The Sri Lankan powercut saga: a history of bad decisions
Jan 22, 2022
Why are we having all these powercuts? And why doesn’t anyone seem to be able to do anything about it? We dig into every aspect of power generation in this country: from the power plants, to their output, to how we ran out of fuel to keep the lights on.

Updated 30th March, 2022.

A note of recent events

Unfortunately, we’ve had a bit of a wild ride with electricity recently.

On the 10th of January, the Ceylon Electricity Board (CEB) confirmed that scheduled powercuts would happen on the back of a complex load shedding schedule.

The reasons given differed by source – many newspapers cited a breakdown of the Sojitz power station, a private thermal power generator that puts out 163 MW of power and counts itself as part of the Kelanitissa power plant. Our own fact-checking showed a slightly broader set of issues. CEB suffered massive power losses to the tune of at least 800MW. [1]

In public statements, the CEB was unsure of its supply of fuel to run even existing infrastructure. A.G.U Nishantha, President of the CEB Technological Engineers and Superintendents' Union, confirmed that the Ceylon Petroleum Corporation (CPC), which supplies CEB with fuel, had asked the CEB to settle their debts. He noted that the CEB’s ongoing inability to purchase fuel put the the country at risk for more long term power failures in the future.

However, just a day later, the Government got together; the President called an emergency meeting, and assured uninterrupted power supply.  CEB would settle that Rs 18 billion of its debt to CPC; the Central Bank of Sri Lanka would release forex for CEB’s urgent coal imports; and everything would go back to being wonderful. The CPC agreed to lob some fuel at the CEB, the CEB announced that power was back on the menu; no more load shedding, no more panic, the public would be able to go about their business.

Despite the promises by both the political infrastructure and the CEB, there have indeed been powercuts. On the 14th of January, we ran a survey asking people whether they had gone through powercuts over the last seven days, and verified that the political propaganda was just that - propaganda [2]. Fast forward to today, and we know power cuts are happening. 10 hours a day, with more to come. It’s gotten so bad that we’re crowdsourcing places to work from. A.G.U. Nishantha’s words now have the tinge of prophecy (or good statistics).

So. Why are we in this mess?

How do we get our electricity, anyway?

Let’s start with what we know.  To understand the state of power generation in Sri Lanka, we dug into three sources: the CEB’s 2019 annual reports, which details the past;  the long-term generation expansion plan for years 2022-2041 by The Generation Planning Unit of the CEB, which details the future and looks back to historic trends; and data from the Global Energy Monitor, a database of power stations.

Here’s a map of generation in the country.

Let’s make this easier to understand. Sri Lanka has three main types of power stations powering our national grid. They are: thermal (that is to say, fuel-powered, be they oil or coal), hydropower, and renewables. These three types of energy either come from CEB or IPPs - independent, private power plants that supply the national grid.

SourceCapacity (MW)
CEB Thermal1559.1
CEB Renewables1382.85

Note that the generation plan mentions 614 MW of independent thermal power and 683 MW of independent renewable energy, but looking into their own detailed charts, we couldn’t find those exact figures - or renewable plants.

So we’re going with whatever detail we have a breakdown for.

Let’s look at the detailed picture. These plants are organized loosely into collections,that are either called stations, or complexes. Within each station/ complex are multiple units that produce power; for example, New Laxapana is a unit within the Laxapana Complex.

Here’s a table of CEB’s thermal power plants (ie: those that require fuel to run). Units are in Megawatts (MW).

Plant NameStation / ComplexInstalled capacity (MW)
Unit 1Norocholai300
Unit 2Norocholai300
Unit 3Norocholai300
Sapugaskanda - ASapugaskanda80
Sapugaskanda - BSapugaskanda80
Kelanitissa CCPP – GTKelanitissa110
Kelanitissa CCPP – STKelanitissa55
Kelanitissa GT - 07Kelanitissa115
Kelanitissa Frame V GTTKelanitissa80
Uthuru JananiOther26.7
Barge MountedOther62.4
Thulhiriya (Emergency)Other10
Kolonnawa  (Emergency)Other20
Mathugama  (Emergency)Other20

And here’s CEB renewables:

Plant NameStation / ComplexInstalled capacity (MW)
Old Laxapana Stage ILaxapana28.8
Old Laxapana Stage IILaxapana25
New LaxapanaLaxapana100
Upper KotmaleMahaweli150
Small hydro (Udawalawa and Inginiyagala)Samanalaweva17.25
Mannar Wind ParkMannar Wind Park103.5

Adding to this are the Independent Power Producers (IPP) - private power plants. For example, we’ve been hearing about Sojitz a lot recently, often in connection with Kelanitissa. Sojitz is a private power plant within the Kelanitissa complex area, but in administrative terms, it’s a separate entity.

PlantInstalled Capacity (MW)
Sojitz Kelanitissa (Pvt.) Ltd163
ACE Power Embilipitiya Ltd*100
ACE Power Matara Ltd*24.8
Asia Power Ltd*51
West Coast (Pvt)Ltd.300

*Contracts of these power plants were scheduled to be terminated in April 2021.

All told, we have 3580.75 MW to play with, with a possible other 600+ MW that we haven’t accounted for here because of the reporting data.Based on the statistical data from CEB Statistical Digests and System Control Centre Annual Reports, this is quite enough to handle peak load. Here’s what the years have demanded of this equipment:

YearPeak Demand

Demand typically scales with GDP, as the CEB themselves note. The better an economy does, the more power it demands.

As for renewable energy – hydro technically falls into that category, but the classical renewables - solar, wind, mini-hydro, biomass – are still in the future. For now, this is what most of us are relying on. And it looks like, at full capacity, the CEB power infrastructure is quite capable of handling the load we put on it as a country.

So why are we in this mess?

Problem one: fuel - and lack of government foresight.

  1. As we can see from the power charts above, the bulk of our energy comes from thermal energy. Hydro is in second place.
  2. The CEB might have plenty of capacity, but that capacity only exists if we have fuel. In 2021, the Finance Ministry cut fuel imports in the face of dwindling forex reserves. This is on top of existing reductions: the government spent US$ 3,677 million on fuel imports in 2019, and reduced it to US$ 2,325 million in 2020, despite getting less fuel for their money – the price of crude oil had risen by then.
  3. This leads to limited stocks on hand. Throughout the first two weeks of January, the CEB warned about fuel shortages multiple times: of how Kelanitissa had less than 3.5 million litres, which was just enough for two days; and then how they only had enough to last until 18 Jan. Powercut updates were – no, still are – a constant back-and-forth, and erratic to say the least.
  4. The CEB also seems to be a bit short sighted when it comes to acquiring these critical fuel stocks. For instance, Udaya Gammanpila, Minister of Energy, communicated that the CEB should inform the CPC of their requirements ahead of time… which the CEB does not seem to have done.We find this general unawareness of needs a little hard to believe, given how we knew about CEB’s low fuel supplies weeks ago; EconomyNext was predicting fuel shortages by Jan 24.
  5. This isn’t CEB’s first tango with fuel demand and failure either – in 2021, a unit from the Norochcholai coal power plant gave out, taking 300 MW of power off the grid. To compensate, the CEB had to start Kelanitissa; this led to an increase in demand for fuel, shortages of fuel to Kelanitissa, and sporadic power cuts.
  6. To add fuel to the proverbial fire, CEB employees are not happy with how the CEB is run, and the CEB union has been threatening power cuts throughout. As of right now, they’re warning of escalating action against their General Manager, M.R. Ranatunga, after Gamini Lokuge, the Minister for Power and Energy, issued a letter extending his term of stay. Ranatunga was due to retire on the 13th of January, smack-bang in the middle of all these issues. The CEB Employes Union doesn’t like the fact that Ranatunga has been backing a deal for US-based New Fortress Energy buying shares of the Yugadanavi power plant and signing up to supply 10 years worth of LNG to the CEB.

Problem two: money.

  1. The CEB acquires fuel from the Ceylon Petroleum Corporation (CPC). The CEB also owes the CPC a lot of money. They’re in debt to the tune of Rs 91 billion. The CPC and the Minister of Energy have refused to hand over fuel to the CEB unless the payment is made, in dollars. As it happens, the CPC is itself in debt. As of seven months ago, statements from the President’s Office note that the CPC owes Rs. 652 billion to the Bank of Ceylon and People’s Bank. This kind of circular debt is fantastically stupid.
  2. The CEB is also bleeding. They have been so for a while. While trying to understand the CEB’s financial woes, we looked at the Chairman’s message for 2019:“...2019 marked the second lowest rainfall for the first half of the year since 1990. That in turn prompted exceptionally low hydro inflow to major reservoirs. As a result, generation from both major hydro and mini-hydro plants dropped to 4,794.5 GWh, a sharp 24.9% reduction compared to the previous year. With reduced hydro generation CEB had to rely on thermal generation. Total thermal power generation was 65% as at the end of the year, a 10.6% increase compared to 2018. This change in the generation mix towards more expensive thermal power, which was beyond the control of CEB, contributed to widening the gap between the production costs and selling price. That caused CEB to lose around Rs 5.85 for every unit sold. These mounting losses had a negative impact on CEB’s financial position. Although PUCSL is required legally to authorize a cost reflective electricity tariff for CEB, such has not been the case since 2014, while the Government required that several classes of consumers be continuously provided with subsidized electricity. As a result, CEB must sell electricity to consumers at a price less than the average cost of production. In addition, with no reimbursement of government mandated subsidies forthcoming during the year, CEB was forced to borrow further, costing the organization in 2019 a staggering Rs 22.5 Billion in finance costs.”
  3. Apparently, consumers also owe CEB money; Janaka Rathnayake, the PUCSL Chairman, noted that Rs 43 billion of debts have piled up since the start of the pandemic, exacerbating a vicious cycle of debt.
  4. Like it or not, Sri Lanka is now in an economic crisis. The Island report mentions that CEB had ordered 10 million litres of fuel on 12 January, and another 15 million the following day. Five ships have reached Colombo, carrying fuel, but the shipments could not be cleared for want of foreign exchange.

Problem three: a lack of actual power plants.

  1. As far back as 2016, the Public Utilities Commission of Sri Lanka (PUCSL) – that’s the regulatory body – seems to have warned the CEB of impending power crises, as the CEB had not commissioned any new power plants since the Norochcholai and Kerawalapitiya plants were added to the system in 2007. The Daily FT reported that a 2019 Cabinet paper recommending action and a purchase scheme from standby power generators went un-implemented, and that “none of the projects planned in Long-Term Generation Plans since 2011 have been implemented, except for the tender called for a 300MW LNG plant in Kerawalapitiya, which should have been in operation by 2019.” The PUCSL predicted capacity shortages in 2018 and 2019, and an increase in demand at 5.5% per year from 2015 to 2022, which subsequently seems to have been ignored.
  2. The same FT report cited that the Power and Renewable Energy Ministry ordered the CEB to get to it. Thirteen power plants were supposed to come online, including three solar plants, a barge in Galle with 100MW storage capacity, and another of similar capacity in Puttalam or Trincomalee.
  3. Nothing happened. In 2019 the Public Utilities Commission geared up to take legal action against the CEB over unapproved power interruptions. Minister Ravi Karunanayake, who was then the Power, Energy and Business Development Minister, alternatively guaranteed no power cuts shortly before then blaming power cuts on drought and the legal inability of CEB to tap agricultural reservoirs like the Mahaweli.
  4. The Kerawalapitiya power plant supposed to come online in 2019 was approved by the PUCSL only in 2021; a 300 MW affair that was just a fragment of the total 13 that were meant to exist by now.
  5. Why these plants, when the grid figures show an electricity generation system that looks capable of handling peak load? Because of the cost. We’re capable of handling peak with the thermal plants, which are expensive to run, which raises the cost of producing electricity. But CEB cannot pass that cost on to consumers, and so remains in debt. The extra plants were aimed (at least partially) at lowering the cost of electricity enough that CEB could dig itself out of the hole it’s in.
  6. What does the CEB have to say about this failure? We return to the Chairman’s message from 2019:“This disheartening gap between cost of production and selling price would have been addressed somewhat in 2019 had the previously approved plans for generating power at much lower cost been implemented without any impediments and delays – in particular, the first of the several 300 MW LNG plants planned. With regret, it is noted that the 300 MW plant has not materialized for reasons beyond the realm of CEB’s control. Notwithstanding the above, in 2019 CEB aggressively pursued the plan to get the much delayed and derailed generation plants off the ground. The base was commenced to establish the two LNG power plants at Kerawalapitiya, and for the extension of the Norochcholai coal power plant at our existing site. Also to upgrade the gas turbine plants with more modern and efficient machinery at our site in Kelanitissa and to facilitate the rapid expansion of solar and wind plants. The sooner the more efficient power plants come into operation, the sooner the State will be rewarded with reduced subsidy, with a pathway created towards making CEB a very profitable State entity. I sincerely urge all concerned to place a major emphasis and urgent focus on implementing the planned generation plans without any further delay as this is a material pathway to reduced operating costs and profitability.

Deja vu: we’ve been here before

What we’re seeing is that the CEB was underfunded by the State, their plans for reaching profitability by building more efficient power stations cut short, and to make matters worse, government spending on fuel imports dropped to the point where the CEB was left unable to handle the demand on the grid.

This is a persistent pattern throughout years of government. In the 1980s, even at the tail end of slow economic growth and the aftermath of violent political uprisings, we were adding significant capacity to the national grid. Nine hydropower generation units were commissioned between 1984 and 1988, adding up to large amounts of the Laxapana and Mahaweli complexes; it must have been a good time to be a CEB engineer.

Fast forward to 1996, and the country faced powercuts; again in 2001. By then it was apparent that there was an electricity crisis; the 2001 report from the Institute of Policy Studies details what it calls ‘the twin crises - the capacity crisis and the financial crisis’, and pointed out the need to add 160MW of power to the grid each year. Their report notes a situation almost identical to what’s happening right now: “the end- result in 2000 and 2001 has been excessive operation of CEB's own expensive thermal generating plants, extensive operation of more expensive standby generation at customer premises, and the expensive purchase of energy from private developers on short-term contracts.

In short, we’ve been down this road, Neo. We know exactly where it leads. Why we haven’t been able to avoid it seems to be because of a disturbing lack of foresight across multiple governments, with nobody stepping up to fix it even when forewarned. Power corrupts, but we need 230v to function. There’s an old adage that comes to mind:

For want of a nail the shoe was lost.

For want of a shoe the horse was lost.

For want of a horse the rider was lost.

For want of a rider the battle was lost.

For want of a battle the kingdom was lost.

And all for the want of a horseshoe nail.

That seems to be exactly what’s happened here. The problem – more power for a grid in need of it, fuel and energy being in the hands of loss-making enterprises that earned in rupees but spent in dollars, subsequently getting trainwrecked by the economic crisis – seems to have been ignored.  And now the economic crisis hits with a vengeance. We’ll do a piece on our economic woes soon, but for now, just remember that we need money to buy fuel with, and we don’t have a lot of money lying around.

The CEB is not an innocent actor in this saga. One thing that shows up in the coverage is that the CEB is a monolith. It controls almost all energy supply in the country. It has a pattern of avoiding due diligence and forecasting when it comes to power and fuel demand – this shows up in everything from their failure to account for failures and emergency fuel demand, despite the 2021 failure of Norochalai being another miniature lesson in fuel crises. The CEB also doesn’t like being regulated; CEB spokespeople have snubbed the PUCSL in public, stating that how they run their shop is not within the purview of, well, the national body to ensure that public utilities work. However, they don’t seem to hesitate to pass the ball back to PUCSL when possible.

But for all these issues, if you dig enough into their reports, the numbers are available. As a country, we’ve let the all-important question of electricity fall by the wayside.

What will happen in the future?

Unfortunately, the future is far from certain. There’s little we can do right now. The CEB has extensive plans that stretch well into 2041. Their documents, while in draft form at the time of writing, detail the retiring of various thermal plants and a multi-decade effort to bring on more renewable energy as replacements.

Whether any of this will actually happen remains to be seen.

Meanwhile, what we can do from our end is to help you with information. Everyone falls into a CEB division in this country, and if something goes wrong, you should know what to keep an eye on and who to call.

The problem is that CEB divisions don’t match the political and administrative boundaries of the country. While the engineers at CEB no doubt use and generate maps such as theirs, there’s a disconnect between the specialized knowledge they have and our public understanding of the country’s geography.

There are some oddities to note in their division schema. For example, Dehiwala and Ratmalana aren’t part of the Colombo Division; they’re Western South 1. Sri Jayawardenapura is Western South 2 - the same division as Awissawella, Horana and Bandaragama.

It’s not perfect - we took a shapefile from the UN OCHA’s Humanitarian Data Exchange program to compare against (a map of the country from the Survey Department of Sri Lanka.) The CEB map doesn’t conform exactly to the boundaries of the Survey Department. In conversations with CEB engineers, they mentioned that this map may also be outdated, as boundaries change from time to time; so suffice it to say that this is more of a very accurate artist’s rendering than a pixel-perfect map.

However, the next time you see a statement about powercuts in a particular division, at least you’ll be able to enter an address and figure out broadly if you’re going to be affected, and potentially which CEB division to call for a fix.

Good luck. Buy candles.

The data for this piece

CEB Annual Report 2019.pdf 5973021

CEB Long Term Generation Expansion Plan (draft).pdf 14412458


[1] At least three power plants, Kelantissa, Kukule-Ganga and Kerawalapitiya (including the controversial Yugadanavi plant in Kerawalapitiya) went offline at various times. Meanwhile, Sapugaskanda and the Barge at Colombo Port keep running out of fuel.

[2] Results of our survey:

This is a survey of 93 people, covering a period of Jan 7 to Jan 14. The map is an average of the amount of total reported power losses (color intensity) and and the number of people reporting from that neighborhood (node size).

Before we start to interpret this, let’s get the context sorted. Powercuts are a phenomena that affect many people in an area at a time – they affect neighborhoods, and so we can make some inferences from a single person reporting from one neighborhood. However, this is not a representative survey:  it does not show us the full scope of damage, merely a partial slice.

Secondly, these aren’t the addresses that people have given us. 200 people responded to our survey, but only 93 of them gave addresses of enough granularity for us to use OpenStreeMaps to generate lat/long points that captures where our respondents live to a street / landmark level (and not at a house level; we’re trying to preserve privacy here). So there are points that might be slightly further apart or closer together in reality than they are – the difference of a lane, for example.

  1. ==There definitely was a country-wide powercut in a band from 01 to 03 hours a day, with Vavuniya climbing to 07 hours a day, despite the assurances from the government and CEB sources==. While we do not have data from all nine provinces, the distribution of data points – in the Northern, Western, Southern, and Central provinces, and a scattering from the larger population centers in the Eastern Province – suggest that the powercuts were islandwide.2) Vavuniya seems to have been hardest hit, with a single node reporting a whopping 49 hours of power loss, seven hours per day across seven days. It’s one single person, though, so this should be interpreted with caution. Further North, Nallur reports smaller, two-hour losses, while Chavakachcheri has two respondents with very different experiences – two hours and nine hours respectively.
  2. ==Colombo is where we have the strongest clustering. Responses are incredibly varied==. Most places around Dehiwala and Mount Lavinia reported six hours of power cuts or a little above three, but there’s a line of responses close together that reported losses of 12, 14 and 30 hours. We suspect some were lucky because they were close to a business center.Kalubowila seems to have gotten by with three hours, with the outlying areas (like Kohuwala) being hit with six. The second largest hospital in the country is there, so it makes sense that by intent this time, or by dint of the original design, Kalubowila got by relatively unscathed.
  3. ==Ratmalana was more evenly hit – six hours for some, ten for some. Moratuwa reported four hours.==
  4. ==Kotte varies immensely, from 21 hours towards Nawala to two hours near Rajagiriya==, with the losses seeming to drop close to the city.
  5. ==Maharagama seems to have been hit pretty hard==, with 28 hours of power loss reported from the Godigamuwa area, close to the Arpico, tapering off to four hours towards the high level road. Again, the closer to the city center, the better the availability of power. ==Homagama reports close to nine hours lost.==
  6. ==Down South seems to have fared better==; a cluster of points around Unawatuna reports three-hour losses. Matara Four Gravets lost between three and seven hours. Losses seem less towards the coast; Thawalama and Thihagoda, for example, lost seven hours.
  7. ==Wellawaya, in Uva, reported 18 hours lost. To the East, Kalmunai and Akkaraipattu report six hour losses.==
  8. ==Kandy central seems to have been hit for between 10 hours and 14 hours== in Gangawata Korale. This time the pattern is reversed; the outlying regions, like Kothmale, Yatinuwara and Udapalatha report a small two-hour loss.

In general, total distribution of power loss was uneven throughout the country, with a few outliers:

For most people, these powercuts happened across a period of two or three days.