Home   Browse database   Free examples   Contact us   Login
Related links
308 Chemical accident in France with propane
Go back to the accident listing

You do not have a membership or you are not logged in
Become a member
General accident information
Class/Quality * * * * *
Year 1966
Summary Release of propane caused fire, explosions of
Storage tanks at refinery, bleve of lpg sphere,
Evacuation and 99 casualties
Country F
Activity STORAGE
Chemicals Butane, Caustic soda (solid), Lpg, Propane,
Propane, Propane, Propane hydrate, Sodium hydroxide (liquid)
Cause Management-failure
Fatalities/Injuries 18 / 81
Occurrences or events Bleve, Blow-away, Burst/Rupture, Burst/Rupture,
Clog, Cooling, Cooling, Crack,
Defective-working, Draining, Evacuation, Explosion,
Fire, Fire, Fire, Fire,
Fire fighting/Emergency response, Fire fighting/Emergency response, Fire fighting/Emergency response, Fire fighting/Emergency response,
Fire fighting/Emergency response, Fireball, Human-operations, Human-operations,
Human-operations, Ignition, Ignition, Management-failure,
Management-failure, Management-failure, Management-failure, Management-failure,
Management-failure, Management-failure, Management-failure, No-action,
Overheating, Physical-explosion, Release, Release,
Release, Release, Release, Remove,
Safety-measures, Safety-measures, Safety-measures, Sampling,
Solidify/Icing, Switch, Traffic-interruption, Vaporize,
Full accident information 
download accident report Click here to view a document containing the full accident information

Country: FRANCE (F)
Date : 1966 0104

Improper sampling procedures on a 13.8/m, 1200/m3 butane sphere which was 3/4 full resulted in the uncontrolled release of liquid butane through a 51/mm connection. Ignition was believed to have been from a passing vehicle at a distance of 125/m. Fire at the relief valves and the base of the sphere burned about an hour before the tank failed massively. Pieces of steel up to 100/tons travelled about 300/m. One fragment landed on a pipe rack, cutting 40 lines. Another piece cut the legs out from under an adjoining sphere causing it to topple and break its 202/mm connecting line.
Hot spots developed on the three nearby butane tanks, and they ruptured.

The refinery was put on stream in 1964; it was to process 1.7E+9/kg/year; it had had auxiliary installations which included mainly overhead tanks of liquid hydrocarbons of 300000/m3 capacity. These stores were in zone B of the refinery area. There were, among others, two spherical tanks which could hold 2000/m3 of butane and 1200/m3 of propane (the four spheres of propane were numbered from T.6l.440 to T.6l.443).

In order to avoid excessive internal pressure in the sphere in case of an accident or a conflagration each sphere is equipped with a safety device consisting of two valves installed in the upper part of the sphere, yielding 73E+3/kg/hour of gaseous products.
Each also contains a cooling device consisting of two rings of vaporisers installed at the top, the median part and the lower part of the tank respectively. This device is directly connected to the fire extinguishing network by a valve. Its average yield is 2.2m3/minute for the propane spheres.
The nature of liquefied hydrocarbons requires frequent draining during storage to eliminate the water and soda mixed with the product which after pouring off accumulate in the lower part of the tank.
The draining of the spheres is effected by means of two valves located at 50/mm distance from each other and are operated by a square key-lock, the lower valve serving as an evacuation pipe that drips into a square draining trap of 500/mm side-length and 1/meter depth, linked to the network for used water from the refinery.
On the other hand, gas samples are taken from time to time for analysis of the manufactured products and checking of their standards.

Pre-cursors of the accident
The draining of manufactured products practised from the start of storage on site (June 12, 1964) had brought to light some problems arising from the device:
The valves were too close to each other, as the passage of propane from liquid to gaseous state, which took place at a temperature of minus 44/C caused an almost simultaneous icing up of both valves.
Their control by removable keys (lever or valve spanner) rather than wheel key-lock presented risks of gas escape in case one of these keys being dropped.
Their diameter (51/mm) was too large.
With the draining trap located at the feet of the operator, which meant that he was frequently splashed and sometimes suffered burns to his face and hands by the gushing of liquid into this opening.
The valves were often difficult to operate.
Finally, the access to the valves was made difficult by the presence of pipes which the operators had to step over in order to carry out the draining.

Employees had told the management about these problems; things remained practically as they were. Two serious incidents had occurred that gave substance to the apprehension and fears expressed:
On August 6, 1964 at about 2300 hours, an operator's mate, had opened - before massively draining the butane sphere 462 - completely the two valves of the tank, letting the liquid flow normally into the drainage tank, and he had climbed on top of the sphere in order to check the gauge there, thinking he had enough time before finishing the draining operation. It was then that the gas shot out in force.
The operator's mate who wanted to shut the valves which were iced up by the passing gas froze his right hand slightly and had to be treated in hospital.
The draining taps were shut by a manufacturing engineer and one of the firemen on duty who were helped by a favourable wind.
On February 26, 1965 at 1105 hours a chemist, had been assigned with his colleague to carry out the draining of the propane sphere 440 to take a gas sample.
In these ill-defined circumstances, after the usual emission of water and soda the propane shot out and burned the two men. Two safety workers, after being alerted, intervened.
The first one was also burned but the second one managed to shut the valve. The alert had been serious.
This last incident which if the wind had not again been favourable could have developed into disaster even though the motorway had not yet been opened to traffic had subsequently caused the issue of a service bulletin on the method of draining the spheres (March 4, 1965) by the Chief of Technical Services. It said in particular that after the keys had been attached to the two draining valves the valve on the sphere side was to be opened completely, then the valve on the atmosphere side partly opened, without ever opening it completely in order to be sure that it could be closed, as soon as gas appeared, the closure of the draining valve or, in normal circumstances, of the valve on the sphere side, and then shut the second valve.
Additionally, this instruction indicated, for the control draining on the bottom of the sphere, the facility of using the piping between the two taps as a lock-chamber i.e. by opening the valve on the sphere side, shutting it again immediately, then opening the second valve to the atmosphere in order to empty out the content of the line.
It finally made it obligatory that the taking of laboratory samples had to be done in the presence of a safety officer and that draining was to be carried out by two people.
This bulletin which was entered into the service manual and posted in the pump rooms was generally known to the staff but had never been backed up by practice exercises. Also, some operators kept to their own ideas about the question and to the procedures previously practised.

The day of the accident it had been decided to clean propane sphere 433 at the end of sample taking.
Taking part were : a plant operator's mate, a safety officer (shift fireman) and a laboratory helper (labortorian technician).
In contravention of the instructions in the service bulletin from the Chief of Technical Services, this operation was carried out at 0640 hours, i.e. in complete darkness; the lower part of the sphere was lit by the diffused light of a candelabrum and horizontal projectors placed at a certain distance. The temperature was between 4-5/C, and there was virtually no wind.
Contrary to instructions the operator's mate first half-opened the lower valve, then fully opened the upper valve, as it emerges from the experts' statements on the pieces recovered as well as from those made by the safety officer. The latter whose function it was to watch the work and to intervene if need be did not budge but looked on from a distance. Some dirt ran into the drainage tank, then suddenly the gas shot out in force and struck the operator in the face and on the body.
The operator's mate, caught in the cloud, lost his safety goggles and involuntarily unhooked the operating key of the upper valve the fixing nut of which had actually not previously been tightened on the operating square. He fall backwards and pulled the valve handle partly of the valve.
The safety officer shouted : "You have opened it too wide." The operator's mate who had recovered slightly tried to shut the upper valve but did not succeed in putting the key back on because of the icing caused by the escape of gas. He forgot to try and close the lower valve on which the key was still fixed and refused to keep trying.
The safety officer (fireman) turned the water supply to the sprays fitted to the sphere.
Meanwhile the safety officer and the laboratory helper had raised the alert over the telephone and the "g??n??phone". Three safety officers tried in turn to stop the escape, without success.
Gas escaped from the sphere which at 0500 hours in the morning had held 693/m3 of propane at the rate of about 3.3/m3/second according to the calculations made by the experts. The gas mixture, being heavier than the air and there being hardly any wind blowing, the propane expanded by gravity in the direction of the motorway. Nobody thought of alerting the fire service, the gendarmerie and the CRS.
The cloud, approximately 1.5/meters high, reached the motorway on which there were a number of vehicles between 0655 hours and 0705 hours.
Employees from the refinery and from the guard of the factory then intervened on the motorway and on the CD 4 road to stop the traffic. Unfortunately, one minor road was not sealed off in time.
At 0715 hours, a man, driving his CV4 Renault, arrived on the scene; he was going to his work to take up his duties in a company working for the refinery. When he arrived at the cross-over linking the CD 4 with the motorway and crossed the gas cloud the latter, no doubt as a result of a spark produced by the vehicle, caught fire.
The panic-stricken driver stopped his car and got out; his clothes caught fire; he ran and threw himself into a ditch a few meters away. He was found a quarter of an hour later, severely burned, and taken to hospital where he died 4/days later.
The scene had been observed By the neighbouring customs post who telephoned the gendarmerie which immediately sent their available staff to the scene. The CRS for their part acting on their own had obtained information on what was happening and shared the work required with the gendarmerie:
stopping vehicles on the exposed roads, isolating the danger zone, evacuating the houses and the school of the nearby village area of the refinery which was in serious danger.
Sphere 443 had caught fire: it was a drinks retailer who telephoned the fire brigade in the city at 0712 hours. Two other phone calls were received from the refinery a bit later. The direct telephone line had not been used.
At the factory general alert was raised by a siren while the three professional firemen on duty who had been unable to plug the escape tried in vain to extinguish the fire of the sphere by attacking it with powder extinguisher and activating the fixed cooling system of the eight spheres and of the two liquefied hydrocarbon towers.

At around 0710 hours, the first refinery firemen arrived with a fire truck and a dry chemical truck and they attempted in vain to close the valves.
The stock of powder (1500/kg) being quickly exhausted, the safety offcier, the chief of the group, tried unsuccessfully to use the foam extinguisher which he had available. This piece of equipment could not function due to lack of water suction; a foam launcher could not be used for lack of pressure.
In fact, while the fire fighting network of the refinery was designed to deliver a maximum of 800/m3/hour of water the simultaneous opening of the cooling systems for the propane and butane tanks by the safety officers required the use of 1.128/m3/hour. Therefore, from the beginning of the fight against the fire, water was in dangerously short supply. The situation was aggravated by the fact that the neighbouring gas company which also used the water supply network of the refinery had, as a precaution, also started the cooling system for its two propane spheres and was hosing them with a fire hose.
The fire brigade from the city arrived on the spot from 0733 hours onward in successive pickets led in turn by the Adjutant Prevost, (from 0743 hours} and the Commander (from 0746 hours). They joined their efforts with those of the professional and auxiliary firemen from the refinery and were in turn joined by members of the fire fighting team of a nearby factory who arrived at 0820 hours and fire pioneers who after being alerted by the Commander from the city arrived at 0828 hours.
As chief of the first intervention picket from the city's adjutant Prevost occupied himself immediately with sphere 443 which he tried to extinguish with the help of the foam launchers. Being unable to succeed he abandoned the burning tank and concentrated his efforts on the neighbouring propane tank 442.
The rescuers giving up the attempt to extinguish the fire devoted themselves exclusively to the cooling of the other tanks to prevent them from catching fire and hoping that sphere 443 would empty its content which burned as soon as it entered the atmosphere.
However, faced with the drop in pressure already mentioned, Adjutant Prevost and subsequently the two Commanders decided to put a special high powered fire engine for hydrocarbon fires on suction in the nearby canal, but for lack of adequate fittings this was sucked in and could only be recovered after some twenty minutes.
On the other hand, the rescuers were handicapped by the customs enclosure the doors of which were padlocked. Employees of the refinery forced the padlocks and then demolished the enclosure with an excavator.
Meanwhile, reinforcements had continued to arrive and authority was passed first to another Commander.

At 0745 hours the important event mentioned above occurred: the release of the safety valve of sphere 443; the gas which escaped through it caught fire immediately causing a fire column of some 10/meters in height. This incident was interpreted as reassuring by some of the people in charge at the refinery: it indicated according to them that the sphere would empty itself completely.
They told the Commander and some of his co-workers so.
However, some of the rescuers were gripped by a mute apprehension born of the considerable increase of flames enveloping sphere 443 and the growing turmoil caused by the conflagration.
As to the manner in which the accident was attacked, the two Commanders had confirmed the measures taken by Adjutant Prevost, restricting themselves to a role of preventing the spread of the accident by hosing the tanks that were likely to catch fire.
The lowering of pressure constrained the rescuers to a dangerously close approach to the tanks as the water from their launchers reached the top only with difficulty. This dangerous situation determined the Commander to pull his men back after they had fixed their launchers in firm hosing positions.
Nearly one hundred and seventy people were then in are a B.7/1 and in the other areas of zone B. They were firemen from the nearby cities, professional and auxiliary firemen from the refinery and from neighbouring companies or companies working for the refinery, the director, department heads, employees of the factory, supervisors and staff from neighbouring factories and spectators.
The explosion of sphere 443 which occurred at 0845 hours struck most of these people. Added to the waves of burning gas caused by the deflagration were pieces of steel, some of them of considerable weight, that were hurled in some instances over several hundred metres.
Seventeen rescuers succumbed to the explosion or later on to their severe burns. Among the eighty four injured (...) forty two suffered complete disablement for work for more than three months.
However, the explosion had extinguished the fire in the whole of areas B.7/1 and B.7/2 and the southern part of area B.ll. The rescuers whose courage had been above praise and some of whom had saved the lives of colleagues in danger while risking their own lives then fell back, taking the injured with them.
On account of the explosion of sphere 442 at 0945 hours did not cause further victims but, like the preceding one, did cause much material damage as far as 16/km away.
Between the two blown-up spheres a crater, 35/meters long, 15.4/meters wide and 2.1/meters deep had opened up.

Lessons learned
The disaster was the worst accident which had occurred in petroleum and petrochemical plants in Western Europe. Since then, many pressurized tanks containing liquefied gases have BLEVE'd.
Many firemen and emergency responders have been killed while trying to control large fires that the cautious philosophy is to evacuate and take shelter until the material burns itself out.
BLEVE's produced intense thermal radiation from the fireball. This and blast damage from the bursting pressure tanks or spheres are relatively localised compared with unconfined vapour cloud explosions. Therefore, evacuation of up to 500/m will usually ensure the saefty of people. Burning hydrocarbon storage tanks or spheres or very spectacular but unpredictable. Therefore, reporters, passers-by, sightseers must be kept well away for their own safety.

En 1966, des d??fauts de conceptions avaient ??t?? mis en ??vidence dans une raffinerie. De plus, les consignes de s??curit?? n'??taient pas respect??es le jour de l'accident.

Lors de la prise d'??chantillon sur une sph??re de propane (gaz lourd), une fuite s'est r??pandue vers l'autoroute bordant l'installation. Une voiture passant, l'??tincelle provoqua un retour de flamme et un feu de jet sous le r??servoir.

fuite de gaz sous la sph??re de propane: formation d'une nappe gazeuse sur 1.5m de hauteur se propageant jusqu'?? l'autoroute.
initiation de l'incendie de la nappe par une voiture circulant sur une route jouxtant l'autoroute.
inflammation, feu de jet, BLEVE apr??s 1 heure (explosion d'un r??servoir), 1 heure plus tard, la seconde sph??re de propane explose.

17 morts et 84 bless??s.
crat??re 35m*16m*2m entre les deux sph??res.
Les responsables du sinistre ??taient:
autorit??s et pompiers: ni p??nals, ni civils mais responsables de la mise en place des moyens de lutte contre le feu chef d'??tablissement: il ??tait le plus apte ?? savoir ce qui se passait dans son usine et pouvait orienter certaines actions ou ??viter certaines erreurs.
En temps de crise, au plan op??rationnel, le chef du d??tachement des pompiers, les ing??nieurs de la DRIRE et le responsable de l'usine doivent collaborer F. ??tait une anomalie administrative et territoriale.
La raffinerie est situ??e aux portes de Lyon, dans le d??partement de l'Is??re ?? l'??poque. C'??tait donc le capitaine des pompiers de Vienne qui ??tait venu prendre la situation en main, avec les moyens dont il disposait c'est-??-dire inadapt??s ?? la situation. Il aurait fallu des moyens plus puissants, comme ceux de L., et, administrativement, les moyens de L. n'avaient rien ?? faire avec la raffinerie. Les secours ??taient intervenus dans de mauvaises conditions. Les secours de Lyon ??tant contact??s, il y eu des cafouillages car le capitaine des sapeurs-pompiers de Vienne ne savait plus qui ??tait en droit de commander quand le colonel des sapeurs-pompiers est arriv??.

Retours d'exp??rience
Pr??paration et mise en place d'une charte s??curit?? qui d??finit la politique, les objectifs et les mesures ?? prendre en cas d'accident (instructions concernant la pr??vision et les d??cisions prises en cas de crise)
Pr??cision des mesures de gestion de crise dans l'arr??t?? minist??riel du 4 septembre 1967 (1 an et demi apr??s l'accident).
Cr??ation d'un code de qualit?? sur l'installation p??troli??re par le minist??re de l'Industrie.
D??finition d'un POI.
L'annexe ORSEC Hydrocarbures est devenue le PPI.
Le chef d'??tablissement, pour la premi??re fois, a ??t?? d??clar?? responsable de la pr??paration et de la lutte contre le sinistre de son installation. Avant, il n'??tait ni responsable, ni chef des secours.
Attribution de la surveillance du site par les sapeurs-pompiers.
Etablissement du "Guide pratique du chef d'??tablissement" pour diminuer le dysfonctionnement du syst??me de gestion de crise: 10 sc??narios classiques avec les mesures d'urgence, sch??ma d'information interne qui s'adresse ?? l'administration, ?? la direction locale qui se charge des relations avec la presse et le voisinage.

Home   Browse database   Free examples   Contact us