Researchers at TU Graz have demonstrated that a relatively overlooked component of aircraft engines, the intermediate turbine duct, holds substantial untapped potential for improving fuel and energy efficiency, according to Tech Xplore. These ducts connect high-pressure and low-pressure turbines and channel hot gas flow, but they are often bulky to meet structural and flow requirements. By making them “as short, small, and light as possible” while still retaining performance, engines can run more efficiently.
The team, working under the ARIADNE project, combined years of experimental flow data and computational fluid simulations, then applied machine learning methods to explore how changing geometric parameters affects engine efficiency rapidly. Among the approaches tried, such as surrogate modeling, physics-informed machine learning, and reduced-order modeling, the reduced-order models stood out. They simplify simulations by distilling only the most relevant flow features, speeding up computations by several orders of magnitude compared with full CFD runs.
Though reduced-order modeling sacrifices some precision, it reliably captures efficiency trends and highlights which geometrical tweaks, e.g., shorter or narrower ducts, yield gains. This makes it a powerful tool for early-stage design optimization and rapid iteration across many configurations without heavy computational cost.
So far, the models are two-dimensional. The team plans to extend them to full 3D simulations and publish the underlying data and modeling tools, enabling other researchers to build on their work.
For engineers designing future aircraft engines, this work underscores a simple but often neglected principle: optimizing the small, less glamorous components can pay off big. By leveraging AI-enabled reduced-order models, it becomes feasible to trim duct weight and dimensions, helping jets burn less fuel, emit less, and operate more efficiently.