Machine-Learning-Based Surrogate Modeling of Aerodynamic Flow Around Distributed Structures

A machine-learning-based surrogate modeling method for distributed fluid systems is proposed in this paper, where a dimensionality reduction technique used to reduce the flowfield dimension and regression model predict reduced coefficients from input parameters. The specifically designed tackle involving aerodynamic structures, its performance illustrated by application on wake flow around wind turbine arrays an atmospheric boundary layer. main idea first decompose whole domain into subdomains, then carry out each subdomain treating both information parameters as parameters, finally obtain combining of with consideration matching condition at interface. applied two test cases: one-dimensional Poisson equation high-fidelity farm model. results demonstrate great efficiency accuracy excellent scalability different scales.