Journal of Hazardous Materials A137 (2006) 1293–1302 Experimental study of thin-layer boilover in large-scale pool fires Fabio Ferrero a,∗ , Miguel Mu ˜ noz a , Bulent Kozanoglu b , Joaquim Casal a , Josep Arnaldos a a Centre d’Estudis del Risc Tecnol` ogic (CERTEC), Chemical Engineering Department, Universitat Polit` ecnica de Catalunya, Diagonal 647, 08028-Barcelona, Catalonia, Spain b Mechanical Engineering Department, Universidad de las Am´ ericas, Puebla, 72820-Santa Catarina M ´ artir, Cholula, Puebla, Mexico Received 20 February 2006; received in revised form 24 April 2006; accepted 27 April 2006 Available online 4 May 2006 Abstract By performing a series of pool fire experiments, the authors attempt to apply knowledge of thin-layer boilover to the large scale (pool diameter from 1.5 to 6m). Two commercial hydrocarbons were used: gasoline and diesel. As expected, only in the case of diesel did the phenomenon of thin-layer boilover occur. Data were used to analyze various features of thin-layer boilover in fires, particularly its intensity and onset time. A comparison with results published in the literature shows the importance of this study. © 2006 Elsevier B.V. All rights reserved. Keywords: Pool fires; Pre-boilover burned mass ratio; Boilover intensity; Onset time 1. Introduction Fire accidents are a common threat in the process industry and in the transportation of hazardous materials. An accidental fire scenario often involves a hydrocarbon burning above a water layer. When the fuel layer is thin and its boiling temperature exceeds that of water, heat transfer from the flame can lead to a phenomenon known as thin-layer boilover. This event involves the eruptive vaporization of the water layer, which causes burn- ing fuel to be ejected and increases the turbulence of the fire, thus giving rise to larger flames and radiation. The phenomenon has mainly been studied in laboratories [1,2], and the results cannot be extrapolated to a real scenario. Thus, our research consisted of a series of outdoor large pool fire experiments, which enabled us to study the onset time, boilover intensity and thermal penetration rate. Data analysis enabled us to derive a set of important param- eters relevant to the thin-layer boilover phenomenon, such as onset time, boilover intensity (the effect on the burning rate) and the evolution of the temperature of the fuel and water lay- ers. The influence of pool diameter and initial layer thickness on these parameters was studied. The onset time of thin-layer boilover was characterized by an increase in the sound level of ∗ Corresponding author. Tel.: +34 93 401 66 75; fax: +34 93 401 71 50. E-mail address: fabio.ferrero@upc.edu (F. Ferrero). the fire, which is associated with the explosion of fuel/water bubbles formed during the vaporization phase. Boilover inten- sity was calculated from the relationship between the burning rate during boilover and the burning rate during the stationary period. Unless otherwise specified, hereinafter we use the word boilover to refer to thin-layer boilover. 2. Experimental facility and methods Experimental tests were carried out in five concentric circu- lar pools made of reinforced concrete (1.5, 3–6 m in diameter, respectively). In order to measure flame temperature, thermo- couples were fixed at different positions on a metal structure built on a concrete base, 1 m from the outer pool. Furthermore, 10 K- type thermocouples, fixed at the pool’s axis at a distance of 2 mm from one another, were used to determine the temperature of the two liquid phases and their interface. The burning rate was deter- mined by measuring the variation in the fuel level using a system of communicating vessels. The tests were filmed with two video cameras, which enabled us to study flame height, and a thermo- graphic camera (IR). The cameras were fixed at pre-calculated distances so that the flames could be viewed in their entirety. The sound level of the fire and the volume of the unburnt residue were measured. Two heat flux sensors were used to measure external radiation on targets at specified distances. Moreover, various 0304-3894/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jhazmat.2006.04.050