INCLINED FOOTBRIDGE ANTENNA (Algiers, Algeria)
Aeroelastic Wind Tunnel Tests on the Footbridge Antenna of the Djamaa El-Djazair Complex
Year: 2018
The wind tunnel campaign addressed the aeroelastic behaviour of the inclined antenna at which the cable-supporting structure of the Passerelle Djamaa el-Djazair in Algiers (Algeria). The main objective was to quantify the wind-induced dynamic response of the antenna in its first cross-wind mode and to evaluate the role of structural damping under both uniform and turbulent wind, considering the potential interaction between vortex shedding (VIV) and galloping-type instability. To achieve this, a dedicated experimental programme was developed around a single scaled physical model tested in two distinct aeroelastic configurations, each reproducing transverse oscillations of the antenna through a different mechanical principle.
The work started with the definition of the wind input to be reproduced in the laboratory. Based on the site layout and on Eurocode 1 criteria, two reference atmospheric boundary-layer exposures were selected as representative envelopes for the antenna location: a smoother profile consistent with onshore flow from the sea (Category 0) and a rougher profile representative of more built-up/upwind terrain (Category II). The model scale was fixed at 1:80, prioritising Reynolds-number representativeness over a perfect match of integral turbulence length scales. A large set of preliminary tests was therefore carried out to design a boundary layer that reproduced, as closely as feasible at 1:80 scale, the target mean velocity and turbulence intensity profiles for both categories. The final boundary-layer generation system relied on distributed roughness elements along the tunnel floor, combined with an upstream barrier to increase turbulence intensity. The resulting profiles provided a good match of the mean velocity law, with the known and explicitly acknowledged compromise of smaller-than-target integral length scales at higher elevations. In parallel, uniform (quasi-laminar) flow conditions were also adopted, both as a conservative condition for VIV lock-in and to avoid uncertainties related to low-speed profile development.
A single physical model of the antenna was then developed and manufactured by 3D printing in ABS at 1:80 scale, reproducing the full external geometry over a model height of approximately 0.55 m and a longitudinal length of about 0.61 m. The same antenna model was used throughout the programme, while the dynamic boundary conditions and the achievable Scruton number were varied by switching between two suspension systems. This choice ensured that differences in response could be attributed to the dynamic setup and damping level, rather than to geometric or manufacturing inconsistencies across different models.
The first aeroelastic configuration, referred to as the IPE setup, was conceived to reproduce a dynamic system with a Scruton number representative of the full-scale structure. In this setup, the model was fixed on top of a steel element shaped as a double-T assembly. The transverse oscillation mechanism was provided by the out-of-plane bending deflection of the I-shaped element, whose geometry was designed so as to tune the system’s natural frequency. A finite-element model (SAP2000) was developed to guide the design of the setup and to target the required dynamic properties. Tests with this configuration were performed in uniform flow, deliberately selecting conditions that typically maximise lock-in response amplitudes. The outcome of this first stage provided a direct assessment of the expected full-scale behaviour at the reference damping level and showed that, for one of the investigated orientations, the cross-wind oscillation amplitude could exceed the serviceability limit set by the Client, motivating a second phase focused on damping requirements.
The second aeroelastic configuration, referred to as the Pivot setup, was specifically designed to explore how the response changes as the effective damping—and therefore the Scruton number—is increased. In this setup, the antenna model was mounted on the upper end of a tube free to rotate about a pivot through a bearing, with all other degrees of freedom restrained. The single active degree of freedom corresponds to the antenna oscillation out of the ideal plane of the bridge, i.e., a transverse horizontal motion at the antenna top. The rotational stiffness was controlled by connecting the oscillating axis to a system of helical springs acting in the plane of rotation. The entire suspension assembly was located beneath the tunnel floor, allowing the model to emerge in the test section while keeping the mechanism outside the flow. The support frame could also be rotated to set the model inclination with respect to the oncoming flow, enabling a controlled investigation over a limited set of relevant wind incidence angles, consistent with the fact that the first mode is predominantly transverse to the bridge deck axis. As for the IPE setup, the dynamic design of the Pivot configuration was supported by finite-element modelling in SAP2000 to obtain the target natural frequency.
To further extend the accessible Scruton-number range, an adjustable eddy-current magnetic damper was integrated into the Pivot setup. By letting an aluminium plate oscillate within the gap between two neodymium magnets, an additional viscous-like damping contribution was introduced, controllable by varying the magnet spacing. This solution enabled systematic tests at substantially higher damping levels, while acknowledging that added components may slightly modify the global dynamic properties and therefore require dedicated identification.
With the Pivot setup, the test matrix covered uniform flow as well as Category 0 and Category II boundary-layer inflows. Uniform flow was retained as a conservative reference for vortex-shedding lock-in, while the two boundary-layer profiles were used to estimate both the wind-induced vibrations under atmospheric turbulence and the galloping critical wind speed under site-consistent inflow conditions. The campaign was intentionally limited to wind incidence angles close to 0° and 180° (and small angular offsets) because these are the configurations that most directly excite the transverse mode of interest; angles far from these conditions were not considered significant for the investigated mode.
Overall, the wind tunnel programme delivered a structured, design-oriented assessment of the antenna aeroelastic response based on a single, consistent physical model and two complementary transverse-oscillation setups. The first configuration reproduced the reference damping condition to quantify the baseline response, while the second enabled a controlled exploration of increased damping—supported by spring tuning and adjustable magnetic damping—to identify the damping levels needed to keep oscillations within the specified limit under turbulent wind and to provide adequate margin between the galloping critical wind speed and the design wind speed.
