Abstract
Belgium has a densely populated territory and counts numerous plants covered by the Seveso II Directive. Regional authorities in Wallonia (Belgium) tend to impose additional safety barriers in order to reduce or prevent the external risk, and the BLEVE phenomenon in particular. This paper presents an innovative solution which was studied on the Belgian Seveso plant of L’Oreal to prevent the BLEVE phenomenon by the immersion of LPG tanks in a pool of water.
This solution is compared with the main classical solutions which are the mounding of the tanks or concrete-spraying on them, in order to prevent or delay the heating due to an external fire. The proposed immersion in a pool is here complemented with a gas detection and aspiration system, which allows to conduct, in case of a leak, the released gas to a safe area.
The comparison of the three solutions firstly focuses qualitatively on the causes of accident, based on the ARAMIS Project generic fault trees. Apart from the BLEVE, which prevention is the main goal, the other consequences of an accidental release (jet fire, explosive cloud dispersion and vapor cloud explosion (VCE)) are then quantified for different weather conditions, for each of the three safety barriers providing an overview of the residual risks.
Qualitatively, the proposed tank immersion shows a reduced number of accidental causes in comparison with the two others solutions. Moreover, for jet fires and VCEs, both the gas detection and the canalization system makes it safer than both concrete-spraying and mounding. The immersion of the tanks also should prove to be more flexible from the practical point of view than mounding (e.g. external inspection of the tanks, replacement of a tank).
However, specific aspects of the innovative immersion solution require a particular attention, such as water treatment and ice formation. Propositions are then made on how to deal with these aspects.
