Development of Pre-Warning and Continuous Fire-Safety Framework for High-Rise Buildings using AHP and Bayesian Networks

Résumé fourni par la source

High-rise buildings in dense urban areas are increasingly vulnerable to fire hazards due to restricted spacing, ageing electrical systems and inconsistent compliance with fire-safety guidelines. This study proposes a probabilistic multi-criteria pre-warning framework that integrates twenty key indicators derived from the National Building Code of India (NBC 2016), IS fire-safety standards and international references. Building spacing was classified into three NBC-aligned categories, with class 3 locations showing significantly higher potential for flame spread and reduced fire-service accessibility. The indicators were grouped into four subsystems and weighted using the Analytic Hierarchy Process, while uncertainty in observational data was managed through an unascertained measurement model. A Bayesian network was constructed to model the probabilistic dependencies among ignition likelihood, smoke severity, spread potential and the final fire-risk state. The framework was applied to ten high-rise buildings across the Greater Hyderabad Municipal Corporation using field data obtained from the Telangana Fire Disaster Response Emergency and Civil Defence Department. Results show that smokeexhaust failures, high fire loads, weak compartmentation and poor electrical reliability are the primary drivers of elevated fire risk. The hybrid AHP-BN approach also outperformed K-means clustering in risk identification. The study provides an effective decision-support tool for enhancing fireprevention planning in high-rise structures.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.