Multi-objective exploration of multistaged ducted thrusters

Solid-state electroaerodynamic propulsion systems can generate thrust by ejecting ionized air as a plasma jet. Although this concept promises nearly silent drones, the achievable thrust density has so far remained low. Experimental exploration of electrode shape and arrangement is costly and slow. Here, we propose an alternative: an efficient, fully automated framework that integrates an evolutionary algorithm, a two-dimensional incompressible Navier–Stokes equation solver, and a dedicated plasma simulator for solid-state electroaerodynamic propulsion to optimize electrode designs that simultaneously yield a higher thrust density and a higher thrust-to-power ratio. In particular, the latent vectors determined by the evolutionary algorithm are used to synthesize electrode geometries and the operating voltage, after which the computational meshes are generated automatically using the hierarchical Cartesian grid method. The governing equations for corona discharge and air flow are then solved to calculate the thrust density and thrust-to-power ratio. These results are then used to update the fitness function of the evolutionary algorithm. This workflow produces a Pareto front of electrode designs that maximize thrust density while also maximizing the thrust-to-power ratio.