Fluid-structure interaction characteristics of tension membrane structures under free-vibration and forced-vibration based on numerical simulation

Tension membrane structures are apt to experience severe vibration under wind action, and may undergo substantial fluid-structure interactions (FSI). Accurate consideration of the FSI effects in the wind-induced response of such structures is critical for their response estimation. Motion-induced aerodynamic force model established by forced-vibration testing is widely used to estimate the wind-induced response considering FSI effects of flexible structures, and is regarded as an efficient substitute for the technically demanding and costly free-vibration test. Nevertheless, the accuracy and applicability of the motion-induced force model on the response estimation of tension membrane structures under wind remain uncertain, due to the strong nonlinearity in wind-induced response of such structures. In this research, a systematic comparison of wind-induced response obtained by free-vibration test and estimated by the motion-induced aerodynamic model established by forced-vibration test is performed for tension membrane structure based on numerical simulations. A closed-type, one-way tensioned flat membrane structure is determined to be the object due to its relatively idealized geometric configuration. Fully-coupled simulations are utilized for the free-vibration model and validated against the reference aeroelastic experimental results, and complementary forced-vibration simulations are performed to establish the motion-induced aerodynamic force model. It is found that the displacement responses of tension membrane structure estimated by motion-induced aerodynamic force model agrees well with those obtained from free-vibration test, while discrepancies exist between forced-vibration model and free-vibration model in the distribution of fluctuating pressures above the membrane. Energy transfer analysis and proper orthogonal decomposition (POD) analysis on the FSI system show that the discrepancies mainly arise from the disturbance and weakening of the coupling between the vortex convection and structural motion of the forced vibrating model, resulting from the influence of low-order body-induced vortex shedding components.