Fire, process safety and risk analytics. I develop physics-based urban fire models, explainable analyses of occupational accidents, and evidence-led systems that turn complex safety information into decisions people can use.
I work where hazard physics, incident evidence and practical safety decisions meet.
Physics-based and reduced-order fire-spread models for city-scale scenarios, tenability reasoning, uncertainty mapping and multi-hazard integration.
Structured learning from accident records using statistical analysis, fault trees and interpretable machine learning to identify prevention priorities.
Research software that converts reports, regulations and evidence chains into reviewable findings, auditable workflows and defensible outputs.
A concise view of what the studies contribute, not only where they were published.
Analysis of 128 fatal cases found high occurrence of inadequate supervision, failure to use PPE, and absent or poorly followed work standards—showing where prevention systems must intervene first.
Read the paper ↗Across 3,321 Malaysian cases, general workers were the most vulnerable group, while roofers and electricians also showed prominent annual fatality shares.
The study demonstrated the feasibility of machine learning for construction safety management. Random forest was selected for its performance and ability to reveal how site, management, individual and agent factors shape predictions.
In 103 agriculture, forestry and fishing fatalities, struck-by-object events were most frequent, but the fault tree ranked falls higher through dominant exposure-source pathways.
Journal of Energy & Safety Technology
IEEJ Journal of Industry Applications
Journal of Safety Research
Safety Science
Selected modelling frameworks and evidence-led systems I have developed or helped shape.
I co-developed a physics-based urban fire-spread model and integrated it into a multi-hazard and fragility workflow for scenario planning. My contribution connected fire behaviour, tenability thresholds, model validation and decision-ready outputs.
The original class-city fire-safety model designed to represent how fire can move through urban building stocks and support comparative city scenarios.
A reproducible exposure-to-fire workflow connecting building footprints, reduced fire physics, FDS-reference templates, district cases and probabilistic maps.
An evidence-led workspace for reviewing Malaysian Environmental Impact Assessment reports against 158 structured findings, with page evidence, rationale, confidence, officer overrides and audit trails.
A Malaysia-first design-safety assurance platform mapping legal duties, prevention-through-design decisions, HIRARC, bowties, evidence chains and controlled deliverables.
A background spanning academic research, engineering education and on-site safety work.
University College London, United Kingdom
Universiti Putra Malaysia
DMA Construction · Bonatti SPA · Sonatrach
Accessible explanations of the methods, findings and limits behind my work.
For research collaboration, teaching, consulting or project discussions, contact me directly.