Hydrocarbon fires vapor cloud explosions

Due to the destructive nature of hydrocarbon forces when handled incorrectly, fire and explosion protection principles should be the prime feature in the risk philosophy of any hydrocarbon facility. Vapor cloud explosions in particular are consider the highest risk at a hydrocarbon facility. Disregarding the importance of protection features or systems will eventually prove to be costly both in economic and human terms should a catastrophic incident occur without adequate safeguards. 

A deflagration can best be described as a combustion mode in which the propagation rate is dominated by both molecular and turbulent transport processes. In the absence of turbulence (i.e., under laminar or near-laminar conditions), flame speeds for normal hydrocarbons are in the order of 5 to 30 meters per second. Such speeds are too low to produce any significant blast overpressure. Thus, under near-laminar-flow conditions, the vapor cloud will merely bum, and the event would simply be described as a large fiash fire. Therefore, turbulence is always present in vapor cloud explosions. Research tests have shown that turbulence will significantly enhance the combustion rate in defiagrations. 

Hasegawa, K., and K. Sato 1987. Experimental investigation of unconfined vapor cloud explosions and hydrocarbons. Technical Memorandum No. 16, Fire Research Institute, Tokyo. 

The hydrocarbon and chemical industries have traditionally been reluctant to immediately invest capital where direct return on the investment to the company is not obvious and apparent, as would any business enterprise. Additionally, fire losses in the petroleum and chemical industries were relatively small up to the 1950s. This was due to the small size of the facilities and the relatively low value of oil, gas, and chemicals to the volume of production. Until 1950, a fire or explosion loss of more than 5 million dollars had not occurred in the refining industry in the US. Also in this period, the capital-intensive offshore oil exploration and production industry was only just beginning. The use of gas was limited in the early 1900s. Typically production gas was immediately flared (i.e., disposed of by being burnt off) or the wall was capped and considered an uneconomical reservoir. Since gas development was limited, large vapor cloud explosions were relatively rare and catastrophic destruction from petroleum incidents was essentially unheard of The outlays for petroleum industry safety features... 

Sharing of past major incidents with other oil and gas industries provides useful input data for similar process industries in order to identify the most critical barriers and improve their safety processes. One poignant example highlights this matter. In 1998 there was an accident in the gas compression stage of a Middle East oil and gas plant which caused 7 dead as a result of fuel accumulation and vapor cloud explosion which was very similar to the Texas City Refinery disaster on March 23, 2005 in which a distillation tower was overfilled and an uncontrolled release of hydrocarbons led to a major explosion and fires. Fifteen people were killed and 180 were injured in the worst disaster in the United States in a decade. In both incidents, excess hydrocarbons were diverted into a pressure relief system that included a blowdown stack. In the Iranian case, it was equipped with a flare, but one which the operator didn t ignite in Texas City the blowdown stack was not equipped with a flare to burn off hydrocarbons as they were released. As a result, the flammable overflow from the tower entered the atmosphere. Ignition of the escaped hydrocarbons was enabled by startup of a nearby vehicle resulted in the explosion and subsequent fires (Hopkins, 2008). This example shows the repetitive patterns of accidents, and root causes of events all over the world in this sector. The lesson of this paper is that accidents in one country, where the scenarios are very similar, can and should serve as lessons to prevent the same scenario being actualized in other countries. 

This accident occurred on March 23, 2005 at the British Petroleum (BP) Texas City Refinery, the third largest oil refinery in fhe United States, when a hydrocarbon vapor cloud exploded at the isomerization process unit [8]. More specifically, on March 23, 2005 at the refinery, an isomerization unit s start-up whose raffinate tower was overfilled resulted in the raffinate overheating and pressure relief devices opening, and then consequently led to a flammable liquid geyser from a blow down stack unequipped with flare and then an explosion and fire [1,9]. The accident killed 15 workers and injured over 170 others [1,8,9]. 

As a rule of thumb one should approach a hydrocarbon spill (non-fire situation) under the assumption that the liquid is vaporizing (the vapors will be invisible) and that the liberated vapors are heavier than air unless proven otherwise. The expected conduct of a heavier-than-air vapor is for it to drop and spread at or below ground level much as a liquid would. The big difference is that a liquid will be visible and its boundaries well defined. One can expect that the invisible heavier-than-air vapor will settle and collect in low spots such as ditches, basements, sewers, etc. As the vapor navels, it will be mixing with the air, thus some portions of the cloud may be too rich to bum, other sections too lean, and still others well within the explosive range. Some typical vapor densities for petroleum products are 3 to 4 for gasoline, 2.5 for naphtha, and 1.1 for methanol. For comparison, the vapor density for hydrogen gas is 0.1. 

---