Additive white noise in power systems

Hi,

I have been trying to find the level of white noise that can be expected when recording current waveforms on a distribution feeder. Suppose I have a 33 kV feeder carrying 400 A peak, what is a realistic range of SNR? Is it around 40dB or more? Or does the white noise affect the signal amplitude in the order of milliamps or amps (say 3% of peak)? Any help from some power system protection engineer will be much appreciated.

Thanks

08-20-2014 10:30 PM

Top #2

Alan Maltz

08-20-2014 10:30 PM

Kapildev,

I suggest you read chapter 3 in "Power Systems Signal Processing for Smart Grids" by Paulo Fernando Ribeiro, Carlos Augusto Duque, Paulo Márcio Ribeiro, Augusto Santiago Cerqueira, for some interesting discussions of the SNR range.

Alan

08-21-2014 12:37 AM

Top #3

Kapildev Lout

08-21-2014 12:37 AM

Hi Alan,

Thank you for your reply. I recall reading the Mathematical modelling of noise section in this book a few days back and I believe the author said the SNR for white noise is greater than 27dB and 40dB is a decent assumption. However, I had a chat with one of our lecturers who used to design protection relays when Alstom used to be called GEC and he believes it should be higher than 40dB. That's why I was looking for a second opinion on this matter.

Kapil

08-21-2014 02:57 AM

Top #4

Alan Maltz

08-21-2014 02:57 AM

Kapildev,

I would agree with your colleague, 40dB implies that the noise is 1/100 of the signal so on your 400A feeder it would be 4A riding on 400A, or roughly a 250kVA (3 ph) load switching on and off. You should probably be looking at somewhere between 60dB and 80dB depending upon the magnitude of the noise you're trying to capture.

I wish I could give you more definitive data, the best I can do is generate heuristic arguments to make or break a case; there's just very little out there on this topic.

I did find this paper that is more concerned with the delay in telemetering utility data via TCP and UDP, but you may find that some of his charts have useful data regarding the magnitude and decay vs. time of some measured data, maybe there's a nugget of info in there that you can use:

https://idea.library.drexel.edu/bitstream/handle/1860/24/carullo_thesis.pdf;jsessionid=11B44A4E378D1C1C698B255F408E425D?sequence=9

Interesting topic. Best of luck on your thesis!

Alan