Atomisation & Spray combustion

High-fidelity prediction of liquid jet breakup, droplet formation, and spray dynamics under complex multiphase, turbulent, and reactive flow conditions across engineering applications

Technical capabilities in MPflow

MPflow provides a unified framework for modelling atomisation and spray processes across regimes, from dense liquid cores to fully dispersed sprays:

  • Dedicated multiphase CFD solvers for primary and secondary atomisation
  • Interface-resolving methods capturing Rayleigh–Taylor and Kelvin–Helmholtz instabilities
  • Natural transition from Eulerian interface tracking & Lagrangian spray modelling
  • Fully coupled LES and RANS turbulence modelling for high-fidelity transient flows
  • Advanced treatment of:
    • Droplet breakup and coalescence
    • Evaporation and heat transfer
    • Two-way coupling between droplets and carrier phase
  • Compatible with OpenFOAM libraries and extended with proprietary MPflow capabilities

Our CFD approach

  • High-resolution simulation of liquid jet instability growth and breakup
  • Direct modelling of interface deformation under turbulence and shear
  • LES-based resolution of transient structures in near-nozzle regions
  • Coupled modelling of:

    • Primary atomisation (jet breakup)
    • Secondary atomisation (droplet fragmentation)
  • Accurate prediction of:

    • Droplet size distribution (SMD)
    • Velocity and temperature fields
    • Spray penetration and dispersion

Physics-based ML acceleration

  • ML models trained on high-fidelity CFD and experimental datasets
  • Surrogate modelling of:

    • Breakup regimes
    • Droplet size distributions
    • Spray evolution dynamics
  • Embedded into solver for:
    • On-the-fly acceleration of LES/RANS closures
    • Rapid prediction of spray characteristics without resolving all scales
  • Hybrid CFD–ML framework preserving physical consistency and conservation laws
  • Generalisation across:
    • Fuels (including cryogenic and reactive fluids)
    • Injector geometries
    • Operating conditions

Why ML-CFD matters for Atomisation & Spray Combustion

  • Simulation time reduced from 8 hours to <1 minute: rapid evaluation of injector and combustion system designs
  • 7–10× acceleration with ML-enhanced turbulence modelling (RANS & LES)
  • High accuracy (<5–7% deviation) in droplet size, spray angle, and penetration
  • Instant exploration of operating conditions, including pressure, temperature, and injection strategies
  • Scalable across fuels and geometries, including complex injector configurations
  • Reduces experimental campaigns and accelerates development cycles in combustion systems