MS24.5 - Wind Induced Vibrations of Slender Structures

Structural monitoring is getting growing attention for system identification, damage identification, and capacity assessments of structures. On the other hand, wind measurement using a network of anemometers has always been a typical feature of meteorological stations. With the growing interest to study transient wind events that last short periods, closely spaced anemometers with a high frequency of measurement are also becoming common. However, the simultaneous measurement of wind and structural response through full-scale monitoring is still very rare in the literature. If available, the data from simultaneous wind and structural monitoring could be crucial to study the impact of wind phenomena that do not have a well-defined theoretical framework for wind load calculation. One of such phenomena is the downburst outflow wind due to thunderstorm, a strong convective meteorological event whose frequency and intensity is increasing because of climate change. Due to their limited dimension in space and time, downburst outflows due to thunderstorm events have been rarely measured in full-scale. In the past 20 years, there were many efforts to study the response of structures under downburst winds through wind tunnel testing, computational fluid dynamics, and theoretical methods. However, the efforts were not supported with full-scale wind and structural response monitoring of structures under real downburst outflow wind. The wind engineering research group in University of Genova (Italy) developed a comprehensive research project addressed to the study of thunderstorm effects on structures, realizing three full-scale monitoring systems measuring simultaneously wind velocity and structural response. This research presents long-term full-scale monitoring of a lighting pole to study its dynamic response under downburst wind. Initially, the dynamic properties of the structure such as modal frequencies, modal shapes, and damping ratios are investigated through operational modal analysis. Then, a method that separates downburst winds from synoptic winds and gust fronts was applied to identify the occurrence of downburst winds during the monitoring period. Through this method, possible downburst events were identified. The response of the pole measured through accelerometers and strain gauges was analyzed to obtain the displacement time history of the pole during the identified downburst events. The simultaneous wind and structural response data were analyzed to study the correlation between wind speed and structural displacement parameters. The time histories of mean wind speed squared and mean structural displacement was found to be highly correlated. Whereas the time histories of turbulence intensity and the root mean square of displacement fluctuation were found to be negatively correlated. Because of the structural simplicity, availability of the dynamic properties, and the provision of wind and structural response registration, this research will serve as a basis for the validation of analytical downburst wind load calculation techniques.

Overhead electrical transmission line conductors are prone to aeolian vibrations, resulting from the alternate shedding of vortices in the wake of the cable. Aeolian vibrations are characterized by small-amplitude high-frequency flexural oscillations and, whenever not properly controlled, can induce wear damage and fatigue failures of the conductor. The standard technical approach to the assessment of aeolian vibrations and residual life of overhead conductors is based on the Energy Balance Method (EBM) and the Poffenberger-Swart formula for bending stresses [1]. This approach relies on the main simplifying assumption of mono-modal oscillations, introduced by the EBM. Typical aeolian vibration records, however, clearly show that several modes can be simultaneously excited due to wind variations in time and along the span [2]. Aiming at overcoming the mono-modal vibration assumption of the EBM, a stochastic model was recently proposed by the authors. Wind forces are modeled as a narrow band stochastic process, centered around the Strouhal frequency of the conductor and with arbitrary cross-correlation in space. The model can also account for the presence of additional dissipation devices attached to the conductors [3]. The model is herein applied to achieve a full probabilistic description of the Poffenberger-Swart bending stresses, making a further step towards a more refined definition of the expected life of overhead conductors. References [1] EPRI - Electric Research Power Institute, 2006. Transmission Line Reference Book: Wind- induced Conductor Motion. U.S., Palo Alto. [2] Denoël V., Andrianne T., Real-scale observations of vortex induced vibrations of stay-cables in the boundary layer, Procedia Engineering (2017) 199: 3109-3114. [3] F. Foti, V. Denoël, L. Martinelli, F. Perotti, 2020. A stochastic and continuous model of aeolian vibrations of conductors equipped with Stockbridge dampers. Proc. of the XI International Conference on Structural Dynamic, (Eds.) M. Papadrakakis, M. Fragiadakis, C. Papadimitriou, Vol. 1, pp. 2088-2102.

For the modeling of stochastic structural responses due to wind gusts, reasonable assumptions for the aerodynamic admittance are elementary. However, still, a large variety of corresponding models exists, and it lacks validation and reliable model estimation studies based on real-data. In this paper, monitoring data from a high-rise tower in Rotterdam is used to inversely identify the aerodynamic admittance (size-effect and joint-acceptance functions, SEF, JAF) based on Maximum Likelihood Estimates (MLE), Markov-Chain Monte Carlo (MCMC) sampling and a Bayesian posterior (BP) implementation to find parametric models for both aspects of aerodynamic admittance. Based on measurements of the high-rise tower "New Orleans," located in the center of Rotterdam, the mentioned aspects of dynamic wind load effects and the corresponding wind-induced vibrations are analyzed in detail. Available pressure data allow for the coherence analysis for a floor level. Overall accelerations are used to identify the joint acceptance functions, representing the wind load coherence in the context of the structural model properties. All analytes are performed in the frequency domain. The determined models are also used to propose a concept of time-domain analyses, taking into account the coherent composition of the load components. Present results show valuable results of coherence structure based on the pressure measurements. As such results are rare for on-site measurements, it is of particular interest to compare the model estimations (based on Bayesian inference) to known prediction models. Further comparisons will be presented with respect to the overall structural response and the causing dynamic load process. The inversely determined joint-acceptance function allows for a model comparison between frequency and time-based analyses (using correspondingly generated artificial time series of wind forces). Finally, the accuracy of the predictions is also discussed on the basis of the identified uncertainties using the Bayesian statistical concepts.

A very slender footbridge across the 32 rails of the central main station with a total length of 240 m had been evaluated for aeroeleastic effects. The drag and lift coefficients and the Strouhal number have been evaluated numerically with RANS. To avoid galloping effects, a certain small damping must be reached. It has been decided to measure the real damping before installing a tuned mass damper. The measurements during construction showed a sufficient damping value, but after completing the bridge better measurements showed a significant smaller value, which would trigger galloping effects. As the installation of tuned mass dampers was not only expensive, but also an issue for the interruption of railway services, a nonlinear wind history analysis has been performed, showing that the accelerations of the oscillations remain within the critical limits for a footbridge.