A principal conclusion of this study is that SMRs have a significant potential to expand the peaceful applications of nuclear power by catering to the energy needs of those market segments that cannot be served by conventional NPPs with large reactors. Such segments could be:

• Niche applications in remote or isolated areas where large generating capacities are not needed, the electrical grids are poorly developed or absent, and where the non-electrical products (such as heat or desalinated water) are as important as the electricity;

• Replacement for the decommissioned small and medium-sized fossil fuel plants, as well as an alternative to newly planned such plants, in the cases when certain siting restrictions exist, such as limited free capacity of the grid, limited spinning reserve, and/or limited supply of water for cooling towers of a power plant;

• Replacement for those decommissioned fossil-fuelled combined heat and power plants, where the SMR power range seems to better fit the requirements of the currently existing heat distribution infrastructure;

• Power plants in liberalised energy markets or those owned by private investors or utilities for whom small upfront capital investments, short on-site construction time (with the accordingly reduced cost of financing), and flexibility in plant configuration and applications matter more than the levelised unit electricity cost.

It should be noted, however, that none of the smaller reactors has yet been licensed for these applications and there remain both development challenges to overcome and regulatory approvals to obtain before deployment, especially in light of the recent accident at Fukushima.

The present study has found no situations where NPPs with SMRs could compete with the NPPs with state-of-the-art large reactors, on LUEC basis. However, it also found that SMRs could be competitive with many non-nuclear technologies in the cases when NPPs with large reactors are, for whatever reason, unable to compete.

[1] Levelised unit cost of electricity (LUEC) is calculated using the discounted cash flow method over the whole lifetime of the plant, and includes the initial investment, operations and maintenance, cost of fuel, financing and decommissioning costs. LUEC is measured in the units of currency per units of energy (e. g. in USD per MWh).

[2] Design simplification. In some advanced SMRs, significant design simplifications could be achieved through broader incorporation of size-specific inherent safety features that would not be possible for large reactors. If such simplifications are achieved, this would make a positive contribution to the competitiveness of SMRs. The vendors estimate that design simplification could reduce capital costs for near-term pressurised water SMRs by at least 15%.

[3] In order to arrive at a heat credit per MWh of electricity, one needs to establish the total value of the heat produced over the lifetime of the plant by multiplying total heat output by its per unit value. The total value of the heat output is then divided by the lifetime electricity production to obtain the per MWh heat credit. For plants operated in a co-generation mode, a heat credit is then subtracted from total unit costs to establish an equivalent of the levelised costs of producing only electricity.

16

[4] The parameter n of the scaling law is not known precisely. Based on reported values and analysis, an interval of n=0.4-0.6 has been considered in the calculations. For example, if the size of the unit is decreased by a factor of 2, the capital cost would decrease only by 25-35%. That would result in an increase of the specific capital cost by a factor 1.3-1.5.

[5] In Brazil, more that 70% of electricity is generated from the hydroelectric power plants offering very low cost electricity. Other sources of electricity, including nuclear power plants (with large reactors or SMRs), have higher electricity generation costs.

[6] *90 MWe (Korea) W 2*35 MWe, barge (Russia)

2*8 MWe, barge (Russia)

[7] According to the vendors and designers, all the advanced SMRs listed in Table 1 have been designed or are being designed in compliance with their current national regulations.

[8] Replacement for the decommissioned small and medium-sized fossil fuel plants, as well as an alternative to newly planned such plants, in the cases when certain siting restrictions exist, such as limited free capacity of the grid, limited spinning reserve, and/or limited supply of water for cooling towers of a power plant;

[9] SMRs constitute an important share of the actual nuclear fleet: 136 of the 441 reactors in operation, mostly those of older design, have a power falling in the SMR range [1.1]. In 2010, nine nuclear power plants under construction were SMRs [1.1].

[10] It is important to underline that co-generation is not unique to SMRs. However, as will be discussed later, the SMR power range corresponds well to the infrastructure requirements for non-electrical products (e. g. district heating).

[11] Taking into account non-electrical applications.

[12] The detailed design specifications for SMRs shown in Table 3.1 are given in Table A1.1 of Appendix 1.

[13] The detailed design specifications for SMRs in Table 4.1 are given in Tables A1.2 (a) and (b) of Appendix 1.

[14] Safety implications of the features of the above mentioned design groups are discussed in Section 8.2.

[15] Late in 2010 the Westinghouse Electric Company stopped the development of the IRIS project and announced it would go with an alternative integral design PWR of a 200 MWe class. Very few technical details of this new SMR were available as of June 2011.

[16] IMR is the only PWR design in which coolant boiling is allowed in upper part of the core [4.1]. Boiling boosts natural convection and makes it possible to use natural circulation of the primary coolant in normal operation, at a relatively high power level of 1 000 MWth (350 MWe).

[17] The projected construction period for advanced SMRs (typically two to five years for PWR SMRs) is typically shorter than the time needed to build SMRs available today.

[18] See the last row in Tables A1.2(a) and (b) of Appendix 1.

41

[19] The detailed design specifications for BWR SMRs are given in Table A1.3 of Appendix 1.

[20] Safety implications of BWR SMRs are further discussed in section 8.3.

[21] Recent deployments of the ABWR in Japan were accomplished with a three-year construction period [4.4].

[22] Including one atypical, dated-design pressure vessel type HWR in Argentina.

[23] The design specifications for the AHWR are given in Table A1.4 of Appendix 1.

[24] An option to increase AHWR unit power up to 500 MWe is being discussed.

[25] The design specifications for these reactors are provided in Table A1.5 of Appendix 1.

[26] All HTGRs provide for concentrated deployment with multi-module plants, although distributed deployment is not excluded for the ‘pin-in-block’ GTHTR300 and GT-MHR.

• The operating helium pressure is between 7 and 9 MPa, with 7 MPa being the preference.

• The average fuel burn-up is between 80 and 120 MWday/kg, being the maximum for the ‘pin-in-block’ designs.

[27] The detailed design specifications of the 4S are provided in Table A1.6 of Appendix 1.

[28] Seven Alfa class submarines (powered with 155 MWth lead-bismuth cooled reactors BM-40A) were in service from 1972 till 1990.

[29] The technology was developed for non-fast spectrum lead-bismuth cooled reactor cores. Applicability of this technology to fast spectrum cores may need additional validations.

[30] The corresponding detailed design specifications are provided in Table A1.7 of Appendix 1.

[31] For the New Hyperion Power module the date of a formal licensing application is still not defined, while licensing pre-application is already in progress.

[32] http://www. powergenworldwide. com/index/display/articledisplay/3386201290/articles/cogeneration-and-on- site-power-production/volume-n/issue-3/features/carbon-free-nuclear. html

[33] http://www. vnipiep. ru/dalnee_teplosnabzhenie. html

[34] http://bilnpp. rosenergoatom. ru/eng/about/info/

[35] For example, the EC6 (see Section 3) has a proven capability of daily load cycles from 100% to 60% of rated power, and can continue operation with loss of line to grid [4.41]. Some of the currently operated large French reactors (e. g. REP-1300 MWe, N4 design of 1 450 MWe) use daily cycles from 100% to 40% of rated power. Moreover, both the EPRI Utility Requirements Document and the European Utility Requirements (EUR) stipulate that the reactors should be capable of daily load cycles from 100% to 50% (or even 20%) of rated power [4.43].

[36] The reactor power level in the 4S is changed only by adjusting the steam flow rate in the power circuit. In this, the electric power dispatched to the grid and the steam flow rate in the power circuit are controlled in an active mode.

[37] From the condition that an unplanned NPP shutdown does not disrupt the stable grid operation.

[38] Simplified decommissioning limited to disconnection and removal of the transportable modules.

[39] However, three attributes that distinguish most of the new US small and modular reactors from other small reactor concepts developed elsewhere in the world, are namely:

— multi-module plant option;

— option of flexible capacity addition/removal; and

— underground reactor modules.

[40] Some recently announced concepts of small fast reactors for waste incineration, such as the fast-spectrum gas cooled EM2 of 240 MWe (being proposed by the General Atomics in the United States [5.7]), abandon fuel reprocessing and suggest that the spent fuel could be stored within the disconnected reactor module on the site or in a repository.

[41] In [6.1], a term LCOE — levelised cost of electricity is used. We use another abbreviation — LUEC — that is equivalent to LCOE, to be consistent with the literature on SMRs.

[42] However, it cannot be guaranteed that the interest rate for the LUEC calculations used by the vendors were all the same.

[43] The VK-300 designers’ data was not included as abnormally low. ** CCGT — Combined cycle gas turbine; SC/USC Coal —

Supercritical/Ultra-supercritical coal-fired plants, PV — Photovoltaic.

There are some important caveats regarding the use of the scaling law (6.3):

• The scaling law is only true if no significant design changes take place on transition to a larger or smaller capacity plant. If such changes take place (for example, the complexity of the plant design is reduced or increased), this results in a transition to another scaling law curve which may be located below or above the original one [6.5].

• According to reference [6.4], “the economy of scale may be limited due to the physical limitation to increase dimensions of some systems or components (e. g. reactor core, fuel rods and turbine blades). …The maximum size of units in an electrical grid is limited in consideration of grid stability, demand pattern, spinning reserve or other specific characteristics of the system”.

• One should keep in mind that an overall power scaling law for the entire plant is only approximate, because different components may have very different scaling exponents (for example, see Table 6.6), and thus the cost as a function of the plant unit power P is actually a polynomial of P, and it is approximated in (6.3) by a monomial.

The value of the scaling factor n is not fixed, and can be quite different for different NPPs:

[46] For example, for Korean NPPs of generally similar design OPR-1000 and APR-1400 this factor is 0.45, see Table 6.5.

• Another study, based on a French experience, gives a more detailed evaluation of the scaling factors, shown in Table 6.6. According to Table 6.6, the scaling parameter n is close to 0.6 for direct costs and is about 0.3 for the indirect costs including contingencies and owner’s costs. Also, it could be noted that the value of n increases with the increase in plant capacity,

[47] It is noted that the contributors to this report were unable to find any reference with the example of a NPP scaling law with n = 0.7 (the upper range suggested in [6.4]).

[48] Characterized by the in-vessel location of the steam generators and by the absence of large diameter piping.

75

The coefficients x, y and z correspond to the “programme” effect, and the coefficient k is related to the “productivity” effect described below. The main assumptions of the algorithm are as follows:

• The first unit built bears all of the extra FOAK cost (expressed as a factor [ 1+x]).

• The cost of engineering specific to each site is assumed to be identical for each site.

• The cost of facilities specific to each site is assumed to be identical for each site.

• The standard cost (excluding extra FOAK cost) of a unit includes the specific engineering and specific facilities for each unit.

Programme effect (construction of several units on the same site):

If T0 is the standard cost (excluding extra FOAK cost) of the sole unit on a site (see Table 6.8):

• Cost of the first unit: T= (1+x)T0

• Cost of the following units (if programme of 1 unit/site): T0.

• Cost of the 2nd unit on a site with one pair: yT0 (6.6)

• Cost of the 3rd unit on a site with two pairs: zT0

[50] Cost of the 4th unit on a site with two pairs: yT0 ,

where it is assumed that the cost of the 2nd unit of a pair is independent of the rank of the pair on the site.

Productivity effect

It is considered that a productivity effect only occurs as of the 3rd unit of a series. If n is the rank of the unit in the series, and Tn is the cost which results from taking into account the sole programme effect, it follows that:

Tn= ^ , as of n>2 (6.7)

n (1+k) n-2

[51] OKBM Afrikantov is a principal Russian design organisation for nuclear propulsion reactors: http://www. okbm. nnov. ru/.

’ However, this explanation is based on a single set of data and thus should be viewed with caution.

[53] Fuel costs also include fuel cycle costs, but here the predominant SMR strategy is to start all reactors in a currently mastered open fuel cycle with low enriched uranium as a fuel load. This strategy is also typical for all SMRs with fast reactors, except the Korean PASCAR (see Tables A1.6 and A1.7 in Appendix 1). Although a closed fuel cycle with the reprocessing of spent fuel is foreseen for most of the advanced SMRs presented in this report, SMRs are most likely to make a transfer to the new fuel cycles only when such cycles are well mastered for all other reactors. Specifically, all of the fast SMRs addressed in this report offer long refuelling intervals from 7 to 30 years, which offers a time lag for the fuel reprocessing technology to be proved and developed on a commercial scale.

[54] CHP = combined heat and power.

[55] Analysis of the Appendix 1 data has shown that most of the designers specify the production rates for non­electrical products without specifying the electric output of the plant matching exactly these production rates.

[56] As discussed in Section 6.3, this assumption is based on the analysis of only one set of consistent data and, therefore, cannot be considered as reliable. Using this assumption may, therefore, add a certain degree of conservatism to LUEC evaluation for barge-mounted plants.

: The input data and the results are presented in Table A3.5 in Appendix 3 for the discount rates of 5% and 10%.

[57] In Table 6.6 n = 0.51 corresponds to an average between the direct and the indirect costs of a NPP.

[58] Corresponds to the range of n in Table 6.7 .

*Taking into account the heat credit