For the purposes of the present chapter and following the discussion in section 6.3, the sums of the O&M and fuel costs for land-based SMRs were taken equal to the corresponding sums for NPPs with the reference large reactors. For barge-mounted plants, the corresponding sums were multiplied by a factor of 1.5 reflecting the assumption of a higher O&M costs owing to the need of periodical factory repairs of a barge[56], see the discussion in Section 6.3. The resulting O&M and fuel costs (components of the LUEC) for SMRs plant configurations considered are given in Appendix 3 (Table A3.6).

4*335 MWe (USA) 5*125 MWe (USA) 2*300 MWe (Russia) 2*300 MWe, barge (Russia) 1 *90 MWe (Korea) 2*35 MWe, barge (Russia) 2*8 MWe, barge (Russia)

О

О

аз

Q_

аз

APWR/ABWR Advanced Gen III+

million USD

Recently the so-called “mini” or small and modular reactors have attracted a lot of attention. Since 2008, several small private companies have been created in the United States to support the design development, patenting, licensing and commercialisation of several new SMR concepts. The design concepts of such reactors were analysed in this study. The conclusions are as follows:

• The new small and modular (‘mini’) reactor concepts being developed in the United States fit well into the known technology lines for nuclear reactors. The new US small and modular reactors (NuScale, mPower, New Hyperion Power Module, and ARC-100) share many of their design attributes with other small reactor design concepts being developed in countries other than the United States. It is their repeated use of the three attributes, namely:

— multi-module plant option;

— option of flexible capacity addition/deletion; and

— underground location of the reactor modules;

that distinguishes most of the new US small and modular reactors from other small reactor concepts developed elsewhere in the world.

• The attributes of small and modular reactors, such as small upfront capital investments, short on-site construction time (with the accordingly reduced cost of financing), and flexibility in plant configuration and applications, could be attractive for private investors.

High temperature gas cooled reactors (HTGRs, see a brief description in the box 4.2) were operated in the past in the United Kingdom, the United States and Germany, and there are currently two small operating experimental reactors of this type in China (HTR-10) and in Japan (HTTR). The previous operating experience, cumulatively stretching from 1965 to 1989 [4.1], is probably too dated to be judged according to the current regulatory norms or safety standards. In 2010 there were no operating commercial reactors of this type anywhere in the world.

Basic characteristics of the HTGR designs considered in this report are given in Table 4.4[24] [25]. All HTGRs are helium cooled reactors. The PBMR appeared to be a promising concept in an advanced development stage, with targeted deployment date in South Africa set for 2013. However in 2010 the vendor company — PBMR Pty — suffered from financial difficulties with the government no longer supporting the project. By that stage they had started to develop an indirect cycle HTGR similar to the Chinese HTR-PM.

As will be discussed in more detail in Section 8.5, all HTGR safety design concepts provide for passive decay heat removal to the outside of the reactor vessel. In view of this, with the currently known reactor vessel materials it appears that ~600 MWth is an upper limit of the unit size for HTGRs, which means that all HTGRs would fall into the SMR category.

Plant configuration with direct Brayton cycle is employed in all of the designs of Table 4.3, except the Chinese HTR-PM which is an indirect cycle HTGR employing the steam generators and a

Rankine cycle with reheating for power conversion. The indirect cycle efficiency of the HTR-PM is also remarkably high, 42%, due to steam reheating.

Because the high-power Brayton cycle gas turbines are currently not available from the industry, the indirect cycle HTR-PM appears today as a leader among all HTGRs, with the construction related actions and licensing started in China, see Section 4.4.

When high temperature non-electric applications are targeted, the HTGR design includes an intermediate heat exchanger to deliver heat to process heat application systems. Because of high temperatures (up to 850-900oC), HTGRs appear to be the only SMR technology line for which complex co-generation is considered, such as, for example, electricity generation with co-production of hydrogen and use of reject heat for seawater desalination.

The main technical characteristics of HTGR SMRs considered are the following:

• All HTGR designs target availability factors of more than 85%. The plant lifetime is typically 60 years for HTGRs with pin-in-block (non-moveable) fuel design and 35-40 years for those with pebble bed (moveable) fuel design.

• On-line refuelling is used in the pebble bed designs (HTR-PM and PBMR [previous design]), while the pin-in-block designs use partial refuelling in batches. [26]

Given the reference data [7.1], the evaluations performed in Sections 7.2.2 and 7.2.3 were limited to the generation of electricity (or the production of electricity and heat) in locations with large interconnected electricity grids. As noted in the previous sections, these evaluations showed that barge-mounted NPPs with small PWR-8 and PWR-35 twin-units (based on the Russian ABV and KLT-40S designs) would not be competitive in these conditions.

However, small and transportable NPPs (such as the KLT-40S and the ABV) are being developed for application in remote or isolated areas with difficult access and with no interconnected electricity grids (or even no grids at all) rather than in populated areas with the established grids of a large overall capacity. This section provides several evaluations of the competitive deployment options for NPPs with SMRs in the remote or isolated areas (conventionally referred to as “off-grid” locations). However, no market analysis has been performed.

Building reactors in series usually leads to a significant per-unit cost reduction. This is due to better construction work organisation, learning effect, larger volumes of orders for the plant equipment and other factors. However, the first-of-a-kind (FOAK) power plant is usually considerably more expensive than subsequent units.

Reference [6.4] suggests an algorithm, based on the French experience, (see Table 6.8) to calculate FOAK plant effects in the overnight capital cost and cost reductions from building more than one serial plant on a site:

The main parameters of this algorithm are:

-x: FOAK extra cost parameter

-y: parameter related to the gain in building a pair of units.

— z: parameter related to the gain in building two pairs of units on the same site.

-k: industrial productivity coefficient.

Productivity effect Cost of the last unit (multiplicative factor) (in a box)
Plant configuration	Total cost of the plant
	(1+x)To
FOAK
(1+x+y)T0 z
1+k 1
ЛШ
T
(1 + k)2

The industrial productivity coefficient k=0%-2%,

FOAK extra cost parameter x=15-55%,

Parameter related to the gain in building a pair of units y=74%-85%,

Parameter related to the gain in building two pairs of units on the same site z=82%-95% [49] [50]

Reference [6.4] suggests the following values of the parameters (based on the French experience):

x=15% to 55%, according to the nature and amount of changes in the design.

y=74%-85%

z=82%-95%

k=0%-2% (6.8)

According to (6.6) and (6.8), the FOAK plant could be 15% to 55% (35% on average) more expensive than the next ones (built at a site).

For the first and second pair of non-FOAK twin-units on the site, based on (6.6) and (6.8), the per-unit cost reduction factors would be:

1+y

Per unit cost reduction factor for twin units (first pair) = —^—=0.87 — 0.93 Per unit cost reduction factor for twin units (second pair)=1 x (Z — + ^ ) =0.76 — 0.9 (6.9)

If two pairs of non-FOAK twin-units are built on the site, the per unit overnight cost reduction may be as substantial as:

A reduction such as (6.10) is quite significant but it would not be sufficient to compensate the specific investment cost increase because of the scaling law (6.3).

Example: The cost of 4 non-FOAK 300 MWe versus 1 non-FOAK 1200 MWe

As an example, let us consider four non-FOAK 300 MWe PWRs (integral design or marine — type) built on the same site, and compare them to one large non-FOAK 1200 MWe PWR. In this case one should include the effects of economy from building subsequent units on the same site (equation [6.10]), simplification of the design (6.4) or (6.5), take into account the decrease of the cost of financing (due to reduction of the construction period from 6 to 3 years, see Figure 6.4), and multiply the result by the scaling factor (from 1 200 MWe to 300 MWe, see Table 6.7). The results are given in the Table 6.9.

From Table 6.9 it could be seen that, within the assumptions made, four integral type or marine derivative PWRs of 300 MWe class (and not FOAK) built on the same site may have the effective per unit specific overnight capital costs of about 10-40% higher (at n=0.5-0.6 and a 5% discount rate) compared to those of a NPP with a single large PWR of 1 200 MWe.

Similar results for almost identical case studies were obtained by the Westinghouse Electric Company [6.5]. In their case, the construction duration was assumed to be five years for the large plant and three years for each of the SMRs, the annual interest rate was 5%, and the scaling factor used (1.74) corresponds to n=0.6. They found that the specific capital cost of a 300 MWe PWR versus specific capital cost of a 1 200 MWe reactor of the same type would be increased by about 4% (compared to 10-22 % in our case, see Table 6.9 at n=0.6 and a 5% discount rate).

The effects defined by the parametric equations (6.6), (6.7) and (6.8) include both, learning in construction (parameters x and z) and in factory fabrication (parameter k), and sharing of common facilities and systems on the site (parameters y and z). An important assumption regarding learning is that the costs of engineering and facilities for each site are identical, which means similarity of the sites. Otherwise, the learning effects may be not observed.

The international and national NPP build experience, specifically, that of the Russian Federation and Canada [6.5] indicates that learning will not apply:

• if NPPs are consequently built in different countries;

• if there are regulatory changes in a country during the next NPP build;

• if siting conditions for the consecutive plants are essentially different; and

• if the interval between building consecutive plants is too long.

The last effect of a “too long” interval between consecutive plants building is illustrated by Figure 6.5. It is based on the OKBM Afrikantov[51] experience in factory fabrication of the marine propulsion reactors in the Russian Federation [6.9]. As it can be seen from the figure, for the case of full factory fabricated nuclear plants the requirements of continuity are quite strict, with notable increase in labour intensity observed even for a one-year break in the production process (unit number 3 on Figure 6.5).

Some of the SMR designs addressed in this report provide for a long-life reactor core operation in a “no on-site refuelling mode”. The targeted refuelling interval for such SMRs is between 5 and 30 calendar years. Although the fuel burn-up in all these designs is quite moderate and does not exceed the typical values for a conventionally refuelled design of the same technology line, the continuous long-time core operation may result in ageing and fatigue of some safety related structures and components. In view of this it would be necessary to justify that the original safety case is retained throughout the whole long period of continuous plant operation. The regulatory norms providing for such a justification may be not readily available in national regulations. Countries having relevant experience with marine propulsion reactors will have an advantage. Otherwise, a “license-by-test” approach highlighted in Section 9.4 could be helpful.

Large NPPs do not fit in any of these specific markets; therefore, SMRs would compete only with the local non-nuclear energy options.

To analyse the deployment potential of SMRs in remote or isolated areas, the LUEC estimates were compared with the electricity tariffs for that area (because the generating costs of other technologies would be extremely difficult to calculate). The analysis of SMR competitiveness in “off-grid” locations has identified a significant potential for their applications in remote areas with severe climatic conditions hosting mining or refinement enterprises, or military bases, and the affiliated small settlements (see Figure E.8).

On a purely economic basis, isolated islands and small off-grid settlements in populated developing countries (e. g. Indonesia, India) could also become potential market.

It has been found that a variety of land-based and barge-mounted SMR plants with LUEC substantially higher compared to large reactors could still be competitive in these niche markets if they meet certain technical and infrastructure requirements, defined by the specific climate, siting and transportation conditions. In particular, co-generation with the production of heat or desalinated water appears to be a common requirement in many of the niche markets analyzed.

In the analysis performed for the “off-grid” locations, many cases were found when small barge-mounted NPPs with the PWR-8 and the PWR-35 twin-unit plants (based on the Russian ABV and KLT-40S designs) would be competitive.

In view of the negligible contribution of decommissioning costs to LUEC [7.1] (see the discussion in section 6.4), the decommissioning costs were set to zero for all SMR plant configurations estimated in this section.

7.1.3 LUEC estimate

First, the investment component of LUEC for SMR plant configurations was estimated. The calculations were performed2 using formula (6.2) given in Section 6.2.

Next, all LUEC components given in Table A3.6 and Table A3.5 in Appendix 3 were summed yielding the final LUEC estimate in USD/MWh, shown in Table 7.3.

* for 5 % discount rate, USD per kWe. ** at 5% discount rate, USD per MWh.

Table 7.4. LUEC estimates for the various SMR plant configurations (10% discount rate)

PWR-90(2) single module plant

The data from Table 7.5 leads to the following observations:

• The estimates of LUEC are higher than the designers’ data for all plant configurations with PWRs based on the Russian marine derivative designs (by 60-70%, and for the PWR-35 by 100-130%).

• The estimates of LUEC are overlapping the designers’ data for the PWR-125 multi-module plant based on the US mPower design.

• The estimates of LUEC are slightly lower than the designers’ data for the integral type PWR based on the Korean design SMR (by 10%).

• LUEC estimates for the PWR-90 obtained with the two different Korean NPPs with large reactors used as reference are nearly equal.

Assuming that some SMR designers may explicitly include the heat credit in the LUEC values specified for their designs, an evaluation of the possible impact of such an inclusion was performed. The heat credit values for several SMRs were taken from Table 6.12. The ratios of the independent LUEC estimates and the designers’ data on LUEC for the corresponding SMRs from Table 7.5 were then adjusted taking into account the heat credit values, with the results shown in Table 7.6.

From Table 7.6 one can conclude that:

• With the assumption of a heat credit explicitly included in the LUEC, the LUEC estimates for PWR-90 and PWR-302 show reasonably good agreement with the designers’ data given in Table 4.14. The estimate for PWR-8 is 26-42% higher than the designer’ values.

• No explanation was found for the observed difference between the LUEC estimate for the barge-mounted plants with the two PWR-35 based on the Russian KLT-40S and the designers’ data for this plant. Even with heat credit taken into account the LUEC estimate appears to be 26-96 % higher than the designers’ data.

An important factor affecting the LUEC estimation carried out in this report is parameter n in the scaling law (6.3) discussed in Section 6.2.1. For the estimates presented in Table 7.1 to Table 7.6 of this section the value n = 0.51 was used corresponding to the average[57] from Table 6.6. To evaluate the uncertainty associated with the selection of n, Figure 7.4 graphically represents the difference between the LUEC estimates and the designers’ data on LUEC (maximal values) for different values of the parameter n, ranging from 0.45 to 0.6[58].

Figure 7.4 shows that the LUEC estimates are quite sensitive to the selection of parameter n in the scaling law (6.3) and taking (or not taking) into account the heat credit. If heat credit is not taken into account (where it could apply) the majority (5 out of 8) of the independent LUEC estimates are significantly higher compared to the designers’ data on LUEC. If heat credit is taken into account, the majority (5 out of 8) of the independent LUEC estimates envelope the designers’ data on the LUEC.

The independent LUEC estimates obtained in this section (Table 7.3) are used in Section 7.2 to evaluate the deployment potential of SMRs in a number of electricity and electricity and heat markets around the world.

Figure 7.4. Difference (in %) between estimated LUEC and the designers’ maximal values for LUEC at

different values of n ranging from 0.45 to 0.6.

200

"(Л

CD

a

O) ш

it

Ь

The general conclusions from the evaluation of the competitiveness of SMRs performed for the electricity markets (in “on-grid” applications) are similar to the general findings on nuclear power presented in the recent OECD study Projected Costs of Electricity Generation, 2010 Edition. However, there are some important SMR-specific conclusions that are summarised below:

• Within the assumptions of the evaluation performed, the nuclear option in general (NPPs with large reactors or with SMRs) is competitive with many other technologies (coal-fired plants, gas-fired plants, renewable plants of the some types) in Brazil[82], Japan, the Republic of Korea, the Russian Federation and the United States, but not in China.

• SMRs, including twin-unit and multi-module plants, generally have higher values of LUEC than NPPs with large reactors. However, like NPPs with large reactors, some SMRs are

expected to be competitive with several of the coal-fired, gas-fired and renewable plants of the various types, including those of small to medium-sized capacity (below 700 MWe).

A plant with SMRs could be a competitive replacement for decommissioned small and medium­sized 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). SMRs (like nuclear in general) could be more competitive if carbon taxes are emplaced.

In other words, SMRs could be more competitive than many non-nuclear technologies for generating electricity in the cases when NPPs with large plants are, for whatever reason, unable to compete.