E. A. Grover

Measurements of Secondary Losses in a Turbine Cascade With the Implementation of Nonaxisymmetric Endwall Contouring

An approach to endwall contouring has been developed with the goal of reducing secondary losses in highly loaded axial flow turbines. The present paper describes an experimental assessment of the performance of the contouring approach implemented in a low-speed linear cascade test facility. The study examines the secondary flows of a cascade composed of Pratt & Whitney PAKB airfoils. This airfoil has been used extensively in low-pressure turbine research, and the present work adds intrapassage pressure and velocity measurements to the existing database. The cascade was tested at design incidence and at an inlet Reynolds number of 126,000 based on inlet midspan velocity and axial chord. Quantitative results include seven-hole pneumatic probe pressure measurements downstream of the cascade to assess blade row losses and detailed seven-hole probe measurements within the blade passage to track the progression of flow structures. Qualitative results take the form of oil surface flow visualization on the endwall and blade suction surface. The application of endwall contouring resulted in lower secondary losses and a reduction in secondary kinetic energy associated with pitchwise flow near the endwall and spanwise flow up the suction surface within the blade passage. The mechanism of loss reduction is discussed in regard to the reduction in secondary kinetic energy.

Application of Nonaxisymmetric Endwall Contouring to Conventional and High-Lift Turbine Airfoils

Here, we report on the application of nonaxisymmetric endwall contouring to mitigate the endwall losses of one conventional and two high-lift low-pressure turbine airfoil designs. The design methodology presented combines a gradient-based optimization algorithm with a three-dimensional computational fluid dynamics (CFD) flow solver to systematically vary a free-form parameterization of the endwall. The ability of the CFD solver employed in this work to predict endwall loss modifications resulting from nonaxisymmetric contouring is demonstrated with previously published data. Based on the validated trend accuracy of the solver for predicting the effects of endwall contouring, the magnitude of predicted viscous losses forms the objective function for the endwall design methodology. This system has subsequently been employed to optimize contours for the conventional-lift Pack B and high-lift Pack D-F and Pack D-A low-pressure turbine airfoil designs. Comparisons between the predicted and measured loss benefits associated with the contouring for Pack D-F design are shown to be in reasonable agreement. Additionally, the predictions and data demonstrate that the Pack D-F endwall contour is effective at reducing losses primarily associated with the passage vortex. However, some deficiencies in predictive capabilities demonstrated here highlight the need for a better understanding of the physics of endwall loss-generation and improved predictive capabilities.

Predicting Transition in Turbomachinery—Part II: Model Validation and Benchmarking

The ability to predict boundary layer transition locations accurately on turbomachinery airfoils is critical both to evaluate aerodynamic performance and to predict local heat-transfer coefficients with accuracy. Here we report on an effort to include empirical transition models developed in Part I of this report in a Reynolds averaged Navier-Stokes (RANS) solver. To validate the new models, two-dimensional design optimizations utilizing transitional RANS simulations were performed to obtain a pair of low-pressure turbine airfoils with the objective of increasing airfoil loading by 25%. Subsequent experimental testing of the two new airfoils confirmed pre-test predictions of both high and low Reynolds number loss levels. In addition, the accuracy of the new transition modeling capability was benchmarked with a number of legacy cascade and low-pressure turbine (LPT) rig data sets. Good agreement between measured and predicted profile losses was found in both cascade and rig environments. However, use of the transition modeling capability has elucidated deficiencies in typical RANS simulations that are conducted to predict component performance. Efficiency-versus-span comparisons between rig data and multi-stage steady and time-accurate LPT simulations indicate that loss levels in the end wall regions are significantly under predicted. Possible causes for the under-predicted end wall losses are discussed as well as suggestions for future improvements that would make RANS-based transitional simulations more accurate.

Comparative Investigation of Three Highly Loaded LP Turbine Airfoils: Part II — Measured Profile and Secondary Losses at Off-Design Incidence

At the 2006 ASME-IGTI Turbo-Expo, low-speed cascade results were presented for the midspan aerodynamic behaviour of a family of three highly loaded low-pressure (LP) turbine airfoils operating over a wide range of Reynolds numbers (25,000 to 150,000 based on the axial chord and inlet velocity), and for values of freestream turbulence intensity of 1.5% and 4%. All three airfoils have the same design inlet and outlet flow angles. The baseline cascade has a Zweifel coefficient of 1.08 and the two additional blade rows have values of 1.37. The new, more highly-loaded blade rows differ mainly in their loading distributions: one is front-loaded while the other is aft-loaded. The new front-loaded airfoil was found to have particularly attractive profile performance. Despite its exceptionally high value of Zweifel coefficient, it was found to be free of a separation bubble on its suction side at Reynolds numbers as low as 50,000, and this was reflected in very good profile loss behaviour. However, it was also noted in the earlier paper that the choice of a particular loading level and loading distribution would be influenced by more than its profile performance at design incidence. The present two-part paper extends the midspan aerodynamic comparison of the three airfoils to the secondary flow performance. The first part of the paper discusses both the profile and secondary flow performance of the three cascades at their design Reynolds number of 80,000 (or ∼ 125,000 based on exit velocity) for two freestream turbulence intensities of 1.5% and 4%. The secondary flow behaviour was determined from detailed flowfield measurements made at 40% axial chord downstream of the trailing edge using a seven-hole pressure probe. In addition to providing total pressure losses, the seven-hole probe measurements were also processed to give the downstream vorticity distributions. As has been found in other secondary flow investigations in turbine cascades, the present front-loaded airfoil showed higher secondary losses than the aft-loaded airfoil with the same value of Zweifel coefficient.Copyright

Measurements of Secondary Losses in a High-Lift Front-Loaded Turbine Cascade With the Implementation of Non-Axisymmetric Endwall Contouring

This paper is the second in a series from the same authors studying the mitigation of endwall losses using the low-speed linear cascade test facility at Carleton University. The previous paper documented the baseline test case for the study. The current work investigates the secondary flow in a cascade of more highly-loaded low-pressure turbine airfoils with and without the implementation of endwall profiling. This study is novel in two regards. First, the contouring is applied to low-pressure turbine airfoils, whereas studies conducted by other researchers have focused their endwall profiling efforts on the high-pressure turbine. Second, while previous researchers have optimized contouring designs for a given airfoil, the current work demonstrates the potential to open the design space by employing high-lift airfoils in conjunction with endwall contouring. Seven-hole pneumatic probe measurements taken within the blade passage and downstream of the trailing edge track the progression of the secondary flow and losses generated. The contouring divides the vorticity associated with the passage vortex into two weaker vortices, and reduces the secondary kinetic energy. Overall the secondary losses are reduced and the loss reduction is discussed with regards to changes in the flow physics. A detailed breakdown of the mixing losses further demonstrates the benefits of endwall contouring.Copyright

Aerodynamics of a Family of Three Highly Loaded Low-Pressure Turbine Airfoils: Measured Effects of Reynolds Number and Turbulence Intensity in Steady Flow

The steady, midspan aerodynamic performance of a family of three low pressure (LP) turbine airfoils has been investigated in a low-speed cascade wind tunnel. The baseline profile has a Zweifel coefficient of 1.08. To examine the influence of increased loading as well as the loading distribution, two additional airfoils were designed, each with 25% higher loading than the baseline version. All three airfoils have the same design inlet and outlet flow angles. The aerodynamic performance was investigated for Reynolds numbers ranging from 25,000 to 150,000 (based on the axial chord and inlet velocity) and for values of freestream turbulence intensity of 1.5% and 4%. The flow field was measured with a three-hole pressure probe. Also, detailed loading distributions were obtained for all three airfoils using surface static pressure taps. The baseline airfoil and the new aft-loaded airfoil showed a separation bubble on the suction side of the airfoil under most of the conditions examined. In addition, a sudden and intermittent stall was observed at low Reynolds numbers for the new aft-loaded airfoil. The relatively short separation bubble would abruptly “burst” and fail to reattach. As the Reynolds number was decreased over a narrow range, the percentage of time that the flow was fully-separated increased to 100%. By comparison, the separation bubble on the baseline airfoil gradually increased in size in an orderly way as the Reynolds number was decreased. The new front-loaded airfoil provided the most encouraging performance: no separation bubble was present except at the very lowest Reynolds numbers. The absence of a separation bubble also had a favourable effect on the loss behaviour of this airfoil: despite its much higher aerodynamic loading, it exhibited very similar midspan losses to those observed for the baseline airfoil.Copyright

Application of Non-Axisymmetric Endwall Contouring to Conventional and High-Lift Turbine Airfoils

Here we report on the application of non-axisymmetric endwall contouring to mitigate the endwall losses of one conventional- and two high-lift low-pressure turbine airfoil designs. The design methodology presented combines a gradient-based optimization algorithm with a three-dimensional CFD flow solver to systematically vary a free-form parameterization of the endwall. The ability of the CFD solver employed in this work to predict endwall loss modifications resulting from non-axisymmetric contouring is demonstrated with previously published data. Based on the validated trend accuracy of the solver for predicting the effects of endwall contouring, the magnitude of predicted viscous losses forms the objective function for the endwall design methodology. This system has subsequently been employed to optimize contours for the conventional-lift Pack B and high-lift Pack D-F and Pack D-A low-pressure turbine airfoil designs. Comparisons between the predicted and measured loss benefits associated with the contouring for Pack D-F design are shown to be in reasonable agreement. Additionally, the predictions and data demonstrate that the Pack D-F endwall contour is effective at reducing losses primarily associated with the passage vortex. However, some deficiencies in predictive capabilities demonstrate here highlight the need for a better understanding of the physics of endwall loss-generation and improved predictive capabilities. More detailed analysis of the contouring results for the Pack B design is presented in a companion paper (Knesevici et al. [1]).Copyright

Toward the Expansion of Low-Pressure-Turbine Airfoil Design Space

Future engine requirements, including high-altitude flight of unmanned air vehicles as well as a movement to reduce engine cost and weight, are challenging the current state of the art in low-pressure-turbine airfoil design. These new requirements present low-Reynolds number challenges as well as the need for high-performance high-lift design concepts. Here we report on an effort to expand the relatively well established design space for low-pressure turbine airfoils. Analytical and experimental mid-span performance data and loadings are presented for four new airfoil designs based on the Pack B velocity triangles. The new designs represent a systematic expansion of low-pressure turbine airfoil design space through the application of high-lift design concepts for frontand aft-loaded airfoils. All four designs performed as predicted across a wide range of Reynolds numbers. Full-span loss data for the new high-lift designs reveal increased endwall losses, which, with the application of non-axisymmetric endwall contouring, have been substantially reduced. Taken holistically, the results presented here demonstrate accurate transition modeling provides a reliable method to develop optimized, very high-lift airfoil designs. INTRODUCTION Low-Pressure Turbines (LPTs) can contribute as much as 30 percent of the weight of an aeroengine [1] and contain as many as 1900 individual airfoils (Figure 1). When one considers the cost of each LPT airfoil, which is typically a precision investment casting, it is clear that reduced-count technology for LPTs can provide considerable cost and weight savings to both the manufacturer and customer. The “historical” characteristic plotted in Figure 1 reflects airfoil counts of well established, in-service LPTs while the “Current” trend is more indicative of modern engines such as the GP7000 shown. One goal of the current work is to investigate closing the gap that exists between current LPT designs and the count levels afforded by “Geared” designs that employ a reduction gear between the LPT and fan to allow the LPT to operate at a more optimal rotational speed. One of the key challenges in producing high-performance LPTs is the low Reynolds numbers associated with flight at cruise conditions. Low-pressure turbine Reynolds numbers are often low enough to result in significant regions of laminar flow on the suction side of the airfoils, which in turn makes them susceptible to laminar separation and even full stall [2]. Additionally, with the emergence of unmanned air vehicles, which may operate at altitudes significantly higher than commercial airliners, the Reynolds-number-related design challenges are becoming even more difficult. 0 50

Performance Impacts Due to Wake Mixing in Axial-Flow Turbomachinery

One of the last loss mechanisms remaining to be quantified and correlated for inclusion in meanline predictive systems concerns the mixing of wakes across downstream airfoil rows. Here, we demonstrate that the unsteady losses incurred as turbomachinery wakes mix in downstream rows are a function of the velocity ratio across the downstream row as calculated in the frame of reference of wake generation. Analytical and computational results, compared to measurements of wakes mixing under variable free-stream velocity conditions, reveal that wake-loss modification is primarily a result of an inviscid dilation of the stream tubes that comprise the wake fluid. Further, simulations of wakes exposed to a range of turbomachinery-specific velocity ratios indicate that wake-loss augmentation caused by stream-tube dilation is significantly more pronounced than wake-loss reductions imparted by stream-tube contraction. It is demonstrated that wakes in turbines are dilated in the adjacent downstream row, whether it is a vane or a blade row, through a work extraction process that occurs in the wake-generation reference frame. Finally, comparisons between rig data and CFD simulations suggest that wake-mixing losses, enhanced by downstream rows, can contribute as much as 1.5 percent of lost efficiency in multistage low-pressure turbines.© 2006 ASME

Predicting Transition in Turbomachinery: Part II — Model Validation and Benchmarking

The ability to predict boundary layer transition locations accurately on turbomachinery airfoils is critical both to evaluate aerodynamic performance and to predict local heat-transfer coefficients with accuracy. In state-of-the-art Reynolds Averaged Navier-Stokes (RANS) simulations used to predict flowfields on turbomachinery airfoils, boundary layers are often assumed to be turbulent over the entire airfoil surface. Consequently, losses are not accurately predicted, particularly in the case of stalled airfoils. Here we report on an effort to include empirical transition models developed in Part I of this report in a RANS solver. To validate the new models, a two-dimensional design optimization was performed to obtain a pair of Low-Pressure Turbine (LPT) airfoils with the objective of increasing airfoil loading by 25%. Subsequent experimental testing of the resulting two new airfoils confirmed pre-test predictions of both high and low Reynolds number loss levels. In addition, the accuracy of the new transition modeling capability was benchmarked with a number of legacy cascade and LPT rig data sets. Good agreement between measured and predicted profile losses was found in both cascade and rig environments. However, use of the transition modeling capability has elucidated deficiencies in typical RANS simulations that are conducted to predict component performance. Efficiency-versus-span comparisons between rig data and multi-stage steady and time-accurate LPT simulation results indicate that loss levels in the endwall regions are significantly under-predicted. Possible causes for the under-predicted endwall looses are discussed as wall as suggestions for future improvements that would make RANS-based transitional simulations more accurate.Copyright