AI designed Birmingham Blade is optimised for urban wind

The collaboration between AI design specialists EvoPhase and precision metal fabricators Kwik Fab provides a solution to the design and production of small-scale, affordable, generators of wind energy. 

EvoPhase used its AI-driven design process to generate and test designs for their efficiency at wind speeds found in Birmingham, which, at 3.6m/s are lower than the 10m/s rating for most turbines. 

The design process employed evolutionary algorithms, mimicking natural selection, to optimize the turbine's performance by generating and evaluating various designs, introducing randomness to avoid local minima in optimisation.

“Most algorithms would be able to take one design and start refining it [but] only change tiny bits around it, and it still looks kind of the same,” explained Leonard Nicusan, EvoPhase’s Chief Technology Officer. “For algorithms with our AI, we don't test one design. We tested hundreds at a time, created 2000 designs that all looked completely different. They allowed us to find these best characteristics. They produced a seven-fold improvement in the amount of kinetic energy that it collects from the wind, as opposed to the industry standard designs that work for cities.”