First, the bad news. In the first two months of this year, Eskom was forced into load shedding 18 times. That’s 18 out of 59 days, or 31%. The first week of January shouldn’t count, so that number’s more like 35%. Things looked bad in early February. For nine days straight, we lurched from Stage 1 to Stage 2 (to Stage 3, once) and back again. This, after Koeberg Unit One tripped due to human error, taking 900MW of supply offline a week earlier than planned.
The positive is that, since the restoration of most of Majuba’s output (that’s been offline since the ‘cracked’ silo), we have had load shedding only three times. Public Enterprises Minister Lynne Brown announced on February 18 that emergency repairs at the power station added 1 800 MW to the grid. This means Majuba is now generating 3 000 MW, compared with the 1 200 MW it’s been producing since the silo collapse in November.
It’s this 1 800 MW that’s staving off day-after-day of load shedding.
The lights are staying on, despite how dire Eskom’s forecasts have sometimes looked in its status bulletins.
But Eskom has been using load curtailment (basically, forcing its large energy customers to cut demand) to balance the fragile supply/demand situation. In February, according to its status bulletins, demand was “reduced down” (sic) on fifteen days. On these days (again, according to the bulletins), open cycle gas turbines (OCGTs) were not used. It also “reduced down” demand on February 9, without explicitly stating that the diesel peaking plants were not used. Either this was an oversight by the communications team, or they did use the diesel turbines to meet already-reduced demand. It also “interrupted load supply” to meet demand on February 23. (I assume power was cut to BHP Billiton’s aluminium smelters during the evening peak, something Eskom is contractually allowed to do for a limited time each month.)
An analysis of February’s supply-demand data shows that these aren’t small numbers. On the days where Eskom “reduced down” demand, it sometimes forced as much as 2 500MW (!) of demand to simply disappear because it didn’t have the generating plant to get anywhere close to meeting it. Some of this was achieved by load shedding, but on some days (particularly February 12, 13, 23) it quietly removed demand to keep the lights on.
The electricity generation situation is not as stable as it seems from the outside.
Remember, Eskom is also running without the 900MW from Koeberg Unit One (undergoing planned refueling and maintenance) which only returns to service at the end of May. That’s a nice boost for winter, but keeps the margins razor thin until then.
Unplanned outages remain stubbornly high. On Thursday, 8 200 MW of capacity was offline. That’s 19% of Eskom’s total generating capacity! Up until the middle of February, most of Majuba’s output was in this figure. But as we now know, Majuba is producing 1 800 MW more than it was a month back because of those “emergency repairs”. Had those not been done, would the unplanned outages number be closer to 10 000MW? You tell me…
The low number of 4 709 MW of planned maintenance is worrying, especially if you take the 900 MW from Koeberg into account. That means Eskom is only able to do 3 800 MW of planned maintenance on its coal fleet. It hasn’t been this low for a long time. And it’s surely not enough? But that’s a story for another day.
Thankfully, Eskom has confirmed that it has enough money to keep burning diesel at the OCGTs until the end of this month (the end of its fiscal year). Small mercies.
The past week has been relatively stable. More than once, Eskom has had significantly more generating capacity than it forecast it would have. On Tuesday and Wednesday, for instance, it saw peak demand of 31 600 MW and 31 337 MW, respectively. In its status bulletin last Monday, it estimated it would only be generating 29 421 MW and 29 273 MW, respectively. In reality, it managed nearly 32 000MW on Tuesday and 31 573MW on Wednesday. Something went right, for once.
Also positive is that it has managed to exceed (or come very very close to) 32 000MW four times in the last three weeks. It hasn’t achieved that since mid-January.
The official policy (at least from what I can divine from its public messaging) is that Eskom will “load shed only as much as is necessary”. There seems to be enormous political pressure on the utility to not load shed.
But we simply don’t know much of what’s actually going on. And, I suspect with the ‘War Room’ now actively managing the entity and GCIS massaging the communications, we’re going to know less and less…
* Hilton Tarrant works at immedia.