Stage 4 as Eskom loses control of breakdowns

Offline coal plants now over a sky-high 17 500MW …
Between Friday night and late Saturday multiple units went down, including two at Lethabo in Vereeniging (pictured). Image: Waldo Swiegers/Bloomberg

Eskom was forced into the unusual situation on Sunday of pre-announcing Stage 3 load shedding scheduled for the evening peak on Monday and Tuesday this week, as breakdowns spiralled beyond 17 500 megawatts (MW) of coal generation capacity.

However, Eskom issued an update on Monday, announcing that load shedding will be increased to Stage 4 from 17:00 – 22:00 on Monday night.

It added that there was 1 904MW on planned maintenance, with another 17 255MW of capacity unavailable due to breakdowns. 

Due to the demand profile in winter, Eskom has trouble fulfilling demand in the morning and evening peaks (versus in summer where elevated breakdowns mean load shedding all day).

On Friday, it said capacity had recovered to a point where it made a late change to the evening load shedding level (from Stage 2 to Stage 1). In truth, breakdowns ‘improved’ from 16 305MW at noon to 15 997MW after 16:00.

Of course, the 300MW helped – but all Eskom did to reduce load shedding to Stage 1 was invoke 897MW of interruptible load supply to large industrial customers (effectively the aluminium smelters) during the peak.

The utility said on Friday evening that “barring any significant breakdowns, there is a low probability of load shedding being implemented during the weekend”.

Less than an hour before the evening peak on Saturday – at short notice – Eskom announced Stage 2 load shedding. It had lost units at its Arnot, Majuba, Hendrina and Kusile power stations, and two units at Lethabo since Friday night. The units at Majuba and Hendrina were returned to service by Sunday, but by then the utility had lost a further unit at Majuba, along with units at Tutuka and Camden.

That’s nine units lost in 48 hours, of which two were put back into service.

Source: Eskom data

To put this in perspective, from lunchtime on Thursday where it had 15 365MW of plant out of service not due to maintenance, Eskom lost a further 2 400MW of coal generating capacity by Saturday afternoon. This is equal to half the normal output from Medupi (before an explosion in August last year left Unit 4 out of service until 2024). This is over and above the 15 000-odd megawatts of capacity already offline.


In the past week, its coal fleet was only able to produce more than 24 000MW once (24 061MW) – at 18:00 on Thursday.

With an installed capacity of around 40 000MW – excluding imports, nuclear, open cycle gas turbines (OCGTs) and so on – this equates to an energy availability factor (for coal) of 60%.

The problem is that Eskom has only managed to produce more than 23 000MW for 16 hours over the last seven days. It is not certain that this is going to improve meaningfully.

This means Eskom has to enforce load shedding during the evening peak and use its diesel-burning OCGTs to match supply and demand.

Peak (MW) Mon May 9 Tue May 10 Wed May 11 Thu May 12 Fri May 13
Demand 32 701 31 946 32 487 32 106 29 961
Eskom capacity 29 753 29 676 31 154 32 252 30 143
Loadshedding 2 043 2 100 2 100 2 100 1 057
VPS/ILS* 261 227 447 179 897
Eskom OCGTs 1 938 1 901 1 498 984 675
IPP OCGTs 792 762 0 0 224
Renewables 449 689 556 794 648

* Virtual power station and/or interruptible load supply

From this data, one can see that in addition to the ±2 000MW of load shedding on Monday and Tuesday nights, Eskom used as much as 2 730MW of supply from OCGTs.


It also relied on the virtual power station mechanism to interrupt load to large customers. In total, this meant 5 034MW of emergency ‘generation’ on Monday at 18:00 (including the demand-side cuts). And this excludes another 1 801MW from its pumped storage schemes at the peak.

That Eskom believes the situation will ‘improve’ to Stage 2 by Wednesday night’s peak suggests that the second of the Kusile units that have been undergoing planned maintenance may return to service.

The current 2 094MW of planned maintenance comprises Koeberg Unit 2 (±900MW set to return in June), a unit at Kusile (±600MW) and Unit 4 at Medupi (±600MW), which exploded in August.

It has to include the last of these in planned maintenance for its capacity planning models (in other words, it can’t count on this returning to service until 2024).

Eskom executives believe that bringing back the two Kusile units that had been out for maintenance (one of which is believed to have returned) and the unit at Koeberg will ensure that load shedding is “limited” in the winter months.

For now, the sky-high breakdowns have edged the transmission unit’s ‘winter plan’ to the worst-case scenario – which sees breakdowns at above 15 000MW and requires Stage 3 load shedding.

In this scenario, there are as many as 104 days of load shedding between April and the end of August.

Diesel costs in this scenario run at between R1.2 billion and R1.5 billion a month for the rest of winter.

Things are not looking good.


Sort by:
  • Oldest first
  • Newest first
  • Top voted

You must be signed in and an Insider Gold subscriber to comment.


Well yes – of course !!!!
A lot of these plants are way beyond their design life and the Mighty ANC chose not to replace them !!!

If we wanted more baseload on stream tomorrow they should have started building the plant at least 6 years ago!!!!

Because they have not even started to think about it expect another DECADE OF DETERIORATION !!!

Do people understand what the F is happening here????????????????????

The 28 years of ANC corruption, mismanagement and destruction of Eskom (and all other SOE’s) and the continuing sabotage by ANC leaches is still derailing Eskom ability to function properly.

In the meantime massive theft and corruption discovered at Kriel power station.
Chile has increased electricity production by 5% EVERY YEAR over the last few years, mainly by projects like these :
A 400 MW CSP + TS Project that will deliver at 3.5 US$ cts/kWh, and includes massive storage.
A combined CSP+TS and PV project of 1 GW.
Morocco plans a massive combined wind,PV combined with battery storage, and CSP + TS Project of 9.8 GW altogether, that can deliver a 3.6 GW through two 3600 km long HVDC cables to the UK, @ £ 4.8 cts/kWh.

People… It’s not the age of the plants that the issue here.. It’s maintenance, or let’s say the lack thereof.. It’s the fact that each powerstation has at least 70 engineers working on site, one knows less than the other.. In 1990s there was not a single engineer on site.. There was a few expert engineers per power station working from Megawatt Park. Issue is that there is little to no skills at Eskom at this stage… If Maintenence schedules, outages and GOs etc is executed as per the OEM. These plants can run indefinitely…

Comment noted but with all due respect – show me an example of a coal fired power station in the world older than 60 years !!!!!!!!!!!

You may have a point casper1 BUT I would not put money on your statement, particularly in the third world. Having been involved with some generation run by old school engineers and artisans it is amazing what can be done BUT you have to spend money sensibly. Bit negative is if nearby coal reserves have run out, obviously. But for the rest, smart maintenance, replacement and understanding can make machinery work beyond its nameplate lifespan although also, producing below its nameplate power.

Sell the things, openly, fairly and smartly. No BEE or monopolies then let’s see what can be done.

It’s easier to fool people than it is to convince them they have been fooled…case and point, ANC = The Clown Factory

When you quote the cost of using diesel to generate power, is this a net cost, after subtracting the cost that would in any case be incurred by using coal, or is it just a gross figure, in which case it is misleading and not so responsible

End of comments.



Subscribe to our mailing list

* indicates required
Moneyweb newsletters

Instrument Details  

You do not have any portfolios, please create one here.
You do not have an alert portfolio, please create one here.

Follow us: