SEL-351 - ALARM output 1

[mvsubstation]

If the SEL-351 fails, the ALARM output contact goes to 1. Is there any condition of failure where the ALARM relay word bit would not go to 1 as well?

I.E. instead of wiring the ALARM output contacts of the SEL-351 into the trip coil of the circuit breaker (relay failure would trip the breaker), why not just use the ALARM relay word bit in the TR trip equation instead of wiring the ALARM contact?

 

[davidbeach]

Loss of voltage to the power supply will cause the ALARM contact to close but will not cause the relay word bit ALARM to assert.

 

[DiscoP]

We would never think of using the ALARM contact to trip the breaker - we get it to raise a SCADA alarm only.

Is this a common practice ?

 

[DTR2011]

Alarm also pulses when logging in (Level 2 I believe), saving settings, etc. I would NOT recommend connecting ALARM to the trip circuit.

 

[mvsubstation]

The application is for an interconnect breaker for distributed generation. The utility is requiring the relay failure alarm contact to trip the interconnect breaker.

Given Loss of voltage to the power supply will not cause the ALARM bit to assert, I would think, therefore, that one could add a low DC voltage word bit to the trip logic equation in addition to the ALARM word bit, and thus, mimic the ALARM output contact being wired to the trip coil of the breaker?

 

[davidbeach]

No, if there is no power to the relay it can't close a trip contact.

 

[jghrist]

What can you do with relay word bits if the relay has no power? When the relay enters Access Level 2 or B, the ALARM Relay Word bit momentarily asserts to 1 and the ALARM output contact coil is deenergized momentarily. Do you really want your intertie breaker to trip when someone is changing settings or just finding out the relay firmware version (VER command requires Level 2)?

 

[stevenal]

The alarm word bit is asserted for any condition that causes the relay to fail it's self testing. Any condition that prevents the relay from performing the self test will prevent the word bit from asserting. In addition to the condition davidbeach mentioned, failure of the power supply and complete failure of the microprocessor are included in the failure modes that not cause the word bit to assert.

To do what you ask, you will need to wire the normally closed alarm output in parallel with the output used for the TRIP word bit. The no DC voltage will cause the alarm contact to go to its normally closed (deenergized coil) state. Of course all the other conditions the postors have mentioned above will also cause the contact to close.

 

[zer0sequence]

Direct wiring a trip to the alarm contact on an SEL (or any relay for that matter) is a bad idea. As mentioned before, the alarm contact will assert for reasons other than simply relay failure. If it were me, I would wire the relay failure alarm b-form contact into another device (SCADA or any sort of logic processor) and have the second device trip the breaker after a decent time delay (5 sec or more). That way, a pulse of the contact (such as when entering 2AC or pulsing a bit) will not cause a trip.

Make sure and put a big sign on the wall that says turning the relay off will trip the breaker!!

 

[cranky108]

Here one better, turn off the DC to the relay to maintain the breaker. And can't figgure out why the breaker won't stay closed.

 

[dpc]

As mentioned before, the alarm contact will assert for reasons other than simply relay failure.

And sometimes for no apparent reason at all.

 

[stevenal]

The logic processor extends the problem. If the utility requires tripping upon failure, now the failure modes of the logic processor controlling the tripping must be considered in addition to the failure modes of the relay.

Maybe it's my utility side bias, but I'm uncomfortable with stating the utility requirement is a bad idea when we have no idea why it is a requirement. Suggest we stick with stating the disadvantages and leave it there.

 

[bacon4life]

In a new substation where SCADA was behind schedule, my utility put in a trip via time delayed ALARM bit as a temporary "fail safe" measure. Even though the trip was time delayed and there was plenty of notices/posting/education, the feeder still managed to get opened incorrectly. Note that I said incorrectly and not accidentally, because accident implies unexpected. I fully expect inadvertent trips to happen when doing unexpected things like this to the trip equation.

 

[cranky108]

If the concern is a relay failure, then you should ask why are we only installing one relay?

So maybe the concern is a lack of redundency, in which case I would look at other issues along the same line. There is an IEEE document on this whole subject (also a NERC document just to be redundit).