Effects of Reynolds Number and Free-Stream Turbulence on Boundary Layer Transition in a Compressor Cascade

An experimental and analytical study has been performed on the effect of Reynolds number and free-stream turbulence on boundary layer transition location on the suction surface of a controlled diffusion airfoil (CDA). The experiments were conducted in a rectilinear cascade facility at Reynolds numbers between 0.7 and 3.0×106 and turbulence intensities from about 0.7 to 4%. An oil streak technique and liquid crystal coatings were used to visualize the boundary layer state. For small turbulence levels and all Reynolds numbers tested the accelerated front portion of the blade is laminar and transition occurs within a laminar separation bubble shortly after the maximum velocity near 35–40% of chord. For high turbulence levels (Tu > 3%) and high Reynolds numbers transition propagates upstream into the accelerated front portion of the CDA blade. For those conditions, the sensitivity to surface roughness increases considerably and at Tu = 4% bypass transition is observed near 7–10% of chord. Experimental results are compared to theoretical predictions using the transition model which is implemented in the MISES code of Youngren and Drela. Overall the results indicate that early bypass transition at high turbulence levels must alter the profile velocity distribution for compressor blades that are designed and optimized for high Reynolds numbers.

Natural Transition Phenomena on an Axial Compressor Blade

Data from a surface hot-film array on the outlet stator of a 1.5 stage axial compressor are analyzed to look for direct evidence of natural transition phenomena. An algorithm is developed to identify instability waves within the Tollmien Schlichting (T-S) frequency range. The algorithm is combined with a turbulent intermittency detection routine to produce space∼time diagrams showing the probability of instability wave occurrence prior to regions of turbulent flow. The paper compares these plots for a range of blade loading, with free-stream conditions corresponding to the maximum and minimum inflow disturbance periodicity produced by inlet guide vane clocking. Extensive regions of amplifying instability waves are identified in nearly all cases. The implications for transition prediction in decelerating flow regions on axial turbomachine blades are discussed.

Rotating Effect on Fluid Flow in a Smooth Duct With a 180-Deg Sharp Turn

Laser-Doppler velocimetry (LDV) measurements are presented of turbulent flow in a two-pass square-sectioned duct simulating the coolant passages employed in gas turbine blades under rotating and non-rotating conditions. For all cases studied, the Reynolds number characterized by duct hydraulic diameter (Dh) and bulk mean velocity (Ub) was fixed at 1 × 104. The rotating case had a range of rotation number (Ro = ΩDh/Ub) from 0 to 0.2. It is found that both the skewness of streamwise mean velocity and magnitude of secondary-flow velocity increase linearly, and the magnitude of turbulence intensity level increases non-linearly with increasing Ro. As Ro is increased, the curvature induced symmetric Dean vortices in the turn for Ro = 0 is gradually dominated by a single vortex most of which impinges directly on the outer part of leading wall. The high turbulent kinetic energy is closely related to the dominant vortex prevailing inside the 180-deg sharp turn. For the first time, the measured flow characteristics account for the reported spanwise heat transfer distributions in the rotating channels, especially the high heat transfer enhancement on the leading wall in the turn. For both rotating and non-rotating cases, the direction and strength of the secondary flow with respect to the wall are the most important fluid dynamic factors affecting local heat transfer distributions inside a 180-deg sharp turn. The role of the turbulent kinetic energy in affecting the overall enhancement of heat transfer is well addressed.

Heat Transfer and Flow on the Squealer Tip of a Gas Turbine Blade

Experimental investigations are performed to measure the detailed heat transfer coefficient and static pressure distributions on the squealer tip of a gas turbine blade in a five-bladed stationary linear cascade. The blade is a 2-dimensional model of a modern first stage gas turbine rotor blade with a blade tip profile of a GE-E3 aircraft gas turbine engine rotor blade. A squealer (recessed) tip with a 3.77% recess is considered here. The data on the squealer tip are also compared with a flat tip case. All measurements are made at three different tip gap clearances of about 1%, 1.5%, and 2.5% of the blade span. Two different turbulence intensities of 6.1% and 9.7% at the cascade inlet are also considered for heat transfer measurements. Static pressure measurements are made in the mid-span and near-tip regions, as well as on the shroud surface opposite to the blade tip surface. The flow condition in the test cascade corresponds to an overall pressure ratio of 1.32 and an exit Reynolds number based on the axial chord of 1.1×106. A transient liquid crystal technique is used to measure the heat transfer coefficients. Results show that the heat transfer coefficient on the cavity surface and rim increases with an increase in tip clearance. The heat transfer coefficient on the rim is higher than the cavity surface. The cavity surface has a higher heat transfer coefficient near the leading edge region than the trailing edge region. The heat transfer coefficient on the pressure side rim and trailing edge region is higher at a higher turbulence intensity level of 9.7% over 6.1% case. However, no significant difference in local heat transfer coefficient is observed inside the cavity and the suction side rim for the two turbulence intensities. The squealer tip blade provides a lower overall heat transfer coefficient when compared to the flat tip blade.

Flow Measurements in a Nozzle Guide Vane Passage With a Low Aspect Ratio and Endwall Contouring

The vast number of turbine cascade studies in the literature has been performed in straight-endwall, high-aspect-ratio, linear cascades. As a result, there has been little appreciation for the role of, and added complexity imposed by, reduced aspect ratios. There also has been little documentation of endwall profiling at these reduced spans. To examine the role of these factors on cascade hydrodynamics, a large-scale nozzle guide vane simulator was constructed at the Heat Transfer Laboratory of the University of Minnesota. This cascade is comprised of three airfoils between one contoured and one flat endwall. The geometries of the airfoils and endwalls, as well as the experimental conditions in the simulator, are representative of those in commercial operation. Measurements with hot-wire anemometry were taken to characterize the flow approaching the cascade. These measurements show that the flow field in this cascade is highly elliptic and influenced by pressure gradients that are established within the cascade. Exit flow field measurements with triple-sensor anemometry and pressure measurements within the cascade indicate that the acceleration imposed by endwall contouring and airfoil turning is able to suppress the size and strength of key secondary flow features. In addition, the flow field near the contoured endwall differs significantly from that adjacent to the straight endwall.

Local Heat Transfer and Flow Structure on and Above a Dimpled Surface in a Channel

Experimental results, measured on and above a dimpled test surface placed on one wall of a channel, are given for Reynolds numbers from 1,250 to 61,500 and ratios of air inlet stagnation temperature to surface temperature ranging from 0.68 to 0.94. These include flow visualizations, surveys of time-averaged total pressure and streamwise velocity, and spatially-resolved local Nusselt numbers, which are measured using infrared thermography, used in conjunction with energy balances, thermocouples, and in situ calibration procedures. The ratio of channel height to dimple print diameter is 0.5. Flow visualizations show vortical fluid and vortex pairs shed from the dimples, including a large upwash region and packets of fluid emanating from the central regions of each dimple, as well as vortex pairs and vortical fluid which form near dimple diagonals. These vortex structures augment local Nusselt numbers near the downstream rims of each dimple, both slightly within each depression, and especially on the flat surface just downstream of each dimple. Such augmentations are spread over larger surface areas and become more pronounced as the ratio of inlet stagnation temperature to local surface temperature decreases. As a result, local and spatially-averaged heat transfer augmentations become larger as this temperature ratio decreases. This is due to the actions of vortical fluid in advecting cool fluid from the central parts of the channel to regions close to the hotter dimpled surface.

Heat Transfer Analysis in a Modern DLN Combustor

The increasingly stringent emissions standards in recent years have mandated low gas turbine emissions and thus changed the approach to combustion chamber design. In particular, lean burners based on highly premixed fuel-air flows have become more important. These combustors, termed Dry Low NOx (DLN), can now achieve emissions of 25 ppm and below in commercial operation. This development together with the inlet turbine temperature increase has resulted in less cooling air for combustion chambers and turbine blade cooling systems. The designer now needs to optimise cooling air flows that control the wall temperature of the components that confine the hot gases. Moreover, much of the air coming from the compressor is used to premix the fuel and only a smaller fraction is now available for cooling processes. In annular combustor configurations the air available for cooling the combustion chamber walls sometimes also has to cool the first stage nozzle. So the pressure loss along the combustor cooling passages has to be limited in order to assure a suitable supply pressure for these downstream cooling passages. We analysed the cooling air flow around the liner of an annular combustion chamber and we investigated the thermal flux and friction losses. In this paper we show the development of a calculation model that allows the critical components heat transfer analysis of a typical annular combustion chamber. The code developed is based on the generalised 1–D flow treatment. We have used experimental correlations for convection, film cooling and impingement borrowed from works found in literature. The code is provided with a graphical interface that helps the user during the calculation. This code was used in practical application to optimize the PGT5B combustion chamber cooling.

Inlet Fogging of Gas Turbine Engines: Part A — Theory, Psychrometrics and Fog Generation

Gas Turbine output is a strong function of the ambient air temperature with power output dropping by 0.3–0.5 % for every 1°F rise in ambient temperature. This loss in output presents a significant problem to utilities, cogenerators and IPPs when electric demands are high during the hot months. In the petrochemical and process industry, the reduction in output of mechanical drive gas turbines curtails plant output. One way to counter this drop in output is to cool the inlet air. The paper contrasts the traditional evaporative cooling technique with direct inlet fogging. The state of the art relating to fog generation and psychrometrics of inlet fogging are described.

Turbine Separation Control Using Pulsed Vortex Generator Jets

The application of pulsed vortex generator jets to control separation on the suction surface of a low pressure turbine blade is reported. Blade Reynolds numbers in the experimental, linear turbine cascade match those for high altitude aircraft engines and aft stages of industrial turbine engines with elevated turbine inlet temperatures. The vortex generator jets have a 30 degree pitch and a 90 degree skew to the freestream direction. Jet flow oscillations up to 100 Hz are produced using a high frequency solenoid feed valve. Results are compared to steady blowing at jet blowing ratios less than 4 and at two chordwise positions upstream of the nominal separation zone. Results show that pulsed vortex generator jets produce a bulk flow effect comparable to that of steady jets with an order of magnitude less massflow. Boundary layer traverses and blade static pressure distributions show that separation is almost completely eliminated with the application of unsteady blowing. Reductions of over 50% in the wake loss profile of the controlled blade were measured. Experimental evidence suggests that the mechanism for unsteady control lies in the starting and ending transitions of the pulsing cycle rather than the injected jet stream itself. Boundary layer spectra support this conclusion and highlight significant differences between the steady and unsteady control techniques. The pulsed vortex generator jets are effective at both chordwise injection locations tested (45% and 63% axial chord) covering a substantial portion of the blade suction surface. This insensitivity to injection location bodes well for practical application of pulsed VGJ control where the separation location may not be accurately known a priori.

Flow and Heat Transfer in a Rotating Cavity With a Stationary Stepped Casing

In this paper, new experimental results are presented for the flow in a co-rotating disc system with a rotating inner cylinder and a stationary stepped outer casing. The configuration is based on a turbine disc-cooling system used in a gas turbine engine. One of the rotating discs can be heated, and cooling air is introduced through discrete holes angled inward at the periphery of this disc. The cooling air leaves the system through axial clearances between the discs and the outer casing. Some features of computed flows, and both measured and computed heat transfer, were reported previously for this system. New velocity measurements, obtained using Laser Doppler Anemometry, are compared with results from axisymmetric, steady, turbulent flow computations obtained using a low-Reynolds-number k-ε turbulence model. The measurements and computations show that the tangential component of velocity is invariant with axial location in much of the cavity, and the data suggest that Rankine (combined free and forced) vortex flow occurs. The computations fail to reproduce this behaviour, and there are differences between measured and computed details of secondary flow recirculations. Possible reasons for these discrepancies, and their importance for the prediction of associated heat transfer, are discussed.