An experienced electricity system modeller has corrected a suggestion that the draft 2018 Integrated Resource Plan (IRP 2018) uses levelised cost assumptions for the various generation technologies tested, explaining that overnight capital cost assumptions for each technology are key inputs.
The draft IRP 2018, which was published on August 27, is currently out for public comment until October 25.
Addressing a recent IRP workshop in Johannesburg hosted by the South African Solar Photovoltaic Industry Association and EE Publications, Eskom’s Keith Bowen said that, while the levelised cost of electricity is an important indicator of power station lifecycle costs, these were not used as input assumptions for the compilation of the IRP.
South African Network for Nuclear Education, Science and Technology national coordinator Dr Anthonie Cilliers argued recently the IRP modellers had been inconsistent in assessing technology costs, as levelised costs had been used for nuclear, while wind and solar photovoltaic (PV) had been tested using the cost outcomes associated with recent renewable-energy auctions.
“We don’t look at levelised cost, we look at capital cost. Because even with renewables, the capital cost is critical. If you are building a system that has a high penetration of solar PV and wind, there will be some curtailment, as there will be points when you are unable to take the energy produced. That curtailment has a real cost and, if you are only calculating a levelised cost, you are under stating that cost to the system,” Bowen explained.
He confirmed that the cost information for wind and solar PV had indeed been taken from the most recent bid window of South Africa’s Renewable Energy Independent Power Producer Procurement Programme (REIPPPP), also known as the expedited round. As a result, the capital costs used for wind and solar PV in the draft IRP 2018 were R18 847/kW and R16 555/kW respectively.
The overnight costs used for the other technologies listed in the IRP were derived from either independent studies, or costs provided by research organisations such as the Electric Power Research Institute (EPRI).
NUCLEAR COST ASSUMPTIONS
EPRI’s nuclear capital cost figure of R98 079/kW was rejected by the Department of Energy (DoE) in favour of the R69 764/kW arising as a result of a study into nuclear technology costs undertaken by Ingerop.
The overnight capital costs for pulverised coal are assumed at R40 031/kW, at R48 319/ kW for fluidised-bed combustion coal, while they are R9 226/kW and R10 131/kW for open cycle gas turbines and combined cycle gas turbines (CCGT) respectively.
Bowen also contested the argument that the costs associated with the REIPPPP projects could not be taken as an approximation of the levelised cost. “The bid programme gives us a pretty good description of levelised cost, because, as long as there is a competitive market environment, the bid price is actually a pretty good indicator of what the total cost is that needs to be recovered by the seller over the life of the plant.”
Where there was insufficient competition, as was the case for concentrated solar power (CSP), it was more difficult to determine whether or not bidders included mark-ups over-and-above what the levelised cost would be in a competitive setting. Therefore, for the IRP 2018 the technology cost of R58 833/kW associated with CSP, while derived from the REIPPPP bids, may overstate the true cost of the technology.
Even though levelised costs are not the main input in the model, the modellers nevertheless produced figures across the technologies, as well as a range of capacity factors.
The outcome is that wind and solar have materially lower levelised costs than is the case for new coal and new nuclear. Therefore, the model shows that a new-build mix comprising wind, solar PV and flexible generators, such as gas, would represent the least cost for South Africa to 2030.
At a capacity factor of 30%, the levelised cost for solar PV is R771/MWh, falling to R661/MWh at a capacity factor of 35%. For wind the levelised cost is R1 030/MWh at a 30% capacity factor, falling to R772/MWh at 40%. At an assumed capacity factor of 90%, the levelised cost for nuclear is R1 148/MWh, while for pulverised coal it is R1 056/MWh, with assumed coal costs of R580/t.
A wide capacity-factor range, from 30% capacity factor to 90%, was used to determine the levelised cost for CCGT, owing to the fact that the technology would be procured as much for the flexibility it provided as for the energy it could produce.
“At a 30% capacity factor, the cost is higher, but that additional cost is more than offset by the flexibility provided to complement the variable energy plant. So, while baseload coal is cheaper than baseload gas, we would still want to build gas for its flexibility.”
However, he also stressed that technology is changing and that other sources of flexible generation, such as battery storage, could soon begin competing with gas in the not too distant future.
DoE chief director for electricity Jacob Mbele said the department was acutely aware of the need not to foreclose on emerging technologies and would, thus, consider using the generation technologies listed in the IRP as “proxies” for other technologies that are able to deliver an equivalent solution to the power system at the same cost.
Mbele said that the technologies listed in the draft IRP 2018 were selected for being the lowest cost, but that, at the point of procurement, bidding could be opened to technologies not listed in the IRP, but which shared the characteristics of the solution sought for the system.
Bowen noted that, while the EPRI numbers used for the IRP assumed a cost of $600/kWh-storage for lithium-ion batteries, prices had already fallen to $300/kWh. “It’s already halved from what we assumed when we did the IRP modelling. So it’s something we do need to pay attention to.”
LIMITS ON RENEWABLES
Meanwhile, Council for Scientific and Industrial Research Energy Centre principal engineer Jarrad Wright said that, while the yearly build limits imposed on wind (1 600 MW) and solar PV (1 000 MW) would not have a material impact on the composition of the electricity mix to 2030, the implications beyond 2030 would be significant.
Owing to the fact that the limits on solar PV and wind deployments did not scale up together with the growth in the system as a whole, the contribution as measured against total system peak demand would fall consistently to 2050.
For solar PV the 1 000 MW constraint translated to 2.5% a year of system peak demand in 2020, which would fall to 1.7% by 2050. For wind, the constraint translated to 4% of total system peak demand in 2020, falling to 2.7% in 2050.
By contrast, Wright reported that, over the past four years, China and India had been deploying solar PV at a yearly rate equivalent to 8% and 6% of total peak demand respectively.
“Therefore, we may not be fully justified in constraining these technologies, particularly given their steeply falling costs.”
The least-cost scenario, dubbed IRP1, outlined in the draft IRP 2018 document, envisages no yearly limits on the integration of variable renewable energy to 2030.
However, the policy-adjusted version published for comment includes the limits and also deviates from IRP1 by including two coal-fired independent power producer projects, with a combined capacity of 1 000 MW, as well as 2 500 MW of imports from the proposed expansion of the Inga hydropower plant, in the Democratic Republic of Congo.
In total, the IRP contains seven scenarios, with the least-cost scenario, or IRP 1, coming in R875-billion cheaper by 2050 than IRP7, which includes a 5 600 MW of nuclear and which caters for both a carbon budget, instead of peak-plateau-decline as a carbon-reduction strategy, as well as market-linked gas price, instead of an inflation-based assumption.