Effects of Area Ratio and Mean Rise Angle on the Aerodynamics of Interturbine Ducts

2018 ◽  
Vol 140 (9) ◽  
Author(s):  
Yanfeng Zhang ◽  
Shuzhen Hu ◽  
Ali Mahallati ◽  
Xue-Feng Zhang ◽  
Edward Vlasic

This work, a continuation of a series of investigations on the aerodynamics of aggressive interturbine ducts (ITD), is aimed at providing detailed understanding of the flow physics and loss mechanisms in four different ITD geometries. A systematic experimental and computational study was carried out by varying duct outlet-to-inlet area ratios (ARs) and mean rise angles while keeping the duct length-to-inlet height ratio, Reynolds number, and inlet swirl constant in all four geometries. The flow structures within the ITDs were found to be dominated by the boundary layer separation and counter-rotating vortices in both the casing and hub regions. The duct mean rise angle determined the severity of adverse pressure gradient in the casing's first bend, whereas the duct AR mainly governed the second bend's static pressure rise. The combination of upstream wake flow and the first bend's adverse pressure gradient caused the boundary layer to separate and intensify the strength of counter-rotating vortices. At high mean rise angle, the separation became stronger at the casing's first bend and moved farther upstream. At high ARs, a two-dimensional separation appeared on the casing and resulted in increased loss. Pressure loss penalties increased significantly with increasing duct mean rise angle and AR.

Author(s):  
Yanfeng Zhang ◽  
Shuzhen Hu ◽  
Ali Mahallati ◽  
Xue-Feng Zhang ◽  
Edward Vlasic

The present work, a continuation of a series of investigations on the aerodynamics of aggressive inter-turbine ducts (ITD), is aimed at providing detailed understanding of the flow physics and loss mechanisms in four different ITD geometries. A systematic experimental and computational study was carried out for varying duct mean rise angles and outlet-to-inlet area ratio while keeping the duct length-to-inlet height ratio, Reynolds number and inlet swirl constant in all four geometries. The flow structures within the ITDs were found to be dominated by the counter-rotating vortices and boundary layer separation in both the casing and hub regions. The duct mean rise angle determined the severity of adverse pressure gradient in the casing’s first bend whereas the duct area ratio mainly governed the second bend’s static pressure rise. The combination of upstream wake flow and the first bend’s adverse pressure gradient caused the boundary layer to separate and intensify the strength of counter-rotating vortices. At high mean rise angle, the separation became stronger at the casing’s first bend and moved farther upstream. At high area ratios, a 2-D separation appeared on the casing. Pressure loss penalties increased significantly with increasing duct mean rise angle and area ratio.


Author(s):  
Shuzhen Hu ◽  
Yanfeng Zhang ◽  
Xue Feng Zhang ◽  
Edward Vlasic

The inter-turbine transition duct (ITD) of a gas turbine engine has significant potential for engine weight reduction and/or aerodynamic performance improvement. This potential arises because very little is understood of the flow behavior in the duct in relation to the hub and casing shapes and the flow entering the duct (e.g., swirl angle, turbulence intensity, periodic unsteadiness and blade tip vortices from upstream HP turbine blade rows). In this study, the flow development in an ITD with different inlet swirl distributions was investigated experimentally and numerically. The current paper, which is the first part of a two-part paper, presents the investigations of the influences of the casing swirl variations on the flow physics in the ITD. The results show a fair agreement between the predicted and experimental data. The radial pressure gradient at the first bend of ITD drives the low momentum hub boundary layer and wake flow radially, which results in a pair of hub counter-rotating vortices. Furthermore, the radially moving low momentum wake flow feeds into the casing region and causes 3D casing boundary layer. At the second bend, the reversed radial pressure gradient together with the 3D casing boundary layer generates a pair of casing counter-rotating vortices. Due to the local adverse pressure gradient, 3D boundary layer separation occurs on both the casing and hub at the second bend and the exit of the ITD, respectively. The casing 3D separation enhances the 3D features of the casing boundary layer as well as the existing casing counter-rotating vortices. With increasing casing swirl angle, the casing 3D boundary layer separation is delayed and the casing counter-rotating vortices are weakened. On the other hand, although the hub swirls are kept constant, the hub counter-rotating vortices get stronger with the increasing inlet swirl gradient. The total pressure coefficients within the ITD are significantly redistributed by the casing and hub counter-rotating vortices.


Author(s):  
Todd Reedy

A turbulent compressible boundary layer in a nominally Mach 4.2 flow was investigated experimentally. Pitot, wall-static pressure, total pressure and temperature measurements were utilized to determine Mach number, temperature, and velocity profiles within the boundary layer. An adverse pressure gradient was observed, resulting in non-uniform flow in the streamwise direction of the test section during development. Alterations were made to the tunnel top and bottom walls to account for the growing boundary layer displacement thickness, resulting in a much improved, uniform Mach number in the freestream and boundary layer. The existence of a slight adverse pressure gradient remained. Flow visualization was conducted via the Schlieren imaging technique. Experimental results were compared against turbulent compressible flow theory and were found to be in excellent agreement, based on an extension of the law-of-the-wall and law-of-the-wake. Velocity profiles and boundary layer thicknesses of the theoretical and experimental results aligned satisfactorily.


Author(s):  
Yubao He ◽  
Hongyan Huang ◽  
Daren Yu

The backpressure propagation mode accompanied by shock-train evolution is investigated numerically in a rectangular duct with an open space. On this basis, the balance mechanism and parametric effects of heat transfer and skin friction for backpressure propagation are revealed to understand the nature of force competition better. As a result, the backpressure propagation mode can be classified into two different flow processes with increased backpressure. In addition, balance property mechanism reveals that both the momentum inside the boundary layer and the shear force which transfers the momentum from the outer core flow to boundary layer are combined to resist the adverse pressure gradient. Further, parametric effect indicates that varying wall temperatures and roughness heights lead to different degrees of changes in balance property. According to quantitative results, both wall temperature and roughness height decrease the local boundary-layer momentum at the starting point of original pressure rise and thus the local adverse pressure gradient wins the force competition. In the subsequently continuous flow, the adverse pressure gradient continues to propagate upstream and then is retarded gradually by the boundary layer with a fuller velocity profile until a new force balance is generated.


Author(s):  
B. Monje ◽  
D. Sánchez ◽  
R. Chacartegui ◽  
T. Sánchez ◽  
M. Savill ◽  
...  

The channels formed between adjacent blades in a turbine/compressor are nothing more than a variable section duct. Hence, the first step of turbomachinery design is to understand the physical processes experienced by a certain fluid when flowing through these nozzles and diffusers. In the main, nozzles are easier to understand since the fluid flows impelled by a favourable pressure gradient whereas for diffusers the flow has to face an adverse pressure gradient. This latter situation brings about the occurrence of stall (boundary layer detachment from the wall) which makes it more complicated to design and operate the component (both the individual stages and the entire compressor). It is thus essential to characterise the performance of diffusers of any type, which is influenced by several parameters such as geometry, Mach and Reynolds number, inlet total pressure and aspect of the boundary layer at the inlet section. Dolan and Runstadler generated very valuable information in 1973 (Pressure recovery performance of conical diffusers at high subsonic Mach numbers, NASA CR-2299) by providing performance maps for the flow of air in diffusers. This work is aimed at complementing the previous one by giving maps that apply to the flow of supercritical carbon dioxide in similar devices. By doing so, an important step towards the design of thermal turbomachinery specific of this singular fluid is taken.


2006 ◽  
Vol 110 (1106) ◽  
pp. 215-226 ◽  
Author(s):  
G. S. Cohen ◽  
F. Motallebi

Abstract The results of an investigation into the effects that sub-boundary layer vortex generators (SBVGs) have on reducing normal shock-induced turbulent boundary-layer separation are presented. The freestream Mach number and Reynolds number were M = 1·45 and 15·9 × 106/m, respectively. Total pressure profiles, static pressure distributions, surface total pressure distributions, oil flow visualisation and Schlieren photographs were used in the results analysis. The effects of SBVG height, lateral spacing and location upstream of the shock were investigated. A novel curved shape SBVG was also evaluated and comparisons against the conventional flat vane type were made. The results show that in all but two cases, separation was completely eliminated. As expected, the largest SBVGs with height, h = 55%δ, provided the greatest pressure recovery and maximum mixing. However, the shock pressure rise was highest for this case. The experiments showed that the mid height SBVG array with the largest spacing provided similar results to the SBVG array with the largest height. Reducing the distance to shock to 10δ upstream also showed some improvement over the SBVG position of 18δ upstream. It was suggested that total elimination of the separated region may not be required to achieve a balance of improved static pressure recovery whilst minimising the pressure rise through the shock. The effect of curving the SBVGs provided an improved near wall mixing with an improved static and surface total pressure recovery downstream of the separation line. The optimum SBVG for the current flow conditions was found to be the curved vanes of h = 40%δ, with the largest spacing, located at 18δ upstream of the shock. Overall, it was apparent from the results that in comparison to larger vortex generators with a height comparable to δ, for SBVGs the parameters involved become more important in order to obtain the highest degree of mixing from a given SBVG configuration.


1972 ◽  
Vol 39 (1) ◽  
pp. 7-12 ◽  
Author(s):  
M. J. Werle ◽  
R. T. Davis

The laminar boundary-layer equations were solved for incompressible flow past a parabola at angle of attack. Such flow experiences a region of adverse pressure gradient and thus can be employed to study the boundary-layer separation process. The present solutions were obtained numerically using both implicit and Crank-Nicolson-type difference schemes. It was found that in all cases the point of vanishing shear stress (the separation point) displayed a Goldstein-type singularity. Based on this evidence, it is concluded that a singularity is always present at separation independent of the mildness of the pressure gradient at that point.


2013 ◽  
Vol 135 (4) ◽  
Author(s):  
Kin Pong Lo ◽  
Christopher J. Elkins ◽  
John K. Eaton

Conical diffusers are often installed downstream of a turbomachine with a central hub. Previous studies showed that nonstreamlined hubs had extended separated wakes that reduced the adverse pressure gradient in the diffuser. Active flow control techniques can rapidly close the central separation bubble, but this restores the adverse pressure gradient, which can cause the outer wall boundary layer to separate. The present study focuses on the use of a step-wall diffuser to stabilize the wall boundary layer separation in the presence of core flow control. Three-component mean velocity data for a set of conical diffusers were acquired using magnetic resonance velocimetry. The results showed the step-wall diffuser stabilized the wall boundary layer separation by fixing its location. An axisymmetric step separation bubble was formed. A step with a periodically varying height reduced the reattachment length of the step separation and allowed the diffuser to be shortened. The step-wall diffuser was found to be robust in a range of core flow velocity profiles. The minimum distance between the core flow control mechanism and the step-wall diffuser as well as the minimum length of the step were determined.


1965 ◽  
Vol 23 (4) ◽  
pp. 767-778 ◽  
Author(s):  
A. A. Townsend

The development of a turbulent boundary layer in a strong adverse pressure gradient can be described by the two-layer model proposed by Stratford (1959), in which the outer part of the flow is assumed to be unmodified by the pressure-rise and the inner part described by two local parameters, the surface stress and the pressure gradient. The description suggests that the modification of the original flow is in some sense self-preserving, and it is shown here that self-preserving development of the modification is consistent with the Reynolds equations of turbulent flow in particular pressure distributions. For these distributions, the predictions of the two-layer model are confirmed without any need to make the sharp and arbitrary distinction between the two parts of the boundary layer.


2008 ◽  
Vol 112 (1133) ◽  
pp. 415-420
Author(s):  
G. S. Cohen ◽  
F. Motallebi

Abstract Experiments have been conducted to assess the effects that sub-boundary-layer vortex generators (SBVGs) have on reducing normal shock-induced turbulent boundary-layer separation. The freestream Mach number and Reynolds number were M = 1·45 and 15·9 × 106/m, respectively. Detailed measurements of a fully developed, flat plate turbulent boundary layer were used in order to assess the performance of ten different SBVG configurations. The SBVG performance was assessed by comparing total pressure profiles measured upstream of separation and downstream of reattachment. Static pressure distributions, near surface total pressure distributions, oil flow visualisation and Schlieren photographs were also used. The effect of SBVG height was investigated. The results show the largest SBVGs with height, h = 55%δ, provided the greatest static pressure recovery and maximum mixing. However, the shock pressure rise (wave drag) was highest for this case.


Sign in / Sign up

Export Citation Format

Share Document