In utility power system protection we have a protection function called "breaker failure" protection. This protection is used to take appropriate action if/when a circuit breaker fails to open after a protective relay commanded it to.

One of the methods used to determine if the breaker is still closed is a simple fault detector that monitors currents from current transformers (CTs) on the breaker. In some cases we need really fast breaker failure to support system stability. We not only may need this breaker failure to operate fast, but we also need it to be secure against incorrect operation. If it takes it a long time to recognize that the primary current is gone then it is subject to incorrect operation.

In CTs, after the high primary fault current has gone, it is typical to have a "subsidence current" flowing in the CT secondary (L-R circuit) as shown below (taken from here):

enter image description here

After the primary current is gone, we need the current detector in the breaker failure relay to reset (to avoid accidentally timing out and tripping).

This subsidence current is a simple decaying L-R transient. There are digital relays now that can detect that the primary fault current is gone in under 1-1.5 cycles (16-25ms) now by recognizing that the secondary current is now only subsidence and not AC driven. But, there is an older analog relay, the Basler BE1-50BF, that boasts a current detector reset time of 1ms.

My question to the analog electronics experts is this: What analog circuit or approach could/would you take to detect that the CT secondary current being measured is no longer AC, but is just subsidence current?

This paper describes some of the challenges that one digital relay manufacturer faced. But, I'm curious about an analog design like I mention above. I have never seen a digital relay get close to that 1ms time. I'm not building anything, I'm just curious.

EDIT: The subsidence looks like below (snipped from above reference). It is a simple L-R transient.

enter image description here

  • \$\begingroup\$ Interesting question. I'm thinking maybe looking at the first derivative of the waveform along with the waveform itself and qualifying based on both of those things might be useful. If the waveform is near zero the derivative should be max, and if the derivative is near zero the waveform should be max. An L-R decay would have a different signature. But that's just off the top of my head, no idea if it would work without a lot of analysis. \$\endgroup\$ – John D Oct 2 '20 at 0:37
  • 1
    \$\begingroup\$ It would help if there were waveforms you could show. \$\endgroup\$ – Andy aka Oct 2 '20 at 8:56
  • \$\begingroup\$ The Basler is said to be a "solid state relay". That being so, perhaps it avoids the L-R time constant by using a direct current shunt measurement? \$\endgroup\$ – user_1818839 Oct 2 '20 at 10:33
  • \$\begingroup\$ If the load current changes abruptly on a cycle by cycle basis wouldn't make this harder to detect? \$\endgroup\$ – Andy aka Oct 2 '20 at 15:12
  • \$\begingroup\$ Yeah, a load change may contain a decaying DC component. But, the AC component will still be present as well - indicating we still have current flow. We need to detect a decaying DC component in the absence of AC. \$\endgroup\$ – relayman357 Oct 2 '20 at 18:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.