Improved Smith Chart for Axial Compressor Design

Abstract

Preliminary design of a turbomachinery stage usually begins with target flow states, represented by a velocity triangle. Guidance for these target velocities comes from a variety of sources, which include some physics-based rules but, more often than not, have a significant empirical basis. The best-known guidelines for axial designs are the so-called “Smith Charts”. This approach was documented in detail by Smith [1] and quickly became a preferred approach for initial design of axial turbines stages. The method was based on a significant set of test data for various turbine designs, corrected for complicating factors, such as various clearance gaps and so on. Similar, but less well-documented methods, have been developed for axial compressor design. Despite the widespread use of the Smith chart method, these approaches have a very limited substantiation and no clear pedigree. Most attempts to develop Smith-like charts for compressor design guidance are based on fairly simplified models that carry significant unknowns, particularly at high loading. This paper details the development of new Smith charts for axial compressor designs, by making extensive use of optimization techniques. Profile designs are optimized for a range of target flow coefficients and work coefficients. The result is a performance map, representing the maximum possible performance for a given set of coefficient values. The analysis was limited to 2D profile design (quasi-3D analysis) and a reaction of 50%, the theoretical ideal reaction rate for 2D flow. Using the newly developed maps of maximum performance potential, the accuracy of various modeling methods is examined, along with their ability to determine the optimum design condition. Improvements to profile loss methods, based on the new data, are suggested.