Session: 34-03 Turbine design methods 1

Paper Number: 124695

124695 - Aerodynamic Optimization of an Axial Turbine Blade and Diffuser 

As turbomachinery design and performance has improved over past decades, achieving further incremental improvement has become more challenging, requiring detailed evaluation of component-level interactions and trade-offs to drive efficiency gains. Optimization offers a direct pathway to performance improvement but has traditionally focused on individual components with design space restricted by shape generation, domain preparation, and computational resource limitations. This paper presents a multi-component optimization approach for improving aerodynamic performance of blade and non-axisymmetric diffuser combination of a single-stage axial turbine of a marine turbocharger. This approach captures the coupling between blade exit flow pattern and diffuser performance, allowing for holistic optimization for highest possible overall efficiency gain. Blade geometry is parametrized by throat distribution, with a spanwise section-based twisting approach used for shape generation. For the diffuser, a flexible, algorithmic, NURBS-based process accepting fundamental design parameters as inputs and allowing for non-axisymmetric shapes is deployed. Aerodynamic evaluation is performed by CFD, with automatic mesh generation using component-specific meshing software for both the blade and the diffuser followed by calculation applying a compressible solver. Optimization by a hill climbing method yielded a best design consisting of a blade with a triple-peaked, W-profiled throat distribution giving accelerated flow adjacent to the diffuser walls, increasing the diffuser’s separation margin, and enabling higher performance diffusers of greater area ratio and pressure recovery. Overall efficiency improvement of approximately 0.4% (absolute) over the baseline was achieved by the best design, which is more than double the best-achieved improvement from optimizing only the blade, and more than four times that from optimizing only the diffuser. This demonstrates the utility of this approach in studying coupled mechanisms between complex geometry turbomachinery components such as turbine stages, diffusers, and ducts. From the perspective of future turbomachinery design, this approach introduces a new framework that could support existing methods such as sequential, part-by-part design with greater understanding of the influence of interdependencies between components on overall performance.

Presenting Author: Kingshuk Dasadhikari Mitsubishi Heavy Industries

Presenting Author Biography: Turbomachinery research engineer focusing on turbine aerodynamics and cooling, as well as optimization and digital transformation of design processes.

Authors:

Kingshuk Dasadhikari Mitsubishi Heavy Industries
Yasunori Kimura Mitsubishi Heavy Industries
Yoshihiro Kuwamura Mitsubishi Heavy Industries
Hokuto Isoda Mitsubishi Heavy Industries
Fumito Hiratani Mitsubishi Heavy Industries