Energy is the critical component of the country’s future policy mix. Electricity is the necessary condition for economic growth but not a sufficient condition. The sufficient condition can only be supped by politicians and government policies.
The argument around electricity really comes down to the questions: Are renewables, primarily wind, efficient and sustainable?; and can they ensure security of electricity supply at a competitive economic price that can enable government to achieve its goal of “enabling economic growth and prosperity”? The theory and the facts of real life strongly suggest not.
There is a perpetual debate in the literature surrounding the approach utilizing Levelised Costs of Electricity (LCOE). The initial superficial analysis suggests that wind and solar are now cheaper than coal and nuclear. However, it is argued by many experts, that these figures do not consider the full additional costs that renewables impose on the system. The fact is that renewables deliver electricity less than 35% of the time, are intermittent, unpredictable, highly variable and have very low energy density. The additional costs these factors impose on the system are not included in the LCOE’s being currently utilized but are payable by the system.
These additional costs should be directly attributed to each energy technology responsible for the increased systems cost they cause. The findings of experts are that dispatchable electricity (coal and nuclear) with far longer 40-year to 60-year lives cannot be directly compared with non-dispatchable electricity (wind and solar) which have lives of only 20 years utilizing only the LCOE methodology. The literature is replete with articles on the subject supporting these arguments.
Literature Review
A report entitled ‘Critical Review of The Levelised Cost of Energy (LCOE) Metric’, by M.D. Sklar-Chik et al published in the South African Journal of Industrial Engineering December 2016 Vol 27(4) is relevant. It concluded that “LCOE neglects certain key terms such as inflation, integration costs, and system costs. The implications of incorporating these additional costs would provide a more comprehensive metric for evaluating electricity generation projects, and for the system as a whole. LCOE must not be used in isolation, but in conjunction with other project metrics and methodologies.”
A particularly important study and survey of modelling was completed by B.P. Heard et al entitled a ‘Burden of proof: A comprehensive review of the feasibility of 100% renewable-electricity systems’. After a review of 24 studies claiming to show that a 100% renewable electricity system is achievable, they concluded that “there is no empirical or historical evidence that demonstrates that such systems are in fact feasible”. Importantly their conclusion read: “The reality is that 100% renewable electricity systems do not satisfy many of the characteristics of an urgent response to climate change: highest certainty and lowest risk-of-failure pathways, safeguarding human development outcomes, having the potential for high consensus and low resistance, and giving the most benefit at the lowest cost.”
Amongst the studies they reviewed was the CSIR proposals regarding the high-penetration renewables in South Africa. The study concluded “both the use of the terms ‘technically feasible’ and the attempted costing of the proposed systems are inappropriate and premature”. Furthermore “a policymaker could have little doubt that the electricity demand proposed by the high-penetration renewable scenario from CSIR is not realistic and not in keeping with the imperative of alleviating wide-spread poverty in South Africa.” In another article Heard stated that the CSIR study was “as originally advertised, a thought experiment. It’s not a plan, it’s not a blueprint, it’s not ‘proof’ of any kind as the ripple of headlines and reporting suggests.”
Finally, the study by the OECD and NEA entitled “NuclearEnergy and Renewables: System Effects in Low-carbon Electricity Systems investigated”. System costs in this study are defined as the total costs above plant-level costs to supply electricity at a given load and given level of security of supply. This was a comprehensive study focusing on the costs that accrue inside the electricity system to producers, consumers and transport grid system operators referred to as “grid-level system costs” or “grid costs”. Total system costs would include those effects that are difficult to monetize and that could affect a country’s wider economy and well-being beyond the power sector itself. These grid costs would increase the LCOE.
Utilizing the average additional grid cost for the various technologies established in the study, and applying to South Africa, the additional grid costs of nuclear would be R0.026/kWh, coal R0.013/kWh, wind R 0.364/kWh and solar R0.493/kWh. More accurate costs must naturally be identified and estimated for South Africa. The above relevant costs must be assigned to each technology before determining any least cost mix.
Essential reading also includes a research report by D. Weißbach et al (Weißbach, et al. (2013)) on energy returned from energy invested (EROI) in Germany. The research shows that renewables are uneconomic and will lead to economic stagnation. Whereas the EROI of coal and nuclear are in territory that fosters growth.
The above research strongly implies that judging by experience overseas, the costs of wind and solar are likely to exceed those of coal and nuclear by substantial margins. This is a strong criticism of basing energy decisions purely on the simplistic LCOE methodology currently in use in the South African context. Certainly, this is the popular view expressed in the media. The issue being made is the so called least cost solutions as set out in the IRP 2016 need to be investigated, debated and thoroughly re-evaluated.
The costs for each technology must be adjusted for the subsidies currently being paid by the public. In this case they are paid by Eskom. However, these must be passed on to consumers. The costs are ultimately paid by South Africa’s long-suffering business, public and particularly the poor. Eskom are correct to resist these expensive contracts. Labour and labour unions are also correct in considering the impact on employment.
The final words in the work by Fritz Vahrenholt entitled ‘Germany’s Energiewende: A Disaster in the Making’ sum up the situation. “It will take a long time to repair the severe damage caused by a misled energy policy”. Energiewende in Germany is called by many experts a catastrophic failure despite the many excuses put up by its supporters. Recent events in South Australia have proved to be economically disastrous and have shown clearly that large wind leads to substantial economic damage. Electricity prices there have recently risen by over 100%.
It is true that Elon Musk has offered to put in storage. It is also true that economic storage will help alleviate the situation. Elon Musk is not being altruistic. His company makes batteries. At the price, this is cheap publicity. The batteries are currently very expensive and South Australians will pay a high price for all the backup wind needs. These facts sum up the future trend in South Africa if current IRP plans go ahead.
Modelling issues to be considered
The above arguments need closer evaluation as they apply to the South African situation. The following factors need to be fully examined, identified and evaluated:
• The real grid costs or additional grid costs and costs associated with grid stability for each technology.
• The cost of maintaining additional spinning reserve for those technologies requiring any increase in spinning reserve.
• The uncertainty and the economic risk to the economy associated with increasing uncertainty. All transitions from one energy source to another appear to assume they take place as perfectly smooth seamless transitions under conditions of certainty. There is always risk and uncertainty and therefore there is always a probability of transition delay and failure and this needs evaluation.
• The probability of any delays in reducing supplies from dispatchable sources or increasing supplies from dispatchable sources. Delays in supply and resulting unmet demand must be costed at the cost of unserved energy(COUE) currently R77.30/kWh.
• The increased dispatchable costs if the dispatchable technologies are forced to change their quantum of supply. In effect, the true additional costs of more frequent ramping up and down need to be evaluated.
• Costs of excess supply over demand supplied by non-dispatchable supply sources.
• It is important that there should be no limits on or constraints on CO2. This would give a proper judgement of actual costs and mixes.
Currently these are absorbed or paid by the grid operator but should be absorbed by the supplier. The additional cost costs for various technologies as set out above must be allocated to each technology. If this is not done, technologies have inaccurate cost allocations are not fairly evaluated.
These affect the so called least cost mixes calculated. Unless these costs are correctly allocated to the technology that causes the additional costs then that technology is in effect receiving an indirect or hidden subsidy. Subsidies, in turn, cause a misallocation of resources. The approach and planning of the IRP 2016 scenarios and the results appear to suggest they take inadequate account or underestimate the above factors. The outcomes bear little or no relationship to the reality in the world if evaluated against prices in other countries.
Externalities and other considerations
It is important that externalities are ignored for the purposes of calculating the LCOE and the cost impacts of technology systems. Invariably when these are measured not all externalities and environmental costs are considered. In addition, the real benefits associated with technologies are not considered.
All technologies are subject to externalities including renewables. They are extremely difficult to properly evaluate and are subject to personal views, bias and emotions. These costs and benefits should be debated by experts outside of the main costing for each technology and for the mix, which is the subject of this paper.
A full or even partial socioeconomic impact study has not been considered. The IRP 2016 involves major technological changes yet full or partial impact studies have not been conducted. A full socioeconomic impact study forms part of this separate process and this has not been done at all.
Extreme events which invariably do occur need to be statistically incorporated into the input. These are likely to occur more frequently with non-dispatchable highly variable, interruptible technologies causing substantial economic repercussions. Dispatchable technologies are prone to extremely events rarely. To exclude such events gives an unfortunate incorrect bias in favour of these non-dispatchable technologies.
Caution is required when major changes of technology are being considered. All externalities, costs and benefits can be classified into known-knowns, known-unknowns, unknown-knowns and unknown-unknowns. Unfortunately, as unknowns increase risks increase and one moves into the field of complex systems. Long term energy planning and complex economies falls into this category.
The work of Professor David Snowdon has been widely cited for developing an award-winning framework for decision making in complex systems in a complicated world. Snowden is the founder and chief scientific officer of Cognitive Edge, a Singapore-based management-consulting firm specializing in complexity and sense making.
This process recommends that innovative technologies should not be speedily deployed if the results are uncertain and decisions could be inflicting long term possible irreversible damage. These systems need multi- disciplined teams to evaluate them properly. Energy is a complex and relatively slow-moving field and caution should be exercised regarding decisions which once made affect large economic processes and have long term irreversible detrimental effects.
Summary
Merely taking the current estimates of LCOE’s for each technology and increasing them for the grid costs alone would change the situation considerably. Using the OECD average grid costs would increase LCOE’s for nuclear to R1.31/kWh, Coal R1.06/kWh, Wind to R0.98/kWh and solarto R1.11/kWh. These are only estimates based on an overseas study. Estimates of costs for South African conditions must be established.
All the other costs from the factors listed must also be attributed to each technology. It is almost certain that these costs will far exceed the grid costs set out earlier. Estimates place the true economic LCOE adjusted for the above factors at nuclear R1.36/kWh, coal R1.15/kWh, wind R1.80/ kWh, and solar R2.55/ kWh.
These numbers will cause outrage amongst renewable supporters, but the best guideline as to whether these prices are realistic or not is to consider pricing overseas. Prices in Denmark, Germany and South Australia have soared because of their moves towards high penetration wind as their major source of supply for their national or state energy supplies. In 2016, it must be noted that the prices paid by industry in Germany were 52% higher than France (nuclear) and 86% higher than Poland (coal).
In Ireland, often given as an example of the success of wind power, the equivalent comparisons were 34% higher compared to France and 64% higher compared to Poland. On average Germany’s price of electricity appears to be 44% higher than the average electricity price in Europe. South Australia evidently pays some of the highest electricity prices in the world. The above arguments suggest that the current arguments that least cost mix prices using the LCOE of each technology are badly flawed.
The purpose of this article is to state that it is necessary to revisit and re-evaluate the IRP 2016 planned and recommended electricity mixes going forward. More importantly the article is intended to contribute to a rational debate about the real energy choices that South Africa needs to face going forward.
