0

IN THIS ISSUE

### Research Papers

J. Turbomach. 2018;140(11):111001-111001-10. doi:10.1115/1.4041374.

Gas turbine components subjected to high temperatures can benefit from improved designs enabled by metal additive manufacturing (AM) with nickel alloys. Previous studies have shown that the impact on fluid flow and heat transfer resulting from surface roughness of additively manufactured parts is significant; these impacts must be understood to design turbine components successfully for AM. This study improves understanding of these impacts by examining the discharge coefficient and the effect of the coolant delivery direction on the performance of additively manufactured shaped film cooling holes. To accomplish this, five test coupons containing a row of baseline shaped film cooling holes were made from a high-temperature nickel alloy using a laser powder bed fusion (L-PBF) process. Flow and pressure drop measurements across the holes were collected to determine the discharge coefficient from the film cooling holes. Temperature measurements were collected to assess the overall effectiveness of the coupon surface as well as the cooling enhancement due to film cooling. The Biot number of the coupon wall was matched to a value one might find in a turbine engine to ensure this data is relevant. It was discovered that the flow experienced greater aerodynamic losses in film cooling holes with greater relative roughness, which resulted in a decreased discharge coefficient. The effectiveness measurements showed that the film cooling performance is better when coolant is fed in a co-flow configuration compared to a counter-flow configuration.

Commentary by Dr. Valentin Fuster
J. Turbomach. 2018;140(11):111002-111002-12. doi:10.1115/1.4041378.

Due to the low level of profile losses reached in low-pressure turbines (LPT) for turbofan applications, a renewed interest is devoted to other sources of loss, e.g., secondary losses. At the same time, the adoption of high-lift profiles has reinforced the importance of these losses. A great attention, therefore, is dedicated to reliable prediction methods and to the understanding of the mechanisms that drive the secondary flows. In this context, a numerical and experimental campaign on a state-of-the-art LPT cascade was carried out focusing on the impact of different inlet boundary layer (BL) profiles. First of all, detailed Reynolds Averaged Navier-Stokes (RANS) analyzes were carried out in order to establish dependable guidelines for the computational setup. Such analyzes also underlined the importance of the shape of the inlet BL very close to the endwall, suggesting tight requirements for the characterization of the experimental environment. The impact of the inlet BL on the secondary flow was experimentally investigated by varying the inlet profile very close to the endwall as well as on the external part of the BL. The effects on the cascade performance were evaluated by measuring the span-wise distributions of flow angle and total pressure losses. For all the inlet conditions, comparisons between Computational Fluid Dynamics (CFD) and experimental results are discussed. Besides providing guidelines for a proper numerical and experimental setup, the present paper underlines the importance of a detailed characterization of the inlet BL for an accurate assessment of the secondary flows.

Commentary by Dr. Valentin Fuster
J. Turbomach. 2018;140(11):111003-111003-11. doi:10.1115/1.4041494.

The degree of complexity in internal cooling designs is tied to the capabilities of the manufacturing process. Additive manufacturing (AM) grants designers increased freedom while offering adequate reproducibility of microsized, unconventional features that can be used to cool the skin of gas turbine components. One such desirable feature can be sourced from nature; a common characteristic of natural transport systems is a network of communicating channels. In an effort to create an engineered design that utilizes the benefits of those natural systems, the current study presents wavy microchannels that were connected using branches. Two different wavelength baseline configurations were designed; then each was numerically optimized using a commercial adjoint-based method. Three objective functions were posed to (1) minimize pressure loss, (2) maximize heat transfer, and (3) maximize the ratio of heat transfer to pressure loss. All baseline and optimized microchannels were manufactured using laser powder bed fusion (L-PBF) for experimental investigation; pressure loss and heat transfer data were collected over a range of Reynolds numbers. The AM process reproduced the desired optimized geometries faithfully. Surface roughness, however, strongly influenced the experimental results; successful replication of the intended flow and heat transfer performance was tied to the optimized design intent. Even still, certain test coupons yielded performances that correlated well with the simulation results.

Commentary by Dr. Valentin Fuster
J. Turbomach. 2018;140(11):111004-111004-12. doi:10.1115/1.4041514.

In turbines, secondary vortices and tip leakage vortices form in the blade passage and interact with each other. In order to understand the flow physics of this vortices interaction, the effects of incoming vortex on the downstream tip leakage flow are investigated by experimental, numerical, and analytical methods. In the experiment, a swirl generator was used upstream of a linear turbine cascade to generate the incoming vortex, which could interact with the downstream tip leakage vortex (TLV). The swirl generator was located at ten different pitchwise locations to simulate the quasi-steady effects. In the numerical study, a Rankine-like vortex was defined at the inlet of the computational domain to simulate the incoming swirling vortex (SV). The effects of the directions of the incoming vortices were investigated. In the case of a positive incoming SV, which has a large vorticity vector in the same direction as that of the TLV, the vortex mixes with the TLV to form one major vortex near the casing as it transports downstream. This vortices interaction reduces the loss by increasing the streamwise momentum within the TLV core. However, the negative incoming SV has little effects on the TLV and the loss. As the negative incoming SV transports downstream, it travels away from the TLV and two vortices can be identified near the casing. A triple-vortices-interaction kinetic model is used to explain the flow physics of vortex interaction, and a one-dimensional mixing analytical model are proposed to explain the loss mechanism.

Commentary by Dr. Valentin Fuster
J. Turbomach. 2018;140(11):111005-111005-17. doi:10.1115/1.4040419.

The aim of this paper is to review, summarize, and record long-term experience with development and application of aerodynamic probes with built-in miniature pressure transducers for unsteady pressure measurement and industrial research in turbomachine components. The focus of the first half of the paper is on the work performed at VZLU Prague, Czech Republic (Secs. 3–8). The latest development in unsteady pressure measurement techniques and data reduction methodology suitable for future research in highly loaded, high-speed turbine engine components performed at NASA GRC Cleveland, OH, is reported in Secs. 8–15 of this paper. Excellent reviews of similar activities at ETH Zürich, Switzerland by Kupferschmied, et al. (2000, “Time-Resolved Flow Measurements With Fast-Response Aerodynamic Probes in Turbomachines,” Meas. Sci. Technol., 11(7), pp. 1036–1054) and at VKI Rhode-Sain-Genèse, Belgium by Sieverding, et al. (2000, “Measurement Techniques for Unsteady Flows in Turbomachines,” Exp. Fluids, 28(4), pp. 285–321) were already reported and are acknowledged here. A short list of reported accomplishments achieved by other researchers at various laboratories is also reported for completeness. The authors apologize to those whose contributions are not reported here. It is just an unfortunate oversight, not an intentional omission.

Commentary by Dr. Valentin Fuster
J. Turbomach. 2018;140(11):111006-111006-12. doi:10.1115/1.4041400.

Unshrouded industrial centrifugal compressor impellers operate at high rotational speeds and volume flow rates. Under such conditions, the main impeller blade excitation is dominated by high frequency interaction with stationary parts, i.e., vaned diffusers or inlet guide vanes (IGVs). However, at severe part load operating conditions, sub-synchronous rotating flow phenomena (rotating stall) can occur and cause resonant blade vibration with significant dynamic (von-Mises) stress in the impeller blades. To ensure high aerodynamic performance and mechanical integrity, part load conditions must be taken into account in the aeromechanical design process via computational fluid dynamics (CFD) and finite element method (FEM) analyzes anchored by experimental verification. The experimental description and quantification of unsteady interaction between rotating stall cells and an unshrouded centrifugal compression stage in two different full scale compression units by Jenny and Bidaut (“Experimental Determination of Mechanical Stress Induced by Rotating Stall in Unshrouded Impellers of Centrifugal Compressors”, ASME J. Turbomach. 2016; 139(3):031011-031011-10) were reproduced in a scaled model test facility to enhance the understanding of the fluid–structure interaction (FSI) mechanisms and to improve design guide lines. Measurements with strain gauges and time-resolved pressure transducers on the stationary and rotating parts at different positions identified similar rotating stall patterns and induced stress levels. Rotating stall cell induced resonant blade vibration was discovered for severe off-design operating conditions and the measured induced dynamic von-Mises stress peaked at 15% of the mechanical endurance limit of the impeller material. Unsteady full annulus CFD simulations predicted the same rotating stall pressure fluctuations as the measurements. The unsteady Reynold's Averaged Navier-Stokes simulations were then used in FEM FSI analyses to predict the stress induced by rotating stall and assess the aerodynamic damping of the corresponding impeller vibration mode shape. Excellent agreement with the measurements was obtained for the stall cell pressure amplitudes at various locations. The relative difference between measured and mean predicted stress from fluid–structure interaction was 17% when resonant blade vibration occurred. The computed aerodynamic damping was 27% higher compared to the measurement.

Commentary by Dr. Valentin Fuster
J. Turbomach. 2018;140(11):111007-111007-12. doi:10.1115/1.4041292.

Testing new turbine cooling schemes at engine conditions becomes cost prohibitive as gas-path temperatures increase. As a result, turbine components are simulated in a laboratory with a large-scale model that is sized and constructed out of a selected material so that the Biot number is matched between the laboratory and engine conditions. Furthermore, the experimental temperatures are lower, so the surface temperature that the metal component would experience is scaled via the overall cooling effectiveness, $ϕ$. Properly measuring $ϕ$ requires that the relevant flow physics must be matched, thus the Reynolds numbers is matched—both those of the freestream and the coolant, as well as the other scaling parameters, such as the mass flux, momentum flux, and velocity ratios. However, if the coolant-to-freestream density ratio does not match that of the engine condition, the mass flux, momentum flux, coolant and freestream Reynolds numbers, and coolant-to-freestream velocity ratios cannot be matched simultaneously to the engine condition. Furthermore, the coolant thermal transfer properties are unaccounted for in these parameters, despite their large influence on the resultant overall effectiveness. While much research has focused on the effects of the coolant-to-freestream density ratio, this study examines the influence of other thermodynamic properties, in particular the specific heat, which differ substantially between experimental and engine conditions. This study demonstrates the influence of various coolant properties on the overall effectiveness distribution on a leading edge by selectively matching $M$, $I$, and $ACR$ with air, argon, and carbon dioxide coolants.

Commentary by Dr. Valentin Fuster