MS24.4 - Wind Induced Vibrations of Slender Structures

Single axis solar trackers are constructions characterized by a series of photovoltaic modules mounted on a supporting structure, consisting of a motorized longitudinal torsional shaft and a series of vertical columns. The purpose of the underlying structure is to withstand acting loads and, by changing the inclination of the solar panels, to follow the position of the sun and maximize the energy production. In order to have competitive solutions, solar trackers need to be as cost effective as possible. Designers therefore seek to minimize the amount of material used by reducing the cross sectional area in the torsional shaft and other supporting elements, leading, consequently, to limited torsional stiffness. Combined with the low degree of torsional constraints that usually characterize solar trackers, large torsional static and dynamic deformations are expected at frequencies typically excited by wind. The dynamic response induced by turbulent wind is relevant and could cause problems for the structural integrity due to fatigue. The objective of this paper concerns the definition of a predictive model for the fatigue damage accumulation due to cyclical dynamic wind effects acting on the large tracker surfaces. In the presented approach, combining the pressure distribution time histories, and eventually information about the aeroelastic response of the tracker, both acquired through wind tunnel testing, with the structural properties derived from a FE model, it is possible to define and numerically integrate the governing problem of the system. In accordance with evidence observed in actual parks, the fatigue damage will be expected to accumulate in correspondence of the elements connecting the photovoltaic modules to the underlying structure: under this assumption, a relationship able to associate the applied loads to an internal response ought to be defined. To this end, a transfer function that relates the pressure acting on the panel surface to the stress state in a specific point is reconstructed by performing a series of analysis on a local FE model reproducing, of the whole tracker, a single PV module and its supporting beams. From the definition of the transfer function, the time history of the loading term, obtained for a specific cross section of the tracker by combining the pressure distribution and the inertial forces, is translated in the time history of the stress acting in the connections. The analysis of this last quantity, which is performed by means of the rainflow cycle count, combined with the SN curve of the structural detail and material investigated, provide useful information for the fatigue design, for example allowing the computation of the damage rate with the Miner’s rule. Finally, in order to validate the approach, a series of cyclical loads laboratory tests are carried out on real modules. From this analysis it is expected that, with minor calibration on the SN curve and transfer function, the developed procedure produces reliable results for the design. In conclusion, it is expected that the approach developed for calculating fatigue damage in PV trackers, specifically in connections, can enable designers of these structures to optimize the verification procedure.

For many tall building forms, habitability requirements associated with excessive acceleration response become a governing design criterion as building heights increase. The application of modular construction methods to high-rise construction is a relatively new concept with limited previous research being conducted on the dynamic properties of tall modular buildings. Further to this, the real contribution of individual modular elements to overall lateral stiffness is largely unknown leading to significant uncertainty in acceleration response predictions. As modular construction continues to be employed in structures of ever-increasing height, the susceptibility of this form of construction to wind induced accelerations requires further investigation. This research considers the comparison and validation of computational models of a tall volumetric corner post modular structure with an RC core. Both ETABs and mathematically-equivalent mechanical models adapting an analytical stepped beam approach are developed and the inherent properties such as the natural frequencies and mode shapes are calculated. The inherent properties predicted by the models are compared to those obtained from the actual measured response as captured through a full-scale monitoring campaign. A full-scale monitoring campaign employing two triaxial accelerometers, a data acquisition system and a data storage system recorded the white noise ambient acceleration response of two tall, slender modular structures with overall heights of 135m and 150m. Wind speed and direction were also recorded throughout the monitoring campaigns. Structural identification techniques were used to process the measured acceleration responses and obtain estimates of the actual natural frequencies and damping ratios of the partially- and fully-complete structures. The acceleration response of the structure was captured at varying stages throughout the construction programme as more storeys of modules were added to the building and the contribution of the modules to the modal properties evolved. The comparison between the measured inherent properties at the different stages of construction and the model results at the equivalent stage provides vital insight into the overall stiffness contribution of modules in high-rise modular structures. This can lead to more efficient modelling and design procedures for a novel form of building. Furthermore, comparison of the modelled properties and the results from the full-scale monitoring campaign helps to provide a better understanding of model accuracy and identifies opportunities for further refinement of the modelling of tall modular buildings to reduce model size, run time and computational expense, without loss of accuracy in wind-induced response prediction. The validation of the model and identification of stiffness contributions of the modules supports structural optimisation analyses and the numerical investigations required to include vibration response mitigation measures in future designs.

Many civil, transport and other engineering structures (chimneys, submarine pipelines, risers, etc) suffer from appearance of vibrations in these systems. It often occurs when the whole structure of its part is an elastic, or elastically mounted rigid bluff body immersed in a fluid flow. The vortices shedding from the body leads to vortex-induced vibrations (VIV). When a structural natural frequency is close to the shedding frequency the bluff body experiences a periodic and oscillates. VIV can lead to fatigue damage accumulation up to the destruction of the structure, but they can also be used to harvest electric energy from the kinetic energy of air or water flow. In this study, vortex-induced vibrations of an elastic cylinder near a finite-length plate are experimentally investigated in aerodynamic tunnel. Rubber cylinder of diameter D = 6 mm was spanned in the test section of a wind tunnel near the plate of length nearly 6D. The oscillation amplitude peak for a single cylinder was 0.3D, the Reynolds number of the peak-amplitude regime based on cylinder diameter was in the range 180...260. For a cylinder located sufficiently upstream from to the plate trailing edge it was found that the oscillation amplitude ratio A/D decreases if the gap ratio G/D between the cylinder surface and the plate reduces. However, for the cylinder location at the same level or downstream from the plate trailing edge, there are regions of essentially larger oscillation amplitudes compared to the isolated cylinder case. The maximum amplitude increase by 39% was obtained. Smoke visualizations revealed complex interference of the plate cylinder wakes were conducted. When the gap ratio decreases, the shift of the lock-in range to higher velocities was observed. The modification of the shedding laws in the proximity of the plate is studied.

High-rise buildings are sensitive to wind-induced vibrations. The occupant comfort under these vibrations is assessed based on the peak acceleration. This peak acceleration is strongly dependent on the natural frequency and the damping of the building. Bronkhorst and Geurts showed that the damping and natural frequencies determined in the design phase of high-rise buildings can deviate significantly from measured values. Previous studies have investigated the influence of soil-structure interaction (SSI) on the damping of high-rise buildings, and observed that the overall damping is significantly influenced by soft soils. However, these studies did not explicitly account for the wind loading and applied a simplified analytical modelling approach for the foundation without consideration of piles. Furthermore, they did not investigate the influence of SSI on the peak acceleration under wind loading. In-situ measurements of both wind loads and building accelerations on the residential tower New Orleans (Rotterdam, the Netherlands), provided a good basis to investigate the influence of SSI on the dynamic behaviour of a high-rise building under wind loads. In the research presented here, a model was developed which consists of a frequency dependent description of the pile foundation, the building and the wind load. The dynamic spring stiffnesses of the foundation were obtained with a boundary element method software (Dynapile 2016). The building was modelled as a section-wise Euler-Bernoulli beam; the section bending stiffnesses were obtained with a detailed FEM model of the New Orleans tower. The Power Spectral Density function of the wind loading was obtained from wind load measurements. Finally the dynamic response is computed through the solution of the system of stochastic differential equations in the frequency domain. The results show that the spectral model accurately predicts the measured dynamic response of the New Orleans tower in the along-wind direction. For all the examined dynamic properties (natural frequency, damping and peak acceleration), the error compared with the measurements was smaller than 7%. A comparison with results obtained following the guidelines in EN.1991.1.4 showed a 30%–35% underestimation of the peak acceleration. The guidelines EN.1991.1.4 overestimate the natural frequency and the overall damping of the New Orleans tower, which results in an underestimation of the peak acceleration. More results on the influence of SSI on the dynamic properties will be presented in the full paper. The full paper will provide a more in-depth discussion of the comparison with the original design calculations and the calculations with the currently applicable design code (EN.1991.1.4). Finally, an additional model, will be studied. This model includes, in a simplified manner, the effects of SSI on the natural frequency and modal damping, to the Eurocode procedure. The results of this research show that the EN.1991.1.4 guidelines result in an underestimation of the peak acceleration compared to measurements on the New Orleans tower. In the full paper it will be demonstrated that soil-structure interaction plays an important role in obtaining a better estimate for the peak acceleration.