Investigation of pool spreading and vaporization behavior in medium- scale LNG tests Nirupama Gopalaswami, Ray A. Mentzer, M. Sam Mannan * Mary Kay O'Connor Process Safety Center, Artie McFerrin Department of Chemical Engineering Department, Texas A&M University, College Station, TX 77843-3122, USA article info Article history: Received 3 July 2014 Received in revised form 21 October 2014 Accepted 21 October 2014 Available online 22 October 2014 Keywords: Consequence modeling Pool spreading Vaporization LNG Source-term Cryogenic spill abstract A failure of a Liquefied Natural Gas (LNG) tanker can occur due to collision or rupture in loading/ unloading lines resulting in spillage of LNG on water. Upon release, a spreading liquid can form a pool with rapid vaporization leading to the formation of a flammable vapor cloud. Safety analysis for the protection of public and property involves the determination of consequences of such accidental re- leases. To address this complex pool spreading and vaporization phenomenon of LNG, an investigation is performed based on the experimental tests that were conducted by the Mary Kay O'Connor Process Safety Center (MKOPSC) in 2007. The 2007 tests are a part of medium-scale experiments carried out at the Brayton Fire Training Field (BFTF), College Station. The dataset represents a semi-continuous spill on water, where LNG is released on a confined area of water for a specified duration of time. The pool spreading and vaporization behavior are validated using empirical models, which involved determina- tion of pool spreading parameters and vaporization rates with respect to time. Knowledge of the pool diameter, pool height and spreading rate are found to be important in calculating the vaporization rates of the liquid pool. The paper also presents a method to determine the vaporization mass flux of LNG using water temperature data that is recorded in the experiment. The vaporization rates are observed to be high initially and tend to decrease once the pool stopped spreading. The results of the analysis indicated that a vaporization mass flux that is varying with time is required for accurate determination of the vaporization rate. Based on the data analysis, sources of uncertainties in the experimental data were identified to arise from ice formation and vapor blocking. © 2014 Elsevier Ltd. All rights reserved. 1. Introduction Consequence modeling for LNG safety deals with three distinct stages in estimating the potential consequences of major hazard accidents. The first stage is to determine the release mode and release rate of the hazardous material. The second stage is to determine the behavior of the material after its release, and the third is to consider the effects of the material on people (Woodward and Pitblado, 2012). The first stage of quantifying the accident scenario is often studied as source-term modeling. An essential part of source-term modeling for LNG releases on water involve pre- diction of pool spreading parameters and vaporization rate. The pool spreading and vaporization of LNG on water has been studied previously in several small- and large-scale experiments by Shell (Boyle and Kneebone, 1973), Bureau of Mines (Burgess et al., 1970) and Esso Tests (May et al., 1973). Accurate data for pool spreading and vaporization is currently limited. The most comprehensive experiments that are carried out are primarily to understand vapor dispersion and these have certain uncertainties in the measurements and models that are used to determine the source-term. This is due to difficulties in direct measurement of the vaporization rates or pool spreading parameters like pool height, pool radius and spreading rate. Unlike vapor dispersion experi- ments, where measured gas concentration and temperature data are used for validation of models, source-term modeling experi- ments involve an additional secondary treatment of experimental data to obtain the required parameters like vaporization rate for validation of prediction models. The vaporization rate in LNG experiments is determined through one of the three methods. The first method is the measure of the loss of LNG mass that is occurring due to vaporization (Boyle and Kneebone, 1973). The vaporization rate is obtained by deter- mining the slope of mass loss data. This method has been applied widely in small-scale experiments. The second method is to * Corresponding author. E-mail address: mannan@tamu.edu (M. Sam Mannan). Contents lists available at ScienceDirect Journal of Loss Prevention in the Process Industries journal homepage: www.elsevier.com/locate/jlp http://dx.doi.org/10.1016/j.jlp.2014.10.012 0950-4230/© 2014 Elsevier Ltd. All rights reserved. Journal of Loss Prevention in the Process Industries 35 (2015) 267e276