Aeronautical turbulent reacting flows involve a wide range of
scales and complexities caused by the specific shapes of engines
and the combustion regimes encountered in these devices. Because
of the space and weight constraints, designers need to develop
burners which ensure maximum efficiency and compactness. Over
the years, manufacturers have gained significant experience and
existing designs largely rely on flow recirculations to increase
mixing and flow-though times despite a reduced size combustion
chamber. In parallel, pollutant emissions and regulations have
induced changes of technology with the emergence of partially
premixed and premixed burners. Multi-point fuel injection systems
and staging are also being implemented as potential solutions to
the new regulations. All these concepts increase the complexity of
the flow and lead to specific flow dynamics and combustion
responses. Although these designs are being routinely evaluated
by Computational Fluid Dynamics (CFD), most present modeling
strategies rely on Reynolds Average Navier-Stokes (RANS)
approaches developed for mean stationary flows [1e10]. Such
models benefit from extensive research and developments from
the scientific community and have been successfully calibrated on
simple fundamental configurations. However, the complexity of
flows in modern gas turbines adds multiple constraints on RANS
and limits their precision, Fig. 1. Alternative numerical solutions are
thus needed to further increase the share of CFD and decrease the
number of real engine tests and design iterations.
CFD alternatives to RANS for aeronautical gas turbine applications
must justify the increase in development, maintenance and
computer costs. These new tools need also to be compatible with
existing industrial knowledge and conception rules. The use of new
CFD approaches and their future in the design chain is still unclear.
It will probably depend on the computing power available to
engineers as well as their ability to master and analyze ever more
Fig. 1.