Geoscience Reference
In-Depth Information
resulting deposit and discussing some of the most
influential parameters in the simulation model.
of  spherical glass beads (ø = 0.040 mm); 2) well-
sorted, very fine sand composed of spherical glass
beads (ø = 0.069 mm); and 3) moderately-sorted,
fine-grained natural sand (ø = 0.235 mm). The sus-
pended sediment concentration in the mixing tank
ranged from 14 vol. % to 35 vol. %.
Laboratory set-up
The laboratory experiments were carried out in the
flume facility at Utrecht University. The laboratory
setup consisted of a straight, inclined channel lead-
ing from the feeder tank to a horizontal expansion
table, all submerged in fresh water (Fig.  8). This
set-up imitates a canyon or a channel-confined cur-
rent expanding upon entering a wide and flat basin
floor. The channel was 4 m long, 0.22 m wide and
0.50 m deep and had an inclination of 8.6°. The
expansion table was 3.5 m long, 3 m wide and free
of sidewalls, to minimise flow reflections. It was
installed in a water tank with an area and depth of
4 × 4 m 2 and 2 m, respectively (Fig. 8). The expan-
sion table was submerged in 0.8 m of water. The
sediment was supplied to the mixing tank by
means of a conveyor belt, which allowed for up to
15 s of constant slurry discharge. The experiments
were run with sand-sized glass beads which had a
density of 2650 kg m -3 ; comparable to siliciclastic
sediments present in nature. Three types of sedi-
ment were used in the laboratory experiments:
1)  moderately well-sorted, coarse silt composed
Numerical simulation set-up
The feeder tank and expansion table were mod-
elled and included in the simulation domain,
which was set up using a Cartesian orthogonal
computational mesh. The resolution of the com-
putational mesh was varied in order to evaluate
the robustness of the simulation code and to assess
whether results obtained are mesh-independent.
Obviously, the use of a finer mesh implies more
detailed results with regard to sediment distribu-
tion in the flow; however, the global behaviour of
the system (velocity and concentration structure
of the flow and deposit geometry) are not affected
in a significant way by the resolution of the mesh,
confirming the robustness of MassFLOW-3D TM .
The computational mesh used for simulations
presented in this chapter consisted of ~1 × 10 6
active cells. The resolution of the mesh was
uniform from the bed up to the top of the feeder
dimensions
expansion tank:
4000 × 4000 × 1000 mm
channel:
4000 × 220 × 500 mm
mixing tank
Ott
Ott
siphons
Ott
cross-section
800
400
0
0
1000
2000
3000
4000
5000
6000
7000
Fig. 8. Experimental setup for Baas et al . (2004) experiments in the Eurotank flume laboratories at Utrecht University.
Units in mm.
Search WWH ::




Custom Search