A computationally-augmented wind tunnel with irrotational gust generation for low-speed aerodynamics

Large-amplitude flow disturbances, such as gusts, can drastically change the aerodynamic loads on machines. Active flow control can mitigate or exploit these disturbances, but developing an effective control strategy is challenging, partly because the interactions are often non-linearly dependent on the gust amplitude. As a result, the parameter space is often too large to explore by wind tunnel experiments alone. One solution is to create a surrogate computational model (a digital twin) of the experiments that can augment the experimental data and model conditions outside the scope of the experiments. To ensure that the model is faithful to the true disturbed flow conditions, it should be able to assimilate data from experiments. We present a digital twin of a low-speed aerodynamics wind tunnel equipped with a system for irrotational gust generation. The digital twin couples a solver for the viscous flow without the wind-tunnel walls with one for a potential flow to correct the normal velocity at those walls, including the gust-generating suction flow. Using an ensemble Kalman filter, we infer the true time-varying gust conditions from the experiments. We compare the simulated gust response to the response predicted by classical aerodynamics and a low-order vortex model.