Session: 37-02 Radial Turbomachinery Optimization

Paper Number: 123250

123250 - Data-Driven Radial Compressor Design Space Mapping 

Estimates of turbomachinery performance trends are required to inform system-level compromises during preliminary design. Existing empirical correlations for efficiency are based on limited experimental data, while analytical loss models require calibration to yield predictive results. This paper uses a set of 1794 radial compressor computations to map efficiency as a function of mean-line aerodynamics, and determine the governing fluid-dynamic mechanisms. An open-source turbomachinery design code creates annulus and blade geometry, then meshes and runs a three-dimensional computational fluid dynamics simulation for points sampled from the mean-line design space. Polynomial surface fits yield a continuous eight-dimensional representation of the design space for further analysis, predicting efficiency with a root-mean-square error of 1%.

A balance between surface dissipation and secondary kinetic energy losses sets optimum values for inlet Mach number, hub-to-tip ratio, de Haller number, and backsweep angle. Surface dissipation dominates the effect of flow coefficient, with high surface areas at low flow coefficients, and high velocities at high flow coefficients. Compact compressor designs are achieved by increasing inlet Mach number, reducing hub-to-tip ratio, and minimising the radial component of stage loading — all of which reduce efficiency approaching the design space boundary.

The paper demonstrates large ensembles of automated designs and simulations as a higher fidelity replacement for legacy empirical correlations in preliminary design.

Presenting Author: James Brind Whittle Laboratory, University of Cambridge

Presenting Author Biography: James Brind is a Research Associate at the Whittle Laboratory. He holds undergraduate, masters, and doctoral degrees from the University of Cambridge. For his PhD, based in the UK EPSRC Centre for Doctoral Training in Gas Turbine Aerodynamics, he worked on the design of turbine film cooling systems with Mitsubishi Heavy Industries. His post-doctoral research has concerned turbine acoustics and turbomachinery design methods.

Authors: