See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/311923753

Assessing the fires on the deck of cable stayed bridges

Conference Paper · June 2016

CITATIONS READS
2 289

4 authors, including:

Panagiotis Kotsovinos Graeme Flint

Arup Arup
24 PUBLICATIONS   87 CITATIONS    12 PUBLICATIONS   132 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Bridge fire safety design View project

Revision of PD7974-1 & PD7974-3 UK engineering standards View project

All content following this page was uploaded by Panagiotis Kotsovinos on 27 December 2016.

The user has requested enhancement of the downloaded file.

Kotsovinos, Panagiotis, Walker, Gary, Flint, Graeme and Lane, Barbara (2016). Assessing the fires on
the deck of cable stayed bridges. In: Proceedings of the Ninth International Conference on Structures in
Fire. Princeton, USA, 8-10 June 2016.

Assessing the fires on the deck of cable stayed bridges

Panagiotis Kotsovinos*, Gary Walker, Graeme Flint & Barbara Lane
Arup, UK

ABSTRACT

Bridges are community critical infrastructure and, therefore, the consequences

arising from extensive structural damage due to a fire on the bridge could be significant.
This paper presents the process for the characterisation of the design fires on a bridge
deck utilised to assess the thermal and mechanical response of a cable-stayed bridge
structure. Future research requirements are also identified.

1 INTRODUCTION/KNOWLEDGE GAP

Bridges are community critical infrastructure and, therefore, the consequences

arising from extensive structural damage due to a fire on the bridge could be significant.
However, the fire resistance requirements for civil structures such as bridges are not
explicitly covered by national building regulations or, typically, by specific
owner/operator requirements and would need to be assessed from first principles.
A limited number of studies have been carried out on the assessment of cable stayed
bridges from fires on their deck [1, 2, 3]. The majority of previous research is based
primarily on steel girder bridges [4, 5, 6] as a response to the MacArthur Maze bridge
collapse in Oakland. However, a number of incidents involving fires in cable stayed
bridges have occurred in the past such as the ones on the Mezcala and Rio-Antirio
bridges that resulted in failure of a stay cable although fortunately without collapse.
This paper will present the first principles approach that was adopted by Arup as
part of a recent commercial cable stayed bridge project in the UK.

2 ROBUSTNESS ASSESSMENT PROCESS

Cable stayed bridges typically are formed with large spans (hundreds of meters) and
rely on complex load paths to support the bridge structure over areas where piers cannot
be provided. The robustness of a cable stayed bridge under fire conditions can be
influenced by the type of deck (box girder, ladder deck, etc.) and the arrangement of the
steel stay cables (spacing, located on the edges or the middle of the deck, etc.). Cable
loss from a fire on the bridge deck could be important given that cable stayed bridges
are not typically designed for the potential loss of multiple cables.
In order to carry out an assessment of the robustness of a cable stayed bridge in fire
it is proposed to follow the process identified in Figure 1. Such an assessment needs to
satisfy the stated fire safety goals and the owner’s/operator’s requirements for property
protection and business continuity. This typically involves understanding potential fire
events on the deck, and in the surrounding area leading to and under the bridge. Figure
2 describes the number of steps that the assessment needs to follow.
This paper will concentrate on the part of the process for the identification of fire
hazards where and the resulting design fire characterization for a cable-stayed bridge
Kotsovinos, Panagiotis, Walker, Gary, Flint, Graeme and Lane, Barbara (2016). Assessing the fires on
the deck of cable stayed bridges. In: Proceedings of the Ninth International Conference on Structures in
Fire. Princeton, USA, 8-10 June 2016.

structure. This is a critical step in the design process since it affects the input to the
thermo-mechanical assessment of the bridge structure. Such as an assessment is
required so that the fire resistance requirements of bridges are determined.
Note that this paper only considers the potential fire scenarios on the deck. Any fires
below the deck or around the site are not presented here.

The fire safety goals are the applicable A fire engineered approach shall be
legislation and to meet the property adopted to determine the optimum
protection/operational continuity performance balance between the goals and the
criteria of the relevant stakeholders. constraints.

Goals

Constraints Solution

The constraints are the characteristics of the The solution are the structural arrangement and
fires that could occur on, below and around its requirements for fire protection and any
the bridge. potential restrictions or mitigation measures.

Figure 1. Project goals, constraints and solution of the design process

Kotsovinos, Panagiotis, Walker, Gary, Flint, Graeme and Lane, Barbara (2016). Assessing the fires on
the deck of cable stayed bridges. In: Proceedings of the Ninth International Conference on Structures in
Fire. Princeton, USA, 8-10 June 2016.

Identification of potential fire hazards.

Design fire characterisation of credible fire hazards

based on the desired reliability (see section 3)

Characterisation of the resulting thermal loads to

the adjacent structure (deck and cables).

Assessment of the thermo-mechanical response of

the structural elements (deck and cables).

Evaluation of the global structural response of the

bridge structure.

Figure 2. Structural Fire Assessment Process.

3 GOALS (RELIABILITY OF THE BRIDGE)

The required reliability for a bridge project is dependent on the fire safety and
property protection/business continuity goals. Depending on the size, use and location
of the bridge (and therefore if other alternatives exist in case of an accident) there may
be different reliability requirements.
The fire safety goals for the project regarding life safety of s users and staff and fire-
fighter life safety will be depended upon local legislation. Note that for critical
infrastructure it is likely that the property protection/business continuity goals will drive
the required reliability for the project.
To determine the property protection/business continuity goals of the project, the
following need to be discussed/agreed with the Owner/Operator of the project.
• Whether damage to the structure from a low likelihood fire scenario leading to
local/global collapse is acceptable.
• An acceptable probability of occurrence of a fire scenario
• Specific business continuity requirements (acceptable level of
disruption/downtime in case of an accident). This will impact the acceptable
damage to the structure requiring significant time and/or resources to repair.
• Willingness to provide restrictions on the traffic of the bridge e.g. HGVs and
petrol tankers to be restricted from using the bridge or to be escorted
Kotsovinos, Panagiotis, Walker, Gary, Flint, Graeme and Lane, Barbara (2016). Assessing the fires on
the deck of cable stayed bridges. In: Proceedings of the Ninth International Conference on Structures in
Fire. Princeton, USA, 8-10 June 2016.

• Willingness to provide an emergency response plan on the control of the traffic

after a fire accident (good management can limit the potential for fire spread
between vehicles).
• Whether fires originating due to arson/terrorism need to be considered (and
therefore fires potentially igniting in more than one location).
• Whether the bridge needs to be designed for multihazard actions such as fire
followed by explosion

4 CONSTRAINTS (POTENTIAL FIRE SCENARIOS ON THE DECK)

In order to determine the potential fire scenarios on the bridge deck the fire
consultant, in conjunction with the security consultant of the project carrying out a threat
and risk assessment, needs to review the following:
• A traffic study indicating also any future demands of the bridge
• The anticipated traffic speeds and density on the bridge
• Statistics relating to the occurrence of traffic accidents and occurrence of fires
in roads in the nearby area and in bridges in general.
• The anticipated types of vehicles crossing the bridge
• The anticipated contents that will be transported by the vehicles crossing the
bridge
• Whether suppressions systems and supplies/hydrants will be provided.
• Meeting with the local fire service to understand the capability of the local fire
crews to engage and suppress the identified fires and the proximity of the
nearby station to the bridge
• Crossfall of the deck and the deck drainage provided to the bridge and the
associated drainage rates
• The wind conditions in the area
• The potential for alternative vehicles such as CNG and LNG to cross the
bridge now or in the future
• Any heavy industry sites (DSEAR zones or COMAH zones) that would
require road transport to use the bridge.
• The potential for ammunition or explosives being transported over or under the
bridge for nearby military bases or ammunition factories
In addition to the above, in order to identify the potential fire hazards and their
severity, previous notable incidents on road bridges were reviewed [7]. In most of the
reported cases, the fire was triggered by an accident (collision, impact etc.) involving a
truck transporting flammable material such as gasoline or heating oil. Additionally, the
design fires used in tunnels were reviewed [8].
The following four fire scenarios have been identified for a commercial project in
the UK:
 a Heavy Goods Vehicle (HGV) fire;
 a petrol tanker fire arising from the early ignition of the fuel at the release
location following puncture of the tank containing the fuel;
 a pool fire on the deck of the bridge arising from the delayed ignition of the fuel
spilled from the location of the localised failure / puncture of the envelope of the
tank of a petrol tanker;
 Kotsovinos, Panagiotis, Walker, Gary, Flint, Graeme and Lane, Barbara (2016). Assessing the fires on
the deck of cable stayed bridges. In: Proceedings of the Ninth International Conference on Structures in
Fire. Princeton, USA, 8-10 June 2016.

 a tanker transporting Liquefied Natural Gas resulting in pool/jet fires.

Bridges are typically designed to sustain the loss of at least a single cable. As a
result, car and van fires, though the most likely to occur, are not considered to lead to
failure of more than one stay cable and, therefore, are not considered further.
The consequence of the potentially structurally significant fires need to be
considered for all possible fire scenarios in order to assess the risk to the structure (and
determine whether such a risk is acceptable to the owners/operators of the bridge).
All available experimental data of HGVs and petrol tankers are from a tunnel
environment. Experiments conducted to ‘replicate’ the environment in a tunnel are
unlikely to be representative of the condition for open air fires on the bridge deck due
to the enhanced radiative feedback in the enclosed geometry of a tunnel and the forced
ventilation applied in the experiments (typical of a tunnel ventilation / smoke extract
system) [9]. Therefore, it is anticipated that the peak heat release rate expected from a
fire on the deck of a bridge would be of lower severity than that of a comparative tunnel
fire. The fire severity of the potential fire scenarios has been assessed from first
principles.

4.1 Heavy Goods Vehicle (HGV)

Table A.11.4.1 of NFPA 502 [10] notes experimental data indicating that the peak
heat release rate varies from 20MW to 200MW for vehicles in tunnels.
Recent incidents involving HGVs in the UK are summarized below:
 A Co-op delivery lorry fire on the A9 on the 13th July 2012. The content of
the lorry was not reported. Photographs taken at the scene (see Figure 3a)
suggest persistent flame heights of up to 10m observed at the scene.
 A lorry fire on the M1 on the 19th June 2014. The reported cargo was
mattresses. Photographic evidence (see Figure 3b) indicates flames reaching
heights of up to 10m from ground surface.
 A lorry loaded with plastics caught fire on the M20 on the 29th November
2012. Photographic evidence (see Figure 3c) suggests flame heights of up
to 12-15m from the road surface.

a) b) c)

Figure 3. Previous incidents of HGV fires in the UK

It is assumed that an HGV fire on the bridge deck will be limited by the geometry
and nature of the fuel that the vehicle is transporting, i.e. that it is the transported goods
that provide the bulk of the combustible fuel for the fire. Experimental measurements
of heat release rate per unit area (at various incident heat flux) for a range of plastic and
wood materials are provided in Babrauskas and Grayson [12] and therefore the mean
heat release rate taken can be determined.
Kotsovinos, Panagiotis, Walker, Gary, Flint, Graeme and Lane, Barbara (2016). Assessing the fires on
the deck of cable stayed bridges. In: Proceedings of the Ninth International Conference on Structures in
Fire. Princeton, USA, 8-10 June 2016.

The dimensions of a HGV vary and, therefore, so does the quantity of combustible
material transported. For the purpose of this assessment, the maximum permissible
dimensions of a HGV in the UK were taken. The fire is assumed to be restricted to the
five sides of the trailer. The maximum heat release area can be determined based on the
HRRUA of a material type and by considering all 5 sides of the trailer.
The assessment presented, demonstrates that onerous fire heat release rates are
185MW for polypropylene, 130MW for polyethylene and 95MW for polyurethane
based on the maximum dimensions of a HGV in the UK.
Based on the above and the peak heat release rate data for tunnels presented, it is
considered that a 150MW (corresponding to a heat release rate of 1130kW/m2) fire is
representative of a very low likelihood HGV fire event. In addition, for the majority of
the cases where the fire size would be around 150MW, the weight of the transported
material would be higher than the maximum allowed in the UK which is 44 tonnes.
A 100MW fire might be described by the following.
 A fully developed fire having a heat release rate of 755kW/m2 on the five
sides (excluding the base) of the HGV trailer identified in Appendix A1 (i.e.
13.6m x 2.6m x 3.0m (L x W x H)). Such a fire emphasises the length over
which the fire base is distributed.
 A higher release rate (1350kW/m2) fire distributed over a smaller area of a
non-articulated lorry (i.e. four sides, excluding the base and the side adjacent
to the cab, of 8.9m x 2.6m x 2.5m (L x W x H)). Such a fire emphasises the
fire intensity and flame height (though it should be noted that this will be
affected by the wind).
The anticipated duration of steady-state burning of a 100MW fire is calculated to be
approximately 4hours. The thermal load incident on the structural cables should be
determined both in the absence and presence of wind. The location of the fire, plus the
magnitude and direction of the wind (with reference to the local wind rose) should be
selected to maximise the exposure of the structural cables to the thermal load from the
fire.

4.2 Petrol Tankers

A peak heat release rate in excess of 200MW was reported in Ingason [8] for a petrol
tanker in a tunnel. As noted before, experiments conducted to ‘replicate’ the
environment in a tunnel are unlikely to be representative of the condition for open air
fires on the bridge deck.

4.2.1 PUNCTURE – IGNITION – RUPTURE

The scenario assumes that rupture occurs along the entire length of the tank and
close to its top. The burning takes place at the subsequent rupture location. Such a
scenario was experienced in 2011, when a petrol tanker caught fire under the Paramount
Boulevard bridge, Montabello, CA, USA (Figure 4).
For the purposes of this assessment it was assumed that the area of the tanker is
11.7m x 2.5m and the height of the fuel is 2.5m (dimensions selected such as the fire
area is maximised). An assessment in accordance to [13] suggests that:
a) the heat release rate is approximately 70MW;
Kotsovinos, Panagiotis, Walker, Gary, Flint, Graeme and Lane, Barbara (2016). Assessing the fires on
the deck of cable stayed bridges. In: Proceedings of the Ninth International Conference on Structures in
Fire. Princeton, USA, 8-10 June 2016.

b) the duration of a steady-state fire is approximately 5hours;

c) the visible flame height is approximately 17.8m (measured from the elevated
base in the absence of wind).

Figure 4. Petrol Tanker Fire at Paramount Boulevard Bridge, Montabello, CA, USA, 2011.

4.2.2 FUEL SPILL – POOL FIRE

The scenario considers a localised failure / puncture of the envelope of the tank
leading to sustained release of fuel onto the deck of the bridge. Subsequent ignition of
the fuel leads to a pool fire.
The spill rate (of fuel from the tanker) is a function of the tanker geometry and the
hole size and hole location. Petrol tankers vary in capacity (and, therefore, dimension).
The maximum capacity of a petrol tanker in the UK is 38m3. It is assumed that the
tanker is not pressurised. To maximise the spill rate, the height of the free surface of
the fuel is taken to be 2.0m initially (and, the cross-section of the tank is adjusted to
ensure that the maximum initial volume of the fuel is 38m3).
The hole is assumed to develop as a result of a failure of the supply / discharge valve
during the initiating event. A review of typical valve sizes has indicated that the
diameter of the valve could vary from 80mm to 100mm. The hole is assumed to be
located at the bottom of the tanker.
Bridge decks typically have a gradient through the width of the deck (crossfall) and
are provided with drainage units (based on the expected rainfall). The deck drainage
rate is a function of the downpipe diameter and the lateral (across deck) gradient. For a
gradient, that the drainage rate exceeds the maximum spill rate it is unlikely that a non-
instantaneous spillage will lead to a significant pool of flammable fuel.
Assuming a low lateral gradient and a drainage rate of 0.018m3/s, taking the mean
volumetric spill rate during which it exceeds the drainage rate, the steady-state area of
the ignited fuel spill is estimated [13] to be 80m2. The heat release rate is estimated to
be 190MW. The duration of the fire should not exceed 30 minutes.
It should be noted that no assessment of the likelihood of the diameter of the hole
(leading to the loss of fuel) has been undertaken. The assessment indicates that a hole
size less than 80mm diameter is unlikely to provide a significant pool fire sufficient to
be a direct threat to the structural cables. (Though it could lead to a Boiling Liquid
Expanding Vapor Explosion also known as BLEVE.)
Kotsovinos, Panagiotis, Walker, Gary, Flint, Graeme and Lane, Barbara (2016). Assessing the fires on
the deck of cable stayed bridges. In: Proceedings of the Ninth International Conference on Structures in
Fire. Princeton, USA, 8-10 June 2016.

Figure 5 Spill from a petrol tanker

4.3 LNG VEHICLES

In some countries such as the UK the transport of LNG through road tankers is
allowed (other countries restrict this transportation). LNG tankers transport Liquefied
Natural Gas over long distances where pipelines do not exist.
One of the threats of vehicles transporting LNG is fire. Fires resulting from LNG
tankers are not frequent but have occurred in the past (in 2002 there was the first incident
involving a LNG tanker that resulted in a fire). Some of these fires have led also to an
explosion (see Figure 6). A LNG tanker fire specifically on a bridge structure has not
been reported in the past. However, where LNG terminal operators have facilities
around the bridge, such a scenario needs to be considered.

Figure 6 LNG Tanker Fire on the A-91 motorway in Spain (2002) before an explosion occurred after
20-25 minutes of burning.

An LNG tanker fire could result in the following type of events:

 Pool fire
 Jet fire producing a long/narrow turbulent jet
 Flash fire where the flame front moves through vapour cloud
 BLEVE - Note that BLEVEs are currently outside the scope of this report
The exposure of a flash fire normally lasts no more than a few seconds. After a flash
fire burns back to the LNG pool, or if ignition begins at the pool, the result could be a
pool fire. Based on the data available in SFPE Handbook [13], LNG has a higher mass
burning rate compared to gasoline and therefore a higher heat release rate could be
expected for the same amount of fuel.
To characterise an LNG fire, CFD and non-CFD (e.g. PHAST) methodologies can
be used. A single peak heat release rate of the LNG fire scenario as for the rest of the
proposed design fires cannot be determined. The type and geometry of fire and the fire
duration are dependent on the release rate from the vehicle following the rupture of the
tank. Assuming a catastrophic tank failure (i.e. a high release rate) would lead to an
intense but short duration fire and therefore it would not necessarily represent a
conservative thermal exposure for a protected structure. A number of release rates need
to be examined to determine the LNG fire scenario that would result in the most onerous
thermal load to the structure.
 Kotsovinos, Panagiotis, Walker, Gary, Flint, Graeme and Lane, Barbara (2016). Assessing the fires on
the deck of cable stayed bridges. In: Proceedings of the Ninth International Conference on Structures in
Fire. Princeton, USA, 8-10 June 2016.

5 CONCLUSIONS AND NEXT STEPS

This paper presents a framework for assessing the robustness of cable stayed bridges
from fires in their deck. The potential fire scenarios on a bridge deck that could affect
the thermal and structural response of the bridge have been determined.
The following areas of future research have been identified:
 Experimental data on open air fires of HGVs and petrol tankers are required.
 A probabilistic framework that relates the likelihood of a fire and its
consequence (i.e. structural damage) is needed. This could allow for the
project stakeholders to define their desired level of reliability/resilience
depending on their specific property protection and business continuity
criteria.

8 REFERENCES
1. Bennetts, I., & Moinuddin, K. (2009). Evaluation of the impact of potential fire scenarios on
structural elements of a cable-stayed bridge. Journal of Fire Protection Engineering, 19(2), 85-106.
2. Quiel, S., Yokoyama, T., Mueller, K., Bregman, L., Marjanishvili S. (2015). Mitigating the effects of
a tanker truck fire on a cable-stayed bridge, 2nd International Conference on Performance-based and
Lifecycle Structural Engineering, Queensland, Australia.
3. Woodworth, M., Hansen, E., McArthur, C., and Abboud, N. (2015) Protection of Cable-Stay Bridges
from Accidental and Man-Made Fire Hazards: A Rational Physics-Based Approach to Analyzing
Vulnerabilities and Mitigations. Structures Congress 2015: pp. 24-37.
4. Aziz, E. M., Kodur, V. K., Glassman, J. D., & Garlock, M. E. M. (2015). Behavior of steel bridge
girders under fire conditions. Journal of Constructional Steel Research, 106, 11-22.
5. Alos-Moya, J., Paya-Zaforteza, I., Garlock, M. E. M., Loma-Ossorio, E., Schiffner, D., &
Hospitaler, A. (2014). Analysis of a bridge failure due to fire using computational fluid dynamics and
finite element models. Engineering Structures, 68, 96-110.
6. Quiel, S., Yokoyama, T., Bregman, L., Mueller, K., Marjanishvili, S. (2015). “A streamlined
framework and case study for calculating the response of bridges to open-air tanker truck fires.” Fire
Safety Journal, 73, 63-75.
7. Garlock, M., Paya-Zaforteza, I., Kodur, V., & Gu, L. (2012). Fire hazard in bridges: Review,
assessment and repair strategies. Engineering Structures, 35, 89-98.
8. Ingason, H. (2006). “Design fires in tunnels,” in Safe and Reliable Tunnels, Innovative European
Achievements, Lausanne, Switzerland.
9. Carvel, R, Beard, A., Jowitt P., & Drysdale, D. (2004). The influence of tunnel geometry and
ventilation on the heat release rate of a fire,” Fire Technology, vol. 40, pp. 5-26.
10. NFPA 502, (2014) Standard for Road Tunnels, Bridges, and Other Limited Access Highways.
12. Babrauskas V., & Grayson, S. (1992) Heat Release In Fires, Taylor & Francis.
13. SFPE (2002). Handbook of Fire Protection Engineering, Third Edition, NFPA.

View publication stats