Rania , BAAZIZ.Enc. Abderraouf, .BOUAFIA.2026-07-142026-07-142026-07-06http://dspace.univ-skikda.dz:4000/handle/123456789/6120This thesis develops an advanced fuzzy Quantitative Risk Assessment (QRA) methodology applied to the depropanizer column (T-301) of the GL1K LNG complex in Skikda. Hazardous scenarios were identified via HAZOP and modeled through Fault Tree/Event Tree Analysis (GRIF), with basic event frequencies revised using a BORA-based Mamdani fuzzy inference system in MATLAB across five Risk Influencing Factors. Consequences (pool fire, flash fire, fireball/BLEVE, VCE) were simulated with PHAST, and individual and societal risks quantified with SAFETI. Results show the fireball/BLEVE dominates risk, with the site-wide individual risk contour and FN curve both exceeding acceptability thresholds, supporting targeted risk-reduction recommendations for GL1KenThesis