A multi-objective CFD optimization of liquid fuel spray injection in dry-low-emission gas-turbine combustors

Highlights: •An Eulerian-Lagrangian model for the fuel spray injection is evaluated. •The drop breakup, spray-vortex interaction, and wall-wetting play the key roles. •The injection location and direction are the most important parameters. •The best design candidates are proposed using multi-objective optimizations. •A large central perpendicular injection with high co-rotating swirls is optimal. -- Abstract: The main goal of this research is to investigate the effects of fuel injection strategy on the performance of the premixing chamber of modern Dry-Low-Emission (DLE) Gas-Turbine (GT) combustors. Here, an Eulerian-Lagrangian model for multi-phase multi-component flows is evaluated and used to investigate the effects of different fuel spray design parameters, including the injection location, direction, mass-flow-rate partitioning, and flow Swirl number, on the performance of the premixing chamber. The analysis is enriched by multi-objective optimizations accounting for several goals, including the evaporation efficiency, mixture stratification, entropy generation, and flow recirculation. It is observed that the droplet breakup, spray-vortex interactions, and wall-wetting have significant influences on the performance objectives while the droplet residence time effect is minor. Among the design parameters, the injection location and direction have a profound impact on the droplet breakup which predominately controls the evaporation efficiency. In addition, the interactions between the spray and the two swirling vertices inside the chamber strongly affect the mixture stratification (uniformity), e.g. the location and direction of the injection should not be chosen such that a large proportion of fuel droplets are trapped in the shear layer between the two vortices (otherwise the evaporation efficiency drops significantly) or trapped in the strong outer swirling vortex (if large mixture non-uniformity should be avoided). Finally, the best designs meeting a series of targets, e.g. simultaneous perfect mixing and uniformity, etc., are introduced.