Effectiveness assessment of road tunnel fire-fighting strategies by ventilation and water mist systems

(1)

Objectives Road tunnel Fire protection strategies Fire scenarios Results Conclusions Appendix Fire ref. Numerical tools

1

XII Convegno nazionale AIIA "I Sistemi di Gestione della Sicurezza Antincendio nella Fire Safety Engineering”

Milano, 29 ottobre 2009

Effectiveness assessment of road

tunnel fire-fighting strategies by

ventilation and water mist systems

Manzini G. Politecnico di Milano Department of Energy Objectives Road tunnel Fire protection strategies Fire scenarios Results Conclusions Appendix Fire ref. Numerical tools

2

Effectiveness assessment of road tunnel fire-fighting strategies by ventilation and water mist systems

XII Convegno nazionale AIIA "I Sistemi di Gestione della Sicurezza Antincendio nella Fire Safety Engineering” Milano, 29 ottobre 2009

Galli E. MM S.p.A. Unità tecnica impianti

Objectives

In order

to assess the effectiveness of fire-fighting strategies

applied

to an urban twin-bore

road tunnel

, a numerical analysis was done,

including

three different numerical tools: a one - dimensional model

(SES), a lumped – parameters model (ECART) and a CFD model

(FDS)

.

The study focuses on verifying the capability of longitudinal and

transversal

ventilation strategies

adopted when fire occurs at a given

location, of

preventing the spread of smoke along the tunnel

.

These simulations have permitted to verify that some of the strategies

considered are sufficient to confine and extract smoke, in order to

determine satisfactory safety condition during the estimated egress time

for people and for the eventual intervention of fire brigades.

In particular, temperatures and combustion products concentration

along the tunnel show that

successful strategies

, in terms of

smoke

confinement and temperature mitigation

, can be achieved with the

ventilation system

. The activation of a

water mist system

can help

fighting the fire but some critical aspects must be taken in account if

discharging while people are still in the tunnel.

(2)

3

Milano, 29 ottobre 2009

Road tunnel

The

twin-bore tunnel

, assumed for this study, is equipped with

jet fans at

portals plus a transversal ventilation system

, to guarantee the fresh air

flow and the vehicle smoke extraction during normal operation and, with

properly sized fans, the smoke control and extraction in case of fire.

The traffic is

unidirectional traffic and each

bore (tube) consists of 2

traffic lines and 1 for emergency.

The tunnel consists in many areas per each bore, with different geometric

characteristics and intake and exhaust air vents.

Effectiveness assessment of road tunnel fire-fighting strategies by ventilation and water mist systems

XII Convegno nazionale AIIA "I Sistemi di Gestione della Sicurezza Antincendio nella Fire Safety Engineering” Milano, 29 ottobre 2009

Road tunnel

The air flows are provided by a ceiling exhaust ventilation ducts running the length of the

tunnel, and by lateral ducts for fresh air intake, one per bore. During the normal exercise

of the tunnel the fresh air income is provided by lateral air vents distributed every 10 m of

length almost at tunnel bottom level. The

exhaust system

extract the same air flow (normal

operation), and it is consisting of

air vents at ceiling disposed every 50 m

.

When fire

occurs, the supply of fresh air is reduced at 20% and the extraction is very increased, but

limited to a 200 – 250 m length area (consisting in four or five activated exhaust vents

identifying the “protected area”)

including the fire and its surrounding. In addition, totally

invertible jet fans are provided in each tunnel portal, used normally to push the air in the

traffic direction; but, in case of fire, is possible to reverse their airflow direction to help the

confinement of smoke in the extraction – protected area, by pushing air only inside the

tunnel.

(3)

5

Milano, 29 ottobre 2009

Fire protection strategies

The objective of the strategies is to confine the smoke in the protected area (in order to

extract them) and mitigate the temperature outside such area, with the aim to impose

satisfactory

safety conditions during the expected egress time for people and for the eventual

intervention of fire brigades

. Without the local extraction, determined by the transversal

ventilation and, eventually, helped by the jet fan inversion, temperatures and smoke

concentration won't be controlled as long as necessary.

Time [s] Event 0 Fire start 120 Fire detection 160 Fire ventilation activation 180 Fire ventilation is at regime

The hypothesized high-pressure water-mist deluge system is aimed to control, suppress and extinguish the

combustion and has the following main characteristics:

- group of 30 m long racks of nozzles disposed at ceiling, consisting of num. 4 lines of 10 nozzles each;

-injection is activated at the same time of the fire ventilation system (160 s from the fire start) with an

operative pressure of 80 bar and a K factor of 3.8 l/(min bar

1/2

), with a characteristic diameter supposed as

Dv

₉₀

= 200

μ

m;

- two consecutive racks of nozzles (corresponding to 60 m of tunnel length) are used.

(In addition, some simulations have considered the application of a low pressure water mist, with the same

configuration and volumic flow, but with a nozzles pressure of 12 bar, and others cases have been inherent

a high pressure water mist system with two of the four ceiling rack lines of nozzles put on the wall

-horizontal injection, close to tunnel bottom).

6

XII Convegno nazionale AIIA "I Sistemi di Gestione della Sicurezza Antincendio nella Fire Safety Engineering” Milano, 29 ottobre 2009

Fire scenario analysis

The fires considered have been determined on the basis of the references available upon car and HGV

fires and are illustrated in the figures. The curve with the maximum HRR of 30 MW (relative to an HGV

fire, HRR

₄

) has been applied as consistent with PIARC and Metropolitana Milanese indications for the

design of fire protection systems. In addition an HRR curve with a maximum of 10 MW (relative to a fire

comprising 2 or 3 cars and derived from the one with the 30 MW peak, HRR

4

’) was adopted for some

simulations.

0 5 10 15 20 25 0 500 1000 1500 2000 2500 3000 H R R [M W ] . t [s] Cars HRR HRR1 HRR2 HRR3 HRR4' 0 20 40 60 80 100 120 140 160 180 0 500 1000 1500 2000 2500 3000 H R R [ M W ] . t [s] HGV HRR HRR4 HRR5 HRR6

(4)

7

Milano, 29 ottobre 2009

Fire scenario analysis

To achieve the objectives, five main tasks have been carried out and they are briefly

enumerated as follows:

• Phase 0: identification of design fire scenarios;

• Phase 1: preliminary analysis with simple correlations to assess critical velocity,

confinement velocity, smoke production and air/smoke velocity causing droplets floating;

• Phase 2: simulations of the relevant fire scenarios by the SES (used to study the ventilation

strategies adopted and the consequent airflows in tunnel sections);

• Phase 3: simulations of the relevant fire scenarios by ECART (focusing on the different

ventilation strategies and on the combinations of different ventilation strategies plus a deluge

water mist system);

• Phase 4: simulations of the most meaningful fire scenarios with only ventilation strategies

(selected on the basis of the previous results) by FDS.

Numerical simulations have been carried out to assess the effect of the different fire

protection strategies. By SES, only the ventilation strategies have been analyzed, while by

ECART the ventilation and the ventilation + fire control strategies have been analyzed;

finally, by FDS, have been analyzed only some ventilation strategies (some ventilation + fire

control strategies have been analyzed only for research purposes, on the basis of insufficient

capabilities of FDS for water mist fire scenario simulations).

Effectiveness assessment of road tunnel fire-fighting strategies by ventilation and water mist systems

Fire scenario analysis results

Segment air/smoke temperatures curves corresponding to different time instants [°C] (HRR

4

, Fire

scenario 2, Fire operation ventilation strategy, 4 extraction vents, Fire segment= 5),

SES

tool

results.

SES

road tunnel discretization

scheme (long segments cases).

(5)

9

Milano, 29 ottobre 2009

Fire scenario analysis results

Segment air/smoke CO

2

concentration [ppm] – red line: under volume, blue line: upper volume

-simulation segments 50 – 500 m (“Progressiva” segment coordinates from the west portal) (HRR

₄

,

Fire scenario 2, Fire in 250 – 300 m segment, Fire operation ventilation strategy, 4 extraction vents),

ECART

tool results.

10

XII Convegno nazionale AIIA "I Sistemi di Gestione della Sicurezza Antincendio nella Fire Safety Engineering” Milano, 29 ottobre 2009

Fire scenario analysis results

Segment air/smoke Temperatures [°C] – t=

498 s from the fire start – south bore (HRR

4

,

Fire scenario 2, from the figures bottom: Fire

operation ventilation strategy num. 8

extraction vents, Fire operation ventilation

strategy num. 4 extraction vents),

ECART

tool results.

Segment air/smoke CO

2

concentrations [ppm] / Temperatures [°C] – t= 1000 s from the fire start –

south bore (HRR

₄

, Fire scenario 2, from the figures bottom: Fire operation ventilation strategy, Fire

operation ventilation + water mist injection strategy, Fire operation ventilation + water mist* injection

strategy, 4 extraction vents),

ECART

tool result.

(6)

11

Milano, 29 ottobre 2009

Fire scenario analysis results

Segment air/smoke Temperatures [ °C] – t= 119, 299 s from the fire start – south bore (HRR

₄

, Fire

scenario 2, Fire operation ventilation, 4 extraction vents),

FDS

tool results.

Effectiveness assessment of road tunnel fire-fighting strategies by ventilation and water mist systems

The hypothesized tunnel ventilation system (which complies with PIARC standards) provides sufficient

safety condition for people to egress the tunnel.

The activation of a water mist system (wheather high or law pressure) can be useful in order to fight the fire

and confine the smoke but provides some difficulties in order to manage it when people are still in the

tunnel due to the destratification that happen after the activation of the system.

The results concerning the cases applying solely the fire ventilation strategy, determined by a reduction of

the fresh air income and an increment of the extraction, localized nearby the fire location (resulting therefore

in a semi-transversal ventilation) are good in terms of limiting and slowing down the spread of smoke along

the tunnel.

The results concerning fire operation ventilation strategies with low pressure water mist system or some

high pressure water mist nozzles on the wall are similar to previous ones. But, the

positioning of some

nozzles close to tunnel bottom and with an horizontal discharge can improve the safety condition

in the area

occupied by people and vehicles.

In general, the solution with fire operation ventilation strategies with

5 exhaust vents instead of 4

(operating

at the same total volume flow rate)

gives better results

. Besides, if the reduction of fresh air income is not

realized, but the normal intake flow rate is continued when the fire ventilation strategy is activated, results

show a modest improvement of safety condition in the area of the fire without a sensible decrease beyond

the protected area. The

smoke destratification is relevant in every cases, in particular beyond the extraction

zone, but the water mist discharge has determined a more and quicker destratification of the combustion

products

. It has to be underlined that the

uncertainty associated to the results about water mist strategies is

very high

, because of the lack of data available (about dimensional and velocities distribution of the

droplets, and the complex phenomena related to spray behavior). Therefore, there are strong limitations in

the modeling approach adopted by the codes that have been used. Although CO and HCN concentrations are

relevant in determining the safety condition for people during the egress, because of the inherent criticality

of the combustion models, this

work has not focused on the quantification of their production and

consequent distribution along the tunnel

.

(7)

13

Milano, 29 ottobre 2009

Appendix

14

XII Convegno nazionale AIIA "I Sistemi di Gestione della Sicurezza Antincendio nella Fire Safety Engineering” Milano, 29 ottobre 2009

Fire scenarios

FIT

(8)

15

Milano, 29 ottobre 2009

Fire scenarios

• Indicazioni internazionali

FIT

Effectiveness assessment of road tunnel fire-fighting strategies by ventilation and water mist systems

Fire scenarios

(9)

17

Milano, 29 ottobre 2009

Fire scenarios

FIT

18

XII Convegno nazionale AIIA "I Sistemi di Gestione della Sicurezza Antincendio nella Fire Safety Engineering” Milano, 29 ottobre 2009

Fire scenarios

SFPE

NFPA 200

MW

Memorial tunnel: HRR= 50

÷

100 MW

UPTUN: HRR= 25 MW

(10)

19

Milano, 29 ottobre 2009

Numerical tools

•

proSES

(vs. 1.0.6, developed by Politecnico di Milano

-Department of Energy - and owned by Metropolitana Milanese

S.p.A.);

•

SES

(Subway Environment Simulation vs. 4.1, one-dimensional

tool developed by the Department of Transportation of United

States of America);

•

ECART

(vs. 4W0F, lumped parameter tool, a property of ERSE);

•

FDS

(Fire Dynamics Simulator vs. 5.2, CFD tool developed at

Building and Fire Research Laboratory of NIST).

Effectiveness assessment of road tunnel fire-fighting strategies by ventilation and water mist systems

SES + proSES

SES

The SES code is based on a one-dimensional, incompressible, turbulent, slug-flow model developed with a

designer-oriented approach by the U.S. Department of Transportation and provide estimates of air flows,

temperatures, and humidity, as well as air conditioning requirements, for multiple-track subway systems. It

provides a dynamic simulation of different arrangement of tunnels, ventilation shafts, fan shafts, and jet fans,

plus non-steady-state heat sources used to model the fire with a prescribed heat release rate. It has been

validated in model tests and in actual practice and it is applicable to a variety of subway operating and design

configurations and has been demonstrated to be a cost-effective tool for evaluating the performance of most

types of environmental control strategies. The SES code was applied with the complementary tool proSES.

proSES

The proSES is a complementary tool to the SES code

developed by the authors of the present work and owned by

Metropolitana Milanese S.p.A., that help the user to easy

setup, modify and run multiple cases and to generate plots

and animations to rapidly compare and show results of SES

simulations. An interactive graph allows a visual

representation of the network elements (tunnels, stations,

vent shafts, ...) and their relative connections through

junctions. Nodes and segments are then described using

prototypes, so their properties values can be easily modified

in any time, simplifying the generation of different inputs

for parametric studies. The tool uses Python as a glue

programming language and for post-processing the VTK

library (

http://www.vtk.org/

) was chosen, so the

post-processor Paraview (

http://www.paraview.org/

) could be

used to visualize the dataset storing the results to perform

an easily comparation of a great number of cases.

(11)

21

Milano, 29 ottobre 2009

ECART

The computer tool ECART is dedicated to predict the consequences of an accident in a risk installation. It was

originally created to calculate the concentration of airborne radio toxic substances inside nuclear power plants in

the case of a severe accident. As it is not related to a specific design, nuclear or not, it can simulate the airborne

transport of dangerous substances throughout a generic system of rooms, pipes or plant components, together

with the removal and the re-entrainment mechanisms which may occur in the presence of structures, liquid

sumps or water sprays.

The problem of integral analysis of complex systems or industrial installations is still quite difficult using CFD

codes that are suitable for the detection of fluid motion field, but are too heavy to run detailed simulations of fire

propagation through several rooms, corridors or tunnels, accounting the flame and smoke propagation, or the

chemistry of combustion together with the thermodynamic response of atmosphere and structures. ECART is a

lump parameter model for aerosol transport with the capability of modeling water mist injectors and fire

sources. It adopts the classic well-mixed hypothesis to describe the transport within each control volume

interconnected with others via explicit junction (prescribing the mass flow rate) and implicit junction

(prescribing equivalent hydraulic diameter and head losses).

This computer tool is currently developed by ERSE, Milan, Italy (

flavio.parozzi@erse-web.it

) and has been also

supported by Italian Government, national agency ENEA, University of Pisa, Politecnico di Milano and Torino,

French EDF and the European Union. Significant efforts were spent to extend its modeling to fire

phenomenology in the recent years in cooperation with the Department of Energy of Politecnico di Milano (ref:

giovanni.manzini@polimi.it; luca.iannantuoni@polimi.it

), paying particular attention to pool fires modeling

and water mist phenomenology.

This work has been financed by the Research Fund for the Italian Electrical System under the Contract

Agreement between CESI RICERCA and the Ministry of Economic Development--General Directorate for

Energy and Mining Resources stipulated on June 21, 2007 in compliance with the Decree n.73 of June 18, 2007.

22

XII Convegno nazionale AIIA "I Sistemi di Gestione della Sicurezza Antincendio nella Fire Safety Engineering” Milano, 29 ottobre 2009

FDS

FDS (developed by BFRL - NIST) is a Computational Fluid Dynamics (CFD) model of fire-driven fluid flow. The

model solves numerically a form of the Navier-Stokes equations appropriate for low-speed, thermally-driven

flow with an emphasis on smoke and heat transport from fires. The code is well known and has been validated

in tunnel fire scenarios, especially with the Memorial Tunnel Fir Tests simulating mechanical ventilation effect

on smoke movement.