Blade tip heat transfer

In recent years, considerable effort has been made in enhancing understanding and predictive capability for high pressure turbine shroudless blade tip heat transfer by the group (Zhang, O’Dowd, Virdi and He) based on the high speed linear cascade (HSLC). The experimental campaign is closely coupled with CFD (HYDRA) for ‘two-way’ validations and insightful analyses (the biggest lesson learned has been that neither side should be taken for granted!). The findings have been much more far reaching than initially expected.

Conjugate Heat Transfer

An essential element of an adequate HP turbine design  is to be able to predict the blade metal temperatures. This involves convection heat transfer in the fluid part and conduction heat transfer in the solid part. The coupled convection-conduction conjugate heat transfer (CHT) modeling is important not only to blade heat transfer, but also to aerothermal performance prediction in general.  A fluid-only solution can become untenable due to the uncertain boundary condition for the energy equation when surface temperature is unknown.

Development of CHT has been so far largely confined to steady flows. The key challenge is the time scale disparity.  In a recent work (He and Oldfield, 2011), a new approach to CHT for periodic unsteady flows is proposed. To realign the greatly mismatched time scales, a hybrid approach of coupling between time-domain fluid solution and frequency-domain solid conduction is adopted. A novel harmonic transfer function is introduced, leading to an efficient and consistent new semi-analytical framework for the wall boundary condition of  unsteady energy equation under  periodic or even non-periodic (e.g. turbulence) disturbances with LES  (He, 2013).

Immersed Mesh Blocks (IMB) for film cooling

The development is motivated by the recognition that there is a huge length scale difference between the mesh resolution for an aerodynamic design and that for a cooling design. Given a common sequence of cooling design follows an aerodynamic design, it would make sense that cooling design could be based on the same baseline aerodynamic mesh for consistence. This is the basic consideration for developing the cooling resolved fine mesh blocks which can be easily ‘super-imposed’ (in the meshing part) on the base-aerodynamic mesh. This itself is similar to the general overset gridding approach. The distinctive features associated with ‘blocking’ for turbine film cooling are:

a) Cooling hole geometries very similar to each other. Hence multi-hole in one row can be reproduced by simple multiplication. In fact, the same mesh blocking topology can be easily deformed to fit into holes with slightly different geometries.

b) Easy adding, removing and moving blocks on a blade surface can be easily done without going back to the solid modeler and mesh generator, which often tends to be source of problems/bottle-necking.