A hybrid physical-informational platform for aerohydrodynamic design of blade systems with multicriterial optimization based on quantum and neural network methods

Abstract

A hybrid physics-informed computing platform for aero-hydrodynamic analysis and multi-criteria optimization of blade system profiles is presented, combining the Neural Foil neural network surrogate with panel methods and gradient-based optimization. A generalized Magnus effect model for rotating profiles is developed and verified on patented rotor designs. Hybrid quantum-classical schemes for solving the Navier-Stokes equations are integrated, including variational quantum solvers for the Poisson equation and QUBO formulations for quantum annealing problems. A parametric NACA model adapted for optimization tasks is proposed. The acceleration of computations by a factor of 8–1000 relative to XFOIL with an error below 0.5% is experimentally confirmed; an increase in aerodynamic quality of up to 18%, rotor propulsor efficiency η=0.62, and a 40% increase in the spreading force of a trawl board are achieved. Directions for the platform’s evolution in the context of the quantum transition in computational hydrodynamics are formulated.