Abstract

Compressor performance is largely set by decisions made during preliminary design. This paper aims to provide an improved method of selecting the compressor topology: axial, mixed or radial at the early stage of the design process. The choice of topology has traditionally been made as some form of the question “How to choose the most efficient meridional topology for a given duty?”  This question is usually answered by consulting the Cordier line together with whatever empirical knowledge is available. This method of selecting geometric topology is limited in a number of ways. Because it is a line, collapsed by being plotted on a log-log axis, it is a relatively low fidelity way of selecting topology. The number of experimental data points on which it is based is relatively small and of an unknown quality. The fact that is doesn’t provide the designer with the necessary geometry information to specify a real compressor (e.g. hub to tip ratio, inlet to exit radius ratio, pitch to chord ratio).

This paper proposed a new method of selecting compressor geometric topology. The method improves on existing methods in three ways. Firstly, the topology selection is undertaken on a duty based smith chart (i.e. flow coefficient and work coefficient) instead of a Cordier line. Secondly, the number of data points on which the method is based is considerably larger (circa 6000 compressors) using 3D RANS CFD. Thirdly, the geometry of each compressor is fully specified and systematically controlled. This allows the designer to use the method to determine a higher fidelity of geometry.  To ensure that all data points are “good” compressors, a set of design rules has been developed from a database of good compressor designs collated from the literature. These rules ensure that all designs have reasonable meridional geometry, blade section design, stacking, and number of blades.

The paper shows what the optimal meridional topology of compressor is for each point on the Smith Chart as well as the line of duty points at which optimal performance is achieved. It also shows how the compressor design rules can be used to manipulate the location of this optimal line.  Finally, a design example illustrates how the new method can be used in the early stage of design to balance multiple design objectives.