[Acoustics] Moving Bed Test - Loop Method Protocol

Wed Mar 15 12:11:22 CST 2006

Dave,

Overall your approach is reasonable but there are a number of potential
errors that can be made. Chad Wagner and I are in the process of publishing
an ASCE proceedings paper on a simplified approach for using the loop
method, an ASCE Journal of Hydraulic Engineering paper covering a more
complex approach, and a USGS Scientific Investigations Report covering all
of the above in more detail with practical instructions.

You have suggested using the Q/A value for the mean water velocity. Realize
that the default value for Area in WinRiver is the area computed parallel to
the average course. Because you have returned to the starting point your
area will be very small and Q/A may not be representative of the cross
section. Although not perfect, using the mean speed is a better alternative
(this also what would be used for a stationary moving bed test).

Details that are really important:

1)     Good compass calibration!!

2)     Good control on starting and stopping positions. Any error here leads
directly to errors in computing the moving bed velocity.

3)     Uniform boat speed. You are using the pings to sample the cross
section, if you spend more time in one section than in another that section
will have more weight in the equation DMG/Time. Typical problems occur when
too much time is spent near the river banks. Uniform speed may be more
important than average speed.

In the papers/reports in review and in press, Chad and I characterized the
random error associated with measuring the moving bed velocity using the
loop method as +/- 0.6 cm/s (from 28 loop measurements at 17 sites with no
moving bed). This random error should be considered when assessing whether
or not a sight has a moving bed.

If you don't have a good solid starting and stopping point, simply deploy a
buoy to mark the spot. This is probably the preferred method as you can put
the buoy far enough from shore that you won't spend a lot of time navigating
the boat to the buoy while worried about running into the bank (see 3
above).

Based on the data presented in the reference papers we have suggested that
maximum boat speed should be the lesser of a boat speed that requires no
less than 3 minutes to complete the loop or a boat speed that is less than
1.5 times the mean water speed. This is based on the data that we had
available. The longer the loop time the smaller the influence of the random
error on the measured moving-bed velocity, provide other criteria are met.

If you use GPS to establish the DMG as suggested you must then also consider
the accuracy of the GPS unit and potential for multipath errors in
determining the accuracy of the moving-bed velocity computed from the loop.
Since you suggest that you have GPS I assume you are only using the loop to
determine if there is a moving bed. A stationary test using difference
between GPS and BT could also be used.

Hope this helps.

I will let this list know when the reference papers/report are available. I
would hope that I can release an approved draft report within a month.

David S. Mueller

U.S. Geological Survey

Office of Surface Water

9818 Bluegrass Parkway

Louisville, KY 40299

dmueller at usgs.gov

Voice: 502-493-1935

Fax: 502-493-1909

Mobile:502-558-5907

  _____  

From: acoustics-bounces at simon.er.usgs.gov
[mailto:acoustics-bounces at simon.er.usgs.gov] On Behalf Of Huston, Dave
Sent: Friday, March 03, 2006 1:21 PM
To: acoustics at simon.er.usgs.gov
Subject: [Acoustics] Moving Bed Test - Loop Method Protocol

Hi All,

I'd like to know the currently accepted protocol for the "loop method" of
moving bed tests and when one can make a reasonable judgment from a
particular test that the bed was "moving", "fluidized", or whatever people
like to call it [i.e. requiring the use of GGA as the reference instead of
BTM, when processing RDI Workhorse transect files].  All of this is assuming
that the test and flow measurements are being performed with a DGPS that's
working properly and that the ADCP was compass-calibrated correctly with no
ferrous metals in the vicinity of the ADCP affecting the compass.

My present understanding is this:

1) Select a good fixed start location, take your edge estimate and start
your loop (hitting F5 in WinRiver and entering in your edge estimate and the
appropriate bank).

2) Crab across to the opposite bank, take your edge estimate, then loop-back
to the starting bank (without hitting F5, which would have ended the
transect).

3) Return to the same start location and take a final edge estimate and hit
F5 to end the "loop"

Note A) Taking the edge estimates were simply to allow the data processor to
separate the loop into two separate transects which could then be used as
part of your flow measurements.

Note B) Preferably, the total loop time should have been around 600 seconds
(10 minutes).

4) Take the DMG (for the BTM ref.) and divide it by the total transect time
in seconds to get "moving bed" velocity  ---  DMG (ft) / transect time (s) =
MB Velocity (ft/s)

5) Go to the Discharge History Tabular (hit F12) and find the value for
Q/Area (ft/s) for the "loop", which is the overall loop water velocity

6) Take your MB Velocity (ft/s) and divide it by the Q/Area (ft/s) for the
"loop" and multiply by 100 to get a percentage of the "moving bed" velocity
to the overall loop water velocity.

MB Velocity (ft/s) / Q/Area (ft/s) * 100 = %

7) If that percentage is equal to 1% or greater, you could make a reasonable
judgment that the bed is "moving/fluidized" and that the data processor
should use the GGA as the reference.

What I'm not sure about are two things that could be related (and you might
find more).

First, we work at quite a few sites that are channelized with rip-rap on
both levees and not too many trees or objects that are easy to fixate upon.
So...we can't be entirely sure that our ending point is the same as our
start point.  Is there a simple way to correct for that within Winriver by
using the difference in the GPS coordinate data between the end ensemble and
start ensemble?

Second, when calculating the "moving bed" velocity...should I ACTUALLY be
using the difference in the distance made good between the GGA reference
data and the BTM reference data (DMG - DMB) and then divide that by the
total "loop"/transect time?  I'm assuming that if the bed is moving, I would
see a large difference in the distance made good (DMG) between the two
references (GGA and BTM).  This might also help me to correct out any
distance differences between my actual start point and end point.  Should I
be calculating the "moving bed" velocity this way?

I'd appreciate any correspondence or help that any of you out there may
offer.  Additionally, this may be good for the collective to discuss and
formalize the protocol in writing.

Thanks All,

Dave

***********************************

Dave Huston, P.E.

Dept. of Water Resources

Central District

3251 S Street, Room E-25

Sacramento, CA. 95816

(916) 227-7583

(916) 227-7600

***********************************

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://simon.er.usgs.gov/pipermail/acoustics/attachments/20060315/c446dc09/attachment-0001.html