Monday, December 3, 2012

Feyzin Disaster: An illustration about BLEVE dangers


View of the Feyzin's spheres after the BLEVE
On the morning of 4th January 1966, in the normal village of Feyzin, situated 10 km away from Lyon, the normal villagers witnessed something unusual, something that had never happened before: an unceasing waltz of sirens, following by a thunderous sound of explosions and a view of a great fire in the premises of the refinery present in the village. Yes, this meant that something really bad happened in there.

Meet the Feyzin refinery

At that time, the Feyzin refinery had been operating for 2 years, employed around 250 workers and its capacity was about two million tons of crude oil per year. Till date, this refinery has been one of the main drivers of the local economy, giving a substantial income for the community (in the form of taxes) and employment for the villagers.
The location of the refining units, and more particularly the LPG storage tanks, where the disaster took place is, is quite intriguing: the sphere tanks containing propane and butane were situated around 400 meters away from the village and close to two important roads: the A7 motorway linking Lyon and Marseille and a local road. As we are going to see later in this article, this fact will play a key role in the chain of events that took place in this accident.

Small operations, big liabilities

On that day, a sampling operation had to be undertaken in one of the propane storage sphere tanks for quality control. Even though this may sound like a piece of cake, this operation follows a strict procedure due to the fact that at propane and ice formation could freeze the sampling valves present in the sphere.
However, this procedure would not be respected by the operator on that day, and after a mistake in the valve opening, a huge amount of propane gushed out from the sampling line. It was the beginning of the end, since the operators wouldn’t be able to close the valve and stop the leakage anymore.
In fifteen minutes, a very inflammable propane cloud was formed and spread out in all directions, including the roads present alongside the storage zone. If the A7 autoroute could be closed on time, which was not the case: a car managed to enter through the zone affected by the gas cloud. And after contact with a hot point present in the car, the gas cloud ignited and transformed into a sea of flames capable of putting hell in shame. It is needless to say that the poor driver was killed in the process but the worst was yet to come, as the fire travelled back to the sphere in a matter of minutes. This was the perfect moment for an unknown yet spectacularly dangerous phenomenon to happen: behold, the Boiling Liquid Vapor Expansion Explosion, aka BLEVE.

Hi, my name is BLEVE, nice to meet ya !

Sphere tanks during the BLEVE
As the fire travelled back, it ignited the escaping propane underneath and around the sphere, resulting in flames of heights up to 60 m. Since the sphere was exposed to sustained heat, the propane stored inside the spherical tank (in the liquid form) was forced to vaporized, thus increasing the pressure inside the tank. Depressurizing valves present in the tank were then activated in order to relief the tank from this increasing of pressure.

Generally, when the level of liquid present in the tank is higher than the height of the flames, the liquid is capable of absorbing the heat produced by the flames, thus protecting the structure of the sphere. However, the ongoing vaporization and the reopening of the depressurizing valves when the pressure had built again inside the tank decreased the level inside the tank. The metal was no more protected and its temperature increased with the contact of the flames. The metal, under effect of temperature and high pressure inside the tank, tore apart and released a huge quantity of liquid and vapor propane, thus causing a huge explosion, launching missiles composed of huge chunks of metal hundreds of meters away. The immense release of heat was capable of literally atomizing the people close to the sphere, leaving only the carbon shadow of a body as a memento of people who were relatively away from the explosion.


Spheres after a BLEVE
 A fragment of the spherical tank knocked the support of another sphere tank, provoking extensive damage to its structure. The heatwave caused by the first explosion, combined with its weakened state made the second sphere BLEVE too. Fragments of the two tanks, severed pipeworks, led to releasing of more inflammable material, which gave more fuel to the fire. It took more than two days for the firefighters to cease the fire and prevent that other tanks to be destroyed. The blast was so strong that it broke some windows in a church located 3 kilometers away from the refinery.




For those who like some action, here is a video which shows what a BLEVE is on live.


 Lessons from the accident

This accident caused 18 deaths, 81 injured and an overall damage estimated at €7 million (refinery and outside damages). 11 storage tanks containing inflammable material (propane and butane) were destroyed and projectiles were found within more than 800 meters away from the initial locations of the vessels.
The Feyzin disaster is a good example about how underestimated (or undiagnosed) dangers can bring chaos in an industrial environment such as a refinery. Before this accident, BLEVE was poorly understood, thus pressurized tanks were not protected against this phenomenon. Even though many pressurized tanks have BLEVE’d since then, storage spheres are now protected against heat exposure by better design. After the accident, many changes in the French legislation were made in order to take into account this risk, such as better identification of BLEVE-susceptible zones and new rules on distances between industrial sites and cities and buildings such as houses and roads.
This accident also shows that there is no easy operation in industry. Even easy-looking operations like sampling need to follow procedures for the sake of safety of operators.

Sources :  
- The Feyzin Disaster - Case study, Loss Prevention Bulletin 077, IChemE, October 1987, available at http://www.icheme.org/~/media/Documents/icheme/Resources/LPB/LPB%20samples/FeyzinDisaster.ashx
- BLEVE dans un dépôt de GPL en raffinerie, Aria - Ministère chargé de l'environnement, fiche septembre 2006, available at http://www.aria.developpement-durable.gouv.fr/ressources/fd_1_feyzin_gc_fr.pdf