Heat Transfer Enhancement of Wedge-Shaped Channels Using the Interpretable Multi-Objective Generative Design Method

Abstract Gas turbine blade cooling is crucial for enhancing turbine efficiency and durability. In particular, wedge-shaped channels are commonly used for trailing edge internal cooling. However, conventional cooling channel structures such as pin fins have limitations in achieving uniform heat transfer. To address this issue, a new approach using bionic cooling structures generated through self-organized equation had been proposed. Additionally, a multi-objective Bayesian approach is utilized to optimize the topological parameters, with comprehensive emphases on two specific design objectives: improving heat transfer and reducing pressure loss. Simultaneously, data mining techniques are applied to automatically extract design principles and clarify the enhanced performance characteristics resulting from optimization. Then the study investigates the performance of optimized self-organized structures in comparison to pin-fin structures. Results show that self-organized structures provided better heat transfer performance with lower pressure drop, and more uniform temperature distribution. Specifically, under the same level of pressure loss, the selected self-organized structure increases heat transfer by 35%. Data mining identifies the material orientation and the material generation on the left side region as critical factors in topological structures, significantly influencing pressure loss and overall heat transfer efficiency. This research highlights the potential of self-organized structures as a design methodology for turbine blade cooling.