Harnessing Instability Mechanisms in Airfoil Flow for Data-Driven Forecasting of Extreme Events

Abstract

For certain Reynolds numbers, airfoils are subject to sporadic high-amplitude fluctuations in the aerodynamic forces. These extreme excursions may be seen as prototypical of the kind of unsteady and intermittent dynamics relevant to the flow around airfoils and wings in a variety of real-world applications. Here we investigate the instability mechanisms at the heart of these extreme events, and how they may be harnessed for efficient data-driven forecasting. Through a wavelet and spectral analysis of the pressure and vorticity, we find that the extreme events arise due to the instability of a specific frequency component distinct from the vortex shedding mode. During these events, this extreme event frequency draws energy from the energetically dominant vortex shedding flow and undergoes an abrupt transfer of energy from small to large scales. We propose a preprocessing algorithm to extract this extreme event frequency from the surface pressure data, which in conjunction with an extreme event-tailored loss function, allows us to avoid the commonly used long short-term memory architecture in favor of a simple feed-forward network—a significant reduction in cost over the previous state-of-the-art. Our model requires only three pressure sensors, and it is robust to their location—showing promise for the use of our model in dynamically varying applications. Finally, we show that relying solely on the statistics of the pressure and drag data for optimal sensor placement fails to improve model prediction over uniform or random sensor placement.