Performance based fire design

Case studies conducted at Markku Kauriala Ltd

- Helsinki Olympic Stadium
- 14-story residential mass timber building

Dr. Mikko Salminen

13. Dani Hrvatske komore inzenjera gradevinnarstva

Opatija, 14.-16.6. 2018
CONTENTS

• INTRODUCTION
• STEEL STRUCTURES IN FIRE
• CASE: HELSINKI OLYMPIC STADIUM
• TIMBER STRUCTURES IN FIRE
• CASE: 14-STORY MASS TIMBER BUILDING
• CONCLUSIONS
INTRODUCTION – MODELLING OF
STRUCTURES AND FIRE
> 90 % structures in
fire are designed this
way (at the moment)

J-M Franssen:
Performance-
based fire ”Make it simple…or
design not! But not both”

A Buchanan:
”Consistent level of
crudeness”

Originally from: [Bisby, Training School for Young

Researchers, Malta, 2012]2012]
 INTRODUCTION – MODELLING OF STRUCTURES
AND FIRE (2)

J-M Franssen:
(Proceedings of the First Workshop Structures in
Fire 2000, Kööpenhamina)
[Buchanan, 2008]

Looking back some 10 or 20 years ago, the scientific community

dealing with the problem of fire in buildings could, schematically, be
divided into two separate groups; one group was dealing with the
fire side of the problem and was considering that the temperature
of 540°C was anything which had to be known concerning the
structure of the building; the other group was investigating the
behaviour of the structure, quite happy with the comfortable feeling
that the ISO curve was a perfect representation of the fire.

= Standard fire, traditional, nominal

fire curve, applied in the classification
of building products
STEEL STRUCTURES IN FIRE - GENERAL

• Reduction in strength and stiffness

• Non-linear stress-strain-curve
• Thermal expansion → compression
• Geometrical non-linearity
• Non-uniform temperature distributions

Analysis in fire are considerably more

challenging than in normal temperature
STEEL STRUCTURES – ADVANCED
CALCULATION METHODS

• Sophisticated tools are required: e.g. FDS, SAFIR,

Vulcan, ABAQUS, ANSYS
• Typical procedure presented below:

2. Temperatures of steel
1. Fire simulation (FDS)
profiles (SAFIR)

3. Behaviour of steel structure

(SAFIR)

R30 → R15
teräsrakenne
CASE: HELSINKI OLYMPIC STADIUM

• Built in 1938, XVth Olympic Games 1952

• Major renovation 2012 – 2019?
• All the stands will be covered – load bearing steel
trusses

Figure: K2S Arkkitehdit Oy

OLYMPIC STADIUM - INTRODUCTION

• Prescribed solution of first fire safety consultant: R 60 steel

trusses and sprinklers
• Final solution of Kauriala Ltd: R 15 steel trusses, no sprinklers,
some passive protection without sacrificing safety

Wooden
cladding
Wooden
seats

Kuva: K2S Arkkitehdit Oy

OLYMPIC STADIUM – STRUCTURES

• Line- and transverse steel trusses ja composite columns

• Line truss was selected to be analysed using advanced
methods (requirement of EN 1993-1-2)
• Close co-operation e.g. with structural designer

: Sweco Finland Oy
OLYMPIC STADIUM – CONDUCTED ANALYSIS

• More than 100 cases were analyzed

• Fire classes of timber (protected or not), locations and
sizes of the initial fire, effect of the wind
• FDS- and SAFIR-softwares were applied
OLYMPIC STADIUM – FIRE SIMULATIONS
OLYMPIC STADIUM – STRUCTURAL MODEL

• Line truss was assessed as the most critical part of structure

in fire and was analyzed using SAFIR-software (3d-model)
• First the behaviour of the truss was checked in normal
temperature (validation of the model with structural designer)
OLYMPIC STADIUM – TEMPERATURES

Adiabatic surface temperatures Steel profile

temperatures
OLYMPIC STADIUM – TRUSS DISPLACEMENTS

• Figure presents the maximum vertical displacement in the truss in

the worst fire scenario as a function of time
• Following sensitivity analysis were made: Case 1: Increase of the
mechanical loads by 50 %, Case 2: Increase of the cumulated heat
flux in the hottest zone by 50 % and Case 3: Extension of the
adiabatic surface temperatures of zone 1 to zone 2
OLYMPIC STADIUM – BEHAVIOUR OF THE TRUSS

• Deformations of the truss are scaled by factor 75

OLYMPIC STADIUM – EVACUATION ANALYSIS
OLYMPIC STADIUM – CONCLUSIONS

• For steel trusses, prescribed solution R 60 can be replaced by R

15 and most of the wooden seats and cladding can be left
unprotected. No sprinkler system is needed for the stands
• Prescribed methods in fire engineering may lead to highly
conservative and expensive solutions in some cases
• Significant benefits (cost-effectiveness, architectural issues,
sustainability, etc.) can be achieved by using performance-
based fire design without sacrificing safety
• Performance-based fire design in this case was highly iterative
process which required smooth co-operation between client,
architects, structural designers and fire safety consultants
TIMBER STRUCTURES IN FIRE: BACKGROUND –
FIRE REGULATIONS

• Many building regulations and standards restrict(ed) the use of

timber as a building material because of the combustibility of
timber
• Research projects on the fire behaviour of timber structures
have been conducted providing basic data on the safe use of
timber
• Many countries have started to revise their fire regulations
allowing greater use of timber
Development in regulations: Number of stories in timber

Schmid, J. 2014. Timber

buildings – research and
experience to get comparative
fire safety. Brand 2014.
Helsingborg.
 FIRE DESIGN OF TIMBER STRUCTURES

• Based mainly on Reduced cross-section method: the

original cross-section is reduced by the charring depth
~300 oC isotherm
Standard fire: constant charring
rates from EN 1995-1-2
mukaisesti

EN 1995-1-2. 2004. Eurocode 5: Design of timber structures – Part 1-

Frangi, A. 2012. Fire resistance assessment of timber structures, Basic
2: General – Structural
designfire design.
methods, Worked CEN. Brussels
examples. Workshop ’Structural Fire Design of
Buildings according to the Eurocodes’ – Brussels, 27-28 November
2012.
EN 1995-1-2, ANNEX A
• Can be applied woth parametric fires of EN 1991-1-2 Annex A
• Takes into account thermal properties of enclosure, openings and fire
load density (at some level)
• Constant charring rate at heating phase of fire, after that charring rate
decreases linearly
• Restrictions given in EN 1991-1-2 and 1995-1-2, too hot temperatures
with wooden enclosure – not applicable for every case

Parametric fire
Standard fire

Charring rate as a function of time

Example of parametric fire curve

EN 1995-1-2, ANNEX B

• Advanced calculation methods (FEM) can be applied to define

the char depth (also in natural fires)
• Thermal properties given in Annex B are applicable only in the
case of standard fire exposure

Thermal conductivity Specific heat Relative density

 PARAMETRIC FIRE MODEL OF HARMATHY & LIE

EN 1995-1-2. 2004. Eurocode 5: Design of timber structures – Part 1-

2: General – Structural fire design. CEN. Brussels
MODIFIED CONDUCTIVITY MODEL

• MCM (= Modified Conductivity Model) was developed based on

fire tests and other studies (e.g. Cachim & Franssen, Hopkin et
al.)

Hakkarainen

König et al.

Kinjo et al. Kinjo et al.

UNPROTECTED WOOD – INCREASE IN FIRE LOAD

• When performance-based approach is applied, the fire load

density [MJ/m2] is needed
• EC1: ”the fire load should consist of all combustible contents
and the relevant combustible parts of the construction…”
• Final fire load and char depth have to be defined iteratively
EXAMPLE OF ITERATIVE CALCULATION METHOD

• 9 % of all surfaces is unprotected (= charring) timber

• Fire load density 600 MJ/m2 (floor) = 141 MJ/m2 (all surfaces)
 ITERATIVE CALCULATION

• In this case 3 iteration rounds are needed to define the

final fire load density and char depth
Iteration qt,d,i [MJ/m2] dchar,i [mm] qadd,i [MJ/m2] qadd,i / qt,d,i
round (i)
1 141,18 44 31,43 22 %
2 172,61 53 6,43 4%
3 179,04 55 1,43 1%

Parametric fire curves (Harmathy & Lie)

1200

1000 Standard fire

800
Temperature [oC]

600

400
1
2
200 3

0
0 15 30 45 60 75 90
Time [min]
CASE: 14-STORY MASS TIMBER BUILDING,
JOENSUU LIGHT HOUSE

• Not possible without performance-based design

• Load-bearing structures and protective layers were
considered in the design
 LIGHT HOUSE - RISK ANALYSIS

• Comparison to acceptable solution

• The ”result” of the risk analysis is design fire load
• In the design scenario the effect of sprinklers was not taken into account
• Exposed timber taken into account in the fire load
 LIGHT HOUSE - RISK ANALYSIS (2)

• In Finland, there is lot of fire statistics available, e.g. fire load density
in apartments (e.g. 80 %-fractile ~ 600 MJ/m2)
• Event trees were applied in the risk analysis
LIGHT HOUSE – CHARRING DEPTHS

• Charring depths were calculated using SAFIR –software

• The increase from the structures was taken into account in the
fire load
• Validation of the calculation method to Canadian fire tests
(Hevia, 2014)
LIGHT HOUSE – VALIDATION OF THE CALCULATION
METHODS
LIGHT HOUSE - CHARRING DEPTHS

• Different amount of exposed timber and different protective

layers were considered in the early phase in the design

Charring depths of exposed

timber

Design fires:
Initial: 420 MJ/m2
10 % exposed timber: 580 MJ/m2
25 % exposed timber: 820 MJ/m2
LIGHT HOUSE - RESULT
• The building can be built with reasonable costs (and the process
has already started)
• Preliminary class of the structures is R 90 (sprinkled)
• Approximately 25 % of the interior surfaces in the apartments
could be exposed timber
CONCLUSIONS
• Performance-based fire design of timber buildings includes
challenges – but is possible
• More research and design methods are needed (natural fire,
charring, sprinkler reliability, exposed timber)
• It is challenging to make the performance-based design when
the final structure is not yet decided at preliminary design ->
fire safety designer is needed from the beginning to the end of
the project
• Smooth interaction is required between fire safety consultant
and architect, structural designer, other designers (HVAC,
acoustic), authorities, 3. party inspectors
 Thank you for your attention!

www.terasrakenneyhdistys.fi
[email protected]

31.5.2018 35