FESG provides explanation on Stavanger car park fire

Stavanger airport car park fire

The fire that occured on 7/01/2020 in the airport car park in Stavanger is raising a lot of questions.

What do we know?

  • The fire is reported to have started in an Opel (diesel) on the ground floor.
  • The fire quickly spread to multiple cars.
  • Parts of the building have collapsed due to the high temperatures.
  • At least 300 vehicles are said to have been destroyed.
  • The fire brigade had to contain the fire from the outside, as the possibilities for internal firefighting were limited due to the quick fire spread and low structural stability (15 minutes).
  • There are no indications that the presence of a large amount of electric cars had a significant impact (compared to combustion engine)

Recent fires in open car parks

  • When looking at fires in open car parks in the last 2 years, this event is not a surprise or an incident that is specific for Norway. Below are listed some of the similar fires.



Specific features of fire behaviour Photo

Liver pool, UK

Dec ‘17

  • t=0: fire starts in range rover (3rd floor) due to electric defect.
  • t=21 min: Fire brigade arrived
  • t=30 min: 8 to 10 cars on fire
  • t=39 min: 18-20 cars (HRR of 200 MW), fire brigade attempts internal attack, radiation to intense to reach fire source or fight the fire (low water pressure).
  • t=114 min: spalling of concrete
  • Eventually 1400 cars destroyed
Brooklyn, USA Sept ‘18
  • Kings Plaza is an open parking next to a shopping center.
  • The fire spread from 2nd to 3rd floor.
  • 17 of 300 fire fighters got injured.
  • 120 cars were destroyed
  • The smoke spread tot he adjacent buildings lead to evacuation of buildings.

New Jersey, USA

Jan  ‘19
  • Newark Liberty airport, rooftop fire with approximately 18 -20 cars involved in the fire at the moment of fire fighting attacking the fire.
  • The fire started due to an electric defect in the car.
  • Difficulties to reach the fire due to high radiation and limited water.
Manchester airport May ‘19
  • Terminal 2: fire on roof floor of open car park.

Cork, UK /

Sept ‘19
  • Origin unknown, fire started in car on 1st floor
  • 20 min: alerting fire brigade
  • Fire under control after 3 hours.
  • Eventually 60 cars destroyed/burnt.

Important parameters to have played a role in open car park fires

When trying to understand the important parameters that play a role in such events, we have to differentiate between 2 type of parameters: the fire behavior and the fire brigade intervention.

For the fire behavior:

  • it seems the most common cause is an electric defect in a car. It could be expected that the older the car, the higher the probability of having electric defects.
  • How long it takes from ignition of the fire to a fully developed fire can be significantly different from case to case (dependent on the type of car, type of windows, ventilation conditions, etc).
  • Once a car is fully burning, it is only a matter of minutes before the adjacent car will start pyrolising due to the intense radiation. This process repeats itself to ignite more adjacent cars.
  • Once a certain number of cars is burning, it is impossible to make an intervention in the car park due to radiation, independently of the structural stability.
  • Dependent on the provided structural stability, the risk of collapse can be expected after 15 to 120 minutes.

For the fire brigade intervention it is important to understand:

  • The later the fire brigade arrives on site, the larger the fire will be and the more water will be required to control the fire. The probability of a succesfull intervention is dependent on the timeline. In order to objectify this, the failure probability of a succesfull intervention can based on the Fire Brigade intervention Model.

                                    Figure 4 Fire Brigade Intervention Model (AFAC) [1]

Impact of fire prevention measures

After severe accidents, it is not uncommon to question the prescriptive fire codes. Typically questions as “why are there no sprinklers in these open car parks” arise and it is important to analyse this in an objective manner. Following questions need to be answered:

  • Which safety measures increase the probability of a succesfull fire brigade intervention and by how much.
  • How much is the end user willing to pay (e.g. parking fee) in order to improve the safety by providing these extra systems?
  • Is the installation cost proportional to the risk-reduction? What is the most cost-effective safety system?
  • How to determine a maximum investment line for both newly built as existing car parks as a risk-informed parameter, defendable from a cost-benefit point of view.
  • How to compare the proposition in the Fire Code and rules of good practice to an implicit acceptable fire safety level from the point of view of the authorities.
  • How to propose project specific performance criteria in terms of failure probability of a succesfull intervention as a function of the parameters of the project.
  • How to select the most cost-effective solution leading to an acceptable safety level.

Tools for the fire safety engineer

The fire brigade intervention model (FBIM) developed by Fire Engineered Solutions Ghent is a quantitative risk assessment tool capable to predict the risk level for a carpark based on its features. When a fire occurs it can become uncontrollable if the HRR exceed certain critical values and the firefighters cannot control it. The time for the fire to become critical and the intervention time of the firefighters depend on the features of the carpark such: occupancy, type of cars, geometry, location and distance from the fire house. On these input data there is often uncertainty and it is not possible to predict correctly the type of car which is going to burn first or the arrival time of the firetruck. Therefore different fire and intervention scenarios are evaluated in order to estimate the risk level. The final result are two distributions of time, one for the fire development and the other for the fire fighters intervention. By comparing both, the residual risk level is determined that the fire brigade will be not able to intervene before untenable conditions are reached.

The time required by the fire to develop and become uncontrollable is function of the fireload inside the carpark and of the safety measures installed. The evolution of the HRR is estimated based on the fire development in a vehicle and the spread among them. This approach was proposed by Tohir [2] who showed that the fire spread between vehicles can be estimated calculating the radiative heat flux emitted by a burning car on the adjacent cars. The HRR of each car is estimated based on statistical data coming from different sources. The ignition of another car depends on the distance between them and the radiative flux produced by the car. This approach has been validated by comparing experimental results with those predicted by the model.

By extending this approach to an entire carpark and using different fire source locations it is possible to evaluate a distribution of HRR which depends on the type of vehicles and the geometry of the carpark. Once the HRR evolution is estimated it is necessary to evaluate when does it become critical or uncontrollable. This is strongly related to the safety measures that are installed in the carpark: none, water suppression (light or full sprinklers [3]), mechanical ventilation, water suppression and mechanical ventilation. The different safety measures are designed in order to guarantee safe conditions in the carpark and they are designed with a specific design fire (currently [4] is included). However, as seen in figure the HRR can grow beyond the design value and the different systems can fail due to lack of maintenance or defects. When a carpark is not equipped with any safety measures then the limit HRR is defined as the one that can be safely handled by the firefighters or the structural stability. The probability of fire spread is severely reduced by a water suppression system (during the research project [3] more than 60 full scale tests looked at the efficiency of different water application rates and fire spread to adjacent cars can be avoided by limited amount of water if applied in the early stages). The light or full sprinklers can avoid fire spread among vehicles. A crucial aspect of these systems has shown to be the reliability of these systems (the simpler, the better). The time under which a mechanical ventilation system will allow a safe intervention of the firefighters can be determined. In case the fire grows too big the smoke can spread beyond the accepted limit. Based on the smoke spread (backlayering length) calculated for different HRR (according to [5]) it is possible to evaluate a critical HRR for which the mechanical ventilation system is still functioning as intended. By combining the results of the fire development with the limit HRR for different safety measures it is possible to convert the HRR distribution in a time distribution. This represent for a specific carpark the time when fire becomes uncontrollable.

This time is compared with the intervention time of the firefighters starting from the detection of the fire until fighters reach the burning cars. The intervention time is calculated as sum of the following components: Detection time, notification time, response time, travel time, setup on site and attack time.

These are estimated based on statistical data, therefore each of these times is described by a distribution. The detection time is calculated based on the distance between the detector and the fire and its failure probability. The notification time is the time between the detection and the communication to the firefighters which can be manual or automatic. The response time is the time between the notification until the trucks leave the firehouse and it is estimated based on empirical data coming from the Belgian firefighters [6]. The travel time is estimated based on statistical data coming from the Belgian government [6] and on the distance between the firehose and the carpark. The setup time depends on the firefighters and it is estimated based on statistical data. The attack time depends on the location of the burning car and the geometry of the carpark. When all these times are summed up together they provide a time distribution which can be compared to the one of the fire development. For each car it is possible to evaluate the critical time for the fire development and the intervention time estimating the final risk level for the carpark.

The current approach allows to evaluate the level of risk for different carparks equipped with different safety measures. This approach goes beyond the prescriptive approach because it evaluates the effective risk for different type of occupancies in the carpark. The possibility to use different safety measures allows to evaluate the risk of different fire safety strategies and to compare both the installation costs and their safety level. Since the described approach takes also into account reliability of the different systems, boundary conditions on the required maintenance intervals can be provided in order to guarantee the required reliability of the safety systems.

This latest research is an extention of the knowledge obtained by previous research on car parks [7] where involved tests set-up of smoke control systems [8] and how CFD-models can be used to correctly predict the smoke spread and temperature evolution [9]. The addition of new data on sprinkler efficiency, the probabilistic approach and combining everything in the FBIM methodology, allow for the FBIM-PARK tool [10] to objectify the risk for succesfull fire brigade intervention. By combining this tool with a cost estimation of the safety systems, it is possible to determine what is the most cost-effective safety system, both for newly built as existing car parks.


The method is currently examined for being applicable in other European countries and comparing it to several challenging case studies. Some of the main aspects will be to feed the important submodels of fire brigade intervention with the country-specific date and continue the verification and validation process of the new methodology.


The author would like to acknowledge the Flemish government VLAIO for the funding of this research through project number HBC3028.0331.

[1] L., Foster.. Fire Brigade Intervention Model V2.2. Australian Fire Authorities Council (AFAC). (2004)

[2] M., Tohir, and M. Spearpoint. "Development of fire scenarios for car parking buildings using risk analysis." Fire Safety Science 11 (2014): 944-957.

[3] FESG O&O report: General overview of the fire tests on sprinkler cars in car parks. Annex 4: impact of water density on sprinkler efficiency, august 2019.

[4] NBN S 21-208-2 : 2014 Fire protection in buildings - Design of smoke and heat exhaust ventilation systems (SHEV) for enclosed car park 2014

[5] N., Tilley, X., Deckers, and B., Merci. "CFD study of relation between ventilation velocity and smoke backlayering distance in large closed car parks." Fire Safety Journal 48 (2012): 11-20.

[6] KCCE. Statistieken – Belgische brandweer 2015. Federale overheidsdienst Binnenlandse zaken, Federaal Kenniscentrum voor de Civiele Veiligheid. (2015)

[7] B. Merci, L. Taerwe, P. Vandevelde, E. Van den Bulck, J. Van den Schoor, v. Beeck and Vantomme, “Fundamental design approaches for improvement of the fire safety in car,” IWT SBO project 080010.

[8] X. Deckers, B. Sette, B. Merci and J. S. Haga, “Smoke control in case of fire in a large car park: Full-scale experiments,” Fire Safety Journal , p. 57:11–21, 2013.

[9] X. Deckers, S. J. Haga, N. Tilley and B. Merci, “Smoke control in case of fire in a large car park: CFD simulations of full-scale configurations,” 2013, p. 57:22–34, Fire Safety Journal .

[10] M. Pachera, S. Bigdeli, “FBIM-PARK: Probabilistic Fire Brigade Intervention Model for Car Parks” revision 02 july 2019, Ghent, Belgium (https://www.fesg.be/en/news accessed on 10/01/2020).


Andere nieuwsberichten