No abstract

Only four parameters are needed to describe the characteristics of turbomachines completely. This concept is used to present maximum obtainable efficiencies and the optimum design geometry of turbines as function of specific speed, specific diameter, Mach number, and Reynolds number, based on the state-of-the-art knowledge. Additionally, other aspects such as unit weight and rotor stresses are discussed. Some of the information is preliminary because of limited information on loss relations for turbines, particularly at low Reynolds number and high Mach numbers.

The similarity concept, presented in Part A of this paper, for describing the characteristics of turbines is used to provide design and performance information on pumps and compressors, based on state-of-the-art knowledge. This information is preliminary in many cases because of incomplete knowledge on loss interrelations for compressors and pumps. The similarity concept is expanded to describe a technique which allows finding the optimum match for turbomachines operating in flow systems using several turbocomponents.

Typical space power units have a tendency to encounter low Reynolds numbers in the last turbine stages. Comprehensive test data on the effect of low Reynolds numbers on the efficiency of turbomachines are lacking. An attempt is made to assess this influence, using conventional aerodynamic arguments. By distinguishing between viscous and nonviscous losses some tentative values have been calculated which are in fair agreement with the few available test data. These considerations indicate that the stage pressure ratio and the specific speed affect the Reynolds number influence significantly.

Using established cascade criteria, such as the aerodynamic blade loading coefficient and the blade surface diffusion ratio, generalized relations for the losses in cascades of axial turbomachines are formulated. These relations have been evaluated numerically for incompressible media, assuming conventional Reynolds numbers. The results are presented in forms of diagrams showing profile, endwall, and clearance loss coefficients. The calculated data are compared with test data, revealing fair agreement, with the possible exception of “impulse” cascades. Extension of the “single cascade” data to “multiple row cascades” shows the operating regime where substantial performance improvements can be obtained by multiple row cascades.