At 5am on Tuesday 26 March, loadshedding came to an abrupt end. Since then, South Africa has enjoyed 190 days without loadshedding, including our first winter in five years where the lights stayed on.

AmaBhungane took a deep dive into Eskom’s data to find out why.

The simple answer is: demand is down and the performance of the coal fleet is up. But a closer look at the individual stations reveals a more complex picture.

Demand is down, coal is up

• Want access to the Eskom data? We’re making the data publicly available order to enrich the debate about Eskom and the energy crisis. Download the per station EAF data, as provided to us by Eskom, from January 2015 to August 2024.

Between 2023 and 2024, the demand for electricity dropped by, on average, 804 MW an hour. That may not sound like much – Eskom has an installed fleet of 46 686 MW – but it’s almost one stage of loadshedding.

We looked at the “RSA contracted demand” from September 2023 to August 2024 and compared that with the previous year – in simple terms, this is the amount of electricity Eskom needs to meet the country’s demand at any given time.

The average hourly demand has dropped from 25 908 MW in the 2022/23 period to 25 105 MW in 2023/24. The drop in demand is particularly noticeable in the middle of the day, which suggests that rooftop solar on residential homes is having a real impact.

At the same time, the coal fleet has been performing better by, on average, 1 012 MW an hour.

The data shows an increase in average “thermal generation” – electricity derived from burning coal – from on average of 18 714 MW an hour in 2022/23 to 19 725 MW in 2023/24.

As these two lines move closer together – demand coming down and coal going up – the gap that needs to be filled by nuclear, renewables, hydro, diesel or with loadshedding gets more manageable. 

And over the past year, the gap has narrowed by, on average, almost two stages of loadshedding.

• Eskom publishes graphs showing how demand is met on a day-to-day basis through the Eskom data portal. In the past week, for instance, we’ve relied heavily on wind, imported electricity from our neighbours and burning diesel.  

What about diesel?

A question many people have is, are we blowing Eskom’s budget on diesel just to keep the lights on? 

We looked at the data for the Open Cycle Gas Turbines (OCGTs), specifically the period from April 2024 when loadshedding stopped. Between 1 April and 31 August 2024, the OCGTs – those owned by Eskom and by Independent Power Producers – generated 510 GWh of electricity. 

That’s not nothing, but it’s a lot less than the same period in 2023, when the OCGTs produced 2 359 GWh of electricity. In other words, we used four times more diesel over winter in 2023 than we did this year.

[Graphic – diesel] 

But despite using less diesel – which normally keeps the loadshedding wolf at bay – the data also shows that “manual load reduction” (aka loadshedding) dropped as well: from 7 212 GWh between April to August 2023, to zero this year.

But the picture is not entirely rosy across the board. 

The coalface

A deeper dive into Eskom’s data indicates that while some coal-fired stations are improving, others are facing challenges.

The Energy Availability Factor (EAF) of an Eskom plant is the amount of time a power plant is available to produce electricity – expressed as a percentage – after planned maintenance (PCLF), breakdowns (UCLF) and other causes (OCLF) have been taken into account. (Strictly speaking, it’s not how much electricity a power station actually produces, but this is how Eskom prefers to display its data.) 

Our analysis of the data shows that the EAF of the coal fleet has improved, particularly over the five months from 1 April to 31 August 2024, from 51% in 2023 to 60% in 2024. This is particularly impressive considering that Eskom also spent marginally more time doing maintenance this winter (+2%) when compared to 2023.

Several coal-fired stations – Kusile, Hendrina, Grootvlei, Matla, Kendal, Majuba, Tutuka, Kriel and Matimba – have improved their EAF in 2024, while others – Arnot, Lethabo, Camden, Medupi and Duvha – have gone the other way.

There can be good reasons for this: Medupi, for instance, had a lower EAF (April to August 2023: 83%, 2024: 70%) but mainly because it was spending more time doing maintenance (April to August 2023: 2%, 2024: 13%). 

On the other hand, a station like Tutuka is still a basket-case: its EAF has improved to 32% (April to August 2024) but it still spends an unacceptable 64% of its time in a state of breakdown. 

But by far the biggest factor in Eskom’s turnaround are the repairs and improved performance at Kusile. 

Kusile has struggled since its launch in 2017. It was designed to be one of the largest coal-fired power plants in the world at 4 800 MW, but seven years later it is still only operating at 2 880 MW. A devastating fire took out unit 5 in September 2022, and it took a year to fully resume operations. 

The data shows that Kusile’s EAF increased from a dismal 11% between April and August 2023 to 79% in April to August 2024. This meant that Kusile was available to produce an additional 7 164 GWh of electricity between April and August this year, compared to last year, which has been a major factor in reducing loadshedding.

On the other end of the spectrum is Duvha, the second worst performing power station in the country – between April and August 2024, its EAF dropped from 58% to 39%, entirely due to a spike in breakdowns (UCLF), which increased by 19percentage points.

Prolonging the inevitable? 

Recently, Eskom announced it is abandoning previous plans to decommission Camden and Grootvlei by 2025, and Hendrina by 2026. Instead, it wants to postpone their shutdown to 2030.

Electricity minister Kgosientsho Ramokgopa has said: “What we need is [for those power stations] to continue to give us those megawatts so that we can sustain the South African economy.”

But how much do they actually contribute?

All three power stations are on the smaller side. Although Grootvlei has shown an impressive improvement in its EAF, with a 20 percentage points increase in April to August 2024 (compared to the same period in 2023), it remains a relatively small power station with a capacity of just 570 MW.

When measured against an installed capacity of 46 686 MW, its contribution to alleviating load shedding is minimal.

Hendrina’s EAF has also improved by an impressive 28 percentage points (April to August 2023 vs 2024) but at 1 098 MW its impact is modest.

Camden is the biggest of the three stations (1 480 MW). Its performance has actually declined from 65% to 56% (April to August 2023 vs 2024), but this in part is because it’s doing more maintenance.

Collectively, these three stations accounted for an additional 1 075 GWh of available power between April and August 2024, out of the additional 12 785 GWh that was available to be added to the grid through improved EAF.

This raises the question: is Eskom merely delaying the shutdown of aging infrastructure because of the politics around decommissioning, or are these efforts genuinely part of a broader plan to stabilise and revitalise the grid? 

What the figures suggest is that the system remains tight, and every megawatt is still needed to keep loadshedding at bay. 

How the turnaround happened

Even so, after years of Eskom’s performance deteriorating, the turnaround – particularly since April this year – looks miraculous. Eskom has pointed to several reasons: the return of the Original Equipment Manufacturers (OEMs), a new incentive scheme, and replacing underperforming power station managers.

Bringing back the OEMs to carry out maintenance means that Eskom is relying on the companies that originally designed and built the power station equipment to handle its maintenance. The OEMs have expert knowledge of the equipment and should have a better understanding of how it operates, which, in theory, means better maintenance and an improved long-term performance. 

During a press briefing on 26 August 2024, Eskom’s chairperson, Mteto Nyati, revealed that Eskom had also begun implementing a monthly incentive bonus for employees at power stations that manage to reduce breakdowns (UCLF).

Between April and August 2024, breakdowns were 10% lower than during the same period in 2023.

In response to follow-up questions, Eskom told us: “The performance incentive is paid only to staff contributing directly to the improved structural performance of the system, which includes all power stations. It is based on achieving daily targets at a station level. The staff at currently under-performing stations will not meet the targets and thus not receive the incentive payment for the relevant period, however these stations are receiving extra resources to improve performance.”

Eskom added that the incentive scheme has been running for more than a year, but that it’s not intended to be permanent and will be reviewed from time to time. 

Eskom famously implemented a “red card” system under former CEO Matshela Koko which was more stick than carrot: managers whose power stations didn’t perform could be summarily dismissed. But the system was also criticised because it was alleged that some power station managers avoided taking units offline to deal with minor breakdowns which ultimately led to bigger problems down the line. 

We asked whether the incentive scheme risked a repeat of this problem, but Eskom dismissed this concern: “There has been a visible, structural shift in the generation fleet performance that demonstrates the quality of work Eskom’s teams, supported by our stakeholders, have delivered to strengthen the infrastructure of this country. We have ensured that what is fixed stays fixed, cost effectively.” 

It added: “The specific performance metric chosen is unplanned energy losses which is annually audited by external auditors. This metric is objective and measurable. A department manages and verifies all technical performance statistics monthly.”

In addition to the incentive program, Eskom has also said that replacing power station managers has contributed to the turnaround.

In a discussion with Foord’s Chief Investment Officer, Nick Balkin, Nyati revealed that 45% of the managers at Eskom’s power stations have been replaced since March 2023. This decision, according to Eskom, is part of the Generation Recovery Plan.

We asked Eskom to clarify which station managers had been replaced so that we could track whether new managers had achieved improved performance, but Eskom declined to give us that information, but told us:

“Some experienced General Managers were redeployed from the clusters to assist with the turnaround at Power Stations… some General Managers lacked the experience to handle the complexities at certain stations.”

Correction: This story has been updated to distinguish between percent and percentage point increases and decreases in the EAF.