Critical flows

High-speed, compressible and multiphase flows characterised by strong gradients, shocks, and phase change, critical in energy, propulsion, safety, and industrial processes

Technical capabilities in MPflow

MPflow provides advanced solvers for simulating critical flows involving compressibility, shocks, turbulence, and phase change in complex geometries.

  • Dedicated pressure–velocity coupling algorithm tailored for high-speed and internal flows
  • Fully compatible with OpenFOAM libraries, enabling flexible and scalable workflows
  • Robust handling of shock waves, expansion fans, and discontinuities
  • Multiphase capabilities including cavitation, flashing, and compressible mixtures
  • Coupled modelling of internal flow and external discharge (jets, plumes, sprays)
  • Support for RANS and LES simulations in highly transient regimes

Our CFD approach

We use high-fidelity CFD to resolve the complex physics governing critical flows across a wide range of applications.

  • Resolution of strong pressure and velocity gradients in compressible regimes
  • Accurate prediction of shock–boundary layer interactions
  • Modelling of phase change and metastable flow behaviour
  • Coupled simulation of internal passages and downstream flow evolution
  • Validation across industrial cases, including high-speed nozzles and safety scenarios

Physics-based ML acceleration

Physics-aware machine learning enhances CFD by accelerating predictions while preserving physical consistency.

  • ML models trained on high-fidelity CFD datasets and experimental data
  • Deployment for full-field prediction or sub-model acceleration
  • On-the-fly integration into solvers for turbulence and source-term modelling
  • Real-time prediction of pressure, velocity, and phase distribution fields
  • Generalisation across operating conditions, geometries, and fluid properties

The validation case of our technology has been featured in the prestigious journal Physics of Fluids

Why ML-CFD matters for Critical Flows

  • 100–1000× faster predictions: near real-time analysis of compressible and high-speed flows
  • 5–10× acceleration with ML-enhanced turbulence and source models
  • <5% deviation from CFD in key flow features including shocks and phase interfaces
  • Instant evaluation of transient events such as rapid depressurisation and jet release
  • Scales across regimes: subsonic, transonic, supersonic, multiphase
  • Enables digital twins and real-time monitoring for safety-critical systems