Session: 32-07 Loss Reduction for Axial Turbines

Submission Number: 177440

Secondary Flow Loss Reduction in a Turbine Cascade With a New Integrated Blade and Endwall Configuration 

Driven by the demand for flow control-oriented design and increased turbine load, the control of secondary flow losses near the endwall region has emerged as a critical research focus. Flow control studies on the topic have traditionally involved the separate design of blades and endwalls, which tends to intensify secondary flow vortices. The current work proposes a new configuration to achieve aerodynamic synergy among the blade, endwall, and flow path, thereby reducing secondary flow losses near the endwall region. A parametric design method for this configuration is proposed, utilizing NURBS surfaces for geometric modeling and a droplet-type function to define the profile. A high-load turbine guide vane cascade was selected as the research object. High-fidelity numerical simulations, which were validated against experimental data, were employed. The Optimal Latin Hypercube Sampling technique was used to construct the sample space, and sensitivity analysis was conducted to identify the key design parameters. To enhance the flow control effectiveness, an intelligent optimization design framework was established by integrating the LightGBM machine learning model and the Harris Hawks Optimization algorithm. The optimization results demonstrate the significant potential of the new configuration in regulating secondary flows, achieving a reduction in total pressure loss of more than 9% across the cascade. Flow field analysis reveals that the optimized blade configuration effectively suppresses the development of the trailing edge shedding vortex and delays the formation of the passage vortex, thereby markedly improving the flow field environment and reducing secondary flow losses near the endwall.

Presenting Author: Li Zhijie Northwestern Polytechnical University

Presenting Author Biography: Phd at Northwestern Polytechnical University

Authors:

Li Zhijie Northwestern Polytechnical University
Wu Yanhui Northwestern Polytechnical University
Shi Xuyang Northwestern Polytechnical University
Li Haohua Northwestern Polytechnical University
Ma Xingyue Northwestern Polytechnical University