A. Méndez
CITEEC, University of Coruña, Spain
CITEEC, University of Coruña, Spain
Fluid Mechanics. CPS. University of Zaragoza, Spain
Civil Engineering School, University of Coruña, Spain
Abstract: This paper presents the results of a series of experiments on “Dam Break” for the purpose of providing data to serve as a comparison for a numerical model. A typical process of “Dam Break”, in which water flows under extreme conditions, was simulated in a tank divided into two parts and equipped with a watergate at the hydraulics facilities of the CITEEC (Research Center in Civil Engineering) in La Coruña (Spain).
As the watergate opens, the water flows from the full part of the tank into the empty area. The experimental design allowed us to measure depths and water velocity throughout the whole process, from the opening of the gate until the water reached a state of relative calm to obtain data on the temporal evolution of the flood to be compared with those acquired numerically. The “Dam-Break” process was simulated by setting several depths of water in the downstream tank allowing the free flood of water in the first series, while in the second series an obstacle was added to the experimental design in the downstream tank. The entire process of controlling the instrumentation and data acquisition was monitored by software.
The numerical model developed by Pilar Brufau [1] solves the Reynolds equations [2] by simulating the energy dissipation using the Manning coefficients and finite volumes [3]. To test the model in an extreme case such as a “Dam Break” process, a series of experiments were performed in the hydraulic facilities of the CITEEC (La Coruña).
A typical “Dam Break” process was simulated in a rectangular tank (see Fig 1.), divided into two parts by a wall, in which a watergate was suddenly opened. The watergate was raised and lowered by means of an air compressed piston controlled electronically. One part of the tank was filled with water (50 cm depth). As the watergate rose, the water, in the full part of the tank, flowed into the other area through the opened gate. Several series of experiments were done with different depths of water in the downstream tank: 40 cm, 35 cm, 30 cm, 25 cm, 20 cm, 15 cm, 10 cm, 5 cm and 2.5 cm. The entire process –from a few seconds before the opening of the watergate until the water reached a state of relative calm was monitored by a 19 Wave Meter (see Fig. 1) and four tridimensional velocimeters located as shown in Fig 2 (see section 2 for details).
In the first series of experimental measurements, the downstream tank was absolutely free. A second series was carried out by placing half of a pyramid beside one of the walls in the downstream tank. Some of the Wave Meters were positioned near this intrusive element (see Fig 2).
Fig. 1 Design and
dimensions of the tank for the “Dam-Break” process.
The numbers represent
Wave gauges. The reference system is also shown.
Fig. 2 Design of the tank for the “Dam-Break” process with an intrusive element in the shape of a pyramid. In this experiment, five probes for measuring the height of the water were placed near the pyramid. Probes marked as V1, V2, V3 and V4, represents the velocimeters. V2 and V3 are laboratory velocimeters (see 2.1); V1 and V4 are field velocimeters.
All instrumentation and data acquisition were software controlled. Data were compiled using a specially designed Visual Basic program, which also controlled the movement of the watergate.
Since data acquisition was sequential,
the experimental measurements were interpolated using linear interpolation and
cubic splines to obtain all the measurements referring to the same temporal
instant to improve the comparison with the results of the numerical model.
The instrumentation used allows us to monitor depths and water velocity during the “Dam Break”. For a complete description of the flood, 19 depth probes and 4 tridimensional velocimeters were used. The Wave Meters are based on the variations of electrical conductivity measurements in two electrodes, which depend on the quantity of the water between the electrodes. The velocimeters, SonTek MicroADV’s, are based on the physical principle of the Doppler effect and enables the velocity to be measured in a sampling volume in 3 directions. Further information is given below.
The SonTek ADV is a single point, Doppler current meter. It measures the velocity of the water using the Doppler effect, that is, the shift of the frequency received with respect to the frequency transmitted when the source is moving relative to the receiver. The velocimeter has one ultrasound transmitter and three receivers. The transmitter generates a short pulse of sound at a known frequency. The energy of the pulse passes through the so-called sampling volume (a small volume of water in which measurements are taken). Part of this energy is reflected by suspended matter along the axis of the receiver, where it is sampled by the velocimeter, whose electronics detect the shift in frequency. According to this, to obtain measurements with a velocimeter based on the Doppler effect, the presence of suspended matter is necessary for an accurate reflection of the pulse.
Two kinds of ADV’s were used in these experiments: Laboratory ADV’s and Field ADV’s. The principal difference between the two is the sampling volume, which is located 6 cm from the transmitter in the Laboratory ADV and at a distance of 11 cm in the Field ADV. In the experimental design of the experiments with pyramid four velocimeters, two Laboratory ADVs and two Field ADVs were used, positioned at the locations as indicated in Fig. 2.
A reliable measurement of the velocity can only be had if both the transmitter and the receiver are covered by water. In order to obtain a reasonable tracking of the whole process for most of the initial depths, the volume was located 10 cm from the ground. Under these conditions, the measurements taken with the laboratory ADV’s were valid starting at a depth of 16 cm, while the Field ADV’s required 21 cm of water. Velocimeters were employed only in the second part of the experiment, when the pyramid was placed in the downstream tank.
To measure the evolution of depth, 19
Wave gauges from The Danish Hydraulic Institute were installed at the locations
indicated in Fig. 1. Five of the probes allowed us to control the depletion of
the dam, while the other 14 provided the evolution of the filling process. In
the second series of experiments five probes were moved to be in the proximity
of the pyramid (see Fig. 2 for details). The distance from the bottom of the
Wave Gauge to the tank bottom is 6.5 cm. This means that measurements are only
valid at a water height of over 6.5 cm.
A complete DHI Wave Meter includes the Wave Gauge, which is connected to the Wave Meter conditioning module. A DHI Wave Meter is based on a conductivity type wave gauge, which is comprised of two thin, parallel stainless steel electrodes. When immersed in water, the meter measures the conductivity of the instantaneous volume between the two electrodes. The conductivity changes proportionally to changes in the surface elevation of the water between the electrodes.
An analogue output is taken from the Wave Meter conditioning module and compiled in the data acquisition system. This makes it possible to automatize and centralize data collection for all probes.
The raising and lowering of the water gate, as well as all instrumentation are computer controlled. Data collection is carried out by means of a specially designed Visual Basic data acquisition program.
Data acquisition is sequential, which means that the measurements for each probe refer to a different time. In order to obtain all the measurements referring to the same time, the data of two consecutive measurements were interpolated. In the case of the Wave Meters, water height was measured every 0.002 s. For velocimeters, the frequency of the measurements was 50 Hz.
Experiments with different ratios between the water height in the upstream and downstream tanks were performed. For each relation, three series of data were collected in order to check the data.
The following ratios of water height were used in experiments without the pyramid: 50/40, 50/35, 50/30, 50/25, 50/20, 50/15, 50/10, 50/5 and 50/2.5 cm.
After inserting the pyramid, experiments having a ratio of 50/30, 50/20 and 50/10 cm were performed.
In order to obtain the measurements of each probe referring to the same instant of time, a data interpolation between two consecutive measurements was made. As mentioned earlier, data acquisition is sequential; 19 consecutive measurements, one for each probe, are taken every 0.002 s. The process is repeated until the data have been collected.
Interpolation time was chosen to be the average point between the time at which the last measurement for one series is received and the beginning of the following series. This enables the interpolation between two points. Two kinds of interpolation were performed; linear and with cubic splines. Fig. 3 shows a comparison of an original series of data and the same data with linear and spline interpolation.
Fig. 3 Comparison of
original and interpolated data in an experiment
with a water height ratio of
50/25 cm with data from probe 4.
Negative times correspond to time prior to the
elevation of the watergate.
Several outputs of various water height ratios are shown below. Each plot includes the three experimental series of data. The full series are available by sending an e-mail to citeec@udc.es, the origin of the data must be cited if used in a publication
A measurement of the evolution of the velocity is plotted below. Here, the mean value of 20 consecutive measurements, is shown. The reference system is shown in Fig 1, taking into account that the z axis is positive upwards.
Some measurements of the water height near the pyramid are given in the following section, where the comparison between the experimental and the numerical data is plotted.
The experimental data were used to test the numerical model under the extreme conditions represented by a Dam Break process. The predictions of the numerical model coincide with the data measured for water height, as shown in Fig. 11, 12, 13 and 14, which are plotted for a ratio of 50/10 cm between the upstream and the downstream tanks, with the pyramid.
This paper presents an experimental approach to the Dam Break problem. A typical process of Dam Break has been simulated in a tank divided in two parts and equipped with a watergate. The experimental design allowed us to measure depths and velocity troughout the whole process. Several series of experiments were made with different depths of water in the downstream tank. In the first series of experimental measurements, the downstream tank was free. A second series was carried out by placing half of a pyramid in this tank.
The experimental
results were used to test a numerical model, whose predictions are in good
agreement with the measured data. These data are available for further
calibrations worldwide.
References
[1] Brufau, P.; Simulación bidimensional de flujos hidrodinámicos transitorios en geometrías irregulares. PhD thesis, Universidad de Zaragoza, 2000.
[2] Warsi, Z. U. A.; Fluid Dynamics, CRC Press, 1998.
[3] Versteeg, H. K.; Malalasekera, W.; An Introduction to Computational Fluid Dynamics, The Finite Element Method, Longman, 1995.