Data vs spin — did Eskom mislead South Africans about the load shedding risk?

‘The system is stable and has sufficient emergency reserves,” Eskom said on Valentine’s Day. “Load shedding remains suspended, supported by a stable power system and adequate emergency reserves,” the utility said a week later (last Friday).

Then, less than a day later, on Saturday, South Africans had their electricity supply interrupted for the second time this year.

Daily Maverick spoke to three energy experts who suggested Eskom might not have been entirely frank with the public about the true state of their power system. Signs that load shedding was close were there if you knew where and how to look.

Apparent in the technical data but perhaps obscured to the broader public by the incomprehensibility of engineering jargon and repeated optimistic public messaging, Eskom’s publicly available data showed that “there were symptoms that we were heading towards load shedding about a week before the load shedding happened”.

This is according to Clyde Mallinson, an independent energy analyst who explained the data to Daily Maverick.

He explained that South Africa has about 3,000 megawatts (MW) of diesel capacity and that capacity is meant to be used sparingly and, if used at all, it’s meant to be used to be able to lift generation output during peak times “because it’s flexible”.

By diesel capacity, Mallinson was collectively referring to the sum total of generation capacity in South Africa provided by both Eskom and independent power producers (IPP) open cycle gas turbines (OCGTs), which are primarily powered by diesel. Eskom’s OCGT fleet, which includes Ankerlig and Gourikwa, as well as IPP OCGTs, combine for a total of just more than 3,000MW, but actual availability fluctuates based on fuel supply and unit status.

https://www.youtube.com/watch?v=l1Nu18VymyI&t=26s

Mallinson explained that during times of peak electricity demand, usually in the morning and evening, “we tend to use 3,000 or 4,000 more megawatts than we use in the middle of the day, and to meet that shortfall in the evening peak we normally use our pumped storage, which can do about 2,700MW, and if that’s not enough, then we can augment that with a bit of diesel.” 

But in the first two weeks of December, Mallinson spotted an odd trend where Eskom was “running diesel, not just at peak times, but sometimes through the whole night on one or two occasions”. 

“About halfway through December it stopped, mainly because there was a massive downturn in demand after about the 16th of December. But we burnt diesel on New Year’s Day, not just at peak times but pretty much 24/7, which was odd again because New Year’s Day is a public holiday. Then very late in January, we had that first little flirt with load shedding.”

Daily Maverick reported on Friday, 31 January that Eskom had implemented load shedding for the first time following more than 300 days of uninterrupted electricity supply.

Read more: Stage 3 load shedding across South Africa this weekend — power station ‘boiler tube failures’

Mallinson pointed to a more recent anomaly that stood out in Eskom’s data. 

“In the week leading to load shedding we were using our diesel as a provider of energy rather than a provider of power.” By this he meant diesel should be used for short bursts of power at peak times, not continuously for baseload energy supply. 

“Diesel should only ever be used to make up a power shortfall at peak times, and if we’re using it to make up an energy shortfall, it’s a sure sign that things are not well and that we’re heading towards load shedding.” 

He expanded, saying that if one looks at the diesel use on the Eskom data portal, “you can see from Sunday [16 February] through to Thursday [20 February] there was a massive amount of diesel used. And then on Friday it became massive. I mean, I don’t think they could have burnt more diesel than they did.

“In my personal opinion, we should have started stage 2 load shedding a week earlier than we started stage 3 [Saturday, 22 February] because we were missing load shedding by a whisker and we were running down diesel storage tanks and were only able to replenish our pumped storage by running diesel at night. 

“So there were symptoms that we were heading towards load shedding about a week before load shedding actually happened.