Toroidal Propeller Design via Sensitivity-Driven Methodology – our findings from ASME Turbo Expo 2026

Aristotle University of Thessaloniki (AUTH) team under Torpropel project is delighted to share that official paper was successfully presented at the ASME Turbo Expo 2026❗ 

The toroidal propeller concept represents a fascinating topological departure from conventional designs. Early studies have documented a 20–32% reduction in wake vorticity for large aircraft, as well as measurable acoustic benefits for small-scale UAVs. These are remarkable numbers. 

However, what is missing from the field is the intermediate-scale application of toroidal propellers—and a methodology to investigate this new design systematically. This is the precise gap we set out to address.

💡 The Roadblock:

Classical BEM methods don't work for toroidal propellers—they assume a tip vortex that doesn't exist. CFD is the only way forward.

💡 The Problem:

Toroidal geometries demand a high-dimensional design space requiring thousands of CFD samples—computationally prohibitive.

💡 Our Solution:

Reduce the design space before optimization. We deployed a global variance-based sensitivity analysis (RANS + Sobol) to rank parameters and separate signal from noise.

💡 Findings:

1️⃣ RPM & Radius are the primary linear drivers of Thrust/Torque.

2️⃣ Specific toroidal parameters induce strong nonlinear interactions—creating sweet spots linear models miss.

The payoff❓

✅ Isolate the drivers. 

✅ Map the interactions. 

✅ Shrink the space. 

✅ Intermediate-scale optimization becomes a practical reality.

📜 Full paper & DOI coming soon.