SA nuclear industry sharply criticizes the IRP 2010-2030 Update Report 2013

March 16th, 2014, Published in Articles: EE Publishers, Articles: Energize


A comprehensive review commissioned by the Nuclear Industry Association of South Africa (NIASA) is sharply critical of the draft Update Report 2013 of the Integrated Resource Plan for Electricity (IRP) 2010-2030 recently issued for comment by the South African Department of Energy (DoE).  

nuclear2The draft IRP 2010-2030  Update Report 2013 has not yet been finally accepted by the DoE and the SA government as an official update to the country’s national electricity plan, which was initially published in 2010 and accepted in 2011, for the period from 2010 to 2030.

The NIASA Review points out significant flaws in the draft Update Report 2013 which it says have “caused several of the recommendations to be faulty to such an extent that implementation of the current version of the IRP would impact very negatively on South Africa’s energy security and its economy”.

The NIASA Review recommends that the calculations of the Update Report  be repeated with the necessary corrections to the modelling assumptions, especially for the Base Case, which would radically change the power plant mixture in the direction of supplying more reliable and dispatcheable power at lower costs.

The executive summary of the NIASA Review is published below, and the full review document is available here.

Review of Integrated Resource Plan for Electricity (IRP) 2010-2030 Update Report 2013

Reviewer: Dr. Dawid E. Serfontein, School of Mechanical and Nuclear Engineering, North West University

Review commissioned by: Nuclear Industry Association of South Africa (NIASA)

(Version: 12 February 2014) 


The IRP Update Report (2013) (DOE, 2013d) lays an excellent foundation for modelling of South Africa’s energy needs in that it updated a comprehensive set of data from the 2011 version of the IRP. It also presents an impressive array of test scenarios that allows the reader to assess South Africa’s electricity future, regardless if his/her specific convictions about future trends. It also expanded the original planning horizon (2010 – 2030) out to 2050, which is an important improvement. It also demonstrates and documents a sophisticated set of tools for modelling the most efficient combination of energy supply options for each time slot.

Unfortunately the Base Case contains a number of unrealistic assumptions which seriously skewed its results. Many of these flaws were corrected in the various test case scenarios. Unfortunately the report based its main recommendations on the Base Case. Therefore these recommendations contain a number of very important statements, but which could not withstand the scrutiny presented in the present review.

The most important of the recommendations of the IRP Update are:

  • The electricity demand in 2030 is now projected to be in the range of 345-416 TWh, as opposed to 454 TWh expected in the policy-adjusted IRP. This means a reduction from 67800 MW to 61200 MW for peak demand, with the consequence that at least 6600 MW less capacity is required (in terms of reliable generating capacity).
  • The Update considers the aspirational economic growth suggested by the National Development Plan in order to reduce unemployment and alleviate poverty in South Africa. This growth rate (an average of 5,4% per year until 2030) is also aligned with a shift in economic development away from energy intensive industries which is assumed to dramatically reduce the electricity intensity of the economy allowing the growth rate to have a less imposing impact on electricity demand to 2030 and beyond. However, this is only an aspirational objective and the reality is that demand may not reach the levels required (especially not in the next five years) which raises the risk of overbuilding generation capacity to meet the target.
  • Additional important variables came to the fore, specifically the potential for shale gas, the extent of other gas developments in the region the uncertainty in the cost of nuclear capacity and future fuel costs (specifically coal and gas), including fuel availability.
  • The nuclear decision can possibly be delayed. The revised demand projections suggest that no new nuclear base-load capacity is required until after 2025 (and for lower demand not until at earliest 2035) and that there are alternative options, such as regional hydro, that can fulfil the requirement and allow further exploration of the shale gas potential before prematurely committing to a technology that may be redundant if the electricity demand expectations do not materialise.
  • Regional hydro projects in Mozambique and Zambia should be realised including the infrastructure developments that may have positive spinoffs in unleashing other potential in the region. Additionally regional coal options are attractive due to the emissions not accruing to South Africa, and in cases where the pricing is competitive with South African options, would be preferred.
  • Regional and domestic gas options should be pursued and shale exploration stepped up.
  • Continue with the current renewable bid programme with additional annual rounds (of 1000 MW PV capacity; 1000 MW wind capacity and 200 MW CSP Concentrated Solar Plant) capacity), with the potential for hydro at competitive rates.
  • Life extension of Eskom’s existing coal fleet may play a very important role, as it may delay the need for large capital expenditure on new plants. However, additional analysis should be undertaken to firm up on the costs involved and to weigh up against the environmental impacts of the greater emissions from these older plants.
  • Flexibility in decisions should be the priority to favour decisions of least regret. This would suggest that commitments to long range large-scale investment decisions should be avoided.

This review of the IRP Update found a number of serious flaws in its input assumptions which skewed its results. This caused several of the abovementioned recommendations to be faulty to such an extent that implementation of the current version of the IRP would impact very negatively on South Africa’s energy security and its economy:

  • An unrealistically high discount rate has been used.

This review of the IRP Update follows on and continues arguments that were worked out in detail in NIASA’s review of the Draft Integrated Energy Plan (IEP) (NIASA, 2013). Therefore the present report will best be understood by first reading the review of the IEP. Arguments that have been worked out in detail there will just be summarised briefly in the present review.

The IEP states that South African National Treasury recommended that it should use a real Economic Opportunity Cost of Capital (EOCK) of 11.3% as the discount rate for all its economic modelling. Our review of the IEP, however, showed on fundamental economic theoretical grounds that this should not be the case and that Treasury intended the 11.3% to be interpreted by the Department of Energy (DoE) as a nominal discount rate, which would amount to a real discount rate of only 5.3%, which would have produced fundamentally different optimisation results. On further theoretical grounds, and assuming technology specific-funding from nuclear vendors, it was argued that the discount rate for nuclear power should be lower than for the other technologies. A real discount rate of 3% was proposed. It was further shown that by using artificially high real discount rates, the cost efficiency of nuclear power for reducing CO2 emissions are drastically underestimated.

Unfortunately the IRP Update used an unrealistically high real discount rate of 8%, as well as a real cost of debt of 4%, which is substantially higher than the slightly less than 3% that Eskom currently borrows at. High discount rates dilute all costs or incomes that lay far into the future to virtually zero at the present time. Since new nuclear plants have the longest economic lives (60 years) of all power sources, they deliver half of their outputs between years 30 and 60, which fell outside the planning window of the IRP Update. The residual value of these plants, after the cut-off date of 2050, was then discounted back to low values in 2050, which discriminated substantially against nuclear. This substantially overestimated the contribution of capital costs to the Levellised Cost of Electricity (LCOE) of nuclear, which
caused it to erroneously appear to produce more expensive power than most of the competing plants.

Other flaws include that:

  • Intermittency costs, i.e. costs of stabilising the outputs of wind and PV (Photo Voltaic) solar in order to obtain dispatchable power, have not been explicitly allocated to each power source, especially not in the Base Case. This greatly underestimates the Levellised Cost of Electricity (LCOE) from these intermittent renewables and thus led to artificially inflated capacities for these technologies.
  • “Time of day” electricity selling prices have not been taken into account, which also unfairly benefitted intermittent sources, especially solar PV.
  • Unrealistically aggressive learning rates were assumed in the Base Case, especially for Concentrated Solar Plants with thermal storage (CSP). This skewed the modelling against nuclear and other power sources that produces base-load or dispatcheable peaking power. This flaw was corrected in the constrained learning rates Scenario, which demonstrated that nuclear will take centre stage when more realistic learning rates for CSP will be used. However, since the Base Case took centre stage in the recommendations, the recommendations were skewed against nuclear and therefore the recommendations argue for a scaling down of nuclear, as opposed to its scaling up, as was predicted by the constrained learning rate test case.
  • The security risks associated with relying on imported power were ignored, which skewed the results in favour of all import options, especially imported hydro, leading to the said scaling down of nuclear.
  • The reduction in coal use, in order to save on CO2 emissions, and the reduction in nuclear power capacity, due to the flawed perception that nuclear power will be prohibitively expensive, forced the models to rely on importing power from natural gas and regional coal power stations. As was quoted above, the IRP Update explicitly states that imported coal power is preferable to local coal power “due to the emissions not accruing to South Africa”, i.e. when we import coal power from say Botswana, Botswana is penalised for the CO2-emmissions, rather than South Africa. However, this view presents a very cynical view of global environmental damage:  While the scientific community is still debating the accuracy of the theories of the Intergovernmental panel on Climate Change (IPCC) on manmade global warming, the South African Government has explicitly adopted the IPCC’s position as correct and has committed to multilateral co-operation with the international community in order to mitigate climate change by reducing the emission of CO2 and other Greenhouse Gasses (GHG). Obviously, whether the CO2 is emitted in South Africa or in Botswana is irrelevant: the environmental damage will be identical. Therefore, when the DOE states that regional emissions of CO2 is preferred, it demonstrates cynical disrespect for the planet, i.e. the purpose of the whole CO2 mitigation program is then no longer to save the planet, but only to cook the environmental books in order to make South Africa look green, while we are in fact acting “brown”. A similar approach in the IEP led to scenarios where South Africa would end up importing 90% of its total energy. In our review of the IEP it was shown that this will seriously undermine South Africa’s energy security and even its national security. However, this senseless exercise will rob South African citizens of job and economic opportunities, which is at odds with accepted Government policy.
  • Real inflation or escalation of coal fuel prices towards export parity was not incorporated in the Base Case. As the High Coal Cost test scenario shows, this massively skewed the Base Case against nuclear, in favour of coal.
  • External costs were not internalised, in spite of explicit Government policy that it should be done. Since the health risk of coal, together with its global warming potential, produces the largest external costs by far, this skewed all modelling in favour of coal and against the sources with very low external costs, especially nuclear. For shale-gas, only its low levels of CO2 emissions were modelled, but the well-known problem of leaking methane gas, which is a twenty times more serious Green House Gas than CO2, during drilling, was completely ignored. This skewed the results in favour of natural gas.
  • The reduction in power demand, initiated by Eskom’s power buy-back program, and the associated reduction in electricity intensity of the economy, due to a switch from primary and secondary to tertiary economic activities, was viewed as an acceptable outcome, which may help SA to get through any power shortages. This reduction in demand was also used as one of the main motivations for delaying the nuclear new-build program or seriously scaling it down. However, in the review below it will be shown from the literature that the primary and secondary economic activities primarily supply jobs to the poor, while the tertiary activities supply work to the more well-to-do professions, such as engineers, lawyers, computer programmers etc. Therefore the downscaling of primary and secondary activities, such as mines and heavy manufacturing will hit the poor especially hard. Therefore it will be argued that, while the growth of the tertiary sector should be encouraged, the primary and secondary activities should also be supported as part of a pro-poor program. Therefore the reduction in demand should thus be mitigated by supplying plentiful and affordable electricity to the market. Government should thus rather err on the safe side by slightly over supplying, as opposed to starving electricity intensive industries, such as mines.
  • The IRP emphasizes its uncertainty about future demand and future Levellised Cost of Electricity (LCOE) for different power sources. The solution the IRP proposes for dealing with this uncertainty is to shy away from long-term commitments and to rather build only the minimum generation capacity required, with a view to then quickly adding capacity as power shortages begin to appear imminent. This is the “fly by the seat of your pants” strategy. This strategy will hit nuclear especially hard because nuclear plants have the longest lead times by far and thus fundamentally require long-term commitment to their construction schedules. The proposed aversion to long-term commitment will thus initially preclude new nuclear construction. When demand eventually bounces back, the long lead times of nuclear plants will mean that it will not be possible to construct them in time to meet this demand. Eskom will then have to resort to the technologies with short lead times, i.e. gas turbines, solar PV, wind turbines and coal plants. Unfortunately all these technologies produce power at higher LCOEs than nuclear, as will be shown below, which will mean that the selling price of electricity will increase unnecessarily, which will stifle economic growth even further.

NIASA’s own model calculations showed that:

  • Based on our more realistic assumptions, including taking into account the high external costs of coal, the cost of nuclear waste management and plant decommissioning, the longer 60 year plant lives of new nuclear plants, nuclear power is substantially cheaper than coal. If these assumptions had been applied in the Base Case of the IRP Update, as it should have been since adding external costs is confirmed Government policy, nuclear would have been the preferred power source, after hydro, which is not always readily available.
  • The LCOEs of Nuclear are substantially lower than coal at low WACCs, even when the nuclear overnight cost is increased to the much dreaded $7,000/kW-installed. This is because the LCOE of nuclear is much more sensitive to the WACC than to the overnight cost. Therefore the cap on the overnight cost for nuclear, proposed in the IRP Update, is largely meaningless and should rather be replaced by a cap on the LCOE, which will for all the plant types be determined by the combined effects of WACC, overnight cost, external costs and system costs, such as expensive pumped storage plants required to move the oversupply of power by PV solar panels during mid-day to the evening demand peak. This cap on the LCOE should then be applied to all the power sources, rather than to only single nuclear out.
  • The external cost of Generation III nuclear accidents is approximately 50 times lower than that of coal. This shows that, in spite of the image of nuclear power as a grave threat to humanity, as it is regularly portrayed by anti-nuclear organisations, Generation III nuclear power plants actually have the potential to save humanity from the grave health threats posed by the toxic emissions from coal power plants.


  • All the flaws in the IRP Update conspired to drastically overestimate the LCOE of nuclear and to drastically underestimate the LCOEs of coal, PV-solar, CSP and wind. Therefore nuclear capacity was suppressed or even absent in most of the scenarios of the IRP. However, our own modelling results showed that, once one corrects these flaws, nuclear easily outperform all its competitors, except hydro which is not readily available in South Africa. Therefore nuclear should neither be scaled down nor be delayed.
  • The result that, when all appropriate factors are included in the modelling assumptions, nuclear produces the cheapest power of all the readily available sources, also resolves the “fly by the seat of your pants” problem: The logical solution is then to make long term construction commitments to supply the minimum expected base-load demand with nuclear plants and the demand peaks with peaking technologies such as gas turbines and hydro. As and when unexpected demand increases appear imminent, the other plant types, with their shorter lead times, can then be rolled out quickly in order to supply this excess demand. Nuclear then become the cheap base-load foundation upon which the other, mostly more expensive, technologies can be added.
  • The abundant and affordable power produced in this way should then be used to stimulate the electricity intensive primary and secondary sectors of the economy, in order to create jobs for unskilled and semi-skilled workers as part of a pro-poor strategy.
  • We thus recommend that the calculations of the IRP Update be repeated with the said corrections to the modelling assumptions, especially for the Base Case. This ought to radically change the power plant mixture in the direction of supplying more reliable and dispatcheable power at lower costs.

Related Articles

  • Pyrrhic victory for fired former Necsa board members
  • Deadly air: How Eskom’s coal-fired power stations are killing people
  • The matrix converter: An option for FACTs devices in transmission networks
  • Little-known tax incentives boost business case for renewable energy
  • Off-grid embedded generation systems: Performance standards