Sensing and Processing of a Shock-Induced Recirculation Bubble Using a Vibratory Physical Reservoir Computer

Sensing and processing unsteady aerodynamics is a significant challenge due to the high sensor density and computational cost required to capture the broad spatial and temporal scales spanned by unsteady flow features. Physical Reservoir Computing (PRC) offers a promising approach to overcoming these challenges by leveraging the nonlinear dynamics of physical structures to sense spatially distributed pressure inputs and process information with minimal digital computation. This study demonstrates the vibratory response of nonlinear metamaterial structure that consists of a network of unit cells with bi-linear bending stiffness to SWBLI, providing a pathway to PRC. The vibratory metamaterial is designed as a mechanical network in which each unit cell exhibits bi-linear bending stiffness. We experimentally excite the metamaterial with unsteady pressures generated by a shock wave to examine its dynamic response. We placed the metamaterial with a membrane in a Mach 1.4 wind tunnel with a cavity beneath the metamaterial. We exposed the metamaterial to two test cases: (1) a Mach 1.4 freestream flow without an induced-shock, and (2) a Mach 1.4 flow with a shock induced by a 5 degree wedge impinging on the midpoint of the metamaterial. By comparing the metamaterial’s dynamic behavior under shock-induced excitation with that under no-shock conditions, our results show that the metamaterial vibrates in response to the unsteady pressure fluctuations present in both the freestream and shock-impingement cases, with a clear shift in frequency content between the two conditions. These observations indicate that the vibratory metamaterial is able to sense a change in aerodynamic conditions, especially distinguishing between shock and no-shock environments. This study provides a step towards using the vibratory metamaterials as a PRC to perform computational tasks, including detecting the presence of shock waves, sensing pressure inputs, and predicting aerodynamically relevant loads, through the PRC framework.