An original variational formulation is developed for the inverse problem of reconstructing full-field structural displacement and pressure distribution of membrane wings subjected to steady loads from membrane strain distribution. A direct solution approach in cosimulation with fluid-dynamics solvers is also presented. Moving least squares are used to smooth and remap surface strain measurements, estimated from digital image correlation, as needed by the inverse solution meshing. The same approach is used to map the structural and fluid interface kinematics and loads during the fluid-structure cosimulation. Both the direct and the inverse analyses are validated by comparing the direct predictions and the reconstructed deformations with experimental data for prestressed rectangular membranes subjectedtostatic and unsteady loads. The load distributions reconstructed using the inverse analysis are compared with the corresponding loads obtained using the direct analysis. The inverse analysis runs on standard off-the-shelf PCs and can be implemented in real time, providing load-distribution estimates at a rate in the order of tens of data sets per second

No abstract

This paper deals with the coupled structural and fluid-dynamics analysis of flexible flapping wings using multibody dynamics. A general-purpose multidisciplinary multibody solver is coupled with a computational fluid dynamics code by means of a general-purpose, meshless boundary interfacing approach based on Moving Least Squares with Radial Basis Functions. The general-purpose, free software multibody solver MBDyn is used. A nonlinear 4-node shell element has been used for the structural model. The fluid dynamics code is based on a stabilized finite element approximation of the unsteady Navier-Stokes equations. The method (often referred to in the literature as G2 method) has been implemented within the programming environment provided by the free software project FEniCS, a collection of libraries specifically designed for the automated and efficient solution of differential equations. FEniCS provides extensive scripting capabilities, with a domain-specific language for the specification of variational formulations of Partial Differential Equations that is embedded within the programming language Python. This approach makes it possible to easily and quickly build complex simulation codes that are, at the same time, extremely efficient and easily adapted to run in parallel. The coupling of the multibody and Navier-Stokes codes is strictly enforced at each time step. The fluid dynamics discretization is automatically refined to keep the error on the overall lift and drag coefficients below a user-defined tolerance. The method is first tested by computing the drag force of a non-oscillating NACA 0012 airfoil traveling in air. Subsequently, the drag and lift forces on a rigid and flexible oscillating NACA 0012 wing are compared with experimental data. Encouraging results obtained from the modeling and analysis of the dynamics and aeroelasticity of flexible oscillating wing models confirm the ability of the structural and fluid dynamics models to capture the physics of the problem.

No abstract

Thin structural components characterize a broad class of Micro-Aerial Vehicles (MAV). This work presents an original approach for the determination of transverse load distribution based on distributed strain measurements. A variational formulation is developed for the inverse problem of the reconstruction of full-field structural displacement of membrane wings subjected to static and unsteady loads. Surface strain measurements are estimated from Digital Image Correlation (DIC). Moving Least Squares are used to smooth and remap measurements as needed by the inverse solution meshing, and to map the structural and fluid interface kinematics and loads during the fluid-structure co-simulation. The inverse analysis is verified by reconstructing the deformed solution obtained with an analogous direct formulation, based on nonlinear membrane structural analysis implemented in a general-purpose multibody solver and tightly coupled in co-simulation with a CFD solver. The direct analysis is performed on a different mesh and subsequently re-sampled. Both the direct and the inverse analyses are validated by comparing the direct predictions and the reconstructed deformations with experimental data for prestressed rectangular membranes subjected to static and unsteady loads. The reconstructed load distributions are compared with the corresponding ones obtained using the direct analysis