Heat Transfer in an Airfoil Trailing Edge Configuration With Shaped Pedestals Mounted Internal Cooling Channel and Pressure Side Cutback

Described in this paper is an experimental study of heat transfer over a trailing edge configuration preceded with an internal cooling channel of pedestal array. The pedestal array consists of both circular pedestals and oblong shaped blocks. Downstream to the pedestal array, the trailing edge features pressure side cutback partitioned by the oblong shaped blocks. The local heat transfer coefficient over the entire wetted surface in the internal cooling chamber has been determined by using a “hybrid” measurement technique based on transient liquid crystal imaging. The hybrid technique employs the transient conduction model in a semi-infinite solid for resolving the heat transfer coefficient on the endwall surface uncovered by the pedestals. The heat transfer coefficient over a pedestal can be resolved by the lumped capacitance method with an assumption of low Biot number. The overall heat transfer for both the pedestals and endwalls combined shows a significant enhancement compared to the case with thermally developed smooth channel. Near the downstream most section of the suction side, the land, due to pressure side cutback, is exposed to the stream mixed with hot gas and discharged coolant. Both the adiabatic effectiveness and heat transfer coefficient on the land section are characterized by using the transient liquid crystal technique.

Automatic Optimisation of Pre-Swirl Nozzle Design

The objective of the research described here is to develop and demonstrate use of automatic design methods for pre-swirl nozzles. Performance of pre-swirled cooling air delivery systems depends critically on the design of these nozzles which is subject to manufacturing and stress constraints. The best solution may be a compromise between cost and performance. Here it is shown that automatic optimisation using computational fluid dynamics (CFD) to evaluate nozzle performance can be useful in design. A parametric geometric model of a nozzle with appropriate constraints is first defined and the CFD meshing and solution are then automated. The mesh generation is found to be the most delicate task in the whole process. Direct hill climbing (DHC) and response surface model (RSM) optimisation methods have been evaluated. For the test case considered, significant nozzle performance improvements were obtained using both methods, but the RSM model was preferred.

An Experimental and Numerical Study of Heat Transfer and Pressure Loss in a Rectangular Channel With V-Shaped Ribs

An experimental and numerical study was conducted to determine the thermal performance of V-shaped ribs in a rectangular channel with an aspect ratio of 2:1. Local heat transfer coefficients were measured using the steady state thermochromic liquid crystal technique. Periodic pressure losses were obtained with pressure taps along the smooth channel sidewall. Reynolds numbers from 95,000 to 500,000 were investigated with V-shaped ribs located on one side or on both sides of the test channel. The rib height-to-hydraulic diameter ratios (e/Dh) were 0.0625 and 0.02, and the rib pitch-to-height ratio (P/e) was 10. In addition, all test cases were investigated numerically. The commercial software FLUENT™ was used with a two-layer k-ε turbulence model. Numerically and experimentally obtained data were compared. It was determined that the heat transfer enhancement based on the heat transfer of a smooth wall levels off for Reynolds numbers over 200,000. The introduction of a second ribbed sidewall slightly increased the heat transfer enhancement whereas the pressure penalty was approximately doubled. Diminishing the rib height at high Reynolds numbers had the disadvantage of a slightly decreased heat transfer enhancement, but benefits in a significantly reduced pressure loss. At high Reynolds numbers small-scale ribs in a one-sided ribbed channel were shown to have the best thermal performance.

3D Flow Prediction and Improvement of Holes Arrangement of a Film-Cooled Turbine Blade Using a Feature-Based Jet Model

A feature-based jet model has been proposed for use in 3D CFD prediction of turbine blade film cooling. The goal of the model is to be able to perform computationally efficient flow prediction and optimization of film-cooled turbine blades. The model reproduces in the near hole region the macro flow features of a coolant jet within a Reynolds-Averaged Navier Stokes (RANS) framework. Numerical predictions of the 3D flow through a linear transonic film-cooled turbine cascade are carried out with the model, with a low computational overhead. Different cooling holes arrangement are computed and the prediction accuracy is evaluated versus experimental data. It shown that the present model provides a reasonably good prediction of the adiabatic film-cooling effectiveness and Nusselt number around the blade. A numerical analysis of the interaction of coolant jets issuing from different rows of holes on the blade pressure side is carried out. It is shown that the upward radial migration of the flow due to the passage secondary flow structure has an impact on the spreading of the coolant and the film cooling effectiveness on the blade surface. Based on this result, a new arrangement of the cooling holes for the present case is proposed that leads to a better spanwise covering of the coolant on the blade pressure side surface.

LES and RANS Investigations Into Buoyancy-Affected Convection in a Rotating Cavity With a Central Axial Throughflow

The buoyancy-affected flow in rotating disc cavities, such as occurs in compressor disc stacks, is known to be complex and difficult to predict. In the present work large eddy simulation (LES) and unsteady Reynolds-averaged Navier-Stokes (RANS) solutions are compared with other workers’ measurements from an engine representative test rig. The Smagorinsky-Lilly model was employed in the LES simulations, and the RNG k-ε turbulence model was used in the RANS modelling. Three test cases were investigated in a range of Grashof number Gr = 1.87 to 7.41×108 and buoyancy number Bo = 1.65 to 11.5. Consistent with experimental observation, strong unsteadiness was clearly observed in the results of both models, however the LES results exhibited a finer flow structure than the RANS solution. The LES model also achieved significantly better agreement with velocity and heat transfer measurements than the RANS model. Also, temperature contours obtained from the LES results have a finer structure than the tangential velocity contours. Based on the results obtained in this work, further application of LES to flows of industrial complexity is recommended.

Physical Interpretation of Flow and Heat Transfer in Pre-Swirl Systems

This paper compares heat transfer measurements from a pre-swirl rotor-stator experiment with 3D steady state results from a commercial CFD code. The measured distribution of Nusselt number on the rotor surface was obtained from a scaled model of a gas turbine rotor-stator system, where the flow structure is representative of that found in an engine. Computations were carried out using a coupled multigrid RANS solver with a high-Reynolds-number k-ε/k-ω turbulence model. Previous work has identified three parameters governing heat transfer: rotational Reynolds number (Reφ), pre-swirl ratio (βp) and the turbulent flow parameter (λT). For this study rotational Reynolds numbers are in the range 0.8×106 < Reφ < 1.2×106. The turbulent flow parameter and pre-swirl ratios varied between 0.12 < λT < 0.38 and 0.5 < βp < 1.5, which are comparable to values that occur in industrial gas turbines. At high coolant flow rates, computations have predicted peaks in heat transfer at the radius of the pre-swirl nozzles. These were discovered during earlier experiments and are associated with the impingement of the pre-swirl flow on the rotor disc. At lower flow rates, the heat transfer is controlled by boundary-layer effects. The Nusselt number on the rotating disc increases as either Reφ or λT increases, and is axisymmetric except in the region of the receiver holes, where significant two-dimensional variations are observed. The computed velocity field is used to explain the heat transfer distributions observed in the experiments. The regions of peak heat transfer around the receiver holes are a consequence of the route taken by the flow. Two routes have been identified: “direct”, whereby flow forms a stream-tube between the inlet and outlet; and “indirect”, whereby flow mixes with the rotating core of fluid. Two performance parameters have been calculated: the adiabatic effectiveness for the system, Θb,ab, and the discharge coefficient for the receiver holes, CD. The computations show that, although Θb,ab increases monotonically as βp increases, there is a critical value of βp at which CD is a maximum.

Assessment of the Impact of Laminar-Turbulent Transition on the Accuracy of Heat Transfer Coefficient Prediction in High Pressure Turbines

This paper presents a thorough assessment for two of the contemporary CFD programs available for modeling and predicting nonfilm-cooled surface heat transfer distributions on turbine airfoil surfaces. The CFD programs are capable of predicting laminar-turbulent transition and have been evaluated and validated against five test cases with experimental data. The suite of test cases considered for this study consists of two flat plat cases at zero and non-zero pressure gradient and three linear-turbine-cascade test cases that are representative of modern high pressure turbine designs. The flat plate test cases are the ERCOFTAC T3A and T3C2, while the linear turbine cascade cases are the MARKII, the Virginia Polytechnic Institute (VPI), and the Von Karman Institute (VKI) turbine cascades. The numerical tools assessed in this study are 3D viscous Reynolds Averaged-Navier-Stokes (RANS) equations programs that employ a variety of one-equation and two-equation models for turbulence closure. The assessment study focuses on the one-equation Spalart and Allmaras and the two-equation shear stress transport K-ω turbulence models with the ability of modeling and predicting laminar-turbulent transition. The RANS 3D viscous codes are Numeca’s Fine Turbo and ANSYS-CFX’ CFX5. Numerical results for skin friction, surface temperature distribution and heat transfer coefficient from the CFD programs are compared to measured experimental data. Sensitivity of the predictions to free stream turbulence and to inlet turbulence boundary conditions is also presented. The results of the study clearly illustrate the superiority of using the laminar-turbulent transition prediction in improving the accuracy of predicting the heat transfer coefficient on the surfaces of high pressure turbine airfoils.

Interactions Between Propagating Wakes and Flow Instabilities in the Presence of a Laminar Separation Bubble

Laminar separation was investigated experimentally on a flat plate under a strongly diffusing self-similar pressure distribution. This gave a long and thin laminar separation bubble. Boundary layer velocity traverses were performed at numerous longitudinal stations. Using a single hot wire a combination of individual traces, phase averaging and time averaging was used. To supplement this, an array of microphones was installed to give instantaneous contours of pressure perturbation and to investigate the time dependent flow features. Microphone data were consistent with the strong amplification, under the adverse pressure gradient, of instabilities predicted far upstream of the separation point. Driven at the Tollmien-Schlichting (T-S) frequency, these instabilities grew into turbulent spots developing in the shear layer of the separation bubble. Reattachment of the bubble was caused by transition of the separated shear layer. The waves were strongest in the later stages of transition. Once the coherence was lost, in a turbulent layer, the amplitude became diminished. Wake disturbances were injected into the flow and traced through the flow field. The wake interaction resulted in turbulent patches which penetrated to the wall. Following the patches was the calmed region, detectable as a region of reduced wave activity in the transition region following each turbulent strip. For a short time at the end of the calmed region the viscous instability waves continued to propagate for a considerable distance downstream, in concert with the calmed region, in an otherwise turbulent zone.

The Influence of Heat Transfer Effects on Turbine Performance Characteristics

A method allowing the evaluation of the effects related to heat transfer to the turbine blades on its performance characteristics is presented. The effects investigated are the change of passage dimensions, resulting from heat transfer and the change in flow field, exhibited mainly as a different boundary layer development. Change of hot gas temperature combined with cooling air temperature and possibly flow rate, result in a change of the temperature of the blade material, leading to dimension changes, because of the thermal expansion (dilatation). The changes in dimensions have a direct effect on turbine performance. An immediate consequence is a modification of the mass flow characteristic, due to a change of the throat area. Heat transfer also influences the properties of the gas flowing through the passage and in particular the characteristics of the boundary layers developing on the nozzle vanes and hub, tip endwals. Change of the thickness of this layer results in a change of blockage through the passage, a fact that influences directly the turbine flow function. The influence of both effects on turbine performance is studied. The study is performance oriented, aiming to the derivation of simplified models, which can be introduced in engine cycle decks.

Blade Tip Heat Transfer and Aerodynamics in a Large Scale Turbine Cascade With Moving Endwall

High resolution Nusselt number (Nu) distributions were measured on the blade tip surface of a large, 1.0 meter-chord, low-speed cascade representative of a high-pressure turbine. Data was obtained at a Reynolds number of 4.0 × 105 based on exit velocity and blade axial chord. Tip clearance levels ranged from 0.56% to 1.68% design span or equally from 1% to 3% of blade chord. An infrared camera, looking through the hollow blade, made detailed temperature measurements on a constant heat flux tip surface. The relative motion between the endwall and the blade tip was simulated by a moving belt. The moving belt endwall significantly to shifts the region of high Nusselt number distribution and reduces the overall averaged Nusselt number on the tip surface by up to 13.3%. The addition of a suction side squealer tip significantly reduced local tip heat transfer and resulted in a 32% reduction in averaged Nusselt number. Analysis of pressure measurements on the blade airfoil surface and tip surface along with PIV velocity flow fields in the gap give an understanding of the heat transfer mechanism.