timber structures - lectures

1

Winter 2003 Design of Wood Structures 8-1

TIMBER STRUCTURES


OVERVIEW

• This section will:– describe the fire behaviour of timber construction – give design methods for heavy timber structural

members exposed to fire– briefly discuss fire behaviour of connections in timber

structures


DESCRIPTION OF TIMBER CONSTRUCTION

• Timber structures are divided into two categories: – heavy timber structures– light timber/wood frame construction

• Heavy timber construction describes all uses of large-dimension timber framing in buildings

• Heavy timber structures are principal structural elements (beams, columns, decks or truss)

• Light timber frame construction uses smaller sizes of wood framing (studs in walls, joists in floors)


Glulam

• 'Glue laminated timber' (glulam) are members made from several laminations glued together

• Fire tests have shown that glulam members exposed to fires behave in the same way as solid sawn-timber members of the same cross section


Fire Behaviour of Timber Structures

• Heavy timber members have good fire resistance• When large timber members are exposed to fires

the wood surface initially burns rapidly• The burned wood becomes a layer of char which

insulates the solid wood below and slow down the burning rate

• The char layer does not usually burn• Above 100°C, moisture in the wood evaporates• Some of this moisture travels out to the burning

face, but some travels into the woodWinter 2003 Design of Wood Structures 8-6

Fire-retardant Treatments

• Fire-retardant chemicals are available for treating wood to reduce its combustibility

• The purpose of the chemical treatments is to reduce the rate of flame spread

• Chemical pressure impregnation is effective• Impregnation can have some negative effects

– loss of wood strength – corrosion of fasteners

• Fire-retardant chemicals do not significantly improve the fire resistance of timber members

2


FIRE-RESISTANCE RATINGS

• Design process for fire-resistance requires that: provided fire-resistance > design fire severity

• Verification is usually in time or strength domain• Temp. domain is not used for timber structures

(no critical temp for fire-exposed timber)• Usually, fire design of heavy timber structures is

by calculation methods • Some countries have generic fire-resistance

ratings for heavy timber construction• There are very few proprietary ratings


WOOD TEMPERATURES

• When heavy timber members are exposed to severe fires, the outer layer of wood chars

• Boundary between the char layer and remaining wood corresponds to about 300°C temp.

• Below the char layer there is a layer of heated wood about 35 mm thick

• Layer above 200°C is the pyrolysis zone (thermal decomposition to gases, see Figure below)

• Moisture evaporates in the wood above 100°C• Structural design of heavy timber members is

based on the rate of charring of the wood surface


WOOD TEMPERATURESChar layer and pyrolysis zone in a timber beam


Temperatures Below the Char

• Temp. in wood below char layer was measured• For semi-infinite solid wood, temp. T (oC) below

the char layer is given by: T = Ti + (Tp - Ti)(1 - x/a)2

• Ti is the wood initial temp. (oC), Tp is the temp. at which charring starts (300°C), x is the distance below the char layer (mm), and a is the thickness of the heat-affected layer (40 mm)

• Janssens and White (1994) show that a better fit to experimental data is obtained with a = 35 mm


Thermal Properties of Wood

• Temp. inside fire-exposed timber members can be calculated using FEM

• Thermal properties are not well defined (especially at 100°C and over 300°C)

• Wood density varies greatly between species• After 100°C, density drops to ∼ 90% of its original

and to ∼ 20% of its original value above 300°C• Thermal conductivity varies greatly between

authors (see Figure below as an example)• Figure below shows specific heat variation with

temp. (spike means moisture evaporation)Winter 2003 Design of Wood Structures 8-12

Thermal Properties of WoodVariation of thermal conductivity of wood with

temperature

3


Thermal Properties of Wood

Variation of specific heat of wood with temperature


MECHANICAL PROPERTIES OF WOOD

• Wood is greatly different from other materials– wood strength is very variable– mechanical properties are different in different

directions– strength and ductility are different in tension

and compression– failure stresses depend on the specimens size– strength reduces under long duration loads

• Figure below shows different ways of loading of wood with different failure modes


MECHANICAL PROPERTIES OF WOOD

Loading of wood in different directions


Mechanical Properties of Wood at Normal Temperatures

• Tension and compression behaviour• Bending behaviour• Design values


Tension and compression behaviour

• Figure below shows typical stress-strain curves for wood specimens with no defects

• Parallel to grain vs. Perpendicular to grain• Compression vs. Tension• The wood is ductile in compression


Tension and compression behaviour

Stress--strain relationships for clear wood

4


Bending behaviour

• Bending behaviour is a combination of tension and compression

• Some ductility is available in timber beams when the material is stronger in tension than in compression


Design values

• Structural design calculations require values of the design strength of the wood material

• For limit states design, design stress is 5th percentile failure stress under short-duration loading

• Due to variations, characteristic stresses are usually obtained from in-grade tests of large numbers of representative samples

• The 5th percentile value for design in normal temp. conditions, may be modified to 20th percentile strength value for fire design


Design values

• Design strength of timber depends on duration of the applied load as a duration-of-load factor

• In limit states design, duration-of-load factor is 1.0 for short-duration loads and 0.8 or 0.6 for medium- and long-duration loads

• In working stress design, duration-of-load factor is 1.0 for long-duration loads and 1.25 or 1.6 for medium- and short-duration loads

• The duration-of-load factor for fire design should be the appropriate value for short-duration loads


Mechanical Properties of Wood at Elevated Temperatures

• Sources• Effect of moisture content• Plasticity• Parallel to the grain properties• Perpendicular to the grain properties• Shear• Derived results


Sources and general behaviour

• Review on the effect of moisture content (MC) and temperature on the mechanical properties of wood is given by Gerhards (1982)

• Wood properties are affected by steam at 100°C, wood begins to pyrolyse at about 200°C and turns into char by 300°C

• The range of interest for fire design is therefore from room temperature to 300°C


Effect of moisture content

• When testing timber at elevated temp., MC is sensitive to the test method and specimen size

• Some test specimens are maintained at constant MC throughout the test

• Some tests specimen are at a certain MC before the test and allowed to dry out when heated

• If wood is heated to a temperature above 100°C in dry air, all moisture will evaporate after some time

5


Parallel to the grain properties -Modulus of elasticity

• Figures below show the modulus of elasticity of wood at elevated temperatures

• The effect of temp. on modulus of elasticity parallel to the grain is roughly linear up to 200°C

• There is a scatter over 200°C• Figure below is another example of results

derived by Konig and Walleij (2000) from tests of 145 x 45 mm timber studs in insulated walls, exposed to ISO 834 fire while loaded in bending



Modulus of elasticity of wood parallel to the grain versus temperature



Modulus of elasticity of wood parallel to the grain versus temperature


Parallel to the grain properties -Tensile strength

• Below is a Figure showing stress-strain curves for temp. of 25°C and 90°C at low and high MC for samples tested by Ostman (1985)

• Failure stress at 90°C and 29.5% moisture content is about 60% of that of dry cool wood

• Another Figure shows a comparison among test data as derived by different researchers



Stress-strain relationships for wood in tension parallel to the grain



Tensile strength parallel to the grain versus temperature

6


Parallel to the grain properties -Compressive strength

• Figure below shows temp. effect on compressive strength parallel to the grain

• These results are for dry wood except the marked shaded region (MC > 12%)

• The Figure also shows the relationship derived by Konig and Walleij (2000)


Parallel to the grain properties -Compressive strength

Compression strength parallel to the grain versus temperature


Parallel to the grain properties -Bending strength

• Figure below shows limited bending test results collected by Gerhards (1982)

• The wood shows different slopes for different test results


Parallel to the grain properties -Bending strength

Bending strength of wood versus temperature


Perpendicular to the grain properties -Modulus of elasticity

• For modulus of elasticity perpendicular to the grain, Gerhards (1982) reports eight studies as shown in Figure below for temp. up to 100°C

• The dependence on temperature tends to be greater for moisture content above 12%, but there is a lot of overlap between the studies


Perpendicular to the grain properties -Modulus of elasticity

Modulus of elasticity perpendicular to grain versus temperature

7


Perpendicular to the grain properties -Tensile strength

• Temp. effect on tensile strength perpendicular to the grain is shown in the Figure below

• There is a wide range of results for different MC, but a trend of a greater strength reduction as the moisture content increases

• There are no results of tests over 100°C


Perpendicular to the grain properties -Tensile strength

Effect of temperature on tensile strength perpendicular to the grain


Perpendicular to the grain properties -Compressive strength

• Figure below shows temp. effect on strength in compression perpendicular to the grain

• This shows data from five studies reported byGerhards (1982) (overlap and scatter)


Perpendicular to the grain properties -Compressive strength

Effect of temperature on compression strength of wood perpendicular to the grain


Derived results - Reduction factors

• Figure (a), temp. effect on mechanical properties– modulus of elasticity is assumed to drop linearly to

50% of its normal temperature value at 300°C– tension strength follows the same relationship to

200°C, then drops to zero at 300°C (wet or dry)– compression strength for dry wood drops linearly to

zero at 300°C– compression strength for wet wood drops to 50% at

100°C and remain constant until it reaches 160°C, after which it follows the relationship for dry wood

• Figure (b) shows temp. profile below char layer• Figure (c) shows drop in wood strength below

char layer (significant reduction below 25 mm) Winter 2003 Design of Wood Structures 8-42

Derived results -Reduction factors(a) Effect of temp. on

mechanical properties of wood

(b) Temp. profile below char layer

(c) Reduction in strength of wood below char layer

8


Derived results -Stress-strain relationship

• Figure below shows stress-strain curves derived by Konig and Walleij from computer modelling

• The curves are idealized in a simple way• In the tension region, linear elastic behaviour has

been assumed until failure• In the compression region, elasto-plastic

behaviour has been assumed• The curves include the creep effects


Derived results -Stress-strain relationship

Derived stress-strain relationships for wood at elevated temperatures


DESIGN CONCEPTS FOR HEAVY TIMBER EXPOSED TO FIRE

• Large timber members have good fire-resistance • Fire-resistance can be calculated if charring rate

is predicted on surfaces exposed to standard fire• Figure below shows common fire exposures

(three-/four-sided of rectangular members)• The original cross section b x d is reduced to

residual cross section bf x df after charring• The depth to the char front c (mm) is given by:

c = β x t• β is the charring rate (mm/min), and t is the fire

exposure time (min)Winter 2003 Design of Wood Structures 8-46


Design concepts for large timber members



• Dimensions of the residual cross section are:bf = b - 2c df = d - c (three-sided exposure)df = d - 2c (four-sided exposure)

• Char temperature is about 300°C• There is a layer of heated wood about 35 mm

thick below the char layer• Structural design of timber members is based on

the strength and stiffness of the residual member


Verification

• Verification of strength during fire exposure:U*

fire ≤ Rfire

• U*fire is the design force and Rfire is the load

capacity• The design force U*

fire may be axial force N*fire ,

bending moment M*fire or shear force V*

fire

• The load capacity is calculated as axial force Nf, bending moment Mf or shear force Vf

9


Verification - Simply supported beams• For a beam under a bending moment:

M*fire ≤ Mf

• M*fire is the bending moment and Mf is the design

flexural capacity under fire conditions given by: Mf = Zf ff

• ff is the design strength of wood in fire conditions (MPa) and Zf is elastic section modulus (mm3)

• The value of ff should always be the strength under short-duration loads

• For rectangular sections with no corner rounding:Zf = bf df

2 / 6Winter 2003 Design of Wood Structures 8-50

Charring rate• Rate of charring (under standard fires) depends

on the density and moisture content of the wood• Many codes specify a constant charring rate of:

– 0.60 - 0.75 mm/min for softwoods – about 0.5 mm/min for hardwoods

• The effect of density, ρ (kg/m3), and MC on the charring rate is shown in Figure below, given by the equation below for charring rate β (mm/min):

β = 0.4 + (280 / ρ)2

• Table below shows recommended charring rates


Charring rateCharring rate as affected by density and MC


Charring rate

Charring rates for design

• β for actual cross sections with rounded corners• β1 (10% larger notional charring rate) is for no

corner rounding

Char rateMaterial Minimum density

(mg/m3)β

(mm/minute)β1

(mm/minute)Glue-laminated softwood timberSolid or glue-laminated hardwood timber

290450

0.640.50

0.700.55

Softwood panel products (plywood, particleboard) minimum thickness 20 mm

450 0.9


Charring rate

• In North America, recommendations for charring rate are given by AFPA (White, 1988)

• The proposed charring rate β is the average charring rate (mm/min) given by:

β = 2.58 βn / t0.187

• βn is a nominal charring rate (βn=0.635 mm/min) and t is the time (min)

• The resulting char layer thickness c (mm) is:c = β t = 2.58 βn t0.813

• Figure below shows the resulting depth of char during 4 hours of standard fire exposure


Charring rate

Depth of char from North American recommendations

10


Corner Rounding

• All fire tests of large rectangular timber sections show some rounding of the corners

• Figure below shows a typical charred cross section• Most design codes assume the radius of the

rounding as equal to the depth of the charred layer• If corner rounding is considered in beams exposed

to fire on 3 sides, the section modulus Zf,r of the reduced cross section is given by:

Zf = bf df2 / 6 - 0.215 r2 df

• bf is the beam residual width, df is beam residual depth, and r is the radius of the charred corner


Corner Rounding

Residual cross section of timber beam exposed to fire


Effect of Heated Wood Below the Char Line

• There are several alternative design methods to allow for heated wood below the char line

• Some codes ignore any reduction of wood strength below the char, which can lead to unsafe results for small cross sections

• There are two methods to account for variable temperature inside the unaffected region:– the effective cross section method – the reduced properties method


Effective cross section method

• The effective cross section method accounts for heated wood below char by removing a nominal layer of zero strength from the cross section

• Wood in the effective cross section is assumed to have normal temperature properties

• The flexural capacity, Mf = Zf ff, is calculated with no corner rounding, with Zf for 3-sided exposure:

Zf,z = (bf - 2z)(df - z)2 / 6 • z is the thickness of zero-strength layer (mm)• The design strength of wood is the strength at

normal temp. fb (MPa) so that ff = fb


Effective cross section method

• In Eurocode, the thickness of zero-strength layer:– z = 7 mm for more than 20 min fire exposure– 0 < z < 7 mm for less than 20 min fire exposure

(reduced proportionately)• The AFPA North American design method

increases the nominal charring rate by 20% to allow for the heated wood below the char line

• Using the effective cross section method in accordance with Eurocode, the charring rate β1 should be used (earlier Table)


Reduced properties method

• The reduced properties method (Eurocode) is based on a strength reduction factor kf applied to all of the wood below the char layer

• The flexural capacity is Mf = Zf ff, with Zf and ff asZf = bf df

2 / 6 - 0.215 r2 dfff = kf fb

• Zf includes corner rounding (β from earlier Table)

11


Reduced properties method

• kf is strength reduction factor for residual sections approximated by: – kf = 0.8 for NA design equations (Lie,1977) – kf = 1.0 - 1/g (Ar / p) for Eurocode design equations,

where p is the perimeter of the fire-exposed residual cross section (m), Ar is the area of the residual cross section (m2), g is a factor (m-1), with the value of 200 for bending, 125 for compression and 330 for tensile strength and modulus of elasticity


Characteristic strength of wood• For normal temp. design, characteristic design

strength is taken as the 5th percentile value• In most limit states design, 5th percentile strength

value f0.05, obtained from tests, is listed in codes• For fire design, most codes use 5th percentile

strength value f0.05 so that fb = f0.05• Some codes modify 5th percentile strength f0.05 to

20th percentile for fire design so that design strength fb for fire conditions is fb = k20 f0.05

• k20 is a correction factor to convert 5th percentile to 20th percentile values (1.25 for solid timber, 1.15 for glulam in Eurocode)


Characteristic strength of wood

• The AFPA method uses the mean value of wood strength for fire design (working stress design)

• In the method, the allowable stress in the code fais modified to give an allowable stress in fire conditions fa,f using

fa,f = kmean fa• fa is the code allowable stress (MPa), kmean is a

correction factor to convert allowable stresses to mean values (2.85 for tension and bending, 2.58 for compression, 2.03 for buckling failures)


WORKED EXAMPLE 1

Consider a softwood glulam beam, 130 mm wide by 720 mm deep, spanning 7.5 m with a dead load G = 4.0 kN/m (including self weight) and live load Q = 7.0 kN/m. The beam is laterally restrained with timber decking nailed to the top edge. Check the design for normal conditions and for 60 minutes fire-resistance rating, exposed to fire on three sides. Use the Eurocodemethod with the charring rates from Table 8-52 and the factor k20 = 1.15.


WORKED EXAMPLE 1The characteristic flexural strength is fb = 17.7 MPa.

The strength reduction factor is Φ = 0.8 for normal design and Φf = 1.0 for fire design. The duration-of-load factor is kd = 0.8 for cold design and kd = 1.0 for fire design.

Check design for normal conditions • Design load

wc = 1.2G+1.6Q = 1.2x4.0+1.6x7.0 = 16.0 kN/m • Bending moment

M* = wcL2/8 = 16.0x7.52/8 = 112 kNm• Section modulus:

Z = bd2/6 = 130x7202/6 = 11.2x106 mm3


WORKED EXAMPLE 1

• Nominal strengthMn = kd f0.05 Z = 0.8x17.7x11.2 = 159 kNm

• Design strengthΦ Mn = 0.8x159 = 127 kNm M* ≤ Φ Mn so design is OK.

Loads for fire conditions • Design load

wf = 1.0G+0.4Q = 1.0x4.0+0.4x7.0 = 6.8 kN/m • Bending moment

M*fire = wfL2/8 = 6.8 x 7.52/8 = 47.8 kNm

12


WORKED EXAMPLE 1Method I (effective cross section, no corner

rounding)• Rate of charring: β1 = 0.7 mm/min • Depth of char: c = 60 x 0.7 = 42 mm • Reduced breadth: bf = 130-2x42 = 46 mm • Reduced depth: df = 720-42 = 678 mm• Thickness of zero-strength layer: z = 7 mm• Effective breadth: be = 46-2x7 = 32 mm• Effective depth de = 678-7 = 671 mm • Section modulus:

Zf = be de2/6 = 32x6712/6 = 2.40 x 106 mm3


WORKED EXAMPLE 1

• Flexural strength:Mf = kd ff Zf = kd k20 f0.05 Zf = 1.0x1.15x17.7x2.4Mf = 48.9 kNm

• M*fire ≤ Mf so design is OK.

Method II (reduced properties, no corner rounding) • Rate of charring: β1 = 0.7 mm/min • Depth of char: c = 60 x 0.7 = 42 mm • Reduced breadth: bf = 130-2x42 = 46 mm• Reduced depth: df =720-42 = 678 mm


WORKED EXAMPLE 1• Section modulus:

Zf = bf df2/6 = 46x6782/6 = 3.52x106 mm3

• Beam area: A = bf df = 46x678/106 = 0.0312 m2

• Beam perimeter:p = bf + 2df = (46+2x678)/103 = 1.40 m

• Reduction factor:kf = 1-p/200A = 1-1.40/(200x0.0312) = 0.775

• Flexural strength:Mf = kfkdk20f0.05Zf = 0.775x1.0x1.15x17.7x3.52 Mf = 55.6 kNm



WORKED EXAMPLE 1

• Method III (reduced properties, corner rounding)• Rate of charring: β = 0.64 mm/minute • Depth of char: c = 60 x 0.64 = 38.4 mm • Reduced breadth: bf = 130-2 x 38.4= 53.2 mm• Reduced depth: df =720-38.4=682 mm• Section modulus:

Zf = bfdf2/6-0.215c2df

Zf =53.2x6822/6-0.215x38.42x682=3.90x106 mm3

• Beam area: A =bf df = 53.2x682/106 = 0.0363 m2


WORKED EXAMPLE 1

• Beam perimeter:p =bf+2df =(53.2+2x682)/103 = 1.42 m

• Reduction factor:kf = 1-p/200A = 1-1.42/(200x0.0363) = 0.805

• Flexural strength:Mnf = kfkdk20f0.05Zf = 0.805x1.0x1.15x17.7x3.9Mnf = 63.9 kNm



WORKED EXAMPLE 2

Repeat Example 1 using the NA charring rate in the working stress design format. The allowable stress under long duration loading in flexure is fa = 8.0 MPa. The factor to convert allowable stress to mean failure stress is kmean = 2.85.

Check design for normal conditions• Design load: w = G+Q = 4.0+7.0 = 11.0 kNm • Bending moment:

M*w = wL2/8 = 11.0x7.52/8 = 77.3 kNm

13



Z = bd2/6 = 130x7202/6 = 11.2x106 mm3

• Flexural stress:f*b = M*

w/Z = 77.3x106/11.2x106 = 6.91 Mpa• f*b ≤ fb so design is OK.

Fire design (NA char rate, no corner rounding)• Time of calculation: t = 60 minutes • Depth of char:

c = 2.58 βn t0.813 = 2.58x0.635X600.813 = 45.7 mm


WORKED EXAMPLE 2

• Reduced breadth: bf = 130-2x45.7 = 38.6 mm• Reduced depth: df = 720-45.7 = 674 mm • Section modulus:

Z = bf df2/6 = 38.6x6742/6 = 2.92x106 mm3

• Flexural stress:fb,f = M*

w/Z = 77.3x106/2.92x106 = 26.4 MPa• Allowable stress:

fa,f = kmean fa = 2.85x8.0 = 22.8 Mpaf*b,f ≤ fa,f so the beam fails in fire.


Design for Real Fires

• For realistic fires the Eurocode gives charring rates and strength reduction factors for a particular class of real fires

• This method is based on the work of Hadvig(1981)

• Table below shows values of char rate for real fires

• More details are given in the textbook


Design for Real Fires

Char rate, char time and char depth for parametric fire exposure

Initial char time to (min) Total char depth c (mm)Fuel load (MJ/m2 total area) Fuel load (MJ/m2 total area)

Openingfactor(ml/2)

Char rateβbar

(mm/min.) 80 160 240 320 400 80 160 240 320 4000.02 0.39 24 190.04 0.70 12 24 36 17 34 500.06 0.85 8.0 16 24 32 40 14 27 41 55 680.08 0.95 6.0 12 18 24 30 11 23 34 45 570.12 1.05 4.0 8.0 12 16 20 8.4 17 25 34 420.20 1.15 2.4 4.8 7.2 10 12 5.5 11 16 22 270.30 1.20 1.6 3.2 4.8 6.4 8.0 3.8 7.7 11 15 19


Empirical Equations• Most NA codes include equations to calculate

fire resistance of beams and columns (Lie, 1977)• Lie’s simple equations assumed:

– a uniform charring rate of 0.6 mm/min– section remains rectangular and the residual core has

80% of initial strength under char layer• For beams, the time to failure tf (min) is given by:

tf = 0.1 z b (4 -b/d) (3-sided exposure)tf = 0.1 z b (4 -2b/d) (4-sided exposure) z = 0.7 + 0.3 / Ra

• Ra is ratio of actual to allowable load at normal temp. and dimensions are in mm


Empirical Equations

• For columns, the time to failure tf (min) is: tf = 0.1 z b (3 - d/2b) (3-sided exposure)tf = 0.1 z b (3 - d/b) (4-sided exposure)

• For long columns, z is calculated by trial and error (see textbook)

• For short columns, z is based on better fitting with experimental results for columns of low slenderness ratio, given by:

z = 0.9 + 0.3 / Ra

14


WORKED EXAMPLE 3Calculate the time to failure for the beam in Worked

Example 1 using NA empirical design equation.• Design bending moment: M* = 112 kNm • Design strength: Φ Mn = Φ k1 fb Z = 127 kNm • Load ratio: Ra = M* / Φ Mn = 112/127 = 0.882 • z factor: z = 0.7 +0.3/Ra = 0.7+0.3/0.882 = 1.04• Time to failure:

tf = 0.1 z b (4-b/d) tf = 0.1x1.04x130 (4-130/720) = 50.1 min

• Time to failure is less than 60 minutes, so the beam fails in the fire.


DESIGN OF HEAVY TIMBER MEMBERS EXPOSED TO FIRE

• Beams• Tension Members• Columns• Beam-columns• Decking• Timber-concrete Composite Structures


Beams

• Beams can be designed using the same design equations as for normal temperature conditions, with modifications for strength and cross section

• It is important to determine which surfaces of the beam are exposed to fire (see Figure below)

• In addition to flexural strength calculations, lateral torsional buckling must also be checked

• Shear stresses are not a concern for rectangular beams, but should be considered for I-beams

• Deflections are not usually of concern


Beams

Three-/four-sided beam exposure


Tension Members

• Tension members are not affected by the possibility of buckling

• The tensile load capacity of a fire-reduced cross section can be calculated using one of the design methods– effective cross section– Reduced properties


Columns• Short columns strength depends on material

crushing strength and reduced cross section• Long columns strength (buckling increases with

time) depends on moment of inertia and modulus of elasticity of reduced cross section

• Lateral stability is very important for columns• Columns built into walls may have better fire

resistance (less charring and lateral restraint)• Tests on 16 columns (Malhotra et al. 1970)

achieved fire-resistance ratings between 30 and 90 min, depending on load and slenderness ratio

15


Beam-columns

• A 'beam-column' is a member subjected to combined bending and axial loading

• The design approach is to check the general interaction formula including both flexural strength and axial load capacity, such as:

(N/Nu)2 + M/Mu ≤ 1• N = applied axial load (kN), Nu = axial load

capacity with buckling effects (kN), M = applied bending moment (kN-m), and Mu = flexural capacity with lateral buckling effects (kN-m)


Decking• Assessment of fire resistance of decking must

consider all three possible failure criteria of stability, integrity and insulation

• Solid wood decking includes solid timber orglulam timber planks laid flat and butted together with tongue and groove edges, and timber planks set on edge and nailed together (see details in textbook)


Decking - Stability

• The stability criterion can be assessed in the same way as for beams and columns

• Janssens (1997) proposed an empirical design formula for structural performance of solid decks (based on a temperature and charring model)

• The time to structural failure tsf (min) is given by: tsf = 1.25 d (1 - √0.4Ra) - 11.3

• d is the thickness of the deck (mm), and Ra is the ratio of the applied load to the allowable design load


Decking - Integrity

• The integrity criterion may be the most difficult to satisfy for wood deck systems

• The difficulties arise at the junctions between the planks, which may increase in width due to shrinkage of wood which often occurs during the life of a building

• Tongue and groove joints between the planks are the best solution


Decking - Insulation

• If the integrity and stability criteria are satisfied, there will be no problem meeting the insulation criterion, because the thickness of remaining wood required to carry applied loads will be greater than that required to prevent excessive temperature rise on the top surface


WORKED EXAMPLE 4

A solid timber deck consists of 150 mm thick planks joined with central splines as shown in Figure 10.35(c). The deck spans 5 m with a superimposed dead load of 1.25 kN/m2 and live load 5.0 kN/m2. Calculate the failure time using Janssen's formula. Use the Eurocode reduced properties method to calculate if the deck has a 90 minute fire-resistance rating.

16


WORKED EXAMPLE 4

The characteristic flexural strength of the decking timber is fb = 25.0 MPa. The density of the wood is 5.0 kN/m3. The strength reduction factor Φ is 0.8 for normal design and Φf = 1.0 for fire design. The duration of load factor is kd = 0.8 for cold design and kd = 1.0 for fire design. The factor kfis 1.15 for fire design.


WORKED EXAMPLE 4Check for normal conditions • Thickness of deck: d = 150 mm • Self weight of deck: ws = ρd =5x0.15=0.75 kN/m2

• Total dead load: G = 0.75+1.25 = 2.0 kN/m2

• Design load:wc = 1.2G+1.6Q = 1.2x2.0+1.6x5.0 = 10.4 kN/m2

• Design a strip 1 m wide. Uniformly distributed load = 1.0 x 10.4 = 10.4 kN/m

• Bending moment:M* = wcL2/8 = 10.4x52/8 = 32.5 kNm



Z = bd2/6 = 1000x1502/6 = 3.75x106 mm3

• Design strength:Φ Mn = Φ k1 fb Z = 0.8x0.8x25x3.75 = 60.0 kNmM* ≤ Φ Mn so design is OK.

Janssen's formula • Load ratio: Ra = M* / Φ Mn = 32.5/60 = 0.54• Time to failure:

tsf = 1.25 d (1-√(0.4Ra))-11.3 tsf = 1.25x150 (1- √(0.4 X 0.54))-11.3 = 89 min


WORKED EXAMPLE 4

Eurocode reduced properties method • Design load:

wc = 1.0G+0.4Q = 1.0x2.0+0.4x5.0 = 4 kN/m2

• Design a strip 1 m wide. Uniformly distributed load = 1.0 x 4.0 = 4.0 kN/m

• Bending moment:M*

fire = wc L2/8 = 4x52/8 = 12.5 kNm • Rate of charring: β = 0.64 mm/min • Depth of char: c = 90 x 0.64 = 57.6 mm


WORKED EXAMPLE 4• Reduced depth: df = 150-57.6 = 92.4 mm • Section modulus:

Zf = bdf2/6 = 1000x92.42/6 = 1.42 x 106 mm3

• Section area:A = b df = 1000x92.4/106 = 0.00924 m2

• Exposed perimeter: p = b = 1.0 m • Reduction factor:

kf = 1-p/200A = 1-1.0/(200x0.00924) = 0.46 • Design strength: Mf = kf kd k20f0.05 Zf

Mf = 0.46x1.0x1.15x25x1.42 = 18.8 kNm M*

fire ≤ Mf so design is OK.Winter 2003 Design of Wood Structures 8-96

BEHAVIOUR OF TIMBER CONNECTIONS IN FIRE

• The ability of a structure to carry loads depends on the strength and stiffness of the structural members and connections between members

• Under fire, both members and connections must perform throughout the fire exposure

• Most connections are either metal fasteners or adhesives (very different fire performance)

• Little research has been done on performance of connections in timber structures exposed to fire

17


Metal Fasteners

• The behaviour of metal fasteners depends on the temperature of the metal because:– it affects the strength of the fastener itself– high temperatures lead to charring or loss of strength

of wood in contact with the metal• Geometry and protection of metal fasteners are

explained in more details in the textbook


Nails and Screws

• Nails are one of the best types of connection in timber structures because they penetrate wood and do not weaken the wood with drilled holes

• Screws have many of the advantages of nails including better gripping capacity than nails

• Noren tested nailed splice joints in tension exposed to the ISO 834 standard test fire

• Time to failure was inversely proportional to applied load, varying from 6 to 21 min


Bolted Connections

• Bolted connections are widely used in timber structures with excellent results

• Fire behaviour of bolted connections depends on the amount of heat able to enter the wood through the bolts

• The theory for nails could be applied to bolted connections but no comprehensive studies have been published


Truss Plates

• Truss plates have been shown to have a poor reputation for fire resistance

• Tests (White et al. 1994) on truss plates under ASTM E-119 standard fire exposure up to 300°C

• In the tests, unprotected plates failed in less than 6 min compared with ∼ 13 min for solid timber with no connection

• Various combinations of protection increased the fire resistance, the best gave over 30 min fire resistance when all 4 sides of the member were protected with 13 mm Type X gypsum plaster


Glued Connections

• Many timber structures and timber members are connected with adhesive

• When exposed to fire, glued wood members generally behave in the same way as solid wood provided that thermosetting adhesives are used

• Some adhesive such as epoxies are sensitive to elevated temperatures and should not be relied on in fire conditions

timber structures - lectures

Documents

design of wood structures

exposed timber

solid wood

timber structureswinter

wood surfacewinter

remaining wood

burned wood

large timber members