NEWS

We are pleased to inform you that the 88^th edition of the FABIG Newsletter is now available on the FABIG website.

FABIG Members are able to download Issue 088 from the Technical Newsletters section of the website (once logged-in) whilst non-Members of FABIG are able to purchase it.

This issue comprises the following:

• Editorial: FABIG to participate in International Process Safety Week (IPSW) in December 2024
  G. Vannier - The Steel Construction Institute (SCI)
• Consequences of the 2020 Beirut explosion
  S. Rigby^1,2, A. Ratcliff¹, S. Clarke¹, D. Farrimond¹, S. Fay³ and W. Wholey⁴ - ¹University of Sheffield, ²Arup Resilience (UK), ³AWE, ⁴Arup Resilience (USA)
  This paper presents a case study of the 2020 Beirut explosion and efforts to better understand its consequences. Video footage posted on social media shortly after the explosion was first used to derive an approximate explosive yield by compiling the radius-time profile of the blast wave and correlating with well-known semi-empirical relations. This was followed by a detailed physics-based simulation of the explosion using the estimated yield. The results were used in a forensic assessment of observed building damage.
• Modelling the behaviour of LPG tanks exposed to partially engulfing pool fires
  G.E. Scarponi, V. Cozzani, G. Antonioni and F. Doghieri - University of Bologna
  This paper presents a parametric analysis to investigate the behaviour of a 1.9 m³ bullet LPG tank in partial engulfment fire scenarios. A set of case studies were defined, varying the positions of the zone exposed to fire and the filling degree of the tank. These were simulated using a CFD modelling approach previously validated against partial engulfment fire test results, with a view to demonstrating how CFD simulations may be considered as “digital twins” of fire tests and used to investigate specific safety aspects in detail.
• Hydrogen jet diffusion modeling using physics-informed graph neural network and sparsely-distributed sensor data
  J. Shi, X. Huang and A.S. Usmani - The Hong Kong Polytechnic University
  This paper introduces a physics-informed graph deep learning approach (Physic_GNN) for efficient and accurate hydrogen release and diffusion prediction using sparsely-distributed sensor data. Publicly available experimental data of hydrogen jet diffusion was applied to demonstrate the higher accuracy and computational efficiency of the proposed approach.

Take care,

The FABIG Team