Cantilever plate flutter in wind tunnel tests and comparison with 2D theory

Abstract

Cantilever plate flutter in axial flow has attracted more and more attention, as it is a typical model for the fluid-structure interaction which can be employed for many applications. When the wind speed exceeds a critical value, the cantilever plate will lose its stability by flutter. The purpose of this project is to illustrate the full 2D flutter mechanism of a cantilever plate, conduct systematic experiments in the wind tunnel to quantify the difference between our 2D theory and the nominally 2D experiment with various inevitable 3D effects, and construct the best possible 2D theory-experiment comparison.

For theoretical studies, a pseudo-spectrum method is employed together with the Galerkin procedure to solve this eigen-value problem. The effects of the boundary condition and the structure damping are studied. By modal analysis, the mechanism of the flutter is illuminated. It is found that the origin of the flutter instability is the inter-modal coupling of the first two in-vacuo modes for the air-structure mass ratio of interest. A simplified model with the same flutter mechanism is proposed and validated by the wind tunnel test.

The following 3D factors in a nominally 2D experiment are studied. (a) Structural corrugation effect which works to stiffen and stabilize the plate, when the span-to-chord aspect ratio is high. The most appropriate aspect ratio is found to be around 0.6. (b) Lateral edge flow effect which leads to less effective fluid loading and hence a stabilization factor in the flow-structure system. With the reduction of the lateral clearance to below 1mm, the experimental result is significantly modified. (c) The last missing factor in our 2D theoretical model is the boundary layer flow around the cantilever plate. By flow visualization and the plate flutter experiments, it is found that a plate with a thick boundary layer is more stable due to the reduction of effective fluid loading. By suppressing these three factors as far as possible, a satisfactory theory-experimental comparison is obtained which is much better than all existing literature. The experimental data in this project can serve as a benchmark for future numerical simulations.