Application of Aeroengine Aerodynamic Design Codes to Industrial Gas Turbine Design

Abstract
In 1992, Westinghouse Electric Corporation and Rolls-Royce signed a technology transfer agreement. The initial application for this aeroengine technology was the design by Westinghouse of the turbine for the 1500°C turbine inlet temperature, 58% (net) combined cycle efficiency, 501G engine. A key ingredient of the large jump in combined cycle efficiency over that of its predecessor, the 501F, was the reduction in the surface area of the first 3 cooled turbine stages. This reduction was made possible by the application of aerodynamic design and analysis which allowed accurate prediction of suction surface boundary layer growth in a 3D viscous environment. Thus, the number of aerofoils or the aerofoil chords were reduced in each cooled row, until the suction surface boundary layers were predicted to be a specified margin from separation at the trailing edge. Since the resulting aerofoil loading (back surface diffusion) was higher than previous Westinghouse experience, a 0.32 scale model of a first stage turbine was built and tested in a cold rig at the National Research Council of Canada (NRCC), to substantiate the design codes when applied to industrial turbines. The purpose of the test was threefold: (1) to verify that higher levels of back surface diffusion were possible without boundary layer separation, (2) to verify that aeroengine turbine empirical loss/efficiency prediction correlations were applicable to industrial turbines, and (3) to see how well the single row, 3D, steady, Navier Stokes analysis code used during design predicted actual swallowing capacity, and radial variations in turning and loss. The paper will describe the aerodynamic design tools and their validation, the test facility, hardware, measurement techniques, test results, and comparisons of prediction vs measurement, thus confirming the applicability of aeroengine aerodynamic technology to the design of large industrial gas turbines.

This publication has 0 references indexed in Scilit: