Session: Student Poster Competition

Submission Number: 185365

Numerical Study and Data-Driven Modeling of Flat-Plate Transpiration Cooling 

Gas turbine engines operating on the Brayton cycle require increasingly high turbine inlet temperatures, which significantly intensifies the thermal loading on turbine components. Transpiration cooling has emerged as a promising blade-cooling technique, injecting coolant through a porous medium to achieve uniform surface protection with minimal coolant consumption.

However, experimental validation is often constrained by the high manufacturing costs of porous media and the technical difficulties of testing under realistic high-temperature and high-pressure conditions. Furthermore, high-resolution Computational Fluid Dynamics (CFD) simulations are computationally expensive. This challenge is compounded by the fact that real porous media properties may vary due to clogging and manufacturing tolerances, necessitating rapid performance re-evaluation across diverse operating points.

 

To address these limitations, this study develops data-driven surrogate models based on Artificial Neural Networks (ANN) and Convolutional Neural Networks (CNN) to predict transpiration cooling effectiveness (η) under unseen operating and geometric conditions. A flat-plate transpiration cooling configuration was numerically investigated using ANSYS Fluent to construct a training database by varying porosity (ϕ = 0.183–0.367), thickness (t = 5–9 mm), length (L = 60–120 mm), and blowing ratio (F = 0.5–2.5). The ANN was designed to predict spatially averaged cooling effectiveness from scalar input parameters, while the CNN was employed to reconstruct the full surface effectiveness distribution from CFD field data.

 

The CFD results demonstrate that cooling effectiveness is primarily governed by the blowing ratio across most investigated geometries. At a fixed porosity (ϕ = 0.367) and length (L = 0.12 m), increasing the porous thickness significantly enhances cooling performance. Specifically, at F = 1.0%, the spatially averaged cooling effectiveness increased by approximately 4.8% as the thickness was varied from t = 3 to 9 mm. Additionally, increasing the porous length from L = 60 to 120 mm resulted in a 2.9% increase in the thermal boundary-layer thickness δ_T = 0.99T_main at the trailing edge (x/L=1). These findings provide quantitative guidance for constructing reliable surrogate models for the rapid design screening of transpiration cooling configurations.

This work was partly supported by the National Research Foundation of Korea(NRF) grant funded by the Korea government(MSIT) (RS-2025-25434013), and by BK21 FOUR Program by Jeonbuk National University Research Grant

Presenting Author: Jisu Park Jeonbuk National University

Presenting Author Biography: Jisu Park is a Ph.D. student in the Department of Mechanical-Aerospace-Electric Convergence Engineering at Jeonbuk National University. His research focuses on film cooling, heat transfer, and computational fluid dynamics (CFD). He has one SCI-indexed journal publication: “Film-Cooling Performance with Various Hole Length-to-Diameter Ratios for Cylindrical and Laidback Fan-Shaped Holes with an Inlet Groove.”

Authors:

Jisu Park Jeonbuk National University
Changwoo Kang Jeonbuk National University