Effects of tidal bores on turbulent mixing: a numerical and physical study in positive surges

Abstract

Tidal bores are positive surges created by the flooding of tides inside estuaries and rivers. They propagate upstream in an estuary sometimes as several metre high waves. Tidal bores are fascinating phenomenon with some impact on the wildlife, the mixing and re-suspension of sediments, and river constructions. Field studies remain challenging due to the violent nature of some bores. Herein, the hydrodynamics of tidal bores was investigated in a complementary study with experimental measurements and numerical simulations of idealised models of tidal bores. While laboratory experiments offer an accurate way to measure the flow properties, numerical simulations provide rather detailed information on the hydrodynamics of bores. These simulations require detailed experimental data to accurately replicate the experiments and validate the numerical methods.Investigations were conducted at the University of Queensland in a large rectangular channel partially covered by a fixed gravel bed. The steady flow condition properties were investigated using a Pitot tube and acoustic Doppler velocimeters (ADVs). The unsteady flow conditions were obtained by the fast closure of a downstream gate, which led to an upstream propagating bore. Free surface characteristics were studied for several types of bores, from undular to breaking. Those measurements showed large free surface fluctuations during the bore passage in agreement with earlier findings. Velocity measurements were conducted at high frequency (200 Hz). In breaking bores, the ensemble averaged data exhibited large velocity reversals near the bed beneath the bore front and in its wake, whereas, in undular bores, a lesser deceleration was observed. The vertical velocity data trend closely followed the water depth time derivative for both breaking and undular bores. Overall, the velocity measurements displayed large velocity fluctuations during the passage of the bore which were believed to induce strong turbulent mixing. The turbulence was further investigated by a two point correlation technique to estimate both turbulent integral time and length scales. The turbulent integral length scales values were similar in the steady and unsteady flows, whereas the turbulent time scales were larger in the unsteady flow. The experiments were limited by the instrumentation since the ADV systems interacted with each other and lacked accuracy for measurements performed in the near vicinity of physical boundaries.A numerical investigation was then performed using the computational fluid dynamics (CFD) ii code Thétis developed in the I2M laboratory of the Université de Bordeaux. Numerical simulations reproduced experiments performed in laboratory prior to this study (Chanson, 2010a,b). The fluid was modelled by the Navier-Stokes equations in their incompressible two-phase form. The turbulence model used a large eddy simulation and the interface air-water was tracked by using a volume of fluid model. Two dimensional and three dimensional simulations were performed to study the comp...

Two-dimensional integral turbulent scales in compression wave in a canal

“…KHEZRI (2014) In summary, the present study agreed qualitatively and quantitatively with past CFD numerical data (KHEZRI 2014, SIMON 2014, LENG et al 2018 in terms of order of magnitude. The two-dimensional cross-correlation data in the y-z plane, formed by the two sampling profiles of the Profiler array, showed the existence of large scale coherent structures underneath the free-surface.…”

CFD modelling of surface wave breaking in a long channel

“…This method requires prescribed eddy size to be convected. [4,12] recommended over-estimation of the eddy size for a more quickly stabilised model and better approximation of the real flow structures. In the present study, the prescribed size of eddies were one order of magnitude larger than the turbulent length scale observed experimentally [7].…”

Transverse velocity profiling under positive surges in channels

“…Spatial variations with transverse distance of maximum cross-correlation coefficient R max and optimum time lag T i of the longitudinal velocity component during the three flow phases of a positive surge -Flow conditions: Q = 0.101 m 3 /s, d 1 = 0.175 m, Fr 1 = 1.47, x = 8.5 m, z/d 1 = 0.17, transverse Profiler sampled alone6.3 Turbulent time and length scales before and during a positive surgeThe integral turbulent time and length scales were calculated using Equations (4) to (7) for measurements with single Profiler and Profiler array, independently. The results are presented in Appendix I, including some comparison to past data(Simon and Chanson 2013, Leng andChanson 2017). The complete data set is detailed inLeng and Chanson (2018).Overall, the integral turbulent length scales were of an order of magnitude of 10 -2 m to 10 -3 m, and turbulent time scales were between 10 -2 s and 10 -1 s, corresponding to dimensionless length scales of L/d 1 ~ 0.01-0.1 scales of T×(g/d 1 ) 1/2 ~ 0.1-1.0.…”