Journal of Turbomachinery

A computational study to define the phenomena that lead to the onset of short length-scale (spike) rotating stall disturbances has been carried out. Based on unsteady simulations, we hypothesize there are two conditions necessary for the formation of spike disturbances, both of which are linked to the tip clearance flow. One is that the interface between the tip clearance and oncoming flows becomes parallel to the leading-edge plane. The second is the initiation of backflow, stemming from the fluid in adjacent passages, at the trailing-edge plane. The two criteria also imply a circumferential length scale for spike disturbances. The hypothesis and scenario developed are consistent with numerical simulations and experimental observations of axial compressor stall inception. A comparison of calculations for multiple blades with those for single passages also allows statements to be made about the utility of single passage computations as a descriptor of compressor stall.

Work on rotating stall and its related disturbances have been in progress since the Second World War. During this period, certain “hot topics” have come to the fore—mostly in response to pressing problems associated with new engine designs. This paper will take a semihistorical look at some of these fields of study (stall, surge, active control, rotating instabilities, etc.) and will examine the ideas which underpin each topic. Good progress can be reported, but the paper will not be an unrestricted celebration of our successes because, after 75 years of research, we are still unable to predict the stalling behavior of a new compressor or to contribute much to the design of a more stall-resistant machine. Looking forward from where we are today, it is clear that future developments will come from CFD in the form of better performance predictions, better flow modeling, and improved interpretation of experimental results. It is also clear that future experimental work will be most effective when focussed on real compressors with real problems—such as stage matching, large tip clearances, eccentricity, and service life degradation. Today’s topics of interest are mostly associated with compressible effects and so further research will require more high-speed testing.

Experimental investigations have shown that the enhancement in heat transfer coefficients for air flow in a channel roughened with angled ribs is on the average higher than that roughened with 90 deg ribs of the same geometry. Secondary flows generated by the angled ribs are believed to be responsible for these higher heat transfer coefficients. These secondary flows also create a spanwise variation in heat transfer coefficient on the roughened wall with high levels of heat transfer coefficient at one end of the rib and low levels at the other end. In an effort basically to double the area of high heat transfer coefficients, the angled rib is broken at the center to form a V-shaped rib, and tests are conducted to investigate the resulting heat transfer coefficients and friction factors. Three different square rib geometries, corresponding to blockage ratios of 0.083, 0.125, and 0.167, with a fixed pitch-to-height ratio of 10, mounted on two opposite walls of a square channel in a staggered configuration, are tested in a stationary channel for 5000 < Re < 30,000. Heat transfer coefficients, friction factors, and thermal performances are compared with those of 90 deg, 45 deg, and discrete angled ribs. The V-shaped ribs are tested for both pointing upstream and downstream of the main flow. Test results show that: (a) 90 deg ribs represent the lowest thermal performance, based on the same pumping power, and is essentially the same for the 2:1 change in blockage ratio, (b) low-blockage-ratio (e/Dh = 0.083) V-shaped ribs pointing downstream produced the highest heat transfer enhancement and friction factors. Among all other geometries with blockage ratios of 0.125 and 0.167, 45 deg ribs showed the highest heat transfer enhancements with friction factors less than those of V-shaped ribs, (c) thermal performance of 45 deg ribs and the lowest blockage discrete ribs are among the highest of the geometries tested in this investigation, and (d) discrete angled ribs, although inferior to 45 deg and V-shaped ribs, produce much higher heat transfer coefficients and lower friction factors compared to 90 deg ribs.

Film cooling effectiveness was studied experimentally in a flat plate test facility with zero pressure gradient using a single row of inclined holes, which injected high-density, cryogenically cooled air. Round holes and holes with a diffusing expanded exit were directed laterally away from the free-stream direction with a compound angle of 60 deg. Comparisons were made with a baseline case of round holes aligned with the free stream. The effects of doubling the hole spacing to six hole diameters for each geometry were also examined. Experiments were performed at a density ratio of 1.6 with a range of blowing ratios from 0.5 to 2.5 and momentum flux ratios from 0.16 to 3.9. Lateral distributions of adiabatic effectiveness results were determined at streamwise distances from 3 D to 15 D downstream of the injection holes. All hole geometries had similar maximum spatially averaged effectiveness at a low momentum flux ratio of I = 0.25, but the round and expanded exit holes with compound angle had significantly greater effectiveness at larger momentum flux ratios. The compound angle holes with expanded exits had a much improved lateral distribution of coolant near the hole for all momentum flux ratios.

Experimental measurements in a linear cascade with tip clearance are complemented by numerical solutions of the three-dimensional Navier–Stokes equations in an investigation of tip leakage flow. Measurements reveal that the clearance flow, which separates near the entry of the tip gap, remains unattached for the majority of the blade chord when the tip clearance is similar to that typical of a machine. The numerical predictions of leakage flow rate agree very well with measurements, and detailed comparisons show that the mechanism of tip leakage is primarily inviscid. It is demonstrated by simple calculation that it is the static pressure field near the end of the blade that controls chordwise distribution of the flow across the tip. Although the presence of a vortex caused by the roll-up of the leakage flow may affect the local pressure field, the overall magnitude of the tip leakage flow remains strongly related to the aerodynamic loading of the blades.

This paper presents detailed measurements of the film-cooling effectiveness for three single, scaled-up film-cooling hole geometries. The hole geometries investigated include a cylindrical hole and two holes with a diffuser-shaped exit portion (i.e., a fan-shaped and a laid-back fan-shaped hole). The flow conditions considered are the crossflow Mach number at the hole entrance side (up to 0.6), the crossflow Mach number at the hole exit side (up to 1.2), and the blowing ratio (up to 2). The coolant-to-mainflow temperature ratio is kept constant at 0.54. The measurements are performed by means of an infrared camera system, which provides a two-dimensional distribution of the film-cooling effectiveness in the near field of the cooling hole down to x/D = 10. As compared to the cylindrical hole, both expanded holes show significantly improved thermal protection of the surface downstream of the ejection location, particularly at high blowing ratios. The laidback fan-shaped hole provides a better lateral spreading of the ejected coolant than the fan-shaped hole, which leads to higher laterally averaged film-cooling effectiveness. Coolant passage cross-flow Mach number and orientation strongly affect the flowfield of the jet being ejected from the hole and, therefore, have an important impact on film-cooling performance.

Flow separations in the corner regions of blade passages are common. The separations are three dimensional and have quite different properties from the two-dimensional separations that are considered in elementary courses of fluid mechanics. In particular, the consequences for the flow may be less severe than the two-dimensional separation. This paper describes the nature of three-dimensional (3D) separation and addresses the way in which topological rules, based on a linear treatment of the Navier-Stokes equations, can predict properties of the limiting streamlines, including the singularities which form. The paper shows measurements of the flow field in a linear cascade of compressor blades and compares these to the results of 3D computational fluid dynamics (CFD). For corners without tip clearance, the presence of three-dimensional separation appears to be universal, and the challenge for the designer is to limit the loss and blockage produced. The CFD appears capable of predicting this.

The paper describes a method of computing nonequilibrium, steady flows of wet steam in two- and quasi-three-dimensional turbine cascades. The mixture conservation equations are solved in an Eulerian reference frame using an inviscid time-marching method that includes the effects of the centrifugal and Coriolis acceleration terms in rotating blade rows. Nucleation and growth of water droplets are computed by integrating the relevant equations along true streamlines in a Lagrangian reference frame. Steam properties are computed using equations that display commercial steam table accuracy for pressures below 10 bar. Special procedures for grouping the range of droplet sizes present are described that allow an accurate representation of the droplet size distribution to be retained without requiring a large increase in CPU time. All types of wet-steam flow, including those involving secondary nucleation, can be computed. Examples are presented that display the sensitivity of the calculation procedure in computing nucleation affected by the shock- and expansion-wave structure in the region of a turbine blade trailing edge. Typical CPU time requirements for nonequilibrium solutions involving primary or secondary nucleations are about three times those for perfect gas calculations.

A three-dimensional code for rotating blade-row flow analysis has been developed. The space discretization uses a cell-centered scheme with eigenvalue scaling for the artificial dissipation. The computational efficiency of a four-stage Runge–Kutta scheme is enhanced by using variable coefficients, implicit residual smoothing, and a full-multigrid method. An application is presented for the NASA rotor 67 transonic fan. Due to the blade stagger and twist, a zonal, nonperiodic H-type grid is used to minimize the mesh skewness. The calculation is validated by comparing it with experiments in the range from the maximum flow rate to a near-stall condition. A detailed study of the flow structure near peak efficiency and near stall is presented by means of pressure distribution and particle traces inside boundary layers.

Recent technological advances in the field of additive manufacturing (AM), particularly with direct metal laser sintering (DMLS), have increased the potential for building gas turbine components with AM. Using the DMLS for turbine components broadens the design space and allows for increasingly small and complex geometries to be fabricated with little increase in time or cost. Challenges arise when attempting to evaluate the advantages of the DMLS for specific applications, particularly because of how little is known regarding the effects of surface roughness. This paper presents pressure drop and heat transfer results of flow through small, as produced channels that have been manufactured using the DMLS in an effort to better understand roughness. Ten different coupons made with the DMLS all having multiple rectangular channels were evaluated in this study. Measurements were collected at various flow conditions and reduced to a friction factor and a Nusselt number. Results showed significant augmentation of these parameters compared to smooth channels, particularly with the friction factor for minichannels with small hydraulic diameters. However, augmentation of Nusselt number did not increase proportionally with the augmentation of the friction factor.