Design and Analysis of a High Pitch to Chord Ratio Cascade Representative of Ducted Propfans

Efforts to reduce the specific fuel consumption of a modern aero engine focus in particular on increasing the by-pass ratio beyond the current level of around 5. One concept is the counterrotating shrouded propfan operating at low overall pressure ratio and having only very few fan blades of extremely high pitch/chord ratios. The relative inlet Mach numbers cover a range from 0.7 at the hub to 1.1 at the tip section of the first rotor. A propfan cascade was designed by taking into account two characteristic features of a propfan blade-blade section: • a very high pitch/chord ratio of s/c = 2.25 • an inlet Mach number of M1 = 0.90 which leads to transonic flow conditions inside the blade passage In the design process a profile generator and a quasi-3D Euler solver were used iteratively to optimize the profile Mach number distribution. Boundary layer behavior was checked with an integral boundary layer code. The cascade design was verified experimentally in the transonic cascade wind tunnel of DLR at Cologne. The extensive experimental results confirm the design goal of roughly 5 degree flow turning. A total pressure loss coefficient of less than 1.5% was measured at design conditions. This validates the very high efficiency level the propfan concept is calling for. A 2D Navier-Stokes flow analysis code yields good results in comparison to the experimental ones.

The Role of Tip Clearance in High-Speed Fan Stall

A numerical experiment has been carried out to define the near stall casing endwall flow field of a high-speed fan rotor. The experiment used a simulation code incorporating a simple clearance model, whose calibration is presented. The results of the simulation show that the interaction of the tip leakage vortex and the in-passage shock plays a major role in determining the fan flow range. More specifically, the computations imply that it is the area increase of this vortex as it passes through the in-passage shock, which is the source of the blockage associated with stall. In addition, for fans of this type, it is the clearance over the forward portion of the fan blade which controls the flow processes leading to stall.

A Unified Variable-Domain Variational Approach to Hybrid Problems of Compressible Blade-to-Blade Flow

A unified theory of various hybrid problems for blade-to-blade compressible flow is developed herein via the functional variation with variable domain. Two variational principle (VP) families for three typical hybrid problems are derived, following a systematic approach (Liu, 1990a). Full advantage is taken of the natural boundary condition and suction/blowing along the blade surface are accommodated. This theory is aimed at offering a new theoretical basis for the finite element method (FEM) and various ways for blade design and/or modification, and it also constitutes an important part of optimal cascade theory (Liu,1987b). Based on these VPs, a new FEM with self-adjusting nodes is also suggested, and the numerical tests yield good results.

Computation of Flow Past a Turbine Blade With and Without Tip Clearance

Three-dimensional solutions of the ensemble-averaged Navier-Stokes equations have been computed for a high-turning turbine rotor passage, both with and without tip clearance effects. The geometry is Pratt & Whitney’s preliminary design for the Generic Gas Generator Turbine (GGGT), having an axial chord of 0.5 inch and turning angle of about 160 degrees. The solutions match the design Reynolds number of 3x 106/inch and design inflow/outflow distributions of flow quantities. The grid contains 627,000 points, including 20 radial points in the clearance gap of 0.015 inch, and has a minimum spacing of 10−4 inch adjacent to all surfaces. The solutions account for relative motion of the blade and shroud surfaces and include a backstep on the shroud. Computed results are presented which show the general flow behavior, especially near the tip clearance and backstep regions. The results are generally consistent with experimental observations for other geometries having thinner blades and smaller turning angles. The leakage flow includes some fluid originally in the freestream at 91 percent span. Downstream, the leakage flow behaves as a wall jet directed at 100 degrees to the main stream, with total pressure and temperature higher than the freestream. Radial distributions of circumferentially-averaged flow quantities are compared for solutions with and without tip leakage flow. Two-dimensional solutions are also presented for the mid-span blade geometry for design and off-design inflow angles.

Measurement and Prediction of Heat Transfer From Compressor Discs With a Radial Inflow of Cooling Air

An integral theory is used to model the flow, and predict heat transfer rates, for corotating compressor discs with a superposed radial inflow of air. Measurements of heat transfer are also made, both in an experimental rig and in an engine. The flow structure comprises source and sink regions, Ekman-type layers and an inviscid central core. Entrainment occurs in the source region, the fluid being distributed into the two nonentraining Ekman-type layers. Fluid leaves the cavity via the sink region. The integral model is validated against the experimental data, although there are some uncertainties in modelling the exact thermal conditions of the experiment. The magnitude of the Nusselt numbers is affected by the rotational Reynolds number and dimensionless flowrate; the maximum value of Nu is found to occur near the edge of the source region. The heat transfer measurements using the engine data show acceptable agreement with theory and experiment. This is very encouraging considering the large levels of uncertainty in the engine data.

Turbulence Measurements in a Centrifugal Pump With a Synchronously Orbiting Impeller

Turbulence profiles were measured in a centrifugal pump with an impeller with backswept blades using a two directional laser velocimeter. Data presented includes radial, tangential, and cross product Reynolds stresses. Blade to blade profiles were measured at four circumferential positions and four radii within and one radius outside the four bladed impeller. The pump was tested in two configurations; with the impeller running centered within the volute, and with the impeller orbiting with a synchronous motion (ε/r2 = 0.016). Flow rates ranged from 40% to 106% of the design flow rate. Variation in profiles among the individual passages in the orbiting impeller were found. For several regions the turbulence was isotropic so that the cross product Reynolds stress was low. At low flow rates the highest cross product Reynolds stress was near the exit. At near design conditions the lowest cross product stress was near the exit, where uniform flow was also observed. Also, near the exit of the impeller the highest turbulence levels were seen near the tongue. For the design flow rate, inlet turbulence intensities were typically 9% and exit turbulence intensities were 6%. For 40% flow capacity the values increased to 18% and 19%, respectively. Large local turbulence intensities correlated with separated regions. The synchronous orbit did not increase the random turbulence, but did affect the turbulence in the individual channels in a systematic pattern.

Computations of Unsteady Multistage Compressor Flows in a Workstation Environment

High-end graphics workstations are becoming a necessary tool in the Computational Fluid Dynamics (CFD) environment. In addition to their graphics capabilities, the latest generation of workstations have powerful floating point operation capabilities. As workstations become common, they could provide valuable computing time for applications, such as turbomachinery flow calculations. This paper discusses the issues involved in implementing an unsteady, viscous multistage turbomachinery code (STAGE-2) on workstations. The workstation version of STAGE-2 has then been used to study the effects of axial-gap spacing on the time-averaged and unsteady flow within a 2 1/2-stage compressor. Results include force polar plots, time-averaged pressure contours, standard deviation of pressure contours, time-averaged surface pressures and pressure amplitudes.

The Effect of Blade Tip Geometry on the Tip Leakage Flow in Axial Turbine Cascades

The phenomenon of tip leakage has been studied in two linear cascades of turbine blades.The investigation includes an examination of the performance of the cascades with a variety of tip geometries. The effects of using plain tips, suction side squealers and pressure side squealers are reported. Traverses of the exit flow field were made in order to determine the overall performance. A method of calculating the tip discharge coefficients for squealer geometries is put forward. In linking the tip discharge coefficient and cascade losses a procedure for predicting the relative performance of tip geometries is developed. The model is used to examine the results obtained using the different tip treatments and to highlight the important aspects of the loss generation process.

Rim Sealing of Rotor-Stator Wheelspaces in the Absence of External Flow

Sealing of the cavity formed between a rotating disc and a stator in the absence of a forced external flow is considered. In these circumstances the pumping action of the rotating disc may draw fluid into the cavity through the rim seal. Minimum cavity throughflow rates required to prevent such ingress are estimated experimentally and from a mathematical model. The results are compared with other workers’ measurements. Measurements for three different types of rim seal are reported for a range of seal clearances and for rotational Reynolds numbers up to 3 × 106. The mathematical model is found to correlate the experimental data reasonably well.

Similarity Transformations for Compressor Blading

Improving the performance of high speed axial compressors through low speed model compressor testing has proved to be economical and effective (Wisler, 1984). The key to this technique is to design low speed blade profiles which are aerodynamically similar to their high speed counterparts. The conventional aerodynamic similarity transformation involves the small disturbance potential flow assumption therefore its application is severely limited and generally not used in practical design. In this paper, a set of higher order transformation rules are presented which can accommodate large disturbances at transonic speed and are therefore applicable to similar transformations between the high speed HP compressor and its low speed model. Local linearization is used in the non–linear equations and the transformation is obtained in an iterative process. The transformation gives the global blading parameters such as camber, incidence and solidity as well as the blade profile. Both numerical and experimental validations of the transformation show that the non–linear similarity transformations do retain satisfactory accuracy for highly loaded blades up to low transonic speeds. Further improvement can be made by only slightly modifing profiles numerically without altering the global similarity parameters.