Monitored vs. Unmonitored Microprocessor Relays 1

[dbgv]

What is the difference between an monitored and an unmonitored microprocessor relays?

Has anyone with an EX2100 GE exciter had discussions with NERC about PRC-005 and whether they consider it to be "an elaborate form of a microprocessor relay" and as such should be tested like a microprocessor relay?

 

[DTR2011]

Most microprocessor relays have a self-test alarm contact. This is generally connected to SCADA/DCS. If connected, then it is hardwired and monitored. This will pick up some, but not all potential failures. It can not determine if an input or output is failed or if the device is metering properly.

Some SEL relays also allow this to be communicated via a comm channel for the same. I would call this a soft point alarm.

 

[ptcabb]

It can not determine if an input or output is failed or if the device is metering properly.

Speaking for transmission level IEDs at least, as lower end IEDs mean reduced price levels, meaning reduced capability and more basic technology:

To a certain extent it can. Binary input and output modules do have self supervision or watchdog supervision and detecting if a particular module failed is achievable. Detecting if the actual output relay contact failed is difficult though, you can only detect that once you try to close it. But there's far more electronics on the module then just the output contact that needs to be supervised and typically this has a higher failure rate, as the output relay is a completely enclosed unit.

Metering is quite easy. One solution is that the analog input has 2 separate measurement elements (low rating / high rating) that are used to measure small currents as accurately as high currents. Additionally the values can then be compared to each other and if they fall outside a certain bandwidth, the measurement / analog module is reported as faulty to the IRF and software system.

It's important that the watchdog driving the IRF lives outside of the software process. If the software halts due to a bug, which can be dealt with by rebooting the IED for example, the watchdog has to report this. It can't if it's part of the software itself.

Doing IRF over software / communication is perhaps a nice addition, and as far as I'm aware any modern IED with self supervision can do so, but as communication is one of the weak spots to begin with, it seems rather pointless to me. Hardwire the IRF to another device, be it an IED or RTU or annunicator, and have that one report it upstream.

Never heard about the terms unmonitored and monitored though. Every IED I've seen has an IRF contact. But I'm not so familiar in the low end segment.

 

[dbgv]

Thank you both for your responses. I think the terms monitored and unmonitored came from NERC. I would have asked my NERC compliance people this question, but seeing as how they were the ones who asked me what it meant, I can see they will be no help.

"Most microprocessor relays have a self-test alarm contact. This is generally connected to SCADA/DCS. If connected, then it is hardwired and monitored."

 

[stevenal]

Is your question whether this is a component of a protection system? See the NERC definition of Protection System to answer this.

If you have decided your exciter is a protection system component, the next question is whether or not it is monitored per the standard.
I would be very careful applying this exception. As I understand it, the standard drafting team was considering a system that may not yet even be possible. What is the uptime of your SCADA? Can you always meet the 24 hr requirement of Table 2 without exception? Is the monitoring system itself monitored per table 2? Every portion of it? If the auditors approve your 12 year cycle, it might very well be the first one. Might want to stick to six years for a while until others can report their experience.

 

[dbgv]

My first question is: "Is the EX2100 exciter a component of the protection system?" - The answer that I'm leaning too would be yes, if it can initiate a trip and the output contacts are tied to a lockout relay.

My second question is: "If it is a component of the protection system, then does it fall under the testing requirement for microprocessor relays?"

My third question is then, "If it is considered to be a microprocessor relay, then does it fall under the requirements for unmonitored relays or does it fall under the requirements for monitored relays?"

I can view the exciter parameters, enabled trips, faults codes, etc. from the DCS. I'm thinking that this would then make it monitored, but I haven't been able to find a clear NERC definition of "monitored relays".

 

[stevenal]

Q1: Agreed.

Q2: Yes. Be sure you don't neglect the other components.

Q3: The unmonitored interval is the safe choice. Be sure to target an even shorter interval so that an emergency does not ever put you past the standard maximum intervals.

Table 2 lists the requirements for monitoring, so a separate definition would only add confusion. The Table 2 requirements are very hard to meet. Yes you can monitor the relay alarms when the DCS is up, but what monitors the DCS as the alarm path per row 2? And is there a common failure mode that would affect both the DCS and the DCS monitoring system so that 24 hours will never pass? And can you prove it and document it?