UCL dam-break benchmark 101013

Description

PIRE workshop on dam-break flows – November 2010 Proposed benchmark: two-dimensional dam-break flow on movable bed S. Soares-Frazão, C. Swartenbroekx and Y. Zech Description of the test case This test aims at investigating the two-dimensional morphological evolution of a movable bed made of uniform coarse sand under the action of a dam-break wave. The experiments were conducted in a 3.6 m wide and about 36 m long flume (Figure 1). The breached dam is represented by two impervious blocks and a 1-m wide gate located between the blocks. To simulate the dam-break, the gate located at about 12 m from the upstream end of the flume is pulled up rapidly. A detailed sketch with the flume dimensions is provided at the end of this document. To simulate the morphological evolution, the fixed bed of the flume was covered with a 85 mm thick sand layer. As illustrated in Figure 1 and in the detailed sketch, this sand layer extends over 9 m downstream of the gate and over about 1 m upstream of the gate. Downstream, the sand layer is blocked by a little weir with a height equal to the layer thickness. Figure 1. General view of the flume with the initial sand layer (view from downstream) Test conditions Two different test conditions are proposed, as summarized in Table 1. The initial water level in the reservoir is denoted by z , while the initial water depth in the downstream part is 0denoted z . Water levels are measured with reference to the fixed bed. ...

Informations

• Publié par: Movong
• Nombre de lectures: 57
• Langue: English

Extrait

PIRE workshop on dam-break flows – November 2010

Proposed benchmark: two-dimensional dam-break flow on movable bed

S. Soares-Frazão, C. Swartenbroekx and Y. Zech

Description of the test case

This test aims at investigating the two-dimensional morphological evolution of a movable bed

made of uniform coarse sand under the action of a dam-break wave. The experiments were

conducted in a 3.6 m wide and about 36 m long flume (Figure 1). The breached dam is

represented by two impervious blocks and a 1-m wide gate located between the blocks. To

simulate the dam-break, the gate located at about 12 m from the upstream end of the flume is

pulled up rapidly. A detailed sketch with the flume dimensions is provided at the end of this

document.

To simulate the morphological evolution, the fixed bed of the flume was covered with a

85 mm thick sand layer. As illustrated in Figure 1 and in the detailed sketch, this sand layer

extends over 9 m downstream of the gate and over about 1 m upstream of the gate.

Downstream, the sand layer is blocked by a little weir with a height equal to the layer

thickness.

Figure 1. General view of the flume with the initial sand layer (view from downstream)

Test conditions

Two different test conditions are proposed, as summarized in Table 1. The initial water level

in the reservoir is denoted by

z

0

, while the initial water depth in the downstream part is

denoted

z

1

. Water levels are measured with reference to the fixed bed. The gate is pulled up

rapidly so that the dam-break can be considered as instantaneous.

Table 1. Test conditions

z

0

(m)

z

1

(m)

h

s

(m)

Test 1

0.47

0.00

0.085

Test 2

0.51

0.15

0.085

In both tests, the sand layer of thickness

h

s

= 0.085 m is fully saturated before the experiment.

Sand characteristics are as follows:

d

50

= 1.61 mm, specific gravity

ρ

s

/

ρ

w

= 2.63, initial bed

porosity

ε

0

= 0.42. A roughness Manning coefficient value can be taken as

n

= 0.0165 on the

sediment; for the fixed bed a value

n

= 0.010 can be adopted.

The upstream boundary condition is a closed wall. The downstream end of the flume consists

of a weir and system for sediment entrapment that cannot be modeled accurately. However,

this downstream boundary condition is far from the downstream end of the sand layer and

does not play any role during the period of interest. A suggestion is thus to simulate it either

as a closed wall or as a transmissive condition.

Available measurements

•

Water level evolution in time at 8 gauging points, measured by means of ultrasonic probes

over a total period of 20 s after gate opening. The location of the gauges is indicated in

Table 2 (the coordinate origin is indicated on the general sketch).

Table 2. Gauges locations for test 1 and test 2

Test 1

Test 2

Gauge n°

x

(m)

y

(m)

x

(m)

y

(m)

1

0.64

-0.5

0.64

-0.5

2

0.64

-0.165

0.64

-0.165

3

0.64

0.165

0.64

0.165

4

0.64

0.5

0.64

0.5

5

1.94

-0.99

2.34

-0.99

6

1.94

-0.33

2.34

-0.33

7

1.94

0.33

2.34

0.33

8

1.94

0.99

2.34

0.99

•

Final bed topography. Longitudinal bed profiles were measured in a continuous way from

x

= 0.5 m to

x

= 8 m by means of a bed profiler. Profiles were measured over the whole

width of the flume, with a

∆

y

spacing of 0.05 m. Combination of all profiles allows the

reconstruction of a two-dimensional view of the final bed topography. An indicative

example of such final topography reconstruction is given in Figure 2.

Figure 2. Measured final bed topography

Simulation results

The following results are expected from the participants:

•

A brief description of the simulation: equations used, numerical model, mesh type and size,

test conditions (including boundary conditions).

•

Snapshots of the flow at times

t

= 1 s,

t

= 2 s,

t

= 5 s,

t

= 10 s,

t

= 20 s, i.e., in an ASCII or

Excel file, for each computational point or cell, the coordinates

x

and

y

of the computational

point or cell, the bed elevation

z

b

, the water level

z

w

, the unit discharges

q

x

and

q

y

, the sediment

transport rates per unit width

q

sx

and

q

sy

.

•

Water level evolution at the 8 gauging points; with

∆

t about 0.1 s, in an ASCII or Excel file,

with clear time values and indication of the gauge number.

Results should be sent directly by email to Sandra Soares-Frazão (

[e-mail]

)

by no later than 15

th

October 2010. Large data sets (>

5Mb) can be sent through the following web

page:

http://transvol.sgsi.ucl.ac.be/index_notucl_UK.php

Comparisons with the measured data will be presented during the PIRE workshop held in Louvain-la-

Neuve (Belgium) on 3-5 November 2010. All participants will receive an access to the experimental

data after this workshop.