A 4-stage axial research compressor, which is representative of the rearmost stages of a highly-loaded military or civil compression system, has been designed and tested at RAE Pyestock. The compressor is of large scale, with extended inter-row gaps, to facilitate the acquisition of detailed aerodynamic data. The unit performed well on test, exceeding its design pressure ratio of 4.0, and achieving a peak polytropic efficiency at design speed of 89%. Flow profiles obtained from area traversing at stator exits are presented and discussed. The measured performance is compared with an S1-S2 calculation incorporating an inviscid-viscous blade-to-blade method.

Transonic fans and compressors are now widely used in gas turbine engines because of their benefits in terms of compactness and reduced weight and cost. However, careful and precise design is essential if high levels of performance are to be achieved. In this paper, the evolution of transonic compressor designs and methods is outlined, followed by more detailed descriptions of current compressor configurations and requirements and modern aerodynamic design methods and philosophies. Current procedures employ a range of methods to allow the designer to refine a new design progressively. Overall parameters, such as specific flow and mean stage loading, the axial matching between the stages at key operating conditions and the radial matching between the blade rows are set in turn, using one- and two-dimensional techniques. Finally, detailed quasi-three-dimensional and three-dimensional computational fluid dynamics (CFD) analyses are employed to refine the design. However, it is important to appreciate that the methods all have significant limitations and designers must take this into account if successful compressors are to be produced.

A recent ASME paper by the authors described a quasi-three-dimensional calculation system for transonic compressor blade rows. The system predicts both the internal flow field and the overall performance of the blade row. It therefore enables the compressor engineer to optimize the blade shapes in order to improve the design point efficiency. This is explored in the present paper. A new type of blade profile has been developed to allow sufficient freedom for the optimization. Application to the design of a high-efficiency, transonic civil fan rotor is discussed.

A set of low speed blading has been designed to represent an embedded stage from the DERA C147 high speed research compressor. The aim of the design was to undertake a careful high-to-low speed transformation of the geometry of a high speed stage and evaluate the transformation process through comparing detailed flow measurements taken in both the high and low speed environments. The high-to-low speed transformation process involves compromises due to both geometric and aerodynamic constraints. Geometric constraints include the parallel annulus of the low speed compressor and also its size and power which restrict the Reynolds number that can be achieved. Aerodynamically the high speed blades have to be subsonic and the effects of Mach number on loss buckets and boundary layer development limit the extent to which a full high-to-low speed match is possible. The low speed blading has been tested at Cranfield University’s 4-stage research compressor facility. Detailed traverse measurements were taken at rotor and stator exit along with blade surface static pressure measurements and oil flow visualisation. These, along with previous traverse measurements from the C147 compressor, have been used to show that a good comparison of pressure, flow angle, and endwall loss distributions can be achieved despite the compromises inherent in the transformation process. However some interesting differences were apparent and these are discussed. In addition, 3D flow calculations have been performed on both the rotor and stator using measured inlet conditions. These predictions model the endwall corner flow well. However, further work is needed to obtain better modelling of the clearance flow of both blade rows.

No abstract