Transcript
SEB GUIDELINES
SEBGL-OTH7 Guidelines on Structural Fire Engineering Part II: Design of Structural Elements and Assessment of Fire-Damaged Fire-Damaged Structures
STRUCTURAL ENGINEERING BRANCH ARCHITECTURAL SERVICES DEPARTMENT August 2013
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page i of ii
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
Contents 1.
Introduction............................................. ................................................................... ............................................ ........................................ .................. 1
2.
Partial safety factors for loads ............................................ ................................................................... ................................. .......... 2
3.
Thermal Response of Structural Members in a Fire ......................................... 5
4.
Behaviour of Structural Members under Fire ................................................. 12
5.
6.
4.1
Models ........................................... ................................................................. ............................................ ...................................... ................ 12
4.2
Fire Engineering Design of Steel Structure ........................................... 14
4.3
Fire Engineering Design of RC Structure.............................................. 21
4.4
Fire Engineering Design of Composite Structure ................................. 30
4.5
Fire Engineering Design of Timber Structure ...................................... 33
Assessment and Repair of Fire Damaged Structure ........................................ 37 5.1
Purpose of Post-fire Assessment ......................................... ............................................................. .................... 37
5.2
Procedures of Post-fire Assessment ........................................................ 37
5.3
Assessment of Residual Strength ............................................ ............................................................ ................ 51
5.4
Structural Structural Appraisal .......................................... ................................................................. ...................................... ............... 55
5.5
.................................................................. ............................................ ........................ 55 Repair Options ............................................
................................................................. ............................................ ............................... ......... 57 List of References ...........................................
Copyright and Disclaimer of Liability This Guideline or any part of it shall not be reproduced, copied or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without the written permission from the Architectural Services Department. Moreover, this Guideline Guideline is intended intended for the internal use of the staff in the Architectural Services Department only, and should not be relied on by by any third party. No liability is therefore undertaken to any third party. While every effort has been made to ensure the accuracy and completeness of the information contained in this Guideline at the time of publication, no guarantee is given nor responsibility taken by the Architectural Services Department for errors err ors or omissions in it. The information is provided solely on the basis that readers will be responsible for making their own assessment or interpretation of the information. Readers are advised to verify all relevant representation, statements st atements and information with their own professional knowledge. The Architectural Services Department accepts no liability for any use of the said information and data or reliance placed on it (including the formulae formulae and data). data). Compliance with this Guideline Guideline does not itself confer immunity from legal obligations.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page ii of ii
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
1.
Introduction
1.1
There are two approaches for complying with the statutory requirements for fire safety, namely: Prescriptive Provisions and Alternative Approach. For structural members, Prescriptive Provisions specify the material, shape and size, thickness of fire protection materials and construction details to be used in order to satisfy the statutory requirements. Compliance of these provisions is deemed to satisfy the statutory requirements laid down for fire resisting construction for buildings in Part XV of the Building (Construction) Regulations. Regulations. Alternative Approach (or fire engineering approach) is a performance based method, which aims at achieving an “equivalent level” of safet y of the building.
1.2
Part I of this set of guidelines describes the fire scenarios development in a fire, the techniques in fire modelling and the procedures to calculate the maximum gas temperature and duration of a fire in a typical fire engineering study. Part I also mentions that in a small compartment with the usual design fire load, the fire will likely to be fully developed. developed. In such circumstance, it may be safely assumed that the results from a structural fire engineering study will not eliminate the fire protection to steelwork, and project officers are advised to adopt Prescriptive Provisions for the structural members. A structural fire engineering study is particularly applicable for large compartments with high headroom and limited fire load or open-sided buildings (e.g. open-sided car park, sports stadium, indoor swimming pool, public transport concourse in ArchSD projects, and casino or cinemas in the private sector projects), ext ernal structural steelwork located outside the facade of the building, and localised fire which is unlikely to flash over. Should a structural fire engineering study be required, Figure 1 shows the typical steps involved.
Fire Behaviour
Thermal Response of Structural Members
Structural Behaviour Figure 1 Steps in a typical structural fire engineering study (Source: Source: modified from IStructE 2007)
1.3
Part II of this set of guidelines (this “Guideline”) will describe:
a) b) c)
the heat transfer mechanisms from the fire to the structural members, and the procedures to obtain the temperature of the members during a fire; the structural design of steel structure, reinforced concrete, composite structure and timber exposed to fire; and the assessment and repair of fire-damaged structure.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 1 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
As project involving structural fire engineering study will usually engage a specialist fire consultant, this Guideline will only describe the principles and will introduce those simplified methods that can be carried out by hand calculation so as to enable enable PSE to vet the submission from from the consultant. For more detail computation, specialist literature should be consulted. In the end of of this Guideline, a list of reference is therefore provided. 2.
Partial safety factors for loads
2.1
An accurate assessment of the performance of a structural member during a fire requires knowledge of both the reduction in material properties at increasing temperature and an accurate assessment of the “realistic” loads acting on the structure at the time of the fire. BS fire. BS 5950: Part 8: Structural use of steelwork steel work in building. Code of practice for fire resistant design (BSI 2003) was a pioneering work specifying lower partial factors on loads at the fire limit state ( Table 1(a)) than for ambient temperature design. Table 1(a) Partial safety safety factors for loads Partial safety factor Partial safety factor at Load Type at ambient state fire limit state Permanent dead loads 1.4 1.0 Non-permanent 1.6 0.8 imposed loads Permanent imposed 1.6 1.0 loads (Source: Source: BS 5950: Part 8) 8)
2.2
Since then, other national national codes have introduced introduced partial safety factors at fire limit state not only in relation to steel and composite structures but also for concrete members. In the US, ASCE 7-05: Minimum Design Loads for Buildings and Other Structures published by ASCE specifies the partial safety factors for dead load and imposed load are 1.2 and 0.5 (for both permanent or non-permanent) respectively, and in Australia and New Zealand, AS/NZS Zealand, AS/NZS 11702:2002: Structural design actions - Part 0: General principles specifies the corresponding factors as 1.0 for dead load and 0.6 for non-permanent imposed load (and 0.4 for permanent imposed load) (Buchanan 2001). In Europe, BS Europe, BS EN 1990: Basis of structural design (BSI 2002) now specifies the partial safety factor for dead load to be 1.0 and the partial safety factor either 1 or or 2 (Table 1(b)) depending on different use of the floors and frequency of the variable imposed loads for the fire limit state. BS EN 1991-1-2: Actions on structures. General actions. Actions on structures exposed to fire (BSI 2002a) recommends the use of of 2; but the UK National Annex to BS EN 1991-1-2 recommends the use of of 1 (the larger of of 1 and 2).
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 2 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
Table 1(b) Partial safety safety factors for loads for for different use use of floors in BS EN 1990 1990 Load Type 1 2 Imposed loads Domestic, residential 0.5 0.3 Office 0.5 0.3 Congregation areas 0.7 0.6 Shopping areas 0.7 0.6 Storage areas 0.9 0.8 0.7 0.6 Traffic area 30kN 0.5 0.3 Traffic area 30-160kN 0 0 Roofs Wind loads 0.2 0 (Source: Source: BS EN 1990) 1990)
2.3
In Hong Kong, Code of Practice for Structural Use of Concrete 2013 (“Concrete Code 2013”) 2013”) (BD 2013) and Code of Practice for the Structural Use of Steel 2011 (“Steel Code 2011”) 2011”) (BD 2011) 2011) also specify the partial safety factors for loads (Table 1(c)). Though the partial safety factor for non permanent imposed load is 0.8, both codes allow designers to reduce the partial safety factor for non-permanent imposed loads to 0.5 when suitable justification is available. Table 1(c) Partial safety factors for different types of load Partial safety factor at fire Load Type limit state Dead loads 1.0 Permanent imposed loads 1.0 Non-permanent imposed loads 1) in escape stairs and lobbies 1.0 2) all other areas 0.8* Wind loads 0.33 Note: The value may be reduced to 0.5 when suitable justification j ustification is available. (Source: Source: Concrete Code 2013 and Steel Code 2011)
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 3 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
2.4
Summary of partial safety factors for loads Based on the above codes, this Guideline therefore recommends the partial safety factors for loads in Table 2. Table 2 Recommended Recommended partial safety factors factors for different different types of load Partial safety factor at fire Load Type limit state Dead loads 1.0 Permanent imposed loads 1.0 Non-permanent imposed loads 1) in escape stairs and lobbies 1.0 2) all other areas 0.7# Wind loads 0.33 #
2.5
1. Concrete Code 2013 and Steel Code 2011 2011 specifies a range range from 0.5 to 0.8, and a factor of 0.7 is therefore recommended as it covers domestic, residential, office, congregation areas, shopping areas, car-parks for vehicles less than 3t (i.e. most of the use of premises of ArchSD), as recommended in the UK National Annex to BS EN 1991-1-2. 1991-1-2 . This factor is also more conservative than those recommended in ASCE 7-05 and AS/NZS 11702:2002. 2:2002. 2. For storage areas, this this Guideline recommends recommends a factor of 1.0, as the imposed imposed load can be regarded as permanent one.
Example The following is an example showing the loading to be applied to the slabs of a hypothetical structure for residential use during a fire with: Dead load gk = 5.9 kPa Imposed load q k = 3.0 kPa Ambient temperature design load nk = 1.4×5.9+1.6×3.0 = 13.06 kPa Partial safety factors at fire limit state for: dead load = 1.0 imposed load = 0.7 Fire limit state design load n k = 1.0×5.9+0.7×3.0 = 8.0 kPa. Therefore, reduction factor ηfi for the design load level in fire limit state = 8.0 / 13.06 = 0.61. The design load at fire limit state is therefore about 61% of the design load at ambient temperature, and this is a very significant reduction, especially in adaptive reuse of of historical buildings. For such adaptive reuse, Carbon Fiber Reinforced Polymer (CFRP) or Glass Fiber Reinforced Polymer (GFRP) has now increasingly been employed employed as strengthening works. Although such reinforcing polymer is not fire resisting and can only be included in strength calculation during ambient temperature; the structural member without such reinforcing polymer may still be adequate at fire limit state as the design load at fire limit state is about 60% that at ambient limit state.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 4 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
3.
Thermal Response of Structural Members in a Fire
3.1
Heat Transfer Mechanisms
3.1.1 To carry out a structural fire engineering calculation to investigate the structural response of the building, it is first important to obtain an accurate estimate of the temperature gradient through the structural members. The temperature distribution through a structural member is dependent on: a) b)
the transfer of heat by convection and radiation from a fire to a member, and the transfer of heat by conduction within a member.
The following paragraphs will therefore explain the principles in calculating the temperature gradient through a structural member with heat transferred by convection, radiation and conduction during a fire. 3.1.2 Convective and radiative heat flux Through convection and radiation, net heat flux h
net
comes from the fire to the
member. BS member. BS EN 1991-1-2 gives the following equations to calculate h net : h net,c α c θ g θ m
h net,r Φ ε m ε f σ θ r 273 θ m 273 4
4
h net h net,c h net,r where
2 h net,c is the net heat flux due to convention (W/m ); 2 h net,r is the net heat flux due to radiation (W/m );
αc
θg θm εm
εf Φ σ
is the coefficient of heat transfer by convection, which may be taken as 25W/m 2K and 35W/m 2K for standard fire and parametric fire respectively; is the gas temperature in the vicinity of the steel member (°C); is the surface temperature of the member (°C); is the surface emissivity of the member and can be taken as the following values: Steel 0.7 Concrete 0.7 Others 0.8 is the emissivity of fire which may be taken as 1.0; is the configuration factor (W/m2K); and is the Boltzman constant = 5.67×10 -8 W/m2K 4.
The configuration factor Φ depends on the position and shadow effects of the fire exposed surface of the member, and can be calculated as follows; Φ
2
Y tan 1 1 X2 2 π 1 X X
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
X Y tan 1 2 1 Y2 1 Y
Page 5 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
where X = A/2C and Y = B/2C with A, B and C defined in Figure 2(a). Yet, rather than carrying out the calculation, Φ may be taken conservatively as 1.0W/m2K (IStructE 2007).
Figure 2(a) Calculation of configuration of configuration factor Φ (Source: Source: BS EN 1991-1-2) 1991-1-2)
3.1.3 Temperature distribution due to conduction Heat transfer by conduction in a structural member is governed by governed by the Fourier’s equation which states that the quantity of heat transferred per unit time across an area A is proportional to the temperature gradient as follows: T q -kA x where A is the area across which heat is transferred (m2); k is the thermal conductivity of the material (W/mK); q is the heat transfer across the area A (W); T is the temperature (K); and x is the distance normal to the area A (m). 3.2
Temperature of Structural Members in a Fire
3.2.1 In Part I of this set of guidelines, two main types of fire have been distinguished, namely standard fire and parametric fire. Part I of this set of guidelines then gives the details in finding the temperature-time graphs for both standard fire and parametric fire. With the temperature-time graph of a fire and the heat transfer equations in Section 3.1, the temperature distribution of each structural element under fire can be found. 3.2.2 To ease the tedious calculation, simple design charts derived from standard fire tests are given in literature defining the temperature distribution through members during standard fire. For parametric fires, either computer analysis or empirical calculations can be used. For computer analysis, commercial computer software usually employs finite element or finite difference techniques, and requires the setting of the boundary conditions (including the heat sources by a temperature-time function), material thermal properties of conductivity, specific heat and emissivity. As this will inevitably require the knowledge, skill and judgment of the specialist fire consultant, it will not be described further further in this Guideline. Guideline. The following following paragraphs paragraphs will, instead, describe the simplified empirical calculations that can be used to get the
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 6 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
maximum temperature of the structural members during the fire, which therefore serves as a quick quic k way to check the validity of the computer anal ysis. 3.2.3 Steel members 3.2.3.1 As steel has a high thermal conductivity, it may be sufficiently sufficientl y accurate to ignore thermal gradients within members and assume an average uniform steel temperature θa. A lumped lumped mass model is therefore adopted. According to BS to BS EN 1993-1-2: Design of steel structures. General rules. Structural fire design (BSI 2005a), the rise in temperature Δθa,t of unprotected steel members at time t in an interval of Δt of Δt is given by the following equation:
A m /V h net Δt ρ c a a
Δθ a, t k sh
where h net is the net heat flux (W/m2); Am
and
is the surface area of the member exposed to fire per unit length (m2/m); V is the volume of the member per unit length (m 3/m); ca is the specific heat of steel at elevated temperature (J/kg/K); ρa is the density of steel at elevated temperature (kg/m3); Δt is the time interval (s); Am/Vis the section factor for unprotected steel ste el per unit length (m-1/m); k sh is the correction factor for the shadow effect. sh
3.2.3.2 The specific heat capacity capac ity of steel ca varies with temperature, and BS EN 1993-1-2 gives the following equation for c a at temperature below 600oC: ca = 425 + 0.773× θa – 0.00169× – 0.00169× θa2 + 2.22×10-6× θa3(J/kg/K) Again, rather than calculating their values as temperature rises, BS rises, BS EN 1993-12 recommends that the specific heat ca may be considered to be independent of the steel temperature, and an average value of 600 J/kg/K may be used. The correction factor k sh sh for shadow effect is to account for the reduced effectiveness to raise the temperature of steel member. In BS In BS EN 1993-1-2, 1993-1-2 , the correction factor for shadow effect is given by: for I-section: for other sections:
A m V b A m k sh sh= 1.0 V b k sh sh= 0.9
A m V A / m V /
A m is the box value of the section factor. The box value for A m is V b
where
the perimeter that covers the periphery of an imaginary box that surrounds the section completely. Table 3 shows the section factor for some unprotected steel members. An example will be followed to demonstrate demonstrate the steps involved in the calculation.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 7 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
A m for different types of structural steel members V b
Table 3 Values of
(Source: Source: BS EN 1993-1-2) 1993-1-2)
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 8 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
3.2.3.3 Example Consider a simply supported Grade S275 356×171×67kg/m UB of span 6m supporting rc slabs of 150mm 150mm thick. Calculate the temperature of the beam subjected to a standard fire BS fire BS EN 1991-1-2 at time 600s. The ambient temperature is assumed to be 20 oC, and the gas temperature g at time t (s) is given by: θg = 20 + 345 log (8t (8 t + + 1) The heat flux to the steel beam by: convection h net,c = αc (θg – θ – θm) = 25 (θ ( θg – θ – θm) radiation h net,r = Φ εm εf σ [(θr + 273)4 – (θ – (θm + 273)4] Change in steel temperature, Δθa,t is given by:
A m /V h net Δt c a ρ a
Δθ a, t k sh
Here, Am/V(/m) = 163.9/m (for three sides exposure), = 7850 kg/m 3, k sh sh = 1. Using Excel spreadsheet, the temperature-time of the steel member till t=600s is shown in Table 4. Therefore, the steel temperature at 600s = 514.4 °C. Table 4 Temperature of steel members subjected to standard fire Time t (s)
g (°C)
h net,c (W/m2)
h net,r (W/m2)
hnet (W/m2)
0
20.0
5
96.5
1913.4
447.6
2361 2361
439.8
0.6
20.6
10
147.0
3159.8
939.7
4099
440.2
1.0
21.5
15
184.6
4076.9
1441.7
5519
440.9
1.3
22.8
20
214.7
4795.9
1940.9
6737
441.8
1.6
24.4
25
239.7
5381.8
2431.9
7814
442.9
1.8
26.3
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
570
670.8
4518.3
18038.1
22556
659.3
3.6
493.7
575
672.1
4461.3
17957.1
22418
661.8
3.5
497.2
580
673.4
4405.0
17874.3
22279
664.4
3.5
500.7
585
674.7
4349.2
17789.9
22139
667.0
3.5
504.2
590
675.9
4294.0
17703.9
21998
669.7
3.4
507.6
595
677.2
4239.5
17616.3
21856
672.3
3.4
511.0
600
678.4
4185.6
17527.3
21713
674.9
3.4
514.4
ca (J/kg/K)
Δθa,t (°C)
(°C) 20
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 9 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
3.2.3.4 Similar to unprotected steel, an empirical calculation method is presented in BS EN 1993-1-2 to calculate the incremental increase in temperature for a steel element protected with a spray or board material, which is related to the specific heat, thermal conductivity, density and moisture content of the protection material. Fire protection materials material s insulate steel ste el structures from the effects of the high temperatures that may be generated in fire. There are three common types of fire protection materials to steel, namely, fire protection boards, sprays, and intumescent coatings. Clauses 15.66 to 15.72 of the General Specification for Building (2012 Building (2012 edition) issued by ArchSD detail the requirements and application procedures of these three types of fire protection materials. Among the three, sprayed protection systems (Figure 2(b)) are the cheapest, and should therefore therefore be the first choice whenever possible. Its principal disadvantage is that the sprayed surfaces are not visually appealing. Fire protection boards offer a clean, boxed appearance which may be prefinished or suitable for further decoration; but are highest cost among the three. Unlike sprays and boards (which are non-reactive systems), intumescent coatings are a reactive system providing insulation by swelling using a charred layer of low conductivity materials such that the steel will not be affected during a fire. For fire-protected steel using any one of the three systems, Prescriptive Provisions will usually be followed, and the specialist contractors will then carry out the calculation to find the thickness of protection that can satisfy the required fire resistance rating. A structural fire engineering study will therefore seldom be used.
Figure 2(b) Sprayed fire protection system
3.2.4 Concrete members Due to the low thermal conductivity, high thermal gradients will occur through concrete members, which together with the effects of the mass transport of water or water vapour, make estimating the temperature distribution through the members very difficult. Usually, it is difficult to calculate the temperature distribution within a rc member using hand calculation. In 1978, 1978, IStructE published Design and Detailing of Concrete Structures for Fire Resistance (IStructE 1978), and this publication was a pioneering work giving temperature profiles for both flat soffit slabs and beams exposed to the standard fire. Annex A of BS of BS EN 1992-1-2: Design of concrete structures. General rules. Structural fire design (BSI 2004a) now gives the temperature profiles (extracts shown in Figure 3) during a standard fire for slabs, beams and columns. One of the main limitations of such temperature profiles is that they are only applicable to a
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 10 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
standard fire. As such, time-equivalence time-equivalence mentioned in Part I of this set of guidelines may be useful, and this will be elaborated in latter paragraph in this Guideline.
(a)200mm (a) 200mm thick slabs at time 30, 60, 90, 120, 180 and 240 mins
(b) 600×300 beam at time 90 mins
(c) 300×300 column at time 60 mins Figure 3 Temperature profiles profiles in rc members during during a standard fire (Source: Source: BS EN 1992-1-2) 1992-1-2) 3.2.5 Composite and timber members members Similar to rc members, the thermal gradients through composite members and timber members are difficult to be estimated. Simplified methods have therefore been developed for defining the temperature within such members. These simplified methods will be elaborated in latter paragraph in this Guideline.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 11 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
4.
Behaviour of Structural Members under Fire
4.1
Models
4.1.1 Figure 4 shows the possible models available for a structural fire engineering design. The following three models that can be used used to predict the behaviour of structural elements in a fire: a) b) c)
member analysis; frame analysis (i.e. analysis of part of the structure); and whole building analysis (i.e. analysis of entire structure).
Figure 4 Models for structural fire engineering design (Source: Source: modified from BS from BS EN 1994-1-2) 1994-1-2)
4.1.2 Member analysis and frame analysis analysis Member analysis and frame analysis are the simplest methods for predicting the structural behaviour of buildings, which analyse individual members at the fire limit state using partial load and material factors. Hand calculation is possible with these methods. However, in these methods the overall stability of the building in a fire fir e should be considered. For braced frames no additional checks are normally required provided a sufficient number of cores or bracing, that provide the lateral resistance, have adequate fire resistance, shielding or containment within fire resisting cores. For sway frames, a frame analysis at elevated temperatures is required to ensure sufficient overall stability during a fire. 4.1.3 Whole building analysis analysis 4.1.3.1 Traditionally, the behaviour of structural members has been investigated by standard fire test. However, the difference between the behaviour of isolated members and the behaviour of the entire building can have a beneficial or detrimental effect on the overall fire resistance of the building. In the last two decades, large-scale fire tests have been carried out in at the BRE test facility (Figure 5) in Cardington in the Borough of Bedford of the UK to investigate whole building behaviour during fire. Compartment fire tests were conducted Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 12 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
on the full-scale steel-composite, timber and concrete buildings, which aimed at identifying modes of structural behaviour that cannot be identified from the standard fire tests.
Figure 5 Full-scale test model of a timber building at the BRE Cardington Laboratory during fire (Source: Source: Lenon 2003)
4.1.3.2 One of the most important observations for composite framed structures from the full-scaled fire tests is that such structures possess great reserves of strength by adopting large deflection configurations due to the mechanism of membrane action (Figure 6). This mechanism is able to transfer the applied loads on the floor to the supporting columns when the supporting steel beams to the floor are damaged and have lost their load carrying capacities. The understanding of tensile membrane action has a significant implication for the fire protection of steel-framed buildings with rc floors. This is because the supporting steel beams may be allowed to fail without endangering the safety of the whole building. Consequently, fire protection to these steel beams can be reduced or eliminated (Wang 2002). 2002).
Figure 6 Membrane action during fire (Source: Source: Newman et al 2000) al 2000)
4.1.3.3 Whole building analysis is now one of the models that can be used in a structural fire engineering design. The model takes into account the membrane action of floor slabs, and allows the beneficial effect of the grillage of beams and floor slab, acting as a unit, to be included within the structural design. Commercial softwares are usually required to carry out whole building analysis. Based on the results the BRE Cardington tests, SCI published Fire published Fire
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 13 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
Safe Design: a New Approach to Multi-storey Framed Buildings (Newman 2000) giving as a series of tables which are applicable to structures of similar construction (i.e. non-sway frames with composite steel decks). 4.2
Fire Engineering Design of Steel Structure
4.2.1 This section will describe the member analysis, which can calculate the t he performance of steel structures exposed to fires by hand calculation. The design based on individual simple members using member analysis (without considering quantitatively the effects or restraints provided by adjacent members) is generally more conservative, and can therefore serve as a quick check on the results generated from computer software. The design code code will be referenced from BS EN 1993-1-2, 1993-1-2, as Steel Code 2011 only provides broad guidelines on the detail design. 4.2.2 Mechanical properties of steel at elevated temperatures The mechanical properties of steel, including the proportional limit, yield strength and elastic modulus, at elevated temperature must be known prior to the design. Figure 6 shows the stress strain curve for steel at elevated temperature.
Figure 6 Stress-strain Stress-strai n curve for steel at elevated temperature (Source: Source: BS EN 1993-1-2) 1993-1-2)
Two critical parameters are to determined for structural design: a) effective yield strength f y,θ y,θ; b) modulus of elasticity at linear elastic range Ea,θ. BS EN 1993-1-1 relates to the values of these two parameters to their respective values at the ambient temperature of 20° by applying a corresponding reduction factor to account for the effect of elevated temperature as follows: f y,θ y,θ = k y,θ y,θ f y Ea,θ = k E,θ E,θ Ea where
f y is the yield strength at 20°C Ea is the slope of the linear elastic range at 20°C
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 14 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
k y,θ y,θ is the reduction factor for yield strength k E,θ E,θ is the reduction factor for the linear elastic slope Table 5 shows the reduction factors to be applied to the characteristic yield strength and elastic modulus of structural steel at elevated temperatures, and Figure 7 plots the reduction of these two properties with temperatures. Table 5 Reduction factors for mechanical properties of hot rolled steel at elevated temperatures Reduction factors relative to the values at 20°C Steel temperature Effective yield strength Modulus of elasticity θa k y,θ k E,θ y,θ = f y,θ y,θ / f y E,θ = Ea,θ / Ea 20°C 1.000 1.000 100°C 1.000 1.000 200°C 1.000 0.900 300°C 1.000 0.800 400°C 1.000 0.700 500°C 0.780 0.600 600°C 0.470 0.310 700°C 0.230 0.130 800°C 0.110 0.090 900°C 0.060 0.0675 1,000°C 0.040 0.0450 1,100°C 0.020 0.0225 1,200°C 0.000 0.0000 (Source: Source: BS EN 1993-1-2) 1993-1-2)
Figure 7 Variation of yield str strength ength and modulus of elasticity elasticity of hot rolled steel with temperature (Source: Source: modified from Lennon et al 2007) al 2007)
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 15 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
4.2.3 Design of Steel Members Exposed to Fire 4.2.3.1 The design can now be carried out by ensuring that a structural member under fire will have design effect of actions in fire condition Ed,fi less than that at ambient temperature Ed, given by the following equation: Ed,fi = ηfi × Ed where ηfi is the reduction factor for the design desi gn load level in fire condition. The reduction factor ηfi can be estimated according to the respective ratio of dead load and live load. The following simplified formula can be used for estimating ηfi: G k ψ i Q k,1 ηfi = γ G G k γ Q,1Q k,1 where Qk,1 Gk γG γQ,1 ψi
is the characteristic imposed load; is the characteristic dead load; is the partial safety factor for dead load at ambient temperature; is the partial safety factor for imposed load at ambient temperature; and is the partial safety factor (Section 2) for imposed load under fire.
The following three methods are available in checking whether the design resistance R fi,d,0 fi,d,0 for a member under tension, compression and bending is adequate to resist the design effect of actions in fire condition Ed,fi: 1. limiting temperature method; 2. critical temperature method; and 3. moment capacity method; 4.2.3.2 Limiting Limiting temperature method BS 5950: Part 8 includes the limiting temperature method providing a very simple but effective procedure using the concept of load ratio – that is, the ratio of the load carried during the fire to the ambient temperature load capacity – capacity – to derive a limiting temperature, which is then compared with the design temperature to assess the need for passive protection. Figure 8 shows the flow chart in carrying out the calculation.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 16 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
Figure 8 Flow chart showing the steps steps in limiting temperature temperature method (Source: Source: Wang 2002)
For a steel beam under bending, the load ratio R is defined as: M R f Mc where Mf is the maximum bending moment in the beam under fire condition and Mc is the bending moment capacity of the beam at ambient temperature. If a steel beam is uniformly heated and can undergo large strains before failure, BS 5950: Part 8 gives the limiting temperature of a beam with the limiting temperature-load ratio relationships (Table 6). The example in Section 2 shows that the design load at fire limit state is about 60% of the design load at ambient temperature, and hence the limiting temperature is therefore about 550oC, which serves as the usual rule of thumb for steel beams in bending during a fire. Table 6 Limiting temperatures temperatures of steel beams beams in bending Load ratio
0.7
0.6
0.5
0.4
0.3
0.2
0.1
Limiting temperature (°C)
520
555
585
620
660
715
N/A
(Source: Source: BS 5950: Part 8) 8) Similar tables are available for columns in compression and members in tension. The limiting temperature method can also be used for beam-columns and cold-formed sections (Wang 2005).
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 17 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
4.2.3.3 Critical Critical temperature method The critical temperature method is given as BS as BS EN 1993-1-2, 1993-1-2 , and is in fact the same method as the limiting temperature method in BS 5950: Part 8 . The critical temperature is the maximum temperature that the beam can sustain and the critical temperature method. In the critical temperature method, method, it is necessary to calculate the degree of utilization o of the beam, which the ratio of the applied maximum bending moment in the beam at the fire limit state to the beam’s bending moment capacity at ambient temperature, temperature , given in the following equation: E fi,d μ0 . R fi,d,0 Indeed, the degree of utilization o is the same as the load ratio R in BS in BS 5950: Part 8. Then, the critical temperature θa,cr of a steel member is calculated using the following equation:
1 482 1 3.833 0.9674μ 0.9674μ 0
θ a,cr 39.19 ln
Provided that the temperature of the steel member during fire to be less than the critical temperature θa,cr of the steel member, the steel beam can survive the fire. 4.2.3.4 Moment Moment capacity capacit y method The moment capacity method is based on the known temperature of the critical element with the relevant strength reduction factor used. If the moment capacity does not exceed that applied at the fire limit state then the beam does not require protection. BS EN 1993-1-2 contains formulae in finding the moment capacity of structural members under tension, compression and bending. This method is, however, not widely used, as knowledge of the temperature profile of the beam is required. Indeed, a usual assumption in the design of steel under fire is that the steel member is of uniform temperature, and as such either the limiting temperature or the critical temperature method can achieve the purpose. 4.2.4 Example Consider a simply supported Grade S275 356×171×67kg/m UB of span 6m carrying a design uniformly distributed load of dead load of 20kN/m and non permanent imposed load of 18kN/m at ambient temperature. Calculate the critical temperature of the beam under a fire. Loadings under fire dead load Gk = 20kN/m imposed load Qk = 18kN/m Ambient temperature design load Nk = 1.4×20+1.6×18 = 56.8kN/m Partial safety factors for: dead load = 1.0 imposed load = 0.7
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 18 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
Therefore, fire limit state design load Nk = 1.0×20+0.7×18 = 32.6 kN/m Under fire, the reduction factor η fi is G k ψi Qk,1 32.6 ηfi = 0.57 γ G G k γ Q,1Qk,1 56.8 Assuming that Ed,fi = R fi,d,0 at ambient temperature, i.e. o or R = 1.0 at ambient temperature, the adequacy of the structural members due to bending is checked by: Limiting temperature method: The load ratio R under fire = 0.57, and the limiting temperature = 564oC. Critical temperature method: The degree of utilization μ0 under fire = 0.57, and the critical temperature 1 = 39.19 ln 1 482 = 529°C 3.833 0.9674 0.57 4.2.5 Connections 4.2.5.1 Traditionally, it was commonly assumed that there was no need to take special provisions for the connections as long as they are protected at least in the same manner as the adjacent members that they connect (Franssen et al 2009). al 2009). This implies that, if none of the connected members is protected, then there is no need to protect the joint. This concept was based on the idea that the thermal massivity of the joint should be higher than the massivity of the members because of the presence of additional mass in i n the connection zone, either from f rom end plates, fin plates, web cleats or stiffeners. Therefore, the joints are usually at a temperature lower than the temperature of the connected elements at fire (Figure 9). It was also based on the observation of numerous unprotected steel structures that completely collapsed in severe fire, and where the steel beams were severely distorted, but rarely detached from the columns (Franssen et al 2009). al 2009).
Figure 9 Temperature distribution distribution of steel structure structure at fire (Source: Source: modified from European Erasmus Mundus Master Course, Education, Audiovisual and Culture Executive Agency)
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 19 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
4.2.5.2 However, the behaviour of connections in a fire or post-fire condition may be critical in maintaining overall structural stability. The importance of connections has been observed and documented in the damaged WTC buildings around Ground Zero as a result of fires initiated after the collapse of twin-towers. BS EN 1993-1-2 now provides the following two methods to calculate bolted or welded joints during fire: a)
The first is based on ensuring that the fire resistance of the joint is greater than or equal to that of the connected members. In general, this is a conservative method as the temperature of the connection is generally less than that of the beams (Lennon et al 2007). al 2007). However, it is also necessary to consider the utilization of the connection compared to the utilization of the member. As a simplification, the utilization of the joint and the connected members may be related to the loading and resistance at ambient temperature.
b)
The second method is to find the resistance of the joint according to Annex D of BS of BS EN 1993-1-2 whereby the temperature of the components are calculated and reduction factors used to determine the resistance of the joint. Annex D of EN of EN 1993-1-2 provides the following equation for calculating the temperature distribution through the joint for beam depths less than 400 mm: h 0.88θ 0 [1 0.3( )] θ h 0.88θ D where h is the temperature at height hei ght h (mm) of the steel beam ( oC); 0 is the bottom flange temperature of the steel beam remote from the connection ( oC); h is the height of the component component being considered above the bottom of the beam (mm); and D is the depth of the beam (mm). Once the temperature distribution has been derived then the capacity in shear, bearing and tension is calculated using appropriate reduction factors to allow for the effects of elevated temperature.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 20 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
4.3
Fire Engineering Design of RC Structure
4.3.1 Mechanical properties at elevated temperature 4.3.1.1 Strength Strength of concrete Figure 10(a)(i) plots the reduction of the strength of different types of concrete with temperatures. Concrete Code 2013, following following BS BS EN 1992-1-2, 1992-1-2, specifies the reduction factors to be applied to the characteristic strength of concrete at elevated temperatures (Table 7).
Figure 10(a)(i) Strength reduction of different types of concrete at elevated temperatures (Source: Source: modified from Lennon et al 2007) al 2007)
Table 7 Reduction factors factors for strength of normal normal weight concrete with siliceous aggregate at elevated temperatures Concrete temperature
Strength reduction factors
20°C 1.00 100°C 1.00 200°C 0.95 300°C 0.85 400°C 0.75 500°C 0.60 600°C 0.45 700°C 0.30 800°C 0.15 900°C 0.08 1,000°C 0.04 1,100°C 0.01 1,200°C 0.00 (Source: Source: Concrete Code 2013 and BS and BS EN 1992-1-2) 1992-1-2)
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 21 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
4.3.1.2 Strength Strength of steel reinforcement Table 8 shows the reduction factors for steel reinforcement in Concrete Code 2013, while BS EN 1992-1-2 specifies to follow the reduction factors for structural steel in BS EN 1993-1-2. 1993-1-2. Table 8 also shows the corresponding values in BS EN 1993-1-2. 1993-1-2. PSE may note that the values given given in Concrete Concrete Code 2013 are more conservative than those in BS EN 1993-1-2. 1993-1-2. There is no distinction made in the reduction factors for mild steel bars and high yield bars (Grade 460 or Grade 500) in Concrete Code 2013 and BS EN 1993-1-2, 1993-1-2 , though BS though BS EN 1993-1-2 gives another set of reduction factors for cold-worked bars. Table 8 Reduction factors for steel reinforcement at elevated temperatures Reduction factors relative to the values at 20°C Steel temperature Concrete Code 2013 BS EN 1992-1-2 1992-1-2 20°C 1.00 1.00 100°C 1.00 1.00 200°C 1.00 1.00 300°C 1.00 1.00 400°C 0.87 1.00 500°C 0.60 0.78 600°C 0.36 0.47 700°C 0.11 0.23 800°C 0.08 0.11 900°C 0.06 0.06 1,000°C 0.04 0.04 1,100°C 0.02 0.02 1,200°C 0.00 0.00 (Source: Source: Concrete Code 2013 and BS and BS EN 1993-1-2) 1993-1-2)
4.3.1.3 Strength of prestressing steel s teel wires Figure 10(a)(ii) plots the reduction of the strength of different types of prestressing steel wires in prestressed concrete with temperatures specified in BS EN 1992-1-2. 1992-1-2. The reduction factors of prestressing steel wires are higher value than those of hot hot rolled reinforcements. Moreover, upon cooling the strength of prestressing steel wires will only partially be recovered.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 22 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
Notes:
1a – 1a – cold cold worked prestressing steel wires (Class A) 1b – 1b – cold cold worked prestressing steel wires (Class B) 2 – quenched – quenched and tempered prestressing steel bars
Figure 10(a)(ii) Reduction factors for strength of prestressing steel at elevated temperatures (Source: Source: BS EN 1992-1-2) 1992-1-2)
4.3.2 Load carrying capacity 4.3.2.1 To determine the load carrying capacity of a concrete member, the temperature profile of the member should be known in order to evaluate the reduction in the cross section area and the steel temperature. However, as stated in the earlier paragraphs, high thermal gradients will occur through concrete members due to their low thermal conductivity. Thus, if Prescriptive Provisions are not followed, the temperature distribution through the members usually requires the use of finite elements or finite difference programs in commercial softwares. Instead of using computer analysis to find the thermal gradient through concrete members, the earlier paragraphs have mentioned that Annex A of BS BS EN 1992-1-2 gives the temperature profiles (Figure 3) for standard fire for slabs, beams and columns. 4.3.2.2 500°C 500°C isotherm method As an alternative, BS alternative, BS EN 1992-1-2 provides the simplified method of “500°C isotherm method” for rc beams and columns. columns. This simplified method uses a general reduction of cross-section with respect to a heat damaged zone at the concrete surface. At the earlier paragraphs, it was mentioned that typically 500°C – 500°C – 600°C steel and concrete has reduced in strength to the point where their residual strength, which provides the factor of safety assumed for design at ambient state, has been lost (Lennon et al 2007). al 2007). That is, the load carrying capacity at around 500°C – 500°C – 600°C 600°C is unable to carry the ultimate load of 1.4G k and 1.6Qk , and is only able to sustain a load of load of 1.0G k and 0.7Q k at fire limit state. The concrete with temperature higher than 500°C is therefore assumed to have no contribution to the strength strength of section during fire. Annex A of BS BS EN 1992-1-2 gives the 500°C isotherms (extracted shown in Figure Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 23 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
10(b)) for different types and and sizes of of members under standard fire. The remaining concrete with temperature less than 500°C is assumed to have full strength and and stiffness as in ambient temperature. The temperature of the reinforcement is then assessed using the temperature profiles given in Annex A of BS of BS EN 1992-1-2 (extracted shown in Figure 3). The reinforcement located beyond the reduced section can still be used in calculating the strength of the member, as the damaged concrete in excess of the 500 oC value is assumed to retain its insulation properties in terms of providing the required cover to the reinforcement. The strength of steel reinforcement should be reduced according to the temperature at the centroid of the steel bar. An example will be given at the end of this section to illustrate the 500°C isotherm method.
(a) 300×300 column at time 30, 60, 90, 120, 180 and 240 mins
(b) 300 column at time 30, 60, 90, 120, 180 and 240 mins
(c) 300×160 beam at time 30, 60 and 90mins Figure 10(b) 500°C isotherms of rc members during a standard fire (Source: Source: BS EN 1992-1-2) 1992-1-2) 4.3.3 Time equivalent
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 24 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
The curves in Figure 3 and Figure 10(b) are obtained in a standard fire, and in (“tPart 1 of this set of guidelines, the concept of time-equivalence (“tequivalence”) has been introduced. The “t“t-equivalent” is “the exposure time in the standard fire resistance test which gives the same heating effect on a structure as a given compartment fire” (Law 1997). 1997). T-equivalent is to relate the exposure of a structural element in a real fire to an equivalent period of heating in the standard fire resistance test (Figure 11). Thus, t-equivalence can be used for relating a real fire to the standard fire.
Figure 11 Graphical representation of t-equivalence t-equival ence
BS EN 1991-1-2 gives the following expression to calculate the time equivalent: te,d= (qf,d ×k b ×wf )×k c where qf,d = design fire load density (MJ/m 2); wf = ventilation factor to take into account vertical and horizontal openings=(6/H)0.3[0.62+90(0.4-α [0.62+90(0.4-αv)4] in the absence of horizontal openings; k c = factor dependent on material=1.0 for protected steel and reinforced concrete; H = the height of the compartment (m); αv = Av/Af ; Av = the total area of the opening; Af = the total floor area; account the thermal properties of the the enclosure and k b = factor to take into account = 0.7 when there are no horizontal openings and bounding surfaces are unknown, or when the bounding surfaces (and hence the thermal inertia b (= λρc )) are known: Thermal inertia b (= λρc ) (J/m²s K)
k b (min. m²/MJ)
≥2500 ≥720 to ≤2500 720
0.04 (0.055) 0.055 (0.07) 0.07 (0.09)
½
Note: Values in in brackets brackets are those given given in the UK National National Annex Annex to BS to BS EN 1991-1-2. 1991-1-2.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 25 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
4.3.4 Example Consider a compartment fire with a parametric fire curve as shown in Figure 12, and a rc simply supported floor beam of 600 × 300 with span 8 m carrying a dead load of 6 kN/m (including its self weight) and an imposed load of 10 kN/m. Determine the beam’s flexural capacity at maximum gas temperature and justify whether the beam is stable or not if only 3T25 is provided at a parametric fire as with the temperature-time graph shown in Figure 12 and with the following parameters: qf,d = design fire load density = 753MJ/m2 wf = 2.364 k c = 1.0 and k b = 0.5. The characteristic concrete cube strength f cu cu is 30 MPa, the characteristic yield strength of reinforcement f y is 460 MPa and the concrete cover is 40 mm.
Figure 12 Parametric fire curve of the compartment
The equivalent fire severity te,d using the T-equivalent model is given by: te,d = qf,d × k b × wf × k c = 753 × 0.055 × 2.364 × 1.0 = 97.9 min (say 90 min) From Figure 3, the 500 oC isotherm will be at 30mm from the concrete surface, and hence the width b of the beam at 90 mins = 300 - 2×30 = 240mm. Distance from concrete surface to the centroid of the steel bars = 40+25/2 = 52.5mm. The maximum temperature of the steel bars will occur occur at the corner of
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 26 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
the section, and using Figure 3, the temperature of such bar will be about 500 oC while the temperature of middle bar will be about 340oC. From Table 8 and using the values given in Concrete Code 2013, the characteristic yield strength of the corner bars at 500oC = 0.60×460 = 276 MPa and that of the middle bars at 340 oC = 0.95×460 = 437 MPa. Therefore, the weighted average strength of bars at this fire condition = (276×2+437)/3 =330MPa. The design of the moment capacity of the beam at 500 oC is the same as that at ambient temperature with the following parameters: b=240, d=600-52.5=547.5, d=600-52.5=547.5, f y=330MPa, f cu cu=30MPa, A s=3×491=1472 Therefore, z = d (0.5 0.25 k/0.9) where d=M/bd2fcu, giving z=482mm The moment capacity of the beam at 500 oC = 0.87×330×1472×482×10 0.87×330×1472×482×10 -6 = 204kNm Fire limit state bending moment in elevated temperature
M ult
1 8
(1.0 6 10.7 10) 82 104kNm
Since the fire limit state bending moment 104kNm, which is less than 204kNm, the beam can survive the parametric fire with maximum gas temperature around 800oC. 4.3.5 High strength concrete Since the 1980s, many fire tests have been conducted to investigate the material properties of high strength concrete (HSC) at elevated temperatures. The following two observations were noted (Lennon et al 2007): al 2007): a) b)
the strength loss of HSC at elevated temperatures is more pronounced; and the susceptibility of HSC to explosive spalling at temperatures below 400oC.
For the first observation, BS EN 1992-1-2 divides HSC into three classes, namely: Class 1 for concrete C55/67 and C60/75, Class 2 for concrete C70/85 and C80/95, and Class 3 for concrete C90/105, and Table 9 gives the strength reduction factors for these three classes of HSC at elevated temperatures. For the second observation, HSC is susceptible to spalling since it has smaller free pore volume, so that the pores become filled with high-pressure water vapour easily in fire. Clause 4.3 of Concrete Code 2013 2013 therefore specifies that the content of silica fume if used in HSC should not exceed 6% by weight of the total cementitious content in order to prevent spalling of HSC at elevated temperature.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 27 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
Table 9 Reduction factors for HSC at elevated temperatures Strength reduction factors Concrete temperature Class 1 Class 2 Class 3 20°C 1.00 1.00 1.00 0.90 0.75 0.75 100°C
0.90 0.85
0.75 0.75
0.70 0.65
0.75 0.60 0.45 0.30 0.15
0.75 0.60 0.45 0.30 0.15
0.45 0.30 0.25 0.20 0.15
0.08 0.113 0.04 0.075 1,000°C 1,100°C 0.01 0.038 1,200°C 0.00 0.00 (Source: Source: BS EN 1992-1-2) 1992-1-2)
0.08 0.04 0.01 0.00
200°C 300°C 400°C 500°C 600°C 700°C 800°C 900°C
4.3.6 Hollow block floor floor 4.3.6.1 Hollow block floor was commonly used in the 1960s due to its good thermal and sound insulation properties. For hollow-block floors, there is no provision in Code of Practice for Fire Safety in Buildings 2011 (“FS (“FS Code”) Code”) (BD 2011a) issued by Buildings Department on this specific type of floor construction. The topping can hardly achieve a fire resistance rating (FRR) of 60 mins for the minimum thickness (100mm) for solid floor construction, and the width of the ribs can again hardly achieve a FRR of 60 mins for the minimum width (200mm) for beam in FS Code. Hence, PSE may consider a structural fire engineering study is required to justify its fire resistance rating for insulation. However, ignoring the contribution of the concrete hollow blocks in fire resistance is not a correct assumption, as one of the advantages of hollow block floors is its good thermal insulation of the blocks. 4.3.6.2 This Guideline recommends PSE to follow f ollow the provision of minimum dimension for hollow-block floors for insulation in BS 8110: Part 2, which specifies that such type of floor construction can be treated as solid construction in calculating its fire resistance by including the contribution of the cement sand or clay tiles using the effective thickness t e given by the following equation: t e h ξ t f where and
h is the overall actual thickness of slab; is the proportion of solid material per unit width of slab; tf is the thickness of non-combustible finish.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 28 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
In most cases, without carrying out detailed measurement of the hollow blocks to calculate , the topping plus the non-combustible finish may already be able to achieve the required FRR for insulation. In addition to non-combustible non-combustible plaster, tiles and floor screed may also be included as the cover to the reinforcement (Koon 2010). As such, hollow block floor can usually satisfy the prescriptive requirements of FS Code, and no structural fire engineering study is required. The example in the following paragraph will show the calculation of the effective thickness. 4.3.6.3For structural integrity, the cover to steel reinforcement in hollow block floor follows that for solid rc slabs, and hence if the cover cannot satisfy the prescriptive requirements, PSE may consider a structural fire engineering study similar to that for rc solid slab construction with the effective thickness te calculated using the above expression. 4.3.6.4 Example Consider a hollow-block floor with a topping of 75mm, and ribs of width 80mm and depth of 350mm at 500mm c/c as shown in the following figure. The thickness of the concrete blocks is 20mm. 20mm. There is also cement sand floor screed of 25mm.
Here, = the proportion of solid material per unit width of sl ab = 1
360 235 500 350
= 0.517
tf = the thickness of non-combustible finish = 25mm h = the overall actual thickness of slab = 350mm Hence, the effective thickness te h ξ t f 350 0.517 25 =277mm
For a FRR of 60 mins, the minimum thickness as specified in FS Code is 100mm, and hence the effective thickness of the hollow-block floor slab far exceeds the minimum requirements. Indeed, even ignoring ignoring the contribution of of the floor screed, the effective thickness of the hollow-block slab is 241mm, which already exceeds the specified specified 100mm. 100mm. Alternatively, ignoring the contribution of the blocks, the effective thickness of the hollow-block slab including the floor screed is 100mm, which also meets the specified 100mm.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 29 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
4.4
Fire Engineering Design of Composite Structure
4.4.1 Composite slab 4.4.1.2 Composite slabs are constructed from reinforced concrete slabs in composite action with steel decking underneath (Figure 13). The steel decking acts as support to the concrete during construction and is generally profiled to maximize structural efficiency. No fire protection is usually provided to the soffit of a composite slab. This means that the profile sheet steel decking is exposed directly to the fire. Observations during actual actual fires also showed that the steel deck had debonded from the concrete slab owing to the release of steam from the the concrete (Bailey and and Moore 2000). A simplified method method is therefore to assume that in a fire the reduced design loads are assumed to be mainly resisted by the steel reinforcement, placed in the concrete slab, with the exposed steel deck being largely sacrificial. Hence, the structural fire engineering design of a composite slab is similar to the fire engineering design of rc slab.
Figure 13 Typical composite slab with profiled steel decking (Source: Source: IStructE 2003)
4.4.1.3 Instead of calculating the fire resistance rating of a composite slab, manufacturers of the proprietary steel decking have specified the fire resistance rating and the design imposed load. Such proprietary system adopts the simplified method in its design, and is based on test results with the steel reinforcement carrying the load during the fire. PSE thus usually does does not need to carry out structural fire engineering calculation for composite slabs. However, it should be noted that the values given by manufacturers are only applicable for the standard time-temperature fire scenario. 4.4.1.4 Should a calculation be required, the usual assumption is that the strength of the slab at elevated temperatures is calculated by contributions of the concrete slab and any flexural reinforcement. Any contribution from the steel decking is ignored in the design, due to its high temperature and observed behaviour in fire where the deck deck debonds from the concrete. Annex D of of BS BS EN 1994-1-2: 1994-1-2 : Design of composite steel and concrete structures. General rules. Structural fire design (BSI 2005b) gives the following formula to calculate the effective thickness of the composite slab:
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 30 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
l1 l 2 h2 h 0.5h ( ), for 1.5 and h 1 40mm 1 2 l1 l 3 h1 h eff l l h h 1 0.75h 1 ( 1 2 ), for 2 1.5 and h 1 40mm l1 l 3 h1 where h1, h2, l1, l2 and l3 are shown in the following figure:
(Source: Source: BS EN 1994-1-2) 1994-1-2) After computation of the effective thickness of the slab, Table 10 shows the temperature distribution across the slab of an equivalent thickness of 100mm in a standard fire given in Annex D of BS of BS EN 1994-1-2. 1994-1-2. Though this table only shows the temperature distribution for a slab sl ab of an equivalent thickness of 100mm, it is a conservative assumption to be used for slab with thicker thickness. Thereafter, the computation computation steps are very similar to that for rc slabs. That is, the method method then calculates the capacity of the slab by using the reduced strength of the reinforcing bars and concrete due to elevated temperatures (Chung and Wang 2006). One assumption of the method is that the steel deck remains bonded to the concrete; but fire tests have observed that this is not a valid assumption (Lennon et al 2007). al 2007). PSE should therefore note that the UK National Annex to BS to BS EN 1994-1-2 states that Annex D of BS BS EN 1994-1-2 should not be used, and advises designers to refer to references at www.steel-ncci.co.uk (which www.steel-ncci.co.uk (which is in preparation). Table 10 Temperature Temperature distribution in a composite composite slab of 100mm thick under a standard fire
(Source: Source: BS EN 1994-1-2) 1994-1-2)
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 31 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
4.4.2 Composite beam 4.4.2.1 During a fire, the composite deck and the rc slab act as a heat sink to transfer away from the steel beam. Also, a composite beam is usually designed as simply supported, and in such case since the compression flange of the steel section is restrained, LTB of the steel section does not occur. Therefore, the design of a composite beam during fire is to ensure that the maximum applied moment in the beam does not exceed the plastic moment capacity of the composite cross-section under fire conditions. The concrete concrete in compression may be assumed to be cold and the the steel temperature is calculated. Similar to steel beams, there are three methods to calculate the moment capacity of a composite beam under fire: 1. 2. 3.
limiting temperature method; critical temperature method; and moment capacity method;
4.4.2.2 Limiting Limiting temperature method The limiting temperature method was introduced by BS 5950: Part 8 , and it requires the calculation of the load ratio rati o R given by: M R f Mc where Mf is the maximum bending moment in the beam under fire condition and Mc is the bending moment capacity of the beam at ambient temperature. BS 5950: Part 8 then gives the limiting temperature of a beam with the limiting temperature-load ratio relationships with complete shear connection and partial shear connection ( Table 11). For unprotected steel beams, the earlier paragraph has quoted the usual rule of thumb for their limiting temperature as 550oC with a load ratio R of 0.6, and for composite beams, their limiting temperature has been raised to about a bout 600oC. Table 11 Limiting temperatures temperatures of unprotected unprotected steel beams beams in bending Limiting Load ratio R temperature 0.7 0.6 0.5 0.4 0.3 0.2 0.1 (°C) 100% degree of 550 580 610 645 685 740 840 shear connection 40% degree of 575 600 635 665 700 760 860 shear connection (Source: Source: BS 5950: Part 8) 8)
4.4.2.3 Critical temperature and full moment capacity methods Both methods are specified in BS EN 1994-1-2. 1994-1-2 . The critical temperature method may be used to estimate the critical temperature of the lower flange of the composite beam under a given sagging bending moment. This method is very simple and can be used to quickly estimate whether the composite beam
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 32 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
is adequate in fire without going going through the detailed calculations. However, BS EN 1994-1-2 does not cover a composite beam with partial shear connection at ambient temperature. Hence, for such beams, PSE may still need to refer to BS to BS 5950: Part 8. 8. The full moment capacity method in BS EN 1994-1-2 is the same as in BS in BS EN 1994-1-1: Design of composite steel and concrete structures. General rules and rules for buildings (BSI 2004b) for a composite beam at ambient temperature; but the compressive capacity of the concrete slab and the tensile capacity of the steel profile should account for non-uniform temperature distributions in the concrete slab and the steel profile. As such, the the full moment capacity method involves quite tedious computation, and either the limiting temperature method or the critical temperature method is preferred.
4.5
Fire Engineering Design of Timber Structure
4.5.1 As combustible material, materi al, timber can contribute rapidly to a fire within a building. Timber may also have been treated with insecticides or anti-fungal treatments or with polishes, waxes and painted finishes, which increase the potential for fire development. During fire, timber t imber ‘browns’ at about 120-150°C, 120 -150°C, ‘blackens’ around 200-250°C and evolves combustible vapours at about 300°C (IStructE 2010). Above 400-450°C (or 300°C if a flame is present) the surface of the timber will ignite and and char at a steady rate (IStructE 2010). However, timber does have a degree of fire resistance which increases with the thickness of the component under attack. When the outer layer chars ( Figure 14), it also acts as an insulating layer to the inner uncharred core. 4.5.2 In the alteration works of historical buildings, such members may require preservation. Two approaches can be adopted: a) b)
applying fire protection material to the timber members; or using a structural fire engineering approach to substantiating the load carrying capacity of the member under fire.
Figure 14 14 Charring of timber during during fire (Source: Source: The Old House Web)
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 33 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
A simplified design of timber members is given BS EN 1995-1-2: Design of timber structures. General. Structural fire design (BSI 2004), which calculates the effective cross-section, the reduced strength and stiffness of the members. The method is based on charring depths and temperature profiles of the member during the fire. During a fire, the outer layer of the timber member chars losing all strength while retaining a role as an insulating layer which prevents excessive temperature rise in the core. The central core is slightly temperature affected with some some small loss of of strength. As such, the simplified method is to perform the calculations using normal ambient design methods for unchatted zone ( Figure 15) (Purkiss 2007).
Figure 15 Simplified method for timber members members at elevated temperatures (Source: Source: Wood Marketing Federation, Ireland)
4.5.3 The rate of charring is the key parameter in determining the effective crosssection during a fire, which in terms depends on the timber type (or the density of the timber) and the moisture content. content. For a standard fire, BS fire, BS EN 1995-1-2: Design of timber structures. General. Structural fire design (BSI 2004c) gives two values (Table 12) of charring rates, namely: β0 for single face exposure and βn for multi-face exposure. Table 12 Charring rate of different types of timber timber Charring rate (mm/min) Timber type β0 βn Softwood 0.65 0.7 Glued laminated timber ( ( 290kgm 290kgm -3) -3 0.65 0.8 Solid ( (290kgm 290kgm ) Hardwood 0.65 0.7 Solid or Glued laminated timber ( ( =290kgm-3) -3 0.50 0.55 Solid or Glued laminated timber ( ( 290kgm 290kgm ) 0.9 Plywood Panel ( ( =450kgm-3 and thickness 20mm) (Source: Source: modified from BS from BS EN 1995-1-2) 1995-1-2)
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 34 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
For single face exposure, the depth of charring d char,0 is given by dchar,0 = β0×t For multi-face exposure, the depth of charring d char,n is given by dchar,n = βn×t Once the depth of charring has been found, it is possible to calculate the total depth of section reduction d ef by the following equation: def = dchar,n + k 0×d0 where d0 is taken as 7mm; and k 0 is given by: k 0 = t/20 (for t < 20 min) k 0 = 1 (for t > 20 min) Similar formulae (e.g. by Hadvig 1981) have also been developed for calculating the depth of charring in a parametric fire. 4.5.4 Other empirical methods In addition to the method in BS EN 1995-1-2, 1995-1-2, a number of empirical approaches to the assessment of the fire performance of timber elements are available, including the method by Lie (1977). For Lie (1977), the fire resistance rating of a beam heated on three sides is given by: B t fi,d 0.1fB(4 ) D where f is a factor allowing for effective over-design, and values of the parameter f are given in the following table: Member type Column Load Factor ( ) (%) Beam L/D 10 L/D 10 1.0 1.0 1.2 75 1.1 1.1 1.3 75 50 1.3 1.3 1.5 50 (Source: Source: Purkiss 2007).
4.5.5 Example A timber softwood beam of 250×75 is simply supported over a span of 4.5 m. The bending strength of the softwood is 27.5MPa. The beam system is at 600mm c/c and carries a permanent dead load of 0.2kN/m and an imposed load of 2.5kPa. It is required to determine whether the beam can survive a standard fire with duration of 30 min. With a partial safety factor of 0.7 for imposed load, the load in the fire limit state is 0.2 + 0.7×(0.6×2.5) = 1.25 kN/m. Maximum bending moment M Ed,fi = 1.25 × 4.5 2/8 = 3.16 kNm.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 35 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
BS EN 1995-1-2 method The charring charring rate βn for softwood is 0.80 mm/min. The depth of charring d char,n = 30 × 0.80 = 24mm As t > 20 min, k 0 = 1,0, k 0d0 = 7 mm. So, def = 24 + 7 = 31mm Reduced width b = B − 2 def = 75 − 2 × 31 = 13mm Reduced depth, d = D − def = 250 − 31 = 219mm Elastic section modulus = bd 3/6 = 13 × 219 3/6 = 104 × 10 3mm3. Strength of softwood = 27.5 MPa. Therefore, MRd,fi = 27.5 × 104 × 10 −3 = 2,86 kNm. As MEd,fi = 3.16 kNm, the beam is not satisfactory. Lie (1977) method The design fire moment M Ed,fi = 3.16 kNm. Stress at fire = 3.16× 10 6/(75 × 250 2/6) = 4.04 MPa. Design strength (assuming medium-term loading) is 0.8 × 22/1.3 = 13.54 MPa. % of allowance = (4.04/13.54)×100 = 0.30 Therefore, f = 1.3. B 75 t fi,d 0.1fB(4 ) 0.1 1.3 75(4 ) = 36 min D 250 The beam is marginally satisfactory.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 36 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
5.
Assessment and Repair of Fire Damaged Structure
5.1
Purpose of Post-fire Assessment Post-fire assessment differs from structural fire engineering design. It is aimed at estimating the severity of a fire that may have damaged the structure and adequacy of the existing building in respect of structural and fire safety after a fire incident. After a fire, the first thing to do is to proceed with a preliminary assessment of the structural integrity of the building to determine whether it is possible to safely enter the building. Then, it is required to assess the extent of the damage and see whether the building can still be repaired. The following two publications are particular useful in carrying out assessment and repair of fire damaged structure: Concrete Society (2008), TR 68: Assessment, design and repair of firedamaged concrete structure (Camberley: Concrete Society); and IStructE (2000), Appraisal (2000), Appraisal of existing structures (London: IStructE, 3 rd ed.).
5.2
Procedures of Post-fire Assessment
5.2.1 The general procedures adopted ( Figure 16) in post-fire assessment of fire damaged structures are as follows (IStructE 2010): • initial site visit • preliminary inspections • desk study • detailed structural assessment • damage assessment • specification of repairs. An information paper illustrating the procedures and the salient features to assess and repair Tai Shing Street Market in Wong Tai Sin due to a fire in April 2013 is being prepared.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 37 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
− −
Initial site visit Verify if structure is safe to enter Take action to secure public safety
Preliminary inspections − Identify follow-up areas − Note area with maximum temperature Detailed evaluation − Computational modeling of fire scenario using CFD method, e.g. modeling using CFAST − Non-destructive tests − Destructive tests Structural appraisal Repairs − Identify extent of repair − Prepare details and specifications of repair Figure 16 Procedures of assessment of fire damaged structure (Source: Source: modified from Gosain and Choudhuri 2008)
5.2.2 Initial site visit and preliminary inspections 5.2.2.1 During an initial inspection, spalling, s palling, the flaking of the concrete, concrete , the formation of major cracks and the distortion of the construction are to be identified so as to assess the structural integrity. Should structural integrity be in doubt, PSE should consider fencing off and/or install sufficient temporary propping to these problematic areas. Excessive deflection, large extensive cracks in structurally sensitive areas, misalignments, and distorted members are indications that the load-carrying capacities may have been seriously impaired. The PSE should consult his respective CSE via his SSE on critical decisions. 5.2.2.2 As the concrete surfaces of the structure are always blackened and visibility in the absence of artificial lighting is poor, it is difficult to ascertain the extent of damage. During the initial inspection, an overview examining the most conspicuously damaged elements and identifying the extent of damaged elements is a practical way to give an indication of the likely scale of the damage and the areas to be under detailed investigation. investigation. It is also essential to gather all possible clues regarding the history of the fire. Witness, fire reports from Fire Services Department, talks with firemen, photographs and videos, etc are important in establishing the fire history, e.g. when the fire started, whether flashover occurred.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 38 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
5.2.2.3 An initial assessment of the gas temperature at the time of the fire is required to determine whether whether structural damage has been resulted. A quick guide guide can be referenced to the position, the condition, the melting and the charring of materials (including non-structural materials) (Figure 17(a)). Table 13(a) and Table 13(b) list the effect of elevated temperature on and the ignition temperature of common construction construction materials. For example, if a brass lock or handle is found to have melted, it is logical to deduce that the gas temperature might have been over 900 oC (a very high temperature indicating that structural damage may have occurred). Consideration of the fire characteristics may also prompt other specific issues, such as whether toxic or deleterious combustion products have been given off. For example, the burning of extensive quantities of PVC may give off enough hydrogen chloride to initiate corrosion of steel elements or reinforcement (Concrete Society 2008). Hence, it is important to visit the site as soon as the building can be entered safely and before the removal of of debris. A comprehensive comprehensive set of photos taken to record the environment and condition of the venue as earlier as possible after the fire can help to assess the fire load and maximum attainable gas temperature at the fire in the environment, which are essential to devise a cost-effective repair proposals
Figure 17(a) Condition of fresh water pipes, drain pipes and air ducts after fire at Tai Shing Street Market
5.2.2.4 Other useful visual observations obser vations include (Anderberg 2009): a)
sooty concrete surfaces which means that the temperature has been below 500°C; b) pieces of charred wood where the charred depth may give you information information about the duration of the fire (charring depth of about 20 mm is equivalent to 30 mins. standard fire). Useful information on the fire severity can also be obtained by the length of time taken to fight the fire, the length of time between the fire being noted and the arrival of the brigade, the operation of any automatic fire detection or firefighting equipment and the degree of effort required to fight the fire.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 39 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
Table 13(a) Effect of elevated temperatures on common construction materials Approximate temperature o ( C) 100 150 120 120-140 150-180 120 120-140 130-200 250 100 150 200 400-500 200-300 240 250 300-350 350-400 400 420 400 600 650 500-600 800 900 950 900-1000 950-1050 900 900-1000 1000-1100 1100-1200 1150-1250
Substance
Examples
Paint Polystyrene
Polyethylene Polymethyl methacrylate PVC
Cellulose wood Solder lead
Zinc Aluminium and alloys Glass
Silver Brass Bronze Copper Cast iron
Condition Deteriorates Destroyed
Thin-wall food containers, foam, light shades, handles, curtain hooks, radio casings Bags, films, bottles, buckets, pipes Handles, covers, skylights, glazing Cables, pipes, ducts, linings, profiles, handles, knobs, house ware, toys, bottles Wood, paper, cotton Plumber joints, plumbing, sanitary installations, toys Sanitary installations, gutters, downpipes Fixtures, casings, brackets, small mechanical parts Glazing, bottles
Jewellery, spoons, cutlery Locks, taps, door handles, clasps Windows, fittings, doorbells, ornamentation Wiring, cables, ornaments Radiators, pipes
Collapse Softens Melts and flows Shrivels Softens and melts Softens Bubbles Degrades Fumes Browns Charring Darkens Ignites Melts Melts, sharp edges rounded Drop formation Drop formations Melt Softens Melts Drop formation Softens, sharp edges rounded Flowing easily, Viscous Melts Drop formation Melts Drop formation Edges rounded Drop formation Melts Melts Drop formation
(Source: Source: IStructE 2010 and Concrete Societ y 2008)
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 40 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
Table 13(b) Ignition temperatures of common construction materials Ignition Auto-ignition Material o 1 o 2 temperature ( C) temperature ( C) Wood 280-310 525 Wool 240 Paper 230 230 Cotton fabrics 230-270 255 Polymethylacrytate (Perspex) 280-300 400-600 Rigid polyurethane foam 310 410 Polyethylene 310 415 Polystyrene 340 350 Polyester (glass-fibre filled) 350-400 480 PVC 390 455 Polyamide 420 425-450 Phenolic resins (glass-fibre filled) 520-540 570-580 Notes: 1 The temperature to which material has to be heated for sustained combustion to be initiated from a pilot source. 2 The temperature at which the heat evolved by a material decomposing under the influence of heat is sufficient to bring about combustion without application of an external source of ignition.
(Source: Source: IStructE 2010) 5.2.2.5 Concrete delamination, spalling and cracks cr acks Spalling of concrete in fire involves the breaking off of layers or pieces of concrete from the surface of the structure, as it is heated ( Figure 17(b)). The low thermal diffusivity of concrete means that the thermal gradients will induce high internal mechanical pressures in the concrete mass, which allows the development of cracks both during heating and cooling. Besides, when the temperature gradients are high, the pressures accumulated inside the pores increase due to the water vapour evolution, and so the risk of spalling of external concrete layers also increases (Alonso 2009; IStructE 2003). One major effect of spalling is that it may significantly reduce or even eliminate the layer of concrete cover on the reinforcement bars, thereby exposing the reinforcement to high temperatures, leading to a reduction of strength of the steel and hence a deterioration of the mechanical properties of the structure as a whole.
Figure 17(b) Concrete spalling during during fire at Tai Shing Street Street Market
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 41 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
There are three common types of spalling, namely: explosive spalling, aggregate spalling, and corner spalling (Concrete Society 2008)). 2008)). Explosive spalling occurs early in the fire (typically within the first 30 minutes) and proceeds with a series of disruptions, disr uptions, each locally removing r emoving layers of shallow shall ow depth. Aggregate spalling also occurring in the early stage, involves the expansion and decomposition of the aggregate at the concrete surface causing small pieces of the aggregate flying off the surface. Such type type of spalling will will only result in superficial damage. Corner spalling occurs in the later stage of the fire, and is due to tensile cracks developing at planes of weakness. However, this type of spalling occurs in the later stage, when the concrete is already significantly weakened, and will not usually affect structural performance. In some damaged areas, delamination at the interface between reinforcement bars and concrete instead of spalling may occur due to thermal expansion of reinforcement bars. A quick way to identify the extent of delamination is by using hammer tapping. PSE should, however, note that large amounts of spalling do not necessarily imply that the reinforcement, or the structure, is substantially weakened since spalling may occur due to quenching effect by the cold water from firemen’s hoses. A quick way to note that spalling occurs during the fire is that the exposed surfaces are smoke blackened (Purkiss 2007). Local buckling of steel reinforcement in a flexural member exposed by spalling usually indicates that the reinforcement has been subjected to direct fire exposure. When steel is exposed to a temperature of 600 oC, the bars lose about 50% of their strength and are then unable to resist the axial thermal restraining forces imposed. Buckling will then occur. The absence of buckled steel reinforcement in flexural members may therefore indicate that the steel is unlikely to have reached 600oC, indicating the spalling has occurred after the fire. Thus, only more severe with exposed buckled reinforcement and blackened concrete surfaces suggests more severe damage. However, for main main steel reinforcement in columns which have been tied by binders, it is very possible that the steel has reached a temperature above of 600 oC without exhibiting signs of buckling. Exposed steel reinforcement in columns after a fire requires further detailed investigation. Besides spalling, concrete concrete may also crack at high temperatures. Cracking is a useful sign in assessing the temperature of the fire, as cracks found in the compression zones of beams or slabs or in columns indicate the existence of severe problems. 5.2.3 Desk study In parallel with the initial site visit, all drawings, calculations, specifications, and information regarding the use of the building should be collected. Using the observations during the initial site visit and the desk study, the duration of the fire, the extent of damage and the desk study, a decision on whether detailed structural assessment of the structure can be made. If a fire is not long or has been suppressed by sprinkler system, the damage will usually be minor or cosmetic. However, in more severe cases, structural damage (such as concrete cracking, spalling, deformed flanges and webs of steel beams, deformations of
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 42 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
structural members) may be evident. evident. considered.
Detailed assessment should then then be be
5.2.4 Detailed assessment of temperature at fire 5.2.4.1 Fire Modelling Since the residual strength of concrete and steel reinforcement depends on the maximum temperature reached, it is important to estimate this temperature. However, if the extent of damage by the fire is large, it may be impractical and unnecessary to determine the maximum temperature of every structural member unless to those members with obvious damage and distortions present. To aid the damage appraisal and the development of a cost-effective repair schedule, a fire model using CFD method may be used to estimate the fire intensity (gas temperature) and the resultant approximate isothermal surfaces. The report from Fire Services Department is vital to input the locations of the fire, the duration and and the spread of of the fire. They can also provide guess of of temperature and whether flashover has occurred. Observation of burnt elements (Table 13(a) and Table 13(b)) and the photos taken at the initial site visit can help in estimating the intensity of fire load at the time of the fire. Part I of this set of guidelines has already described the various models to predict the temperature distribution in a fire during design, and these models can also be employed to back analyse the variation of gas temperature with time during an actual fire. Usually a zone model (e.g. by CFAST), the compartment is divided using a system of differential equations that express the conservation of mass and energy, assuming valid the ideal gas law and defining the density and the internal internal energy. Such fire modelling can help to establish the progression of the fire and provide information of the supposedly major damaged zone. An example of such modeling technique can also be found in the information paper on the assessment and repair of Tai Shing Street Market in Wong Tai Sin after fire fir e (being prepared as at 8 August 2013).
5.2.4.2 Colour of concrete at fire In Hong Kong, most of the buildings have been constructed with concrete. Concrete is made from aggregate, and its colour changes when subjected to heat. The change of colour is due to the presence of ferrous components in the cement paste, coarse and fine aggregate. At above 300 oC, a red discolouration is important as it coincides approximately with the onset of significant strength loss. However, this change of colour is most pronounced for siliceous aggregates but not so for granitic aggregates, since the red colour change is a function of the ferrous content which varies with different types of aggregates. This modification in colour is permanent: it is therefore possible, on the basis of the colour of the concrete, to make an approximate assessment of the maximum attainable temperature and temperature profile reached during the fire. Figure 18(a) shows the colours of the concrete at different heating temperatures, and Table 14 provides an overview of the colours of o f concrete at different temperature ranges.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 43 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
Figure 18(a) Colours of concrete at different heating temperatures (Source: Source: Hager 2013) Table 14 Summary of colours colours of concrete in different different temperature temperature ranges Heating Colour Description Temperature 300 to 600°C pink or red 600 to 900°C whitish gray over 900°C buff (Source: Source: International Federation for Structural Concrete 2008, Felicetti 2004)
This means that it is possible to assess maximum attainable temperature of concrete at the fire by observing the colours of the concrete. Figure 18(b) shows that the colour changes gradually from heating face to inner of the concrete. With reference to the above correlated temperature, the residual resistance of the concrete after a fire can then be be deduced. In practice, PSE can confirm that any concrete that turns pink is suspicious. A temperature of 300°C corresponds, more or less, to concrete that has lost a permanent part of its resistance (Concrete Society 2008). A grey‐white colour indicates concrete that is fragile and porous. porous. Furthermore, a permanent distortion of the construction indicates an overheating of the reinforcement. However, colour changes are most pronounced for siliceous aggregates and less so for granitic aggregates, which are predominant in Hong Kong. Also, due consideration should always given to the possibility that the pink/red colour may be a natural feature of the aggregate rather than heat-induced (Concrete Society 2008).
Figure 18(b) Change in colour of concrete heated from the left face (Source: Source: Short et al 2001) al 2001)
To minimise the colour changes due to aggregate type, PSE may develop a baseline colour chart from a set of control samples (Figure 18(c)) obtained from in-situ concretes in non-fire damaged areas in different elevated temperatures. This set of control samples should be similar to the in-situ damaged concrete in respect of mixing proportion, concrete grade, age and effects from external environment. A pair of concrete slices will be heated in Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 44 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
different elevated temperatures, e.g. 200°C, 300°C, 450°C, 600°C and 800°C. A chart (Figure 18(d)) showing colours of the concrete samples in different temperatures together with colours of the concrete sample at ambient temperature can be established as reference to determine the depth of damage of the in-situ concrete in fire.
Figure 18(c) Sliced concrete cores (Source: Source: Felicetti 2004)
Figure 18(d) Colours of sliced concrete concrete cores cores taken at Tai Shing Street Street Market at different temperatures
5.2.4.3 Petrographic examination Another method is petrographic examination, which involves the examination of concrete thin sections cut from the core to determine aggregate and paste mineralogy, and micrcostructure. As heating concrete causes a progressive progressive series of mineralogical changes, the thin sections can be investigated by petrographic examination to determine dete rmine the maximum temperature te mperature attained and deduce the depth to which the concrete has been damaged. Identification of microscopically observed features can also allow the temperature profile to be plotted through the depth depth of concrete members. Petrographic examination can examine the colour change in aggregates to estimate the maximum attained temperature, and the estimated temperature can be cross-checked with changes in the cement paste and evidence of physical distress such as cracking and microcraking (Ingham 2009). An example is quoted by Concrete Society showing the use of petrographic images (Figure 18(e)). In the example, some aggregate particles have become red indicating that the concrete has reached at least 300ºC at that point. Particles of flint have been calcined (brown mottled) and so have been heated to 250 – 450ºC. 450ºC. The cement matrix is bisected by numerous fine cracks (white) within the cement matrix (dark), some of which radiate from quartz grains
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 45 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
(white) in the fine aggregate fraction. This deep cracking and cracking associated with quartz suggest that the concrete has reached 550 – 575ºC. 575ºC. The concrete is therefore identified to have been heated to approximately 600ºC in the area represented by the sample. Further information on this topic can be obtained from TR 71: Concrete petrography - an introductory guide for the non-specialist (Concrete non-specialist (Concrete Society 2010) published by Concrete Societ y.
Notes:
C denotes calcined flint aggregate R denotes reddened flint aggregate
Figure 18(e) Example of petrographic petrographic image (Source: Source: Ingham 2008)
Petrographical examination, however, has the following limitations: a) b) c)
d)
colour changes are most pronounced for siliceous aggregates and less so for granitic aggregates which whic h are commonly found in Hong Kong; the preparation of sample slice is expensive and time-consuming; lack of experts in petrographic examination in Hong Kong, and the Public Works Central Laboratory can only provide interpretation on petrographic images related to the alkali-aggregate reaction and alkali-silica reaction; the examination is usually limited to particular micro-locations instead of overall condition of structural elements. That means that plenty of sample slices from different level from the exposed concrete surface and locations of structural elements are required to represent the heating temperature during the fire and the depth of damage after the fire. fi re.
In the UK, Applied Petrography Group of the Geological Society of London keeps a list of petrographers (available: www.appliedpetrographygroup.com www.appliedpetrographygroup.com;; accessed: 8 July 2013), and in the US the Society of Concrete Petrographers also keeps a list of petrographers (available: www.societyofconcretepetrographers.org;; accessed: 8 July 2013). www.societyofconcretepetrographers.org 5.2.4.4 Colour Colour image analysis In earlier paragraphs, it has been mentioned that using the change of colour in concrete is one of the most frequently used technique to assess the maximum attained temperature and to determine the depth of damage in the concrete from the fire incident. However, assessment on the colour by visual method may not be reliable. To minimise the subjective judgment, an objective approach is to use colour description systems such as RGB and HSI colour
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 46 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
spaces (Figure 19(a) ) to determine the colour change in the concrete for assessing the depth of the damage.
(i) RGB colour space (ii) HSI colour space (Source: Source: Blue Lobster Art and Design) (Source: Source: Black Ice Software) Figure 19(a) Colour description systems RGB colour space is a system most commonly used in most devices displaying images. Every colour can be represented by three elements in terms of amounts of Red (R), Green (G) and Blue (B). It is now also possible to convert the temperature distribution in a concrete element by using colour image analysis analysis in HSI colour colour space. The colour image analysis analysis aims at determining the temperature of concrete by the change in hue (H) ( 色相，顏 色，色彩), saturation (S) (飽和度，色度) and intensity (I) (白光光量，亮度) when concrete is heated. In order to convert the RGB RGB colour space into HSI colour space, the values of H, S and I can be calculated mathematically as follows: I
1 3
(R G B)
S 1-
min{R, G, B}
H cos 1 {
I 0.5 [(R G) (R B)] B)] (R B) 2 (R B)(G B)
}
Short et al (2001) al (2001) proposed to adopt optical microscopy combined with colour image analysis to quantify changes in colour for concrete in elevated temperatures in terms of H, S and I as per the above definition. A number of concrete samples with different mixing components were heated to equilibrium temperatures of 175, 250, 300, 350, 400, 450, 500 and 700°C and then cooled to ambient temperature. The samples were impregnated with a colourless resin, cut, ground and polished for examination in reflected light (Figure19(c)). By using a polarizing microscope together with image analysis workstation and associate software, the colour parameters including values of H, S and I on the prepared samples can be determi ned.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 47 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
(i) ambient temperature
(ii) at 350°C
(iii) heated from the left
Figure 19(c) Colour of polished sections sections of concrete concrete (Source: Source: Short et al 2001) al 2001)
In the above measurement, Short et al (2001) al (2001) observed that the major changes occurred in values of H. Figure19(d) illustrates that values in levels 0-19 of a sample heated to 350°C is more than that of the control sample and further shows that red colour (0-19 levels) started to develop significantly for samples between 250-350°C. This trend of the colour change can be used as a reference to determine the thermal history of concrete and the depth of the heat affected zone, particularly the 300°C isotherm.
Figure 19(d) Distribution Distribu tion of H at different elevated temperatures (Source: Source: Short et al 2001) al 2001)
Felicetti (2004) followed the similar method conducted by Short et al (2001) but adopted a common low-cost digital camera to take picture of the concrete samples. Felicetii (2004) found that the digital camera proved to be quite a sensitive device for the assessment of chromatic change of opaque materials with the sizable effect of illuminant, white balance and hue combination on colour measure accuracy. A single high resolution digital image allows to separately analyse the cement mortar and the aggregate and to outline statistical trend of its colour parameters. Lin et al (2004) further carried out colour image analysis on a number of mortar specimens by using an ordinary digital camera and his own developed image colour intensity analyser, and obtained the variation of H, S and I of three primary colours R, G and B ( Figure 19(e)) at different elevated temperatures. They observed that the numerical values of H decrease as temperature increases, but the variation is not significant. Unlike the results of Short et al (2001), they observed that S shows a marked increase with increasingly temperature. I shows little changes in the range range 0 – 200°C, 200°C, decreases with increasing temperatures in the range 200 – 800°C, 800°C, and increases with temperatures in the range 800 – 1000°C. 1000°C. The variation of these three
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 48 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
properties with temperature therefore serves as a useful way to deduce the temperature gradient across concrete cross section.
Figure 19(e) Variation of H, S and I with temperature (Source: Source: Lin et al 2004) al 2004)
The above summarises the colour image analysis carried out by Short et al (2001), Felicetti (2004) and Lin et al (2004). Advantages of colour image analysis include (Felicetti 2004): a) b) c) d)
no special preparation of concrete sample is needed and a faster in-situ analysis is feasible; the whole colour profile can be reconstructed from one side picture of a concrete core; the cement paste and the aggregate can be recognised on the picture and analysed separately; the method can now be carried out by low-cost camera.
However, PSE should note that photos taken for both the control samples and damaged concretes should be under the same external environments and equipments in order to minimise the deviations from external lighting, quality of digital camera and flash intensity. In addition, resolution of photos should be set as high as possible for showing the colour properties of a particular layer or element in the concrete sample. Moreover, as mentioned mentioned in earlier paragraphs, the change of colour of granitic concrete may not be large, and thus the method should be further development for concrete in Hong Kong. ImageJ, a free Java-based image processing program (available: rsbweb.nih.gov/ij)) developed at the National Institutes of Health in the US, rsbweb.nih.gov/ij can be used to carry out the colour image analysis. This program is capable to analyse 3D live-cell imaging and radiological image by user-written plugins originally in medical and health care industry. Since the user-written plugins allow adding special features in this Java-based program, the program has then been widely applied in other industries to analyse anal yse images. By using ImageJ, a particular location, layer or element of concrete samples can be selected to analyse the colour properties in respect of R, G, B, H, S and I ( Figure 19(f) ).
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 49 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
Figure 19(f) Colour image analysis using ImageJ ImageJ
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 50 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
5.3
Assessment of Residual Strength
5.3.1 The maximum temperature and temperature are of use to determine the residual strength of the structural materials after the fire. Earlier paragraphs describe the effect on elevated temperatures on the strength of concrete, steel reinforcement, prestressing wires, and structural steel during a fire. However, the properties after heating (i.e. upon cooling) are of interest from the point of view of reinstatement. The following paragraphs will describe the effect of elevated temperature on the residual strength of concrete, steel reinforcement, prestressing wires, and structural steel. 5.3.2 Concrete Figure 20(a) shows the residual strength of Grade 20 and Grade 30 unstressed concrete upon cooling with the corresponding changes of its colour. Concrete temperatures up to 95°C have little effect on the strength and other properties of concrete. Above this threshold cement paste shrinks due to dehydration and aggregates expand due to temperature rise. For normal weight concrete, the the aggregate expansion exceeds the paste shrinkage resulting in an overall expansion of the concrete. In addition to the expansion, reductions in strength, modulus of elasticity, and thermal conductivity occur, as well as increased rate of creep as temperature rises. For temperatures up to 300°C, the residual compressive strength of structural quality concrete is not significantly reduced. Strength loss at high temperatures is due to the dehydration of the cement paste in the concrete matrix. As temperature rises further, the dehydration of the paste will lead to a loss of essentially all of the concrete strength.
Upon cooling to ambient temperatures the strength of concrete may be further reduced from its strength at high temperature because of continuing disintegration of the microstructure (Hertz 2005). As such, for the sake of assessment of fire damaged concrete structure Concrete Society (2008) recommends a more conservative approach by discounting the residual strength for concrete exposed to temperatures above 300°C ( Figure 20(b)). .
Figure 20(a) Residual strength of concrete (Source: Source: Concrete Society 1978)
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Figure 20(b) Recommended Recommend ed residual strength of concrete after a fire (Source: Source: Concrete Society 2008)
Page 51 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
5.3.3 Steel reinforcement Figure 21(a) shows the residual strength of steel reinforcement. The original yield stress of hot rolled steel bars is almost completely recovered on cooling from temperatures of 500°C to 600°C, and on cooling from 800°C it is only reduced by 5%. Concrete Society (2008) therefore therefore recommends to assume conservatively that there is no loss in residual strength for hot-rolled steel reinforcement for a temperature up to 600 oC (Figure 21(b)).
Figure 21(a) Residual strength of steel reinforcement reinforcement and prestressing prestressing wires w ires (Source: IStructE 2010)
Figure 21(b) Recommended residual strength of hot-rolled steel reinforcement after a fire (Source: Source: Concrete Society 2008)
5.3.4 Prestressing wires Prestressed concrete slabs are now increasingly common in ArchSD projects, especially in D&B contracts, with its advantage advantage of larger span. However, the behaviour of prestressing steel after a fire is much more critical than that of reinforcing steel. First, the tensile properties of prestressing steel wires lose more markedly at elevated temperatures (Figure 10(a)(ii)). Secondly, unlike hot rolled steel reinforcement, the strength of prestressing steel wires will only partially be recovered upon cooling. Figure 21(a) shows the residual strength of prestressing wires. During the fire, the reduction in the tensile strength strength of cold-worked prestressing steel starts at temperatures between 100°C and 200°C. Its strength reduces to about 50% of room temperature strength at 400°C and to less than 10% at around 700°C. After the fire, the residual strength of prestressing steel cannot be completely recovered if its temperature during fire fi re is over 200°C. The residual strength of the prestressing steel with its temperature of 450°C during fire is approximately 75% of the original tensile strength. Moreover, the reduction reduction of elastic modulus in the concrete, increased relaxation due to creep and non-recoverable extension of tensioned steel occur as a result of increased temperatures. All effects contribute to losses in tension. Maximum temperatures reached in the tendons and their durations, together with the temperature distribution, are therefore more important in the assessment of fire-damaged prestressed concrete members than in the case of reinforced concrete. Concrete Society (2008) recommends to the residual strength for prestressed wires as shown in Figure 21(c).
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 52 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
Figure 21(c) Recommended residual strength of prestressing wires after a fire (Source: Source: modified from Concrete Society 2008)
5.3.5 Structural steelwork Figure 22(a) shows the residual strength upon cooling after fire exposure of structural steel. Grade S275 hot rolled structural steel section, subjected to a temperature above 600ºC, may suffer losses in residual properties on cooling, and at 1000°C it is 10% or less. Young’s modulus decreases with temperature rise at a slightly higher rate than does does yield strength. Grade S355 hot hot rolled structural steel also suffers losses in residual yield and tensile strength when subjected to temperature over 600ºC. Grade S355 S355 usually vanadium and niobium, and at high temperatures these elements tend to precipitate out of the matrix creating a coarse distribution. Yet, when cooled back to the room temperature its yield stress or the tensile strength will only fall not greater than 10% below their original values. Thus, it may be concluded that steel members which is only utilised to less than 90% of their maximum load bearing capacity will have sufficient strength when cooled back to room temperature. Figure 22(b) shows the residual strength properties upon cooling of bolts. The residual strength properties upon cooling show a substantial recovery for grade 4.6 if heated to only only 600°C and for grade 8.8 if heated to only 400°C. However, for high-strength friction grip bolts, as their strength behaviour is highly dependent on the contact surface, their residual performance has to be assessed by their distortion.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 53 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
(a) Structural steel (b) Bolts Figure 22 Residual strength of steel (Source: Source: IStructE 2010) 5.3.6 Timber In the earlier paragraphs, it was noted that any charred part of a timber section must be assumed to have lost all strength; but any timber beneath the charred layer may be assumed to have no significant loss of strength because the thermal conductivity of charred timber is low. 5.3.7 Structural masonry or brickwork The physical properties and mechanisms of failure of brick walls exposed to fire are not known in detail (IStructE 2010). As with concrete there may be a loss of compressive strength. Brick give much better performance if plaster is present because of its improved insulation and reduction of thermal shock. IStructE (2010) advises that if brickwork does not show visible damage (e.g. appreciable deformation, cracking or spalling), the strength of the bricks may be taken to be similar to the original value, as bricks have been manufactured at high temperatures. Mortar containing sand with a given mineral composition behaves similarly to concrete with aggregate of comparable mineral composition, and thus inspection of its colour can indicate the maximum t emperature at the fire. 5.3.8 Testing Table 15 summarises the types of tests that may be carried out in fire-damaged rc structure and the information that may be obtained (Concrete Society 2008). The primary test to determine residual strength of concrete is to carry out compressive tests on concrete cores from the fire-damaged zone. Strength tests on cores suffer a major limitation that they average the strength of concrete throughout the core, which may contain both damaged and undamaged concrete.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 54 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
Table 15 Tests for fire damaged rc structures Information gained
Test location
u t i s n I
Test type e v i t n c o u r N t s e d
Test method
Colour changes
Extent of damage
Depth of damage
Compressive strength of concrete
Tensile strength of steel reinforcement
Visual inspection Hammer tapping Rebound hammer Ultrasonic pulse velocity Open-up
y r o t a r o b a L
e v i t c u r t s e D
Petrographic examination Thermoluminescence Colour image analysis Core test Tensile test
(Source: Source: modified from Concrete Society 2008) For structural steel and steel reinforcement, the primary one is carry out tensile tests to determine its yield and ultimate tensile strength. Microscopic examination of fracture surface is not commonly carried out. 5.4
Structural Appraisal With the establishment of the temperature profile and distribution, and the strength of the concrete, steel reinforcement and structural steel, calculation can be carried out to assess structural capacity and the need for repairs. Usually, member design (unless the stability of the structure is in doubt) is adequate.
5.5
Repair Options
5.5.1 After a fire, it is sometimes necessary to carry out major repair works. It is practically impossible to provide standard solutions. Each situation must be examined individually and the best solution chosen for each case. Two options, namely, repair or demolition, are commonly adopted in post-fire post-fire repair. In this respect, the following factors must nevertheless be taken into account: − the strength of the structure after the fire; − the permanent distortions; − the durability after the fire and repairs; − the aesthetic aspect. 5.5.2 Repair of timber structure As those uncharred timber does not lose strength, the repair of timber structures after fire depends on the degree of charring. If the depth of char is insignificant the remaining section may still be able to resist the design loads, no special structural repair is required, though the charred layer is to be removed by the use of a scraping plane. plane. For connections, repairs can be carried out by nailing, bolting, screwing, steel plating or gluing. If the depth of char affects the structural integrity of the building, such timber members will generally need total replacement. Structural strengthening may also be used after the removal
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 55 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
of the charred layer, and in some cases, demolition or replacement of the damaged parts may be a more economical solution. 5.5.3 Repair of rc structure After a fire, steel reinforcement will regain most of of the original strength. Thus, unlike timber or structural steel, damage to rc structures during fire will not generally affect its structural integrity, and thus most rc structures can be repaired and returned to service, though some structural members may require demolition and and replacement (Concrete Society 2008). Indeed, the Concrete Society (2008) investigated a number of different concrete structures (including residential houses, offices, warehouses, factories and car parks) damaged by fire in the UK, and found that the majority of structures were repaired and that among those that were not, very many could have been but they were demolished for reasons other than the damage they suffered. Sooty concrete surfaces need only cleaning because the maximum temperature has been below about 500°C. Demolition of part of the structure with extensive damaged concrete is most economical repair option, especially for those structural elements having been distorted in some way following a fire. For repair, the following information is required:
• the extent of breaking out of fire damaged concrete and removal of fire • • • •
damaged steel reinforcement; requirements for preparation of concrete surfaces that are to receive repair concrete, including special requirements to prevent feathered edges; details of new steel reinforcement including lap length and splicing with original bars, mechanical anchorage, cover etc; any fabric reinforcement or wire mesh that may be required to hold the repair concrete in place in the temporary condition, including means of supporting the fabric/wire mesh and the required concrete cover; and the thickness and the properties of the repair materials.
Comprehensive repair methods and procedures of rc structures can be referenced to TR 68: Assessment, Design and repair of fire-damaged concrete structure published by Concrete Society. Recasting, spraying concrete and/or patch repair with polymer modified cementitious mortar are commonly used materials to repair the damaged concrete. SEBGL-MT2 Causes of Concrete Deterioration, Investigation and Repair Methods (available: http://asdiis/sebiis/2k/resource_centre/)) details these repair options. http://asdiis/sebiis/2k/resource_centre/ For prestressed concrete members, as described in earlier paragraphs there is a significant loss of strength and modulus of elasticity of the prestressing wires during fire and on cooling, and hence it is difficult to repair prestressed concrete structure without replacing the tendons should the duration of the fire be long enough to raise the temperature of the wires to over 200 oC. As such, a practical method is by providing cover generally provided to the tendons in prestressed concrete (Concrete Society 2008), and General Notes to Drawings are now being amended to specify such requirement.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 56 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
5.5.4 Repair of steel structures Steel remains straight after a fire and which had been carrying an appreciable load was probably not heated beyond 600ºC, and would not have undergone any metallurgical changes. The general rule rule is therefore that if the steel is straight and there are no obvious distortions then the steel is steel fit for use (Corus 2006). Moreover, if the distortion of steel section is within acceptable limit after being heated, it is not necessary to replace the existing steel structure. The recommended deflection limit in the report is 1mm/m long steel section and the maximum deflection limit of the steel section is 5mm (Kirby et al 1993). al 1993). Where deflections are visible, PSE should calculate the load carrying capacity and then devise appropriate appropriate repair option. option. If steel exposed to a fire shows roughened appearance due to excessive scaling and grain coarsening, then it may suggest that the steel has been exposed to a temperature of around 900°C. Steel so modified is commonly called “burnt” steel (Hammond and DeCicco 2003), and their suitability for further use should be carefull y studied. Examples of reinstatement works of fire-damaged steel structures were reported in The Reinstatement of Fire Damaged Steel and Iron Framed Structures (Kirby et al 1993) al 1993) published by BS Swinden Laboratories.
6.
List of References
Codes and Standards Buildings Department (2011), Code of Practice for the Structural Use of Steel (Hong Kong: Buildings Department). Buildings Department (2011a), Code of Practice for Fire Safety in Building (Hong Kong: Buildings Department). Buildings Department (2013), Code of Practice for the Structural Use of Concrete (Hong Kong: Buildings Department). BSI (1985), BS 8110: Part 2: Structural use of concrete. Code of practice for special circumstances (London: BSI). BSI (2002), BS (2002), BS EN 1990: Basis of structural design (London: BSI). BSI (2002a), BS EN 1991-1-2: Actions on structures. General actions. Actions on structures exposed to fire (London: BSI). BSI (2003), BS (2003), BS 5950: Part 8: Structural use of steelwork s teelwork in buildings. Code of practice in fire resistant design (London: BSI). BSI (2004), BS (2004), BS EN 1992-1-1: Design of concrete structures. General rules and rules for buildings (London: BSI).
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 57 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
BSI (2004a), BS EN 1992-1-2: Design of concrete structures. General rules. Structural fire design (London: BSI). BSI (2004b), BS EN 1994-1-1: Design of composite steel and concrete structures. General rules and rules for buildings (London: BSI). BSI (2004c), BS EN 1995-1-2: Design of timber structures. General rules. Structural fire design (London: BSI). BSI (2005), BS EN 1993-1-1: Design of steel structures. General rules and rules for buildings (London: BSI). BSI (2005a), BS EN 1993-1-2: Design of steel structures. General rules. Structural fire design (London: BSI). BSI (2005b), BS EN 1994-1-2: Design of composite steel and concrete structures. General rules. Structural fire design (London: BSI). Structural Design Bailey, C G and Moore, D B (2000), “The structural behaviour of steel frames with composite floor slabs subject to fire: Part 1: theory”, t heory”, The Structural Engineer , 78(11), pp. 19-27. Buchanan, A H (2001), Structural design for fire safety (New York: Wiley). Chung, K F and Wang, A J (2006), “Fire resistance design of composite slabs in building structures: from fro m research to practice”, The Structural Engineer , 84(20), pp. 30-6. Corus (2006), Fire resistance of steel-framed buildings (North Lincolnshire: Corus). Franssen, J M, Kodur, V and Zaharia, R (2009), Designing steel structures for fire safety (Boca Raton: CRC Press/Balkema). Hadvig, S (1981), Charring of wood in building fires – practice, theory, instrumentation, measurement (Lyngby: measurement (Lyngby: Technical University of Denmark). IStructE (2003), Introduction to the fire safety engineering of structures (London: IStructE). IStructE (2007), Guide to the advanced fire safety engineering of structures (London: IStructE). Law, M (1997), “Review of formula for T-equivalent”, T-equivalent”, Fire Safety Science Proceedings of the Fifth International Symposium, 3-7 March, Melbourne, Australia, p. 985-96.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 58 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
Lennon, T, Moore, D B, Wang, Y C and Bailey G G (2007), Designers’ guide to EN 1991-1-2, 1992-1-2, 1993-1-2 and EN 1994-1-2: handbook for the fire design of steel, composite and concrete structures to the Eurocodes (London: Thomas Telford). Lennon, T (2011), Structural fire engineering (London: engineering (London: Thomas Telford). Lie, T T (1977), (1977) , “A “ A method of assessing the fire resistance of laminated timber beams and columns”, columns”, Canadian Journal of Civil Engineering , 4, pp. 161 – 161 – 9. 9. Purkiss, J A (2007), Fire safety engineering: design of structures (Oxford: Butterworth-Heinemann, 2nd ed). Wang, Y C (2002), Steel and composite structures: behaviour and design for fire safety (London; New York: Spon Press). Assessment and Repair of Fire-Damaged Structures Alonso, C (2009), “Assessment of post-fire post -fire reinforced concrete structures: Determination of depth of temperature penetration and associated damage”, Alexander, MG et al (eds) (2009), Concrete Repair, Rehabilitation and Retrofitting II (London: II (London: Taylor & Francis Group), pp. 471-4. Anderberg, Y (2009), “Assessment of fire-damaged fire -damaged concrete structures and the corresponding repair measures”, Alexander, MG M G et al (eds) (2009), Concrete Repair, Rehabilitation and Retrofitting II (London: II (London: Taylor & Francis Group), pp. 631-6. Concrete Society (1978), TR 15: Assessment of fire-damaged concrete structures and repair by gunite (Camberley: Concrete Society). Concrete Society (2008), TR 68: Assessment, design and repair of fire-damaged concrete structure (Camberley: Concrete Society). Concrete Society (2010), TR 71: Concrete petrography - an introductory guide for the non-specialist (Camberley: non-specialist (Camberley: Concrete Society). Felicetti, R (2004), “Digital-camera colorimetry for the assessment of firedamaged concrete”, concrete”, Proceedings of the Workshop: Fire Design of Concrete Structures, Milan, 2-3 December 2004, 2004 , pp. 211 – 20. 20. Hager, I (2013), “Colour “ Colour Change in Heated Concrete”, Concrete ”, Fire Fire Technology, Technology, 49, pp. 1-14. Hammond, D J and DeCicco, P (2003), “Evaluating structural damage”, damage”, Cote, AE et al (eds), al (eds), Fire Fire protection handbook (Quincy, handbook (Quincy, Massachusetts: NFFPA, 19 th ed.), 12p.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 59 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
Ingham, J P (2008), “Application of petrographic examination techniques to the assessment of fire-damaged fire-damaged concrete and masonry structures”, Materials Characterization, Characterization, 60, pp. 700-9. International Federation for Structural Concrete (2008), Fire design of concrete structures – structural behaviour and assessment (Lausanne, Switzerland: International Federation for Structural Concrete). IStructE (2000), Appraisal (2000), Appraisal of existing structures (London: IStructE, 3 rd ed.). Kirby, B R, Lapwood, D G and Thomson, G (1993), The reinstatement of fire damaged steel and iron framed structures (London: BS Swinden Laboratories). Gosain, N K, Drexler, R E and Choudhuri, D (20 08), “Evaluation and repair of fire-damaged fire-damaged buildings”, Structure Magazine, Magazine, September, pp. 18-22. Lin, D F, Wang, H Y and Luo, H L (2004), “Assessment of fire-damaged mortar using digital image process”, process”, Journal Journal of Materials in Civil Engineering , 16(4), pp. 383-6. Short, N R, Purkiss, J A and Guise, S E (2001), “Assessment of fire damaged concrete using colour image analysis” analysis”, Construction and Building Materials, Materials , 15(1), p p. 9-15. Other References Bond, A J, Harrison, T, Brooker, O, Moss, R, Narayanan, R, Webster, R and Harris, A J (2011), How (2011), How to design concrete structures using Eurocode 2 (Surrey: The Concrete Centre). IStructE (1978), Design and detailing of concrete structures for fire resistance (London: IStructE). Hertz, K D (2005), “Concrete “C oncrete strength for fire safety design”, design”, Magazine of Concrete Research, Research, 57(8), pp. 445-53. K oon, oon, C M (2010), “Structural appraisal of reinforced concrete buildings with historic values,” values,” Presented at Seminar on Concrete Damage Assessment, Concrete Repair and Concrete Mix Technology, Hong Kong, China, 2 February 2010. 2010 . Lennon, T (2003), “Whole building behaviour: results from a series of large scale tests”, Presented at the CIB-CTBUH International Conference on Tall Buildings, 8-10 May 2003, 2003, Malaysia. Malaysia . Newman, G M, Robinson, J T and Bailey, C G (2000), Fire safe design: a new approach to multi-storey steel-framed buildings (Ascot, Berkshire: SCI). Pang, P T C (2006), “Fire engineering design and post fire assessment”, a ssessment”, The Structural Engineer , 84(16), pp. 23-9.
Structural Engineering Branch, ArchSD Guidelines on Structural Fire Engineering Part II Edition No./Revision No. : 1/-
Page 60 of 60
File code : SEBGL-OTH7 SEBGL-OTH7 CTW/MKL/MFY/SCF First Edition: August 2013
