به حداقل رساندن هزینه های اقتصادی و ریسک برای تکنولوژی رآکتور زیر مرحله بحرانی مبتنی بر شتاب دهنده :مورد طراحی برای انعطاف پذیری: قسمت 1

Demonstrating the generation of electricity for commercial markets with accelerator-driven subcritical reactor (ADSR) technology will incur substantial financial risk. This risk will arise from traditional uncertainties associated with the construction of nuclear power stations and also from new technology uncertainties such as the reliability of the required accelerator system. The sensitivity of the economic value of ADSRs to the reliability of the accelerator system is assessed. Using linear accelerators as an example of choice for the accelerator technology, the economic assessment considers an ADSR with either one or two accelerators driving it. The extent to which a second accelerator improves the accelerator system reliability is determined, as are the costs for that reliability improvement. Two flexible designs for the accelerator system are also considered, derived from the real options analysis technique. One seeks to achieve the benefits of both the single and dual accelerator ADSR configurations through initially planning to build a second accelerator, but only actually constructing it once it is determined to be economically beneficial to do so. The other builds and tests an accelerator before committing to constructing a reactor. Finally, a phased multiple-reactor park with an integrated system of accelerators is suggested and discussed. The park uses the principles of redundancy as for the Dual accelerator ADSR and flexibility as for the real options design, but for a lower cost per unit of electricity produced.

In an attempt to meet energy needs in a responsible and sustainable way, a revolutionary nuclear reactor concept is having its engineering feasibility re-assessed. The design in question is the accelerator-driven subcritical reactor (ADSR), the concept for which dates back to the 1990s (Bowman et al., 1992 and Carminati et al., 1993). If hopes for ADSRs are fulfilled then they will provide the world with electricity while: emitting minimal amounts of CO2; ensuring a high level of safety during operation due to the use of an accelerator and a subcritical reactor core; achieving a significant reduction in backend radioactive waste compared to contemporary reactors – they may even consume waste from other reactors; and extending the consumption time of the world's uranium and thorium resources by multiple orders of magnitude. Inevitably a power station that promises so many benefits is not without its challenges. Multiple aspects of the engineering requirements of the design are the subject of challenging Research and Development (R&D) programmes (ENEA, 2001); chief areas of concern are the reliability of the accelerator system, the reliability of the beam target (the interface between the accelerator and reactor core) and long-term corrosion of the steel structure due to the presence of heavy liquid metal. A poor outcome from this R&D would be the finding that the design requirements of ADSRs are so extreme that they are untenably expensive. In the commercial electricity market, all the nuclear power stations ever constructed have self-sustained fission reactions during operation – they are all critical reactors. When a critical reactor is operating, electricity is generated. For the ADSR (a subcritical reactor) only when the nuclear core and its accelerator system are operating is energy generation sustained and electricity produced. To date no attempt has been made to couple together an accelerator, beam target and nuclear reactor as a single system to produce a sustained nuclear chain reaction for greater than a nominal power output. A proposal for doing this at the Belgian nuclear research facility, StudeCentrum voor Kernenerge Centre d’etude de l’Energie Nucleaire (SCK·CEN), has recently received support from the Belgian government (SCK·CEN, 2010). The study is intended to be complete by the year 2024. The financing of any nuclear power station is dominated by capital costs. There is therefore a significant financial risk associated with the construction of a nuclear power station. The risk is particularly large when demonstrating the first-of-a-kind of a technology; this issue is exemplified by the escalating costs and delays currently being experienced at the Finnish Olkiluoto facility (WNA, 2010), which is constructing the world's first European pressurised water reactor (EPR). It now appears as if the Finnish EPR will be a loss leader (Harding, 2007). It is not unheard of for first-of-a-kind nuclear reactors to be loss leaders; there have even been instances in the past where vendors planned from the outset to make their new design as such (Kaijser, 1992). In addition to typical economic construction risks, ADSRs add unique new risks. These are due to the required accelerated proton beam and the beam target. Only the accelerator, and not the beam target, is the subject of the presented work. Contemporary accelerator systems are less powerful and less reliable than the specifications quoted for ADSRs (Burgazzi and Pierini, 2007 and ENEA, 2001). Accelerator-specific R&D is being carried out to bridge this technological gap (Burgazzi and Pierini, 2007, Pierini et al., 2003 and Teng, 2001). Even if R&D predictions are optimistic enough such that ADSRs do appear worth pursuing as a commercial proposition, there will still be risks associated with whether accelerator performance will meet the predictions. With similarity to how unanticipated problems in the first-of-a-kind EPR have led to delays in its construction, an unexpectedly high rate of unplanned shutdowns of the first-of-a-kind ADSR accelerator system will affect its performance throughout its operational lifetime. If the reliability of a realised ADSR accelerator is poor then either the revenue of the ADSR will be low or the cost of failing to fulfil electricity contracts will be high. Regardless, the ADSR will return less marginal profit to offset the capital expenditure. This is not desirable for nuclear power stations as they typically operate as base-load electricity generators with low marginal costs of generation (Pouret et al., 2009). In this paper an economic analysis of the benefits and costs associated with designing increased multiplicity for ADSR accelerators is deliberated. The aim is twofold. The first aim is to scrutinise formally an assumption that to the authors’ knowledge has yet to be addressed in peer reviewed literature. The assumption is that designing an ADSR to have multiple LINear ACcelerators (LINACs) will untenably raise the cost of the ADSR. The analysis is mindful of, and therefore lends itself to, the possibility that types of accelerator other than LINACs might be the preferred choice for an ADSR; the cost of other accelerator types may be significantly less than LINACs and therefore the construction of multiple devices more reasonable. The second aim is to recognise that, given the large capital that is at risk, a second accelerator will significantly reduce investment uncertainty, even though it will increase the cost of constructing the ADSR. This second aim is considered to be of particular interest for the first-of-a-kind ADSR. This is because, following R&D, this will be the time when there is greatest uncertainty regarding accelerator reliability. Treating the reactor vending and operating firms as a single company, it may be that a vendor–operator's strategy is to demonstrate the technology with a less risky ADSR driven by two accelerators. The long-term aim being that the nth-of-a-kind ADSR will be driven only by a single accelerator, should the technology prove to be successful. The paper is structured as follows. In Section 2 there is a review of the demands on an accelerator used to drive a nuclear reactor. The performance achieved by contemporary high-power accelerators is detailed along with expectations of future performance from accelerator R&D literature. In Section 3 an ADSR designed with two accelerators is described whose primary aim is to reduce the reliability demands on the individual accelerators. A 4-step real options design framework is then used along with an economic model to assess the expected value of ADSRs designed with either one or two accelerators. In particular, the real options framework enables the recognition of an accelerator system design that is a balance between the one and two accelerator designs; this and a second flexible design that builds and tests an accelerator before constructing a reactor are discussed and also assessed in the economic model. In Section 4 a qualitative discussion is given, which highlights for all of the designs recognised pros and cons not captured by the presented economic analysis. At the end of the discussion a speculative design idea for a phased and integrated “park” of multiple reactors is suggested. The reactor park is motivated by the economic concerns of minimising both the levelised cost of electricity and the capital at risk. In Section 5 conclusions from the investigation are given. Appendix A explains possible methods by which dual accelerators might best be operated and Appendix B indicates the coinciding unplanned shutdown frequency of an accelerator network.

If an ADSR operator were to sell electricity into the contemporary UK electricity market, it would only return a marginal profit if it successfully delivered electricity for more than 3 days in every 10 of its contracts. By assuming that all other systems are 100% reliable and that an accelerator system failure will cause a shutdown 24 h in duration, it is established that it is not worth scheduling to sell any ADSR-produced electricity unless the accelerator fails less than ∼200 times per year. Assuming that the relative frequency of failure durations of future accelerator systems express similar behaviour to that of contemporary ones, it has been demonstrated that an ADSR driven by two accelerators will shut down ∼200 times per year with accelerators that individually trip many hundreds to low thousands of times per year. The feasibility of the Dual accelerator design hinges on there being a system in place that enables one accelerator to compensate for the other within a second of it experiencing a fault. A top-down financial cost-benefit-analysis has been performed for ADSRs. The financial model has been used to examine by how much a second accelerator escalates the cost of generating electricity with an ADSR. Without commenting on the future price of electricity it has been shown that if accelerator performance is poor, Dual accelerator ADSRs do become more valuable than Single accelerator designs. It has also been shown that Dual accelerator ADSRs eliminate nearly all of the investment uncertainty when considering only risks associated with accelerator performance. The down-side of this is an increase in CapExBFR of ∼17.5%. Once other risk factors are taken into account, such as construction delays, the increase in initial capital at risk might be considered more important than increased reliability during operation. The sensitivity of the frequency of accelerator trips to the duration of a trip in turn makes the economic case for ADSRs highly sensitive to the duration that the reactor core can tolerate an interruption in the supply of protons. In relation to the presented analysis: if reactors can tolerate accelerator failures for either greater than, or less than, the 1 s duration used as a reference in this paper, the likelihood of accelerator trips leading to reactor shutdowns will fall, or rise, respectively. Importantly the −2/3 power law that governs the frequency of different beam interruption durations in existing accelerator systems makes the economic sensitivity asymmetric. The value lost if only trip durations briefer than 1 s can be tolerated is greater in magnitude than the value gained if equivalently longer trip durations can be tolerated. The duration of beam failure that a reactor can sustain without shutting down does not significantly affect the total number of trips it can tolerate – instead it affects the likelihood that unplanned reactor shutdowns will occur, i.e. in the context of the presented analysis, it significantly affects the decision tree parameters: ρO, ρC and ρP, but only negligibly affects the value returned by the scenarios: LCOEO, LCOEC and LCOEP. Following a failure, the time it takes to restart an accelerator or the time it takes to divert the beam from a redundant accelerator compared with the duration an ADSR core can tolerate beam interruptions is of crucial importance to its performance. Real options analysis has identified two more first-of-a-kind ADSR designs. One is an expandable design, where initially a single accelerator is built and a second is planned for should it be required. The cost of planning to later build a second accelerator is expected to negligibly increase the LCOE if it is not constructed. The Expandable design does not increase significantly the CapExBFR. This design provides the option to improve reliability, should that be required. The most significant trade-off in the presented economic analysis is that reliability can only be improved after 5 years of reactor operation, which significantly increases the LCOE compared to the Dual design in outcomes where accelerator performance is poor. Qualitatively the issue has also been raised that building the accelerators at different points in time, rather than in parallel, may result in missing out on potentially large economies of scale savings. The second flexible design is the Accelerator Test design. In this case the capital required to test accelerator performance is minimised through only constructing an accelerator at first and then building the reactor (and possibly a second accelerator) later if it is determined to be beneficial to do so. For this design the CapExBFR and the LCOE are worse than for all of the other designs in nearly all circumstances. The very high tolerance of poor accelerator reliability of the Dual and Expandable designs means that in effect they can also test accelerator performance for a similar investment cost as for the Accelerator Test design. This is because they (nearly) guarantee successful electricity sales after construction. In addition to testing the accelerator system these designs will also test all other facets of the design of an ADSR while the Accelerator Test design will not. The concept of a reactor park has been suggested as a way of improving the economic case for generating electricity from an ADSR. The reactor park is expected to cope with accelerator trips in a more financially efficient way than is possible for a single reactor. Greater flexibility in the construction of accelerators and reactors is anticipated. In a follow-on from this study the real options technique and the decision analysis methodology have been applied with greater scope to the design suggestion of an ADSR reactor park (Cardin et al., 2012).