Explosion of a Hot-Oil Circuit as a Consequence of Process, Plant Engineering, and Management Shortcomings N. Piccinini and M. Demichela Centro Studi su Sicurezza, Affidabilita ` e Rischi (SAfeR), Dipartimento Scienza dei Materiali e Ingegneria Chimica, Politecnico di Torino, Corso Duca degli Abruzzi, 24 –10129 Torino, Italy; norberto.piccinini@polito.it (for correspondence) Published online 16 November 2005 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/prs.10108 On June 3, 1993, an explosion occurred at the Mediterranea refinery, in Milazzo, Sicily. Seven people died and eighteen people were injured with burns. An explosion in the accumulation tank 27V-16 involved hot-oil circuits subserving the refining units Topping 3, Topping 4, and HDT-RC. The explosion was caused by a sudden overpressure in the 27V-16 tank, probably as a result of the sudden vaporization of a small quantity of water. Shortcomings in the process design and in the management were at the origin of the accident, in particular: • Some process choices undertaken a couple of years before the accident, such as the use of gas oil from a Topping column in place of diathermic oil and the use of a combustible gas in place of nitrogen. • Some operational mistakes in the 13F-2 oven con- duction and the failure (or inadequacy or exclu- sion) of control systems or automatic and manual safety devices. A detailed Fault Tree connects the process, design and management shortcomings identified. © 2005 American Institute of Chemical Engineers Process Saf Prog 25: 21–32, 2006 1. INTRODUCTION On June 3, 1993, at 13.15, an explosion occurred at the Mediterranea Refinery in Milazzo, Sicily. Seven peo- ple died and eighteen people were injured with burns. The technical and managerial factors that led to the collapse of a hot-oil circuit expansion tank while it was being started up are discussed in the paper. The tank collapsed because a modest quantity of water (a few hundred liters) entered the tank, which contained gas oil at a temperature of 250° C. This kind of accident is known to have occurred before. The FACTS database, maintained by TNO (The Netherlands), records at least six accidents of this type. A report by DuPont [1] warns about risks of this kind. Finally, the American Petroleum Institute (API), in Sec- tion 2 of Report 758, on “Safety Digest of Lessons Learned” [2], discusses explosions resulting from the mixing of hot oil with vaporized water during the start-up phase of a plant. As a consequence of its analysis of these explosions, API suggests following specific procedures when starting up hot-oil units. Sim- ilar accidents are also described in Kletz [3] and in the BP Booklet “Hazards of Water” [4]. This literature shows that this type of accident is well known. In the accident analyzed herein, API’s recommenda- tions were not heeded. In addition, there were process, plant engineering, and managerial shortcomings that are worth studying, to learn some lessons from this case. Sources for the present work include documents presented at the trial that followed the accident [that is, piping and instrumentation diagrams (P&ID), work- sheets, the supervisors’ consignment register, inter- views, photographs, etc.], technical reports drawn up by a succession of different consultants, and the sen- tence issued by the Court of Barcellona P.G., Messina, on June 20, 1998 [5, 6, 8 –11]. © 2005 American Institute of Chemical Engineers Process Safety Progress (Vol.25, No.1) March 2006 21