Increasing energy demands and mission challenges often require jet engines to operate in particle laden hostile environments, especially, sand, ash and dirt. Large amounts of particle ingestion, sand in particular, can lead to severe damage to various engine components.The presence of particulates in ingested air affects the three major components of jet engines: compressor, combustor and turbine.While many previous studies have investigated the effect of particulates on compressors, the focus of this study is to investigating the effect of sand transport in the internal cooling passages of turbine blades. 



A simplified geometry, a U-shaped duct with square cross section, is considered in the present study to simulate a two pass internal cooling duct.



An effort is made to answer the following questions:

1. How do sand particles impinge the walls and ribs in a two pass duct?
2. Which regions in the two pass duct are most prone to erosion and deposition under prolonged sand ingestion?









































Figure 1. Computational domain (a) Side view of two pass with pitch numbering shown (b) Bottom view, ribs at a section





Carrier Phase: The turbulent flow field is simulated with Wall Modeled Large Eddy Simulations (WMLES) with our in-house conservative finite volume code, Generalized Incompressible Direct and Large Eddy Simulations of Turbulence (GenIDLEST). The governing transport equations for the carrier phase (fluid) are discretized using a second-order central (SOC) difference scheme on a nonstaggered grid topology.  The SOC discretization has been shown to be suitable for LES computations due to its minimal dissipation. The Cartesian velocities and pressure are calculated and stored at the cell center, whereas contravariant fluxes are stored and calculated at cell faces. The discretized continuity and momentum equations are integrated using a projection method. The subgrid stresses are modeled using the Dynamic Smagorinksy Model (DSM).





































Figure 2. Mean streamline distribution at a plane parallel to ribbed wall, 0.5D away (left) and mean streamline distribution in a pitch length (10) at the z symmetry plane (right).



  

















Figure 3. Coherent structures in the computational domain (inflow at top left)





​

Dispersed Phase: The dispersed phase is modeled by following the trajectories of individual particles in a Langrangian framework. The model is implemented in an unstructured multiblock, multiprocessor framework and validation in turbulent channel flow has been reported in our previous work.













































































Figure 4. Particle impingement recorded in simulations at the ribbed wall, at different sections, experimental on left and CFD on right  (a) 180 deg turn (b) mid-section (c) inlet/outlet region (d) side-wall in second pass near bend region



​

The study identifies damage prone regions in the two pass cooling duct with rib turbulators. Along with investigation of sand transport, effect of different particle sizes and collisions models is also studied. This information can help modify the geometry of the blade or location of film cooling holes to avoid hole blockage and degradation of heat transfer at the walls.





All the details are not presented on the website due to copyright and proprietary issues.