The objectives of the thesis were, on the one hand, to test the possibilities of a well-known experimental configuration (oscillating grid mixing box), used to study stratified fluids in turbulent flows with zero-mean, on the study of sediment transport, and to find out what aspects involved in this phenomena could be explored from this new perspective. On the other hand, compare a well-established criterion, such as Shields parameter, with the local turbulent energy ( ), as well as to compare the differences between zero-mean and sheared induced lift-off.Verification and adjustments to the parameterizations given in the bibliography for the r.m.s. turbulent velocity, , generated by an oscillating grid were done as well as the calibration of the different mixing boxes and grids to be used through a series of experiments. This gave a clear view of the behaviour of the flow in all the mixing boxes used. A special experimental configuration was designed and performed for the first time, and measurements were taken to get to know the flow properties in the region between the grid and the bottom (solid boundary). Experiments were performed with actual sediment with different characteristics, for instance, different sediment sizes (from clay to medium sands) including samples made of single sediment size and samples extracted from the seabed. An experiment to compare the sediment behaviour under a sheared flow (recirculation tank) and under a no-sheared flow (oscillating grid mixing box) was designed and performed. The present work concluded that the experimental configuration used could be employed in further works to investigate aspects of sediment transport under a turbulent flow with great accuracy. This experimental configuration, in combination with techniques of image analysis, greatly improves the capabilities of the grid stirred experimental configurations, mainly on the study of sediment behaviour. The flow behaviour, between the grid and the free surface, and between the grid and the bottom, is, in general terms, the same. Other important contributions of the present work are the measurements taken near the boundary. Vortices do not distort before impingement. After vortex impingement, no constant mean flow is distinguished. The former enhanced the idea that these experimental configurations are suitable for studying the behaviour of a sediment bed under the influence of a turbulent flow. A smaller magnitude of than of was required to start sediment motion, through comparing the theoretical critical friction velocity, (Critical Shields Parameter), needed to start sediment motion with the measured critical for different sediment sizes. The results obtained show the importance that the time span between extraction and analysis has on the measured value of , and that cohesiveness of sediment play a more important role than sediment size hampering the erosion processes that cause sediment lift off from the sediment bed. Under the same circumstances, a turbulent flow is more efficient than eroding the sediment bed, since it does not depend on the roughness of the sediment bed. It was possible to measure sediment velocities during lift-off and settling, using grid stirred experimental configurations in conjunction with image analysis techniques. It is even possible to determine the sediment size (size mode), even in the range of clays and silts, with an error as small as 6 or 8 percent. It was possible to quantify the sediment flux at a given turbulence intensity. Velocity fluctuations are about the same in the three directions and in a random way above the sediment bed, meaning no preferential direction of stresses. The former hampers the process of particle imbrication making it easier to dislocate the particles from the bed.