Time-series estimation using transition networks on realistic, sparse flow data

Inspired by biological swimming and flying that utilizes distributed haptic sensors to control complex flows, we propose a data-driven approach for load estimation relying on complex networks. We exploit sparse, real-time pressure inputs combined with pre-trained transition networks to estimate aerodynamic loads in unsteady and highly-separated flows. The transition networks contain the aerodynamic states of the system as nodes along with the underlying dynamics as links. Two network strategies are proposed and tested on realistic experimental data from the flow around an accelerating elliptical plate at various angles of attack. Aerodynamic loads are then estimated for angles of attack cases not included in the training dataset to simulate the estimation process. Performance and limitations of the two network strategies are discussed, showing that transition networks can represent a versatile data-driven alternative for real-time signal estimation using sparse and noisy signals in realistic flows.