Sacha Parneix

Effect of Showerhead Injection on Superposition of Multi-Row Pressure Side Film Cooling With Fan Shaped Holes

In the present study the effect of approach boundary layer conditions on film cooling effectiveness superposition for pressure side fan shaped holes on a first stage vane is investigated. More particularly, the effect of showerhead cooling on the film effectiveness of downstream pressure side rows is addressed. The experimental test facility used is a continuously running , two passage, linear cascade wind tunnel equipped with a central vane and contoured side walls. Main stream stagnation conditions and pressure measurements on the vane external surface were taken to determine the isentropic Mach number distribution. The turbulence level generated with a bar grid is around 15%. Film cooling effectiveness has been determined with the narrow banded thermochromic liquid crystal steady state technique. Mainstream as well as coolant flow could be heated to shift the iso-temperature contours across the vane. Carbon dioxide was used as coolant gas to better match the density ratio in the experimental facility to engine conditions. The test model is a research vane equipped with four rows of cylindrical holes showerhead and three rows of fan shaped holes along the pressure side at typical inclination angles to the surface. The study incorporates single row blowing with the following row approach conditions: 1) no showerhead injection, 2) boundary layer trip, 3) isothermal showerhead blowing, and multi-row blowing with and without showerhead blowing. The results are used to investigate the applicability of the single row results superposition approach for multiple-row injection. Simulating the appropriate aerodynamic conditions during individual row measurements improves the superposition prediction in comparison to the multi-row results.Copyright

Prediction of pressure loss and heat transfer in internal cooling passages

Abstract: This paper reports CFD‐simulations of the turbulent flow, pressure loss and heat transfer occurring in ribbed passages. The channel section is rectangular, with an aspect ratio of 2.04. Ribs are square cross‐section, their height is 10% of the channel height, and their inclination is varied from 90° to 33°. Reynolds number is 30,000. Three turbulence models (k‐ε wall functions and 2‐layer, V2F) are used and compared to the experimental data of Cho et al. 1 . All three models accurately predict the pressure losses due to the ribs and the qualitative heat transfer distribution on the ribbed wall. However, only the V2F model can accurately reproduce the absolute heat transfer levels, this at all inclination angles. The correlation developed by Han and co‐workers for smaller rib‐heights under‐predicts the friction factor and wall heat transfer level on the current configuration. This shows the danger of using a correlation outside of its application range.

Predictions of External Heat Transfer for Turbine Vanes and Blades With Secondary Flowfields

Detailed heat transfer distributions on the endwall and along the vane/blade surface are essential for component mechanical integrity and life predictions. Due to secondary flows, high gradients in heat transfer are present at the endwall and at the vane or blade surface itself where the passage vortex influences the mainstream flow. This paper documents the benchmarking of three turbulence models; (1) k-e realizable with wall functions (2) k-e realizable with two layer model, and (3) the V2F model for endwall and surface heat transfer and flowfield predictions. Benchmark experimental data from a scaled-up low speed rig for both a stator and rotor geometry are used for comparisons of heat transfer and flowfield. While the k-e realizable turbulence models give a good prediction of the secondary flow pattern, the heat transfer at the endwall and at the surface is not well predicted due to the inadequate modeling of near wall turbulence. The V2F model gives better agreement with the experiments on the endwall and vane midspan heat transfer is also well predicted, although transition occurs too far upstream on the suction surface. The results from this study represent the feasibility of CFD utilization as a predictive tool for local heat transfer distributions on a vane/blade endwall.Copyright