Multi-hole injectors

High-fidelity simulation of multi-jet spray interaction, atomisation, and fuel-air mixing, enabling optimisation of injector design for efficient combustion and reduced emissions in engines

Technical capabilities in MPflow

MPflow provides advanced capabilities for resolving complex multi-hole injector flows and spray interactions:

  • Dedicated solvers for internal nozzle flow and external spray formation
  • High-resolution modelling of multi-jet interaction and spray overlap
  • Interface-capturing methods for primary atomisation inside injector holes
  • Eulerian–Lagrangian framework for spray evolution and droplet tracking
  • Fully coupled RANS and LES turbulence modelling
  • Accurate prediction of:

    • Spray cone angle and penetration
    • Droplet size distribution (SMD)
    • Jet-to-jet interaction effects
  • Fully compatible with OpenFOAM libraries and extended MPflow solvers

Our CFD approach

  • Detailed simulation of flow inside injector holes, including cavitation and turbulence effects
  • Resolution of primary breakup at nozzle exit and transition to spray regime
  • Modelling of multi-hole spray dispersion and interaction in large combustion chambers
  • LES-based prediction of transient jet structures and turbulent mixing with air.
  • Captures key mechanisms:

    • Jet spreading and merging
    • Spray cone widening due to multiple holes
    • Enhanced fuel-air mixing

Physics-based ML acceleration

  • ML models trained on high-fidelity CFD injector datasets
  • Surrogate prediction of spray angle and penetration and droplet size distributions.
  • Embedded ML acceleration for turbulence closures (RANS/LES) and spray sub-models (breakup, dispersion).
  • Enables rapid evaluation across:
    • Number of holes and hole diameters
    • Injection pressures
    • Fuel types and operating conditions
  • Maintains physics consistency and generalisation across injector designs

Why ML-CFD matters for Multi-hole Injectors

  • Up to 1000× faster predictions: rapid evaluation of injector configurations and spray performance
  • 5–10× acceleration with ML-enhanced turbulence and spray modelling (RANS & LES)
  • High accuracy (<5–7% deviation) in spray angle, penetration, and droplet size distribution
  • Instant exploration of injector design space: number of holes, diameter, spacing, and pressure
  • Improves fuel-air mixing and combustion efficiency through rapid optimisation
  • Reduces experimental testing and engine development time