Multi-Objective Optimisation: Mapping the Trade-Off Space for Toroidal Propellers

The core paradox of engineering design is that by the time you fully understand the system trade-offs, your freedom to act on them has collapsed. In complex engineering cases, the true challenge is never a single goal. It's the balance: efficiency vs. noise, performance vs. weight, innovation vs. manufacturability. This is the field of multi-objective optimization (MOO).

The Core Theory:

For any system with competing objectives, a Pareto front exists — the set of designs where improving one metric would require sacrificing another. Instead of finding a single "best" answer, MOO maps the entire space of optimal compromises, transforming subjective trade-offs into quantified choices.

Our Project Approach:

In the TorPropel project, the LFMT Lab at AUTH University collaborated with Limmat Scientific to integrate MOO at the preliminary design stage. Here's how:

1.     Parametric Model: We parametrically defined the toroidal propeller geometry.

2.     Objective Search: The algorithm is set to address all objective goals simultaneously.

3.     Pareto Front Generation & Strategic Choice: The output is a Pareto Front of optimal designs. 

This process enables us to make foundational strategic decisions early in the design process, such as, "For the mission profile, how much dB noise reduction is worth a specific efficiency trade-off?" In short, we can choose not just a shape, but also a validated performance strategy.