Nightmare at Megawatt Park – 23 November 2018

Although Eskom CEO Phakamani Hadebe at his system update on Friday 16 November said he did not anticipate any further load shedding, just two days later it had to implement load shedding on what is normally a reduced-demand Sunday afternoon.

 What caused this nightmare at Megawatt Park at a time when Eskom said it had 5 200 Megawatts (MW) of excess capacity?

     The reason for the nightmare was a slow return to service of those units that had been taken off line for maintenance. The weekly status reports that were re-introduced in June 2018 give a snap shot of peak load and available capacity, but load and capacity are dynamic so what happens at 2am and 8pm are substantially different as illustrated below.


To cater for the peaks, Eskom has pumped storage and Open Cycle Gas Turbines (OCGT), which despite their nomenclature actually run on diesel not gas. The pumped storage system uses two dams with a generation/pumping station in between the two dams. Pumped storage uses off-peak power to pump water to an upper dam and then releases this water to a lower dam to generate power during peak periods. OCGT uses diesel to power turbines, but this is more than 20 times more expensive than coal-fired power.

The problem on Sunday was that there was such a slow return to service of coal-fired power stations so that around noon the system operators needed the peaking power units to come online to meet the shortfall. In a Parliamentary session this week however Hadebe said that “we had problems with water and diesel”. That normally means that the upper storage dam had not been filled to capacity overnight, so there was not enough water to flow through the turbines in generation mode, while the OCGT did not have enough diesel in their tanks to fuel the turbines. That is why despite a nominal 5 200 MW excess capacity load shedding had to be instituted. During the week excess capacity had been far lower and substantially below the international benchmark of having 15% excess capacity, yet no load shedding was instituted.

The problem going forward is that former CEO Brian Molefe used emergency measures to fulfil his mandate of “keeping the lights on”. That meant he cut back on capital expenditure at the Eskom-tied coalmines, refused to sign the grid connection contracts for Independent Power Powers Producers (IPP) and used the peaking power OCGT units during the day instead of only during peak periods, with the result that OCGT costs shot up.

Blame should however not be attached to severely to Brian Molefe as he inherited a problem not of his making.

The original sin took place in the early 2000s, as the advice of the 1998 Energy White Paper was ignored. The White Paper in section 7.1 stated: “Although growth in electricity demand is only projected to exceed generation capacity by approximately the year 2007, long capacity-expansion lead times require strategies to be in place in the mid-term, in order to meet the needs of the growing economy.”

As it normally takes about six years to construct a coal fired power station, the decision to add to capacity should have been taken in 2001, but it was only in 2004 that cabinet approved a five-year investment plan in South Africa’s electricity infrastructure amounting to R93bn.

I asked former Eskom CEO Brain Dames why the new build programme did not include gas, as Sasol was using gas from Mozambique to supply its petro-chemical complex at Secunda, but he said Eskom would “stick to its knitting” and only construct large coal-fired power stations.

What happened next is however a case of what not to do in project management.

In 2009 the Eskom plan was to have the new build capacity starting to add to the grid in 2011 as shown below.


Source: Eskom

Ingula was only completed in 2016 and Medupi and Kusile are not yet completed.

Some of the problems that Medupi encountered are shown below.

Source: Eskom

Due to the delay in the start of this new capacity, the Department of Energy in 2011 launched the Renewable Energy Independent Power Producer Procurement Programme (REIPPPP) to supplement Eskom power. It was a resounding success and hailed as being a model for renewable energy programmes throughout the world, spurring investment into SA’s energy sector with R194bn in commitments. At the end of 2015, 6 376 Megawatts (MW) of power was successfully procured from 102 IPPs in four bid rounds of the REIPPPP. Of the 6 376 MW procured, just over 2 000 MW of electrical generation capacity has been connected to the national grid by the end of 2016. This is equivalent to just under half of the capacity of an additional coal-powered station, delivered in a third of the time.

What is also heartening is the reduction in cost over time.


Source: IPP Office


Source: IPP Office

The key to this success was risk mitigation as shown in the slide below from Paddy Padmanathan of ACWA Power who said the key to successful IPPs was the allocation of risk to those best able to mitigate the risk.


Source: ACWA Power

A recent successful IPP project in neighbouring Mozambique shows what can be done to address Africa’s power needs. In February 2016 Mozambique President Filipe Nyusi inaugurated the 120 MW Gigawatt Park gas-fired power station in Ressano Garcia. The project, which took 18 months to complete, was developed by South African investment group and majority shareholder in the project Gigajoule International.

The stalling by state-owned electricity utility Eskom on signing PPAs with 37 IPPs since July 2016 has undermined the credibility of the South African government. On 9 February 2017 President Jacob Zuma in his State of the Nation Address instructed Eskom to sign the outstanding PPAs.

“Eskom will sign the outstanding power purchase agreements for renewable energy in line with the procured rounds,” Zuma said.

Industry players said however that the damage has been done and IPPs would in future be more wary of participating in South Africa’s power procurement.

Politicians and bureaucrats do not seem to understand the concept of “time is money” so the millions of dollars in interest charges while no revenue was being generated since July 2016 will make private sector investors wary of preparing bids where there was little certainty.

“The changing of the rules in midstream has done severe damage to South Africa’s hard won reputation as a reliable partner. What you need from governments as an IPP partner is certainty, predictability and trust that commitments will be met,” Matthew Ash from Ash Konzult said.

Fazeel Moosa from Investec said that smaller projects have greater flexibility, while a modular approach to project development could also help establish trust between all parties concerned.

“What we are seeing is that phased projects have a greater chance of success so Noor 1 for instance leads to Noor 2, Noor 3 and so on,” Moosa said.

So that brings us back to the present day and whether load shedding is back with us as it was in 2014 and 2015.

The prospects are not good as Eskom faces three problems, two short term and one longer term.

In the short term it needs to improve its plant performance, in other words reduce the  percentage of unplanned outages to below 10% after a substantial deterioration this year.

This then impacts the percentage of excess capacity which should be above or near the international benchmark of 15%.


The second major short term problem is coal stocks. Due to Brian Molefe’s focus on cutting costs, there was not sufficient capital expenditure at the Eskom-tied coalmines, which has been compounded by the business rescue of Optimum, which used to supply 400 000 tons per month to Eskom.

The result is that Eskom spokesman Khulu Phasiwe said the risk of load-shedding is there and it is rising as South Africa heads into the summer rainfall period when some open pit mines cannot produce if there is excessive rain nor can boilers burn coal with a very high moisture content.


“In total, we have 11 coal-fired power stations that have less than the required minimum amount of 20 days stockpile. Out of the 11, five of them have less than 10 days of coal stock and that is the challenge,” he said.


This is in contravention of the South African Grid Code, which requires Eskom to maintain stock levels at a minimum of 20 days.


The last major period of load shedding was in September 2015, but there was periodic load shedding in June and August this year as Eskom workers went on an illegal strike as they are classified as “essential workers”.  During this time they blockaded entrances to power stations and did other acts that constrained the operations.

In November 2018 the Johannesburg High Court dismissed Eskom’s application to urgently interdict Kuyasa to resume contractual coal supply to the Majuba power station in Mpumalanga. Kuyasa supplied 140 000 tons of coal per month to Majuba, which is one of the stations that are below minimum level. The other ten affected power stations are: Arnot, Camden, Duvha, Hendrina, Komati, Kriel, Kendal, Matla, Tutuka and Kusile, which are all located in Mpumalanga. Eskom warned that should Kuyasa be allowed to cancel its coal supply contract because it wants a higher price, other suppliers will follow suit and export their coal, leaving Eskom without coal.

One of the best performing stations, Lethabo in the Free State, had an industrial accident on October 10 when the live steam pressure pipe from the boiler of Unit 5 ruptured. Based on the preliminary investigation, Eskom expects that Unit 5 will remain out of service for a minimum of three months.

Of the 12 coal supply contracts that Eskom CEO Phakamani Hadebe told Parliament in August that were being negotiated, six have been signed off and two of the six are already delivering coal, according to Phasiwe. Of the other four signed contracts, one will start delivering in December and the others early next year. Eskom is in talks with Exxaro, Glencore, South32 and others to provide as much extra coal as possible. In a typical year, Eskom buys around 114 million tonnes of coal with the majority at fixed prices below ZAR350 per tonne.

The longer-term problem is the so-called “utility death spiral”.  The utility death spiral takes place because utilities have large fixed costs, which means they issue many bonds that have to be serviced. If they get the mix wrong between supply and demand then the death spiral starts, as either they do not have enough of the utility, whether power, telecommunications or water, to satisfy demand, so customers look for alternatives, or they have too much, but once basic needs are satisfied such as a bath a day, then the spare capacity still needs its financing obligations to be serviced.

This problem is reflected in key metrics such as stagnant demand, escalating debt and ballooning costs and tariff increases, which drive away customers.

In the 2008 financial year, the utility had an installed nominal capacity of 43,037 MW, employed 35,404 people, and sold 224,366 Gigawatt-hours (GWh) of electricity. Eskom’s net profit was R974m, and total debt was R106.4bn. In the 2018 financial year, after tariff hikes of 350% over the preceding decade, nominal installed capacity was only 6% more at 45,561 MW, employee numbers had jumped by 37% to 48,628 and debt had mushroomed to R568.8bn.

Source: Statistics South Africa

To cope with the short term problem, OCGTs are once again being used more often, as they were in 2016.

Eskom expects a further R1bn to be used until the end of March or around R200m per month.

On this basis there will still be plenty of sleepless nights at Megawatt Park.

This contribution to Finance Friday was made by

Helmo Preuss

Forecaster Ecosa cc

+27 79 773 2458

Leave a Reply

Your email address will not be published. Required fields are marked *