Information Processing Dynamics With Perturbed Cylinder-Flap Fluid-Structure Interactions

Dynamic interactions between fluids and structures are common in aerodynamic systems. Generally, design principles are structured to minimize the coupling between fluid dynamic modes and structural dynamic modes to ensure the stability and integrity of the system. Traditionally this is done by stiffening structural members to raise the natural frequencies. Flexible morphing structures, however, often include compliant regions that may experience increased fluid-structure interaction, increasing the control demands of the system. On the other hand, the field of embodied intelligence has advanced our understanding of the role mechanics of physical systems play in sensing, computation, and control. The framework of Physical Reservoir Computing has provided recent advances in leveraging the inherent nonlinear dynamics of systems for computation. Thus a foundational understanding of the information processing (i.e. analog mechanical computing) characteristics of fluid-structure interactions will be key to enabling flexible morphing aerodynamic design. In this work, we have built on past studies of mechanical reservoir computers and fluids-based reservoir computers with simulations of a simple fluid-structure interaction reservoir computer. We used a canonical 2-D cylinder and flap geometry in a channel as an initial testbed, focusing on how the variation in mean inlet velocity, shear modulus, and flow perturbation amplitude affect the dynamic response and information processing ability. We found that at very low flow speeds, the shear modulus of the flap had a negligible effect on the information processing of the system. However, at higher speeds and finite shear modulus, the more compliant system showed increased information processing.