Design of Multi-Storey Steel Structures for Dependable Performance in Fully Developed Fires:The Slab Panel Method

Presented on behalf of Associate Professor G Charles Clifton, The University of Auckland

20th April 2010

Unprotected Steel Behaviour in Severe Fires

• Exposed steel elements may reach close to the temperature of the surrounding gases

• Beams will sag downwards at an initially rapid rate, then a decreased and constant rate with time

• Columns will undergo increased compression loading due to restrained thermal expansion

• The extent of deformation depends on:– Is member shielded by an effective radiation barrier ?– If directly exposed, what is expected maximum fire temperature and duration ?

• How can this be used in design? – Details of New Zealand methods now given

Exposed steel elements are likely to reach close to the temperature of the surrounding gases. Typically where the principal heat transfer into the steel is by radiation [8], the elements of exposed beams and columns will reach: for I section flanges in contact with a concrete slab, �g – 150ºC for I section webs and flanges not in contact with a concrete slab, 0.95�g Steel beams that are exposed to fully developed fire will undergo initial rapid vertical downwards deflection driven by thermal gradients and then, once the temperature rises above the NZS 3404 limiting temperature, by loss of strength. For beams that stay below the limiting temperature, most if not all of this deflection is recovered on cooling down. For unprotected beams, the initial rate of deflection is high, as described in Appendix A section CA4.2.3 of HERA Report R4-131 [9] and for other than Fire Hazard Category 1 fire loads, permanent deflection will occur. Columns will undergo increased compression loading due to restrained thermal expansion. If the temperatures exceed the limiting temperature (see sessions 1 or 3 for how this is calculated) then local or member buckling is likely. Below this temperature any change in length is recovered on cooling. The extent of shielding and the structural fire severity are therefore critical parameters in design of unprotected steelwork

FED of Office and Retail Buildings:Slab Panel Fire Design MethodCovers design and detailing of steel framed

buildings with unprotected secondary beams

or joists for dependable inelastic response in

severe fires

Applicable to:–Fire Hazard Category 1 to 4

–te from 30 to 240 minutes

–Most forms of floor system made integral with the supporting steel beams

High temperature design procedure, based on the inelastic reserve of strength available from the floor system

Fire Hazard Category is the system used in the Compliance Document for Fire Safety C/AS1 to denote the different structural fire severities possible from the different Purpose Groups. It is based principally on the Fire Load Energy Density (FLED) expressed as MJ/m^2 floor area. The 4 categories are as follows: FHC 1 FLED from 0 to 500 MJ/m^2 floor area (or higher if the combustible material is very slow burning (threshold rate of burning is stated in C/AS1)). Design fire load for FHC1 in calcs is 400 MJ/m^2 floor area FHC 2 FLED from 500 to 1000 MJ/m^2 floor area. Design fire load for FHC2 in calcs is 800 MJ/m^2 floor area FHC 3 FLED from 1000 to 1500 MJ/m^2 floor area. Design fire load for FHC3 in calcs is 1200 MJ/m^2 floor area FHC 4 FLED is greater than MJ/m^2 floor area (or lower if the combustible material is very fast burning (threshold rate of burning is stated). Design fire load for FHC4 is required to be determined for the specific application. Apartments, hotel rooms hospitals are FHC1, most offices are FHC2, areas with combustible storage between 2 and 3 m high are FHC3 and FhC4 is typically high fire load single storey buildings. The FHC is used directly in the te calculations and the setting of the FRRs required from C/AS1.

Basis of

Design Procedure

Under ambient temperature conditions:

• The beams support the floor slab

• One way action prevails

• Load path:

slab → 20 beams →

10 beams → columns

Under severe fire conditions:

• Unprotected secondary beams lose strength

• Two way action prevails (slab panel)

• Slab panel supports the beams

• Load path : slab panel → supporting beams → columns

• Slab panel axial forces are in in-plane equilibrium

The basis of the design method is presented in detail in the report[1] by Bailey and in brief summary form in section 2.3 of DCB Issue No. 59, which is included in this session’s notes. The design method is applied to a slab panel, which is defined in section A2.1 of Appendix A. The slab panel resists applied load by two-way action back to the supports, through; yieldline moment action, plus tensile membrane action, plus enhanced yieldline moment capacity from in-plane axial forces An explanation of this mechanism will be given in the seminars and in DCB Issue No. 60. Readers who want a detailed explanation should study [1].

New Zealand Application of BRE Slab PanelDesign Methodology

Concept developed by Dr Colin Bailey of UK BRE; now Uni of Manchester

published in 2000

New Zealand Application covers:

• No explicit limitation on slab panel size

• Design for elevated temperature moment and shear

• Development of slab negative moment resistance where

practicable

• Slab reinforcement and detailing requirements

In the report on the ambient temperature test, Bailey notes some factors required to be considered for general application. These are [9]: 1. Elevated temperatures of components near the fire-exposed face need determination, for general application, to account for expected strength loss of materials which are at high temperatures. 2. High temperature shear capacity at the slab panel supports needs to be determined 3. Detailing for effective force transfer and integrity at supports needs to be considered, especially under conditions of high structural fire severity. The initial application of the UK procedure, as formulated in SCI Publication P288 [2], puts conservative restrictons on its use to avoid these factors exerting significant influence. The New Zealand application of the UK procedure, as developed herein, addresses these factors directly and thus has a wider range of application.

Structural Performance to be Delivered by the

Procedure - 1 of 2

Under severe fire conditions:

• Slab and secondary beams may

undergo appreciable deformation

• Support beams and columns undergo

minimal deformation

• Tensile membrane response may be activated

• Load-carrying capacity and integrity are preserved

for full burnout

• Insulation is met for required period

The maximum extent of inelastic deflection of the floor system that would be expected is described in section 2.2 of DCB Issue No. 59 - see details in this session’s notes. Inelastic response of floor systems in practice would be expected to be less in the event of fully developed fire, for the following reasons: lower fire load presence of shielding linings non-fire rated enclosures reducing fully developed fire size fire service intervention

Structural Performance to be Delivered by the Procedure - 2 of 2

Suppression of structural damage controlled by:

• Shielding linings (limited effectiveness)

• Sprinkler protection (extremely effective)

Effective compartmentation is maintained:

• Between floors

• Between firecells, same floor

More details on control of structural damage is covered later in this session. Effective compartmentation will be maintained: between floors, by the requirements of this procedure by fire seperating walls on the same floor (special details may be required for these walls depending on their location relative to the slab panel forming the ceiling ; see later in this session) Shielding linings in the context of this slide are linings which shield the steel beams from "seeing" the fire at the moment it first reaches full development. They comprise e.g. suspended ceilings or wall linings. The intention is that as long as these remain in place they keep the temperatures in unprotected steel beams behind the linings sufficiently low to minimise any fire related beam movement or permanent deflection. This concept has been experimentally tested and forms the basis of the Radiation Barrier Method which is used here.

Building Structure Characteristics Required

for Implementation of Slab Panel Design Procedure

(1) Floor slabs– concrete: structural grade, NWC or LWC– mesh/reinforcement: within slab panel, any grade

over supports ≥ 15% uniform elongation

(2) Steel beams– UB, WB, light steel joists, cellular beams

(3) Columns– UC, WC require passive protection

(4) Connections– must maintain integrity during heating and cooling down

– connector failure (bolts or welds) to be suppressed

– same detailing as required for earthquake; NZ standard practice

(4) Overall building stability– no limitations on lateral load resisting systems

– building stability not endangered by use of SPM

Building characteristics required for implementation of slab panel design procedure. Details of this are given in DCB Issue No. 54 and in HERA Report R4-90-DD-Rev 2 [12]. In summary: (1) Floor slabs (1.1) Concrete is normal weight, � 20 MPa (1.2) Mesh reinforcement hard-drawn wire mesh to NZS 3421 [24] can only be used if the pitch of the mesh bars is 300mm; mesh with lesser pitches do not have the ductility required mesh formed from welded grade 430 bars to NZS 3402 [26] must be used where the area of mesh required is such that the required pitch < 300 mm. Plain or deformed bars may be used for this mesh, plain bars are easier to weld into mesh. position, covers are as specified in Appendix A (1.3) Bar reinforcement DH12 grade 430 reinforcement position, covers are as specified in Appendix A (2) Steel beams / joists typically will be composite with the floor slab if these beams are not composite, then shear studs to NZS 3404 Clause 13.3.2.3 (h) are required; ie maximum stud spacing at 4xslab thickness hot-rolled beams, welded beams, Speedfloor Joists, beams with web openings are all suitable see Figs 60.1 and 60.2 and Appendix A for extent of protection, unprotection No limitations on lateral load resisting systems means that the stability of the slab panel is not dependent on external lateral support to the structural system supporting the slab panel. This point is made because some fire design systems being developed, eg that by David Proe in Australia, are based on a minimum lateral support being provided to the perimeter of the fire floor by the surrounding structural system

Floor Systems Applicable

• All composite

floors on steel

beams

• Some precast concrete

floors with

suitable topping

Reinforcing mesh

65mm minimum

Fire emergency reinforcement

210mm Metal Deck

Reinforcing mesh

Clipped Pan Profile

Fire emergency reinforcement

Reinforcing mesh

Slab thickness

Trapezoidal (W) Profile

Reinforcing mesh

Joist

Light Steel Joist75m m or 90m m

Negative reinforcement when required

(Comflor rib bars)

See section 2.1(6) from HERA Report R4-131 for the restrictions on application to precast floor systems and the reasons for these.

Detailing Requirements

(1) Floor slab

– Decking fastened to beams; typically composite

– Slab tied to edge beams

– Detailing requirements given in procedure

(2) Protection to slab panel edge support beams

– When specified, apply over full length

– Details given for application around connections to secondary beams

(3) Protection to columns

– Apply over full length

As with any system designed to deliver a dependable level of inelastic response, the detailing is as important as the design. (1) This especially relates to the floor slab, where: decking must be fastened to beams to NZS 3404 Clause 13.3.2.4 mesh must be lapped to NZS 3101 Clause 7.3.21 bars must be lapped to NZS 3101 Clause 7.3.17 see figs 60.4 to 60.8 for detailing of reinforcement covers are important (2) When passive fire protection is specified, it must be placed as specified, especially: full length of beams full height of columns (3) When unprotected concrete-filled structural hollow section columns are used FRR is provided to DCB Issue No. 58, pp25-30, and Canadian method for protected beams to these columns, treat the connection region as shown in DCB Issue No. 42, Fig. 42.2, by running the passive protection over the column within the depth of the connection region

Detailing Requirements

Primary Support

Passive

Protected Primary Support

Beam

Development Work Undertaken

• 13 stage experimental and analytical

development programme undertaken

• Key stages presented in following slides (not

all steps covered)

• overview given in section 8.2 of HERA Report

R4-131

• Further development work planned as noted later

Step 1: Cardington Fire Tests

• Demonstrated

performance of large scale composite floor systems

• Showed systems with unprotected beams and protected columns have high fire resistance

Step 2: BRE Design Model and Test

• Colin Bailey Tensile Membrane Model, UK BRE

• Large scale ambient

temperature tests on

lightly reinforced

slabs to validate

behaviour

Step 3: First Edition of SPM

• Generalised application of Bailey model

• HERA DCB No 60, February 2001

• Incorporating moment capacity of secondary beams

• General formula for yieldline determination

includes support moment contribution

• Limits on application set by Bailey for:– integrity

– maximum deflection

enhancement factor, e

Step 5: Furnace Testing of Six Slab Panels

• part of PhD research

project

• details as shown opposite

• all slabs withstood 180 minutes ISO fire without failure

Results of tests

Load ratio ≤ 1.0 � no tensile membrane enhancement required

Load ratio > 1.0 � tensile membrane enhancement is required

D147 top surface crack

pattern

Step 6: Second Edition of SPM

• Incorporating results of

furnace tests

• HERA DCB No 71, February 2003

• Improved determination of slab and reinforcement

temperatures

• Revised reinforcement limits for integrity

• Relaxation of maximum deflection and limits on e

Step 7: Development and Validation of FE Model

• 6 slab panels modelled

• Best fit to mid-span

deflection made for each case

• Accuracy of models also compared with:

– reinforcement strains

– edge deflections and rotations

Example shown for

Speedfloor slab

The slab panel method must work with deforming edge beams. Our FEM work, which is reported in HERA Report R4-118.1:2004, showed that the performance of the slab panel is not diminished by deflection of the supporting edge beams, at least up to edge beam vertical deflection of up to span/75, and that some 65% of this deflection is added to the slab panel central deflection. These findings are incorporated directly into the procedure. They are based on Finite Element Modelling which has been undertaken by New Zealand’s most highly qualified and one of the most experienced FEM engineers. The fact that only 65% of the support beam vertical deformation is seen in the slab panel central deformation shows that some load is carried by more direct means into the corner supporting columns of a slab panel. Note that the tensile membrane enhancement is dependent on the vertical height difference between an element of yieldline and the corresponding position along the support beam, so it is not changed by allowing for the supporting beam deformations

Steps 8 and 9: Extending Validation Using FEM

FEM used to extend experimental testing to determine the influence of:

• contribution of the unprotected secondary beams: assumptions confirmed

• effect of deformation in slab panel edge supports (no effect on capacity; increases panel midspan deformation, 65% contribution)

• included validation of FE model against Standard Fire Tests on composite and non composite beams

Steps to Implementing a Slab Panel Design

First design the floor and structural system for gravity and lateral loading conditions, then:

Step 1: Determine the size of the slab panel and location of the slab panel supports

Step 2: Determine which of the internal supports can carry negative moment

Step 3: Start with recommended reinforcement contents

Step 4: Input all variables and check capacity

If doesn’t work, follow the recommendations of the report for increasing the slab panel fire emergency load carrying capacity

The design of the floor and structural system for gravity and lateral loading determines the beam size, spacing, slab depth, concrete strength etc. Then implement SPM as follows: Step 1: Determine the size of the slab panel and location of the slab panel supports. These are the dimensions Lx and Ly . The length Lx is the distance between adjacent primary beams, which are support beams for sides 1 and 3 of the slab panel. The length Ly is the distance between points of effective slab panel secondary beam support – i.e. sides 2 and 4. Step 2: Determine which of the interior supports can carry negative moment, especially over the primary support beams. If sides 1 or 3 are interior, this negative moment capacity can be especially beneficial Step 3: Start with the reinforcement contents recommended by the report. These are for mesh as required for other purposes such as shrinkage and temperature crack control and interior support bars over internal supports carrying negative moment Step 4: Input all other variables and do the first check on the moment/tensile membrane load-carrying capacity of the slab panel. If this is satisfied, check shear capacity. If this is also satisfied, the design is complete If the design load carrying capacity is not adequate, follow the advice given at the end of section 3.1 to increase the load carrying capacity.

Moment/Tensile Membrane ResistanceThis uses the modified Bailey model, ie:

w* = G + Qu from Loadings Standard

wu ≥ w* required

where:

w* = fire emergency distributed loadwu = slab panel load carrying capacitywylθ = yieldline load carrying capacity in firewylθ,ss = simply supported yieldline load carrying capacity

in firee = tensile membrane enhancement factor

= fn (Lx, Ly, mx, my, teq, to, hrc fyr,θ, Eyr,θ)

to, hrc are slab thickness, deck rib height

fyr,θ, Eyr,θ are for reinforcement

( ) eww-w w ss,ylss,ylylu θθθ +=

The modifications to the Bailey method are covered in the following slides (only the modifications to the method are covered - features such as allowing for deck trough bars or two layers of slab reinforcement are not modifications to the procedure and so are not included in the list)

Differences between SPM & Bailey Method

• Moment resistance at the slab panel supports is accounted

for in the yieldline load carrying capacity calculation

• The contribution of composite secondary beams is included directly in the calculation of the slab panel capacity on

the basis of 100% composite action between the secondary

beams and the slabs

• Floor systems covered by the method include reinforced concrete slabs, slabs on profiled steel decks (clip angle

and trapezoidal decking) and slabs supported on closely

spaced cold formed steel joists.

• Modifications to the heat path method of reinforcement

determination within the concrete slab given by ECCS Tech

Note 82 and incorporated into EC4-1-2 have been made

based on the furnace testing of 6 slab panels to give

more accurate steel temperatures

Differences between SPM & Bailey Method

• The procedure is applicable to a wide range of composite floor systems including reinforced concrete flat slabs,

slabs on steel decking and flat slabs supported on

closely spaced rolled formed steel frame joists

• Bailey’s method assumes failure is a fracture across the short span of the slab panel. However, if the tensile

resistance in the Lx (short span) direction is less than

in the Ly direction, final fracture can be in the other

direction, ie involving a crack developing along the

midspan of the slab panel in the Ly direction, in which

case the Bailey method will overestimate the slab panel

capacity. The modified SPM (2009 modifications) accounts

for this.

• Orthotropic strength under fire conditions is taken into account .

• Equilibrium is maintained in the yieldlines at their intersections within the slab panel. This is not checked

in the Bailey model

Differences between SPM & Bailey Method

• Deflection limits originally proposed by Bailey

have been modified on the basis of furnace testing

of 6 slab panels and the increase in slab panel

with length of standard fire exposure is included

• The limits on reinforcement required for integrity

have been modified based on the furnace testing of

6 slab panels

• Detailing requirements to ensure that the slab

panel will dependably deform to the extent

required by the method without failure have been

included

• A shear check is included to preclude shear failure of the slab panel

Shear ResistanceThis is additional to the Bailey

model:

w* = G + Qu

φfire = 0.89 from standard

vc = conc. slab shear capacity

dv = effective shear depth

Vu,θ,sb= shear capacity of

secondary beam in fire

Ssb = spacing of secondary

beams

)2/(** xLwv =

vcfireslabu dvv φ=,

requiredS

Vvv

sb

sbu,

slabu,

,∗+≤

θ

The shear resistance of a reinforced concrete flat slab is almost independent of the reinforcement content, while the flexural/tensile membrane resistance is directly proportional to the reinforcement content. This means that as the reinforcement content is increased the latter increases while the shear resistance does not. This could potentially lead to a shear failure, which is brittle and sudden, at the supports instead of the desired flexural/tensile membrane failure which is non sudden and ductile. Bailey found this occurred in his ambient temperature testing of slab panels. It is less likely in fire but still is a potential mode of failure that this check suppresses. Having come from a seismic engineering background in which shear failure of concrete is undesirable and suppressed by suitable means it is natural for me to apply the same check in this fire case. The 0.89 factor comes from the phi factor for fire given in the steel standard which is phi,fire = (phi, ambient/0.85)

Example of Model: Parametric Fire

Slab Panel Support Beam

Secondary Beams

Reflected Floor Plan

Region modelled in FEA

Step 10: Distribution of Slab Panel Loads into Supporting Members

• Based on yieldline pattern

• Important is realistic

• FEM modelling showed

more realistic than ambient temperature

design practice

G+Q Fire - 44minHand calc.(HC) ABAQUS (ABQ) ((ABQ-HC)/ABQ)*100 SPM ABAQUS ((ABQ-SPM)/ABQ)*100

Column-1 (A-5) 64.8 43.5 -49.0% 55.0 71.8 23.4%

Column-2 (B-5) 159.9 180.2 11.3% 148.8 130.0 -14.5%

50% of Column A-4 18.9 29.6 36.1% 32.6 31.2 -4.5%

Total 243.6 253.3 3.8% 236.4 233.0 -1.5%

Distribution of slab panel loads into the supporting beams. This is important because these beams have to be designed to carry the loads distributed into them by slab panel action. This is assumed to be on the basis of the yieldline pattern tributary area. The question was asked by the international reviewer of the method how valid this assumption was and the research shown in that slide was undertaken to test the accuracy of that assumption. It is important because failure to correctly determine the loads being carried by the slab panel supporting beams and their columns in fire could lead to these being underdesigned or overdesigned.

Steps 11 and 12: Extending SPM to Panels With Unprotected Support Beams

• Current procedure requires slab panel support

beams to be protected unless very strong

• Need for this was investigated

• Potential relaxation of rules is possible,

however

• Will require experimental testing to confirm

adequacy

• Flexural torsional buckling of support beamslateral stiffening needed of edge support beams

Step 13: Review of Demands on Columns

• Columns designed for equivalent fire severity, te

• All SPM analyses undertaken on fire in one level only

• No column inelastic demand under fire

• Earlier Cardington test building monitoring looked at

fire from one floor to next

• Confident columns will support structure if fire

spreads from one floor to another

• Floors are effective as fire separations but won’t stop

spread up outside or un stopped internal voids

Design Application: Report and Computer Program SPM0306

Input Screen

Output

ScreenPrinted I/O Also

Available

Effect of secondary beam

Effect of removing secondary beam

Example of

SPM

application

to office building