Reliability starts with eliminating design, construction & commissioning errors

Most of the power industry would generally agree with the notion that wire-based technologies are inherently reliable. After all, we have been engineering protection and control schemes for "100 years" and the lights are generally on!

But is this really true and is there room for improvement?

Wire-based engineering starts with opening a blank CAD screen, then drawing new boxes, new terminal numbers, new wires, new wire numbers, then giving that to the factory to connect them up, then do point-to-point wire testing, then IED testing, then function testing then pack it all up to site, then install and connect and test the connections and the functions ….

Despite all that, I had to investigate a blackout which occurred in 450 ms of four 275Kv substations connected by double circuit lines over 500 km.
The root cause was ONE incorrectly connected wire which was not discovered in any of the several stages of checking and testing by highly qualified technicians (including the factory point-to-point!?!?!?).

Note that when the secondary system is going to be replaced in X years’ time, they will start with a new blank screen because the available IEDs are four generations past the original ones. Even if the IEDs were the same, if the same wiring error was not detected again (noting several levels of testing failed to find it the first time), the blackout would occur all over AGAIN ….

IEC 61850 technology can dramatically IMPROVE RELIABILITY.

Once you have engineered a PTOC.Op to be sent to an XCBR as a GOOSE message (provided there is no annoying vendor-proprietary requirement for mapping the GOOSE via GGIO refer: Say No to GGIO), you don't have to re-engineer that every time you replace the relays or the entire protection system or just add a new bay -
what was working for the last X years in service as a configuration of the messages copied into a new IED is obviously FAR more reliable than starting again with a new blank CAD screen.

This is REUSABLE ENGINEERING giving rise to RELIABLE ENGINEERING.

 

Reliability is also about Availability.

Availability = MTBF / (MTBF + MTTR)

  • Mean Time Between Failures
  • Mean Time To Restore

If a relay fails in wire based systems, it may take anything from an hour to several hours for the technician to get to site. Then the bay has to be switched off or isolated in some way whilst the faulty relay is disconnected, then the new relay connected, tested and re-commissioned (assuming a spare was in fact readily to hand).
All up could be eight hours or more before the system is restored.

Some national electricity rules/codes would require a line to be switched off if the protection redundancy has been compromised for more than eight hours!

But what if you could do all that in FIVE MINUTES flat with the technician not even leaving home?
refer this presentation I recently gave to CIGRE B3 in 2012: My Papers and Contributions

Providing mechanisms for secure remote access from anywhere with confidence of having the right IED tools , such as the DCU,  and eliminating the need to physically be at the IED to disconnect/connect wires with the use of GOOSE and Sampled Values, gives the possibility to dramatically reduce the Mean Time To Restore.


 



Extra Notes: