with those of
the old. It is essential to provide a sufficient depth of low
permeability concrete cover to any reinforcing steel in order to
inhibit corrosion for as long as possible. Meticulous
preparatory work and application of the new concrete or mortar
are mandatory. The results of these prescriptions and wet curing
to develop strength and reduce initial drying shrinkage are
evident in the long term durability of the shotcrete repairs at
Bronte.
Speaking generally, no new breakthrough in materials or
methods, proprietary or generic, has yet appeared to displace
shotcrete {gunite) as the first choice for restoring fire
damaged concrete sections {2), {3). Nor is there any substitute
for good workmanship in sight.
CONCLUDING REMARKS
Clearly there are gaps in knowledge about the fire
resistant properties of concrete at early ages still to be
filled. Improved investigative and repair techniques will always
be welcome. Yet, those faced with the consequences of a construc-
tion fire may take heart from this feedback report in the
validity of past and present practice.
Once again, concrete has demonstrated that it is the
Phoenix of construction materials, even when immature.
Irrespective of the age of the concrete, knowledge and
techniques sufficient to determine the extent of the fire
damage, assess structural adequacy and devise remedial measures
have been proof tested by time and all too frequent use. More
often than not, standard concrete repair methods can restore the
new structure to full serviceability at less cost than replace-
ment. Judged by performance in a highway bridge exposed to a
harsh environment, the life expectancy of such remedial work is
at least 25 years. All provided, of course, that the best of
engineering knowledge and judgement, concrete practice and work-
manship govern.
Smith 41
REFERENCES
1. Smith, Peter, "New Concrete is Not for Burning - Investiga-
tion and Repair of Damage to Concrete Caused by Formwork and
Falsework Fire," Journal ACI, V. 60, No. 11, Nov. 1963,
pp. 1535-1566.
2. "The Assessment of Fire Damaged Concrete Structures and
Repair by Gunite," Technical Report No. 15, The Concrete Society
U.K., 1978.
3. Dore, E., "The Assessment of Fire Damage to Concrete
Structures," Concrete, Sept. 1984, pp. 49-51.
4. Smith, Peter, "Resistance to High Temperatures," STP 169B,
ASTM, 1978, pp. 388-419.
5. Abrams, Melvin S., "Performance of Concrete Structures
Exposed to Fire," Proceedings, "Materials and Processes - In
Service Performance," Ninth National SAMPE Technical Conference,
1977, Reprinted as Research and Development Bulletin 060.01D by
PCA.
6. Gustaferro, Armand, "Experiences from Evaluating Fire Damaged
Concrete Structure," SP 80, ACI, 1983, pp. 269-278.
7. Logothetis, L., and Economou, C., "Influence of High
Temperature on Calibration of Non Destructive Testing of
Concrete," RILEM Bulletin, Materials and Structures, Vol. 14,
No. 70, Jan.-Feb. 1981, pp. 39-43.
8. "In Situ/Nondestructive Testing of Concrete," {edited by
V.M. Malhotra), SP 82, ACI 1984.
9. Clifton, James, "Nondestructive Evaluation in Rehabilitation
and Preservation of Concrete and Masonry Materials," SP 85,
ACI 1985, pp. 19-29.
10. Harmathy, T.Z., "Determining the Temperative History of
Concrete Constructions Following Fire Exposure," Journal ACI,
Vol. 65, No. 11, Nov. 1968, pp. 959-964.
11. Smith, Leslie, and Placido, Francis, "Thermoluminescence: A
Comparison with the Residual Strength of Various Concretes,"
SP 80, ACI 1983, pp. 293-304
12. Suter, G.T., "Can Extremely Green Concrete Withstand Fire
with Minor Damage?," SP 80, ACI 1983, pp. 255-268.
13. Dougill, John W., "Material Dominated Aspects of Design for
Structural Fire Resistance of Concrete Structures," SP 80,
ACI 1983, pp. 151-176.
4·2 Young Concrete
14. Diederichs, U., and Schneider, U., "Bond Strength at High

Temperatures," Magazine of Concrete Research, Vol. 33, No. 115,
June 1981.
15. Royles, R., Morley, P.O., and Khan, M.R., "The Behaviour of
Reinforced Concrete at Elevated Temperatures with Particular
Reference to Bond," Proceedings, Conference at Paisley- Bond in
Concrete, June 1982, pp. 217-228.
16. "Guide for Determining the Fire Endurance of Concrete
Elements," ACI 216 R-81, 1981.
17. Chynoweth, Gary L., "Properties of Latex-Modified Shotcrete
Beneficial to Concrete Repairs," Transportation Research Record
1003, TRB 1984, pp. 42-46.
Smith 4~3
Fig. I.-Scene after the f.ire (1959)
Fig. 2-20-yenr-old concrete arch prior to shotcrete repair (1959)

44 Young Concrete
Fig. 3-Preparatory work on the 19-day-old concrete arch (1959)
Fig. 4-Shotcrete application and finishing in progress (1959)

Smith 45
Fig. 5-Completed structure showing part of the repairs and

strengthening (1959)
Fig, 6-Condition of the repaired structure after 18 years (1977)

4·6 Young Concrete
Fig. 7-Repa.ir work in progress on the old inner arch (1983)
I
Fig, 8-Condition of shotcrete repairs after 25 years exposure (1984)
SP 92-4
Assessment and Repair of Fire-Damaged

Concrete Structures-an Update
by A. K. Tovey
Synopsis: The purpose of this paper is to outline some of the

proposed changes to the Concrete Society Technical Report 15 which
is the most comprehensive UK document dealing with assessment and
repair of fire-damaged concrete structures. The publication is
being updated to take into account changes of practice and methods
of assessment since 1978. Repair techniques other than Gunite are
to be included with particular comment in respect to resin
repairs.
Keywords: buildings; concrete construction; damage; estimates;

evaluation; fires; plastics, polymers, and resins; repairs; shotcrete
47
48 Assessment and Repair
A. K. Tovey is Principal Structural Engineer in the Building

Group, Cement and Concrete Association, England. Member of the
Institution of Structural Engineers and Chartered Institute of
Arbitrators. Involved with various British Standards committees
and the Concrete Society Fire Resistance Committee.
IN'rRODUCTION
A number of publications (1-4) within the UK advise on the

assessment of the effect of a fire on concrete structures. The
most comprehensive, with regard to concrete structures, is the
Concrete Society Technical Report 15 (1 ), "Assessment of
fire-damaged concrete structures and repair by Guni te". That
publication is currently being revised by the Fire Resistance
Committee of the Concrete Society UK and this paper outlines the
major rev1s1ons. It is expected that the revised report will be
available mid-1986.
The principal objectives have been to update the information

given in the light of changes of practice and methods of
assessment since 1978 and to include repair techniques other than
Gunite (sprayed concrete). This has led to the change of title to
"Assessment and repair of fire-damaged concrete structures".
The revised Report is split into three main sections -

1. Assessment, 2. Design, 3. Repair methods. The detailed section
on assessment follows that given previously; a general
qualitative assessment followed by a quantitative investigation of
individual members. A flow chart (Figure 1) has been added to
assist in developing a programme for assessment.
QUALITATIVE ASSESSMENT
The general qualitative assessment method gives a simple way

to indicate the degree of damage and likely repair of individual
members and overall structure. Figure 2 shows part of a first
survey on a four-storey warehouse. The number adjacent to each
member indicates the visual degree of damage according to the
classification system given in Technical Report 15 (1) which is to
be standardized to make it simplier to use (Table 1). A detailed
first assessment of this type enables a very good initial picture
to be drawn of the overall condition of the structure and the
likely repair.
An example is given in Figure 3 which shows the whole of the

first and second floors of the part structure used in Figure 2.
Tovey 49
This clearly shows that the main damage was sustained·in the upper
left-hand quarter of the first floor with the most notably
affected members in the area bounded by columns 5, 8, 19 and 22.
The degree of damage is again clearly shown to reduce away from
this area towards the lower right-hand corner (col. 59). This
corner was the position that the fire started but the major
combustible material was located to the rear of the building
adjoining a single-storey steel framed warehouse which was totally
destroyed. The fire can be seen to have spread to the second
floor but was brought under control before a major fire or
extensive damage had developed. A comprehensive report giving a
good indication of both damage and likely repair can be prepared
in only a few days using the simple visual assessment procedure.
At this stage, any surprising aspect in the performance of a
structure would also be noted.
QUANTITATIVE ASSESSMENT
The detailed quantitative investigation using various tests

is to determine the likely temperatures to which the member has
been subjected and hence the loss, if any, in strength of
reinforcement and concrete. From examination of the building
contents, it is possible to assess the maximum likely fire
compartment temperatures and, from eye witness reports and
evidence of combustible content, to assess the equivalent standard
fire period. This can be compared with the temperature
distribution and strength loss determined from test evidence.
Report 15 ( 1 ) indica ted three main tests which could be carried
out in order to determine strength loss in concrete. These are:
color determination, Schmidt hammer and core testing. The revised
report deals with these and, in addition, gives guidance on the
use and selection of other tests such as ultrasonic pulse
velocity (5), Windsor probe (5), BRE internal fracture (5) and
thermoluminescence tests (6). The information obtained from tests
is then used to assess the strength of the member. An example of
tests carried out on a structure is shown in Figure 4. This
figure shows the position and results of a series of different
tests on a section of wall and floor slab. Each of the results
can be compared together to improve the confidence of the
assessment. The pundit readings on the floor, for example, show
the degree of damage from the maximum occurring at the seat of the
fire (readings around 200) to that where the concrete is undamaged
(readings around 34) and hence indicate the area requiring repair.
The strength reduction curves given in the previous report

were based on research on unstressed concrete (7). It is known
that the strength reduction of concrete unstressed at the time of
heating is greater than for concrete under compression. Since the
majority of structural concrete will be stressed (at least under
dead load) at the time of heating, then some modification to the
previous data is appropriate. The strength loss data has been
simplified to a straight line plot for consistency with other

documents (8,9) (Figure 5).
The color of concrete can change as a result of heating (10)

and therefore may be used to indicate the maximum temperature
attained and the equivalent fire duration. In some cases at above
3000C a pink discoloration may be readily observed. The onset of
noticeable pink discoloration is important since it coincides
approximately with the onset of significant loss of strength due
to heating. Therefore, any pink discolored concrete should be
regarded as being suspect.
Pink discoloration is due to the presence of ferrous salts

in the aggregate and/ or in the sand, and in some cases these are
not present. Therefore, concrete which has not turned pink is not
necessarily undamaged by fire. The color change to pink tends to
be more prominent with siliceous aggregates, Calcareous and
igneous crushed rock aggregates are less susceptible to this
effect.
The residual strength of reinforcement can be judged from

the assessed temperature of the surrounding concrete or, if
spalling has occurred, from the period of exposure and likely fire
compartment temperature.
The effect upon steel during elevated temperatures and after

subsequent cooling has been researched in detail (11,12).
Significant loss of strength may occur while the steel is at high
temperature and this is usually responsible for any excessive
residual deflections. However, recovery of yield strength after
cooling is generally complete for temperatures up to 450°C for
cold worked steel and 600°C for hot rolled steel. Above these
temperatures, there will be a loss in yield strength after
cooling. The actual loss in strength depends upon the heating
condi tiona and type of steel but the simplified values given in
Figure 6 will be sufficient for most purposes. Where this aspect
is critical to the assessment, the matter should be discussed with
the reinforcement manufacturer if known or, alternatively, tests
carried out on samples taken from the member. Loss in ductility
may occur after exposure to particularly high temperatures.
Buckling of reinforcing bars often occurs as a result of
compressive stress induced at high temperatures by restraint
against thermal expansion.
A damage factor may be determined for each member although

normally it will only be necessary to consider Class 3 and Class 4
damaged members in detail. The damage factor is determined from
Figure 5 using the estimated average temperature within the
compression block obtained from temperature profile curves as, for
example, are given in Figure 7.
The general process for this is to record the depth of the

layer which has reached a temperature of 300°C by reference, where
appropriate, to the depth of the pink zone. Estimate the increase
Tovey 51
beyond this depth to reach a zone which has not exceeded 1000C by
reference to Figure 7. The average damage factor may then be
determined for the compression block taking a factor of 1 .0 for
all concrete subjected to temperatures less than 1oooc, and a
factor of 0.85 for concrete within the 3000C to 100°C zone.
It may also be necessary to determine a damage factor to

allow for reduction in bond or anchorage of steel reinforcement.
A number of papers have been written (13,14,15,16). These
indicate that bond is influenced not only by temperature but also
by bar diameter, type of aggregate, section size and the
compressive strength of the concrete. The presence of stirrups or
other confining steel is also likely to influence bond. A
conservative damage factor for bond of 0.7 may be applied to
reinforcement within the 3000C to 1000C zone although a value of
0.8 or better might be considered for lower temperatures, small
diameter bars (12 mm or less), -concrete of lower compressive
strength (25-30 N/mm2), concrete containing calcareous or
lightweight aggregates and for reinforcement contained by
stirrups.
REPAIR TECHNIQUES
In addition to reviewing the sprayed concrete process

(Gunite) which is the usual method of repairing major fire-damaged
concrete structures, the working group has considered and included
in the revision of Report 15 other repair methods such as resins,
polymer modified and cement mortars, plaster, sprayed mineral
preparations, as well as the use of alternative supports.
It is essential that a repair must restore any loss of

strength, maintain durability and fire protection. In situations
where, following a fire, there is still sufficient strength and
cover for durability then a thin hand or spray applied material
could be used to restore a loss in fire protection. Equivalent
thickness of concrete cover for various materials may be found
from manufacturers. Some information is given in reference 21 and
in BS 8110 (22) the revision to CP 110 (8).
Resin Repairs
As a result of the study, it has been necessary to make

particular comment on resin repairs. These are commonly used to
overcome problems of reinforcement corrosion but may not provide
the necessary protection in the event of a subsequent fire.
Resin repairs may consist of a variety of configurations of

patch or infill of Epoxy, Polyester and Acrylic mortar. Resins
are often used for repairs to lightly spalled areas and, though
they may perform quite satisfactorily in normal service, there is
no comprehensive information on the performance of such repairs or
that of the materials when subject to heat or an actual fire test.
What information does exist, including some published

papers (17,18,19), indicates that these materials may soften at
relatively low temperatures (sooc). As a consequence, it is
possible that some resin repairs may not provide adequate fire
protection to the reinforcement and may fail to retain adequacy in
compression zones. Accordingly it is recommended that resin
repairs only be used when either:
(i) performance data can be supplied to show that the

particular formulation has adequate fire resistance and
retains its structural properties under the envisaged fire
condition, or
(ii) the material is adequately fire protected by other

materials and retains its structural properties at the
expected fire temperatures at the relevant depth in the
section, or
(iii) loss of strength or other properties of the material will

not cause unacceptable loss of structural section or fire
resistance.
The designer should refer to the manufacturers' literature

for details of the performance of the various materials available.
A wide variety of materials exists and their specifications are
liable to changes at relatively short notice. In general, the
materials will be capable of providing good bond and compressive
strengths, the flexural and tensile strengths may exceed that of
concrete but the thermal expansion is considerably larger than
concrete and this may be a point to be considered where the
temperature range is large. For further information, the designer
should refer to the Concrete Society Technical Report No. 26,
"Repair of concrete damaged by reinforcement corrosion" (20).
Polymer Modified Mortars
In many instances, hand repairs of small areas can be

effective by use of polymer-modified cementitious mortars. These
repairs will generally be to areas or patches of between 12 mm and
30 mm depth. In particular, styrene butadiene rubber (SBR)
modified mortars appear sui table. There is limited test
information on such mortars but it is expected that they will be
satisfactory in a fire as they should behave as a cementi tious
product. It is to be expected that the use of a small sized
aggregate will improve the performance of mortars compared to
concretes due to a lesser tendency for damage caused by aggregate
splitting in the event of a subsequent fire. In other respects,
these mortars are also described in the Concrete Society Technical
Report No.26, "Repair of concrete damaged by reinforcement
corrosion" (20) to which the designer should refer.
Tovey 53
Cement Mortars
These may be hand applied to damaged areas but great care in

surface preparation is necessary in order to ensure adequate
adhesion. Generally, mortars will be applicable to well-defined
areas placed in layers using good rendering practice, up to a
total 30 mm thickness.
Plaster
This may be readily applied to both plain and roughened

concrete surfaces. It can restore a degree of fire resistance but
will not replace cover requirements for durability.
Sprayed Mineral Preparations
This technique will not assist where strength restoration or

replace cover requirements for durability are required. Where
internal repairs of a minor nature are necessary, these systems
will restore fire resistance and shape to damaged members. For
particulars, the designer should consult specialist contractors.
Alternative Supports
The designer can consider the use of alternative supports

such as further columns or new beams to sub-divide floor spans.
Such schemes may well prove economic as they may allow lesser
restoration to damaged members. Ideally, new supports would be in
reinforced concrete but there is no reason why steel, timber or
masonry should not be used providing the required strength, fire
resistance, durability and appearance are achieved.
ACKNOWLEDGEMENTS
In addition to the author, members of the current working group

revising Report 15 are:
Mr R H Jackson Andrews, Kent & Stone

Mr C D Jones Freeman Fox (Wales) Ltd
Mr E Mellor De Leuw Chadwick 0 hEocha
REFERENCES
1. The Concrete Society., Assessment of fire-damaged concrete

structures and repair by gunite, Report of a Concrete Society
Working Party, London, The Concrete Society, 1978, 28 pp.
2. Green, J.K., Technical Study : reinstatement of concrete

structures after fire. The Architects' Journal. Vol.141, No.2,
14 July 1971, pp.93-99, No.3, 21 July 1971. pp.151-155.
3· Smith, L.M., The assessment of fire damage to concrete

structures, Thesis for PhD degree, Paisley College of Technology,
Paisley, Scotland, September 1983. (Bibliography contains 259
relevant references)
4. Tucker, D.M. and Read, R.E.H., Assessment of fire-damaged

structures, Garston, Building Research Establishment, November
1981, 4 pp, IP 24/81.
5. Keiller, A.P., A preliminary investigation of test methods

for assessment of strength of in situ concrete, Wexham Springs,
Cement and Concrete Association, 1981, 36 pp, Technical
Report 42.551.
6. Placido, F., Thermoluminescence test for fire-damaged

concrete, Magazine of Concrete Research, Vol.32, No.111,
June 1980, pp.112-116.
7. Malhotra, H.L., The effect of temperature on the compressive

strength of concrete, Magazine of Concrete Research, Vol.8, No.23,
August 1956, pp.85-94.
8. British Standards Institution, CP 110 : Part 1 : 1972, The

structural use of concrete, Part 1 : Design, materials and
workmanship, London, 156 pp.
9. Design and detailing of concrete structures for fire

resistance, Interim guidance by a Joint Committee of the
Institution of Structural Engineers and The Concrete Society,
London, The Institution of Structural Engineers, 1978, 59 pp.
10. Bessey, G.E., Investigations on building fires, Part 2 : The

visible changes in concrete or mortar exposed to high
temperatures, London, H.M. Stationery Office, 1950, pp.6-18,
National Building Studies Technical Paper No.4.
11. Stevens, R.F., Contribution to discussion on: Steel

reinforcement, by R.I. Lancaster, Structural Concrete, Vol.3,
No.4, July/August 1966, pp.184-185.
12. Holmes, M., Anchor, R.D., Cook, G.M.E., and Crook, R.N., The
effects of elevated temperature on the strength properties of
reinforcing and prestressing steels, The Structural Engineer,
Vol.60B, No.1, March 1982, pp.7-13.
13. Morley, P. D., and Royles, R., The influence of high

temperatures on the bond in reinforced concrete. Fire Safety
Journal, 2(1979/80), pp.243-255.
14. Sager, H., and Rostasy, F.S., The effect of elevated

temperatures on the bond behaviour of embedded reinforcing bars,
Bond in Concrete : Proceedings of the International Conference,
Paisley, 14-16 June 1982, London, Applied Science Publishers,
1982, pp.206-216.
'lbvey 55
15. Royles, R., Morley, P.D., and Khan, M.R., The behaviour
reinforced concrete at elevated temperatures with particular
reference to bond strength, Bond in Concrete : Proceedings of the
International Conference, Paisley, 14-16 June 1982, London,
Applied Science Publishers, 1982, pp.217-228.
16. Morley, P. D., and Royles, R., Response of the bond in

reinforced concrete to high temperatures, Mazagine of Concrete
Research, Vol.35, No.123, June 1983, pp.67-74.
17. Levitt, M., The fire resistance of resin jointed concrete,

Proceedings of Conference on Plastics in Building Structures
(1965), Plastics Institute, Pergamon Press, 1966, Paper 12,
pp.77-81.
18. Plecnik, J,M., Bresler, R., Chjan, H.M., Pham, M., and Choa,
J,, Epoxy repaired concrete walls under fire exposure, Proceedings
of the American Society of Civil Engineers, Structural Division,
Vol.108, No.ST8, August 1982, pp.1894-1908.
19. Plecnik, J,M., Bresler, B., Cunningham, J.D., and !ding, R.,
Temperature effects on epoxy adhesives, Proceedings of the
American Society of Civil Engineers, Structural Division, Vol.106,
No.ST1, January 1980, pp.99-113.
20. The Concrete Society., Repair of concrete damaged by

reinforcement corrosion, London, 1984, Technical Report No.26.
21. FIP/CEB Report on methods of assessment of the fire

resistance of concrete structural members, Wexham Springs, Slough,
Federation Internationale de la Precontrainte, 1978.
22. British Standards Institution., BS 8110:1985, The structural

use of concrete, Part 2: Recommendations for use in special
circumstances, London.
TABLE 1 - VISUAL DAMAGE CLASSIFICATION FOR REINFORCED CONCRETE ELEMENTS c.n
0\
Surface Appearance Structural Appearance
Class Element Plaster/ Color Crazing Spalling Condition of Cracks Deflection
>
Finish
I main reinforcement 1:rJ
1:rJ
~
1:rJ
0 Any Unaffected or beyond extent of fire 1:rJ
1 Column some normal slight minor none exposed none none

3~
Floor
peeling
.. .. .. .. .. .. .. .. =
......
~
Beam .. .. .. minor very minor exposure . .. =

0...
~
"'0
2 Column substantial
loss
..
pink
..
noticeable
..
localized
to corners
up to 257..
none buckled
none
..
none
..
=·
'"l
Floor localized up to 107..

to patches all adhering
Beam .. .. .. localized up to 257.. .. ..

to corners none buckled
minor to
soffit
3 Column total buff/ extensive considerable up to 507.. minor none

loss friable to corners not more than
one buckled
Floor .. .. .. considerable up to 20%. small not significant

to soffit generally adhering
Beam .. .. .. considerable up to SO%. . .. .

to corners not more than
sides one buckled
soffit
----
'lbvey 57
1-
z
w
>
w
CARRY OUT DAMAGE

CLASSIFICATION SUrlVEY u.
0
1-
z
w
::i
if)
if)
w
if)
if)
<(
NO
PRELIMINAFW
DESIGN FOf1
MAJOH
YES z
~
if)
w
D
NO
-'
<(
>
0
a:
Q_
Q.
<(
1-
0
<(
a:
1-
z
0
0
Fig, !-Assessment procedure for fire damaged structures

c.n
co
2 3 3 4 3 5 3 6 3 7 3
)
1
I 2
1 3
3
3 2
2
1 2 2
3
3
3 3
3
3 3
3
3
3 >
00
00
1 2 1 1 3 2 3 3 3 3 3 2
1L 2 3 1 3 3 3 3 ('l)
00
15
1
I 2
2
16
2
3-
3
17
2
1
1
18 2 3
2
19 3 3
3
20 3 3
3
21 2
3
3
-
00
a
('l)
2 3 1 1 3 3 2 2 :::::
3 3 1 2 2 3 .......
2 3 2 2 1 1 1 3 2 2 1
p;
3 1 2
3 3 1 2 1 2 :::::
3
a..
30 31 2 3 32 1 33 1 1 34 1 35 1 36 1
~
3 1 2 2 2
)
2 1 1 1 1 1 '
2 2/_ 3 1 2
1
1
1
1
1
1
1
1
1
1 '"i:l
2 2:.
L 1
1
2
1 1
1
1 1
1
1 1
1
1 1
1
1 0
0 ""
°
-
44 ot,5 1 46 1 47 1 1 48 1 1 49 1 1 50 1 1 51 0
. 1
)
0
Ref. Classification factors - Table 1
E
4
Col
---r
'
'
.
I
Member Ref
Fig. 2--Member classification (Grd - lst)

Tovey 59
Ref. repair Classification

GROUND -1st i·: :·:-. ·: < >J Class 0 Deco rat ion
Wff~ 1 Superficial
0 Most notably affected members &~~ 2 General repair
~ 3 Principal repair
1 2 3 4 5 6 7 8 9 10 11 12 13 14
r---r ---¥' -¥ -P- -r ~ -r--f· 1i .l';i w- -iH.-01;;,'.97,

rjT·-r·T·T·T·T·T·T:T.-.T:·r··l7\l
'
~--f1·-P~-f~~-T-· ~-
' . ' ' .
3
•. . . . . . . . . . . . '·.· .· ......1
F\>:f.~:~i.,.:/t~<·;.-:J~->~->:k~- 43
~~-f-~?_ha__19_h9_h1~2___li;·.:·:.1~·~.··. :·.:;r~~\··/:l~6~··~~/~~-~j59
Fig. 3-Tnit:ial classification of repair

134 133 132 202 126 124

• • 46• • • 11~11
X
ELEVATION
• Pundit
0 BRE
I:B Windsor
x Schmidt
I I
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - _ _ _ __J ~
"L..r--------- ----,_J----- -------- -,__r----------- ---.,_,J
X 48 52' 48 X X 47
35 38 52 74 67 85 225 153 216 44
35 37 40 45 43 45 77 123 104 38
35 36 37 38 55 52 47 36
36 34 35 36 35 35
33 33 34
PLAN
_ _ _ _ _ _ _ _ _ _ _ _ _ __r-,__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ......r-,
----------------------------------------·r
!I
Fig. 4-Tests to wall and 1~affle

1·2r-------------------------------------------;
Typical hot- rolled
1·0 ----.,
O·B
0·6
0·41-- REPORT 15
0·2
0 200 400 600 800 1000

0 200 400 600 BOO 1000 Temp °C
Temp °C
'"~
1·2
'::rt
Typical hot- rolled
O·B
0·6
\
\
\
\
REVISION ~"" ~.. _.,,..
\
0·4 \ REVISION
\
\
\
0·2 \ 0·2
Normal Pmk White grey ,Buff
0 200 400 600 800 1000

0 200 400 600 BOO 1000
Temp °C
o-3
Temp °C
<
('!)
'<
Fig. 5--Compressive strength on cooling Fig. 6--Yield strength of steels at room temperature
after heating to an elevated temperature 0\
.......
CcJ T ..... ptnh.....iotribvtOooo;., • .,....... .,,_,..,, M ............
0"1
(a) Slabs exposeo ~o hea: oo one face liOO,--,-,---,---,-,,.-,--r-r-r-o-T,;-T-r-l ~
~··· :·.::~
eoo
( ) >
=
000
'.....
::· (fJ
(fJ
('!)
I (fJ
- I OC;
JOO
: :1
Joel
i\imtiLJ I l f"l =t'- 1--.....;,_ "'"'""'I == =r----i- --+==-:....___
1 ~~ ~
i i E
3('l)
:::;
,..-
,.
~
:::;
""""--- I I
c..
2oo! 1 ! ! f"'-- ~ \ ::=t-- --L: :=--r=--9 :zn~
\OQ I .::lVU :-:1m Ul"t uv:.K '>"YI;..;=:; Kl= I =+==-=+- -LJ ~
;:::::;
('l)
v
7~ 100
O•srane~ !rQ<II
1::0::.
~~oos~d T.oee o! si•C·"'"'
(OJ 'hm!'<'falur~ <lo"'robuhun on a d~n·~ concrorT~ rib<>' ~><'•m
=·
>-;
(b) 380 mm columns
~(':''~ Sur!ac~s ~•Po

\0 l1r~
j s~d
300 1 ! i I 1- ... 1'-L r=+ ....d: I T='. :-· I

2001 :;--- -L::;:---.: --L:l ~ 1
,I·:~ 'ool I
I I II -r=r----r1- II IY:r.!
I
o n m n • m ~ •
Dostanc~ !rom surlaceot CCIUMI'I •""" 0 I
o 10 20 Jo .:.o so Go 10 so so :oo 110 120 tJO t.:.o 1so
Monomum distonc• !rom sur!oc• •~<POS•d to lir• (mm}
Fig. 7-Temperature gradients at various times from start of BS furnace test

SP 92-5
Nondestructive Testing Methods

for Evaluating Damage and Repair
of Concrete Exposed to Fire
by Richard A. Muenow and Melvin S. Abrams
Synopsis: Evaluation of fire-damaged concrete has been made using

field-applicable test techniques which are valuable but limited.
With the advent of reliable non-destructive testing methods, a
much more comprehensive assessment of damage is available. The
objective of this paper is to present. information on Pulse-Echo,
magnetic and microwave testing techniques.
Pulse-Echo non-destructive testing can be conducted on insitu

concrete for:
measuring thickness of damaged concrete
assessing bond loss between concrete and reinforcing steel
detecting internal cracks and their orientation
determining degree and extent of cement matrix microcracking
estimating insitu concrete compressive strength
Accompanying the Pulse-Echo method are steel reinforcement loca-

tion techniques that involve microwave and magnetic principles.
These tests can determine remaining concrete cover and additional
data for engineering the required repairs.
Of equal importance to the assessment of damage is the evaluation

of the performed repairs to determine that the repairs satisfy all
the structural requirements of the original specifications.
Keywords: bond (concrete to reinforcement); concretes; damage;

evaluation; fires; impact hammer tests; nondestructive tests;
repairs; ultrasonic tests
63
64 Nondestructive Testing
Richard A. Muenow has specialized in the field of non-destructive

testing. As part of his experience, structures exposed to fire,
including nuclear containments and high-rise buildings have been
evaluated for structural damage and repaired by the use of non-
destructive testing techniques.
Melvin S. Abrams is now an Associate in the firm of Muenow and

Associates, a non-destructive testing company. His experience
includes major research projects to develop information and data
on design of more fire-resistive concrete structures, and inves-
tigation of structures damaged by fire.
INTRODUCTION
Fires may occur in concrete structures during construction as well

as after the structures are in use. Many construction fires are
severe and cause extensive visible damage due to large fuel loads,
long duration, high ventilation, high-water content and low
strength of "green" concrete, and the impingement on hot concrete
of water under high pressure from fire hoses.
The type of damage that often meets the eye after a structure is
exposed to fire is shown in Figures 1 and 2. Figure 1 shows a
large wall after exposure to a severe hydrocarbon fire. The con-
crete is highly discolored, spalling is extensive and deep, and
some of the steel is bent. Extensive spalling of concrete is
evident on the wall shown in Figure 2. Much of the horizontal
and vertical reinforcement is exposed, and some of the steel is
damaged.
Although damage to the concrete as the result of a severe fire

appears extensive, as shown in the photographs, the real extent
of the damage and the feasibility of repair is not known until a
comprehensive damage evaluation is conducted. Yet, a decision to
demolish fire-damaged concrete members is often made by regula-
tory officials and by engineers only on an emotional response to
the initial visual appearance of the structure and without benefit
of a real damage evaluation. Such decisions are usually premature.
Action taken as a result of a pren1ature decision delays construc-
tion or causes business interruption for long periods of time.
The result of these delays leads to drastic economical consequen-
ces that could have been avoided if the structure had been pro-
perly evaluated for possible repairs. The benefits of a compre-
hensive fire-damage evaluation are readily apparent.
Muenow and Abrams 65
TRADITIONAL FIRE-DAMAGE EVALUATION TECHNIQUES
For many years several techniques have been used to evaluate dama-
ge to concrete exposed to fire. lvhile valuable in certain res-
pects, most of these techniques are often costly and time consu-
ming. They are also limited in that they are more qualitative
rather than quantitative. They do not provide specific informa-
tion as to location, type, and degree of damage. Without such
information, it is impossible for even highly-trained structural
engineers to determine the effects of damage on structural capa-
city of the member or on the structural integrity of a building.
Consequently, the remedial work required to restore the building
to its original capability is inconclusive.
FollO\qing is a brief discussion of conunon techniques currently in

use in evaluating fire damage.
Visual Inspection
The value of this method depends chiefly on the experience of

the investigator and his ability to conduct an intelligent, rather
than emotional, evaluation. Often it is possible to decide which
members should be replaced, rather than repaired. For instance,
excessive deflection of slabs, of primary members, and large ex-
tensive cracks in structurally sensitive areas usually indicate
the need for removal and replacement of the concrete member. In
most cases, however, absence of excessive deflection and cracking
indicates that the concrete can be repaired. Rarely do damaged
concrete members need to be removed and replaced.
The visual inspection, by itself, rarely- if - ever defines

damage to a useful level. The most important function of the
investigator is to recommend ,qJwt additional evaluation techniques
must be used to determine the type, location, and extent of damage
so that proper repair techniques can be recommended.
Auditory Methods
These methods employ a hammer or archaeological pick, or

employ a metal chain to rap on the damaged concrete members or to
be drug across flat concrete slabs. At best, an experienced in-
vestigator can determine if the damage is cosmetic or structural.
These methods do not pinpoint location, type, or degree. As
before, areas that require additional testing can be determined.
Schmidt Impact Hammer
Use of a Schmidt Impact Hammer produces limited information

regarding location, type and extent of fire damage to concrete.
In order to use this method, a reading on the impact hammer scale
must be obtained on undamaged concrete. Lower readings on the

scale obtained on concrete damaged by fire indicate the degree of
damage. The greater the difference in scale readings, the greater
the damage.
If a calibration curve that relates scale reading to compres-

sive strength is available for the type of concrete being tested,
estimates of the compressive strength of the damaged concrete can
be obtained. It is evident that this method is highly qualita-
tive.
Damage Evaluation by Concrete Discoloration
Damage to concrete can be estimated by noting the discolora-

tion resulting from exposure to elevated temperatures (1). Dam-
aged concrete that is tinted red indicates exposure to tempera-
tures of 316 to 593C (600 to llOOF). Grey and buff discolorations
occur at temperatures above 593C (llOOF).
During the last fifteen to twenty years, a great deal of in-

formation has been developed and published concerning the beha-
vior of concrete exposed to fire (2) and the effect of elevated
temperatures on the strength properties of concrete (3, 4, 5).
Hith this information, and a knmo1ledge of fire-exposure duration,
it is possible for an investigator to make a reasonable estiamte
of the type and extent of damage. This approach might be one of
the more-useful methods of evaluating concrete damaged by fire.
Measurement of Pulse-Veloc:l.ty with Soniscope
A method often used to assess damage of concrete exposed to

fire depends on information obtained from pulse-velocity measure-
ments through the concrete. Instrumentation to accomplish such
measurements has been developed over the last twenty five years.
The method is a pass-through technique requiring a signal-
transmitting transducer on one si.de of the test member, and a
signal-recei.vi.ng transducer directly opposi.te on the other side of
the test member. Also required is calibration information on the
relationshi.p of the pulse-velocity to concrete compressive
strength and general condition of the concrete under investi.ga-
tion.
This method has several drawbacks that limit the usefulness

of informati.on obtained. Alignment of the transducers is critical
and time consuming. \Vhen the pulse-veloci.ty is lower than that of
undamaged concrete, it is not known if the reduction is due to
deterioration of the cement matrix, a single or few large crack(s),
microcracking, or a combination of these phenomena. Obviously,
this method is limited because it does not ind:l.cate location,
type, and degree of damage.
Core Sampling
Of the commonly-used methods under discussion for evaluating

damage to concrete due to fire, core sampling yields the most
reliable information as to location, type and degree of damage.
Even this technique, however, has serious drawbacks. IVithout some
parallel testing, it is not simple to determine the best locations
from which to obtain core samples. In a structure where exten-
sive damage has occured, a large number of cores must be produced.
All samples must be examined by a qualified petrographer to eval-
uate the extent of the physico-chemical changes, bond loss bet\veen
cement matrix and reinforcing steel, and cracking. The operations
of obtaining a sufficient number of core samples and petrographic
analyses are expensive and time consuming. Thus, even though this
method does produce more information on location, type and degree
of fire damage, economic factors of time and money limit its use.
PULSE-ECHO, HAGNETIC, AND HICRO\VAVE TESTING TECHNIQUES
During the past two decades, sophisticated non-destructive testing

methods have been developed for locating and identifying specific
damage to the concrete as a result of exposure to fire. Type,
degree, and location of damage can be determined rapidly using the
Pulse-Echo technique. Remaining undamaged concrete cover over
reinforcing steel can be determined rapidly using magnetic and
microwave testing techniques. The use of these non-destructive
techniques provide a rapid, accurate, and comprehensive damage
survey which is required to evaluate the integrity of the insitu
structure. Analysis of the information obtained indicates the
type of repairs required to restore the structure to the original
specifications.
Pulse-Echo Non-Destructive Testing
This procedure is a systematic method for performing non-

destructive volumetric examination of in-place concrete. The test
method is based on the laws of wave propagation through an elastic
medium and is performed electronically with equipment which util-
izes the theory of random signal analysis. The location, type,
and extent of internal discontinuities in existing concrete can be
analytically determined by interpreting the microseismic test
results.
The non-destructive testing method (NDT) to be discussed is a

Pulse-Echo microseismic technique. A device is used to transmit
a mechanical low-frequency stress wave into the concrete from one
surface of the concrete member as shown in Figure 3.
lo/ave energy reflections occur at discontinuities within the

member and at the front and back surfaces of the member. All re-
flections are measured with a receiving transducer on the same
concrete surface as the transmitter. Signals are recorded by an
68 Nondestructive 'Jesting
oscilloscope. Data recorded include time between reflections,

frequency of reflections, attenuation, absorption, and decay time
for the wave energy.
Conditions and properties of concrete such as compressive

strength, modulus of elasticity, Poisson's ratio, voiding, honey-
combing, micro- and macro-cracking, loss of cement matrix, and
loss of bond to aggregate can be located using NDT data. The
structural integrity of the concrete member can be evaluated us-
ing NDT results.
The equipment used for the procedure has been developed in

recent years for the unique purpose of inspecting concrete from
one side. The equipment is not wholly available commercially.
Most other concrete non-destructive equipment uses pass-through
transmission techniques that require access to both sides of the
test member. This NDT equipment and procedures are significantly
different from those used in NDT of metals.
To accomplish a Pulse-Echo NDT program and analysis, the

following equipment is used: oscilloscope with extensive modifi-
cations to the electronics; piezoelectric transducer with wave
shaping, filtering, and rejection capabilities; Schmidt Impact
Ilanuner; coaxial cables; camera and assorted adaptors.
The test procedure is initiated by calibrating the electron-

ics according to the manufacturer's instructions. An undamaged
test area is selected in the concrete structure for base-line
tests. The piezoelectric transducer is placed on the concrete
surface at the test location. The Schmidt hammer is placed 1 to 2
inches from the transducer and activated, Data will appear on the
Cathode Ray Tube (CRT) of the oscilloscope. The CRT image is
photographed and then erased from the oscilloscope. The test
procedure is repeated at all locations designated in the pre-
determined test grid.
Data from the microseismic tests is interpreted utilizing

laws of wave reflection. The sketches designated "A" and "B" in
Figure 3, illustrate oscilloscope signals obtained from two spe-
cimens of concrete, In sketches "A" and "B", the signal feature
noted as (a) is an indication of the \vave generated at the sur-
face on which the receiver (transducer) and impactor (mechanical
wave producer) are placed, The signal feature noted as (b) is an
indication of the wave reflection from the opposite surface,
Sketch "A" represents a signal from a concrete specimen with

no internal discontinuities. Sketch "B" represents a signal from
a concrete specimen \vith an internal discontinuity at approximate-
ly one-half the distance between the front and back surfaces.
This signal feature, noted as "reflector" is an indication of the
wave reflection from the discontinuity.
Figure 4 shows a floor of a fire-damaged structure being

tested using the non-destructive Pulse-Echo method. A mechanical
low-frequency stress wave is transmitted into the concrete using
an impact hammer. The path of the 'qave through the concrete is
picked up by the transducer and is displayed on an oscilloscope.
Typical patterns obtained from tests on concrete damaged by

fire are shown in Figures 5 through 15. In all of the photographs,
the left-hand reflection is the front or entry surface of the test
member. Figure 5 represents solid, undamaged concrete. Note the
absence of reflections between the front and back surfaces of the
test member. Figure 7 shows cracking and micro-cracking in the
top 2.54 em (1 in.) of a 19-cm (7~-in.) thick floor slab with
additional slight micro-cracking down to 5.72 em (2~ in.) from
the top. The remainder of the concrete was not damaged by the
fire. Figure 12 is for a 106.68-cm (42-in.) thick wall and indi-
cates severe cracking and delaminating for a depth of 22.86 em
(9 in.) from the front face, and a depth of 12.7 em (5 in.) from
the back face plus moderate micro-cracking throughout the thick-
ness. Figure 13 shows a small crack 8.89 em (3~ in.) from the
back face of an 88.9-cm (35-in.) thick wall. Figure 14 shows
multiple cracks for a depth of 10.16 em (4 in.) from the front
and back faces, and a crack about 33.02 em (13 in.) from the back
face. Figure 15 shmqs cracking throughout an 88.9 -em (35 -in.)
thick 'qall.
It can be readily seen that type and extent of damage to con-

crete can rapidly and accurately be identified and located using
the Pulse-Echo system. It should also be noted that the NDT can
be conducted from a conunon surface.
Once the damage to the concrete is defined, removal and re-

placement of the damaged material can be initiated. The proced-
ures outlined in ACI 506 (6) often are follmqed. Since the extent
of damaged material has been accurately defined, chipping opera-
tions to remove this concrete proceed rapidly with a minimum re-
moval of solid undamaged concrete as shown in Figures 16 to 19.
Figure 16 shows a section of a floor from which the damaged con-
crete was removed. Pulse-Echo testing of this area verified that
all damaged concrete was removed. Columns with damaged material
removed are shown in Figure 18. Hith the information obtained
using the Pulse-Echo testing technique, it 'qas possible to remove
only damaged material. Pulse-Echo test results indicate that
damage exists to a depth of 25.4 em (10 in.) from the surfaces of
a member shown in Figure 19. As seen in Figure 19, this thickness
of damaged concrete was removed before reaching solid undamaged
material.
Another important function of the Pulse-Echo technique is the

evaluation of repairs to determine the bonding of new concrete to
old and new concrete to reinforcing steel. Figures 20 (a) and
21 (a) show delaminations due to shrinkage at both faces of a wall
repaired by shot-creting. The surfaces of the wall at the loca-
tion of the delamination were epoxy injected. Figures 20 (b) and
21 (b)show that the delaminations were filled, and that the con-
crete lvas solid throughout the wall. Patterns shown in Figures
20 (b) and 21 (b) are identical to that shown in Figure 5 which is
typical for solid, undamaged material.
As indicated earlier, compressive strength of concrete, modu-

lus of elasticity and Poisson's ratio can be estimated or calcu-
lated from information obtained with the Pulse-Echo technique.
Compressive strength is accurately estimated from relationships
established between the velocity of the stress wave through the
concrete and concrete compressive strength. The vast amount of
Pulse-Echo testing over a twenty-year period has resulted in velo-
city-strength relationships for concretes made with many types of
aggregates. To calculate the elastic constants, velocity measure-
ments obtained lYith the Pulse-Echo technique and concrete density
information are used in well-known formulas.
Magnetic and Nicrowave (Radar) Test Techniques
When damage due to fire does not extend down to the reinfor-
cing steel, it is necessary to determine the depth of undamaged
concrete cover remaining over the steel. This information is re-
quired to restore concrete cover thickness to original specifica-
tions. Magnetic and Microwave non-destructive testing techniques
are now available to obtain this information rapidly and accurate-
ly.
Instrumentation used for locating reinforcing steel and for

measuring thickness of concrete cover is a JANES "R" NETER. This
instrument is referred to as a pachometer and is an industrial
standard test method. As described in the JAMES INSTRUMENTS,
INCORPORATED Field Manual 4956(7), this equipment is a rugged and
precise instrument for locating and measuring depth of cover of
reinforcing bars in reinforced concrete. The R-Meter measures
disturbances caused by the presence of an external magnetic mat-
erial being introduced into a known magnetic field generated by
the R-Meter. Magnitude of this disturbance is indicated by the
deflection of a needle on the meter which is calibrated to read
directly in bar sizes versus distance from the probe.
Another ND'l' method for determini.ng concrete cover thickness

is the radar or n1icrowave technique. Radar utilizes electromag-
netic wave energy or radio waves to detect and locate internal
discontinuities by reflection in non-metallic construction mater-
ials. Ground radar or subsurface radar has been used for years
to do geologic mapping and shallow subsurface foundation mapping.
Only recently has the construction industry recognized radar as a
potential NDT technique for one important inspection; namely,
detecting, locating, and sizing reinforcing steel bars and other
concrete embedments.
To detect an internal discontinuity an internal change must

occur. Hith radar, the changes are in dialectric constants which
relate to electrical characteristics rather than distinctive
density material changes. For example, the reason that radar lo-
cates reinforcing steel is because there is an 80 percent change
in dialectric characteristics, 1~hereas radar will have difficul-
ties recognizing air-filled voids in concrete since the dialectric
change is minor.
SUMMARY
Several commonly-used methods for determining damage to concrete

exposed to fire and to elevated temperatures are discussed. \olhile
these methods are useful, they generally yield information of a
qualitative nature. Use of this information is severely limited
in assessing the structural impact of the damage because it does
not define the type, specific locations, and degree of damage.
In contrast, a non-destructive Pulse-Echo technique for detecting

damage to concrete exposed to fire is no1~ available. This method
rapidly and accurately detects the type, extent, and specific lo-
cation of damage. This information is precisely what is required
by a qualified engineer to assess the effect of fire damage on
structural capacity. The information is also useful in determin-
ing the type of repair required. The Pulse-Echo technique also
serves the indispensable function of evaluating the performed re-
pairs. Bonding of new concrete to parent material and to rein-
forcing steel can be verified.
\~hendamage does not extend down to the reinforcing steel, magnet-

ic and radar microwave techniques are now available to determine
concrete cover thickness remaining over the steel. These non-
destructive test methods, along 1~ith the Pulse-Echo non-destruc-
tive technique, provide a comprehensive, rapid, economical, and
accurate system for performing a damage survey.
REFERENCES
1. Bessy, G. E., "Investigation of Building Fires, Part 2: The

Visible Changes in Concrete or Mortar Exposed to High Temper-
atures", National Building Studies Technical Paper 4, 1950,
Her Majesty's Stationery Office, London, England.
2. Abrams, H. S., "Behavior of Inorganic Materials in Fire", PCA

Bulletin RD067-01H, Portland Cement Association, Skokie,
Illinois.
3. Abrams, M. S., "Compressive Strength of Concrete at Tempera-

tures to 1600F", PCA Bulletin RD016-0TT, Portland Cement
Association, Skokie, Illinois.
72 Nondestructive 'H:~sting
4. Schneider, U. "Zur Kinetic Festigkeitsmindernder Reaktionen in

Normalbetonen bei Ilohen Temperatures", Thesis, T. U. Braun-
schweig, 1973.
5. Anderberg, Y., and Thelandersson, S., Division of Structural

Mechanics and Concrete Cons true t:l.on, Bulletin 51,, Lund Instit-
ute of Technology, Lund, Sweden, 1976.
6. ACI 506, "Recommended Practice for Shot-Creting", American

Concrete Institute, Detroit, Michigan.
7. James Instruments, Incorporated, "Field Manual '•956", Chicago,

Illinois.
'>, • j I
,f\,'
,.,
'
Fig. 1-Damage to concrete wall due to hydrocarbon fire
F:i.g, 2-Damage to heavily reinforced wall from fire

74 Nondestructive 'lesting
SKETCH A
(a ) ( b)
I
I OSCILLOSCOPE SIGNAL
I
I I
l-< distance D >i

II
.,. .• b: •.• J:l.·~ . ".d"\. .;:- •
•/~------~--~~~~~
)/,
a. · · 4'' · · k·. ~·
A • • • B
. ;·
. "::\ . . . ~·.A.;
., ) £'. .... .a 1 .. . .· . . . A • • . f,l . • .
IMPACTOR
CONCRETE SPECIMEN
SKETCH 8
REFLECTOR
¥
(a) (b)
OSCILLOSCOPE SIGNAL
,{:'" .
IMPACTOR
CONCRETE SPECIMEN 2
Fig. 3-Pulse-echo diagrams

Fig. 4-Pulse-echo testing of fire damaged concrete floor
Fig. 5-Solid concrete

Fig. 6-Very slight microcracking at top of floor
Fig. 7-Cracking and m:icrocracking in a 19.0 em (7! in.) floor

Fig. 8-Cracking and microcracking through total thickness of floor
Fig, 9-Delam:inations 0 to 15.24 em (0 to 6 in.) from front face

of wall
78 Nondestructive 1esting
Fig. 10-Microcracking for 10.16 em (4 in.) depth from both faces

of wall
Fig. 11-Microcracking for depth of 5.08 to 7.62 em (2 to 3 in.)

from both surfaces of girder
Fig. 12-Severe cracking, delamination or moderate microcracking

throughout
Fig. 13-Small crack near back face of wall

(I ,, ' ·~··
... ,
II. I
t- ,
J I
•:. I,
~liu
n,.....___
',.
...
.. ,. ..• ~
,~·.'.
r ~._,
·•.·.. ·..
Fig. 14-Multiple cracks near both faces, and internal crack in

floor
Fig. IS-Cracking throughout total thickness of wall

"C1
(!)
:>
0
8
(!)
H
Ul
(lj
:.:
(!)
.w
(!)
H
u
~
0
u
'0
(!)
00
~
(lj
"C1
.c:
u
•rl
.c:
:.:
8
0
H
4-;
(lj
(!)
H
(lj
H
0
0
rl
~
~
I
rl
00
•rl
~
Fig. 17-Precise removal of damaged concrete from severely damaged

area
Fig. 18-Removal of fire damaged concrete columns us:i.ng pulse-echo

information
84· Nondestructive Testing
,I
Q)
-1-'
Q)
1-<
u
10:
0
u
'0
Q)
oo
<U
e<U
'0
4-<
0
1-<
Q)
:>.
<U
.--1
10:
•.-I
.
0
......
~
u
e
"'".
If)
N
4-<
0
.--i
<U
>
0
eQ)
~
0\
I
.--<
00
.
•rl
"'"'
Fig, 20(a)-Delaminations at both faces of wall
Fig. 20(b)-Solid concrete condition after epoxy injection repair

Fig. 2l(a)-Sharp delaminations at surfaces of wall
Fig. 2l(b)-Solid condition after epoxy repair

SP 92-6
Analysis of the Developing Damage

of Concrete Due to Heating and Cooling
by K. Kordina, W. Wydra, and C. Ehm
Synopsis: Experiments were carried out on normal concrete speci-

mens with measurements of the total deformation and of the acou-
stic emission in order to investigate the correlation between
crack development and deformation behaviour of concrete during
unsteady temperature conditions.
For this purpose specimens were loaded with different stress le-
vels and heated up to maximum temperatures between 150 °C and
750 °C. After a holding period the specimens were cooled under
load.
The creep deformations showed strongly increasing rates at about

450 °C during heating and at the beginning of the cooling phase.
In the heating phase the activity of acoustic emissions increased

considerably. During the holding period no remarkable activity
could be observed. A new increase of the acoustic activity could
be clearly observed at the beginning of the cooling phase.
In a normal concrete specimen, the main activity in the damaging

process occurs during unsteady temperature conditions, and is
dependent on the maximum temperature reached.
The comparison between acoustic emission activity and creep de-

formations shows that the acoustic emission activities are high at
the same time that creep deformations are high.
Keywords: concretes; cooling; crack propagation; creep properties;

damage; deformation; heating; high temperature; stresses; ultrasonic
tests
87
88 Heating and Cooling
ACI member Karl Kordina is professor in the field of reinforced

and prestressed concrete structures at the Technical University
Braunschweig, West Germany. His main interests are different items
in the field of structural engineering, especially the behaviour
of structural elements in fire.
Wlodzimierz Wydra is docent at the university of Wroclaw/Poland,

where he received his PhD in structural engineering. During a
scientific stay at the Technical University of Braunschweig, West
Germany, he was engaged in the theoretical treatment of the re-
sults of investigations concerning the behaviour of concrete at
elevated temperatures.
Christian Ehm is a physicist in the field of material sciences.

He received his PhD at the Technical University of Braunschweig,
West Germany. He is working mainly on the behaviour and proper-
ties of concrete at elevated temperatures.
INTRODUCTION
Problems of crack formation, crack propagation and crack

detection in structural concrete members have long been centre of
interest of many experimental and theoretical investigations. At
first stress induced cracks were most often examined, see for
example /1, 2, 3, 4/.
When investigating the structural behaviour of concrete ex-

posed to fire attack however, it is also necessary to include the
influence of elevated and high temperatures on the crack behavi-
our. Consequently it is of special interest, when the cracks
appear, what their effects on the deformation behaviour of con-
crete are, whether the cracks appear as a consequence of mechani-
cal or thermal stresses and what kind of stress is dominant for
the crack behaviour.
Finding exact, clear solutions for the tasks involved in

answering these questions causes some experimental and analytical
difficulties, even at normal temperatures. These difficulties be-
come considerably greater, when test temperatures rise. This
applies not only for the experimental procedures but also when the
results of the tests with specific specimens and specific con-
cretes should be generalized.
Due to the experimental difficulties, observations and exact

measurements of crack widths using optical methods were often only
carried out on the specimens' free superficial surfaces.
However, it must be taken into consideration that the ob-

served crack patterns are a two-dimensional picture of the volu-
metric crack distribution.
Kordina, Wydra, and Ehm 89
One possibility to investigate the SWll of deteriorations in

the volume of a tested specimen is the measurement of the porosity
with regard to temperature exposure. For this the cracks are in-
terpreted as great pores.
The disadvantages of the porosity measurement method and -

to some extent the optical method - are that the measurements can
not be carried out when the specimens are exposed to high tempera-
tures. The measurements must take place after cooling down to room
temperature.
The point in time when measurements are taken plays a great

role. Because the specimens alter their properties with time based
on various physical and chemical reactions, in particular the ab-
sorption of humidity.
Therefore, with the results of these measurements it is not

possible to say at exactly what time the measured values - crack
widths or porosities - had their origin, whether it was during the
heating time, the holding period or the cooling time.
These problems occured particularly, when the specimens were

exposed to temperature cycles.
The aim of the present work is the analysis of the formation

and propagation of cracks in mechanically loaded and unloaded
specimens during high temperature exposures and the answering of
the question at what moment of a heating-cooling cycle will cracks
begin.
In addition, informations will be given about the moment,

when the process of crack formation and propagation is particular-
ly intense and what the effects on the deformation behaviour are.
For this purpose, deformation measurements in the three axes

of two different specimen geometries were carried out. Thus, the
volumetric deformation behaviour could be presented.
The method used to determine the kinetics of crack propaga-

tion was that of acoustic emission analysis. With this method
crack formations and propagations could be registered in concrete
under specific mechanical and thermal stresses in real-time,
whereby the asswnption is made that a clear correlation exists
between crack kinetics and acoustic emission activity.
The percentage of structural destruction which comes from

heating, holding time or cooling in dependence on temperature and
stress level was determined by evaluating the acoustic emission
measurements. Since different test specimen geometries and magni-
tudes were used, it is possible to generalize the test results on
concrete.
TEST PROGRAMME
Scope of the Investigations
A survey of the tests carried out is given in Table 1. The

stress level a was varied between 0 and 60 per cent of the uni-
axial compressive strength of the specimens at room temperature.
The heating under the stress level 0 per cent means that the ther-
mal expansion was measured.
The maximum temperatures were chosen for reasons which main-

ly concern temperature dependent alterations in the mechanical
properties of concrete. The main emphasis of the investigations
was at temperatures of 300 - 450 °C. Tests with high stress levels
and high maximum temperatures are not useful, because the danger
exists that the specimens will be destroyed during one of the tem-
perature phases.
Testing Procedure
A sketch of the testing procedure is given in figure 1.

First, the specimens were loaded uniaxially with the specified
stress level a. 'rhis compressive stress was kept constant during
the entire temperature course. Then the specimens were heated up
to a maximum temperature between 150 °C and 750 °C. The heating
rate was 2 K/min. After a holding period of 2 hours at the maxi-
mUIII temperature the specimens were cooled under load. During the
entire test the compressive load and the temperature were re-
gistered. The total deformation in the loaded axis and as well
as in the unloaded axis were measured in order to investigate
the volUIIletric deformation behaviour.
The acoustic emission activity was registered during loading,

heating, temperature holding and cooling. Thus, the mechanical
changes in the internal structure of the concrete could be studied
in real-time.
After cooling the surfaces of the specimens were photographed

and investigated for cracks.
EXPERIMENTAL DETAILS
Specimens
Two types of specimens were investigated in the course< of the

experiments:
- discs (200 x 200 x 50 mm 3 ) ;
-prisms (110 x 110 x 330 mm 3 ).
The discs were sawn out of cubes measuring 200 x 200 x

200 m3 • '!'hey were loaded on two of their narrow sides, which were
ground in order to have smooth surfaces. The prisms were cast in
special moulds. The loaded surfaces were also ground.
The mix proportions and some important properties of the con-

crete used in the experiments are given in Table 2.
After demoulding, the specimens were stored for 7 days under

water and then at a temperature of 20 °C and a relative humidity
of 65 per cent. They were not tested before an age of 90 days.
Loading Equipment and Load Measurements
The loading frame of the test equipment and the essential

loading and measuring devices are presented in Figure 2. The frame
is designed as a welded steel construction of high stiffness. The
mechanical loading of the specimens is achieved by two stress rate
controlled hydraulic cylinders each with a maximum force of
1000 kN. The heat resistance pressure pistons are cooled by water
(for a more detailed description of this equipment see /5/).
The disc-shaped specimens were loaded by so-called "loading

-brushes" in order to provide a defined, homogeneous stress state.
The prisms were loaded by stiff loading platens. Two load cells
constructed between the cylinders and the specimen were available
for measuring the load forces.
Deformation Measurements
The deformations of the concrete specimens in the three axes

were measured directly by means of a specially developed high
temperature - deformation measuring device illustrated in Figure 3
on a disc-shaped specimen. With this device, the strains are
transferred by transmitters out of the hot furnace regions. In-
ductive displacement transducers are installed in the region of
ambient temperatures. 'rhe deformation measuring system for the
determination of the defonnations in the loaded axis is designed
according to the principle of a dilatometer, that means thermal
strains of the transmitters are compensated by a special geometric
arrangement.
There are two measurement systems for the strains in the un-
stressed axes. One is similar to the measurement system for the
loaded axis. The other system is conducted according to a lateral
extensometer. The thermal strains of the transmitters, separately
determined by calibration tests, must be taken into considera-
tion.
Equipment for Measuring the Acoustic Waves
A piezo-electric transducer was used to convert the mechani-

cal waves, which were emitted, when the concrete was stressed me-
chanically or thennally, into electric signals. Due to the high
temperatures, the transducer could not be .installed directly on
the surface of the specimens. The acoustic waves had to be guided
out of the hot region of the furnace with the aid of a waveguide.
The waveguide consisted of a quartz-glass rod, which was glued on

to the specimens' surface with a thermal resistant resin. 'I'he
coupling of the transducer on the other end of the wave guide was
done by a contact agent.
The electric signal given by the transducer trav~lled along

a very short cable to a preamplifier. The signal then travelled
through frequency filters and the main amplifier to the essential
measuring system.
In this study, two methods were used to evaluate the acoustic

emission signals. The first method was "ring-down-counting", that
is, all pulses whose voltage exceeds a certain predetermined
threshold value were transformed into normalized square wave im-
pulses and counted by a pulse counter. '!'he total number of these
pulses is called "pulse sum". The other method was the counting
of the pulses per a unit of time, here per minute. The resulting
value is called "pulse rate".
An extensive description of the application of acoustic emis-

sion analyses on concrete is given in /6/.
Heating System and Temperature Measurements
The heating of the specimens was realized by an electrical

furnace. A 3-zone-PID programme controler regulated the heating
and cooling of the specimens and the realization of the tempera-
ture curves.
The temperature of the specimens was measured with NiCr-Ni-

thermocouples, which were applied on the specimens' surfaces by
means of a heat-resistant glue. The moderate temperature increase
of 2 K/min and the holding period of 2 hours provided for the
equalisation of possible temperature gradients.
Equipment for Photographing Cracks
The cracks on the surfaces of the specimens were photographed

with the aid of special microscopic camera equipment, which
allowed magnifications between si~ and 50 times. The most fre-
quently used magnifications were six times and 25 times (see Figu-
res 4 a- f). Thus, cracks with a minimum width of approximately
0,01 millimeters could be observed and measured.
RESULTS AND DISCUSSIONS
In the following, some typical results of the experiments are

presented in pictures. First, microphotographs of crack distribu-
tions are discussed. Then, from four selected experiments the tem-
perature courses, measured values of total strains and ~caustic
pulse sums and the calculated values of total creeps and acoustic
pulse rates are presented as typical examples for all experiments
(Figures 5 - 8).
Surface Cracks
In Figures 4 a - d and S a - b some examples of surface

cracks after thermal and mechanical stresses are shown. In Figures
4a and 4b regions of the unloaded surfaces of specimens are shown,
which were heated to 600 °C under a compressive stress level a of
40 per cent.
The magnification of the microscope was 6 times. In Figure 4a

a local spallation (this means a spatial phenomenon) is seen,
which is caused by the great thermal expansion of a quartzitic
gravel-stone. Several cracks (width 0,01 - 0,02 mm) run mainly in
the load direction. When pores lie in their path, the cracks go
through rather than around them.
In Figure 4b a plain spallation resulting from two broad

cracks (width~ 0,1 mm) can be seen.
Figures 4c and 4d are photographed with a higher magnifica-

tion of 2S times. In Figure 4c, thermally induced cracks (Tmax =
7SO oc, a = 0 per cent) can be observed. These cracks (widths be-
tween O,OOS and 0,02 mm) are distributed over the surface without
giving preference to a specific direction. When stress was applied
(a = 20 per cent) to a specimen heated to 7SO °C, broad cracks in
direction of the load were formed in addition to the thermally
induced, arbitrarity distributed narrow cracks (Fig. 4d).
In Figures Sa and Sb areas of the specimens' loaded surfaces

are shown. These surfaces are ground, so that the gravel-stones
embedded in the cementstone are visible.
In Figure Sa (magnification 6 times) a fine crack (width

~ 0,01 mm) runs in the bond-zone (interfacial region) between one
aggregate stone and cement stone (hardened cement paste) around
this aggregate stone and then crosses the cement stone to the tip
of another aggregate stone.
With higher magnification (2S times) it can be seen that

this type of crack branches and then runs from one small sand
grain to others (Fig. Sb).
Deformation Behaviour
Total Strains
'rhe total strains (diagrams b in Figures 6 - 9) were measured
during the experiments on the surfaces of the specimens. The
strains in all three disc axes were registered. Due to the symme-
trical form of the prisms, it was only necessary to measure in the
load direction (longitudinal strain) and in one of the perpendi-
cular directions (lateral strain).
The volumetric strains £tot,v were calculated as given by the

formula:
for discs and
£tot,v(t,T1) = £tot,y(t,T1) + 2 £tot,z(t,T1)

for prisms with
x,y,z = directionsof the specimens,

y: parallel to the load
x,z: perpendicular to the load.
The measured values for longitudinal and lateral strains are

in good agreement for the nonloaded heating of a prism (Fig. 6).
They reach values of about 12.5 °~ 0 at 750 oc.
During the holding and cooling periods both values differ

somewhat. In both cases no increase could be observed. At the be-
ginning of the cooling period, a sharp decrease was observable.
Then the strains decrease continuously to residual strains of
about 4 °/00 •
Calculated from the equations shown above, the temperature-

dependent volumetric strains from heating to different maximum
temperatures and then cooling are aillustrated in Fig. 10. No ef-
fect of the holding period could be observed after heating to 350
and 450 °C. The residual strains after cooling are negative due to
a slight compressive stress on the specimens induced by the mea-
suring equipment. After heating to 650 °C, the effect of the
holding period is clear: the strains increase without an increase
in temperature. Another effect can be observed, namely after
heating to ~ 750 °C: the strains remain constant although the
temperature rises. The strains also remain constant when the tem-
perature first decreases. After cooling to room temperature posi-
tive strains around 10 o/00 remain.
Longitudinal and lateral strains differed, when the specimens

were loaded during heating.
In the load direction the strain is much less than in the la-
teral directions. In comparison with the nonloaded specimen (Fig.
6), £ 1 is less and £ 3 is somewhat greater.
The differences between the strains in the loaded and the

nonloaded axis can clearly be seen, when considering the results
of the experiments on discs. There are also differences between
the strains in the two nonloaded axes. This probably occurs be-
cause of the unsymmetrical shape of the specimens. Due to the re-
latively low temperature of 450 °C the strains are also not very
large.
When the specimens were loaded with a high stress level (a =

0.6, Fig. 9), the strain in the loaded axis remained negative
during the entire course of the experiment. It reaches a residual
value of about - 8 per mille, whereas the strains in the unloaded
axes reach great positive values. During the holding period an in-
crease of the strains occurs.
In Table 3, the total volumetric strains during the nonsteady

temperature course are shown for some specific temperatures. It
can be seen that during heating the influence of the stress level
(if it is below 40 per cent) on the volumetric strains is not
very great particularly at higher temperatures.
The volumetric strains with the same temperature both during

heating and cooling were nearly the same, when the specimens were
heated to maximum temperatures not higher than 450 °C.
However, when the maximum temperature was 600 - 700 °C 1 the

volumetric strains during cooling were 2 - 3 times greater than
during heating at the same temperature. This means that the ma-
ximum temperature has both an extraordinarily great and fundamen-
tal influence on the volumetric strains.
The limit for the maximum temperature must be drawn at 450°C.

Therefore the remarkable differences concerning the volumetric
strains during cooling, which can be seen in the upper righthand
corner of Table 3 must be explained by strong physical and chemi-
cal alterations in the structure of concrete at about 450 °C.
'!:~~~~-~~~~~
The diagrams "c" in Figures 7, 8 and 9 show the total creep
strains during the entire experimental cycle. The total creep of
concrete (linear in direction x, y, z or volrnnetric) in the time
interval (t - T1 ) (see Fig. 1) is defined as the time- and tempe-
rature dependent increase of the strains (in direction x,y,z or
volumetric) under the influence of a steady load. In order to
evaluate the total creep Ec i(t,T) in the time interval from
t = T1 to t < T2 the followlng equation was used:
£c,i(t,T1) = £tot,i(t,T1) - Eo,i(T1) - ET,i,m(T,t,T1)

with
£tot,i (t,Tl) total strains at the moment t for a specimen

stressed at the concrete age T 1 ;
stress induced strains at room temperature, im-

mediately after loading;
£T 11.. ,m (T I t , TI ) mean stress independent thermal strains during

the time interval (t- TI).
The strains ET,i,m(T,t,TI) were calculated on the basis of

measurements with nonloaded specimens, which were exposed to
exactly the same temperature conditions as the loaded specimens.
The measured values ET,i(T,t,TI) were mathematically represented
with the aid of exponential spline functions (see also /7/). The
mathematically smoothed course of the volumetric strains with
dependence on the temperature is given in Figure 10.
The diagrams show that there are different regions in the

creep behaviour, as can clearly be seen in the course of the vo-
lumetric creep curve Ec,v· There are regions with increasing, de-
creasing and constant strains. This includes points of inflections
and discontinuities and holds for discs and for prisms.
Sometimes the discontinuities were so small that the danger

existed that they could be overlooked. Particularly, when the
curves had smoothed or the scale of a diagram was too large.
Therefore, it is necessary to look at small, specific time in-
tervals.
Acoustic Emission Activity
Pulse Rates and Pulse Sums
The acoustic emission activity is of fundamental importance

for the present investigations. It is assumed that there is an
exact, clear correlation between crack formation and propagation
processes and acoustic emissions. Therefore, acoustic emissions
can be considered as an informant of damaging processes in
stressed concrete.
The last two diagrams in Pictures 6 - 9 show the results of

the acoustic emission analyses, which were done in real-time
during the experiments. The curves of the acoustic emission acti-
vities (pulse rates) have two clear regions. The first region of
activity begins around 10 min after the onset of heating in the
discs and 100 min after the onset of heating in the prisms. One
cause of this difference in time is assumed to be the difference
in the temperature distribution in the specimens. In the discs,
the temperature equalisation happens more quickly than in the
prisms. A clear amount of the pulse sums can first be seen after
around 100 min and 150 min respectively. After that the acoustic
emission activities are very high (maximums at 200 min for discs
and 220 min for prisms) and at the same time the pulse sums in-
crease considerably. When the temperature is constant the acti-
vities nearly vanish and the pulse sums show only a small in-
crease.
Independent of the specimen form and the maximum temperature

the second region of activities could be observed at the beginning
of the cooling phase. Then the activities vanish and the pulse
sums remain nearly constant until cooling down to room tempera-
ture.
The dependence of the acoustic emission pulse sum on tempera-

ture and stress level is illustrated in Figure 11. For all cases
the increase of the pulse s~Js is very slight at first. It can be
seen that the stress level clearly influences the course of the
pulse sum curve.
In Table 4, the measured acoustic emission pulse sums are

divided into the pulse sums during heating, holding period and
cooling. For the definitions of the symbols used see Figure 12.
The distribution of the pulse sums into specific time periods
shows a strong dependence on the maximum temperature reached.
When the maximum temperature was below 350 °C the main part of
the acoustic emissions was registered during cooling. That means
that 15 per cent of the internal damages occur during this time.
When the maximum temperature reached around 450 °C there was an
equal amount of the acoustic emissions during heating and cooling.
Only a small part of the damages occur during the holding period.
When the specimens were heated beyond 600 °C, the main part of the
damages was caused during the heating period. Only 25 per cent
occured during cooling.
In all cases, the percentage of damages during the holding

period was small and equal. It can be said, that much damage
occurs during heating, when high temperatures are reached. On the
other hand, the percentage of damages during cooling is great,
when the maximum temperature is not so high (Tmax < 450 °C).
The influence of the stress level could not be as clearly

defined as the influence of the maximum temperature.
Correlations between Deforrnation Behaviour and Acoustic Emissions
A very interesting point in the discussion of the experimen-

tal results is to observe the correlations between the deformation
behaviour and the acoustic emission activity. In all cases, when
high deformation rates appeared, high acoustic emission activities
could be observed both in the heating and cooling phases. When the
deformations were constant however, the activity nearly vanished.
The total creep strains showed a particular behaviour, when

reaching and leaving the holding period. In all cases, a reversal
of the creep strains direction occured. This reversal was accoom-
panied by high acoustic activity. Shortly thereafter, the total
creep strains developed continuously. During the holding period
the creep behaviour was as known from tests at constant tempera-
tures. The changes in the temperature course particularly effected
creep strains in the lateral direction. Whereas the changes in the
longitudinal creep were not so great, the lateral creep was very
sensitive to the temperature course changes and showed qualitati-
ve alterations.
Based on the above information, it can be said, that the vo-

lumetric creep strain behaviour is mainly due to the lateral creep
strains.
CONCLUSIONS
The main aim of the present work was to investigate at what

moment of a heating-cooling-cycle damages - that is crack forma-
tion and propagation processes - in plain concrete occur and how
extensive these damages are.
In order to observe the damaging processes the method of

Acoustic Emission Analysis was used with the assumption, that
there is a clear relationship between acoustic emission and crack
formation and propagation.
When sustained load and maximrnn temperature are compared

concerning their effects on damaging processes, the maximum tem-
perature is shown to be the most important magnitude. At maximum
temperatures of up to 350 °C the most damages occured during the
cooling phase.
When maximum temperatures around 450 °C were reached the da-

mages were proportionately distributed. When the specimens were
heated beyond temperatures of 450 oc the most damages occured
during heating.
The actual damaging processes were always accompanied by a

high magnitude of acoustic emission activity. This high activity
could be correlated to the deformation behaviour, in particular
the creep behaviour, whereby the influences on the longitudinal
strains were somewhat less, the influences on the lateral strains
were very clear. 'l'his means, that there were activities of crack
formations in the longitudinal stressed directions and crack ex-
tensions in the lateral unstressed directions.
As already mentioned, large amounts of the damages occured

during the cooling phase, when the maximum temperature was below
450 °C. Observing the acoustic emission activity, it can be seen
that at the beginning of the cooling phase the activity is high
and so the damaging processes are high. Thus, it can be assumed,
that the crack formation processes will be low, when the tempe-
rature changes are low.
In the present work, quantitative information about the de-

formation behaviour is given along with more qualitative informa-
tion about crack formation processes. In order to obtain quanti-
tative data about the structural alterations, it is planned to
investigate the volrnues of crack systems with the aid of porosity
measurements in addition to the measurements described here. Then
it will be possible to say, when and at what magnitude, cracks
have influence on the strength and deformation behaviour of con-
crete duriny nonsteady temperature conditions.
HEFERENCES
/1/ Kaplan, M.F.: Crack propagation and the Fracture of Concrete.

Journal of the American Concrete Institute, Title No. 58-28,
Nov. 1961.
/2/ American Concrete Institute: Causes, Mechanism and Control of

Cracking in Concrete. ACI Special Publication SP-20, 1968.
/3/ Mindess, S.; Diamond, S.: The Cracking and Fracture of Mor-
tar. Materiaux et Constructions, Vol. 15 - No. 86, 1982.
/4/ Zaitsev, Y.B.; Wittmann, F.H.: Simulation of Crack Propaga-

tion and Failure of Concrete. Materiaux et Constructions.
Vol. 14- No. 83, 1981.
/5/ Ehm, C.; Kordina, K.; Schneider, U.: The Behaviour of Con-
crete under Biaxial Conditions and High Temperatures. RILEM-
CEB-CNHS International Conference on Concrete under Multi-
axial Conditions. Toulouse, 1984.
/6/ Diederichs, U.; Schneider, U.; Terrien, M.: Formation and

Propagation of Cracks and Acoustic Emission. In: Fracture
Mechanics of Concrete, ed. by F.H. Wittmann, Elsevier
Science Publishers B.V., Amsterdam, 1983.
/7 I Wydra, W. ; Diederichs, U. ; Schneider, U.: Deformation Beha-

viour and Creep Effects of Concrete during a Heating-Cooling-
Cycle: Proceedings of the 8th International Conference on
"Structural Mechanics in Reactor 'l'echnology" , Paper H 5/6,
Brussels, 1985.
Table 1: Loading conditions and experimental scope
Stress level a Maximum temperature Tmax (°C]

[per cent of
strength at 20 °C) 150 300 450 600 750
0 X )(
• )( )(
20 )(
• )( )(
40 )( )(
• )(
60 )(
•
x prisms •discs
Table 2: Mix proportions and concrete properties
Component in kg/m 3
Cement PZ 35 F 260
Quartzitic sand 0 - 2 mm 923
Quartzitic gravel 2 - 8 mm 191
Quartzitic gravel 8 - 16 mm 768
Water 190
Fly ash 60
Density 2,31 kg/dm 3
Compressive strength 33,0 N/mm 2

(28 days, 150 mm cube)
Table 3: Total volumetric strains Ev in per mille
Specimen Tmax At the heating temperature After cooling to temperature

a
•c %
Form No. 100 •c 250 •c 400 •c sso •c 2so •c so •c
433 460 0 3.08 7.52 12.42 - 5. 95 0.27

VI
u
VI
441
465
463 20
450 20
3.15
3.03
8.31
7.01
13.36
11.60
-- 5.25
5.27
- 1.40
- 1.57
·~
c 426 459 40 3.44 7.90 13.55 - 5.54 - 0.78
443 450 60 2.44 6.06 12.08 - 7.10 0.53
(Tmax "' 450 •c)

mean value 3.03 7.36 12.60 - 5.82 - 0.59
552
551
172
323 0
0 2.81
2.58
-
7.71
-
-
-- -
8.52
0.95
- 0.27
VI
e
479
488
330 20
325 20
2.40
2.18
6.63
6.18
-
- -- 6.49
6.13 -
-
1. 93
VI
·;: 533 185 40 1.86 - - - - - 0.81
---
0. 480 320 40 1.74 6.20 - 5.54 - 1.64
519
546
340 40
336 60
2.59
1.87
6.30
6.48
-- 5.60
7.81
- 1.91
o. 31
(Tmax < "' 350 •c)

mean value 2.25 6.58 - - 6.68 - o. 76
550 630 0 3.08 7.53 12.80 22.89 16.05 10.22

520 750 0 3.03 8.07 13.81 23.51 17.47 11.79
VI
e
VI
477
515
600
617
20
20
2.32
2.50
4.99
6.39
9.03
10.39
17.22
18.63
11.62
14.21
-
9.98
·;:
0.
523
478
760
600
20
40
2. 73
1.61
7.22
5.25
11.63
9.66
19.10
18.48
-
22.08
-
19.24
517 609 40 2.13 6.26 12.75 22.65 23.09 27.60
(600°C< Tmax < "'7so•c 2.49 6.53 11.44 20.35 17.42 15.77
mean value
Prisms , genera 11 y 6.68 1 )/17.42 2 ) -0.76 1 )/15.77 2 )
2.36 6.55 11.44 20.35
mean value
Concrete generally,
independent of 2.53 6.78 II. 92 20.35 6.29 1 )/17.42 2 ) -0.69 1 )/15.77 2 )
specimen form
Remarks: I) Mean value for specimens with Tmax ~ 450 •c

2) Mean value for specimens with 600 •c < Tmax < 750 •c
Table 4: Acoustic emission pulse sums
Specimen Heating Holdtno period Cooling

T a s 11 •10" 3 s 111 ·10" 3 s 1y·1o" 3 s y-s
form No.
•c s S II • 100 S
sll1-sll
--s-;-. 100 s
1 111
--s-;-. 100 s
"SJY
433 460 0 25.10 31.45 55.00 45.6 11.5 42.8
441 463 20 121,60 180.00 285.00 42.1 20.5 36.8
"'
:.l 465 450 20 52.00 11.30 121.98 40.6 19,8 39.6
Q 426 459 40 129.00 145.40 236.00 54.7 6,9 38.4
443 450 60 65,00 111.50 240.90 27 .o 21.8 51.2
<r • 450 •c) mean value 42.1 16.1 41.8
552 172 0 0,55 1,51 4,15 11.6 20,2 68.2
551 323 0 0.57 1.14 20.46 2.8 2.8 94.4
479 330 20 2.08 6,22 11.90 17,5 34.8 47,7
N
·;:
488
533
325
185
20
40
0.85
1.60
~.95
1.11
18.og
5.58
4.7
28.7
11.6
2.0
83.7
69.4
a. 480 320 40 2.27 4. 39 15.25 14.9 13.9 11,2
51g 340 40 1.17 ~.52 11.21 10.4 12.0 17.5
546 336 60 3.40 6.65 50.20 6.8 6.5 86,8
(T s " 350 •c) mean value 12.2 13.0 74,g
550 630 0 185.10 249.65 303.36 61.0 21.3 11.7
520 750 0 266.00 267 .oo 294.90 90.2 0,3 9.5
411 600 20 194.00 254.12 320.56 60.5 18.8 20.7
N
·;:
515 617 20 309.00 351.00 469.50 65.8 8.9 25.2
523 160 20 223.15 236.95 3~0.05 69.9 4,1 26.0
a.
478 600 40 18.00 og,1o 188.70 41.3 5.9 52.8
517 609 40 10g,5o 115.00 151.15 12.2 3.6 24.2
(6oo •c < r s " 750 •c) mean value 65.8 9.0 25.2
Prisms, generally mean value 37.2 11.1 51.7
Concrete, generally (independent
of specimens geometric and 38.4 12.4 49.2
temperature)
Su • pulse sum at the end of the heating (t=t 11 l

s 111 • pulse sum at the end of the holding period (t • till)
s,v • pulse sum at the end of the cooling (t • llyl
t:l
Qi
>
~
en
en
-
Cll
'-
en
a.
"t, timet
1-
I
CII
'-
.a
~
Cll
Tmax --'--
I
I
--·---....
a.
-
E
Cll
I
timet
Fig, !-Fundamental testing procedure

Fig. 2-Test rig for high temperature concrete tests

deformation transmitters
(unloaded axis I
rod I deformation transmillers

tube\ (loaded axis!
/inductive displacement
transducer
prtncipal directions
Fig. 3-Instrumentation for high temperature deformation and

acoustic emission measurements
2mm
(a) a 40 percent; Tmax = 600 C (b) a= 0 percent; Tmax
(c) a = 0 percent; Tmax 750 c (d) Cl. 20 percent;
Fig. 4-Examples for surface crack patterns in dependence of

stress level a and maximum temperature Tmax
(a) Ct = 0 percent; Tmax = 750 C (h) et = 20 percent; Tmax = 600 C
Fig. 5-Examples for surface crack patterns in dependence of

stress level a and maximum temperature Tmax
a) ~ 1000 Prism no. 520

.!; stress level ct= 0.0
0
b) -;!!.
.!;
- - - - Ey
"'c: .......... Ez
·a'- - Ey=Ey+2•£z
.....
0 "'
:s
c) "'S2
I
)( 300
E
:J
"'
QJ
-s"'a.
d)
800 1000 1200

time in minutes
Fig. 6-Temperature course and experimental results during a

heating-cooling cycle for prism with stress level a = 0.00
a) u BOO
0 Prism no.477
.G stress level a=0.2
Qj
:; 600
.....
0 ,_
Qj
a. 400
E
$
200
b) 0~
0
.b - - - - Ey
<II ......... Ez
c
·a - Ey= £y+2·tz
.1::
<II
0
:§
..·························...
.·..
······ .,...,---------,,
c) 0
---
~
0
.G
a.
0 .....................·····························
Qj
,_
Qj -2
u
0 -4
:§
d) "'
I
S2
e:::J"
Ill
Qj
.!!!
:::J
a.
e) j'
c
·e
.G
~,_
Qj
Ill
"5
a.
0
0 100 400 500 60
time in minutes

heating-cooling cycle for prism with stress level a = 0.20
a) ~1000
Disc no. 465
.!:
~800
stress level ct= 0.2
~
~6
E
$!
b) 0~
0
- - - - - Ey
.!: - · - · - - Ex
Ill ••••••·••• Ez
c ---Ey:Ey•Ex•Ez
£Ill
:S
·~·'!'::'~::::':·~~:·:.-:=::-:::. . . . . . -.-.-. -::.:-::.:-:-:.~·.:-::: ~::-:-.:::-.: .
.s ---------------------
0
c) ';!.
.s
0.
u
~
..............
000000 0 00 •:-..:.:\•M.,.....-•U".•-~"'=:::: ;':'.::T:o":"'o~':": H"To"oftlo~: :r:o-:'0 ::~: ::0:: :"::.-:". 00 : : : ; : • ; ; : : • . : : : . : ' . ".:::'.: : : : . : - . :
._
............ _
~
d) "''52
X
E
iil
QJ
Ill
:;
0.
e)
.'~
,!;
15000·~~~==~==========~==~====~
time in minutes

heating-cooling cycle for disc with stress level a = 0.20
Kordina, Wydra, and Ehm lll
a) \;' 1000
Disc no. 443
.!:
f! 800 stress level ct=0.6
:I
0
~ 600
E
$
1.00
0
b) '$. - - - - Ey
.!: -·-·-Ex
Ill
c: ............ Ez
~Ill - - Ey= Ey+Ex + Ez
"5
2 -·-·-·-·-·-·-·-·-·-·-·-·-·-·
------- ~~ ... ..... _
-10
---- ---------------------
........···"··..............................................................................
0
c) '$. 6
.!:
.f!
a. 4
2
:
-·-·-·-·-·-·-·-·-·-·-·-·-·-·-·-·-·-·-·-·-·
u
:§ ···:.::::.:·:;;:< .....·
.s
-2 '', ......
-4 ''
' ... ... _
-6
-8
-------- ---------------
....I
d)
!:? 300
E
•
..
:I
Ill 200
~
:I
a. 100
0
e) I
c:
·e,!; 1
600 800 1000 1200

time in minutes

heating-cooling cycle for disc with stress level a = 0.60
t.O
0~
Prism
.£
c Disc
·a
....Ill
L.
u
£Qj
E
::J
g
0
E
L.
Qj
:5
0~~~--~J_----L---~----~--~~--~----~~
0 t.OO 600 800
temperature in °C
Fig. 10-ThennaJ volumetric strains during heating and cooling down

from different maximum temperatures (calculated form)
..., 500
I
-
0
)(
---0.=0.0
E
:J
Ul
............. a.= 0.2
<II
Ul
-----a.= 0.4
3 -·-·-·-a.= 0.6
a.
.
.
I
;
I
I
temperature in °C
Fig. 11-Mean acoustic emission pulse sum during heating for prism
Ill
E
::J
Ill
1r tn tm 1rv time t
14 ~ hold ~4 ~
heating cooling
time
Fig. 12-Sc hematic representation of acoustic emission pulse

sum course
SP 92-7
Possibilities of Assessing
the Temperatures Reached by Concrete
Building Elements During a Fire
by L. Krampf and A. Haksever
Sy~o~: The present paper is intended to give an overall i nfor-

mation on temperatures which may have occured in concrete cross-
sections affected by a fire
Firstly the paper contains a catalogue of circumstances and vesti-

ges which can be found when the affected building is visited and
which allow conclusions concerning the intensity of the fire and
the resulting value of damages to the structural elements. For
this purpose, residual contents of the building (fire load), the
state of building materials, and the evident condition of the con-
crete structure itself are used.
In the second part of the paper, general informations are given on

the development of natural fires and on the parameters which in-
fluence it. On the basis of two series of experiments - Metz and
Lehrtc -which gave temperature-time developments in compartments
where wood cribs or furniture were burnt, temperature fields are
presented for square concrete cross-sections affected from all
sides, and forT-shaped sections affected from three sides. It is
stated that the presented figures can only be rough approaches for
practical work.
Short, only introductory informations on residual changes of mate-

rial properties due to temperature influences are added to facili-
tate judgements and decisions.
It stands to reason that, if findings won by this guide lead to

the conclusion that the structure may be worthy to be kept, more
intense and detailed investigations have to be done.
Keywords: buildings; concretes; construction materials; damage;

deformation; estimates; evaluation; fires; fire tests; high temper-
ature; temperature; thermal properties
115
116 'lemperatures
L Krampf is a civil engineer and scientific co-1~orker in the In-

stitut fUr Baustoffe Massivbau und Brandschutz of the Technical
University of Braunschweig. She is working in the field of fire
research, and her main interest is the load-bearing and deforma-
tion behaviour of concrete elements and structural assemblies ex-
posed to fire.
A. Haksever is a scientific collaborator in the Special Research

Department "Brandverhalten von Bauteilen" at the Technical Univer-
sity of Braunschweig. He was a structural engineer before turning
to research and is author of a number of papers on the behaviour
of structural elements and structural total systems under fire and
problems of heat balance calculations.
INTRODUCTION
The present paper may serve to support the engineer who is forced
to decide if a reinforced or prestressed concrete structure being
affected by a fire can be reused, if need be after repair, or if
the building must be demolished.
Following the interest to conserve valuable buildings - the inte-

rest of their owners as well as of the public- it is not only ne-
cessary to guarantee an adequate behaviour of the structure during
a fire but to investigate and judge carefully the condition of a
building and its structural elements after a fire has occured.
This judgement will often draw far-reaching consequences. In any
case, safety must take precedence of economy, and the responsible
engineer may be confronted with a very difficult task, the more so
if he is not involved in fire protection questions, in his normal
daily work. The paper shall provide data which will enable the
engineer to value roughly the temperature which may have occured
within concrete cross-sections during a fire. In a number of ca-
ses a decision concerning repairability and reuseability should be
possible with the help of this guide though it will become evident
that in many cases, it cannot replace detailed investigations.
Krampf and Haksever 117
CIRCUMSTANCES AND VESTIGES AT THE PLACE OF CONFLAGRATION
A visit to the place of conflagration should be undertaken as soon

as it will be possible without too much danger. Normally, valua-
ble vestiges are removed when the place is being tidied up. When
revealing vestiges are found, one must be aware of the fact that
only the sum of all circumstances will give sufficient
information- if need be in combination with more detailed
investigations - for the judgement of repairability.
Fire load, combustible material
The temperature development in a fire compartment is mainly depen-

dant on the kind and amount of the burning material. In chapter 3
some examples are given which have been derived from fires with
wood cribs and wooden furniture, and which show clearly the in-
fluence of the fire load to the temperature-time develop-
ment /3,5,6/. The examples make evident that the knowledge of the
burnt material is of main interest.
The temperature curves given in chapter 3 were reached when a com-

plete burn-out had been possible and permitted. If by interven-
tion of the fire brigade or by other circumstances, the fire load
could not burn up completely the temperature may have kept at a
significantly lower level.
Building Materials
Timber-- If wooden e1ements or parts of them the origin a1 eros s

section of which is still known, can be found near the concrete
element in question, the depth of charring can provide a clue to
the temperature which may have been reached in the vicinity /9/.
Fig. 2.1 shows results won with samples subjected to the ISO 834
standard fire and extinguished immediately after the burners had
been switched off. The scatter of the results is considerable due
to the condition of the wood, knots, cracks, humidity etc. In ad-
dition, the side flanges normally are less affected than the bot-
tom side of a cross-section. Strictly speaking, the depth of
charring occuring during a standard fire is not equal to the d,pth
reached with the same temperatures occuring in a real fire.
A number of factors influences the process. But for the present
purpose it might be allowed to do some conclusions by fig. 2.1.
Due to the fact that during the cooling down period of a real
fire, the depth of charring will increase which has not been in-
118 Temperatures
eluded in fig. 2.1 the estimation of the relevant room temperature

by the aid of the figure, is on the safe side.
Metals-- Aluminium is the most usual light metal in buildings, and

its melting products can be found very often. In this case, tem-
peratures of 650 °C within the metal, corresponding to
about 700 - 750 °C in the surrounding air have to be assumed.
If small light-blue flames occuring at unprotected light metal

members have been observed during the fire that should not lead to
the assumption of higher temperatures: the magnesium included in
the alloy will have burnt already whilst the aluminium was mel-
ting.
After the plastic insulation of copper-cables which normally con-

sists of polyvenylchloride has--melted and burnt away at about
300 °C, the copper-core covers with a black oxid which penetrates
into the core rather quickly and shows green flames at its surfa-
ce. The melting point of copper is at 1084 °C.
Steel structural members as well as nonstructural ones may pre-

serve help for judgement, in a limited range. The material pro-
perties of steel -mainly strength and deformability - are
strongly influenced by temperature. Stability problems excluded,
steel members will collapse as soon as they are reaching their
"critical temperature", that means that temperature at which the
yield stress has decreased to the level of the working stress/20/.
Stress depending critical temperature of constructional steel is
given in fig. 2.2.
Before reaching the critical temperature, steel elements undergo

extreme deformations which, as a rule, are not fully reversible.
But it is evident, that taking into consideration the residual de-
formation of steel members after a fire, their working stress dur-
ing that fire should be known.
If an originally unprotected steel element can be taken for such

reference it may roughly be assumed that the surrounding gas tem-
perature has been 100 °C higher than the steel temperature.
From heat tinting which sometimes is shown by steel surfaces, con-

clusions to temperatures can be done. But only the blue hue is of
practical interest in our context. It signifies that temperatures
of 300- 400 °C will have been reached; with higher temperatures
the blue colour disappears.
High temperature sealing of the surface of steel profiles may oc-

cur with temperatures of about 500 °C already. However, it is
strongly dependant on the exposure time and therefore should not
be taken to prove temperature. But it is important to know that
it does not signify extremely high temperature, in any case.
If it is reported on yellow-greenish flames on the surface of gal-

vanized steel surfaces it may be concluded that the zinc will have
burnt away at about 800 °C, that means about 900 °C of the sur-
rounding gas.
Glass-- Normal glass panes are bursting very early in case of

a fire. From single undestroyed planes no conclusions to low tem-
peratures must be done; however, if larger areas of sound flat
glass planes are found it will follow that the temperature in
their surrounding will not have exceeded 400- 500 °C.
The melting point of usual silicate glass is about 1000 °C and lo-
wer. With special fire protection glass the melting point increa-
ses. Numbers cannot be given here since the products are differ-
ing from each other.
Lightweight Boards-- Lightweight building boards in walls or sus-

pended ceilings will often be affected by a fire but still remain-
ing in their original position. In such cases, their condition
may provide additional reference to the fire attack /13/.
Fig. 2.3 shows the burning-through of chip-boards of different
thickness (approximately corresponding to plywood) - a1 - and the
bri ttl i ng of gyp sum-p 1asterboa rd - az - versus the temperature
reached by the ISO Standard fire. There is no scatter shown, but
it should be regarded as to be considerable due to different ori-
ginal humidity, to cracks and product-specific properties.
Plaster-- Plaster, mainly ceiling plaster, will often fall away in

case of fire. But this may happen in early stages of the fire at-
tack, depending on better or worse bond to the concrete or brick-
work. No conclusions to the working temperature should be done
from the phenomenon.
Rough conclusions to temperature, derived from the condition of

plaster which is still remaining in site may be done with the aid
of figs. 2.3 and 5.1.
120 Temperatures
Brickwork-- Clay-bricks undergo high temperatures during the manu-

facturlng process. From this fact, it can be concluded that in
most cases, they will show no deterioration after a fire has
occured. However, if sintering of bricks is found the temperature
should be valued not lower than about 1200 °C though the composi-
tion of the clay influences the sintering temperature.
Sometimes, brickwork walls, mainly of perforated brick, have suf-

fered spalling of the outermost shell. Reasons of this phenomenon
are unfavourable load redistributions and possibly high initial
moisture content, more than decreasing strength caused by high
temperature.
Aerated Concrete-- The exposed area of aerated concrete elements

often changes colour during a fire, tending to a pale pink, and
show a net of cracks due to the temperature dependant bri ttl i ng of
the material. From the depth of the cracks raw conclusions to the
temperature may be done: with an ambient temperature of about
1000 °C residual cracks of about 20 mm depth will occur.
Due to the heating, aerated concrete suffers decrease of strength.

Roughly, the residual strength may be assumed to be 80 %with
400 °C and 50% with 600 °C reached before /14/.
Concrete Elements-- The deformation of a concrete structure after

a f1re gives an overall impress1on of the intensity of the fire
attack. However, since the value of residual deformations depend~
on many parameters, such as the material characteristics, the sta-
tical system and the working mechanical stresses, direct numerical
conclusions to the fire and the temperatures which may have occur-
ed must not be done.
Soot on concrete surfaces points to relatively low temperature

since with more than 800 °C soot burns away without residue. Of
course, it cannot be excluded that sooting may have happened dur-
ing the cooling down phase of the fire. But the difference can be
found out by the condition of the concrete surface, in most cases
(mellowness, brittleness).
Under heat influence certain concrete aggregates release water or

gas. This process may be a sudden one and cause bursting of
grains thus destroying small concrete areas near the surface of
the member.
There is a wide range of temperature at which such aggregate

spallings can occur, beginning with about 150 °C and up to about
rruo~-715/. It follows that aggregate spalling does not neces-
sarily indicate high ambient temperature.
Spalling of greater depth, often affecting large areas and someti-

mes laying open the reinforcement, results from the co-operation
of a number of unfavourable parameters, but the temperature at
which they may occur is rather moderate and can be taken as
300- 450 oc at the concrete surface, corresponding to about
700- 750 °C of the ambient gas /15/.
Spalling can easily be differentiated from concrete layers which

have fallen away from concrete members which have suffered high
temperature (about 1000 °C) lasting for a rather long while cau-
sing mellowness, brittling and cracking.
In fig. 2.4 the reasons for the temperature depending deteoration

of concrete are given: with about 350 oC dehydration of hematite;
with about 450 °C dehydration of lime, with about 575 °C quartz
transformation /18/.
Concrete members often show a changed colour after a fire. Howe-

ver, contrary to 1iterature /4/, it is not recommended to draw
far-reaching conclusions from this fact since the colours do not
only depend on the temperature reached before but from a number of
other parameters, and they undergo further changes after cooling.
HEATING OF CONCRETE MEMBERS IN NATURAL FIRES
Development of natural fires
The most important boundary conditions, which influence the deve·-·

lopment of a natural fire, are
a) Quantity and type of combustibles in the compartment,

b) Fire load configuration,
c) Distribution of the fire loads in a fire compartment,
d) Geometry and size of the fire room,
e) Thermal properties of the enclosures and the
f) Ventilation conditions, which determine the air supply into
the fire compartment.
122 Temperatures
The parameters given above influence mainly the developing of a

natural fire in a fire compartment.
The figure 3.1 shows principally the three phases of a natural

fire by the development of hot gas temperatures.
-The first phase, which is the so-called ignition phase, includes

generally a rapid increase of the temperatures in the hot gas
layers. The temperature development during this phase is mainly
a function of the burning behaviour of the fire loads as well as
of the fire spread.
The following second phase is defined as burning phase and is

influenced essentially by the size of fire room, the physical
properties of walls and the size and place of openings and con-
sequently by the amount of oxygen available at fire place. Ge-
nerally the temperatures reach a maximum during this period of
fire and the duration of the fully burning phase is limited by
the combustion energy present in fire compartment.
-The third phase of a natural fire is the so-called cooling down

phase. During this phase the heat energy released is insuffi-
cient to cause an additional increase of temperatures or at
least to hold the temperature level constant. This case implies
that a certain heat flow occurs from the hot surroundings back
to the fire room. The heat energy leaving the enclosing walls
influences thoroughly the hot gas temperatures during this pha-
se.
In figure 3.2 the temperature-time-curves of some natural fires

are representatively illustrated and the standard fire curve of
ISO 834 is also added to the graph.
The parameter of the curves in fig. 3.2 is given as fire load of

wood cribs. The corresponding natural fires originate from the
tests carried out in Metz /3/. The size of the fire room amounts
to 12.6 m2 and 39.0 m3 respectively.
The fires in fig. 3.2 belong to a ventilation factor of P = 0.25.

This factor defines the ratio between the openings area (AF.) and
the total area (Aw) of the surface of that wall which inclutles the
opening. The defined ratio is used as a characteristic parameter
for the natural fires in Metz and is to be obtained with respect
to equation (3.1)
Compared with fig. 3.2 the figure 3.3 provides the results for a
ventilation factor 0.6 using the else same boundary conditions as
in fig. 3.2. Obviously a ventilation increase causes lower hot
gas temperatures in the fire room as it is being evident in
fig. 3.3. This apperance can also be confirmed by means of theo-
retical investigations/18/.
In figures 3.2 - 3.3 the temperature-time-curves are illustrated

with mean values. However the measurements show that the tempera-
tures scatter, and a range of 200 °C was sometimes exceeded bet-
ween the max. and min. values. The enclosing walls of the fire
furnace from Metz were highly insulated; consequently relatively
high hot gas temperatures were measured. During the tests atmos-
pherical influences were excluded.
The investigations carried out by Institut fUr Baustoffe, Massiv-

bau und Brandschutz of the Technical University of Braunschweig
under realistic conditions in a residential house in Lehrte showed
that conditions in real fires are obviously more unpredic-
table /6/. This evidence is demonstrated in the figures 3.4. to
3.6. The size of the fire room from Lehrte with a ground floor of
45m 2 was comparable with the fire room in Metz, however the thsu-
lation of the enclosing walls and ceilings made of masonry and
concrete with cement finish and lime plaster was obviously lower.
Consequently lower temperatures were expected during the fire.
In fact this prediction was verified by the temperature measure-

ments during the tests and the figs. 3.4 - 3.6 show this effect in
general.
The results of the tests from Lehrte, especially concerning the

temperature measurements are differing considerably from each oth-
er though the same amount of fire loads and the ventilation factor
(p = 0.30) were used for the tests. The temperature development
in a real fire is influenced mainly by the wind conditions like
the wind speed and direction, but also some other effects can
disturb the theoretical predictions.
Generally it is not reasonable to carry out a heat balance calcu-

lation with the aim to investigate precisely the temperature deve-
lopment of hot gases and then to obtain the probable temperature
fields in structural members, after a fire attack. But in order
to estimate the possible damage of a real fire and to determine
the necessary fire protection measures, such theoretical investi-
124 Temperatures
gations are helpful /8/.
TEMPERATURE FIELDS IN CONCRETE CROSS SECTIONS
In order to predict the temperature distribution in concrete mem-

bers under nonsteady heating conditions a computer program is
written, the working operation of which based on the two dimensio-
nal methods of differences /7, 16/. The method applicated enables
to calculate with a good convergence the temperature fields in
cross sections using the nonreversible thermal properties of mate-
ria 1s.
The temperature calculations in this paper are based on experi-

ments done with wood cribs in compartments the size of which is
comparable to a moderate sitting-room. A number of parameters,
like: fire load, the size of compartment, extinguishing measures,
different athmosperic conditions etc. could not be included , and
thus the limited validity of the examples is evident. If they
shall be used for practical purposes, that should be done with ca-
re, taking into account the possible influences on fire develop-
ment and doing comparisons to the conclusions won by evaluations
following chapter 2 of this paper.
Thermal exposure of concrete members
In order to investigate representatively the thermal exposure of

the concrete members the following temperature-time-curves are
considered
Test series from Metz:

Fire load 30 kg/m2, ventilation p = 0.25 (Fig. 4.1}
30 kg/m2, p = 0.60 (Fig. 4.2}
60 kg/m 2 , p = 0.25 (Fig. 4.3}
Test series from Lehrte:

Fire load 60 kg/m2, ventilation p = 0.30 (Fig. 4.4}
Fire rooms from Lehrte had normal insulation on the walls. At··
mospherical influences contrary to Metz should and could not be
avoided during the tests.
The ventilation factor as p = 0.30 was somewhat unfavourable since

in modern buildings rather p = 0.60 is a normal case.
Temperature fields are plotted for the following cross sections.
An allside exposure is assumed for the cross sections:

b x d = 200 mm x 200 mm and
b x d = 300 mm x 300 mm,
and for a three side exposure from below a T-beam is chosen with
dimensions
b0 /d 0 /d = 300/500/160 (mm)
It is assumed that concrete contains mainly quartzite aggregate.
It is also assumed that the cross sections investigated are not
insulated. Normal plaster and finishes as isolation can be taken
into account by enlarging the cross section as much as the thick-
ness of plaster, if the plaster sheet did not fall down in big
pieces during the fire.
Maximum temperatures
Since the temperature dependant material properties of concrete as

well as of the cold drawn steel are not reversible, the max. tem-
peratures attained in a cross section indicate the residual ther-
mal damage of concrete members after a fire/11/. In order to de-
termine the thermal damage on fire affected concrete elements
isotherms are plotted for max. temperatures attained during the
exposure of the cross sections given above.
In order to determine the damage of fire exposed concrete members

the 300 °C isotherm can be accepted as lower limit by which a cer-
tain loss of strength of material begins/19/. In the following
comparisons this isotherm has been taken into account as basis
boundary curve to predict damaged area in the cross section. The
following conclusions can be drawn from figures plotted:
a) In the test from Metz with a fire load of 30 kg/m2 and p = 0.25
it is being evident that the exposed cross section of
200 mm x 200 mm was heated up to core even more than 300 °C,
whilst for p = 0.60 the penetration depth of the same isotherm
on the major axis of the cross section amounts to only 10 mm
(comp. figs. 4.1 and 4.2).
126 Temperatures
b) The comparison of the fire loads 30 and 60 kg/m2 from Metz for
a ventilation factor p = 0.25 shows that the increase of fire
1oads causes the heating of total cross section
area (200 x 200) more than 450 oc (comp. figs. 4.1 and 4.3).
c) Under the assumption that the thermal properties of a fire room

are kept unchanged and there is enough air flow into the fire
room during the fire it can be concluded that the size of fire
loads has a major effect on fire severity, but the ventilation
conditions influence also the fire attack.
d) Besides that the thermal properties of the walls have a dis-

tinct influence on the exposure of the structural members.
This conclusion can be drawn by the comparison of the
tests Metz-Lehrte for a fire load 60 kg/m 2, p = 0.25. In Metz
the total cross section area (200 x 200) is heated up over
450 °C, whilst in Lehrte the same isotherm has only a depth of
about 15 mm (comp. figs. 4.3 and 4.4).
Obviously, to use the temperature fields from Metz for practi-

cal purposes, provides at least the prediction of the tempera-
ture fields of exposed concrete members on the safe side.
e) Additional comparisons have shown that for greater cross sec-

tions more favourable conditions exist in case of fire. Even
for an unfavourable fire load of 60 kg/m 2 from
Lehrte (fig. 4.4) the 300 °C isotherm did not reach a depth of
30 mm in a cross section from 300 mm x 300 mm. This effect is
caused by the higher heat capacity of the massive structural
members.
Temperature fields during fire
The development of temperature fields in exposed cross sections

(200 x 200) are illustrated as plot figures for some typical heat-
ing phases of fire. The figures show that the temperature distri-
bution in a cross section is in agreement with the developing hot
gas temperatures during the ignition and fully developed phases of
fire (comp. fig.3.1 to 4.5). During these phases of fire steep
temperature gradients exist at boundary zones of exposed cross
sections while the temperature development in core remains almost
uninfluenced.
During the cooling down period however an interesting effect oc-

curs; although the inner parts of the cross section still show a
temperature rise, the cooling of the edge zone is being affected
by the decrease of ambient temperatures. In some cases temperatu-
re hills can be observed between the outer and inner regions of
the cross sections. These effects imply, as mentioned previously,
a change in the direction of heat flow between the exposed con-
crete member and its surrounding.
DAMAGES, REUSEABILITY
This paper does not aim at detailed recommendations on the judge-

ment of repairability and reuseability of concrete structures,
which have been affected by a fire. But some general information
on possible deteoration of the materials and the structural ele-
ments should complete what has been said before.
Concrete
The residual strength of concrete can be tested by the well known

measures. The dependance of decreased strength on the temperature
reached before as well as on the cooling down conditions can be
taken from fig. 5.1 /21/.
As a rough solution, the decrease of the elastic modulus versus

temperature may be taken as equal to that of strength.
Steel
The behaviour of reinforcing and prestressing steels during a fire

and their condition after it, is affected by their composition and
processes of manufacture. Hot rolled steels get back their origi-
nal strength after they have been cooled down. From the
"after-treatment" caused by the suffered temperature even a
hardening effect can occur. However, this phenomenon can also be
caused by a chill effect from extinguishing water, possibly accom-
panied by embrittlement of the material. In practice, this effect
should be taken into consideration only if the reinforcement of
columns has become layed open and possibly met by water.
Work-hardened steels undergo a decrease of strength when they are

heated. This decrease is not fully reversible after cooling. De-
crease of strength down to that of the original - unhardened - ma-
terial may occur if that temperature ·which removes the after-
treatment effects, has lasted long enough.
128 Temperatures
The injurious temperature range and the percentage of deteoration

depends on the kind and intensity of work-hardening; high tensile
prestressing steels are most susceptible /2/. Figs. 5.2 and 5.3
can give an impression of the effect, only; in case of doubt spe-
cial investigations have to be done.
Since the mechanical stress working during the heating process can
influence the residual characteristics, considerable differences
from the courses given in figs. 5.2 and 5.3 may occur.
It should be considered that steel arranged near the surface, es-

pecially near the corners of a cross-section are affected to a
higher degree than those deeper within the concrete. This can
easily be seen in the figures of Chapter 4 (isothermes).
Bond
The bond between concrete and reinforcement is deteorated by heat-

ing. The residual bond strength is not only depending on the tem-
perature reached before but in addition on the time of its action.
However, with preceeding temperatures of 300 °C the residual bond
strength should be estimated to be not higher than 85% and with
500 °C not higher than 50% of the initial one.
Prestressing
Prestressing which possibly decreases during a fire can partly be

increased again during the cooling period. Fig. 5.4 can give an
impression of this process, only. It shows test results won with
high tensile prestressing wires of 9 mm diameter /1/.
Chemical influences
In case of a fire, corrosive decomposition products may be releas-

ed which can penetrate into concrete cross-sections and affect the
reinforcement /12/. Above all, this danger occurs if larger por-
tions of polyvenylchloride have been involved in the fire, and it
should be taken into account that long term damages can be initia-
ted.
Spalling, mellowness, brittleness, cracks, deformations of the

structure
Spalling as described in chapter 2 has to be expected in most ca-

ses, to a lower or higher extent.
Mellowness of concrete, brittleness and falling away of concrete

is affecting mainly the corners of concrete cross-sections.
It should be looked after residual cracks which may have formed

not only in the tensile zone of the members.
Residual deformations are mainly dependant not only on the inten-

sity of the fire attack but on the statical system as well/10,17/.
The decision to which degree they may be tolerated must have re-
gard to this system.
REFERENCES
/1/ Abrams, M.S. und Cruz, C.R.: The Behaviour at high tempera-
ture of steel strand for prestressed concrete. Portland Ce-
ment Ass., Skokie, Ill., 1961.
/2/ Abrams, M.S. und Erlin, B.: Estimating post-fire strength
and exposure temperature of prestressing steel by a metallo-
graphic method. Portland Cement Ass., Skokie, Ill., 1967.
/3/ Arnault, P., et al.: Rapport Experimental sur les Essai s
avec des Feux Naturels executes dans la petite Installation.
/4/ anonym: Assessment of fire-damaged concrete structures and
repair by gunite. Concrete society, Techn. Report Nr. 15,
London, 1978.
/5/ Bechtold, R.: Tragverhalten von AuBenstUtzen unter Brandbe-
anspruchung. Bericht des Instituts fUr Baustoffe, Massivbau
und Brandschutz der Technischen Universitat Braunschweig,
1977.
/6/ Bechtold, R., Ehlert, K.P. und Wesche, J.: Brandversuche an
einem zum Abbruch bestimmten, viergeschossigen modernen Wohn-
haus in Lehrte. Bericht des Instituts fUr Baustoffe, Massiv-
bau und Brandschutz der Technischen Universitat Braunschweig,
1977.
/7/ Haksever, A.: StahlbetonstUtzen mit Rechteckquerschnitten
bei natUrlichen Branden. Institut fUr Baustoffe, Massivbau
und Brandschutz der Technischen Universitat Braunschweig,
Heft 52, 1982.
130 1emperatures
/8/ Hanmathy, T.Z.: A new look at compartments fire, Part I,

Fire Techn. Vol. 8 No. 3, 1972.
/9/ Kordina, K., Krampf, L.und Meyer-Ottens, C.: Das Brandver-
halten von Holzbauteilen. Holzbau-Taschenbuch, Verlag
w. Ernst u. Sohn, Berlin, 1974.
/10/ Kordina, K., et al .: Jahresbericht 1975/77 des Sonderfor-
schungsbereichs 148 "Brandverhalten von Bauteilen". Tech-
nische Universitat Braunschweig, 1977.
/11/ Kordina, K. und Meyer-Ottens, C.: Beton-Brandschutz-Hand-
buch, Betonverlag, 1981.
/12/ Locher, F.W. und Sprung, S.: Einwirkung von salzsaurehalti-
gen PVC-Brandgasen auf Beton. beton, Heft 2, 1970.
/13/ Meyer-Ottens, C.: Brandverhalten verschiedener Bauplatten
aus Baustoffen der Klassen A und B. Die Bauwirtschaft, Heft
6 - 8, 1969.
/14/ Meyer-Ottens, C., und Kordina, K.: Brandverhalten von Bau-
teilen aus dampfgehartetem Gasbeton. Fachverband Gasbeton,
Wiesbaden, 1974.
/15/ Meyer-Ottens, C.: Zur Frage der Abplatzungen an Betonbautei-
len aus Nonmalbeton bei Brandbeanspruchung. Deutscher Aus-
schuB fUr Stahlbeton, Heft 248, 1977.
/16/ Schmidt, E.: Uber die Anwendung der Differenzenrechnung auf
technische Aufheiz- und AbkUhlprobleme. Beitrag zur techni-
schen Mechanik und technischen Physik (Foppl-Festschrift),
Berlin, 1974.
/17/ Schneider, U. und Haksever, A.: Bestimmung der aquivalenten
Branddauer von statisch bestimmt gelagerten Stahlbetonbautei-
len bei natUrlichen Branden. Technische Universitat
Braunschweig (1976), Heft 46, CIB W14/78/1 (0).
/18/ Schneider, U. und Haksever, A.: Warmebilanzberechnung fUr
Brandraume mit unterschiedlichen Randbedingungen, Teil I.
Technische Universitat Braunschweig, CIB W 14/80/55 (D).
/19/ Schneider, U.: Effect of fire on concrete. RILEM-Committee
44-PHT, Draft, 1980.
/20/ Twilt, L.: European recommendations for fire Safety of buil-
ding structures. Part II, Delft, Marz 1975.
/21/ Roitman, M.: Grundlagen der Brandschutz-Nonmierung im Bauwe-
sen. Berlin, 1972.
T[°C) a[mmJ
1200 120
100
~ 80
:J
0
~
a. £
~ 60 a.
<II
"0
~
Ol
.!;;
40 L..
L..
t 0
8. £
u
E
8 20
0 0 30 60 90 120 150 100 t [min)

standard fire duration
Fig. 2.1-Charring depth of timber exposed to the ISO 834

Standard fire
1.0
-
0.8
.....~- 0.6
"' 1\
"\
.,J'
1 0.4 \
~
Vi 0.2 1\
0
0 200 400 600 BOO T [oc)
critical steeltemperature
Fig. 2.2-Critical temperature of constructional and reinforcing

steel in dependence on the stress level
132 Temperatures
T [°C] a [mm]
1200
_. l-- >--
lOCO
rV
.... 800
C1J v ...
~.... I "E
C1J
0.
E 600
I
I /
v-- chip -board
v-- gypsum plaster- board 30
E
~
VI 25
15
~ at/ foz !JlC1J
E 400 c
20 -"'
t
8. I v 15 £
-~
E
8 200
1
/ I 10
~- 5
0
0 :l> 60 90 lal l!XJ 180 t {min]
Fig. 2.3--Burning through of chip boards (a 1 ) and brittleness of

gypsum plaster boards (a 2 ) exposed to the ISO 834 Standard fire
a[mm]
---
120
1000 T..,....... 100

/ hemalrte - H 2 0~
~ !OJ
v Ca(OH) 2 -H,O
ct-ll-quarlz·~
80
e /
C1J
0.
E I ,. /
x
_,-"
t
"0
~
-g
600 I temperature
f - - within /
3500<:
cross -section 4&:/'t

v L- 60 ~
~
~ L{1J 7 ~- --- 40 0
8.
E
/ / ~~3oC
8 200 // / /
.~~
---
20
/ '// ,/
.,. 756oc ~~~~~wness
0
/h / /
0 30 60 90 120 150 180 t [min]
Fig, 2.4--Physicochemical transformations due to temperature

(ISO 834 Standard fire) in different layers of a concrete section
rr
ignition phase
bumrg phase cootirg down phase
t
- - - fire duration
Fig. 3.1-Development of a real fire
---1------ _ A I A =p =0,25
ISO 834
~ 1000 ·················
..Y ······················
~
~
C1l
BOO
0.
E
2
i
~
600
60 kg/m 2
E
8
~ 400 --- 30 kg/ m2--
0
Q;
~
200 __ 15 kgL~-!--------!----
00 30 60 90 120 150
- - - - - - fire duration I min I
Fig. 3.2-Temperature-time development of natural fires (wood cribs)

compared to the ISO 834 Standard fire (Metz-tests)
134 Temperatures
1100 t-------J----t---J-----+---+---j
7()()1-----
0 20 30 40 50
- - - - fire duration t [ min l
Fig, 3.3-Temperature-time development of natural fires (wood cribs)

compared to the ISO 834 Standard fire (Metz-tests)
~1000 openong in the southern 'MJII

'\\ speed of wind - 0,5 m/sec
C1l from south
5 800~--------~~~--~~~~~~~------~
~
~ 600
.9!
15 30 45 60
fire duration I min J
Fig, 3.4-Temperature-time development in a compartment; Lehrte-tests;
fire load 30 kg wood/m2; ventilation factor 0.30
(JlOOO openong on the southern wall
~
0
/1 speed ot Wind 110rying about
,~1 4m/sec from south - Past
I•
((/ II
1\
1\
,, .\
t! \.
,,rlrv ~ ...... , l.---

'-.;, 1\.....--
bottom
Cl.'iling
I/:'
400
I
I
I half heoght .......
200
~·-·
ui
I
·-·-.":::.;.;::;;
30 60 90 120
Fig. 3.5-Temperature-time development in a compartment; Lehrte-tests;
c;woo ,.. openong on thl.' southffn wall
\I
0 I
spl.'ed of wond \Oryong about
~C1l
C1l
.... 800
/
·"
. /
-\
4 m/sec from west
r'
~ 600
i - I\
~ I
I I( \...--"
~ '<..bottom
'·
400
!
I '·-:::---- 1---
200
!
Tf
ceilong ' · - .
·- ................
30 60 90 135 180
Fig, 3,6-Temperature-time development in a compartment; Lehrte-tests;
136 Temperatures
:EI200
... 1000
800
I~
\
f
'I/ /
I
r;,
/
I I
v
. . . . ·n
100------1-1
I
r----..
~ "" .\
----.::---\
~I\
\ I /
/
"' \
"
> ]00 ~ 0
t.OO
I I'.. .._
I 1---
Jl
\ '\ .:.; I
200 [\' I" v~~ ,.r- 0 6
0
1/
0 15 30 45 60 90
,\ "'' t'...: I'- v/~d' IJ
±>~~_:,p)
heat exposure I (min]
square section 200°
~
]00
IL
Iii w-
so•
tOO
-150
_,
''
H
IS •20
I
'
• I I I I I
~
-200 !'..
§~
- JOO
-400
- II v I'-.
75 500
1/
Ill/
t: ISO
1\
-
_TI
';: 8
.. ·-
0
"'!::
.,_
~
lw
ISO
-·zoo
JOO
-p.
-
It
T -section
300/500/160
70 0
100
I\
square section 300

.
, ..:---
-150-l--
<00 ]QQ~
°
I/
]
I· II II 01~01 ••• I I l
Fig. 4.1-Maximum temperature occurring in different cross sections
exposed to a fire like: fire load 30 kg wood cribs/m2; ventilation
factor 0.60; high heat insulation of surrounding elements {Metz-tests)
Kramp£ and Haksever 137
, ..
...
000
I I I I t \ChlO
I
::-.:::
000 /:
'// v r-...,
" "'
I~
-·
400
7 \
t\
I
I
I
/
1/
1\
v
' \
I
""' 1\
I
1
TI
I \__ t-... \ I'-.
.f
...-'~' 17
/
...... ,\" '
J
v
11
.........
Jl
0 30 60
I---
90
---
~
square section 200

~ 1->oo-
--,..'-no-
°
//I/
_....__._,~
heat exposure [min)
I
' - ,~
-f---t-+--"-+-+-+~>7•,• I I I I I I I
- ·100
150 I--
-~
'"" I/
I/
- ., ..
- "
II -~ 8
n
- I\
- 75 --
- 3=J ...
110 n•c-
100 - ,..
-~~~-r-
~ -
- - T- section 10
'
::1*" - lOO ))Q/500/160 square section J:X> 0

exposed to fire like: fire load 30 kg wood cribs/m2; ventilation
factor 0.60; high heat insulation of surrounding elements (Metz-tests)
138 Temperatures
...
1300
I ZOO
_,--. t=H--t--+ ~ IQ I I I I
I \ /.
I \ 'l/ v v t::=,1=::: r--:::: ~ "','\

_____
'I v 1\
1\ ..........
'I
II
1/
> '"'
" \ 1\\.\' -
'"' 1---- \ \ J 1/ I
\\ 1'--. / '!J
""' .\I". i'-<Tc-'
I'-.. v
oU0 ,"-. I"' 1::-- ~~ r-:::: ? V·;.
~ ~700
15 30 45 60 90 ;..,
heat exposu-e I [min]
square section 200 O
L .....100
' I· I I I I I ..!~;_,,~1+1-tl_,l~!+t
~
I
100 c;:.
~
f'lv
K
400
500
~
600 11 \'
700 .
\
I
l
~".#
~ .c:-- 150
1~~~
200 I'-..
~I\: t--...
r;:::
700
- 300
T- section
~00 1100
300/500/160 square section D:l 0
t~ "
< I I I 115-::nltJI I I I ~ I t
exposed to a fire like: fire load 60 kg wood cribs/m2; ventilation
factor 0.25; high heat insulation of surrounding elements (Metz-tests)
--~----200,-----41
~--+--+10•~ I J I I
I
.... 800 1/1\ ttr;v

600
\ I
{\
I
-
"' ' ............
r--...._
- r--
\
n
0
0 30 60 90 135 180
[min]
heat exposure t
square section 200 °
I!-
1-1" J
'
•
I
+-<lf-+-1__._.,.,__._. 100--------+,
-+•••• ~. • • • • •
~ M- ISO r. ~
zoo v
we lOO
'00 ~ v ......
§tf -
1/ 1/
ll
110
1\
A- •o•
I./ 1/
li ~- - - I'-
- --
100 ' v
150 - f-1-=-l~ ~
~ - 700
T- section
!~ - 500
300 /500 /160 square section 300°
, ••• t I 11~10111
' I I 'I
Fig. 4 .4-Maximum temperature occurring in different ·Cross sections
when exposed to a fire like: fire load 60 kg wood cribs/m2; venti-
lation factor 0.30; high heat insulation of surrounding elements
(Metz-tests)
140 Temperatures
I ' t t I
t t I "'"
lDO
j
I ~,-,--,--,-,-~~.--,-,--,-,~~
?:
~ v '::;>-- ·--c-,:::: '~
If' \ llt~~~ -1-f-- r=:::f01\\
v
f
f- -\ - --
- . -1- /17 ""1\ \
·-- -~
b)'
1':
-~!!
,, ,_ -~ ~
-- --- - - -
- 1- t-- - -- - 1- - - -1 - - -
'- :::::.1 = I~ Fc100'C /
-
1\ J, / V;
--
!10
/
'- ,&, l-J \\~ ~ i-ldo- I-- ~ ' / 1/J
100 [\_~ != ~~~-
ooo-- 1=::: ~ r.:::: /
exposure 15 min 30 min
100
I 1L \\
I '"I 20~
If' v- \
(....-- 1------
~ -p....._ I-- r----:::
~
~·
.L t' .rl " """
TI
1-
1--- r---- -~/ 11
1/ L~<~' .1f.
----
" 1--
.:
/'~
,,<$
~>,; <J- ""'"- ~
I
\
{ /
\ \ 11-- 8- 1)~-
1- 1- -- ·--1.-
lp
rs \:: v
li
'1- -
/
\ ['-..'-- __.-' /
- I
~
1--
1--- v / 1\
·""
-
-
I/ D IJ.
'-- " !"-. ____, /
60 min
_) ::::: - ~<O'C-
llO
1-
----~
90 min
Fig. 4.5-Tempcrature fields within a squnre section 200 exposed

to a fire like Fig. 4.1
1.0
.c 0.8
OJ
c
<lJ
L.
1i) 0.6
0::J sudden
u cooling down
.iii
<II
L.
0.4
....0
.Q
0 0.2
L.
0 T[°C]
0 200 400 600 800
temperature reached before
Fig. 5 .1-l{esidual compress:i. ve strength of normal dense concrete :i.n

dependence on the temperature reached before and on the cooling
conditions
!}) 0.8
!})
<lJ
.....
lil
0.6
~
<lJ
>.
0
::J
0.4
u
.iii
<lJ
L..
0 0.2
0
0
.....
0 T[°C)
0 200 LIJO 600 800
temp!:'rature rf:'ached before
Fig. 5.2-Residual yield strength after heating of cold-worked
reinforcing steel; exposure time to the maximum temperature 30
minutes
142 Temperatures
1.0
.!:
.,_. ~~
~ 0.8
4.>
.....
1il
~ 0.6
\\
8~ ~
Ill
c
4.>
.,_.
0 04
~
:J
'0
iii
4.>
.....
0 0.2
0
0 200 400
temperature reached be fore
Fig. J,3-ResidunJ tensile strength after heating of high tensile

prestressing steel (US-American); exposure time to the maximum
temperature 1 or 8 hours
0.8
(Jl
c
'iii
Ill
4.>
0.6
.....
iil
4.>
.....
Q_
0.4
'+-
0
0
.,_.
0
.....
0 200 400 600 800

temperature reached before
Fig. 5.4-Decrease of prestressing force under elevated temperature (a)
and after cooling down (b)
SP92-8
Residual Properties
of Concrete Heated Rapidly
by K. D. Hertz
Synopsis: The paper describes a new technique developed by the au-

thor for heating concrete rapidly by application of microwave po-
wer. Using an appropriate thermal insulation it is found to be
possible to heat dry concrete specimens 10°c per minute and 20°c
per minute avoiding the development of thermal stresses within the
specimens. A series of 90 specimens has been heated 10°c per minute
to various maximum temperature levels. The residual compressive
strength was measured, and the results are compared to the similar
results of the same concrete heated slowly. Applications for the
technique are outlined, and a possible future development is pre-
sented.
Keywords: compressive strength; concretes; fire tests; heating;

high temperature tests; microwaves; thermal insulation
143
144 Hesidual Properties
K.D. Hertz for a decade has made research on fire exposed concrete
structures at the Institute of Building Design at the •reclmical
University of Denmark, and has simultaneously been active as con-
sulting engineer designing structures in Denmark, Greenland and
Saudi Arabia. He is a M.Sc. Ph.D. Struct. Eng.
INTRODUC'riON
Since World War II the development of concrete structures in gener-

al has been towards more slender cross-sections using less concrete.
This development is beneficial from an economical point of view for

the builder, but it also makes the structures more susceptible to
various influences of the environment and among these to the effect
of fire.
Slender cross-sections are heated more rapidly and to higher tem-

perature levels than heavy massive cross-sections, and this is one
of the reasons why modern concrete structures often are difficult
or impossible to repair after fire exposure.
If fire damaged concrete structures should be restored, it is nec-

essary to know the mechanical properties of the concrete and espe-
cially this is a problem for the rapid heated concrete of the slen-
der cross-sections.
A concrete specimen intended for mechanical testing should be of a

certain size compared to the size of the aggregates, if the materi-
al in the specimen should be representative for the material used
in the structure.
In Denmark cylinders of diametres of 100 mm and 150 mm are stand-

ardized and similar measures are used in other countries.
Heating such specimens in a normal oven at a heating rate of about

10°C per minute, large internal stresses will develop and most
likely damage the concrete before it is tested.
Of course, international stresses are also present in the rapid

heated structure, but they will depend on the geometry of the
structures and therefore they must be taken into account by calcu-
lation or simple considerations. (llertz [ 4]) .
Bacause of these test problems, it has until recently been impos-

sible to find the mechanical properties of rapid heated concrete.
By means of microwave power it is now shown that it is possible to

heat concrete specimens 10°C and 20°C per minute at l000°c without
introducing large temperature differences or thermal stresses.
Hertz 145
This supplies us for the first time with information on the resid-
ual compressive strength of rapid heated concrete.
The testings show a difference in residual strength of rapid and

slow heated concrete. And the difference becomes a maximum about
300°c where it is 20%.
Since especially the temperature level 300°c is found in large

parts of many fire exposed slender cross-sections, the deviation is
essential to the calculation of the residual load carrying capacity
of slender concrete structures.
This paper describes the heating method and the test results and
possible new applications for the heating method.
DIELECTRIC HEATING
Suppose a material consisting of molecules, which can be consider-

ed as electric dipoles, and suppose an electric field is applied to
this material, the molecules will now attempt to orient themselves
in the direction of the applied field.
If the field oscillates the dipoles will tend to follow the oscil-
lations. But if the field oscillates sufficiently rapid, which is
approximately 1 GHz, the hindrance of the movement due to the mate-
rial structure will cause a considerable heat development.
These quick electric oscillations having a wave length of no more

than about the size of the electric circuit are called microwaves.
Some materials cannot be heated by means of dielectric power either

because they do not consist of dipoles or because their electric
phasedisplacements are too small.
Such materials are for example benzene, carbon, glass and porce-
lain.
Other materials such as water and many organic compounds are easily
heated by microwaves.
The microwave field can be applied to the material by means of two

conductors or by placing the object to be heated, which is called
the load, in the cavity of a microwave oven.
The walls of the cavity are all made of metal in order to reflect
the microwaves produced by one or several magnetrones, and a metal
stirrer can be used to disperse them and ensure a homogeneous power
distribution.
As microwave-heating until now mostly has been utilized in cooking

and biological research, only little information is available a-
bout concrete.
146 Residual Properties
However, some investigations are made on the dielectric constants

of building materials by Hasted and Shah [ 1], especially concern-
ing the influence of the moisture content.
It is obvious that the concrete is more susceptible to microwave

power when the moisture content is high, but extrapolation of the
values indicates the possibility of heating a dry concrete.
From the extrapolated values the power distribution in a cylindrial

specimen was calculated in Hertz [ 2].
Assuming a thermal insulation on the surface of the cylinder, the

thickness of this insulation could be determined, so that the heat
loss from the hot spesimen to the cold oven will answer a tempera-
ture distribution in the cylinder, which is almost the opposite of
the distribution caused by the damping of the penetrating micro-
waves, and the final superimposed distribution would be almost
uniform.
By calculation it was therefore concluded, that it seemed possible

to eliminate temperature differences in concrete specimens almost
completely during fire tests by using microwave power combined with
a regulation of the surface heat loss by using an insulating mate-
rial. (Figure 1) ,
MICROWAVE TESTINGS
At the very first heating in 1982 a concrete cylinder of diametre

10 em and height 20 em was heated 10°C per minute with an almost
uniform temperature distribution in a 2,15 Gl!z microwave oven at
600 W until the temperature of the specimen was about 900°c, and
the radiation from the surface of the insulation was likely to ig-
nite the plastic ceiling of the cold oven cavity. (Hertz [ 3]).
At the second heating the specimen was heated 20°C per minute at
1200 W until B00°c, and still the temperature distribution was al-
most uniform. ( I'igure 2) .
'l'his proved that the method was advantageous in many respects.
A series of 90 cylindrical concrete specimens was heated 10°C per

minute to various maximum temperatures, and their residual compres-
sive strengths of the same concrete heated in previous testings 1°C
per minute to the same maximum temperatures. Figure 3).
The cylinders had a diameter of 10 em, a height of 20 em and were

provided with two circular sheets of aluminium foil at the ends in
order to create the same uniform penetration conditions for the mi-
crowaves by reflection as if the cylinder was of infinite height.
As thermal insulation 5 em ceramic fibre wool was used, because

this material was not susceptible to microwave power.
Hertz 147
The concrete was based on Danish sea gravel (a mixture of limestone,

sandstone and granite), and it had the initial compressive strength
at the time of heating 20 MPa and a water/cement ratio of 0.87.
Thus, it was porous and likely to decompose when exposed to high

temperatures, and the values of the residual strength of this con-
crete are used as a reference making a conservative design. Hertz
[ 4].
In addition, the relative high porosity diminishes the risk of

steam explosion that otherwise could be a problem due to the rapid
heating.
For the same reason the 14 days old concrete specimens were kept
for 3 days at 105°c before heating in the microwave oven.
The specimens were heated 10°C per minute two by two at 1200 W and
2.45 GHz the time necessary to achieve the maximum temperature
wanted.
At the end of heating the insulated specimens were placed in a box

insulated by 15 em mineral wool to ensure a slow natural cooling,
and thereafter they were stored one week at 20°c and 65 pet. rela-
tive humidity prior to the compression test in order to ensure that
the strength of the concrete was a minimum at the time of the com-
pression test.
From the test results it can be seen that the residual compressive
strength of rapid heated concrete decreases almost linearly as the
temperature increase.
This straight line curve is quite different from the typical S-

shaped curve, which is the result of a slow heating rate for the
same concrete. (Figure 4).
At maximum temperature of 200°C the residual compressive strength

of rapid heated concrete is about 10 pet. less than the slow-heated,
20 pet. less at 300°C and about 10 pet. more at 600°c.
A possible explanation to these observations can be deduced from

the nature of the processes decomposing the concrete, while it is
heated.
The free water and from approximately 150°C the water of the hy-
drated calcium silicates evaporates, and consequently the hydrated
cement paste shrinks.
At the same time the aggregate expands, and at 300°c the stresses
between the aggregate and the matrix are so large that microcracks
develop through the matrix.
Heating the concrete rapidly the stresses are introduced much

quicker, and they are not reduced by creeping of the matrix to the
same extent as possible at slow heating rates.
Therefore, from 220°C to 400°C a more destructive cracking takes

place, and thus the residual strength is less than the one of a
slowly heated concrete.
From approximately 400°c the calcium hydroxide in the matrix will

decompose, and while cooling and curing within the seven day peri-
od after the heat exposure the residual calcium oxide absorbs wa-
ter from the air and expands, and the expansion widens the cracks
of the matrix.
This is the main reason why the residual strength drops about 65
pet. from 400°c to 600°c for slow-heated specimens.
However, the dehydration of the calcium hydroxide as well as the

water absorption of the calcium oxide needs time to take place.
By heating at 10°C per minute it seems that the process does not
develop fully and the residual strength drops only about 35 pet.
from 400°c to 600°c.
The fact that the rapidly heated concrete has a lower residual
strength than the slowly heated at maximum temperatures from 200°c
to 400°c is often critical to the structural use of the test re-
sults. Large parts of the cross-sections of a concrete structure
reach maximum temperatures of this level by fire exposure, and the
parts achieving higher temperature levels will have strengths,
which are often negligible for calculation of the load-bearing ca-
pacity of the structure.
This means that the use of test results of ordinary slow-heated

specimens may give rise to an error of 20 pet. in estimating the
load-bearing capacity of a structure exposed to a fully developed
fire course, and perhaps even more in case instability phenomena
are decisive for the estimation.
By using microwave heating, the specimens can be heated rapidly

and consequently these errors are avoided.
Microwave heating is not only advantageous for determination of

the mechanical properties of rapid heated concrete, but also the
specific capacity of heat can easily be determined.
Hitherto the usual procedure has been to measure the heat loss of
a heated specimen during a slow coolinq in a calorimeter.
By heating an insulated concrete spec~nen in a microwave oven, all

the output power of the magnetrones is used for heatinq the con-
crete and the oven cavity remains cold.
Above 105°C 600 W output power is found to heat a cylinder of dry

weight 3.5 kg exactly 10°C per minute.
From these data the specific capacity of heat will be 1.03 kJ/kg°C.
Hertz 149
This confirms that a reasonable value to be used in calculations

of the temperature distribution of fire exposed concrete sections
would be 1.0 kJ/kg 0 c, when used related to the density of the dry
concrete, and the heat of the evaporating water is treated sepa-
rately.
By means of microwave power the specific capacity of heat is de-

termined for concrete heated at a realistic heating rate in stead
of slowly cooled concrete.
Taking into consideration that the processes of dehydratisation of

the cement paste needs time to take place, and that the reversed
processes probably do not occur at the same temperature levels and
to the same extent by cooling as by heating, the value found by
heating must be the most correct for the purpose of calculation of
the temperature distribution of a concrete cross-section heated by
fire.
FU'rURE APPLICATIONS OF MICROWAVE HEATING
One obvious possibility is to establish a silmuntaneous heating

and compression, which is essential to the study of the mechanical
properties of concrete at transient conditions, i.e. heating and
loading at the same time.
The transient properties of concrete are to be used for the esti-

mation of the deflection of fire exposed structural members of
concrete, force distributions of hyperstatic structures and resid-
ual prestressing forces and much more.
The simultaneous heating and compression can be achieved by using

the pistons of the compression test machine as conductors applying
the microwave power directly on the concrete specimen. (Figure 5),
This direct dielectric heating by means of conductors was one of

the first applications of microwave power, so the technique is
well known.
The concrete specimen should be surrounded by a metal cabinet so

that the microwaves are not dispersed into the laboratory, and the
cabinet should be provided with a thermal insulation in order to
diminish the heat loss from the hot specimen and to avoid tempera-
ture differences in the specimen.
Naturally the direct dielectric heating is suitable for testing of

concrete in biaxial as well as monoaxial compression at high tem-
peratures, and very likely this is the simplest heating method for
transient testings even at slow heating rates.
Microwave power is not only a possibility for rapid and slowly

heatings at constant heating rates.
By means of an automatic ON-OFF control of the magnetrons the con-

crete can be heated by almost any heating curve.
A PID regulated ON-OFF control unit could be used for measuring

the temperature of a reference specimen and comparing this meas-
ured temperature to the temperature wanted at the time.
'l'his set point temperature could be given as u curve drawn by hand

and read optically or simulated by a mechanical system or calcul-
ated by a programmable computer.
By these procedures almost any fire course can be followed provid-

ed that the capacity of the oven is sufficient for the maximum
heating rate, and that the possible rate of heat release from the
specimen avoiding unacceptable temperature differences is suffi-
cient for the maximum cooling rate.
Still, the risk of explosive spalling should be considered at the

fastest heating rates.
REFERENCES
( 1) HASTED 1 J . B. SHAH 1 M.A. :

Microwave Absorption by Water in Building Materials
Brit. J. Appl. Phys., Vol 15,
Number 7, pp. 825-836. July 1964.
[ 2) HERTZ, K.D.:
Microwave Heating for Fire Material •resting of Concrete
- A Theoretical Study.
Report No. 144.
Institute of Building Design,
Technical University of Denmark.
Lyngby 1981. 25 p.
CIB W14/81/3 (DK).
[ 3) HERTZ, K.D.:
Microwave Heating for fire Material 'l'esting of Concrete
- An Experimental Study.
Report No. 164 .
Lyngby 1983. 38 p.
CIB W14/83/19 (DK).
[ 4) HERTZ, K.D.:
Analyses of Prestressed Concrete Structures Exposed to
Fire.
Report No. 174.
Lyngby 1985. 154 p.
CIB W14/85/9 (DK)
Hertz 151
•c
600
400
200
Fig, !--Temperature distributions in concrete cylinders of diameter

150 mm heated 10 C per minute normally left and by microwaves right
0(
1000
600
600
400
200
10°( per min.
oc
~
1000
BOO
600
400
200
20°( per min.
Fig. 2--Measured temperature distributions in concrete cylinders

(100 mm) heated 10 C per minute and 20 C per minute normally left
and by microwaves right
10 20 30 40 SO 60 70 80 90 100 min.
Fig,3--Temperature rise in time of the final microwave heating

procedure
_h_
fcc20
1.0
& - micn:1-1Klve heated
0.9
:---normal heated
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
100 200 300 400 500 600 700 000 °(
Fig. 4-Residual ultimate stress of concrete wi.th Danish sea gravel

heated by microwave power 10 C per minute and heated normally 1 C
per minute
Fig. 5-Simultaneous compression and heating usi.ng the pistons as

conductors
SP 92-9
Residual Strength of Fire-Exposed

Reinforced Concrete Columns
by 'f. T. Lie, T. J. Rowe, and T. D. Lin
Synopsis: A study was carried out to assess the residual strength

of reinforced concrete columns after exposure to a standard fire
for various lengths of time, and cooling. The use of a
mathematical model, an ultrasonic pulse test method and a load test
method are investigated. Calculated temperatures and residual
strengths of test columns were compared with those measured.
Comparisons were also made between calculated and measured pulse
velocities. The results indicated that using the calculation
procedure and the method of measuring pulse velocity described in
the study, the residual strength of concrete columns can be
assessed with an accuracy sufficient for practical purposes.
Keywords: columns (suppor~; compressive strength; deformation;

fires; fire tests; loads (forces); reinforced concrete; strength;
temperature; ultrasonic tests
153
154 Reinforced Columns
T.T. LIE is a research officer with the Division of Building

Research, National Research Council of Canada. He worked for
several years in Europe and Japan before joining the NRC in 1967.
He is currently studying structural fire resistance, which includes
testing and calculation of the fire resistance of structural
members. T.J. ROWE is Manager, Fire Research Section, Construction
Technology Laboratories, a division of the Portland Cement
Association, Skokie, Illinois. He is involved in investigations of
structural problems and failures, large-scale structural and fire
testing, nondestructive testing of concrete, and evaluations
involving fire-damaged structures. Rowe is a member of ACI
Committee 216, Fire Resistance and Fire Protection of Structures.
T.D. LIN is principal research engineer with the Construction
Technology Laboratories of PCA. His work involves theoretical
analysis, design and fire tests of full scale structural elements.
He has 15 years experience in the study of reinforced concrete
structures exposed to fire.
INTRODUCTION
The residual strength of fire-damaged concrete structural

members is an important factor in determining the feasibility of
repair of the structure. The restoration of load-carrying capacity
and fire resistance of a fire-damaged structure to an acceptable
level is required by relevant building codes. Often, repair is
more economical, both in terms of cost and time, than demolishing
and rebuilding the structure.
To assess the residual strength of fire-exposed reinforced

concrete columns, studies sponsored jointly by the National
Research Council of Canada (NRCC) and the Portland Cement
Association (PCA) were recently carried out. Both theoretical and
experimental studies were performed. These studies include:
• Development of a mathematical model to calculate the temperatures

in the columns, and their strength, during and after exposure to
fire.
• Subjecting loaded test columns to a standard fire for various
lengths of time, and cooling of the columns in an environment of
standardized decreasing temperatures; measurement of column
temperatures during and after exposure to fire until near ambient
temperatures are reached in the column.
• Comparing the measured internal temperatures of the column during
and after exposure to fire with calculated temperatures.
• Applying ultrasonic methods of nondestructive testing to monitor
changes in concrete material properties of the columns before and
after exposure to fire; evaluation of relationships between
measured pulse velocities and concrete quality, compressive
strength, and modulus of elasticity.
• Determining the ultimate axial strength of the columns after
exposure to fire, under conditions of near ambient internal
concrete temperatures.
Lie, Rowe, and Lin 155
• Comparing measured residual strengths of the columns with

calculated values.
In the current studies, two 305 x 305 mm reinforced concrete

columns made with siliceous aggregate were investigated. The
columns were constructed by the Construction Technology
Laboratories of PCA and tested in the laboratories of the Division
of Building Research of NRCC. The measured results of these tests
and the calculated results are discussed in this paper.
CALCULATION PROCEDURE
The analysis of the performance of a column during and after

exposure to fire involves the calculation of temperatures,
deformations, and the stresses of the column. The calculation
procedure is described in detail in Reference 1 and will not be
further discussed here. Information is provided with regard to the
environment temperatures and the properties of the concrete and the
reinforcing steel during and after exposure to fire.
Environment Temperatures
As in the previous study [1], the temperature course of the

environment, i.e. the fire to which the column is exposed, is
assumed to follow that of the standard fire described in ASTM E-119
[2]. After exposure to fire the temperature of the environment is
assumed to decrease to room temperature according to the relations
specified in ISO 834 Standard [3]. Depending on the duration of
exposure to the fire, the rate of decrease of temperature is as
follows (Fig. 1):
dT/dt for th ,.; 0.5 h (l)
dT/dt 250 (3 - th) °C/h for 0.5 < th < 2h (2)
dT/dt 250°C/h for th ) 2 h (3)
dT/dt 0 i f T = 20°C (4)
where
t the time in hours,

th the duration of the fire in hours,
T the fire temperature, at time t, in hours.
Properties of Steel and Concrete
In the calculations, the same equations were used for the

thermal properties and stress-strain relations of concrete and
steel as those described in Reference 1. However, the studies
described in Reference 1 were concerned only with the period of

rising fire temperatures. For periods in which the steel or
concrete contracts, as is the case during cooling, other
stress-strain relations were assumed, in which the stress reduces
with the strain according to a straight line that is parallel to
the tangent of the stress-strain curve at the origin (Figs. 3
and 5). The stress-strain curves for the reinforcing steel and the
concrete at elevated temperatures are illustrated in Figs. 2-5.
As can be seen in Figs. 2 and 4, steel and concrete lose

strength when heated to high temperatures. In the calculations it
was assumed that steel retains its original tensile strength after
cooling. The compressive strength of concrete after cooling,
however, is reduced. The reduction in strength of the concrete
depends to a high degree on the temperature to which the concrete
has been heated and to some extent also on the load to which the
concrete was subjected during heating. In the calculations,
approximate values derived from literature [4,5] have been used for
estimating residual compressive strength of siliceous aggregate
concrete. Because information on the dependance of the concrete
strength on load is meagre and there is considerable scatter in the
values given in the literature for the residual strength, the
influence of load was not considered. This tends to lower the
values of the estimated residual strength of the concrete. These
values can be given as a function of the maximum temperature
attained by the concrete, by the following relations (Fig. 6)
fr = (1 - 0.001 T) fco (5)
for 500°C < T < 700°C
fr = (1.375 - 0.00175 T) fco (6)
f = 0 (7)
r
where
fr residual strength of the concrete,

fco initial strength of the concrete,
T highest temperature attained by the concrete.
TEST SPECIMENS*
The specimens were square, tied, reinforced concrete columns,

made with siliceous aggregate. They were 3810 mm long and had a
cross-section size of 305 x 305 mm. Twenty-five-mm diameter longi-
*Detailed information on the test specimens is available.

tudinal reinforcing bars and 10-mm diameter ties were used. The
location of the main reinforcing bars, which were welded to steel
end plates, and the locations of the ties are shown in Fig. 7.
The yield stress of the main reinforcing bars was 444 HPa and
that of the ties was 427 Hl'a. The ultimate tensile strength was
730 HPa for the main bars and 671 HPa for the ties.
The concrete mix was designed for a compressive strength of

approximately 35 HPa. The mix proportions were as follows:
Cement 325 kg/m3

\~ater 140 kg/m3
Sand 874 kg/m3
Coarse Aggregate 1058 kg/m3.
The average compressive cylinder strength of the concrete of

the two columns tested, measured on the day of the fire test, was
38.9 HPa for Column A and 41.8 HPa for Column B. The moisture
condition at the centre of Column A was approximately equivalent to
that in equilibrium with air of 87% relative humidity at room
temperature, and of Column B, with air of 82% RH.
Chromel-alumel thermocouples, 0.91 mm thick, were installed at

mid-height of the columns for measuring concrete temperatures at
different locations in the cross-section.
TgsT APPARATUS
Fire Tests
TI1e fire tests were carried out by exposing the columns to

fire in a furnace specially built for the testing of loaded columns
and walls. The test furnace was designed to produce the conditions
to which a member might be subjected during a fire, with respect to
temperature, structural load, and heat transfer. It consists of a
steel framework, supported by four steel columns, and the furnace
chamber inside the framework. The characteristics and instrumen-
tation of the furnace are described in detail in Reference 6.
Ultrasonic Tests
Ultrasonic tests to measure the pulse velocity of concrete in

the column were made using a commercially available instrument,
known as a V-Heter [7]. The test arrangement utilizing this
equipment is shown schematically in Fig. 8. The V-Heter supplies a
train of sharp, electrical pulses to a transmitting transducer in
acoustic contact with the concrete surface. The electrical pulses
are converted to low frequency (54 kHz) mechanical pulses by
piezoelectric elements in the transducer assembly. The mechanical
pulses propagate through the concrete and are sensed by the
receiving transducer located at a known distance. The mechanical

energy is reconverted to electric signal by the transducer. The
electrical signals are subsequently amplified and displayed on an
oscilloscope. TI1e V-Meter utilizes a precise calibrated time delay
circuitry to calculate and display the transit time of the pulse
through the concrete. From the transit time and the thickness of
the concrete as measured as the straight line distance between the
two transducers, the pulse velocity can be calculated. This
velocity is a measure of the quality of the concrete, and is
related to compressive strength and dynamic modulus of elasticity.
TEST CONDITIONS AND PROCEDURE
Fire Tests
TI1e columns were installed in the test furnace by bolting

their steel end-plates to a loading head at the top and a hydraulic
jack at the bottom. Concentric loads were applied to the columns
about one hour before the fire tests. The load on Column A was
992 kN and that on Column B, 1022 kN.
During the tests the heat input into the furnace was
controlled so that the average temperature followed as closely as
possible the standard temperature-time relations described earlier
in the present paper. The accuracy of control during both tests
was such that the duration of the average furnace temperature from
that given by the standard temperature-time relations was less than
10°C, except in the first 3-5 minutes of the test and after the
furnace had cooled down to below l00°C. Column A was exposed to
fire for one hour and Column B for two hours before the start of
the cooling period. Measurements were made of the furnace
temperatures during the fire exposure and cooling periods until the
average furnace temperature reached near ambient temperature.
Measurements were also made of the concrete temperatures at various
locations during the fire exposure and cooling periods. During
these periods the load was kept constant and the axial deformation
of the column was measured until the temperatures in the column
reached values close to ambient temperature. At that stage, which
was reached about one day after the cooling period started, the
load on the column was increased at a rate of 12.5 kN per minute
until the column failed.
Ultrasonic Tests
Prior to the start of the fire test, over 90 pulse velocity

measurements were made through the cross-section at various loca-
tions along a vertical profile of the column. A layout of loca-
tions of the pulse velocity measurements is presented in Fig. 9.
Pulse velocity measurements were also made on companion

152 x 305 mm concrete cylinders of Column A and Column B. The
cylinders were fabricated at the time the columns were cast, and
then maintained at curing and storage conditions identical to those
of the columns. The pulse velocity measurements were made along
both the transverse and the longitudinal axis of the cylinders.
The cylinders were subsequently tested for compressive strength on
the day of the fire test.
Approximately 20 hours after the start of the fire test, pulse

velocity measurements were attempted at each of the test grid
locations identified in Fig. 9. At approximately 25 hours after
the start of the test, a region near mid-height of the column,
between Elevations F and G, was chipped to a depth of 1 to 3 in.
(25 to 75 mm) to remove loose, delaminated material. Chipping was
performed with hand tools. The resulting exposed shape of the
column was nearly round, as shown in Fig. 10. Pulse velocity
measurements were also made through the column in the region.
Rf<:SULTS AND DISCUSSION
Column Temperatures
Previous studies [1], in which several columns were tested,

showed that the mathematical model used for the calculation of
temperatures in siliceous concrete columns during fire exposure
gives reasonably accurate predictions. In general, predicted
temperatures were slightly higher than those measured, probably
because of conservative assumptions for the mechanism of heat
transfer from the furnace to the column.
The columns tested in the present study were made from the
same siliceous aggregate concrete. Using the same procedure and
thermal properties described in Reference 1, the temperature
history of the columns was calculated. There was again good
agreement between the calculated and the measured temperature
curves. Figures 11 and 12 present measured and calculated
temperatures of the concrete as a function of time for various
depths along the centerline of the column.
Ultrasonic Test Method
The use of ultrasonic test methods to evaluate material prop-

erties and defects in concrete is widespread [8]. These techniques
are based on detecting changes in amplitude, phase, and direction
of mechanical waves as they propagate through a concrete member.
Changes in wave characteristics generally indicate a corresponding
change in internal structural make-up of the concrete.
Pulse velocity measurements made prior to exposure to fire

ranged from 4360 to 4665 m/s over the height of Columns A and B.
The average measured reading was 4510 and 4560 m/s for Columns A
and B, respectively.
Results of pulse velocity tests and corresponding compressive

strengths for 152 x 305 mm concrete control cylinders are presented
in Table 1,
TABLE 1 Average measured pulse velocity and compressive strength

for 152 x 305 mm companion cylinders
Average Compress:l,ve
Cylinder Pulse Velocity Strength
Column No, (m/s) (MPa)
A 1 4570 37.7
A 2 4600 39.0
A 3 4615 40.0
B 1 4665 42.1
Relationships between pulse velocity and compressive strength

as determined by tests of cylindrical specimens have been compiled
by many investigators, It is important to realize, however, that
these relationships can vary significantly between concrete mixes,
due to differences in aggregates, water/cement ratios, air content,
moisture content, and density, Plots of pulse velocity versus
compressive strength from references 9 and 10 are presented as
Figs, 13 and 14, respectively. Figure 14 additionally presents a
curve identifying "fire damaged" concrete, Due to dehydration,
matrix restructuring, and micro cracking, there is a significant
decrease in measured pulse velocities for a given compressive
strength for "fire damaged" concrete compared to "undamaged"
material.
The relationship between residual compressive strength and

temperature exposure for siliceous concretes is presented in
Fig, 15. This plot, from Reference 4, indicates an approximate
residual strength of 70, 50, and 20% of original for concrete
exposed to 300, 500, and 700°C, respectively.
Calculations were performed to predict pulse velocity values

for Columns A and B exposed to a one-hour and a two-hour fire,
respectively. The following sequence was employed:
1. Maximum internal temperatures at various distances from the

exposed face were derived from recorded values.
2. Measured temperatures were related to anticipated strength for
each zone of temperature exposure within the column, based on
data presented in Fig, 15,
3, Strength values for each zone of temperature exposure within
the column were related to anticipated pulse velocity values
based on data presented in Figs, 13 and 14.
4, Utilizing the "total transit time" computational approach, in-
dividual pulse velocity values were combined for each tempera-
ture zone to calculate total, through-section pulse velocities,
Post-fire velocity values calculated utilizing the above

approach ranged from 1500-2000 m/s for Column A and from
1100-1500 m/s for Column B. These results compare favourably to
actual measured values of 1200-2000 m/s for Column A and
1000-1600 m/s for Column B,
Axial Deformation and Strength
Deformation The axial deformations measured during the tests

and the calculated deformations are shown for Column A in Fig. 16
and for Column Bin Fig. 17. Up to a test time of four hours,
there is good agreement between calculated and measured
deformations. After this time, the differences between calculated
and measured deformations become considerable,
A good agreement between calculated and measured deformations

was also found in earlier tests on reinforced concrete columns,
which were exposed to fire until the column failed [1], These
tests generally lasted from two to five hours and did not include a
cooling period. The good agreement for tests without a cooling
period and for the present tests in the first few hours may be
attributed to the fact that, for rising temperatures and strains of
the concrete and steel, the effect of creep is included in the
stress-strain relations used in the mathematical model, Because of
insufficient data and complexity of the model, the effect of creep
is not included in the stress-strain relations for decreasing
temperatures and strains of the concrete and steel, It is likely
that this, and shrinkage of the concrete, are the main reasons for
the considerable difference between calculated and measured
deformations after about four hours,
Strength The calculated residual axial strength of Column A

was 1725 kN and of Column B, 2470 kN, The strengths measured after
cooling of the columns to near ambient temperatures, were 1987 kN
for Column A and 2671 kN for Column B, Calculated strengths are
about ten percent lower than those measured.
Factors that may have contributed to the differences include

the somewhat conservative prediction of column temperatures and
neglect of creep in the cooling period, Implied in this relation
is the assumption that the residual strength of the concrete is
only a function of the maximum temperature attained by the
concrete, It is known that the residual strength is to some extent
affected by the load on the concrete during heating. Studies
[4,11,12] showed that the residual strength of concrete specimens
heated under stress was higher than that of specimens heated
unstressed, Above a critical stress of about one-third of the
compressive strength of the concrete, however, the residual
strength was no longer affected by the stress,
The neglect of the influence of creep and load in the

calculations is expected to result in a lower calculated residual
strength, It appears, however, that the neglect of the influence
of these factors does not introduce large errors. Thus the results
suggest that it is justified to consider the residual strength of a
concrete column that has been exposed to fire as principally
dependent on the maximum temperatures attained in the concrete.
In Figs. 18 and 19, calculated maximum temperatures reached at

various depths in the column are shown, as well as the residual
strength of the concrete at various depths after cooling.
Figure 18 shows the temperature and strength distribution for
Column A, which was exposed for one hour to fire, and Fig. 19 for
Column B, which was exposed to fire for two hours.
In these figures also the profiles of the sections are shown

after further chipping off of the weaker concrete with handtools a
few weeks after the test. The profiles are rather irregular, with
most of the concrete coming off at the corners. The concrete at
the corners separated from the column along vertical cracks and
came off rather easily in essentially triangular-shaped pieces.
The cracks were formed after the exposure to fire, and can be
attributed to the fact that the inner part of the column was still
increasing in temperature, whereas the outer part was cooling
substantially.
A comparison between the profiles of the remaining concrete

sections and the calculated isotherms indicates that a calculated
maximum concrete temperature of about 550°C is a reasonable limit
at which the concrete possesses significant residual strength.
Because calculated temperatures are about ten percent conservative,
the actual temperature at the locations where this limiting
temperature is reached, is about 500°C. At this temperature the
residual strength of the concrete, immediately after cooling down
is still about fifty percent of the original strength. After this,
the strength reduces somewhat with time for several days [4,13-15]
but the concrete will later rehydrate and gradually regain most of
its original strength (Fig. 20).
CONCLUSIONS
Various methods can be used to assess the residual strength of

fire-exposed reinforced concrete columns. In this study the use of
a mathematical model, an ultrasonic pulse test method and a load
test method were investigated. The results of the study indicate
that:
1. The column temperatures predicted by the model are in good

agreement with those measured, although the predicted
temperatures are somewhat conservative.
2. The axial deformations predicted by the model are fairly
accurate for the first few hours, but deviate from those
measured in the cooling period. The deviation can be
attributed to creep, which is taken into account in the model
only for the period of rising column temperatures. The creep
did not significantly affect the residual strength of the

column.
3. Pulse velocity measured by the ultrasonic test method
correlated well with the residual strength and quality of the
fire-damaged concrete.
4. Calculated residual strengths of the columns are about ten
percent lower than those measured. In the light of the
complexity of residual strength evaluation, the agreement
between the calculated and measured strengths can be regarded
as satisfactory.
In summary, using the calculation procedure and the method of

measuring pulse velocity described in this study, the residual
strength of concrete columns can be assessed with an accuracy that
is sufficient for practical purposes.
REFERENCES
1. Lie, T.T., Lin, T.D., Allen, D. E. and Abrams, M.S., Fire

Resistance of Reinforced Concrete Columns, National Research
Council Canada, Division of Building Research, DBR Paper 1166,
NRCC 23065, Ottawa, 1984.
2. Standard Methods of Fire Tests of Building Construction and

Materials, ANSI/ASTM E119-83, American Society for Testing and
Materials, Philadelphia, 1983. ·
3. Fire-resistance Tests - Elements of Building Construction,

International Standard IS0-834, Geneva, 1975.
4. Abrams, M.S., Compressive Strength of Concrete at Temperatures

to 1600 F, American Concrete Institute, Special Publication
SP-25, Temperature and Concrete, Detroit, pp. 33-58.
s. Assessment of Fire-damaged Concrete Structures and Repair by

Gunnite, Concrete Society Technical Report No. 15, The
Concrete Society, London, 1978.
6. Lie, T.T., New Facility to Determine Fire Resistance of

Columns, Canadian Journal of Civil Engineering, Vol. 7, No. 3,
1980.
7. James Electronics, Inc., "Instruction Manual for Model C-4899

V-Meter", Manual No. F-0062, Chicago, May 1977.
8. Hanson, N.W., Rowe, T.J., Anstedt, D., and Corley, W.G.,

Vibration Techniques for Evaluation of Defects in Concrete,
Vol. 52, Productive Applications of Mechanical Vibrations,
Applied Mechanics Division (ASME), New York, 1982, p. 107.
9. Rowe, T.J., Evaluation of Concrete in Columns at the Parish

Station Unit No. 8 Cooling Tower, Report to Marley Cooling
Tower Company (Mission, Kansas 66201), submitted by

Construction Technology Laboratories, a division of the
Portland Cement Association, Sept., 1981.
10. Watkeys, D.G., Non-destructive Testing of Concrete Subject to

Fire Attack, M.Sc. (Eng.) Thesis, University of London, UK,
1955.
11. Malhotra, H. L., The Effect of Temperature on the Compressive

Strength of Concrete, Magazine of Concrete Research, Vol. 8,
No. 23, 1956.
12. Weigler, H., and Fisher, R., Beton bei Temperaturen von 100 C
his 750 C, Beton Herstellung Verwendung, V. 18, No. 2, 1968.
13. Zoldners, N.G., Effect of High Temperatures on Concretes

Incorporating Different Aggregates, Mines Branch Research
Report No. 64, Department of Mines and Technical Surveys,
Ottawa, 1960.
14. Weigler, H., and Fisher, R., Effect of Temperatures above

100 C on the Compressive Strength of Cement Mortar, Behaviour
of Concrete at High Temperatures, Bulletin 164, Deutscher
Ausschuss fur Stahlbeton, 1964.
15. Harada, T., Research on Fire Proof of Concrete and Reinforced

Concrete Construction, Tokyo Institute of Technology, 1961.
u
0
"'::>
1-
600
""w
"'
Q.
:;: 400
w
1-
200
0
0 100 200 300 400
TIME, min
Fig. !-Temperature of environment as a function of time during and

after fire for various durations (th) of the heating period
600
T • 20 o C
500
T • 200°C
ro
0..
400
:;:
T • 400°C
vl 300
w
"'..... T • 600°C
200
T • 800°C
100
0.02 0.03 0.04 0.05

STRAIN, E 5
Fig. 2-Stress-strain curves for the reinforcing steel at various

temperatures (yield strength = 443 MPa)
600
500
"'
"- 400
::0
T • 400'C
vi 300
....
Vl
"'
.....
Vl
200
100
0.01 0.02 o. 03 0. 04 0. 05
STRAIN, Es
Fig. 3-Stress-strain relationship for steel after decrease of strain
40
T • 20" C
30
"'
"-
::0
vi 20
....
V>
"'
.....
Vl
10
STRAIN,£c
Fig. 4-Stress-strain curves for concrete at various temperatures

(compressive strength = 35 MPa)
40
Ll NE
..
"-
::;;
30
~ 20
....
Vl
,_
"'
Vl
10
0.03 o. 05
STRAIN, Ec
Fig. 5--Stress-strain relationship for concrete after decrease

of strain
;!:
C)
z 0.9
ll;!
,_
Vl
..... 0.8
<(
z
C) o. 7
ii<
0
l5 0, 6
z
0
>= 0.5
(J
~ 0.4
,_
"'
C)
15 0. 3
"',_
Vl
0.2
~
e 0. I
....
Vl
"' 0
0 100 200 300 400 500 600 700 800
TEMPERATURE, "C
Fig. 6--Residual strength of siliceous concrete after cooling as a

function of the temperature attained by the concrete [4,5]
48 mm COVER TO MAIN
EINFORCING BAR
]305 mm
533 x 533 x 25 mm
u
305 mm
TIIICK STEEL PLAIT-
I<Zfflt--ZZffZ"'_j 50 mm
I--
f--
E
E
f--
~
f-- E
E E
E
s I-- ~
~ >--
<(
V>
I-- "'
<(
CD
u•
f-- >--
~
f--
f--
f--
Fig. 7-Test column and location of reinforcing bars
TRANSMITIING RECEIVING
TRANSDUCER
TRANSDUCER\- o.... .•. •
FROM ~~-c·.~.·->-'-'-----~~>---'----11'1
V-METER • ·; • , .
~~··.· ~: ...
Fig. 8-Schematic of ultrasonic test arrangement

152 mm
203
203
203
203
203
203
203
COLUMN CENTRHI NE
203
203 ISIICIINI~
I 2 3
203 10 mm
'""~f: w
E
E
ICI 5 5
-N
lEI 4 4 h•
I 2 3
FRONT IOOOR/ EASTI
ELEVATION PLAN
Fig. 9-Location of pulse velocity measurements
WEST
Dl RECTI ON TIIICKNESS OF
CONCRETE, mm
E -w 292
N-S 279
SE- NW 276
NE- SW 287
1-1 307
2- 2 286
3- 3 285
4-4 286
SE BAR- NE BAR 200
SW BAR - SE BAR 201
NW BAR- SW BAR 199
NE BAR - NW BAR 192
EAST
Fig. 10-Cross section of column after exploratory chipping

600
500 CALCULA TEO

MEASURED
400
300
200
u
0
100
w
"'
:::>
,__.
<( 300
"'
w
"- 200 64 mm OEPTH
""
w
,__.
100
200
100
600 720 840 960 1080 1200 1320 1440

TIME, min
Fig. 11-Temperature of concrete as a function of time for various

depths along center line of column A (fire exposure time: 1 hour)
800
CA lCUlA TEO
600
MEASURED
400
200
u
0
600
w
"',__.:::> 400
<(
"'a..
w
200
""
w
,__
600
400
152 mm OEPTH
200
o~~~~-L~~~-L~-L~~~
0 200 400 600 800 1000 1200 1400 1600

TIME, min
Fig. 12-Temperature of concrete as a function of time for various

depths along center line of Column B (fire exposure time: 2 hours)
........ .....
.... , AVERAGE CURVE
FOR MANY MIXES
''
''
''
e"' ''
'
>-
,_ ''
u
~
ANTIC I PAlEO
RANGE OF DATA
''
....
>
....
'
VI
.....
:::>
a.
COMPRESSIVE STRENGTH, MPa
Fig. 13--Pulse velocity-compressive strength relationship [8]
5000
----
UNDAMAGED .,.
;
...... ... --- ~-
e"' ...... ;
/
>-.
,_
u
----
0
.....
....
>
....
VI
.....
:::>
a.
00 10 20 30 40 50 60
COMPRESSIVE STRENGTH, MPa
Fig, 14--Pulse velocity-compressive strength relationship [9]

125r----~~--~--~----~---r--~r---,----,
~100-<>~.
~ 75 0
" ~~I DUAL STRENGTII
~ ~0 • •
u.
0
50
25 -
~o...._
MODUlUS 0 "·
...........0
~
-
----.....:0
OOL--~1~0~0--~2~0~0--~30~0--~40~0~~5~07
0--~6~0~0---7~0~0--~800
·-o
TEMPERATURE, °C
Fig. 15-Residua1 strength and modulus relation [4]
12
10 - - CAlCUlATED
E 8
E -----MEASURED
6
z" 4
0
....
0
""
::;:
-2 ',,
"'a ... , .... ____________ ,
-4
....
u.
c -6
.... -8
I
I
""
X
-10
-12
I
I
"" -14
-16
0 200 400 600 800 1000 1200 1400 1600
TIME, ml n
Fig, 16-Axial deformation as a function of time (fire exposure

time: 1 hour, Column A)
o~--1.~~----~~~------------1
-2
l
I
E -4 I
E I
-6 I
I
z I
0 \
1-- \
<( \
:;;: \
\
""
0 \
U-
~ -18 -
' ',
-' -20 '
'' ........ ,,
<(
X
<(
-22
-24
......... _______ ,
-26 I
I
-28 --CALCULATED I
I
-30 -----MEASURED I
-32 I
-34
-36L-~--~~~-L--~~---L--~~
0 2 00 4 00
TIME, min
Fig, 17-Axial deformation as a function of time (fire exposure

time: 2 hours, Column B)
MAXIMUM RESIDUAL STRENGTil

TEMPERATURE DEPTH, (FRACTION OF
REACHED, "C mm ORIGINAL STRENGTIH
• + +
PROFILE OF SECTION AFTER

REMOVAL OF WEAKER CONCRETE
Fig. 18-Calculated maximum temperatures reached at various depths

during heating and cooling period and residual strength at various
depths after cooling; profile of section after weaker concrete has
been removed (fire exposure time: 1 hour, Column A)
MAXIMUM RESIDUAL STRENGTH

TEMPERATURE DEPTH, (FRACTION OF
REACHED, •c mm ORIGINAL STRENGTIII
PROFILE OF SECTION AFTER

REMOVAL OF WEAKER CONCRETE
Fig. 19-Calculated maximum temperatures reached at various depths

during heating and cooling period and residual strength at various
depths after cooling; profile of section after weaker concrete has
been removed (fire exposure time: 2 hours, Column B)
120
,_
:I:
(!)
z
.....
<>:
,_
V>
.....
<(
z
(!)
""
0
..._
0
.....
(!)
,_
<(
z
.....
u
"".....
B.
:.:
,_
(!)
z
.....
,_
""
V>
0
0 4 10 II 12
TIME, months
Fig. 20-Natural recovery of compressive strength of concrete, heated

at various temperatures [12]
SP92-10
Fire Performance of Various Concretes

Deduced from Strain Measurements
During First Heating
by P. J. E. Sullivan and G. A. Khoury
Synopsis: Generalised response curves for the transient

thermal strain behaviour of concrete have been develo9ed
from a series of tests employing a wide range of
materials and a lower than normal heating rate. The
latter allowed detailed assessment of underlying
"material" behaviour to be made which was not
complicated by ''structural" effects that develop at a
fast heating rate. The temperature, stress and moisture
conditions within a cylindrical test specimen have been
investigated and a study of the behaviour of individual
constituents has confirmed that aggregate ther1nal
stability is a critical factor.
Thermal strains during virgin heating were separated

into "Free" and "Load Induced" components possessing
different and distinct properties allowing successful
prediction of residual strains.
A master curve connecting Load Induced Thermal Strain

with temperature up to 450°C was found to exist during
first .heating for different concretes thus simplifying
analys1s of heated concrete structures. Departure from
the master curve signified onset of concrete damage.
While transient creep did not occur during cooling, the
effect of cracking caused by thermal incompatibility was
clearly evident. Strain measurements, the~efore, proved
sensit1ve indicators of damage taking place during the
thermal cycle.
Keywords: concretes; fires; heating; measurement; performance;

permeability; strains, stresses; temperature; thermal properties
175
176 Strain Measurements
Dr. Patrick Sullivan is on the research and teaching

staff at Imperial College and has headed research on
concrete at elevated temperatures since 1966 and has
also supervised a number of doctorate students on
cryogenics and durability of concrete. He is also
Manaqinq Director of Probe Technical Services, an
independent laboratory for testing concrete structures
and developing of new materials.
Dr. Gabriel Khoury is a Research Fellow at Imperial

College and has investigated the effacts of heating,
crude oil and bacteria on concrete on which subjects he
has published jointly with Dr. Sullivan. He is also
involved in the supervision of a number of doctorate
students and is a member of the Technical Committee of
the Airship Association and has published on his concept
of the solar powered airship.
IN'I'UODUCTION
For nearly a century, the thermal stability of concrete

as a material has been assessed by means of strength
measurements taken at elevated temperatures or after
cooling. Strain measurements during transient thermal
states can, however, provide a more sensitive and
detailed indication of the thermal stability of
concrete. Assessment of cracking and loss of strength of
the concrete immediately after cooling can also be
better appreciated if the processes operative during the
temperature cycle are understood. Furthermore, analysis
of heated concrete structures requires data on the
transient thermal strain behaviour of concrete,
information on which is lacking.
Previous tests for fire applications have,
understandably, employed relatively high heating rates
up to 8°C/minute /l-3/. "Structural" effects /7 I can
develo~ within the test specimen under such heating
condit~ons which obscure and/or distort fundamental
"material" trends /7/. Such effects have been minimized
in the present test programme /4-9/ by employing heating
rates below 1°C/minute and by optimising specimen size,
both of which reduce temperature differentials within
the specimen. Analysis of results was also simplified by
(a) applying constant 9ravity loading, (b) ensuring as
linear a heating rate as possible and (c) performing
replicate tests which made it possible to employ
statistical analysis. Furthermore, by employing a range
of materials with different a')gregates, cements and
cement replacements, it was ?Ossible to develop a
generalized model for a wide ran~e of concrete mixes.
Sullivan and Khoury 177
APPARATUS
The apparatus /5/ has been designed to investigate the

transient and steady-state thermal strain behaviour of
unsealed concrete cylindrical specimens subjected to a
constant uniaxial compressive load up to 10 tonnes and
temperatures up to 700°C. It consists of three identical
"dead weight" creep rigs each with a three zone furnace
(figure l) operated by an automatic temperature
controller. This enabled constant load (figure 2) to be
maintained on the specimen during linear heating subject
to minimal axial temperature variation (<l2°C).
The specimen's surface temperature is measured by three
t¥pe K thermocouples (sheathed in insuquartz) attached
w1th a high-temperature-resistant cement to the top,
middle and bottom sections. The total axial
displacement of the specimen is transmitted from two end
platens via four invar connecting rods to an lvdt
beneath the furnace at ambient temperatures. Both
temperature and displacement are monitored by a
computerized data logging system.
Calibration of the strain measuring device, using copper
samples indicated that the coefficient of thermal
expansion of the copper sample ranges between 98% and
102% of the known value of th~) copper material. The
errors of 100 P,E (=strain x 10- or less are within the
variability in strain obtained from concrete under
similar conditions /5/.
TEST SPECIMENS AND MATERIALS
The samples were unsealed cylinders (62mm dia x l86mm

long) with a diameter small enough to minimize radial
temperature differentials during thermal transients but
allow the use of l2mm aggregate. A 3:1 specimen
length:diameter ratio was chosen to reduce contributions
from end effects at the central portion of the specimen.
The top surfaces of the specimens were ground plane,
parallel and orthogonal to their vertical axis.
Details of the five concretes, and their designations,
are given in table l. Further information is given in
references 7 and 8.
The aggregates were crushed dolomitic limestone, crushed
quartz dolerite basalt, natural weathered Thames River
siliceous gravel and lightweight (sintered pulverized
fuel ash Lytag). The fine aggregate was natural
siliceous sand (zone 2) in all concretes except for the
BII and lightweight mixes which employed calcareous
quartz sand (zone 2) and Lytag fines respectively.
Two types of cement were used, Ordinary Portland (OPC)

and Sulphate Resisting (SRC). Cement replacements
included pulverised fuel ash (pfa) and ground granulated
slag (Cemsave).
TEST PROCEDURE AND PARAMETERS
t-1ain Tests
All specimens were subjected to a constant uniaxial

compressive load and immediately afterwards heated at a
constant rate of l°C/minute to 600°C at which
temperature they were maintained for one hour. This was
followed by natural cooling under load, begining at a
rate of l.9°C/minute at 600°C and declining by about
0.34°C for every 100°C drop in temperature. The overall
axial deformation and surface temperature of the
specimen were measured throughout the thermal cycle.
The specimens were cured at 100% RH and 20°C for at
least 6-8 months before testing.
The test regime is simplified to constant uniaxial
compression for ease of testin9 and analysis of t~e
results. The important point 1s that the concrete lS
loaded before the thermal transient. Four load levels
(0,10,20 and 30%) were applied as fractions of the
specimen strength prior to the strain tests.
The maximum test temperature of 600°C represents the
upper limit for the structural usefulness of Portland
cement based concretes. Althou9h concrete surface
temperature in a fire exceeds th1s limit, the bulk of
the concrete is likely to remain within the test
temperature range.
The l°C/minute heating rate was low enough to ensure
that "structural" effects were not significant. This was
confirmed by similar trends obtained during heating at
0.2°C/minute /4,8/.
Additional Tests
The temperature distribution during heating was
determined using 10 chromel-alumel thermocouples cast
within the concrete specimens.
Weight loss and rate of weight loss measure~ents were
made on the the concrete samples as used in the strain
tests. They were heated in a separate sinqle zone
furnace mounted over a Mettler Pll balance.
Dilatometry tests were performed on individual samples
(20-30mm rectangular blocks) of aggregate and cement
paste during constant heating and subsequent cooling at
a rate of 2 C/minute in rough vacuum (-0.1 mm Hg) using
a Netzsch 402E quartz tube dilatometer.
CONDITIONS WITHIN SPECIMEN AND STHUC'l'URAL EFFECTS
Structural effects are defined here as those induced by

significant spatial variation in stress, strain,
moisture content, pore pressure, material properties
etc., within the specimen caused by large temperature
gradients during rapid heating or cooling and which
distort its thermal strain behaviour.
Temperature Distribution
Analytical solution of the Fourier transient heat

conduction equation /6/ for a specimen of radius "a" and
thermal diffusivity "D" during heating at a constant
rate "R", indicate that the maximum radial temperature
difference AT is:
AT= Ra2/4D ....................... (1)
Radial temperature gradients for a given material can

therefore be minimized by reducing the rate of heating
and particularly the rad1us of the specimen. Both these
measures have been implemented in this work.
A finite difference solution /6/ of the Fourier equation
confirms these relationships for a specimen of finite
length. The temperature distribution at the centre of
the specimen remains unaffected by end heating providing
the length/diameter ratio .is greater than 1-1.5 during
heating at 1°C/minute. ·
An experimental investigation confirms these trends /6/
and shows that the effect of variation in material
properties, such as thermal diffusivity, is to cause two
peaks in the AT versus surface temperature relationship
at about 160°C and 550°C (figure 3) corres9onding to the
processes of moisture loss and dissociation of Ca(OH) 2
respectively. Peak values of AT for gravel and
lightweight-concrete specimens, representing· upper and
lower bounds of thermal diffusivities, did not exceed
24°C and 12°C respective!~, durina heating of 60mm
diameter specimens at 1 C/minute. 'l'he peak AT at
0.2°C/minute would be a fifth of these values and hence
negligible.
Thermal Stress Distribution
Although thermal stresses are relaxed by transient creep
during virgin heating, they can still be significant
/10/ at hiqh rates of heatina. 'l'hey were, however,
estimated 12/ to be negligible- in unsealed specimens
dur.in9 heating above 100 C at rates even higher than
l°C/m1nute due to the magnitude and rapid rate of
development of transient creep.
An upper bound of thermal stresses during heating can be

obtained from an elastic thermal. stress analysis for the
specimen /4,7/. An analytical solution shows that the
maximum axial thermal stress az, resulting from the
temperatures given in eqn. l, 1s at the centre of the
specimen and can be expressed in terms of the modulus of
elasticity "E", poisson ratio "v" and coefficient of
thermal expansion "a" as well as "a", "R" and "D":
a EAT
K ••••••••• ( 2)
80(1-V) 2 ( 1- V)
which is tensile (figure 4) and directly proportional to

the maximum radial temperature difference AT. Likewise,
it can be shown that all maximum axial, radial and hoop
stresses are fractions of K and hence directl~
proportional to AT (=Ra2/4D). Thermal stresses can,
therefore, also be minimized by reducing the heating
rate and particularly the specimen radius.
Thermal stresses assume greater importance during
coolinq and reheating because of the absence of
transient creep. The elastic analysis given above is
more applicable here except that the stresses during
cooling would be reversed. Published data /11/ sugqest
that the residual strength of concrete specimens similar
to those employed in the current tests are unlikely to
be affected for rates of cooling below 2°C/minute.
Moisture Condition and Pore Pressure

The specimen weight loss during first time heating to
500-600°C is taken to represent the amount of moisture
escaping from its surface: A typical result for rate of
weight loss per unit surface area of the specimen is
shown in figure (5) for limestone concrete heated at
l°C/minute.
Only 3% of the evaporable water content escapes during
initial heating between 20°C and l00°C at l°C/minute.
Even at 0.2°C/minute only 9% escapes. Therefore, over
90% of the evaporable water is still retained when the
specimen reaches l00°C thus implying that hydrothermal
(autoclave) conditions exist during first heating even
in small unsealed samples. This may cause a
transformation of the CSH into a weaker or stronger
hydrate depending on the Ca0/Sio 2 ratio.
The sharp increase in the rate of moisture loss above
l00°C for all mixes confirms that the permeability of
concrete increases markedly above that temperature as
reported recently by other workers /12,13/.
The rate of moisture loss reaches a peak at about

l35-200°C depending on the heating rate /7/. Explosion
of the specimen can take place at these temperat~re at
higher heating rates /14,15,3/ than those experlenced
here. On further heating, the rate of moisture loss
declines sharply and for temperatures above 250-300°C it
is independent of the initial moisture content /7/. The
total amount lost for any given temperature during the
transient also approaches the maximum level attained at
constant temperature /7/.
Maximum pore pressure developed within the specimen at
l°C/minute has been estimated, ~n the basis of published
data, to be less than 0.2N/mm /4/· This is less than
about 7% of the tensile strength of the concretes tested
and would be even lower at 0.2°C/minute.
INDIVIDUAL CONSTI'l'UENT BEHAVIOUR
Dilatometry of the a9gregates indicates that thermal

stability increases ln order for the gravel, limestone,
basalt and lightweight aggregates (figure 6).
The expansive cracking of the gravel aggregate at about
350°C renders it unsuitable for use in concrete destined
for heating above that temperature. Damage, though less
severe, of the limestone and basalt aggregates is
evident from their residual expansions, the former being
less stable and experiences a residual expansion after
heating to 350°C. Lightweight aggregate is expected to
be dimensionally stable even at 600°C as it is
manufactured above l000°c.
The dilatometry tests on the cement pastes (figure 6)
show that their free thermal strain is significantly
influenced by the type of cement and cement replacement
employed. This is due to differences in thermal
deformations of the various anhydrous and hydrated
constituents /16/ present in different proportlons in
each blend. Following initial expansion, the cement
~astes undergo considerable shrinkage above l00°C as
they lose evaporable water and dehydration of the
varlous hydrates proceeds but reaches a minimum at about
500-600°C /16/. Thermal shrinkage increases in order
for the OPC/pfa, OPC and SRC/Cemsave mixes, but thermal
movement on subsequent cooling is at least an order of
magnitude less.
Differential Thermal Analysis (DTA) and
Thermogravimetric Analysis (TGA) indicate that the
a9gregates and sands employed were chemically stable
Wlthin the test temperature range while the cement
pastes exhibit typical characteristic trends showing, ~s
expected, the Ca(OHl 2 content of the blend cement
pastes to be lower than that of the Portland cement
paste /7/.
'l'IIEHMJ\f, STRAIN RESUVl'S
Examples of the complete strain history at four load

levels during first heating at l°C/minute and subsequent
coolin9 are given in figure (7) for 2 concretes.
Strains During First Heating
An important feature of the transient thermal strains
shown in fiqure (7) is the considerable contraction
under load ~ompared with the free (unloaded) thermal
strain (FTS). This contraction is is referred to as the
"load induced thermal strain" (LI'l'S). 'l'he total thermal
strain is, therefore, considered to· consist of two
components (figure 7):
Total Thermal Strain FTS(T,O) - LITS (T,s) ...... (3)
exDressed as functions of temperature (T) and stress (s)

al~~ough, as will be shown later, they are also
dependent upon several other factors.
The thermal strains are zero at the begining of heating
since the initial "elastic" strains were deducted.
Free Thermal Strain (FTS) Although the different
concretes had s1m1lar aggregate contents by volume
(65-67%), FTS varied considerably with type of concrete
(figure 8) because of the dominant role of aggregate
type. The large expansion of the gravel concrete
contrasts with the lightweight concrete which deforms by
a small amount up to the test temperature of 600°C.
The large increase in expansion of the gravel concrete
above 350°C is caused by the expansive break up of the
aggregate 6 while that of the II, 81 and BII concretes
above 550 C is at least in part caused by the a-~ inver-
sion of quartz which takes place at 573°C. Such
behaviour is absent in the liqhtweiqht mix which does
not contain quartz. - -
The influence of type of cement, although small by
co~Jarison, is not negligible as can be deduced from
figure 8 where the greater shrinkage of the SRC/Cemsave
paste compared with that of OPC (figure 6) results in
less thermal expansion of the BII than the BI basalt
concrete.
Load Induced Thermal Strain (LITS) The importance of
LITS l1es 1n 1ts very large magn1tude as indicated by a
typical set of results shown in figure (9).
In one of the most revealing results of this work, the

LITS curves of the concretes, with the exception of
gravel, are statistically the same for temperatures uo
to about 450°C (figure 10). This is despite their
initial strengths varying by a factor of 2 (table l) and
the different ag9regates and cement blends used. To
simplify analys1s, a "master" LITS curve is, therefore,
produced for each stress level (figure 9) to regresent
the results for concrete in general up to 450 C. The
gravel concrete departs from the trend at about 400°C
because of an increase in creep resulting from concrete
damage.
LITS is a non-linear function of temperature (figure 9).
A peak occurs in its coefficient at about l50°C /8. The
rat~ of moisture loss also peaks at this temperature
(figure 5) which suggests that drying creep contributes
to the peak in LITS coefficient.
The results imply that the capacity of different
concretes to rela~ stresses (thermal and parasitic) are
similar, for a g1ven strength, and verv large. It
follows that the actual stresses which may aevelop after
relaxation would still be in proportion to the thermal
expansion of the aggregates.
The concept of the master LITS curve is valid up to
450°C providing the following criteria are met:
- The concretes contain the same aggregate content by
volume since the presence of the aggregate restrains
LITS /8/.
Comparison is made on the basis of the same
stress/cold strength ratio. Application of the same
stress level does not produce a master LITS curve /4/.
- The concrete does not experience damage during heating

which can result in siqnificant departure from the
master curve (figure 10).-
Preheating conditions are similar since preheating
can reduce or eliminate transient creep within the
temperature range of previous exposure /8/.
The heating rate is the same although this effect is
not very marked for 0.2-l°C/minute since a five-fold
reduction in heating rate has resulted in an increase of
only 10-20% strain /8/.
- The cement blends produce a calcium silicate hydrate.
- The age at heating and the curing conditions are
consistent.
Strains During First Cooling
A comparative analysis of the contraction of the

specimens during cooling is simplified by taking all
strains to be zero at the initial 600°C point of
cooling.
For the thermally stable lightweight (LI) mix, the
cooling curves are statistically unaffected by stress
level (figure ll) confirming that transient creep,
structural and incompatibility effects are negligible
during cooling.
'fhe cooling curves of the limestone (H) concrete,
however, begin to diverge below about 500-450°C (fi9ure
ll). The curve for the unloaded specimen turns 1nto
expansion even while the temperature is still falling
which is clearly caused by the expansive effect of the
limestone aggregate which registers a large residual
expansion on cooling (figure 6). This behaviour becomes
less evident as the load increases and for 30% load,
there is a tendency for the concrete to experience some
creep, all of which imply damage to the concrete. The
divergence in strains takes place after a given
temperature drop (l00-l50°C in this case) when thermal
incompatibility in the lightly loaded specimens may have
created tensile stresses exceeding the tensile strength
of the bond region or the cement paste. These tensile
stresses increase as cooling proceeds while, by
contrast, thermal stresses diminish since the cooling
rate reduces with temperature.
The basalt concretes proved more stable during cooling
than the limestone concrete. The cooling curves of the
BI mix begin to diverge slightly below 300-200°C, i.e.
after a temperature drop of 300-400°C /4/. The improved
performance compared with limestone concrete is
primarily due to the smaller residual expansion of the
basalt aggregate (figure 6) which allows the concrete to
experience a larger temperature drop before the build up
of stresses causes divergence of the cooling curves.
The contraction of the concretes increase in order for
the LI, BII, BI, H and GI mixes (fiqures 12) which is
the same order as the expansion during initial heating
(figure 8).
Residual Strains and their Prediction

The residual strains (figure 13) represent the total sum
of all irrecoverable contractions and expansions taking
place during the thermal cycle /9/.
The residual strains of the loaded specimen RS(s,T) is
estimated in terms of the unloaded specimen RS(O,T),
LITS(s,T) and any creep that takes place at constant
temperature {c(s,T)}. Any expansive damage during
cooling in the unloaded specimen, which may be

suppressed in the loaded condition, can be accommodated
by adding the difference in the cooling curves between
the loaded and unloaded specimens {e(s,T)}. The residual
strain for a maximum temperature (T) and constant load
(s) then becomes:
RS(s,T) RS(O,T) + LITS(s,T) + c(s,T) + e(s,T) ..... (4)
which predicts residual strains (figure 13) even when

the unloaded specimen experiences cracking during
cooling. It follows that the superposition of strain
approach, and separation of strains into FTS and LITS to
explain those of loaded concrete, are justified.
Furthermore, once RS(O,T) is known, and one test is
sufficient, the normalized master LITS curve can be
employed to predict RS(s,T) for any stress level up to
30% and temperatures up to 450°C. It should, however, be
emphasized that the strains measured during the thermal
cycle were for h~ating and cooling rates which did not
introduce significant structural effects.
A comparative study of the residual strains for the
concretes (eg. figure 13) in the unloaded condition
shows that only the lightweight mix exhibits a residual
shrinkage. The basalt concretes produce negligible
(BII), or slight (BI), net residual expansion. In
contrast, the ·limestone and gravel conc~etes, being
seriously damaged, produce residual expansions of 5500
and 11500 microstrains respectively. All loaded
concretes, however, have residual contractions which
increase with load level /9/.
CONCLUSIONS
Structural Effects:
Thermal strains representing underlying "material"
behaviour were measured uncomplicated by "structural"
effects by employing a relatively slow rate of heatinq
(<1°C/minute) and small specimen diameter (60mm) which
minimized radial temperature gradients and thermal
stresses during thermal transients.
Conditions Within Specimen:
Two peaks in the maximum radial temperature difference
(AT) occur at about 160°C and 550°C during heating at
l°C/minute due to moisture loss and dissociation of
Ca(OH)2. The tensile thermal stresses that develop in
the specimen centre during heating are proportional to
AT. ~1oisture loss increases sharply above 100°C to
reach a maximum at about l35-200°C depending on the
heating rate. ·
Individual Constituent Behaviour:
Limestone and basalt aggregates were dimensionally

unstable after heating to 600°C registering residual
expansions. 'rhe considerable thermal shrinkage of cement
paste above l00°C increased in order for OPC/pfa, OPC
and SRC/slag mixes.
Thermal Strains:
Are considered to consist of Free (FTS) and Load Induced
(LITS) comoonents. FTS is dominated by the thermal
expansion of the aggregate while LITS is identical
irrespective of type of aggregate used for temperatures
up to 450°C. A master LITS curve, therefore, exists for
concrete in general providing certain criteria are met.
Cooling strains of thermaily stable concrete are
unaffected by stress level indicating absence of
transient creep. For basalt, and particularly limestone,
concrete the development of cracking during cooling, due
to excessive thermal incompatibility caused by aggregate
residual expansion, is clearly evident. Residual
contraction (expansion) is obtained for thermally stable
(unstable) unloaded concretes. Residual strains can be
predicted by suoerposition of irrecoverable strains
occurring throughout the thermal cycle.
Implications:
By reducing the heating rate to low levels, it has been
possible to ascertain the underlying "material" trends
of behaviour and grade the different types of concrete
as to their thermal stability during the first thermal
cycle. Once the "material" mode of behaviour is known,
it becomes possible to estimate the distortions, and
additional stresses, that develop at the higher heating
rates experienced in fire.
ACKNOWLEDGEMENTS
The authors thank Mr. Brain Grainger of the Central

Electricity Research Laboratories for his generous help
throughout-the period of this research.
REFERENCES
1. Schneider, U. and Kordina, K. On the behaviour of

normal concrete under steady state and transient
temperature conditions. Proceedings of the third int.
Conf. on "Structural Mechanics in Reactor Technology"
London 1975. North-Holland Publ. Co.
2. Anderberg, Y. and Thelanderson, S. Stress and

deformation characteristics of concrete at high
temperatures: 2. Experimental investigation and
material behaviour model. Lund Institute of
Technology, Sweden, Bulletin 54, 1976, pp. 81.
3. Thelanderson, s. Mechanical behaviour of concrete
under torsional loading at transient,
high-temperature conditions. Lund Institute of
Technology, Sweden, Bulletin 54, 1976, pp. 84.
4. Khoury, G.A. Transient thermal creep of nuclear

reactor pressure vessel type concretes. Ph.D thesis
submitted to the University of London, 1983.
5. Sullivan, P.J.E., Khoury, G.A. and Grainger, B.N.

Apparatus for measuring the transient thermal strain
behaviour of unsealed concrete under constant load
for temperatures up to 700°C. Magazine of Concrete
Research, Vol. 35, No. 125, Dec. 1983.
6. Khourv, G.A., Sullivan, P.J.E. and Grain9er, B.N.
Radial temperature distributions with1n solid
concrete cylinders under transient thermal states.
Magazine of Concrete Research, Vol. 36, No. 128,
Sept. 1984, pp. 146-156.
7. Khour¥, G.A., Grainger, B.N. and Sullivan, P.J.E.
Trans1ent thermal strains of concrete: Literature
surve¥, conditions within specimen and behaviour of
indiv1dual constituents. Magazine of Concrete
Research, Vol. 37, No. 132, September, 1985, pp.
131-144.
8. Khoury, G.A., Grainger, B.N. and Sullivan, P.J.E.
Strain of concrete during first heating to 600°C
under load. Magazine of Concrete Research, Vol. 37,
No. 133, December 1985, pp. 195-215.
9. Khoury, G.A., Grainger, B.N. and Sullivan, P.J.E.
Strain of concrete during first cooling from 600°C
under load. Magazine of Concrete Research, Vol. 38,
No. 134, March 1986.
10. Parrott, L.J. A study of transitional thermal
creep in hardened cement paste. Magazine of Concrete
Research, Vol. 31, No. 107, June 1979, pp. 99-103.
11. Campbell-Allen, D. and Desai, D. The influence of
aggregate on the behaviour of concrete at high
temperatures. Nuclear Engineering and Design 3, 1966,
pp. 54-65 & 246-255.
12. Chapman, D.A. and England, G.L. Effects of
moisture migration on shrinkage, pore pressure and
other concrete properties. 4th. Int. Conf. on
"Structural Mechanics in Reactor Technology",
San-Fransisco, 1977, paper H5/3.
13. Bazant, Z. Pore pressure and drying of concrete at

high temperatures. Journal of the Engineering
Mechanics Division. Proc. of ASCE, Vol. 104, October
1978, pp. 1059-1079.
14. Harmathy, T.Z. Design of concrete masonary walls for

fire endurance ACI special publication SP-39,
"Behaviour of concrete under temperature extremes",
1973.
15. Sullivan, P.J.E. and Zaman, A.A.A. Explosive

spalling of concrete exposed to high temperatures
Proc. 1st. Int. Conf. on "Structural mechanics in
reactor technology", Berlin, 1971, paper Hl/5.
16. Piasta, J. Heat deformations of cement paste phases

and the microstructure of cement paste. Materiaux et
Constructions, Vol. 17, No. 104, October 1978, pp.
415-420.
TABLE 1 - DETAILS OF FIVE CONCRETE MIXES
proportions by dry weight

Limestone basalt gravel lightweight
I II I I
Designation
Constituent
- II BI BII GI LI
cement (OPCl 1.0 1.0 - 1.0 1.0

cement (SRC)
pfa
-
0.25
-- 0.5
- -
-
-
-
Cemsave
sand
-
2.3
-
1.8
0.5
1.8
-
-
-
-
fine sand
coarse sand
-
-- -- -
--
0.65
1. 55
-
-
fine aggregate
medium aggregate -
-- -
-- 0.87
0.69
coarse aggregate 3.45 2.7 2.7 2.39 0.69
water 0.555 0.45 0.45 0.60 0.80
plasticizer 0.005 0.005 0.005 - -
water/blend 0.444 0. 45 0. 45 0.6 0.33
agg./blend 4.6 4. 5 4.5 4.59 2.25
cement ~ontent 388 415 420 365 447
(K<]/m )
dens1ty 3as cast 2 346 2472 2498 2259 1810
(Kg/m )
ayg· content 67.8% 67.3% 67.2% 65.0% 65.3%
by volume)
Strength <NLmm2l
28 day (cube) 51.0 61.0 50.9 34.1 47
before test (cylinder):
moist
air-dry
_, 64.2
71.9
61.3
74.0
56.5
67.0
31.0
39.3
49.5
46.6
oven-dry 54.9 62.1 70.4 34.8 45.6
--
control
measuring
lhermocouples \
_¥==; J
lever arm
5
p
~;-
(15: 1 ratio)
fumace
~ heating
zones
~
specimen
-<-
\fJ
•nv•r conneeting rods
:::::
micrometer
e
¥
~
::s
~
transducer
dead weights
=
0....
~
::=-'
0
'- :::::
'"I
'-<
Fig. !--Schematic diagram of furnace,

specimen and strain measuring device
Fig. 2--Mechanical loading system of
creep testing machine -
00
\0
28
24
~ 20
1 16
~ 12
,..-,,
8
4 ~ .,..... .......... "' -----
gravel -
1
0 100 200 300 400
SURFACE TEMPERATURE-•C
Fig. 3-Maximum radial temperature difference curves for lightweight

and gravel concrete specimens during first heating at 1 C/minute from
the initially air-dry condition
2 N
a
.....,0
~ a
~ ClJ
f<
".,ClJ
.,
0
-1
~
...
s:
f<
ClJ
I<
.,
liClJ"'
+a
.....,0
..
ClJ
I<
§
u
Radial Distance (%)
Fig. 4-Calculated thermal stress profiles, 1~ithout relaxation by

creep, in the central cross section of gravel concrete specimen
(60mm diameter) during heating at 1 C/minute
7
~
I:
QJ 1.5
u 6
H
5
2!.
Ill 4 1.0
Ill
\-z-
0
..1 3
~
.<! 2
...__ __ o.5
....QJ
"
)
....QJ ..,0
~ 1
QJ
~
0 0
0 100 200 300 400 500 "'
P:
Temperature- 0 c
Fig. 5-Weight loss of limestone concrete specimen during heating at

1 C/minute from the initially moist condition
10000 500 700

AGGREGATES
"'...0
IC
8000
"'...
0
CEMENT PASTES
...
1: IC
.."
H ..." - OPC/PFA PP2
I
6000
b."
I CPl
...."
0 4000 -OPC BI
" ...."
0 GI
LI
! 2000 H
u
.." - SRC/Cemsava Ion
CPJ
"
0
0
-50000
Temper11ture - 0c
Fig. 6-Free thermal strains of small rectangular blocks (20-30mm) of

aggregates and cement pastes (initially moist) during the first heat
cycle (2 C/minute) to 600 C.
1-'
1.0
t-:.l
soool- Heating J\ Cooling
I sooot Heating
I Cooling
I
UJ
..
'II)
c
K
Stress
~
......
-·a:
""!
$lj
::s
....,= 0
......
Ill
~
$lj
riJ
.:::
""!
~
3~
..... ::s
......
f\~ ~ \\~.,.
-sooot- riJ
Basalt II
......51 t eon:::-et~
tl
...fl -1000
u
= D
-1500
Fig. 7--Thermal strain of two initially moist concretes during first

heating at 1 C/minute under load to 600 C and subsequent cooling
10000~----------------~------~
,
t
r
I
.i Ir
~ sooo / I
I I .BII
''b
...
...,
I(
1:: Ll
..
b
Temperature - 0 c
...,
1::
0
....u
~
....
1::
E -4000~------------------------~
Fig. 8-Free Thermal Strain (FTS) of 5 concretes during first heating

at 1 C/minute
Temperature - 0 c
~emperature - 0 c ,__..
200 300 400 500 600 0 600
0
'e
-1.000
r:J)
-2000
g
s·
..·f.. ,£""
~I -Jooo a=
...
000
IJC
CJ)
!);J
w
IC
"·J.. no. of curvel .s,., -4000
=
.....c included ....... ~
CJ)
...... 3CJ)
:1 .. . '(6-11) -, -5000
c
....
0 BI ::s
.......
0'
::Iu
.,
......c
-60
E• standard
.l.t:. .
~
t,.,
......c
81
-6000
-7000
w
... .
0 deviation
u
""(10-11) -8000
.
-9000
:t lE'':
Fig. 10--Load Induced Thermal
.. Strains of 5 concretes during
"'(eJ first heating at 1 C/minute
under 10 percent compressive
Fig. 9--Master Load Induced Thermal load
Strain (LITS) curves representing
5 different concretes each con-
taining 67 percent aggregate by
volume, during heating at 1 C/minute
under 3 load levels
100 200 300 00 . _100 _200
I
-1000
(a)
-1000
\0
0
~
-2000
)( \0
~
... -''''[ •«•••
0
c ~
......."' -3000 ~ Lightweight

)(' -3000
"' concrete "'c

.....
-4000 30% ~~-4000[
+> stress
level
"' ---
g -5000
..... '- 0%
~
0
0
u.
-6000
Limestone
en
-7000
-8000
concrete
-
:::::
..,<-·
..,::s
::s
Fig. 11--Strains during cooling from 600 C under four load 0..
levels of thermally (a) stable and (b) unstable concrete ~
::s"'
0
:::::
~
'-<
1-'
\0
C,Jl
Fig. 12-FTS during cooling from 600 C of 5 unloaded concretes
Fig. 13-Actual and predicted residual strains of 3 concretes at

4 stress levels
SP 92-11
Chloride Absorption
from Concrete Surfaces
by A. Lammke
Synopsis: After actual fires with plastic materials involved con-

crete surfaces are often infested with high chloride concentra-
tions, causing a serious risk of corrosion of the reinforcement.
Absorption techniques with lime paste have been recommended to
remove chlorides from concrete surfaces. To study this reaction
concrete beams with different chloride contents, added as sodium
chloride, were cast. The chloride distribution in the concrete
before and after lime paste treatment and the chloride content of
the used lime were determined analytically. The results show that
the lime does not absorb chlorides, and within the concrete, the
chlorides are redistributed. Further tests were made to study the
rate of chloride absorption from aqueous solutions by lime, acti-
vated carbon and ion exchange resins. The amount of absorbed
chloride was very low. Corresponding the absorption of chlorides
by fresh pastes of different cements and completely hydrated ce-
ments was studied. An effect of absorption but much more of che-
mical bond was shown.
Keywords: absorption; calcium chlorides; chlorides; concretes; lime
The author, Axel Lammke, holds degrees in chemistry and civil

engineering. He is head of the chemical department of the Insti-
tute of Testing Materials for the Building Trade at Braunschweig,
West Germany, and member of several DIN-standard committees. His
primary work includes testing of materials, development of test
procedures, diagnostic studies of failures in building construc-
tions and materials. He is a lecturer for chemistry of building
materials and chemistry of cements at the Technical University
of Braunschweig.
197
198 Chloride Absorption
1. Introduction
In actual fires chloride containing organic materials, as present

almost everywhere in buildings, form by pyrolysis (apart from
volatile organic compounds) hydrochloric acid. This acid reacts
with metals causing serious corrosive attack. In highly equipped
industrial or other buildings with technical purpose this type
of secondary damage may be more serious than the direct destruc-
tion by the fire itself.
Hydrochloric acid reacts with the calcium compounds of hardened

cement in concrete and related materials as well, forming calcium
chloride in a neutralization reaction. Within reinforced concrete
the soluble calcium chloride may cause severe corrosion of the
steel even in alcaline surroundings. This attack is influenced
by the chloride concentration itself, by the amount of humidity
and oxygen present and particularly by the type of cement in-
volved. As a safe margin we accept 0,4 % of chloride by weight
of the cement [1).
When this margin is exceeded, consideration has to be given to

the safety and durability aspects of the affected concrete. If a
dangerous risk of corrosion is to be expected, chipping off of
the infested concrete and replacing it is common remedial prac-
tice [2) since neither simple washing [3) nor electroosmosis [4)
seem to be successful means to remove chloride from concrete.
Another removing technique has been widely used: absorption of
chlorides from concrete surfaces with hydrated lime paste [5)
[6) (7]. Concrete surfaces containing higher amounts of chlorides
are treated almost perpetually with pastes of hydrated lime. A
considerable reduction of the chloride concentration in the outer
concrete layers due to absorption of chloride by the lime paste
is claimed, without - in our opinion - reliable proof of effi-
ciency. Even the authors of the lime curing method have become
a bit doubtful and recently published less promising results [7).
2. Lime curing of concrete specimens

2.1 Preparation of specimens
To test the lime curing method under laboratory conditions con-

crete specimens size 10 x 15 x 70 em' were cast. Normal portland
cement was used in an equivalent amount of 300 kg/m' • Sodium
chloride was added dissolved in water in an amount corresponding
to 1, 2 and 4 %by weight of the cement. The beams were condi-
tioned in plastic sacks for 28 days, then dried naturally in a
covered shelter in the open air for three months. At that age the
humidity of the concrete was found by oven drying (105 °C) to be
~ 4 %, The porosity determined according to DIN 52 102 [8) was
about 20 % by volume. Except one longitudinal surface all other
surfaces were covered with an epoxy coating.
A part of the samples was used to determine the original chloride

distribution. With a special drilling technique, using a hard
metal tool, concrete powder from different depths in steps of
Lammke 199
5 mm up to a total of 30 mm was obtained, following a statistical

pattern over the surface. The dust was dried and the chloride
content determined by potentiometric titration after dissolving
in diluted nitric acid. 'I'he analytical procedure closely followed
the methods published by "Deutscher AusschuB fiir Stahlbeton" [9] •
The results in table 1 show that the chloride is unevenly distri-

buted. During the concrete's drying process the soluble chloride
is partly transported to the outer zones. There, a higher content
than calculated is found, corresponding to a lower amount in the
inner parts.
2.2 Lime curing and chloride content of the concrete surface
The prepared specimens with known chloride distribution were

treated on their free surface with a paste made of hydrated lime
and water, ratio 1 : 1. The natural chloride content of the lime
was determined, distilled water was used. The paste was spread
as a 6 - 8 mm thick layer on the concrete surface. The specimens
were covered with plastic foil for 8 hours to retain the moisture
for a sufficient amount of time. Then they dried for 16 hours in
laboratory air. The dry lime could be brushed off easily and the
procedure was repeated another four times. The dry lime was col-
lected and kept.
After this treatment the chloride distribution in the concrete

was determined exactly as described in part 2.1. The results are
shown in table 2. The results in table 1 and 2 are combined and
presented as graphs nr. 1, 2 and 3.
At first sight, these graphs show a striking decrease of chloride

concentration in the outer concrete layers just to a depth im-
portant for reinforcement. However, the right edge of graphs 2
and 3 may raise the curiosity of the interested watcher. How
might the full and the dotted lines continue? If the remarkable
decrease of chloride in the treated concrete is due to absorption
by the lime, the chloride must be found there. An analysis of
the used lime proved the chloride content to be very low and in
fact no remarkable amount of chloride had been absorbed. The
total loss of chloride in the outer 3 em of the concrete spe-
cimens and the total amount of chloride in the used lime were
calculated. The absolute values expressed in grams are combined
in table 3. The discrepancy is obvious.
2.3 Lime curing and chloride content of the total concrete

cross section
Since the lime curing caused a decrease of chloride in the outer

concrete layer without a corresponding absorption by the lime,
it is most likely, that the chloride had been redistributed
within the concrete. Following this idea, the described investi-
gations had to be extended over the entire cross section of the
specimens.
The concrete beams were treated 5 times with lime paste on one
half of their surface, then cut with a stone saw into blocks and
slices. The slices from the treated and untreated halves were
analysed. The results are given in table 4 and in the graphs 4,
5 and 6.
Again a decrease in chloride content in the outer zone is shown;

however, regarding the total cross section, the outer decrease
corresponds to an increase in chloride concentration of the inner
parts. Comparing roughly the area between the full and the dotted
line on the left side of the diagrams and the enclosed area in
the middle, they seem to be very much alike in size. 'I'hat means:
the lime treatment does not reduce the chloride content of the
outer layers of concrete, but transports the chloride deeper into
the inner depths. The only efficient medium of the lime paste is
the water. It dissolves the chlorides and transports them into
the concrete. The lime used in the second series of concrete
treatment was analysed as before. Again no considerable amount
of chloride was found.
3. Theoretical aspects
The reported results plain, and straight forward, about this

method of lime curing deserve a more detailed approach. 'l'his
refers to general aspects of absorption as well as to the be-
haviour of cements towards chlorides.
3.1 Absorption of chlorides
Absorption is defined as a change in concentration, normally in-

creasing, of one material on the surface of another. It can be
described by the Freundlich equation [10).
k . cn
x mass of absorbed material
m mass of absorbing material
c equilibrium concentration
of absorbed material in solution
C = a - x
a start concentration
of absorbed material
k constant depending on the

absorbing material
n constant depending on the

absorbed material
In order to check the absorbing capacity of lime for chlorides

in aqueous solutions some tests were made, including ion exchange
Lammke 201
resins and activated carbon as absorbants. With chloride solutions

of different molarity and at a fixed temperature of 20 °C absorp-
tion isotherms according to the Freundlich equation were regi-
stered. '!'he results are shown in graph 7. In this diagram only
the ion exchange resin shows a considerable absorption of chlo-
ride, while calciumhydroxide and activated carbon show almost no
effect at all.
3.2 Reactions between cements and chlorides
When trying to extract chlorides from concrete by means of ab-

sorption, one must take into account, that the hardened cement
paste with its high microporosity is a well absorbing system in
itself. It will always be in competition with any other absorbant.
Moreover, chlorides react chemically with alumina phases of the
cements [1) [11). Some uncertainty prevails, as to whether the
portland cement alone and its hydrates are able to react alike.
To study this, three different cements, normal portland, high
alumina ("white") portland and low alumina portland were chosen.
The characteristics of the cements are shown in table 5. Parts
of the cements were hydrated for three months with a surplus of
water, then dried and ground to normal cement fineness. The un-
hydrated and hydrated cements were mixed with water, containing
a known amount of chloride. During the reaction time of 1, 2, 3,
4, 7 and 30 days the samples were turned around continuously to
avoid hardening. Parts of the solution were filtered and analyzed
for chloride content after the reaction time. From these results
the total amount of absorbed and chemically bound chloride, % by
weight of the cement, was calculated. All results are shown in
graph 8. The general conclusion from the diagram is, that chlo-
rides are bound of cements to a by much higher extent than by cal-
ciumhydroxide. This kind of bond was found with unhydrated and
hydrated cements. It occurs to a different extent with different
types of portland cements.
4. Conclusions
The well known risk of chloride attack to concrete and reinforce-

ment after actual fires with chloride containing plastic materials
involved has sponsored the research for remedial measures. One
of them, the treatment of affected surfaces with lime paste is
widely used. At first, it seemed to have a reducing effect on
chloride concentration in outer concrete layers. But, as shown,
the chloride is not absorbed by the lime, as claimed. In fact
a redistribution of chloride within the concrete takes place.
Tests comparing the absorption of chlorides show the poor effect
of lime in this respect, and cements were shown to bind chlorides
to much higher extent than lime does. The high inner surface of
hardened cement makes it a very good absorband, better by far than
lime. If there should be any chance to extract chlorides from
concrete by absorption, the absorbant must be much more effective
than lime.
5. Bibliography
[1] Richartz, W.
Die Bindung von Chlorid bei der Zementhartung
Zement - Kalk - Gips, 22, 10, S. 447-456 (1969)
(2] Ruffert, G.
Betonschaden durch PVC-Brande und ihre Sanierung
Deutsche Bauzeitung, 5. S. 70-71 (1980)
[3] Frey, R., Weber, B.

Zur Entfernung von in Betone eingedrungenen Chloriden
mittels verschieden temperiertem Wasser
tiz Fachberichte, 106, 6, S. 402-406 (1982)
(4] Kupfer, H.
Entfernung von Chlorid aus Beton durch Elektroosmose
Forschungsvorhaben B I 5- 80 01 81 - 1/330-20
Deutscher AusschuB fur Stahlbeton 1980
[5] Lowicki, C., Schuh, R., Reiter, C., Spalke, F.

Die Kalksanierung. Verfahren zur Sanierung chlor-
wasserstoffgasbeaufschlagter Stahlbeton-Elemente
schaden prisma, 6, 3, S. 37-41 (1977)
[6] Bernhardt, H.
Eindringverhalten von Chlorwasserstoff bzw.
Calciumchlorid in Leichtbeton
schaden prisma, 11, 3, S. 33-37 (1982)
[7] Oehme, A., Bernhardt, H.

Weiterentwicklung der Extraktionssanierung
schaden prisma, 12, 2, S. 26-31 (1983)
(8] Deutscher NormausschuB DIN 52 102

Prufung von Naturstein, 9 (1965)
(9] Deutscher AusschuB fur Stahlbeton

Arbeitskreis Chlorideindringtiefe
Vorlaufige Richtlinie zur Bestimmung der Chlorid-
verteilung in Beton und erganzende Untersuchungen
Fassung Juni 1982
[10] Freundlich, H.
Kapillarchemie, Bd. 1 und 2
(Leipzig 1930, 1932)
[11] Roberts, M.H.

Effect of calcium chloride on the durability of
pre-tensioned wire in prestressed concrete Magazine
of Concrete Research, 14, s. 143-154 (1962)
Lammke 203
Table 1
Chloride distribution in test specimens
Specimen
1 % 2 % 4 %
% NaCl by weight of cement
Chloride content of concrete

(calculated) 0107406 0114813 0129626
% Cl- by weight of concrete
Found content of chloride

in different depths
0 015 em 01 1154 012222 013921
015 - 1 em 011270 012002 013602
1 - 115 em 010897 011604 012945
115 - 2 em 010710 011422 012731
2 - 215 em 010681 011225 012495
215 - 3 em 010630 011128 012525
Table 2
Chloride distribution in test specimens
after 5 lime paste treatments
Specimen
% NaCl by weight of cement
1 %
I 2 %
I 4 %
Found content of chloride

in different depths after
5 lime paste treatments
0 - 015 em 010141 010236 010414
0, 5 - 1 em 010337 010898 011294
1 - 115 em 010542 0 I 12 S] 01 '2046
115 - 2 em 010585 011380 012176
2 - 215 em 010551 011383 012470
215 - 3 em 010621 011407 012583

Table 3
Loss of chloride in treated concrete ./.
total absorption of chloride in used lime
loss of chloride gain of chloride

concrete lime
[g] [g]
specimen
1,3147 0,0352
1 % NaCl
specimen
1,8009 0,1077
2 % NaCl
specimen
3,9241 0,1069
4 % NaCl
Table 4
Chloride distribution of the total concrete cross section
before and after lime treatment [% Cl] by weight of concrete
specimen specimen specimen

layer 1 % NaCl 2 % NaCl 4 % NaCl
em untreated treated untreated treated untreated treated
0 - 1, 1 0,0854 0,0587 0,1744 0,0890 0,3063 o, 1379

1, 1 - 2,2 0,0681 0,0806 0,1326 0,1509 0,2630 0,2761
2,2 - 3,3 0,0665 0,0646 0, 1229 0,1370 o, 2343 0,2816
3,3 - 4,4 0,0633 0,0672 0, 1197 0,1267 0,2305 0,2672
4,4 - 5,5 0,0599 0,0659 o, 1237 0,1271 0,2174 0,2481
5,5 - 6,6 0,0618 0,0592 0,1167 0,1153 0,2213 0,2401
6,6 - 7,7 0,0595 0,0613 0,1242 0,1208 0,2215 0,2281
7,7 - 8,8 0,0624 0,0616 0,1221 0,1171 0,2453 0,2480
8,8 - rest 0,0656 0,0590 0,1327 0,1391 0,2592 0,2749

Lammke 205
Table 5
Chemical and mineral composition of cements
Portland Portland Portland HS

35 F "white" low alumina
Loss on ignition 0,89 2,55 2,59
Insoluble 0,32 0,18 0,51
Sio 2 20,82 21,78 19,86
CaO 63,95 65,64 62,55
MgO 1,52 0,56 2,34
Fe 2 o 3 4,21 0,58 5,70
Ti0 2 0,07 0,08 0,12
Al 2 o 3 3,94 4,38 3,17
so 3 2,93 2,28 2,26
Minerals calcu-
lated according
to Boque
c 3s 61,3 64,9 66,8
c 2s 13,5 13,5 6,6
c 3A 3,3 10,6 -
c 4 AF 12,8 1,8 15,1
C 2F - - 1,3
~
Q u f:
QJ E
--·--!Y-
0 u
z
~
QJ
~-
00 .r:.
---~- ~
"' ....
QJ QJ 0.
c
OJ
c::> E
·..:: u
QJ
~I
I
I ~
I
/\
/
v --~
\
\
'{:
~ ''
~ ''
u
c '
~
0
"'
I .r:.
0.
u
0 ~
sd g
o'
12
d
Ol
~
Q u ·r, E
QJ
I E
z0 uQJ 0
~ I
~ 0 -l ~
.c
~ ~ QJ I a.
c
QJ :J
§ c I u
QJ
E I I
u -1 ~
QJ
lI
/J-,-· -
0.
I
"'
-- 'C
I \
\ ,_,,. \'!
\
\
\
\
f------L
.r:.
I 0.
u 0 ~
sd g
d
12
d
0 Ol
Lammke 207
0
$!,......
u0 I E
z u
Ql
-- I o
(11
E
.....
•u-
:: ~o -~ ~ .c
c "' - o-
c "'E
~ ~
"' ij}
u
~~~ ....
0
VI I I
(J
"I
0
.....
~
,, ,..,
0
!':"' /
u
-- (
I
u
5
...... /' ...... , .c
u a.
;;--0 - ·0
,..,o e
0"1
N
o· "'d 0
;;· g.
~
iJ I E
u I E
~ u
...,~ 2"' ~
lC£
c ~
.l= a.
f
\
"' c .§"'
E
\ u "'
::J \
u I
"'a.
"' I I
I
I
I
I
.... ....
/ .... ....
' ....
.... ....
~
!':
u
c
0
u
I ....
.... ....
....
u
g 9
d d
% Cl /concrele
0,25
Spt'C lm~:>n ?'1, NoCI
-unlreol~:>d
---lime lrrol~:>d
0,15
1\ \ ''
I
" .... /
""
I ....__ /
I
I
~ fv ~ ~ --:: -J
I
I
0. 10
I
Q08
0 10 20 30 '0 so 60 70 80 90 100
deplh (mm)
graph 5
% Cl /conuele
0,)}
spec•mE>n l'/.NoCI
\ .,..- ........
'
- unlrE>OIE>d
--- limE> lr~:>oiE>d
~
I
"' /
'I
0 .•.
I
I
I
I
I
\
I'-- ~
' '-
~
'" 'v
../
v
I
Q2
I
I
I
I
I
I
0, IS
I
I
Q13
0 10 20 30 '0 so 70 80 90 100
graph 6 deplh (mm)
Lammke 209
0
E .'!
E m
5.0 · I - ---~--~--,..----,--·--,---,---.----.----,
~ (.0 1 - - - - - - - - - ·
-~ p
£ lO -·-·- - · ·· - - - - - - - 1 - - - - - 1 - ~-;..-_...,T---1
c
6 ~.0 ----- --· .. - - --~~ - - 1 - ---11---i
v--v--
~ v--
l-g Lv_·.- ·-~.-~=~" " '; !-·~·:::d:-==±::::::::~~:~'l==f==:f==:L-
"'
1.0
0
0
.....
10 .70 50 60 10 eo 90 100
concenlrol10n C I mmol I
obsorpl,onisolhermes for chlorides from aqueous solul,ons

obsorbonls o--a 10n exchange res1n
T - calcium hydrox1de
. - - . activated carbon
graph 7
0.4
c
"'
E
"'u 0.3.
0
.I:
~
"'3: 0.2
>-
.D
"'
"0
0,1
()
:cu
"0
c
::>
0
.D 0
0 5 10 15 20 25 30
lime of reaction I days I
por !land cement
norma I 35 F o--a
high alumino '<"---<>
low alumino o---o
--
hydro led
graph 8

SP 092 PDF

Uploaded by

Copyright:

Available Formats

You might also like

SP 092 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SP 092 PDF

Uploaded by

Copyright:

Available Formats

SP92-l

Synopsis: This paper outlines the procedures adopted in obtaining

Keywords: buildings; concrete construction; damage; fire resistance;

A. K. Tovey is Principal Structural Engineer in the Building

A small number of major fires in concrete structures have

An initial questionnaire, as given in Appendix A, was

Having obtained the contact, the next stage was to obtain

The record summary sheet is divided into six sections. The

that occurred. It may be that either the fire brigade or

As a result of this investigation, detailed information has

Examining the items under damage and repair shows that:

(2) much spalling occurred but most structures still performed

(3) almost without exception the structures performed well

It is hoped that these investigations will provide essential

information gathered contains details of tests on components and

An example of the information that is being obtained may be

Another case which illustrates a detailing deficiency is

The building fires investigated varied from minor fires of

The evidence to date shows all structures to have behaved

INVESTIGATION OF BEHAVIOUR OF STRUCTURE

The working group responsible for the review is seeking to

One such area being investigated is the overall effect of

Another area of concern in the UK is the loss of strength or

The evidence so far indicates that the major reserve in

The authors thank The Concrete Society and the Chairman of

1. The Institution of Structural Engineers., Fire resistance of

2. Forrest, J.C.M., An international review of the fire

4. Malhotra, H.L., Some recent experiences of fires in concrete

5· Cembureau., Concrete for fire resistant construction, Wexham

1. Have you had experience of a fire-damaged structure

2. What type of structure was involved:

3. What was the type of construction:

4. What type of concrete was damaged:

5. How much of the structure was damaged:

6. Was the structure:

7. Please give the name and address of a contact for

8. Are there any other comments you wish to add.

FIRES IN CONCRETE STRUCTURES

Summary Record No, ....

1.1 Name of the building:

2.1 Type of occupancy:

3.1 Date of fire:

4.1 Extent of fire spread (No. of storeys, floor areas involved):

5. STRUCTURAL DAHAGE ASSESSHENT

5.1 Name of assessor:

6.1 Was the damaged part repaired or replaced?

7.1 Has there been a previous fire in the building?

CONCRETE SOCIETY - FIRE RESISTANCE COMMITTEE

Ref Type of Building Type of Construction Damage Repaired ?

Office block In situ hollow pot Localized spalling Yes - reconcreted

Supermarket Prestressed double Spalling of beams Yes - gunite/

Office block In situ waffle Spalling to ground Yes - gunite

Office block In situ RC Spalling to soffit Yes - reconcreted

Office block In situ RC Yes

8 College Precast Localized to Yes

9 Bus Garage Steel frame Severe No - demolished

10 Office block Steel frame Spalling of beams Yes - reconcreted

11 Furniture shop In situ RC Spalling to ground Load test