6.2 Licensing efforts

    It is acknowledged by all sides that for sales to most markets outside South Africa to be possible, certification by a highly experienced, high credibility nuclear safety regulatory agency is required. This is not to denigrate the competence of the South African regulatory authorities, but reflects the risk aversion of electric utilities and those that supply finance to power station construction particularly as electric utilities are exposed more to investment risk. One of Exelon’s main contributions to the venture was their role in piloting the design through the US NRC procedures. The NRC had begun to review the design and had collaborated with the South African National Nuclear Regulator (NNR) on design issues but when Exelon withdrew, the NRC quickly wound down licensing activities.  It has been reported that PBMR (Pty) Ltd officials met with NRC officials in October 2004 to discuss design progress  but it does not appear that NRC is carrying out any substantial design evaluation.
    Without NRC approval for its design, it is not clear that the Demonstration Plant would have much value in promoting foreign sales. Until the design had been approved by the NRC and finalised, construction cost of the commercial export design cannot be estimated accurately. If the Demonstration Plant design differed significantly from what was required by the NRC (for example if the Demonstration Plant was built without a pressure containment and the NRC indicated it would require one for any plant built in the USA) potential buyers would see construction and operation of the Demonstration Plant as having only limited demonstration value
Required information
    The FEIR should state what strategy has been developed to obtain internationally credible regulatory clearance for the commercial PBMR design and how this would fit in with the Demonstration Plant.


6.3 Construction cost and cost of associated facilities

     Repaying the cost of construction of the plant has always been expected to be the major element in the overall cost of power from any nuclear power plant. Its importance has increased in the last decade as attempts to introduce competition to the electricity industry have increased the cost of capital raising the charge for repaying the construction cost.
     The FEIR contains no information on the expected construction cost of the Demonstration Plant or on the commercial plants. It merely states: ‘The cost to build the PBMR demonstration module will probably be available on completion of the project business plan (year end 2002).’ The DFR contained no details on the cost of the Demonstration Plant.
     In 1999, Nicholls (Nicholls, 2000) forecast that the construction cost would be about US$100m (then equivalent to about R600m) for a single commercial module, presumably as one of 8-10 units installed on one site. The strategic importance of this estimate was that it placed the price of the PBMR at around the US$1000/kW of installed capacity, a level above which it was widely assumed that nuclear could not compete with gas-fired technology.
     Nicholls  was quoted separately as estimating the cost of the Demonstration Plant as double the settled down commercial cost with a further US$100m for a fuel sphere production plant. The total cost of the Demonstration Plant was therefore then estimated to be about US$300m or a little less than about R2bn.
     In 2002, the DFS (PBMR (Pty) Ltd, 2002b, p 23) suggested some cost increases had occurred and the target construction cost for commercial units was now placed at US$1000-1200/kW. However, there appear to have been major cost increases. These have been masked by three factors. First, it is not clear whether the current cost estimates cover as full a range of costs as the original estimates, for example, if the cost of the first fuel load was omitted (conventionally this is included in the construction cost), the apparent cost would fall masking real cost increases. Also, it is also not clear whether the new estimates are now a cost or a price (i.e. including the profit). Second, there has been some depreciation (about 10 per cent) of the Rand against the US dollar between 1998 and 2004. However, the third factor is the most important. In 1998, the design was expected to produce a net output of 110MW but commercial plants are now expected to have an output of 165MW, an increase of 50 per cent. This would allow the cost of a module to rise by 50 per cent without increasing the cost per kW.
     In September 2001, Nicholls  admitted the original schedule for the Demonstration Plant had slipped. He then projected start of construction for 2002, with completion expected in 2005 and commercial sales to begin in 2009. There was discussion about up-rating the output of the plant to 130MW to be achieved without significant cost increases.  The Chief Executive of one of the partners in the project, Corbin McNeil of Exelon, was quoted in the same article as saying the upper limit on output was 150MW but he assumed the final figure would be 130MW. McNeil also stated the cost of the first module had risen to about US$300m. This article also acknowledged delays in the design work particularly with the turbine and the graphite liner.
     In 2002, the DFR, (PBMR (Pty) Ltd, 2002a, p 50) stated the design could be up-rated to 137MW ‘without a significant increase in cost’. This meant that costs per module could increase by nearly 20 per cent whilst still remaining within the US$1000/kW target.
     In April 2002, Exelon withdrew from the PBMR venture , although it agreed to fulfil its commitment to fund the venture until completion of the feasibility study phase, then expected to be finished in September 2002. Forecast start of construction of the Demonstration Plant had by then slipped to 2004.
By May 2002, Nicholls  was much less precise in his estimate of the cost of the Demonstration Plant, estimating a cost of between US$2000-5000/kW. At the bottom end of the range, assuming a unit size of 110MW and US$2000/kW and an exchange rate of US$1=R6, this would translate into a total cost of R1.3bn, while at the upper end, with 130MW and US$5000/kW, it would translate into R4bn. It is not clear whether these estimates included the cost of a fuel production facility. Nicholls still adhered to the US$1000/kW estimate for commercial orders provided these were built in groups of 8-10 per site and only after 20 units had been sold.
     By December 2002, the target output of commercial units had increased to 165MW, 50 per cent higher than originally planned. Nicholls  admitted that the US$1000/kW would not be achieved until 32 units had been sold. Further delays were announced in the programme. Earlier in 2002, the shareholders of PBMR (Pty) Ltd had expected to announce whether they would proceed beyond the feasibility stage by the end of 2002. This decision was postponed into an unspecified date in 2003 and appeared still not to have been taken in December 2004. In July 2003, the Demonstration Plant was expected to be 125MW with subsequent units producing 165MW.
     A particular issue was the supplier of the gas turbine. This would be the first-of-a-kind and would be the first commercial gas turbine to use helium gas as the energy carrier (normally gas turbines are driven by the exhaust gas from the combustion of the oil or gas fuel) and represents a significant engineering challenge. The contract to design the turbine was originally placed with the French company, Alstom but they were replaced in 2001 by Mitsubishi for unspecified reasons. It is not clear how far development problems with the gas turbine have delayed the programme and have increased costs. In November 2004, PBMR (Pty) Ltd announced a major design change in the gas turbine moving to a horizontal turbine generator set rather than the three-shaft vertical configuration that had been planned.  It should also be noted that the frequency of the North American electrical system is 60Hz, compared to 50Hz in Europe and South Africa. China is 50Hz, but Japan is part 50Hz and part 60Hz. This means the speed of rotation of the gas turbine is different and generally gas turbines that produce power at 60Hz are of a significantly different design to those that produce power at 50Hz. It is not clear who would pay the cost of development of 60Hz machines for exports to the USA.
    The main extra cost for the demonstration programme apart from the generating plant itself was the fuel manufacture plant expected to be built at Pelindaba. In 1999, Nicholls estimated this would cost about US$100m (R600m) but more recent forecasts for the demonstration programme have not separated the fuel plant from the reactor, so it is impossible to determine how far escalation in the cost of the demonstration programme has been the result of increases in the cost of the fuel plant.
     Once the end of the feasibility phase had been reached, the partners’ commitment to fund the venture came to an end and essentially PBMR (Pty) Ltd had no further guaranteed access to funding. It was planned that in the demonstration phase, PBMR (Pty) Ltd would be reconstituted and the previous partners would have the right to take up a shareholding in proportion to the funding they had provided for the feasibility phase. It is not clear how PBMR (Pty) Ltd has been funded since the end of the feasibility phase. It appears most likely that a combination of government and Eskom money has allowed PBMR (Pty) Ltd to continue operations, albeit on a severely reduced scale.
     By August 2003, PBMR (Pty) Ltd was seriously short of cash and was appealing to the South African government for support.  A review of the project was begun by the government in January 2004 and it gave PBMR (Pty) Ltd ‘two months to propose a way forward for the PBMR.’  The Demonstration Plant was then projected to cost US$1.3bn (R8bn) and it was still hoped to begin site work at the Demonstration Plant in 2004. In March 2004, Terblanche estimated the cost of the Demonstration Plant would be R10bn and it could not be in full operation before 2010, implying a 2007 construction start and the launching of commercial sales after 2012.  Ferreira  broadly confirmed these figures in September 2004 and in November 2004, these remained the most recent published estimates.
However, a June 2005 press report appears to suggest that the cost of the demo phase may have increased again to R14-15bn.  If this increase of about 50 per cent in a little over a year is confirmed, this would add to the evidence that costs are seriously out of control. It is not clear whether the US$1000-1200/kW estimated cost for commercial units still stands.
     In the period 1999-2004, the estimated cost of the demonstration programme appears to have escalated by a factor of five and, since 2004, may have escalated by a further 50 per cent. Until the detailed design is completed: equipment design development, for example on the turbine, has been carried out; design approval by the National Nuclear Regulator (NNR) is given; and the plant has actually been built, the cost estimates must be treated with scepticism. Experience with other nuclear projects shows these processes provide ample scope for further major cost escalation.
     A particular regulatory issue is that of containment/confinement to the reactor. The containment serves to prevent the contents of the reactor escaping into the environment if there is an accident in the reactor or if there is an external accident, for example, an aircraft hitting the plant. The arguments are complex, but, in essence, it is argued (PBMR, 2002b, p 29) that a pressure producing accident is implausible so an expensive pressure-retaining containment would not be necessary. PBMR (Pty) Ltd argues that a containment that need only withstand, for example, aircraft impact would be much cheaper.
     In September 2003, a spokesman for the NNR said ‘''At this stage, we don't have the answer'' about whether a pressure-resistant containment is required, the NNR executive said. ''It's a long shot to say the regulator has accepted'' that confinement suffices.’  However, PBMR (Pty) Ltd (for Eskom) not only has to convince the South African NNC, it also has to convince a high credibility international regulator, most likely the US Nuclear Regulatory Commission (NRC). It would make no economic sense nor would it be politically acceptable for PBMR (Pty) Ltd to design one model for South African use and another (apparently safer) for international orders. So until this issue is resolved, there must be a significant risk that construction cost estimates will increase. The issue of containment is by no means the only significant licensing issue still to be resolved.

Required information

An up-to-date estimate of the cost of the Demonstration Plant is required, broken down into the cost of the plant itself, the fuel supply plant and any other significant facilities. An analysis of the cause of the delays to the programme and of the factors behind the massive cost escalation that has occurred is required. An analysis of the remaining risks of cost escalation, for example from design changes, unexpected equipment development problems, should also be provided.


6.4 The cost of capital

     While the construction cost of the plant has been of continual concern, there has been little debate about the cost of capital. Traditionally, the cost of capital for power plants was very low, typically a real annual rate of 5-8 per cent. This low cost of capital reflected the fact that, as monopolies, electric utilities were generally able to pass on whatever costs they incurred to consumers, so there was very little risk that the loan would not be repaid. Of course, this did not make constructing new power plants a low economic risk, it simply meant that electricity consumers were bearing the risk rather than the company. Also government-owned utilities were regarded as being fully underwritten by government and the credit rating of government owned utilities was generally the same (very high) as that of the government itself and the cost of borrowing correspondingly low.
     In the past decade, with the opening up worldwide of the electricity industry to competition and the privatisation, at least in part, of many utilities, the position has changed dramatically. Many electric utilities, the potential customers for the PBMR, have been privatised and wholesale electricity markets introduced. This is planned to take place in South Africa with the splitting up of Eskom into regional distribution companies, a transmission company and a requirement to sell 30 per cent of its generation. This plan, notably the sell off of generation, appeared to be under review in October 2004 and it may be that Eskom will continue to be able to pass on the costs of its investments to consumers no matter how ill-conceived these decisions turn out to be.
However, in other markets, investment in generating plants is now a high risk to the owners of companies and the companies providing them with finance. The privatised utilities can no longer rely on government backing to support their credit rating. In Britain, the country that pioneered electricity privatisation and opening to competition of electric utilities, this risk is very real. In 2003, about 40 per cent of Britain’s generating capacity was owned by financially distressed companies.  Half of this capacity was the nuclear plants while the rest was a mixture of coal and gas-fired plants. At one point, the second largest owner of power plants in Britain was the consortium of banks that had lent money to investors and had repossessed the plants when they began to lose money.
     Even before this stark demonstration of the economic risk of owning power plants, the real annual cost of capital for new generation plants in Britain was in excess of 15 per cent compared to about 6-7 per cent for investment in the parts of the industry that remained a regulated monopoly (essentially the distribution and transmission networks). In developing countries where currencies are less stable, there would be an additional risk premium on capital and, for example, the real cost of capital in Brazil would be at least 20 per cent. Given that repaying the capital charges is the largest element of the cost of nuclear power, it is easy to see if this cost is increased by a factor of 2-3, the impact on the economics of nuclear is going to significant and probably disastrous.
     Nicholls (Nicholls, 2000) used a real cost of capital of 6 per cent and although this appears to have been increased to 8 per cent for subsequent analyses, this is far below the level that will be applied in many of the PBMR’s target markets.
     A decision to allow use of too low real cost of capital would have significant consequences, especially in a country like South Africa that has limited access to capital and very heavy demands for public spending in areas such as health and education where the returns on investment would be high and the risks low. Using capital on a low-return, high-risk project like the PBMR would risk crowding out more attractive and socially useful projects.
The issue of rate of return was raised by the Legal Resources Centre (Register of Comments (2002), 28.137), but the response suggests the person replying either did not understand the question or chose not to answer it: ‘The PBMR project has been thoroughly evaluated by the respective investors on a commercial basis.  Although their required Return on Investment (ROI) varies, normal commercial benchmarks were used in this evaluation process.’

Required information

The FEIR economic assessment  should specify and justify the cost of capital that will apply to the Demonstration Plant and the associated facilities

6.5 Maximum electrical output

     There has been considerable confusion about the output of the Demonstration Plant, which has been variously reported as 110MW, 125MW, 137MW and 165MW. The DFR (PBMR, 2002a, p 25), stated the Demonstration Plant would be 110MW but would be modified in service to produce 125MW. The extent of the modifications necessary was not specified. It was implied that the first 10 commercial units would produce 125MW, but later units would produce 137MW. The DFR spoke of a later move to a core producing a thermal output of 400MW core and improvements in the conversion efficiency so that this would generate 200MW of electricity. The design changes necessary to achieve the 137MW output were expected to be such that earlier units could not be retrofitted to produce this higher level of output. In September 2003, Nicholls  was quoted as saying the Demonstration Plant would produce 125MW, while a year later, Nucleonics Week  reported ‘the first unit would be limited to 110 MW’. In November 2004, Nucleonics Week  reported the thermal output of the plant would be 400MW, sufficient to generate 165MW. It reported: ‘Eskom will file for revision of the EIA to take account of the higher electrical capacity’ after final Record of Decision (ROD) was given.
     This confusion needs to be resolved to clarify exactly what the Demonstration Plant will prove. Up-rating the output of a plant by 50 per cent is clearly not a trivial step and the International Panel discussed in detail the implications of the increase from 110MW to 125MW. If the design of the Demonstration Plant is significantly different to that of the commercial units, there must be doubts about how far the Demonstration Plant will indeed be a useful demonstration of the technology. Alternatively, if the design is the same but only operating at two thirds of its capability, potential buyers may not be convinced that the Demonstration Plant does demonstrate the commercial technology.
     Clarification is also needed on how far regulatory approval for a 110MW unit would be transferable to a 165MW unit. In this context it should be noted that Westinghouse obtained regulatory for its new AP600 design in 1999 but this design proved not to be economic. Westinghouse up-rated the output by about 50 per cent to gain scale economies and had to begin again the process of gaining license approval in March 2002 for the replacement AP1000. Final approval by the US regulatory body, the NRC, is not expected before December 2005.
It is not clear how far the up-ratings to the PBMR are due to simple changes to optimise the output of the plant (for example, operating at a higher temperature) and how far it is due to attempts to use scale economies to compensate for failing economics. It should be noted however that the design taken on from HTR produced a thermal output  of 226MWth, this was up-rated to 265MWth, then 300MWth and now commercial plants are expected to produce more than 400MWth, an increase on the original design of nearly 80 per cent.

Required information

Clarification is required on the expected output of the Demonstration Plant, how the design will relate to that of any subsequent commercial units. In particular it should show extent to which the Demonstration Plant will ‘demonstrate’ the commercial technology and how far safety licensing for the Demonstration Plant will be applicable to the commercial units.

6.6 Operating performance

     For any technology with high up-front costs, operating reliability is essential for good economic performance. To illustrate this, let us assume that the load factor  of a nuclear plant is expected to be 90 per cent and at this level, fixed costs will represent two thirds of the overall cost of power per kWh. If load factor is actually 60 per cent, this alone will raise the overall kWh cost by a third. Extra repair and maintenance costs to reflect the issues that produced this poor performance will increase costs even more.
     Reliability of nuclear power plants worldwide has been extremely variable and has generally been well below the levels forecast. For example, the Dungeness B nuclear power plant in Britain, which was selected ahead of other options partly on the basis that it would have a high lifetime load factor of 85 per cent has, after 20 years of operation, a lifetime load factor of only 36 per cent. The two existing Koeberg PWR units, also after nearly 20 years of operation, have lifetime load factors of only about 65 per cent.
    Nicholls  forecast that the lifetime load factor of the PBMR would be 94 per cent. This is hard to justify on a number of grounds. First, it would make the PBMR more reliable than any operating reactor worldwide. In 2004, the best lifetime load factor for any nuclear plant was 93.5 per cent and only 6 out of more than 400 operating units had achieved a lifetime load factor over 90 per cent. Second, much is made by PBMR (Pty) Ltd and Eskom of PBMR’s ability to ‘load-follow’, in other words vary its output as demand changes (PBMR (Pty Ltd, 2002a, p II and PBMR (Pty) Ltd, 2002b, p 24). Clearly if the units are operating at below their design rating ‘load-following’ for any significant part of the year it will be impossible to achieve load factors as high as forecast and the economic performance will be similarly reduced. The ability to load-follow would be an optional feature that would also increase the construction cost.
    For the Demonstration Plant, it might be expected that reliability would be poorer than for commercial units partly because of the need to carry out testing and demonstration activities, and partly because the Demonstration Plant will inevitably throw up technical problems that will only become apparent when a real plant is actually operated, and these will require shutdown for repair. If operating performance is expected to be significantly poorer than for the commercial units, this will make the power from the Demonstration Plant very expensive because the fixed costs will be spread over fewer saleable units of electrical output

Operating performance

The forecast load factor for the Demonstration Plant should be specified and justified, and its impact on the cost of power identified.

6.7 Operations & maintenance cost

    There is a common perception that once a nuclear power plant is built, the electricity is essentially free. Nuclear plants are assumed to be largely automatic and fuel costs are assumed to be low. While fuel costs are generally low, operations & maintenance (O&M) costs can be high. For example, a number of US nuclear power plants were closed down in the 1990s because it was judged it would be cheaper to pay the cost of building and operating a new gas-fired plant than paying the cost of simply operating an existing nuclear plant. Since then extensive efforts have been made in the USA to reduce costs. The USA is the only country to publish properly accounted O&M costs. In 2003, the cheapest plant to operate generated at about US 1.2c/kWh (US cents) of which, about US 0.4c/kWh was fuel cost. The most expensive plant cost US 2.6c/kWh and the median was about US 1.65c/kWh.
     No estimates of the operating cost of the PBMR have been published but Nicholls (Nicholls, 2000) estimated fuel costs at 0.4c/kWh, comparable to US figures. Given that in the same paper he forecast that total generating cost would be US 1.43c/kWh including repayment of capital, it seems likely Nicholls assumes the non-fuel O&M costs will be negligible. Given the non-fuel O&M costs alone for US plants average about US 1.2c/kWh, this assumption seems highly optimistic and cannot be accepted without detailed justification.

Required information

The O&M costs for the Demonstration Plant should be specified and justified, broken down by fuel and non-fuel costs.

6.8 Decommissioning cost

    Decommissioning is an immensely complex area that cannot be fully covered here. If the South African government allows the PBMR project to proceed to the demonstration phase, it is important to note that this commits it not just to the cost of the facilities required, but also to pay for the decommissioning of the Demonstration Plant and other associated facilities such as the fuel manufacturing plant.
    Decommissioning has significant economic, ethical and social dimensions as well as technical aspects. It is assumed that the ‘polluter pays’ principle should apply to the funding of decommissioning and this means:
    
• There should be clear plans to return the site to ‘green-field’ status after plant closure and decommissioning, i.e., the land should be fit to be released for unrestricted use including food production;
• Those that consume the electricity from the plant should pay for its decommissioning. This is generally done
by creating a ‘segregated’ account  that accumulates funds provided by consumers throughout the life of the plant to pay for its ultimate decommissioning;
• Provision needs to be made for large scale waste disposal to deal with the large amounts of radioactive material generated during decommissioning;
• Clear plans need to be put in place to document the location of all radioactive materials and plant design so that those that decommission the plant have full knowledge of what they will encounter;
• Careful ongoing monitoring of skills availability needs to be carried out to ensure that vital skills remain available until decommissioning is complete. This is particularly important where a long delay between plant closure and completion of decommissioning is anticipated. For example, in the UK it is planned to delay completion of decommissioning until more than 100 years after plant closure.

     In one respect, the polluter pays principle cannot be followed. There is no way to prevent a future generation having to carry out the potentially dangerous task of decommissioning nuclear power plants.
Decommissioning is conventionally assumed to be carried out in three phases: removal of fuel; removal of uncontaminated or lightly contaminated structures; and removal of contaminated structures, essentially the reactor itself. From a purely economic viewpoint, the incentives are always to carry out stage one as quickly as possible. A plant with nuclear fuel in it must be fully staffed because of the risk of criticality and once the fuel has been removed, the staffing level can be significantly reduced saving the labour costs. The economic incentives are to assume as long a delay for stages 2 and 3 as possible. Any fund created to pay for decommissioning will have longer to earn interest, reducing the provisions consumers must make to achieve the required sum. In practice, social and technological factors may over-ride this incentive. For example, it may be politically unacceptable to leave a potentially hazardous facility in place for several decades simply to allow the fund to accumulate sufficient interest to pay for decommissioning.
    The DFR (PBMR, 2002a, p 27) anticipates two possible strategies, early plant dismantling or ‘safe enclosure’, in which stages 2 and 3 would be delayed. The DFR does not specify the length of the delay, but it should be noted that the THTR plant in Germany is expected to be in safe enclosure for at least 30 years. The DFR states that: ‘if the demonstration module is not successful, the plant will be mothballed in ‘safestore’ until the decommissioning of Koeberg I and II. However, negotiations with Eskom in this regard have not been finalized.’
Typically, it is assumed that the cost of decommissioning represents about a third of the construction cost. Since the decommissioning cost clearly has little direct relation to the construction cost, this indicates the immaturity of decommissioning technology and the only plants fully decommissioned worldwide are not representative. For example, they may have operated for only a short time and are little contaminated, or the plant may have been disposed of in a large hole without dismantling (Trojan, USA) or the plant is very small.
    The FEIR (PBMR (Pty) Ltd, 2002b, p 201) states that 1.5 per cent of the capital cost is provided for decommissioning. It is not clear what is meant by this. Subsequent clarification by consultants (Register of Comments, 2002, 28.149) has suggested that: ‘the PBMR Operator will provide 1.5 per cent of the capital cost of the plant on an annual basis over the useful life of the plant.’ And that the proposed minimum provision would be based on a 15 per cent of original yet escalated, construction costs, (sic) be made available for decommissioning at the economic end of the plant (Register of Comments, 2002, 28.149).
    This is still far from clear and the reliance on estimating the decommissioning as a percentage of the construction cost betrays the fact that little work has been done on estimating decommissioning costs. The FEIR does specify that a segregate (sic) fund will be set up.
    Experience with the plants of similar technology to the PBMR in Germany is particularly salutary. The 15MWth pilot AVR plant (it produced heat but no power) is of similar technology to the PBMR and operated from 1967-88 before engineering problems caused its closure. The estimated cost of decommissioning and dismantling the AVR escalated from about €20-million during the early 1990s to as much as €490-million in 2002 (about R7bn).  So even after closure of the plant, decommissioning costs were subject to huge price escalation and if any provisions had been collected, they would have proved totally inadequate, leaving later generations to meet the cost.
    The THTR 300 demonstration plant, also using pebble bed technology, was in service for only six years to 1989 but produced minimal amounts of power and is therefore likely to be lightly contaminated. It was de-fuelled only in 1995, placed in ‘safe enclosure’ in 1997 and it is not expected that decommissioning of the contaminated parts of the plant will start before about 2020. No recent cost estimates for decommissioning have been published. Again, if it had been assumed the plant would operate for, say 20 years and decommissioning provisions had been collected from electricity consumers on that assumption, any provisions would have been totally inadequate.
    For a demonstration plant, which inevitably has a very uncertain length of operating life, it would seem more prudent to include the necessary provisions in the initial cost to reduce the risk of a shortfall in decommissioning funds if the plant operates for a shorter period than expected.

Required information

The estimated decommissioning cost for the Demonstration Plant should be published broken down into the three main stages. The assumed timing of the three phases should also be specified and the arrangements for funding the process (how the money would be collected and kept, what rate of interest is assumed) given.

6.9 Operating life

    The expected operating life of the plant will determine how long the owner has to repay the construction costs. The longer the life, the lower the annual repayments are. In practice, expected operating life is not as important as might be expected. Generally, commercial loans do not have a repayment period longer than 20 years so this is the maximum ‘amortisation’ period for a commercial facility.
    Nicholls (Nicholls, 2000) projected a 40-year life for a commercial PBMR module. This would appear to be rather optimistic. No estimate has been given for the Demonstration Plant’s lifetime. Demonstration plants often have quite a short life because they tend to be expensive to operate and once they have demonstrated (or failed to demonstrate as in the case of THTR 300) the technology, they are retired to reduce the losses consumers must bear. This is of particular concern if the decommissioning provisions are collected over the forecast operating life of the plant and this forecast proves too long.

Required information

The FEIR economic assessment  should specify and justify the expected economic life (the time over which construction costs will be recovered and decommissioning provisions collected) of the Demonstration Plant

6.10 Who will pay the extra cost of power from the Demonstration Plant?

     It is clear that the overall cost of power from the Demonstration Plant will be much higher than the current average cost of power to South African consumers. Foreign partners would be unlikely to want to subsidise power to South African consumers. There are plans to break Eskom’s effective generation monopoly, selling off some of their plant to competing companies. In this situation, Eskom is unlikely to want to be saddled with an uneconomic generating plant when it cannot pass on the additional costs to consumers. The DFR (PBMR, 2002a, p 32) states: ‘Eskom will, upon successful commissioning, purchase it from PBMR (Pty) Ltd on normal commercial terms. The cost of development of the PBMR technology and generic design, as well as the up-front programme development costs, will therefore not be recovered through the sale of the demonstration plant, and will be capitalized in the books of the proposed PBMRCo enterprise for amortization over the 25-year review period’. It is not clear what ‘normal commercial terms’ means. It could mean: paying the forecast cost of a commercial PBMR plant; paying the equivalent cost of the cheapest generating option; paying the full cost of the Demonstration Plant, minus some development costs and the cost of the fuel plant.
     No mention is made of the operating costs. It could well be that with a relatively small fuel plant, operating unreliability and inexperience with operating PBMRs, the operating costs could be higher than those of, say, a coal plant. In this situation, Eskom would be left with a facility that would not be economic to operate even on a marginal cost basis and it would be left unused.
     In evidence to the South African Parliament's Minerals and Energy Affairs Portfolio Committee, the CEO of Eskom, Thulani Gcabashe, only committed that Eskom would ‘host’ the demonstration unit.  It remains to be seen whether government is willing to provide subsidies or whether it will try to force Eskom to pass the extra costs on consumers.

Required information

The FEIR economic assessment  should indicate precisely what Eskom will be expected to pay for the Demonstration Plant, how much the additional cost of power from the Demonstration Plant over and above the cost that would have been incurred if the power had been generated by commercial plants will be and who will pay these additional costs.

6.11 Analysis of risk

     The PBMR project has always been a high-risk project. Thomas (Thomas, 1999) writing in 1999 said:
‘The development of the PBMR by Eskom would represent a highly risky venture which would be underwritten by tax-payers and electricity consumers.’
     These risks have been amply demonstrated over the following six years. The cost of the Demonstration Plant has increased by a factor of five and completion of the Demonstration Plant, expected in 1999 to be in 2003, is now still six years off. If the risks had, by now, all been incurred, this poor history of technology development would be of limited relevance to the decision whether to go ahead with the Demonstration Plant. In economists’ jargon, ‘bygones are bygones’. In other words, the development costs have been incurred and cannot now be ‘unspent’: what matters for decisions being taken now are the remaining costs and risks. Of course the failure to control costs and the huge slippage in the time-table must be taken into account in judging the competence of the developers, PBMR (Pty) Ltd and the likelihood that the remainder of the programme can be completed to time and cost.
     The previous analysis has shown that there are still many risks. The design is far from complete, for example, a major change to the turbine generator design was announced in October 2004, the design has not received South African NNR approval, nor has substantive progress been made with approval by the US NRC. Even when these processes are complete, the history of nuclear power amply demonstrates the large risk of cost escalation during the construction phase. So the risk that costs will escalate even further is high. The statement in the Register of Comments (Register of Comments, 2002, 28.144) that ‘the PBMR detailed design has been finalised.’ cannot be justified. Since then, the turbine generator design has been changed, the plant output upgraded, apparently requiring significant design changes and until NNR approval is given, clarifying, for example, whether a pressure containment is needed, the design cannot be regarded as finalised. The problems in completing the design also do not provide confidence in the abilities of PBMR (Pty) Ltd nor do they augur well for the technological success of the Demonstration Plant.
     Attempts to reduce the risk to the South African public have had some success, with about a third of the development cost in the feasibility phase being met by foreign companies, notably Exelon, but also BNFL. However, for the much more expensive (at least five-fold) demonstration phase, Exelon will not participate and BNFL seems unlikely to be in a position to make a substantive contribution. Attempts to bring in other foreign investors, such as US utilities, the French company Areva and Chinese interests have not yet succeeded and it now appears likely that if the Demonstration Plant is to go ahead, it will be largely underwritten by South African public money through the government, Eskom, or IDC. This will include not only the estimate of at least R10bn to build the plant and associated facilities, it will also include the cost of decommissioning the plant and the extra cost of buying the electrical output over and above the cost of generating in commercial power stations.
     The FEIR was seriously inaccurate even before it was published. It acknowledged the withdrawal of Exelon but the sales projections were still heavily dependent on Exelon. Exelon would buy the first commercial unit, before Eskom, and in the crucial first five years of the commercial phase when the business has to establish itself, it assumed Exelon would buy half the units sold. In the three years since the FEIR was published, the date when the first commercial units are expected to be sold has slipped by eight years and no replacement for Exelon has been found. Inevitably, the pressure is on Eskom, underwritten by South African taxpayers and electricity consumers, to step in to fill the gap.

6.12 The cost of a catastrophic accident

This report does not examine the costs that would arise if the Demonstration Plant were to cause a catastrophic accident. However, it should be noted that the 1986 Chernobyl accident in Ukraine is expected to result in costs of US$235bn in the 30 years after the accident. . It is therefore essential that the promoter’s claims that such an accident is totally impossible should be evaluated fully, and if the probability is not zero, consideration needs to be given on how such astronomic costs could be met.

6.13 The cost of waste and spent fuel disposal

     This report does not examine the cost of waste and spent fuel disposal. However, a number of points should be made.
     First, worldwide, no spent fuel has been disposed of yet. All fuel used to date remains in temporary surface stores or has been reprocessed to produce plutonium. Note that reprocessing does not reduce the amount of waste to be disposed of,  it merely splits it up into different ‘packages’. Until facilities have been designed and built that give the public full confidence that spent fuel can be disposed of in such a way that there is no risk that this material will be exposed to the human environment over the millions of years that it will take for the material to become harmless, the costs must be regarded as speculative.
     Second, worldwide, very few waste disposal facilities for low-level and intermediate-level waste have been built in recent years and the waste that is being disposed of is mainly going to old sites designed fifty or more years ago. Until there is more evidence of the cost of designing, building and operating waste disposal facilities that meet current safety standards and are publicly acceptable, the cost of waste disposal must also be regarded as uncertain.
     Third, as with decommissioning, the cost of waste and spent fuel disposal will be incurred decades after the waste is created. If funds are put aside at the time the waste is created, these funds can be invested and can be expected to grow substantially. For example, a fund that is invested for 40 years, earning an annual real interest rate of 2.5 per cent will grow by a factor of 2.7. However, this does point to the need to establish clear procedures to take money from consumers to pay for these activities and to keep it in secure investments so the risk that it is lost is minimised.

7. The commercial programme

     Construction of the Demonstration Plant only makes sense if there is a high probability that it will lead to a profitable (to South African interests) stream of orders for commercial PBMRs. It is therefore essential to examine the prospects for such sales if the economic case for the Demonstration Plant is to be properly assessed.

7.1 The economic competitiveness of the PBMR
The economic competitiveness was assessed in detail by the International Panel of experts in 2002 and their report would provide a proper basis to analyse the economic prospects for the PBMR programme. The estimates given by Nicholls in 2000 (Nicholls, 2000) are clearly out of date. The information required for commercial units is:
 
    • Construction cost;
    • The cost of capital;
    • The plant’s maximum electrical output;
    • Operating performance especially reliability;
    • Operations & maintenance (O&M) cost, including fuel supply and spent fuel disposal;
    • Decommissioning cost and;
    • Operating life.

    In some cases, for example, maximum electrical output, the information will comparable for all markets, but in others it might vary. For example: PBMR (Pty) Ltd might sell units to Eskom at a discount to the cost other customers; construction cost will vary depending on how many units are being built on the site; the cost of capital will vary from country to country according to the commercial position of the customer and the economic conditions in the export country; operating performance will vary according to whether the plant is expected to be base-load or load-following; decommissioning cost will vary according to the cost of waste disposal in the country of installation.
    A key assumption will be the construction cost. Let us assume the Demonstration Plant alone (not including the fuel plant) will cost about US$1bn (two thirds of the R10bn that the demonstration programme was estimated to cost in 2004) or about US$9000/kW if the plant produces 110MW, the gap to commercial units costing US$1000-1200 is huge. If the design can be stretched to produce 165MW at no extra cost, the cost per kW would be about US$6000/kW. This still leaves a huge reduction in costs to get down to the target levels. Some of this will come from not having to incur the technology start-up costs the Demonstration Plant would require. The rest must come from various scale economies and learning effects. These include: building ten units on a site; scale economies in manufacturing if a minimum number of units are sold. The DFR did not publish any details of these scale economies claiming the information was commercially confidential (PBMR (Pty) Ltd, 2002a, p 56)

Required information

    The government should publish the report by the international Panel of Experts. Eskom should publish the latest cost and performance estimates for the commercial plants as well as the assumptions on factors such as cost of capital by market. It should also specify how the unit cost is expected to be reduced by a factor of at least five from the Demonstration Plant to a fully commercial unit.