Geography Reference
In-Depth Information
sheet production and migration (Recking et al., 2009). It
has also been adapted and improved to study step-pools
experimentally (see Zimmermann et al., 2008).
the first study, images of a 2D bed captured in plan-view
were used to characterise the trajectories of bedload
particles entrained by turbulent flow in a channel that was
wide compared to the grainsize (Lajeunesse et al., 2010).
In a second study, a high-speed camera was used to film
the movement of particles flowing down a steep narrow
channel only slightly wider than the grain diameter. In this
study the 2D trajectories of the particles were captured
in side-view of a transparent flume (Bohm et al., 2006).
13.2.2 Particle trajectoriesandvelocities
usingPTV
Particle tracking velocimetry (PTV) and a closely related
technique - particle imagery velocimetry (PIV) were
developed and are widely used in the field of fluid
mechanics (for a review see Adrian, 2005 and Chapter 16,
this volume). The use of these image-based techniques
to study sediment transport has been steadily expanding;
such techniques were first applied to studies of suspended
sediment transport in turbulent flows and more recently
to bedload transport. Tracking a single detected particle
in a series of images is relatively easy and a number of
authors have used image analysis to measure the trajec-
tories of saltating and rolling particles (e.g., Nino et al.,
1994; Hu and Hui, 1996; Lee et al., 2000; Ancey et al.,
2002; Ancey et al., 2003). Tracking sediment flux is more
complex and requires segmentation of all the particles in
the image as a first step. Segmentation can be achieved
by applying algorithms commonly used in PTV. Such
algorithms were specifically developed to track tracers in
a flow in regions where standard PIV algorithms based on
cross-correlations were not well adapted because of large
velocity gradients. These algorithms were then applied to
tracking suspended sediments. Sechet and Le Guennec
(1999) investigated the role of near wall turbulent
structures. Nezu and Azuma (2004) building on a
technique described by Okamoto et al. (1995) used PTV
to characterise particle-laden free surface flows. Studies
tracking coarse material are rare: Pilotti et al. (1997)
analysed incipient motion of dark coloured grains on a
light coloured smooth bed, Papanicolaou et al. (1999)
tracked green glass beads over a layer of fixed transparent
ones. In both cases the number of particles being tracked
was low and the contrast between the grains and the bed
was sufficiently high to allow segmentation using simple
colour thresholding procedures. In these studies only the
motion of segmented particles was tracked, rather than
the entire bed. In a more sophisticated application, Capart
et al. (1997) investigated water-sediment interaction
following a dam-break using segmentation to track the
motion of 6 mm plastic beads. More recently, Spinewine
et al. (2003) extended the work of Capart et al. (2002) to
the study of granular flows.
We present two flume studies that use PTV to recon-
struct 2D trajectories of grains transported as bedload. In
13.2.2.1 2D trajectories in plan-view of individual
grains transported as bedload
The goal of the study presented here was to investigate
the motion of bedload particles over a flat bed of uniform
grainsize under steady, spatially uniform, turbulent flow.
The methodology consists of capturing images in plan-
view of bedload particles moving across a bed using a
high-speed camera mounted directly above the flume.
From image analysis, particle velocities, step (or flight)
lengths and durations, and the surface-density of moving
particles can be measured (vertical velocity and saltation
height cannot be measured in plan-view). Details of the
study and the methodology can be found in Lajeunesse
et al. (2010).
The experiments were carried out in a rectangular
tilting flume with a width of 0.096 m and length 0.24 m.
Water was injected by a pump at the upstream inlet of
the flume and flowed over a sediment bed composed
of quartz grains several centimeters thick. Three series
of experiments were performed with a D
50
of 1.92, 2.6
and 5.5 mm respectively. No additional sediment was
fed at the upstream and as a result an erosion wave
slowly propagated from the inlet towards the outlet of
the flume. All the experiments were stopped well before
the erosion wave reached the study reach situated in the
middle of the flume and therefore it did not influence
the results (this was also verified by measurements of the
bed slope at the start and end of the experiment which
showed that it remained constant). Approximately 10%
of the sediment particles in the bed were dyed black.
The rest were the natural colour of quartz (i.e., clear to
white). A high-speed camera (250 frames per second,
1024
1024 pixels) mounted vertically directly above the
bed was used to track the motion of the sediment particles.
The position of the black particles could be tracked
between successive frames with an interval of 0.04 s and
the goal was to determine velocities and trajectories of
each dyed bead from the temporal sequence of images
(Figure 13.3). The spatial resolution of the camera was
×
Search WWH ::
Custom Search