In this opinion piece, Professor Dawid Serfontein enters the debate on the costs and benefits of South Africa’s renewable-energy fleet, responds to recent media articles on the matter and makes the case for a revision to the sequence in which new generation technologies are added to the electricity mix

I am impressed with the dramatic reductions in the prices of solar photovoltaic (PV), wind and gas power and therefore agree that it would be wise for South Africa to build a fleet of these renewables, backed up by gas turbines, fuelled with affordable imported liquefied natural gas (LNG). But then we must build it in the right sequence. The Renewable Energy Independent Power Producer Procurement Programme (REIPPPP), so far, had the cart before the horses and thus wasted billions of Rands. South Africa, however, can save billions of rands by simply interchanging the build sequence for the next three years or so to gas turbines first, then wind and then PV-solar. This article is not about promoting nuclear: if new nuclear can be brought online below about R1.00/kWh, it will be an extremely valuable low carbon and non-polluting power source. However, it will take between seven and 12 years to build the first nuclear plant. We will also only know whether we can get the nuclear plants below R1.00/kWh once the tender prices are in at the end of 2017. Therefore nuclear will not be discussed here.

Signing the new 591 MW PV-solar REIPPPP contracts for the 2016 Expedited Bid Window will cost Eskom R0.62/kWh. But what will this power save Eskom? Acting Eskom CEO Matshela Koko said in a recent Fin24 article, under the headline ‘Koko unpacks Eskom’s renewable costs, but experts disagree’, that it will be only R0.13/kWh worth of coal fuel for existing coal plants. Therefore, Eskom will buy this REIPPPP power at a loss of R0.62 less R0.13, which is R0.49/kWh.

However, in a Council for Scientific and Industrial Research (CSIR) study Professor Tobias Bischof-Niemz states that, in 2014, Eskom bought 2.19-billion kWh of PV-solar and wind power via Bid Windows 1, 2 and 3 contracts of the REIPPPP for R4 529-million. This saved South Africa a total of R5 305-million-worth of extremely expensive load shedding (R87/kWh), very expensive diesel fuel (R3.10/kWh) for Eskom’s open cycle gas turbines (OCGTs) and some affordable coal fuel (R0.23-R0.35/kWh). The country thus gained a net financial benefit (profit) of R780-million on the REIPPPP in 2014. I do not dispute these results. However, this report is misleading in that it selectively emphasises those results and arguments that support the profitability of these REIPPPP contracts and selectively neglects to emphasise those arguments and even its own results that show that specifically the PV-solar contracts ran at a loss to South Africa, namely: the 1.12-billion kWh of PV-solar produced only R2 806-million of value, but cost R3 084-million. The PV-solar thus produced a net loss to the country of R278-million. This loss is not mentioned. All the profit was thus generated by wind power. Their finding that wind generated a profit, but PV-solar a loss is key to improving the sequence in which we roll out renewables. However, the report camouflaged this crucial finding by adding the loss for PV-solar to the profit for wind so that only this joint profit was explicitly mentioned. More seriously, although this fact was not explicitly hidden, the report neglected to explicitly warn that 2014 was a year of exceptional power shortages, due to the collapse of the coal silo at Majuba power station. Due to these shortages Eskom suffered very high levels of extremely expensive load shedding and OCGT diesel burning, which created the said opportunities for the wind power to run at a profit by displacing these costs. The report did not warn readers that, once the situation normalises (as it did about six months after the report was published), load shedding would disappear and OCGT diesel burning would became very rare, so that the REIPPPP renewables would then almost certainly not displace much load shedding or OCGT diesel burning, and that therefore the whole “REIPPPP renewables are profitable” narrative of the report would then fall flat. The narrative of this report actually did fall flat on about August 8, 2015, when Eskom restored sufficient power supply. I calculate that PV-solar alone now generates the following loss: 1.12-billion kWh/year x (R3.28/kWh REIPPPP tariff in 2017 rands, less R0.57/kWh (displacement cost of coal and global warming in 2017), which equates to 1.12-billion kWh/year times R2.71/kWh, which is R3.03-billion a year.

The situation with wind power is slightly better, as the wind normally blows strongest during the early evening demand peak and therefore there is still a chance that wind power will still prevent some expensive OCGT diesel burning on some evenings. However, also due to the addition of pumped storage from Ingula, South Africa now faces very little OCGT diesel burning and therefore the gain from displacing this diesel burning will be small. Therefore the loss from wind power will also be larger. So the total loss may be closer to R6-billion/year.

However, as Bid Windows 1, 2 and 3 are now water under the bridge, I return to the latest expedited REIPPPP Bid Window. In commentaries submitted to Fin24, Professor Harold Winkler and South African Renewable Energy Council chairperson Brenda Martin argued that the REIPPPP PV-solar and wind power save full coal power cost, including power plant capital cost. I argue that the pollution cost of coal burning should be added to this. The draft 2016 Integrated Resource Plan (IRP) Base Case ignored the global warming cost of coal’s release of carbon dioxide (CO₂). However the Draft 2016 Integrated Energy Plan (IEP) priced CO₂ cost at R0.26/kWh. If we add that to the draft IRP Base Case’s costs, the full cost of new coal power becomes R1.12/kWh, made up of R0.25/kWh fuel cost plus R0.26/kWh global warming cost and R0.61/kWh capital and other cost (of which the cost of coal fuel and global warming cost are R0.51/kWh). Subtracting the latest R0.62/kWh PV-solar price from the R1.12/kWh total cost theoretically gives a profit to South Africa of R0.50/kWh.

So with all these conflicting numbers, who is right?

South Africa uses little power during the middle of the night, then develops a small power demand peak in the early morning when everyone cooks breakfast, a slight reduction in power demand during the middle of the day and finally a large power demand peak in the early evening when everyone cooks dinner and watches television. Eskom’s fleet of baseload coal and nuclear plants, with the help of a little wind power and pumped storage, can normally cheaply supply all demand up to the level of the morning demand peak. It is thus only during the early evening demand peak that Eskom is on some days forced to run its very expensive diesel-fuelled OCGTs. This is also the time when load shedding in exceptional cases occurs.

However, in 2014 the coal silo at Eskom’s Majuba coal plant collapsed, forcing South Africa into extremely expensive load shedding and almost round the clock very expensive OCGT diesel burning. In this period PV-solar power from the REIPPPP created substantial value by staving off the fraction of this load shedding and OCGT diesel burning that occurred during the middle of the day, when the sun shines. This situation continued through the winter of 2015, as delaying maintenance to run their coal plants 24/7 to try and keep the lights on, caused some of Eskom’s coal plants to break down. However, Eskom worked hard to catch up on its maintenance backlog and continuously succeeded in bringing new coal units online at Medupi and Kusile and recently at the Ingula pumped storage plant. Therefore, Eskom ended the crisis and brought sufficient power supply online on August 8 2015. Since then South Africa has had no load shedding and only a small amount of OCGT diesel burning and only during the evening demand peak, when the sun never shines. Therefore the CSIR study’s vision of PV-solar displacing substantial amounts of load shedding and OCGT diesel burning is now simply wrong and is likely to remain wrong for the foreseeable future.

So now PV-solar power displaces only coal power. The question remains: does this save Eskom the full cost of new coal power, including capital cost, or just the coal fuel and global warming cost? A well-balanced combination of wind turbines and PV-solar panels, backed up by gas turbines can supply fully dependable and dispatchable power, as the gas turbines can be started up if the wind drops or the sun sets. Therefore such a combination could replace a coal plant or a nuclear plant and therefore it could save South Africa the capital cost of the coal plant. However, the problem is that the contracts for new coal plants are fixed many years ahead of their delivery date. Therefore, even if Eskom were now to procure a large amount of PV-solar, wind and gas power through the REIPPPP, it will take many years before this will result in capital cost savings. In the short term, both Winkler and Martin are thus wrong: the latest round of REIPPPP PV-solar and wind power will not save Eskom one cent worth of capital cost for new plants. In the short term, the only benefit of these REIPPPP plants will be the displacement of the coal cost plus global warming cost. For our coal plants with the cheapest fuel cost this amounts to R0.13/kWh coal plus R0.26/kWh global warming, which equals R0.39/kWh. For the coal plant with the most expensive fuel this will amount to about R0.51/kWh. So the PV-solar cost of R0.62/kWh for the Expedited Bid Round will displace coal power at a loss of between R0.11/kWh and R0.23/kWh. However, there is no realistic scenario in which this new PV-solar can generate a net benefit to South Africa.

At a 26% load factor, the 591 MW PV-solar of the Expedited REIPPPP bid window will produce 1 350-million kWh a year, at an initial loss to our country of R0.23/kWh, or R310-million a year. The loss/kWh on new wind power will be smaller than on PV-solar. However, as the amount of wind power generated is much larger than for solar, the total loss for wind will be similar than to PV-solar. However, as more PV-solar and wind will obviously be added with each subsequent Bid Round, the total accumulated loss will quickly snowball to several billion Rand. Therefore, the policy of forcing Eskom to buy this PV-solar power through the REIPPPP is clearly irrational. Meanwhile, the South African Constitution states that all laws, regulations and policies taken by Government must be “reasonable”.

The following strategy could thus save the country several billion over the next three years or so:

• Honour all existing REIPPPP contracts (Bid Windows 1, 2 and 3). However, have the determinations for the REIPPPP PV-solar and perhaps also some of the wind power in the Expedited Bid Window set aside in a court of law by showing that these are irrational and thus unconstitutional. Therefore, refuse to sign these new contracts.
• Expedite the procurement of preferable combined cycle gas turbines (CCGT), fuelled with affordable imported liquefied natural gas (LNG), possibly through the REIPPPP. As LNG is cheaper than diesel and CCGTs use less fuel than the current OCGTs, such CCGT power can probably be acquired at roughly R1.50/kWh, even after the global warming cost of the methane gas that leaks into the atmosphere during shale gas mining has been added. R1.50/kWh is much too expensive to be used as baseload power. Furthermore, importing large amounts of LNG will expose South Africa to strategic risks, such as volatility of the gas price and a negative impact on our balance of payment. Therefore LNG imports should always be limited, for instance by also using pumped storage and hydro power as peaking sources.
• The main point of this argument is that if you are going to build the gas turbines anyway, it will be much more economical to build them first and the wind and PV-solar later.
• This gas power would then result in an affordable back-up power for the PV-solar and wind power contracts that have already been procured during Bid Rounds 1, 2 and 3. Since the combination of this PV-solar, wind and gas power will be fully dependable and dispatchable, this will result in a large strengthening of the country’s power reserve margin, which has been inadequate since 2008.
• However, as Eskom currently has enough power to supply all South Africa’s needs without load shedding and with minimal OCGT burning, the CCGTs should initially not be run, but just be kept in reserve. So instead of signing take-or-pay contracts, the owners of these plants should initially rather just be compensated for a reasonable rate of return on their capital costs.
• This arrangement would give South Africa considerable security of supply, i.e. if an accident were to again take out one of our coal stations, we could immediately start up the CCGT and supply power at the specific times of day that it was most needed. This would much more effectively prevent load shedding or OCGT diesel burning than PV-solar that can only do so during the middle of the day when the sun is shining or wind that can only do so when the wind is blowing.
• As and when the demand picks up, the CCGT can then burn the LNG, with the existing PV-solar and wind power from Rounds 1, 2 and 3 of the REIPPPP as fuel savers for the CCGT.
• When demand picks up even more, a few years down the line, we can then restart the bid processes for PV-solar and wind via the REIPPPP. As PV and wind power prices drop continuously and dramatically, we can then look forward to obtaining the same PV-solar and wind power contracts much cheaper than is currently the case. It is likely that the fully dependable and dispatchable new wind, PV-solar and CCGT combination that we will build at that stage, might deliver power cheaper than new coal. So at that stage construction of coal plants can be stopped and the PV-solar, wind and gas combination can then also displace the capital cost of coal, which would make the renewable option fully profitable to Eskom.

So while the full details of this plans still need to be worked out, the point is that the current REIPPPP strategy to deploy intermittent wind and especially PV-solar power before gas backup is putting the cart before the horse and is wasting billions of rands, without substantially improving our secure supply of power. However, by prioritising the deployment of gas backup, hydro or pumped storage and only then deploying new PV-solar and wind power plants, we will quickly improve our security of supply. As the wind and PV-solar will be added later, this move will also not threaten the dream of building a very strong fleet of renewables. It will just reschedule it in a manner that will actually work.

Serfontein is associate professor in the School of Mechanical and Nuclear Engineering, North-West University.