Different mechanisms of vortex induced vibration of bridges

Resumen

Resumen en Castellano (extended version in English follows below) La mayor eficacia de nuevos materiales y metodologías de diseño hacen posible puentes de geometrías estilizadas y mayor flexibilidad. El Puente del Tercer Milenio en Zaragoza es un ejemplo de ello. La mayor flexibilidad de estas estructuras las hace más vulnerables frente a la acción del viento en general y a los fenómenos aeroelásticos en particular. La compleja aerodinámica de estas formas no está resuelta. La vibración inducida por vórtices es un fenómeno aeroelástico que ha sido observado en distintos tipos estructurales: chimeneas de aireación, edificios, puentes, etc. En puentes puede afectar a su conjunto a través de la acción sobre el tablero, pilonas o arcos o parcialmente a cables u otros miembros en su estado final o durante la construcción. Así, este fenómeno fue responsable de vibraciones inaceptables en modos de flexión vertical en el Puente de Storebælt en Dinamarca poco antes de su inauguración. El Puente de Storebælt es el objeto del núcleo de este trabajo que se centra en el estudio de los modos de vibración vertical y de torsión inducida por vórtices. La transferencia de energía y mecanismos de la interacción dinámica fueron investigados a partir de ensayos de un modelo de sección a escala 1:70. En este marco se diseñó y construyó un nuevo Túnel de Viento de Capa Límite en el Centro Andaluz de Medioambiente. El túnel fue equipado con un dispositivo de ensayo dinámico y estático de modelos de sección. Este sistema, diseñado según el estado del arte, además de puentes puede ser utilizado para ensayos diversos de distintos tipos de estructuras lineales. El nuevo túnel y el dispositivo de ensayo fueron evaluados satisfactoriamente mediante comparación con resultados de referencia del Túnel de Viento canadiense de ambos Puentes del Tercer Milenio y Storebaelt. Los ensayos del modelo de sección del Puente de Storebælt, objeto de esta investigación, se llevaron a cabo en este túnel y en el Boundary Layer Wind Tunnel I de la Universidad de Western Ontario en Canadá. La respuesta general observada presentó una histéresis considerable, de modo que la amplitud de la respuesta depende sensiblemente si la velocidad crítica se fija aumentando desde velocidades menores o al contrario. El inicio de la vibración se ha asociado al desprendimiento de vórtices en el borde de salida del tablero para ambos modos de vibración. No obstante, los mecanismos responsables de la evolución de la amplitud de las mismas difieren. Así, en el caso de la vibración en modo de torsión se han identificado estructuras asociadas a vorticidad moviéndose por convección con el flujo medio, como responsables de las vibraciones de máxima amplitud. El cambio en la estructura del campo de presiones, de la zona de separación en particular, a partir de un determinado ángulo de ataque se ha asociado con el origen de las estructuras señaladas. La capacidad de disipación de energía del sistema ha sido caracterizada para las distintas fases de evolución de la respuesta. Esta caracterización permite la evaluación del nivel de amortiguamiento aerodinámico, concepto ampliamente utilizado en los modelos de predicción de la respuesta. Finalmente y en contraste con los mecanismos de vibración anteriores que involucran al puente de forma íntegra, se presenta el caso una tipología particular de puentes en construcción. En concreto el estudio se centra en un problema recurrente de tableros formados por vigas doble T Vigas antes de que se ejecute la sola sobre la mismas. Se ha identificado la emisión de vórtices a sotavento del alma como responsable de vibraciones -en línea- en flexión lateral. Por último se proponen simples medidas aerodinámicas correctivas. The vortex induced vibration of the cylinder has been studied for over a century. Though significant progress has been made, there are fundamental questions that still do not have a clear answer. The aerodynamics of other bluff bodies with an afterbody involve mechanisms of a different nature, including the vortex shedding in the near wake responsible for the cylinder vibration. Such mechanisms are: impinging shear layer instability, leading edge vortex shedding, trailing edge vortex shedding and the interactions between them. The present study focuses on the vortex induced vibration (VIV) of modern bridges. In the framework of this study a new boundary layer wind tunnel was built at CEAMA, University of Granada. The wind tunnel was equipped with a rig for both dynamic and static testing of section models. The wind tunnel itself and the section model test rigs were benchmarked with available results from BLWTL, The University of Western Ontario for the two bridges being studied: the Storebaelt Bridge (Denmark) and the 3rd Millenium Bridge (Spain). The experimental work was carried in both wind tunnels I and II at the BLWTL and CEAMA wind tunnel I. The interest in the above mentioned bridges is due to their different nature. The Storebaelt bridge (1998) has a shape, common nowadays in single deck long span bridges. Despite the numerous studies that had been carried out, the Storebaelt Bridge exhibited unacceptable vertical VIV in March, 1998 when it was just about to be inaugurated. The 3rd Millenium deck cross section is fully curved, incorporating carefully aerodynamically shaped leading and trailing edges in an attempt to reduce the vorticity generation. Modern bridges have shapes that are a cross between bluff bodies and airfoils. There is a need for fundamental physical investigation of the mechanisms that trigger the VIV of these shapes with complex aerodynamics. Some of the mechanisms involved are known to be motion induced; consequently, simultaneous measurements of force and response of the bridge undergoing both vertical and torsional VIV were carried out. The forcing is studied in detail, looking at the simultaneous pressure field throughout the different phases of the response for increasing and decreasing wind speeds. The pressure field is analyzed in detail by means of the phase averaging technique. Vortex shedding at the trailing edge (TEV) seems to be the only source of significant pressure fluctuation of the motionless deck. The motion induced by these trailing edges vortices may be responsible for the control of the impinging shear layer instability at the leading edge. The onset of the oscillation is therefore attributed to TEV. Significant pressure fluctuations are observed at the leading edge when the motion takes place. The build-up of larger amplitude vibrations appears to be caused by leading edge vortices generated by motion induced instability at the natural frequency. A mechamism that has been referred to as vortex interaction of these vortices and those forming at the trailing edge could be responsible for the enhancement of maximum amplitude oscillations, however this requires further investigation. An increasing phase is observed in the force, with respect to the response, reaching values of 90 deg. consistent with a typical resonant response. The physics of the torsional VIV appears to be of a somewhat similar nature. It is observed that trailing edge vortex shedding may be responsible for the onset of the torsional vibration. The build-up of larger amplitude oscillations is however attributed to a motion induced impinging shear layer instability on the upper flange of the leading edge. Such a mechanism is observed to trigger impinging leading edge vortices that are convected by the free stream and is indeed responsible for the enhancement of maximum amplitude oscillations. A constant ¿ phase is observed throughout the build-up stages between force and response consistent with the described mechanism. It is noted that the convective wind speed is somewhat higher than for the vertical mode. This could be possibly explained by the additional inertia that may be provided by the considerable deck rotation to the flow structure. Significant hysteresis was observed in both the vertical and torsional response when the maximum amplitude oscillation wind speed was approached from higher and lower wind speeds. The nature of this hysteresis may be related to the effect of the free-stream wind speed in the instability of the motion induced structures. Nevertheless, the phenomena requires further investigation. No VIV was observed in the tests of the 3rd Millenium Bridge. The spectra of hotwire measurements in the near wake did not show any peak that one could relate to vortex activity. Furthermore, the pressure measurements on the trailing edge did not present a peak either. Finally, numerical simulation of the flow around the structure did not show any significant vorticity in the near wake, in good agreement with the experimental observations. The consistency of the experimental and numerical pressure field on the leading and trailing edges was remarkable. The design of the edges of the curved deck of the 3rd Millenium bridge proved to be successful in mitigating VIV. The vortex induced vibration may be observed in different bridge components other than the deck, such as cables or at different stages of the construction. In contrast with the above mentioned phenomena, VIV of structural member of a Bridge under construction is studied. Evidence is presented of along wind VIV of slender beams set on top of the pylons waiting for the deck to sit over them. The design of such beams provides them with significant stiffness in vertical bending to support the deck while very little in lateral bending. A computed modal frequency equal to two times the shedding frequency was found for the corresponding Strouhal number and measured wind speeds of anemometers nearby.