Surrogate-based optimization for evaluating turbine cooling impacts on aero-engine performance

Abstract

Traditionally, Gas Turbine Engines (GTEs) have been optimized at the component level, with a primary focus on maximizing individual efficiencies to ensure functionality and reliability. However, this approach does not always result in engines with low Specific Fuel Consumption (SFC). Given rising fuel costs and global initiatives targeting net-zero carbon dioxide (CO₂) emissions by 2050, it is increasingly critical to design engines with minimized fuel consumption at the system level. This research presents the development of an integrated optimization tool that shifts the focus from maximizing component efficiency toward minimizing engine-level SFC. The work concentrates on turbine design, with particular emphasis on airfoil cooling, which represents the largest contributor to turbine cooling demand and has a direct impact on SFC. While cooling is essential for durability, excessive cooling reduces engine performance due to the extraction of compressor bleed air and mixing losses that degrade turbine work output. The developed tool integrates several in-house platforms from Pratt & Whitney Canada (P&WC): the Turbine Aerodynamic Meanline (TAML) solver for meanline calculation, the Cooling Flow Prediction Tool (CFPT) for estimating airfoil cooling requirements, and the Framework for Design Exploration (FDE) for optimization execution. The workflows combine Design of Experiments (DOE) and Surrogate-Assisted Optimization (SAO) to efficiently explore the design space and accelerate convergence. SFC exchange factors and sensitivity factors are incorporated to estimate the impact of cooling on turbine efficiency and SFC during the preliminary design phase, where rapid but reliable estimates are required. Two test cases representing turboprop engine configurations were analyzed, exploring variations in airfoil count, axial chord, and turbine inlet tip radius. The tool supports workflows for both turbofan and turboprop/turboshaft engines, including modes where the power of the Power Turbine (PT) may be fixed or allowed to vary for more design flexibility. Results were validated against manually optimized configurations and full-engine computational analyses, demonstrating strong agreement. Sensitivity analyses identified axial chord and tip radius as key drivers influencing SFC through their effects on cooling demand and aerodynamic performance. The developed tool provides a valuable capability for preliminary turbine design, reducing optimization time, improving accuracy, and minimizing manual iteration between engineering groups. Its application supports P&WC’s efforts to deliver high-performance, fuel-efficient, and environmentally sustainable engines. In order to provide Pratt & Whitney Canada with a competitive advantage, certain proprietary technical data have been withheld.

Date	7 Aug 2025
Original language	American English
Awarding Institution	École de technologie supérieure
Supervisor	Saïd Hany Moustapha (Supervisor) & Patrick Germain (Co-supervisor)