Hazardous Materials Releases
High-fidelity prediction of gas dispersion, toxic cloud evolution, and explosion risk in complex environments, enabling safer design and rapid response in industrial and energy systems
Technical capabilities in MPflow
MPflow provides advanced CFD tools for simulating hazardous material releases across a wide range of scenarios, including gas leaks, LNG dispersion, and accidental releases in industrial environments.
Our capabilities include:
• Simulation of compressible and incompressible multi-species flows
• Modelling of buoyancy-driven dispersion (heavy and light gases)
• Treatment of turbulent mixing and atmospheric boundary layer effects
• Accurate prediction of dense gas behaviour (e.g. LNG, CO₂, hydrocarbons)
• Coupling with terrain, obstacles, and built environments
• Support for transient releases, jet dispersion, and continuous leaks
MPflow solvers are fully compatible with OpenFOAM libraries and incorporate robust numerical schemes for stability and accuracy in highly transient dispersion problems.
Our CFD approach
Using MPflow, MultiFluidX delivers high-fidelity simulations of hazardous releases, capturing the full physics of dispersion and risk formation:
- Resolution of transient gas clouds and concentration fields
- Accurate modelling of jet release, flashing, and phase change effects
- Prediction of flammable and toxic concentration thresholds
- Integration of wind conditions and atmospheric stability classes
- Simulation of complex geometries including industrial facilities and urban layouts
- Support for risk assessment, safety distances, and mitigation design
Physics-based ML acceleration
Our ML-CFD framework enhances traditional CFD by learning from high-fidelity simulations and experimental data:
- Development of surrogate models for rapid dispersion prediction
- ML-based closure models for turbulence and mixing processes
- Real-time estimation of gas concentration fields and hazard zones
- Deployment of trained models for scenario screening and digital twins
- Hybrid workflows combining CFD accuracy with ML speed
These models retain physical consistency while dramatically reducing computational cost.
Why ML-CFD matters for Hazardous Materials Releases
- Up to 5000× faster simulations, enabling near real-time assessment of accidental releases and emergency scenarios
- 5–7× speed-up with on-the-fly ML deployment for RANS/LES turbulence modelling
- High accuracy (<5% deviation) in concentration fields and dispersion behaviour compared to validated CFD
- Instant evaluation of multiple leak scenarios, supporting rapid risk analysis and safety planning
- Scalable across conditions, including different gases, weather scenarios, and facility layouts
- Real-time decision support, enabling integration into monitoring systems and digital twins
