OT: A small example of a cascade fault

greenspun.com : LUSENET : Electric Utilities and Y2K : One Thread

I was at work on night shift last night when I experienced a very unusual fault. Although this fault is not rleated to Y2k in any way, the sequence of events may help to describe what controllers and operators may experience at or after the rollover.

At 05:13 this morning, the SCADA alarm sounded. My first reaction was to look at the frequency meter which read 49.84 Hz. (we run at 50 Hz here in NZ rather than the 60Hz in USA). I silenced the alarm then opened the SCADA alarm list to see that the only 2 generators that had running at one of our stations had both tripped off line. They were not connected to the same 200 KV bus, so that immediately told me that it was a fault external to the station, and something quite major. The alarms listed, and the fault indications both showed that they went out on transformer protection, rather than generator protection, indicating that the fault was a serious earth fault, that had not cleared close to where the source.

I rang the Grid controller to determine whether it was OK to restart plant at the same station, or whether to provide the required power from another station. He was suprised to hear that we had lost generation, as his indication of a grid fault was in a city some 120 miles away. Between us we were able to work out just what had happened. (fortunately I have had 6 years experience as a grid controller so we were able to discuss events on an equal footing).

It appears that there was a fault in the 33KV distribution sytem in the city which was not cleared by the earth fault protection relay. A 220/110 Kv interconnecting transformer bank at the closest substation saw the fault and tripped, leaving only one other bank in service. It did eventually trip on overcurrent, but not before the fault started to cascade through the grid. There are 4 lines (at 2 different voltages 220 kV and 110KV) connecting our power station to the city where the fault occurred, and one of the 110 KV lines tripped at our end. Our 2 generators were both connected at 220KV and somehow saw large circulating currents in the transformer HV neutral earth connections and so tripped. Finally clearing the fault.

Generation was restored within 5 minutes of the fault occuring. The company that operates the transmission system will be asking some questions of the company that operates the distribution in the city involved about their protection and why it failed to operate correctly.

However the main point about this post is to demonstrate the speed at which events occured, and how SCADA was only able to help us determine what had happened after it was all over. The stand alone protection systems are what operated to finally isolate the fault, and by the time SCADA saw it, everything was already over.

Malcolm

-- Anonymous, September 02, 1999


Moderation questions? read the FAQ