wiki:WikiStart

Version 205 (modified by peakall3je, 4 years ago) (diff)

--

Project Name

InfrastructureCNRS_Coriolis
Project (long title)Coriolis and Rotational effects on Stratified Turbulence
Campaign Title (name data folder)16CREST
Lead AuthorJeffrey Peakall
ContributorStephen Darby, Robert Michael Dorrell, Shahrzad Davarpanah Jazi, Gareth Mark Keevil, Jeffrey Peakall, Anna Wåhlin, Mathew Graeme Wells, Joel Sommeria, Samuel Viboud
Date Campaign Start12/09/2016
Date Campaign End21/10/2016


1 - Objectives

Our primary objective is to measure detailed turbulence distributions within channelised gravity currents, as a function of Coriolis forces, concentrating on: i) the bottom boundary layer, ii) redistribution of turbulence within bends, and, iii) redistribution of turbulence at the interface between the gravity current and the ambient. These datasets will enable existing theory on the presence and influence of Ekman boundary layers to be tested, with important implication for the basal shear stress distributions, erosion, and the evolution of channels. These data on the distribution of turbulence will then be applied to i) examine the turbulence distribution in straight channels, ii) provide an analysis of secondary flow and associated turbulence around bends for the first time, and an assessment of how channelized flows alter as a function of Rossby numbers and therefore latitude, iii) assess how the morphodynamics of submarine channels vary as a function of the Rossby number, iv) explain the observed patterns of submarine channel sinuosity with latitude (Peakall et al., 2012; Cossu and Wells, 2013; Cossu et al., 2015), and, v) incorporate the entrainment data into numerical models of submarine channels, in order to address the unanswered question of how these flows traverse such large-distances across very low-angle slopes (Dorrell et al., 2014).Our primary objective is to measure detailed turbulence distributions within channelised gravity currents, as a function of Coriolis forces, concentrating on: i) the bottom boundary layer, ii) redistribution of turbulence within bends, and, iii) redistribution of turbulence at the interface between the gravity current and the ambient. These datasets will enable existing theory on the presence and influence of Ekman boundary layers to be tested, with important implication for the basal shear stress distributions, erosion, and the evolution of channels. These data on the distribution of turbulence will then be applied to i) examine the turbulence distribution in straight channels, ii) provide an analysis of secondary flow and associated turbulence around bends for the first time, and an assessment of how channelized flows alter as a function of Rossby numbers and therefore latitude, iii) assess how the morphodynamics of submarine channels vary as a function of the Rossby number, iv) explain the observed patterns of submarine channel sinuosity with latitude (Peakall et al., 2012; Cossu and Wells, 2013; Cossu et al., 2015), and, v) incorporate the entrainment data into numerical models of submarine channels, in order to address the unanswered question of how these flows traverse such large-distances across very low-angle slopes (Dorrell et al., 2014).


2 - Experimental setup:

2.1 General description

A channel model is positioned within the Coriolis facility. The channel model consists of an initial tapered input section with a honeycomb baffle for flow straightening and turbulence control, a 3.2 m straight channel section, and two bends with a mid-channel radius of 1.5 m. The channel is made of acrylic and is 60 cm wide and 50 cm high; the sinuous section has a sinuosity of 1.2. The slope is 3/50 radians (3.5 degrees, 6% gradient) and the channel terminates 10 cm off of the floor. Saline fluid is pumped into the top of the channel, forming a gravity current, which flows along the channel, and off the end. The basal 10 cm of the flume operates as a sump for the denser saline fluid to accumulate. In turn, this fluid can be drawn down in one of two ways: i) whilst recirculating the fluid, though this is limited to 20 $m3$/hr (5.55 l/s), and ii) through emptying to the drain, in which case any flow rate is possible. Two long metal rails are positioned to either side of the channel model across the full width of the flume. These carry a computerized gantry, which can be positioned at any point along the channel. The gantry itself contains the controls for two Schneider slides, one orientated transverse to the model, and the other connected slide, orientated in the vertical. Thus the system enables xyz control.

2.2 Definition of the co-ordinate system

The co-ordinate system is defined as follows. X is defined as zero at a position 56 cm upstream from the downstream end of the straight channel. This is the position that the basal ADV probe (ADV1) is located at when the traverse is at position X1. Position X4 (at the apex of bend 2) is 7.01 m downstream of position X1. The Y-direction is defined relative to the local channel cross-section at any given X-position. The zero position is defined as the right bank of the channel, as looking downstream. The z-position is defined locally from the bed of the channel, for any given XY position, with zero representing the bed.

2.3 Fixed Parameters

NotationDefinitionValuesRemarks
$Q_0$ Input Discharge $6 \ ls-1$
$\Delta\rho$ Density Difference $20 \ kg \ m-3$
$W$ Channel Width $0.6 \ m$
$\nu$ Viscosity $10-6m2s-1$
$S$ Slope $3.5{\circ}$

2.4 Variable Parameters

NotationDefinitionUnitInitial Estimated ValuesRemarks
$\Omega$ Rotation Rate $rads-1$ -0.18 - 0.15
$H_w$ Water Depth $m$ 1-1.1
$Q_o_u_t_p_u_t$ Output Flow Rate $ls-1$ 5.5 - 17
$k$ Roughness - -

2.5 Additional Parameters

NotationDefinitionUnitInitial Estimated Values
$h$ Depth of gravity current $m$ 0.2-0.4
$U$ Mean downslope velocity $ms-1$ 0.1-0.15
$\delta$ Thickness of Ekman boundary layer $mm$ ~10
$R$ Centreline radius of curvature $m$ 1.5

2.6 Definition of the relevant non-dimensional numbers

Flow Reynolds number across the obstruction, $Re = Uh/\nu$.

Densimetric Froude number, $Fr = U/(g'h){1/2}$, $g' = g(\Delta\rho)/\rho_0$.

Rossby number (width), $Ro_W = U/fW$.

Rossby number (radius of curvature), $Ro_R = U/fR$

Canyon number, $\beta = sW/\delta$.

Bulk Richardson number, $Ri = 1/(Fr{2})$

Keulegan number, $Ke = (Ri/Re?){1/3}$


3 - Instrumentation and data acquisition

3.1 Instruments

Two ultrasonic systems are used to measure velocity.

Ultrasonic Velocimetry Profiling (UVP) Ultrasonic velocimetry profiling (UVP) is a technique that measures a single component of velocity at up to several hundred points along a line. A transducer sends out an ultrasonic pulse, and then gates the return signal into a series of spatial bins. The individual transducers can be multiplexed in order to provide pseudo-velocity fields. The transducers are linked via a multiplexer with a delay of 15 ms, so the two-dimensional velocity field is not instantaneous, however velocity fields can be collected at 3-4 Hz. Two frequencies of transducers were used. An array of ten 4 MHz UVP probes (with 10 m long cables) was used for collecting downstream velocity profiles. Initial test experiments used probes positioned at heights of (centre point of each probe): 7, 16, 26, 56, 86, 116, 146, 176, 206 and 236 mm. The actual experiments used probes are positioned at heights (centre point of each probe) of 10, 25, 50, 75, 100, 150, 200, 250, 350 and 450 mm from the base of the channel. These probes are positioned in a custom made plastic holder, in turn connected to a bar strapped to the channel top. Initially, the probes are positioned on the channel centreline, 80 mm downstream of the apex of bend 2, looking upstream. An array of ten 2 MHz UVP probes (with 4 m long cables) is used to examine the nature of secondary flow at the second bend apex. This involves drilling holes in the apex of bend 2 and inserting the UVP probes. The probes are positioned at heights of 45, 90, 135, 180, 225, 270, 315, 360, 404, and 450 mm from the base. 2 MHz probes are required for the cross-section measurements since the measurement range needs to be much larger (60 cm) than is required for the axial velocity measurements.

Profiling Acoustic Doppler Velocimetry (ADV) Early experiments used 3 co-mounted Nortek Vectrino Profiler Acoustic Doppler Velocimetry (ADV) probes ADVs for mapping three-dimensional flow velocities, whilst later experiments including the main experiments used 2 ADVs to minimize side lobe interference, and because of the loss of one probe. The ADVs were run at 100Hz in all experiments, both in the initial tests and in the actual experiments. The ADV probes measure three component velocities over a depth range of around 30 mm (up to 34 mm), with this zone starting 40 mm below the probe head. Bottom tracking by the instruments enables this depth to be precisely known and controlled, provided sufficient seeding is present in the flow. ADV collection was fully synchronised with the traverse through sending of the required voltage offset signal to them, thus enabling individual velocity files to be collected for each probe at each traverse. Movement of these probes on the traverse in the y and z planes allows detailed vertical profiles and flow mapping to be undertaken. Initial test experiments used the 3 probes, 1 was a stem probe with additional shielding, and 2 were flexible probes. The basal probe was numbered 1, middle 2, and upper probe 3, and were at heights of 7.2, 10 and 13 cm respectively. Lateral offset in the x-direction was 8.5 cm between probes 1 (most downstream probe) and 2, and 7.5 cm between probes 2 and 3 (most upstream). The traverse positions for each cross-section were based on probe 2 being positioned directly above the cross-section. These 3 probes were co-mounted on the traverse in order to collectively measure over a height of approximately 6 cm. ADV dwell times varied between 30 and 60 seconds.


In the actual data collection experiments, two ADV probes were used. The basal probe is numbered 1, and the upper probe is numbered 3 [keeping with the earlier experimental protocol]. The traverse positions are based on the position of ADV1, the basal and most downstream ADV. The offset in the x-direction is 11 cm between probe 1 (most downstream probe) and probe 3 (most upstream) [i.e. when the traverse is positioned in the zero position (56 cm upstream of the end of the straight channel section) then ADV1 is in the zero position, and ADV3 is at -11 cm]. In the Z-direction they are positioned 14 mm apart. The dwell time at each point is 60 seconds. In the initial position the ADVs start 15 cm from the sidewall. Note that in these experiments the conductivity probe is located 7.3 cm downstream from ADV1, and is offset by 5.5 cm such that its initial position is 20.5 cm from the wall. The ADVs are positioned at the following positions from the floor for each of the measurement positions:

Run PositionADV1 (Z-distance above floor)ADV3 (Z-distance above floor)
Straight Section X1
Y=15 cm 7.2 cm 8.6 cm
Y=30 cm 7.2 cm 8.6 cm
Y=45 cm 7.2 cm 8.7 cm
Apex bend 2 – X4
Y=15 cm 6.8 cm 8.6 cm
Y=30 cm 6.7 cm 8.7 cm
Y=45 cm 6.6 cm 8.0 cm

These data suggest that there is some small-scale variation in lateral slope in the bend apex region.

Siphon rig The siphon rig consists of 12 siphon tubes, each with a nominal internal diameter of 1.6 mm (3.2. mm external diameter), and 15 m in length. The siphons are held in place by a plastic holder which is in turn connected to a rod attached to the channel. The siphon rig is positioned 80 mm downstream of the inflection downstream of bend 2, offset laterally 78 mm off-axis from the position of the UVP (on the right hand side as looking downstream). The siphons are positioned at heights of 10, 25, 50, 75, 100, 125, 150, 175, 200, 250, 350 and 450 mm above the bed. The siphon tubes are connected to a 12 head peristaltic pump. Samples are taken when the traverse (with ADVs and conductivity probe) is positioned upstream measuring the straight section. Samples are collected in an array of 12 plastic containers over a 60 second period, starting at 5, 10, 15, 20, 25, 30 and 35 minutes into the experiment (the start of the experiment is defined as that point where the input is steady). Samples are subsequently measured for conductivity using an Anton Paar DMA35 Portable Density Meter.

PME Microscale Conductivity and Temperature Instrument (MSCTI) The Microscale Conductivity and Temperature Instrument (MSCTI) is designed to measure the electrical conductivity and temperature of solutions containing conductive ions. This instrument provides two analog outputs, one linearly proportional to the solution conductivity, and one non-linearly proportional to the solution temperature. One probe is positioned on the traverse, alongside the ADVs, whilst a second probe is positioned at the base of the input box.

Particle Imaging Velocimetry (PIV) A xxxx laser in conjunction with 3 cameras was used to provide PIV images. The laser light sheet was brought in parallel to the floor of the channel. The light sheet can then be racked in the vertical through a series of steps through the use of a motorized mirror, and synchronized with the PIV cameras. The field of view extends from close to the upstream end of the first bend, towards the mid-point of the second bend.

3.2 Definition of time origin and instrument synchronisation


6 - Table of Experiments:

Experiment No.Experiment NameDownstream (x)PositionDensity Excess Input flowRotation RateRossby NumberInitial Water DepthOutflow RateRun TimeADV Dwell Time
$(kg \ m-3)$ $(l \ s-1)$ $(rad \ s-1)$ $Ro_W$ $(m)$ $(l \ s-1)$ $(Minutes)$ $(s)$
Test Experiments
0 fixstr1_1909a X1 20 12 0 $\infty$ 1 5.5 30 continuous
1 fixstr1_2109a X1 10.3 12 0 $\infty$ 1 5.5 21 60
2 fixstr1_2109b X1 10.3 10 0 $\infty$ 1 5.5 21 60
3 fixstr1_2209a X1 19.2 6 to 8 0 $\infty$ 1 5.5 21 60
4 fixstr1_2209b X1 19.2 10 0 $\infty$ 1 5.5 22 60
5 fixstr1_2609a X1 19 5.9 0 $\infty$ 1 5.5 27 30
6 fixstr1_2609b X1 19 5.9 0 $\infty$ 1 5.5 30 30
7 fixstr1_2709a X1 20 6 0 $\infty$ 1 5.5 20 30
8 fixstr1_2709b X1 18.4 15.3 to 6 0 $\infty$ 1 5.5 15 30
9 fixstr1_2809a X1 20 20 to 6 0 $\infty$ 1 5.5 15 30
10 fixapex2_2809b X4 20 20 to 6 0 $\infty$ 1 5.5 10 30
11 fixapex2_2809c X4 18.4 20 to 6 0 $\infty$ 1 5.5 5 30
12 fixapex2_2809d X4 18.4 20 to 6 0 $\infty$ 1 5.5 5 30
13 rotstr1_2909a X1 20.3 20 to 5.64 0.083 1 5.5 30 30
14 rotstr1_2909b X1 20.5 20 to 6 0.083 1 5.5 15 30
15 rotstr1_3009a X1 20.4 20 to 5.83 0.083 1 5.5 30 30
16 rotstr1_3009b X1 20.4 20 to 5.5 0.083 1 5.5 35 30
17 rotstr1_3009c X1 20.4 20 to 5.5 0.083 1 5.5 10 Visualization experiment
Main Experiments
18 fixstr1_0410a X1 18.8 13.1 to 5.8 0 $\infty$ 1 5.5 36 60
19 fixapex2_0410b X4 18.8 13.1 to 5.8 0 $\infty$ 1 5.5 30 60
20 rotstr1_0510a X1 20 13.1 to 5.8 +0.083 1 5.5 36 60
21 rotapex2_0510b X4 18.8 13.1 to 5.8 +0.083 1 5.5 30 60
22 rotstr1_0610a X1 20 10.97 to 5.75 +0.167 1 5.5 36 60
23 rotapex2_0610b X4 19.8 10.8 to 5.9 +0.167 1 5.5 30 60
24 rotstr1_1010a X1 19.7 10.86 to 5.92 +0.041 1 5.5 36 60
25 rotapex2_1010b X4 20.4 10.86 to 5.78 +0.041 1 5.5 30 60
26 rotstr1_1010c X1 20.2 10.86 to 5.75 +0.021 1 5.5 36 60
27 rotapex2_1010d X4 19.4 10.89 to 5.78 +0.021 1 5.5 30 60
28 rotstr1_1210a X1 to 1 5.5 36 60
29 rotapex2_1210b X4 to 1 5.5 30 60
30 rotstr1_1310a X1 to 1 5.5 36 60
31 rotapex2_1310b X4 to 1 5.5 30 60
32 rotstr1_1310c X1 to 1 5.5 36 60
33 rotapex2_1310d X4 to 1 5.5 30 60


7 - Diary:

Monday, September 19th 2016

Experiment name: fixstr1_1909a. Filenames: fixstr1_1909a1, fixstr1_1909a2. Location: Position X1 (75% of the way down the straight section, 56 cm upstream from the end of straight section). Input rate: 12 l/s, density excess: 20 kg/m3. Water was very cloudy to the extent that we were not able to use the laser. No siphon rig was used. Running basal ADV with ADV #1 located 7.2 cm from the base. ADV just measured at one point (no traverse measurement).

Experiment started at 2:45pm and stopped at 3:15 pm. The flow automatically stopped part way through as a valve was not open for recirculating water. Part way through red dye was added to visualise the current. Dye visualisation suggested pretty thin flow on the inner bank and significant super elevation on the outer bank. Some perturbation was observed on the surface of the flow at the outer bank, but otherwise the flow surface appears quite smooth, and mixing appeared to be very low. Mean maximum flow velocity from raw output was around 20 - 25 cm/s. ADV and UVP measured twice, the first is referred to as fixstr1_1909a1 and this had a velocity range of 0.3 m/s on the ADV setting, and 0.25 m/s for the UVP. Instantaneous flow velocities were faster than anticipated, as a result of the steep slope (3.5 degrees) with flow wrapping on both instruments so the velocity range was increased on the ADV and the UVP. A second run period fixstr1_1909a2 had a velocity range of 0.5 m/s on the ADV setting. The UVP setting was 680 mm/s (0.68 m/s).

Tuesday, September 20th 2016

Paint was applied to the tank floor to address a leak in the flume, and the tank floor left to dry and seal. The siphon rig was completed and tested. The laser system was aligned. The position of the basal ADV was refined, Orgasol was required as additional seeding in order for bottom tracking to work effectively. Interesting, the stem (fixed) ADV picks this bottom point up better than the flexible ADV in the absence of seeding. Tested synchronization of ADV with traverse; a problem was identified with the nature of the required input signal. Investigation in progress as to how to address this.

Wednesday, September 21st 2016

Experiment name: fixstr_2109a, b. File names: fixtr1_2109a, fixstr1_2109b. Location: Position X1 (75% down straight section, 56 cm upstream from the end of the straight section). Density excess 10.3 kg/m3.

fixstr1_2109a: The input flow rate was 12 L/s and the total experiment time was 21 minutes. The UVP was turned on as the gravity current entered the inlet box. The ADV was turned on 2 minutes after the UVP saw the flow. Two sets of siphon samples were collected three minutes after the current hit the end of the tank. Sampling was done every 10 minutes: Siphon Set #1 (4:30 to 5:30) and Siphon Set #2 (14:30 to 15:30). The flow was too thick and the velocity was about 18 cm/s. It was speculated that the flow rate should be decreased. Siphon samples showed a density decrease in depth, going from the bottom to the top, and also a decrease over time.

fixstr1_2109b: Since the flow was fast and thick, the flow rate was lowered from 10 L/s (t = 0 to 6 ) to 8 L/s (t = 8 to 13) to 6 L/s (t = 15 to 30). 3 sets of siphon samples were collected for the 3 flow rates at 3 different times (4:55 to 5:55, 12:02 to 13:02, and 19:26 to 20:26). The experiment lasted for 21 minutes.

Thursday, September 22nd 2016

Experiment name: fixstr_2209a, b. File names: fixstr1_2909a, fixstr1_2909b. Location: Position X1 (75% down straight section, 56 cm upstream from the end of the straight section). Density excess 19.2 kg/m3.

Before the experiments samples were collected from the storage tanks which store 10 kg/m3 and 20 kg/m3 salt water in order to check the density. Also, samples from the base and the top of the rotating Coriolis tank were collected. No laser was used since the quality of the water was not good enough.

fixstr1_2909a: Flow rate started with 6 L/s for about 16 minutes. Siphon sample Set #1 was collected around 5 minutes after the start of the experiment for 1 minute. After 7 minutes, a sample was collected from the tank on the third floor where salt water enters from the storage tank. Then the flow rate was increased to 8 L/s. At around 16 minutes constant flow rate was reached and a second set of siphon samples were collected at about 19 minutes for 1 minute deviation. The flow rate then was increased to 10 L/s, but since it took a long time to become constant the UVP stopped. So a second experiment was conducted with a flow rate 10 L/s. Also, the Nikon Camera took 2 videos, 10 minutes each one at the start and one almost right at the end.

fixstr1_2209b: Flow rate was 10 L/s and the total experiment timing was 16 min. The flow rate became constant after about 7 minute, so potentially it was not constant before. One siphon sample was collected at 4:27 for one minute deviation. After measuring the density, it was concluded that the density of the inlet box is much lower than 20 kg/m3 meaning a lot of mixing is occurring there. Therefore, there are two issues that need to be considered for future experiments.

  1. How to control the density of the inlet box and reduce mixing.

  1. Start every experiment with a constant flow rate, which takes at least 5 min to occur.

Friday, September 23rd 2016

The tank was drained and cleaned in order to locate, drill and install the 2 Mhz UVPs at the apex of the first curve. The probes were located at 10 points evenly spaced 45 mm apart, starting 45 mm from the bottom. The results from the previous experiments were revised to plan new experiments and updated experimental plan and protocol were provided. A calibration for the conductivity probe on the traverse was plotted. Concentration profiles and plots were updated for the latest experiments. Post-processing of the UVP and ADP data for the conducted experiments were carried out. All of our data on the network was updated including the UVP and ADV.

Monday, September 26th 2016

Experiment name: fixstr1_2609a,b. Filenames: fixstr1_2609a, fixstr1_2609b. Location: Position X1 (75% down straight section, 56 cm upstream from the end of the straight section). Input rate 5.9 l/s. Density excess 19 kg/m3. Laser was not used in the experiments today. Running basal ADV with ADV #1 located 7.2 cm, ADV #2 located at 10 cm and ADV #3 located at 12 cm above channel bed.

fixstr1_2609a: The main goal was to investigate if the concentration of the gravity current, inlet box and tank D were equal (19 kg/m3). The traverse sent one TTL signal to the ADV and didn’t send a return signal so only one measurement at one point was made by the ADVs (all ADVs collected data simultaneously). On the middle ADV (#2) wrapping occurred which indicates that velocity needs to be increased to 0.6 m/s. UVP files showed a fair amount of noise in the data which requires further investigation. We waited about 7:20 minutes for the flow rate to reach a constant value. One person took 6 samples (every 3 minutes, starting at t = 0) from the mixing tank on the third floor which contained the salt water mixture coming from tank D. A second person took 6 samples (every 3 minutes, starting at t = 0) from the inlet box at the upstream end of the channel. A third person took 3 sets of 12 siphon samples every 5 minutes (starting at t = 0, duration = 1 min).

GoPros were installed at three positions as follows: Position 1 - On top of inlet box looking downstream. Position 2 - Inside inlet box focused on inflow pipe. Position 3 - straight section of channel, looking cross-channel. Density measurements of the samples showed that the density in the mixing tank on the third floor and tank D were equal (19 km/m3). However, there was a major decrease of density in the inlet box (max density = 14.8 kg/m3). Density of inlet box needs further investigation.

fixstr1_2609b: Goal was to determine flushing characteristics of the inlet box, and how long it takes to reach steady-state density conditions. Running basal ADV at 5 positions in the bottom cross section only at traverse positions y = 0.016, 0.116, 0.216, 0.316 and 0.416 m (with 5 repetitions). GoPro cameras were mounted at three different positions (Position 1 - Outside rotating tank focused on channel outflow looking upstream. Position 2 - above tank, looking down, focused on bends. Position 3 - straight section of channel, looking cross-channel). Dye was released to visualize the flow. One set of 12 siphon samples (duration = 1 min) was taken 5 minutes after flow rate stabilized. It took 2.5 min for the flow rate to stabilize. The ADV was initialized 5 minutes after flow stabilized. The ADV needed more seeding to improve the signal-to=noise ratio. The GoPro in the inlet box showed a lot of mixing between the current and the ambient which explains the decrease in density in the gravity current in the inlet box. The suggestion was to reduce the inlet box volume to reduce the time to reach steady-state density (by reduce the flushing time).

Tuesday, September 27th 2016

Density of siphon samples from Monday, Sept 26 were measured. Two large PVC tubes (dia ~ 40 cm) were placed in the inlet box and weighted down to reduce the volume (new vol ~ 150 L).

Experiment name: fixstr1_2709a,b. Filenames: fixstr1_2709a, fixstr1_2709b. Location: Position X1 (75% down straight section, 56 cm upstream from the end of the straight section). Input rate 6 l/s. Density excess 20 kg/m3. Laser was not used in the experiments today. Running basal ADV with ADV #1 located 7.2 cm, ADV #2 located at 10 cm and ADV #3 located at 12 cm above channel bed.

Fixstr1_2709a: The goal was to examine the flushing time of the inlet box to see how it behaves once the PVC tubes were installed. One conductivity probe (C1T1) was mounted at the centreline just before the flow straightener (metallic honeycomb). Started with 6 l/s and waited for 5 min to see how inlet box, UVP and pump behaved. The UVP and siphon pump were turned on to fix the suspected pulse that the pump imparted to the UVP. This did not resolve the Signal to Noise Ratio (SNR) issue. The inlet box took a long time to reach a steady conductivity value. It was suggested to start with a higher flow rate and then reduce the flow rate to 6 l/s. The ADV and UVP data were still noisy. No siphon sampling was done.

fixstr1_2709b: It was suggested to install an array of three conductivity probes in a cross-section just before the honeycomb in the inlet box. Only one additional probe was available. As such, two conductivity probes were mounted in the cross section just before the flow straightener (metallic honeycomb): one at the centerline (C2 green) and one 5 cm?? from the right wall (C1T1 black), looking downstream. The traverse was able to move vertically in every cross-section position and the previously outstanding issues have been resolved. Water was initially pumped with a flow rate of 15.28 l/s (55 m3/h). This was the highest achievable flow rate. The water was pumped at this rate for 2 minutes and then reduced to 6 l/s. This seems to achieve the desired effect of reaching a steady conductivity value in a short amount of time in the inlet box. The UVP was set over 10 minutes, but still showed noisy data. No siphon sampling was conducted.

Wednesday, September 28th 2016

Experiment name: fixstr1_2809a and fixapex2_2809b, c, d. Filenames: fixstr1_2809a, fixapex2_2809b, fixapex2_2809c, fixapex_2809d. Location for fixstr1_2809a: Position X1 (75% down straight section, 56 cm upstream from the end of the straight section). Location for fixapex2_2809b, c, d: Position X4 (at the centre of the second apex). Input rate 20 l/s (initial ~2 minutes of each run) then reduced to 6 l/s. Density excess 18.4 kg/m3. Laser was not used in the experiments today. Running basal ADV with ADV #1 located 7.2 cm, ADV #2 located at 10 cm and ADV #3 located at 12 cm above channel bed.

fixstr1_2809a: The goal is to increase the seeding and test the ADVs and UVP for less noise. A new stem ADV was mounted to reduce the noise. The UVP didn’t change in terms of noise issue, so the problem does not lie with the seeding density. Another suggestion was to move the UVP box back to the back bench to reduce interference from the electronics on the traverse. Minimal improvement was seen on the ADV profile. The seeding seems to be getting stuck in the inlet box behind the flow straightening baffles - this was visually observed by a buildup of foam in the inlet box.

fixapex_2809b: The siphons were moved up by 20 cm and the UVP 17 cm. The traverse was also moved to the second apex. The goal for this experiment was to take siphon and UVP data higher than 23.6 cm above the channel bed (previously not possible due to the siphon and UVP configuration), because we want to be able to draw velocity and concentration profiles for 40 cm flow thickness. It was also intended to verify if moving the physical position of the UVP box would impact the data quality. Traverse moved to apex 2 position and siphon samples were taken every 1 minute after flow rate stabilized, for 1 minute sampling time (5-6 min, 7-8 min, 9-10 min). When the flow was released a large cloud of seeding flowed through the channel as a gravity current (presumably seeding caught in inlet box from previous run) and the data on the ADV improved substantially. It was thus concluded that seeding is a viable solution to the noisy data problem (provided a reliable mechanism for seeding the flow, without it getting caught in the inlet box, can be devised). The other suggestion is to increase their distance further apart to avoid side lobe interference. UVP data remained noisy. It is speculated from pressing and GoPro videos that surface waves are produced where the channel sides plunge under the free surface, and that this is the source of the noise in the UVP data.

fixapex_2809c: The goal for this experiment was to turn off the traverse completely then turn on ADV and UVP to check if noise can be reduced. Almost all electronics were turned off on the traverse. Both the UVP and ADV still show a lot of noise. This implies there is either seeding or side lobe effect for the ADV and speculation on surface wave for the UVP.

fixapex_2809d: The new stem ADV was damaged, so the old cable ADV was used instead. Only one experiment was run with just one ADV to see if it will show noise. It didn’t show noise. Since running multiple ADVs showed noise, this means that they are talking to each other. SNR was low, but according to Nortek, in the arms it is still possible to take good data even with low SNR.

Thursday, September 29th 2016

The tank was drained, cleaned, washed, and refilled. The tank was spun with 0.083 rad/s rotation rate (0.8 rpm).

Experiment name: rotstr_2909a, b. Filenames: rotstr1_2909a, rotstr1_2909b. Location: Position X1 (75% down straight section, 56 cm upstream from the end of the straight section). Input rate started from 20 L/s for the first two minutes and then was decreased to 5.64 L/s.

rotstr1_2909a: Density excess was 20.3 kg/m3. Two sets of siphon samples were collected every 5 to 6 minutes and 15 to 16 minutes after the start of experiment. Siphons and UVPs were at the top part of the flow. Gopros took movies from the ping pong balls moving on top of the rotating tank.

rotstr1_2909b: Density contrast was 20.5 kg/m3. Siphons and UVPs were moved down to their original location. Three set of siphon samples were collected every 5 minutes after the start of the experiment. The experiment lasted for 15 minutes.

Friday, September 30th 2016

Experiment name: rotstr1_3009a, b, c. Filenames: rotstr1_3009a, rotstr1_3009b, rotstr1_3009c.

rotstr1_3009a: The total experiment time was 30 minutes. The goal was to use two UVPs and switch between them within the experiment. ADV measurements were collected at the straight section (Position X1 -75% down straight section, 56 cm upstream from the end of the straight section). The stream wise UVP collected data for the first 10 minutes of the experiment. In the second 10 minutes, the UVP was switched to the cross stream and collected data in the cross stream section for the third 10 minutes. It should be noted that the cycle for the cross stream UVP was set to 1500 mistakenly, instead of 666. Bottom check data was collected for both ADVs. Siphon samples were collected every 5 minutes after the flow rate became constant. The 3 sets were sampled at 5 to 6 minutes (Set #1), 15 to 16 minutes (Set #2) and 25 to 26 minutes (Set #3). The laser was used in this experiment.

rotstr1_3009b: The goal was to see how much better the UVP data would be later in the experiment. Input flow rate started from 20 L/s for the first 2 minutes and then was decreased to 5.5 L/s. Density excess and temperature were 20.4 kg/m3 and 23.7 degrees C respectively. The experiment lasted for 35 minutes as opposed to 30 minutes because it was concluded from previous experiments that would show better results later in the experiment. Therefore 5 minutes after the flow rate became constant, the stream wise UVP collected data for 10 minutes. Then, the two UVPs were switched in the next 10 minutes and, finally, the cross stream UVP collected data from 25 to 35 min. A complete sequence of data was collected by the ADV, but it did not restart for some time. Then, it started collecting a second sequence later which was not complete because the experiment ended. The laser did not work in this experiment.

rotstr1_3009c: Input flow rate started from 20 L/s for the first 2 minutes and then was decreased to 5.5 L/s. The goal for this experiment was to add dye and visualize the behaviour of the current at different locations. GoPros? were placed at different locations to take movies of the current and to test which location is the best fit for taking movies.

Monday, October 3rd 2016

The tank was drained, washed and cleaned. New taller probes were built and installed for the stream wise UVP and the siphons. A new sequence for the traverse was considered and calibration for both conductivity probes was completed. All the data and the diary on the Wiki were updated and a new revised experimental plan was proposed for the next three weeks.

Tuesday, October 4th 2016

Test experiments have been completed, and the instrumentation and experimental sequence and conditions optimised. Today, the main data gathering for the experimental programme started. The tank was filled with water. Calibration curves for both probes were prepared (C1 is the conductivity probe in the inlet box and C0 is the one on the traverse). The traverse was reprogrammed for the new sequence. This has been programmed such that the total timing is about 15 minutes, instead of the proposed 19.5 minutes. This was achieved by decreasing the stationary time and the move time, but the ADV dwell time stayed the same (60 s). Two experiments were conducted for Rossby number infinity.

Experiment name: fixstr1_0410a and fixapex2_0410b. File names: fixstr1_0410a, fixapex2_0410b. Density excess and temperature for both experiments were 18.8 kg/m3 and 22.4°C respectively. Input flow rate started with 13.1 L/s and was decreased to 5.8 L/s. The start of the experiments (t=0 on the stopwatch) was when the flow rate and conductivity of the inlet box became constant.

fixstr1_0410a: The total experiment time was 36 minutes. PIV measurements were conducted for the first 10 minutes of the experiment. ADV data was collected at the straight section (Position X1 - 75% down straight section, 56 cm upstream from the end of the straight section) from 15 to 25 min after the flow rate became constant. 7 sets of siphon samples were collected at 5, 10, 15, 20, 25, 30 and 35 minutes after the start of the experiment each with a duration of 1 minute. The cross-stream UVP collected data from 5 to 15 minutes, and the down-stream UVP collected data from 25 to 35 minutes.

fixapex2_0410b: The total experiment time was 30 minutes. ADV measurements were collected from 15 to 35 minutes after the start of the experiment at the second apex (Position X4).

Wednesday, October 5th 2016

The tank was rotated with a rate of 0.083 rad/s (0.8 rpm) in the counter-clockwise direction, which corresponds to a Rossby number of +1. Two experiments were conducted.

Experiment name: rotstr1_0510a and rotapex2_0510b. File names: rotstr1_0510a, rotapex2_0510b. Inflow rate for both runs started with 13.1 L/s and then was decreased to 5.84 L/s until it reached a constant level.

rotstr1_0510a: The total experiment time was 36 mins. PIV measurements were conducted for the first 10 minutes of the experiment. However, the water was not as clear as expected and there were a lot of bubbles on the tank’s wall, so the quality of the PIV data may be low. Density excess and temperature were 20 kg/m3 and 22.6°C, respectively. ADV data were collected at the straight section (Position X1 - 75% down straight section, 56 cm upstream from the end of the straight section) from 15 to 25 min after the flow rate became constant. 7 sets of siphon samples were collected at 5, 10, 15, 20, 25, 30, and 35 minutes after the start of the experiment, with a duration of 1 minute. The cross stream UVP collected data from 5 to 15 minutes, and the down-stream UVP collected data from 25 to 35 minutes.

rotapex2_0510b: The total experiment was 30 minutes. Temperature and density excess were 23°C and 18.8 kg/m3, respectively. ADV measurements were collected from 15 to 35 minutes after the start of the experiment at the second apex (Position X4).

Thursday, October 6th 2016

The tank was rotated with a rate of 0.167 rad/s (1.6 rpm) in the counter-clockwise direction, which corresponds to a Rossby number of +0.5. Two experiments were conducted.

Experiment name: rotstr1_0610a and rotapex2_0610b. File names: rotstr1_0610a, rotapex2_0610b.

rotstr1_0610a: The total experiment time was 36 mins. Inflow rate started with 10.97 L/s and then was decreased to 5.75 L/s until it reached a constant level. The water was dirty so no PIV measurements were conducted. Density excess and temperature were 20 kg/m3 and 22.7°C, respectively. ADV data were collected at the straight section (Position X1 - 75% down straight section, 56 cm upstream from the end of the straight section) from 15 to 25 min after the flow rate became constant. 7 sets of siphon samples were collected at 5, 10, 15, 20, 25, 30, and 35 minutes after the start of the experiment, with a duration of 1 minute. The cross stream UVP collected data from 5 to 15 minutes, and the down-stream UVP collected data from 25 to 35 minutes.

rotapex2_0610b: The total experiment was 30 minutes. Temperature and density excess were 22.4°C and 19.8 kg/m3, respectively. Inflow rate started with 10.8 L/s and then was decreased to 5.9 L/s until it reached a constant level. ADV measurements were collected from 15 to 35 minutes after the start of the experiment at the second apex (Position X4).

Friday, October 7th 2016

The tank was drained, washed and cleaned. Then it was filled and rotated with 0.041 rad/s (0.4 rpm) corresponding to Rossby number of +2. Wiki notes were prepared and the website was updated. All files on the LEGI network were also updated.

Monday, October 10th 2016

Experiment name: rotstr1_1010a, rotapex2_1010b, rotstr1_1010c, rotapex2_1010d. File names: rotstr1_1010a, rotapex2_1010b, rotstr1_1010c, rotapex2_1010d.

Four experiments were conducted. For the first two experiments the tank was rotated with a rate of 0.041 rad/s (0.4 rpm) in the counter-clockwise direction, which corresponds to a Rossby number of +2.

rotstr1_1010a: The total experiment time was 36 mins. Inflow rate started with 10.86 L/s and then was decreased to 5.92 L/s until it reached a constant level. The water was dirty so no PIV measurements were conducted. Density excess and temperature were 19.7 kg/m3 and 23.3°C, respectively. ADV data were collected at the straight section (Position X1 - 75% down straight section, 56 cm upstream from the end of the straight section) from 15 to 25 min after the flow rate became constant. 7 sets of siphon samples were collected at 5, 10, 15, 20, 25, 30, and 35 minutes after the start of the experiment, with duration of 1 minute. The cross stream UVP collected data from 5 to 15 minutes, and the down-stream UVP collected data from 25 to 35 minutes.

rotapex2_1010b: The total experiment was 30 minutes. Temperature and density excess were 23.8°C and 20.4 kg/m3, respectively. Inflow rate started with 10.86 L/s and then was decreased to 5.78 L/s until it reached a constant level. ADV measurements were collected from 15 to 35 minutes after the start of the experiment at the second apex (Position X4).

For the second two experiments the tank was rotated with a rate of 0.021 rad/s (0.2 rpm) in the counter-clockwise direction, which corresponds to a Rossby number of +4.

rotstr1_1010c: The total experiment time was 36 mins. Inflow rate started with 10.86 L/s and then was decreased to 5.75 L/s until it reached a constant level. The water was dirty so no PIV measurements were conducted. Density excess and temperature were 20.2 kg/m3 and 24.3°C, respectively. ADV data were collected at the straight section (Position X1 - 75% down straight section, 56 cm upstream from the end of the straight section) from 15 to 25 min after the flow rate became constant. 7 sets of siphon samples were collected at 5, 10, 15, 20, 25, 30, and 35 minutes after the start of the experiment, with duration of 1 minute. The cross stream UVP collected data from 5 to 15 minutes, and the down-stream UVP collected data from 25 to 35 minutes.

rotapex2_1010d: The total experiment was 30 minutes. Temperature and density excess were 24.1°C and 19.4 kg/m3, respectively. Inflow rate started with 10.89 L/s and then was decreased to 5.78 L/s until it reached a constant level. ADV measurements were collected from 15 to 35 minutes after the start of the experiment at the second apex (Position X4).

Tuesday, October 11th 2016

The tank was drained, washed and cleaned very carefully. Density measurements of siphon samples from previous experiment were done. A revised experimental plan was prepared based on the conducted experiments so far and all the data on the network was updated to the latest experiment. Fresh water is being prepared (heated up) for new experiments tomorrow.

Wednesday, October 11th 2016

The tank was filled and simultaneously rotated with a rate of 0.8 rpm in the clockwise direction corresponding to Rossby number -1. Density measurements of siphon samples from previous experiments were made. Two experiments were conducted.

Experiment name: rotstr1_1210a, rotapex2_1210b

Attachments (5)

Download all attachments as: .zip