A comparative study of optimization models for the gas detector layout in offshore platform. (15th April 2022)

• Record Type: Journal Article
• Authors: Zhang, Bo; Liu, Xin; Sun, Lin; Xiao, Fei
• Is Part Of: Ocean engineering. Volume 250(2022)
• Publication Date: 2022-04-15
• Subjects: Offshore platform -- Leakage scenarios -- Gas detector layout -- Computational fluid dynamics (CFD) -- Stochastic programming (SP) -- Particle swarm optimization (PSO) algorithm; Ocean engineering -- Periodicals; Ocean engineering; Periodicals; 620.4162
• DOI: 10.1016/j.oceaneng.2022.110880
• Languages: English
• ISSNs: 0029-8018

Abstract

Abstract: The layout of gas detector is an important factor affecting whether hazardous gas can be detected successfully. However, existing standards only put forward general requirements, and the practical application effect is not satisfying. In this paper, three p-median optimization models with different objectives have been proposed, including the minimal cumulative detection time model (MCDT), the minimal cumulative leakage concentration model (MCLC), and the minimal cumulative individual risk model (MCIR). To achieve the balance of computation cost and optimization accuracy, the complete accident scenario set (CASS) has been filtered according to the cumulative probability of all scenarios and further more reduced by the K-means cluster analysis. Since some small leakage which could not reach lower explosion limit (LEL) results in limited risk, a scenario screening has been carried out to take out these scenarios. The gas detector layout optimization of an offshore platform has been carried out as a case study. After that, the results have been contrasted, which demonstrates that these models can effectively improve the detection efficiency and each has its own advantages. MCDT model can provide the fastest detection speed which means it suites the situations that require more time to conduct emergency response and evacuation. MCLC model pays more attention to the overall risk of the region. On the other hand, considering the individual risk as the primary focus, MCIR model is more suitable for toxic gas leakage and poisoning situations. Highlights: Build a complete accident scenario set (CASS) which contains all potential accident scenarios. Reduce the computation cost through cumulative probability screening and K-means cluster analysis. An offshore platform is adopted as a case study. The leakage consequences are simulated by Fluent software. Lower explosion limit (LEL) screening is adopted to retain the scenarios which has higher risk potential. Three SP layout optimization models with different objectives are proposed and compared.