Integrated RANS/LES computations of turbulent o w through a turbofan jet engine

The interaction between di erent components of a jet engine represents a very impor-tant aspect of the engine design process. Sudden mass ow-rate changes induced by owseparation and pressure waves, interaction of the unsteady wakes originating from thefan blades with the low-pressure compressor, high temperature streaks interacting withthe rst stages of the turbine are all complex unsteady phenomena that cannot be sim-ply accounted for through boundary conditions of a single component simulation. Onlysimulations that integrate multiple engine components can describe these ow featuresaccurately.Today’s use of Computational Fluid Dynamics (CFD) in gas turbine design is usuallylimited to component simulations. The demand on the models to represent the largevariety of physical phenomena encountered in the ow path of a gas turbine mandatesthe use of a specialized and optimized approach for each component. The ow- eld in theturbomachinery portions of the domain is characterized by both high Reynolds numbersand high Mach numbers. The prediction of the ow requires the precise description ofthe turbulent boundary layers around the rotor and stator blades, including tip gaps andleakage ows. A number of ow solvers that have been developed to deal with this kindof problem have been in use in industry for many years. These ow solvers are typicallybased on the Reynolds-Averaged Navier-Stokes (RANS) approach. Here, the unsteadyow- eld is ensemble-averaged, removing all the details of the small scale turbulence; aturbulence model becomes necessary to represent the e ects of turbulence on the meanow.The ow in the combustor, on the other hand, is characterized by multi-phase ow,intense mixing, and chemical reactions. The prediction of turbulent mixing is greatlyimproved using ow solvers based on Large-Eddy Simulations (LES). While the useof LES increases the computational cost, LES has been the only predictive tool able tosimulate consistently these complex ows. LES resolves the large-scale turbulent motionsin time and space, and only the inuence of the smallest scales, which are usually moreuniversal and hence, easier to represent, has to be modeled (Ferziger, 1996, and Sagaut,2002). Since the energy-containing part of the turbulent scales is resolved, a more accuratedescription of scalar mixing is achieved, leading to improved predictions of the combustionprocess, as shown in Raman & Pitsch (2005). LES ow solvers have been shown in thepast to be able to model simple ames and are currently being adapted for use in gasturbine combustors, e.g., Poinsot et al. (2001) and Constantinescu et al. (2003).In order to compute the ow in the entire jet engine, one needs to couple RANS andLES solvers. We have developed a software environment that allows a simulation of multi-component e ects by executing multiple solvers simultaneously. Each of these solverscomputes a portion of a given ow domain and exchanges ow data at the interfaces with