Partial similarity modeling for the performance of an axial turbine

Abstract

By utilizing the similarity characteristics of turbines, their performance can be tested under conditions more relaxed than the design conditions. The use of the similarity model provides the benefit of saving time and the cost required to build a turbine test bench for the performance measurements. This study employs a partial similarity model, rather than a complete similarity model, with a rotational speed of 6000 rpm and air as the working fluid. Condition selection for this model was based on Buckingham’s Pi theorem to nondimensionalize the parameters affecting turbine performance. Particularly, the pressure ratio was derived using in-house code. Based on mean-line design results, three-dimensional modeling and numerical analysis were performed using commercial software to compare the performances of the design and partial similarity models. The numerical analysis predicted the isentropic efficiencies of 86.1 % and 80.4 % for the design and partial similarity models, respectively. This study is significant as it enables turbine performance testing at a lower rotational speed with air as the working fluid.