LPG leakage and explosion accident analysis based on a new SAA method. (July 2021)

• Record Type: Journal Article
• Authors: Jia, Qingsong; Fu, Gui; Xie, Xuecai; Hu, Shihan; Wu, Yali; Li, Jing
• Is Part Of: Journal of loss prevention in the process industries. Volume 71(2021)
• Publication Date: 2021-07
• Subjects: 24Model -- LPG -- Accident analysis -- SAA; Chemical industries -- Safety measures -- Periodicals; 660.2804
• DOI: 10.1016/j.jlp.2021.104467
• Languages: English
• ISSNs: 0950-4230

Abstract

Abstract: The petrochemical industry has inherently high safety risks. On 5 th June 2017, a liquefied petroleum gas tanker in the storage and transportation department of Linyi Jinyu Petrochemical Co., Ltd. leaked during unloading, causing a major explosion and fire. Ten people were killed and nine were injured. To prevent the recurrence of similar accidents, a retrospective accident analysis was conducted using the system accident analysis method known as '24Model'. This analysis proved the coupling relationship between accident causes involving multiple organisations and multiple factors. The analysis results show that the five modules of maintenance and design, that is, loading and unloading valves, emergency training, management of subsidiaries and contractors, multi-level supervision, and task arrangement need to be checked from top to bottom by relevant organisations. In order to verify the reliability of 24Model analysis results, three SAA methods, namely STAMP, FRAM, AcciMap, were used to analyse the accident, and results obtained by the four methods were evaluated qualitatively and quantitatively. The results show that the output of 24Model is reliable and has a more understandable framework, which can be used as reference for the development of other linear accident cause models and application of system accident analysis methods to real applications. Highlights: A simpler SAA method was used to analyse an LPG accident and prevention suggestions were made based on the results. The concept of 'basic unit' was proposed to describe the expression that SAA repeatedly uses to model the accident causes. The reliability of the output of the 24Model and the other three SAA methods were compared. Some opinions regarding development of linear accident causation theory were presented.