Session: 13-03 Heat Transfer Modeling Methods and Analysis
Submission Number: 178384
Cross-Scale Uncertainty Propagation and Robust Design of Heat Transfer Under Combustor–Turbine Coupled Effects
In aeroengines, turbines accomplish energy conversion, and their thermal protection critically impacts overall engine performance and life. Stable operation under harsh, high-temperature inflows demands complex cooling, yet combustor–turbine coupling introduces inlet variability that challenges reliable blade thermal-load prediction. This work aims to understand the cross scale uncertainty propogation and its rost design due to different scale uncertainties, such as film cooling, hot streak and blade geometry. We proposes and validates an integrated framework that combines a source-term equivalence method with graph neural network (GNN) surrogates for high-fidelity turbine-cooling conjugate heat transfer (CHT). Within the CHT setting, calibrated volumetric and boundary source terms replace explicit resolution of cooling geometry, maintaining agreement with high-fidelity thermal responses while markedly improving nonlinear convergence and numerical robustness. Treating combustor–turbine–induced inlet-temperature variability as the primary disturbance, a GNN surrogate is trained to predict node-level blade-surface temperatures from inlet fields, enabling efficient inference on large meshes. Systematic comparisons against high-fidelity CHT across multiple operating points, geometries, and resolutions assess accuracy, extrapolation capability, scalability, and quantify sensitivity to inflow-mode perturbations. Results show that the integrated use of source-term equivalence and GNN surrogates preserves predictive consistency while substantially reducing computational cost and iteration instabilities, providing a practical and methodologically transparent pathway for engineering-grade turbine thermal protection and digital-twin development.
Presenting Author: Zhihao Shao Zhejiang University
Presenting Author Biography: PhD Candidate, majoring in aero-thermal performance of turbines
Authors:
Zhihao Shao Zhejiang UniversityHao Chen Zhejiang University
Junying Wang AECC Sichuan Gas Turbine Eatablishment
Kai Zhou Zhejiang University
Jianping Yan Zhejiang University
Cross-Scale Uncertainty Propagation and Robust Design of Heat Transfer Under Combustor–Turbine Coupled Effects
Paper Type
Technical Paper Publication