Application de la méthode du fil chaud à la mesure de la turbulence dans l'eau deuxième partie exécution des mesures et résultats

The measurement of turbulence in water by the hot wire méthod Part Il : Measurement procedure and results

¹ Docteur ès Seiences, Laboratoire de Mécanique des fluides, Université de Grenoble.
² Ingénieur, Laboratoire de Mécanique des fluides, Université de Grenoble.

Résumé

Once the general measurement procedure had been finalised during the tests described in Part I, measurements were carried out to determine the characteristics of the turbulence produced in the tank. This took place in the following stages : A) Choice of final parameters, especially rotating arm speed. Different turbulence values are obtained by varying the speed of the rotating arm as the turbulent system is determined by the mechanical generator characteristics. B) Actual measurements. 1. Fluctuating velocity. As might be expected a substantial decrease was observed in the turbulent velocities from the bottom part of the tank near the generator to the free surface. 2. Spectra. These measurements too are more difficult than with conventional wind tunnel screen layouts as the characteristic apparent frequencies of the motion resulting from the spatial frequencies of the turbulent motion and average (rotating arm) speed are low. The spectrum shows a very typical Kolmogoroff zone, but the dissipation zone does not conform to the (-7) exponent law predicted by the Heisenberg theory. 3. Correlations. The correlation between longitudinal velocity fluctuations and distance between the two measurement points was measured and the resulting curve used to define the integral transverse correlation length. 4. Micro-scales. These were measured directly with the aid of a differentiation system and calculated from the spectra. A certain number of measurement difficulties were experienced with probe dimensions as compared to measurement lengths and differentiation system pass bands. C) Interpretation of measurement results. 1. Effective turbulent velocities. These velocities were found to satisfy an equation of the 'diffusion' type surprisingly accurately. The energy diffusion 'coefficient' was determined by matching this equation to the experimental curve. 2 and 3. Relationship between characteristic turbulence quantities. The study of the (approximately isotropic) turbulence behind the screen yielded certain relationships between the characteristic quantities. Comparison with the tank measurements showed that these relationships are at least approximately preserved in spite of differences between the basic parameters of the movements obtained, especially Reynolds number. 4. The Kolmogoroff constant was determined by spectrum measurement and by calculating dissipated power from the micro-scale measurement data. Despite fairly appreciable scatter the values found agree well with other researchers' data. As the Reynolds number for turbulent flow in the tank is about 300, it is seen that the critical Reynolds number for the existence of a Kolmogoroff zone is invariably less than 500, which is the figure some researchers have suggested. 5. The measurements also especially showed up the need for very different special instrumentation from that used in air flows, even though the definitions of the measured quantities may be the same in both cases.