Urban fire-spread modelling, multi-hazard decision support, environmental impact assessment review and safety-assurance systems designed around traceable evidence and human oversight.
The fire-modelling projects connect hazard physics to exposure, fragility, uncertainty and the decisions made by planners, engineers and risk stakeholders.
As a Research Fellow in Physics-based Fire Modelling at UCL, I co-developed an urban fire-spread model and integrated it into a wider multi-hazard and fragility decision framework. The contribution included tenability and threshold reasoning, validation and versioning discipline, and example-first materials that made the model easier to interpret within scenario planning.
The original class-city / city-scale fire-safety model in GitHub represents the spread of fire through an urban building stock. It was developed to compare scenarios, explore how local fire behaviour can accumulate into city-scale consequences, and provide a base for later open and uncertainty-aware workflows.
A reproducible, scenario-driven workflow connecting GeoJSON-style building footprints, local coordinates, building-use classification, compartment assumptions, fuel-load uncertainty, a reduced fire-spread physics kernel, formal verification and validation reports, FDS comparison templates, district case studies and probabilistic burn or ignition maps.
The urban-fire work includes explicit scientific cautions: comparison templates are not presented as external validation until real experimental, FDS or incident-reference outputs are inserted. Release audits, scenario integrity checks and reproducibility manifests are used to keep claims aligned with the available evidence.
These tools apply the same research principle in a different form: every conclusion should remain connected to its evidence, assumptions and reviewer judgement.
SEED EIA is an invite-only workspace for reviewing Environmental Impact Assessment reports against a workbook-aligned framework. The system organises 158 findings across nine review tabs and retains evidence excerpts, page references, confidence, assessment rationale, officer corrections and a complete audit trail. It distinguishes presence checks, cross-checks, qualitative review, Annex 5 criteria, digital authenticity and requirements that still need manual review.
A design-safety assurance platform structured around Malaysia's Occupational Safety and Health (Construction Work) (Design and Management) Regulations 2024. The system maps Regulations 1–39 to accountable actions and connects prevention-through-design decisions, DOSH-style HIRARC, bowtie assurance, controlled project records, revision intelligence, BCF issue exchange and regulatory-change impact assessment.
A research, architecture and data-foundation programme for Malaysian QHSE, environmental compliance, sustainability and integrated management systems. The work includes a federal, Sabah and Sarawak jurisdiction model, a 120-record legal and standards seed register, a 40-workflow operating catalogue, sector applicability mapping, controlled source archiving and legal or technical review gates.
An Algeria-focused multilingual QHSE platform concept supporting secure multi-organisation and multi-site operations, legal applicability assessment, draft compliance obligations, calendars, incidents and near misses, permits to work, inspections, training, reporting and Arabic, French and English profile preferences.
Page references, model inputs, official records and assumptions are retained so an output can be checked rather than merely accepted.
Automation assists analysis, while reviewer overrides, uncertainty and unsupported criteria remain explicit.
The end product is a scenario, matrix, control record, review finding or report that supports a real decision.
Some repositories remain private because they contain development work, regulated content or research that is still being prepared for publication. I can discuss the methods, architecture and collaboration opportunities directly.
Contact Dr Abderrahim Zermane