By Kumar David –

Prof. Kumar David

The public is king so it is necessary in matters of importance to put across issues in a way that most people can understand. I will simplify the Interim Report of the Committee investigating the 17 August, 2020 blackout and add my own comments as a guide discourse. Let’s see if I can manage to keep it both intelligible to the layman and technically acceptable.

Point 1: The maintenance crew at Kerawalapitiya switched on a 220kV circuit breaker energising a bus-bar that was earthed for maintenance. (Bus-bars are three long tubes (3-phase), which serve as junctions to connect together numerous items such as generators, transformers and transmission lines). The crew did not follow correct procedures and the interlock which should have prevented the error did no function or was wrongfully bypassed. The Committee, rightly, reprimands the CEB, which I guess is already, on an all Island basis, updating manuals, retraining crew and checking interlocks. This much is now common knowledge.

Point 2: Protection systems are coordinated. Circuit-breakers must trip around a faulted zone isolating only a small portion to “quarantine the disease” (rush of current, collapse of voltage). In the case of a crucial element like a 220kV grid station this should have been completed within four AC cycles (80ms) but took eight cycles (160ms). Enough time for all hell to break loose – generators were disconnected by their protection systems, transmission lines switched out and loads (chaps like you and I) kicked out, all in an attempt to save the system – that is keep some of it up and running. This is automatic; the poor buggers in the Control Centre can do no more than wet their pants and break into a cold-sweat in the 0.5 or so seconds in which all this happened. However, the big question is why did the coordination of system protection fail? Over to you CEB.

Point 3: When transmission circuits trip and loads a rejected there is too much generation on-line and not enough connected load. Its rather like a car engine running at full power and not enough traction load on the wheels. The car accelerates. In electrical systems if the input power is too much, system frequency goes up as generators speed up! Protection at Norochcholi promptly did its job to save its own generators. The rate at which speed was rising was too high so it cut back its generators and finally tripped out the station. Oh dear, the opposite effect followed (not enough generation) so frequency started falling!  This prompted ‘under frequency load-shedding’ to disconnect loads making other “load-less” generators trip. The system died in minutes

Now we come to two points, the first irritating, the second new and unfamiliar. 

Point 4: If Norochcholi trips out suddenly it takes a heck of a long time (days) to bring it up to full power again. There is a part of the boiler known as the drum which has to cool down and that takes days. What is irritating is that this has been known for 12 years and the CEB has been told ever so often to fit auxiliary systems to help expedite restart. Not much seems to have been done. That’s why we had rotating power-cuts for four days as generators elsewhere tried to cope with the absence of Big Fellow, the 900MW Norochcholi plant.

Point 5: This is the really odd one. The Committee seems baffled; the report is replete with “could be”, “maybe”, “can be attributed to” and “internal fault of the generator” (I don’t buy this last one). Nevertheless, I don’t fault it because neither the Committee nor CEB engineers who must have been advising it have got to the bottom of the jinx. The problem was repetitive and goes like this. Controllers started Victoria and connected to Biyagama. When they attempted to restore supply from there to load-centres the generators tripped. This happened three times. In desperation they tried again using Kothmale and three times again failed though this time the system held-on, on ‘life-support’ as it were, for up to 10 minutes. On the seventh attempt, again using Kotmale and after disconnecting some automatic controls, the system survived, but not without a minute or two of jack-in-the-box frequency oscillations between 48Hz and 54Hz; huge. Unlike the impact of the initial short-circuit which is in the millisecond time domain, these reconnection difficulties are in the slower many-second domain – stability, control-damping, load-mismatch and such electro-mechanical glitches. A very different ball-game.

Ok that’s the story in simple words. What needs to be done? Maintenance crews have to be better trained and manuals and procedures improved; interlocking mechanisms rechecked, protection coordination revisited, and the bloody mess regarding Norochcholi restarting has to be fixed. Intellectually, the challenge is to accomplish credible in-depth analyses of the start-shut drama on reenergisation from black.