Lifting the Fog Off Hydrocarbon Mist Explosions

Throughout the years, thousands of explosions have been reported in the chemical and petrochemical industry. These hazards did not only include gases, vapors and dusts, but also mists. Indeed, in 2009 Santon reported 37 separate incidents, including 20 mist explosions, in an incident survey for the Health and Safety Laboratory. In most of the cases, the incidents arose from the ignition of mist at temperatures near or below the liquid’s flashpoint. The ongoing occurrence of mist explosions demonstrates the necessity to classify hazardous mist explosive areas. Indeed, the importance of hazardous area classification (HAC), as well as the lack of tools which correlate the dispersion of mists with their flammability and explosion severity, is abundantly expressed in literature. For instance, European ATEX regulations require assessing the risk of formation and ignition of explosive atmosphere (ATEX) associated with the production of a flammable mists. Nevertheless, due to the lack of tools, risk analysis and area classification prove difficult. In order to mitigate such explosions and to assess the flammability and explosion severity of hydrocarbon mists, the factors and criteria of liquid handling, as well as the fluids’ safety parameters, should be identified.

For this study, different fluids, with a high industrial interest, have been selected (e.g. ethanol, kerosene, and diesel) to be tested in a new apparatus based on the standardized 20L explosion sphere. A new fluid injection system was deployed using siphon / gravity-fed spray nozzles comprising a Venturi junction and proposing a wide variety of dispersion performances. This system was controlled using a specifically dedicated program which ensures the versatility of the apparatus and its adaptability to the different liquids tested. It allows a fine control of the gas carrier flow, the liquid flow and both the injection and ignition times, which makes it possible to change the dilution rate for a desired particle size distribution. Ignition can be performed by using chemical igniters or permanent sparks.

Using an in-situ laser diffraction sensor, the droplet size distribution (DSD) was determined for the various fluids dispersed into the 20L sphere. Particle Image Velocimetry was implemented to assess the mist turbulence. Several tests were performed in the 20L sphere to determine the minimum ignition energy, the lower explosivity limit, the maximum explosion pressure (P_max) and the maximum rate of pressure rise (dP/dt _max) of hydrocarbon mists. For instance, for a DSD centered at 55µm, a maximum pressure of 9.17 bar and a maximum rate of pressure rise of 1590 bar.s^-1 were found for ethanol. As for kerosene, for a DSD centered at 80µm, 8.76 bar and 570 bar.s^-1 were found respectively.

Experimental results (P_max and burnt gas composition) were compared with numerical data obtained with Chemical Equilibrium with Applications. Design of experiments methods were applied to stress the sensitivity of parameters such as the chemical nature of the fuel, the DSD, the turbulence level or the mist temperature (gas/mist ratio).

These first results already allow to propose a new protocol to determine safety parameters for hydrocarbon mists and provide tools leading to hazardous area classification.