Correlation drops

Up to Velocimeters - single point

Correlation drops

Posted by Iain MacDonald at July 01. 2005
Dear Nortek

I am having problems using multiple vectors in a vertical array. I have attached a figure that shows some systematic drops in correlation from a recent deployment. The drop we attribute to cross talk between the vectors. Two vectors and one sontek ADV were used in the experiment. The vectors sampling volumes were located at 160 and 600 mm above bed, the sontek sampling volume was located 1420 mm above the bed. All probes were downwards looking (I have had a similar problem when the probes are sideways looking). The data shown was collected at 16 Hz with 9600 samples per burst every 15 mins.  I know we have discussed this in the past but I would like to revisit it as we have more data.

Could you please comment on the data and suggest what we could try next (as we have tried the side-mounted configuration).

Thanks Iain

Correlation drops

Posted by Iain MacDonald at July 01. 2005
Attachments?

For displaying images, please see here:
http://www.nortek-as.com/cgi-bin....f=4;t=8

- Atle Lohrmann

Correlation drops

Posted by Iain MacDonald at July 04. 2005
Dear Atle


Here is the image.



Iain

Correlation drops

Posted by Iain MacDonald at July 04. 2005
The image made it fine - the only problem was a capital letter (UNIX server).  I also took the liberty to make the figure a little smaller.

The data gaps certainly look deterministic and I would think that the bottom echo from one probe is getting into the data from the other probe.  I suspect the problem is worse with a hard bottom than it is with a soft bottom.  Also, I would think the problem is greater when using a down-looking orientation compared to the results we saw from the side-looking probes.

Other than this I am not quite sure I have anything sensible to say.  This is a limitation of the technology and I am not quite sure if there are other researchers who have found a way around the problem?

Best regards,

Atle Lohrmann

Correlation drops

Posted by Iain MacDonald at July 05. 2005
Dear Atle

I would like to continue with this topic as more information/questions comes to light. I have included a further two images from the data set. These images are from the very start of the deployment before the vectors are submerged. For bursts 1,16, 31 (which I must add are really burst numbers 0, 1 and 2 as generated by the vector software, I just use a different burst numbering system) both probes are out of the water. For burst 46 the bottom probe is submerged while the middle probe is still out of the water. For burst 61 the middle probe goes underwater after about 150 s. Does this still fit with echo of the bed hypothesis? I cannot see how as only one vector is in the water. Do you agree? I look forward to your comments.

Thanks Iain


(The images are oow_corr_lr.jpg and oow_snr_lr.jpg which are in the nortek ftp site if they don't appear with this message.)

Correlation drops

Posted by Iain MacDonald at July 21. 2005
Dear Iain

Sorry for the wait - the summer holiday is really taking its toll  <img src=:)" />

I think you are quite right, this does not look like acoustic interference.  Instead it looks like electrical interference where the transmit pulse from one system gets into the receivers in the second system.

My first suggestion would be to look at the position of the probe cables.  If they are lying next to each other, the transmit pulse in Vector #1 can induce a current in the receive wires in system #2. I have seen this happen before in the NDV/Vectrino but I have never really looked for it in the Vector.

Does it make a difference if you simply mount the probe cables as far apart as possible while keeping the probes relatively close.  I would think you can do this test with both probes out of the water.

Best regards,

Atle Lohrmann

Correlation drops

Posted by Iain MacDonald at July 22. 2005
Dear Atle,

Thanks for your reply. We have investigated the electrical interference problem a little further. When the proble cables are coiled together the problem is at its most serve. The problem, however, is still evident (to a much less degree) when the cables are not coiled and separated from each other by at least 30 cm. We would now like to try to shield the probe cables. Do you have any ideas on the best way to do this?

Iain

Re: Correlation drops

Posted by TannerLWilson at October 18. 2019

This is the best forum for informative discussions and I like the way you create this. We get different ideas from different users. Visit the https://essaysoriginreview.com/grabmyessay-review/ to see more latest blogs on this forum. Hope you like to read from here.

Powered by Ploneboard
Log in


Forgot your password?
New user?