Assessment of digital image correlation as a method of obtaining deformations of a structure under fluid load

摘要

Digital Image Correlation (DIC) is employed for the measurement of full-field deformation during fluid-structure interaction experiments in a wind tunnel. The methodology developed for the wind tunnel environment is quantitatively assessed. The static deformation error of the system is shown to be less than 0.8% when applied to a curved aerofoil specimen moved through known displacements using a micrometer. Enclosed camera fairings were shown to be required to minimise error due to wind induced camera vibration under aerodynamic loading. The methodology was demonstrated using a high performance curved foil, from a NACRA F20 sailing catamaran, tested within the University of Southampton RJ Mitchell, 3.5m x 2.4m, wind tunnel. The aerodynamic forces induced in the wind tunnel are relatively small, compared with typical hydrodynamic loading, resulting in small deformations. The coupled deflection and blade twist is evaluated over the tip region (80-100% Span, measured from the root) for a range of wind speeds and angles of attack. Steady deformations at low angles of attack were shown to be well captured however unsteady deformations at higher angles of attack were observed as an increase in variability due to hardware limitations in the current DIC system. It is concluded that higher DIC sample rates are required to assess unsteady deformations in the future. The full field deformation data reveals limited blade twist for low angles of attack, below the stall angle. For larger angles, however, there is a tendency to reduce the effective angle of attack at the tip of the structure, combined with an unsteady structural response. This capability highlights the benefits of the presented methodology over fixed-point measurements as the three dimensional foil deflections can be assessed over a large tip region. In addition, the methodology demonstrates that very small deformations and twist angles can be resolved.