The Public Utilities Commission of Sri Lanka has approved one-hour evening electricity interruptions for three days — Monday 19, Tuesday 20 and Wednesday 21 October — with the island split into four groups that are cut in turn between 4.30 p.m. and 9.00 p.m.

The approval closes a question left open since Friday’s islandwide blackout: whether the loss of the Lakvijaya coal plant at Norochcholai would force scheduled load shedding. It will.

Why the cuts

The schedules released by the Ceylon Electricity Board name the cause directly. Demand management is authorised, the documents state, in the event the CEB “is compelled to take demand management measures due to insufficient generation resulting from unavailability of Lakvijaya Power Station caused by a blackout.”

Hiru’s Gold FM desk, reporting the CEB announcement, described the cuts as rotational and following “a breakdown” that left the plant unavailable and cut generation. The two framings differ in a way this desk has tracked since Saturday: the regulator’s document traces Lakvijaya’s unavailability to the blackout, while the CEB account and last week’s expert commentary call it a breakdown. The National System Operator has maintained throughout that no plant failed and that Norochcholai was shut down to standard procedure. The plant accounts for 810 MW.

Earlier in the evening the NSO had asked the commission — and, Hiru reported from a special media briefing called by the Energy Ministry at the Government Information Department, the subject ministry as well — for permission to cut supply. The NSO attributed the step to recent instability and sudden failures in the power system, and to the need to hold the system in balance. The Daily Mirror carried the request at 5.14 p.m., citing the NSO chief executive.

The schedule

Each group loses supply for one hour. Disconnection and restoration both fall inside half-hour windows:

DisconnectedRestored19 Oct20 Oct21 Oct
4.30–5.00 p.m.5.30–6.00 p.m.Group 1Group 4Group 3
5.30–6.00 p.m.6.30–7.00 p.m.Group 2Group 1Group 4
6.30–7.00 p.m.7.30–8.00 p.m.Group 3Group 2Group 1
7.30–8.00 p.m.8.30–9.00 p.m.Group 4Group 3Group 2

The rotation is deliberate. Every group’s cut moves one hour later each evening and wraps around from the last slot back to the first, so no group sits in the same slot twice across the three days.

That also explains an apparent discrepancy between filings. At the request stage both the Daily Mirror and Hiru reported a window ending at 8.30 p.m.; the approved documents and the later filings say 9.00 p.m. Both figures are in the schedule — the final group is restored “between 08:30 p.m & 09:00 p.m.” The earlier number was the start of that last restoration window, not the end.

Checking your group

Allocation is by grid substation and feeder number, not by town, and the published lists bear that out: Ratnapura, Matara, Panadura and Kiribathkumbura appear under both Group 1 and Group 4, and Anuradhapura, Dehiwala, Vavuniya and Trincomalee under both Group 2 and Group 3 — different feeders in each case. A town name alone will not identify a consumer’s group. The CEB directs consumers to the grid substation and feeder number that supplies them.

Two caveats are printed on each day’s schedule. The demand management capacity assigned to a group may change with actual demand, available generation and weather. And disconnection and restoration times may vary by up to 30 minutes either side of those shown.

Update, 20 October: day two proceeds, and a restart date arrives

The cuts went ahead on Tuesday, and the National System Operator has for the first time given a date for a Norochcholai unit returning to service.

NewsFirst reported at 7.03 a.m. that the NSO would again divide the country into four zones for one hour each. NSO Chief Executive Buddhika Samarasekara said one generator at Lakvijaya is expected to be re-added to the system by the day after tomorrow — that is, Thursday 22 October.

The filing also puts a figure on the shortfall: the loss to the national grid from the three Norochcholai generators being out of service stands at 810 MW. That matches the plant’s full output, so all three units are down.

What this answers, and what it opens

Two of the three questions this article recorded as unreported are now answered. A restart date exists for one unit, and the 810 MW figure is confirmed as the whole plant rather than a partial loss.

But the dates do not meet. The approved schedule ends on Wednesday 21 October; the first unit is expected back on Thursday 22 October — one day after the last authorised cut. No filing says what happens in between, whether the schedule will be extended, or whether the remaining two units have any timeline at all. One unit of three also does not replace 810 MW, so the grid is not whole on Thursday either — and no filing gives the returning unit’s output.

A word that keeps changing

Tuesday’s filing attributes the 810 MW loss to “the breakdown of the three generators” at Norochcholai. That is NewsFirst’s own wording in a report otherwise built on the NSO’s account, and it sits against the position this desk has tracked since Saturday — that the NSO has maintained no plant failed and that Norochcholai was shut down to standard procedure. The regulator’s schedules trace the unavailability to the blackout; the CEB and now NewsFirst call it a breakdown. Only the restart timeline is directly attributed to Samarasekara; the characterisation is the outlet’s.

The time window, again

NewsFirst gives Tuesday’s window as 4.30 p.m. to 8.30 p.m., the earlier of the two figures that circulated at the request stage. That is consistent with the resolution recorded above: 8.30 p.m. is the start of the final group’s restoration window, and 9.00 p.m. is its end. The two numbers remain two ends of one window, not a disagreement.

Update, 21 October: the first unit returns today, a day earlier than announced

The first Norochcholai generator is now expected back on the grid on Wednesday — the final day of the authorised schedule, and one day earlier than the National System Operator said on Tuesday. It will add 300 MW.

NewsFirst reported at 8.24 a.m. that the NSO states one generator at the Lakvijaya Power Plant “is scheduled to be reconnected to the national grid today,” adding 300 MW of capacity.

The date has moved forward

This is a change from what was reported a day earlier. On Tuesday, NSO Chief Executive Buddhika Samarasekara said a generator was expected back “by the day after tomorrow” — Thursday 22 October. It is now expected today, Wednesday 21 October.

That closes the gap this article flagged on Tuesday. The approved schedule runs 19–21 October and the restart had been placed one day beyond its last authorised cut, with nothing published about the interval. On the revised timing the unit returns within the schedule rather than after it. No filing explains the change, and none says whether the earlier date was revised or simply described loosely.

Wednesday’s cuts go ahead regardless. The island is again divided into four zones, with supply suspended between 4.30 p.m. and 8.30 p.m. — the same window, and the same two-ended figure, resolved above.

The capacity figure, and what it does not reconcile with

The 300 MW answers a question this article recorded as unreported: what the single returning unit would contribute. But it does not sit cleanly against the loss figure.

NewsFirst repeats that 810 MW was lost to the grid from the plant’s three generators. Three units sharing an 810 MW loss average 270 MW each, yet the returning unit is given as 300 MW. The two numbers are evidently on different bases — 300 MW is a unit’s rated output, while 810 MW is the loss as it reached the grid — but no filing reconciles them, and none states the plant’s rated total.

Taken at face value, 300 MW against an 810 MW loss restores about 37 per cent of what went missing. The grid is not whole on Wednesday evening, which is consistent with the cuts proceeding on the same day the unit returns. No filing gives any timeline for the second and third units.

The word has not settled

Tuesday’s filing called it a “breakdown.” Wednesday’s calls it a “total shutdown of the Norochcholai Power Plant due to a sudden power failure,” and again attributes the 810 MW loss to “the breakdown of the Norochcholai Power Plant.”

That remains NewsFirst’s characterisation rather than a quoted NSO position. The operator’s stated account throughout has been that no plant failed and that Norochcholai was shut down according to standard procedure following Friday’s blackout — the same split between “blackout” and “breakdown” this desk has tracked since Saturday, now in its fifth day without resolution.

Tuesday’s schedule, confirmed

The Daily Mirror published Tuesday’s group timings at 1.55 p.m. on 20 October, listing Group 4 first at 4.30 p.m., then Groups 1, 2 and 3 at hourly intervals. That matches the rotation in the table above exactly, and confirms the schedule ran as approved on day two. The Daily Mirror also states the cuts apply “if electricity generation becomes insufficient” — a conditional the CEB documents carry and the wire reports generally drop.

Wednesday’s rotation, confirmed by the CEB — and a headline that hedges its own lede

The Daily Mirror published Wednesday’s schedule at 10.48 a.m., and it matches the 21 October column of the table above exactly: Group 3 disconnected 4.30–5.00 p.m. and restored 5.30–6.00; Group 4 at 5.30–6.00, restored 6.30–7.00; Group 1 at 6.30–7.00, restored 7.30–8.00; Group 2 at 7.30–8.00, restored 8.30–9.00.

That confirms the rotation ran as approved on all three days, with each group moving one slot later each evening. It also confirms the 9.00 p.m. figure as the end of the final restoration window rather than a cut time — the point resolved at the top of this article.

The paper repeats the CEB’s two conditions verbatim: the cuts will be implemented only if necessary, depending on real-time demand, available generation and weather; and actual interruption and restoration times may vary by up to 30 minutes either side.

One thing to note about the framing. The Daily Mirror’s headline asks a question — “Will today be the final day of scheduled power cuts?” — while its own first sentence states it as fact: “The final day of scheduled power cuts will be implemented today.” The lede is not new information. It restates the end of the authorised schedule, which has been known since Sunday. The question mark is doing the real work, and it is the honest half: nothing published says what happens on Thursday, and with two of three Norochcholai units still down and only 300 MW of 810 returning, the schedule ending is not the same as the shortfall ending.

Update, 21 October: the National System Operator expects Norochcholai back — and gives a third shortfall figure

The operator has now spoken on the record, and it answers part of what this article had listed as unreported.

B. L. Pradeep Priyadarshan Perera, chairman of National System Operator (Pvt) Ltd, told the Daily Mirror by telephone that the Norochcholai coal plant is expected to resume generation by Thursday, the paper reported at 11.49 a.m. on Wednesday.

He put the shortfall caused by the shutdown at 250 MW, and gave the demand picture directly: “These days, we need 2,800 megawatts of electricity to meet nighttime demand, but we are currently facing a shortfall. On Monday night, we were able to supply only 2,570 megawatts.” Furnace oil is being burned to make up additional generation.

Three figures, three bases — and one that does not close

The chairman’s own numbers do not quite reconcile with each other. 2,800 MW of night demand against 2,570 MW supplied on Monday is a gap of 230 MW, not the 250 MW he gives as the shortfall. The two differ by 20 MW, and the filing does not say whether they describe the same night or the same measure.

The 250 MW also sits well below the 810 MW NewsFirst has repeatedly attributed to the loss of the plant’s three generators. These are evidently different bases — 810 MW is generation lost to the grid, 250 MW is the residual deficit after furnace oil and other plant is substituted in — but no filing states that distinction, and a reader moving between the two outlets is given a shortfall that shrinks by two thirds with no explanation.

The day the plant returns is ambiguous in the filing itself

The Daily Mirror published on Wednesday, says the chairman spoke “yesterday” — that is, Tuesday — and glosses his “tomorrow” as Thursday.

Spoken on Tuesday, “tomorrow” is Wednesday. Published on Wednesday, the paper’s own parenthesis makes it Thursday. One of the two is wrong, and the filing itself does not resolve it.

The archive does. NSO Chief Executive Buddhika Samarasekara told NewsFirst on 20 October that a Lakvijaya generator was expected back “by the day after tomorrow” — Thursday 22 October, as recorded above. The chairman’s “tomorrow” and the chief executive’s “day after tomorrow,” spoken a day apart, point at the same date. The Thursday gloss is therefore the one corroborated, and the Daily Mirror’s “yesterday” attribution is the likelier slip: a Tuesday speaker meaning Thursday would have said the day after tomorrow, as his own chief executive did.

Note also that these are two different officials of the same body — Samarasekara as chief executive, Perera as chairman — giving the grid’s position on separate days without either filing acknowledging the other.

The distinction matters because the authorised cut schedule ends Wednesday evening: a plant returning Thursday leaves Wednesday night dependent on the single 300 MW unit and furnace oil, with no authorised schedule beyond it.

This is also a different claim from the one already recorded above, where a single 300 MW unit was reported returning on Wednesday. “The plant resuming generation” and “one of three units returning” are not the same event, and no filing says whether the chairman meant the first unit, a second, or the full station.

Update, 22 October: the second unit returns today, and no extension is expected

Three days after the authorised schedule ended, Hiru News reported at 9.35 a.m. that the second generator at the idled Norochcholai thermal plant will be connected to the national grid today (22). The statement is attributed to Lakvijaya Power Station Director Engineer M.D. Nalaka.

Dr Pradeep Perera, chairman of the private National System Operator company, said it is expected that there will be no extension of power cuts after today.

A senior Public Utilities Commission official added a third voice: the National System Operator has made no request to extend the cuts so far.

What this closes, and what it does not

This article’s “Not reported” list asked whether cuts would be needed beyond Wednesday and when the second and third Norochcholai units would return. Thursday’s filing answers the first two:

It also partly resolves the ambiguity recorded above over whether the chairman’s “plant resuming generation” meant one unit or the station. Hiru’s Thursday filing is explicit that this is the second generator, implying the first had already returned and that the station is being brought back unit by unit. No filing states the first unit’s return date directly.

Note the structure of the assurance: the regulator’s officer confirms only that no request has been made, which is a weaker statement than a confirmation that no cuts will occur. The three statements are consistent but they are not the same claim.

Not reported

No filing says when the third Norochcholai unit will return, or whether the station is back to full rated output with two units running. None says whether cuts will be needed beyond Wednesday, or when the second and third Norochcholai units will return. None gives the plant’s rated total against the 810 MW loss figure, explains why the first unit’s return moved forward by a day, or states whether Wednesday’s cuts were shortened once the unit was reconnected. None gives the volume of capacity being shed, the cost of the oil and diesel generation being run in Lakvijaya’s place, or who bears it.

Ada Derana published on both the request and the approval, but the site has returned a CloudFront block to this desk for 30 consecutive days and neither filing could be retrieved or dated, so neither is cited here.