Phénomènes vibratoires observés sur des vannes-segment déversantes

Docteur ès Sciences Chargé de cours à l'Ecole National Supérieur des Mines de Saint-Etienne Chef du Département Recherches de la Société B.V.S à Grenoble.

Résumé

Vibrations in the vertical stiffeners of the gate lip (section located below the last main girder) have been discovered on a 15 m wide, 11.30 m high radial gate topped with a 12 m wide, 2.50 m high overflowing flap. It is known that for an overflowing jet as wide as the channel, a free surface higher than the tail-bay level under river flow conditions is created within the space between the spillway gate and the jet (Figs. 1 and 2). This increase in level, which can be calculated by formulae (3) or (4), depends on the discharge, the head and the angle between the jet axis and the invert at the impact point. For an overflowing jet less wide than the channel (Figs. 3 and 4), there can be no increase in level because the flow diverted towards the upstream side is divided into two parts on either side of the sluice axis ; then it passes along the gate towards the two banks where it deviates and flows away into the tail-bay through the two side passages between the overflowing jet and the channel walls. Thus, two cross currents sweep the gate-lip and all its vertical stiffeners. A "Karman street" type of eddy wake appears downstream of each stiffener but not exactly with the same periodicity. In fact, the vortices concerned are affected not only by the movements of the jump but also by the random pulsations peculiar to the separation points due to asymmetrical singularities. However, experience has shown the presence of a prevailing frequency which can be defined by the theory of "periodic wakes" and which satisfactorily complies, with Strouhal's law. By simplifying the phenomenon, it can be said that the hydraulic excitation exists about bout the lower frequency of the jump (0.5 to 1.5 Hz) and the higher frequency of "periodic wakes" (8 to 11 Hz for the gate concerned). If the natural frequency of the vertical stiffeners is much greater than the excitation frequencies, the resulting oscillations are forced and consequently these structures are rarely endangered. On the contrary, if the natural frequency of the stitffeners is close to the excitation frequency, there is some risk of failure even if the resonance is intermittent over short periods. This paper gives the methods of calculation to avoid such risks of failure. The results have been confirmed by model and prototype tests. The velocity pulsations of the cross currents along the gate can be found on the diagram measured on a 1:20 scale model (Fig. 5). The two prevailing frequencies can be clearly seen and the transverse velocities at the vertical stiffeners vary from 0 to 9,8 m/sec. The measurements made on the prototype (Figs. 7 and 8) have confirmed the prevaling frequencies on the model. The natural frequency of stiffeners has been measured on a 1/4 scale model (Fig. 6). It amounts to 8.5 Hz, which varies somewhat from the steel plate prestressing data (Table n° 2). The excitation magnitude is considerably influenced by the profile of the stilling basin. The two profides tested are shown in Figure 1). Profile 1 leads to the failure of the structure (Fig. 10), while profile 2 causes no appreciable stress on the structure (Fig. 11).