Use of Immature Technology Causes Major Delays at the RRP

In November 2007, the Rokkasho Reprocessing Plant (RRP) encountered several difficulties in combining liquid high-level waste (HLW) with borosilicate glass. HLW is waste that remains after uranium and plutonium are recovered from used fuel. Operators had difficulty in controlling the vitrification furnace temperature, problems with accumulation of platinum group elements that impact glass pourability, bricks falling to the bottom of the furnace and clogging of nozzles that prevented the waste liquid from flowing into the vitrification canister below (See Figure 3.7).³¹ The problems were like those which had been encountered at the Tokai Reprocessing Plant. Joule

heating of the furnace could not uniformly heat and liquefy the wastes (Yoshioka, 2013; Okubo).

 In January 2004, the R&D Problem Evaluation Committee of the Japan Nuclear Cycle Development Institute (JNC), the PNC’s successor, reported that the committee had learnt for the first time that the PNC’s glass melting furnace was an immature technology that had not been fully developed. There were many problems that needed to be resolved. The furnace was too small for the RRP’s AREVA plant and needed to be scaled up to function at the RRP. More surprisingly to the committee, the vitrification furnace design did not address any of the problems that had occurred earlier. The committee judged that the technology would be mature in five years at the latest (JNC, 2004; JCA, 2017).

31 The vitrification process occurs at the separation stage. The separated highly radioactive fission materials travel downward into the glass melting furnace. The fission materials are melted and flow down into a canister that will store the vitrified waste (See Figure 3.7).

106 Figure 3.7 Glass Melting Furnace

Note: 1) HALW is highly radioactive liquid waste. 2) In the glass melting furnace, electric current passes directly through the glass by means of electrodes installed in the furnace. Joule heating melts the glass. The molten glass combines with HALW and is poured into the canister by heating the flow-down nozzle located at the lower part of the casing.

Source: JNFL, 2014.

The problems with the vitrification process were finally resolved in 2013, apparently by replacing it with La Hague’s furnace. The JNFL described its experience with the glass solidification problem at a meeting of METI’s Nuclear Energy Business Environment Preparation Study Special Working Group as follows:

 The RRP acquired the main segment of its reprocessing machinery from AVEVA (except for the PNC’s glass vitrification segment). The RRP imported the main segment after the La Hague UP2-400, UP2-800 and UP3 plants had successfully tested and proven the validity and operational soundness of the segment. UP2-400 conducted the test operation in 1986 and completed the rated operation in 1987, the UP2-800 conducted the same test from 1994 to 1995 and UP3 from 1989 to 1993

 In 1988, the glass vitrification technology was still under development at both PNC and AREVA and both were immature. But the RRP introduced the PNC’s glass vitrification technology first because it was indigenously developed.

 The RRP started the final, pre-operative test of the entire process in March 2006 and completed the test successfully in February 2008, except for the glass vitrification section.

It processed 425 tons of spent fuel, as planned.

 The RRP did not validate the glass vitrification technology. The RRP conducted an active test without going through an orderly development process, from concept design, mockup test, basic design and detail design. It immediately conducted an active test. The RRP

107

started to install the device without waiting for the findings of the operational test at the Tokai Reprocessing Plant and without obtaining sufficient validation from its own real-size mockup test.

 The now-defunct PNC had no obligation to guarantee the performance of the technology and it is ambiguous who is responsible for technology verification after the technology transfer. Moreover, the RRP depended on the PNC and its affiliated manufacturers for the operation of glass vitrification and scarcely involved itself in the management of the process. Furthermore, there were few full-time RRP employees at the work site and the reprocessing work was basically run by loaned-out contract workers who had little experience in reprocessing (Kamata, Saito, 2011).

In addition to the glass vitrification problem, the RRP had problems with high-level waste

concentration technology that was imported from UK’s BNFL. The liquid concentration technology entails the collection and concentration of waste liquid from various sources into a can, from which the liquid is separated from solids and would be siphoned into the glass vitrification device. On 30 July 2010, the RRP discovered a leak from the can for the first time, followed by two more similar incidents. The RRP investigation found that the bottom of the can had corroded due to buildup of solid waste at the can’s bottom (Mihama no Kai, 2011). The reprocessing plants in France and UK reportedly had similar problems with their glass vitrification facilities (ICRC, 2005).

These problems of course delayed commissioning of the RRP. The reason these problems emerged stemmed partly from Japan’s determination to rely on indigenously-designed technologies for the RRP and partly on management’s over-confidence that it could improve upon imported foreign technologies. In 2010, the JNFL finally decided to remodel the glass vitrification unit to better control temperature of the molten glass by using the segment supplied by AREVA. Also, in June 2013, AREVA signed a new strategic agreement with JNFL to bring the RRP into commercial operation, including active testing, the start-up itself, capacity ramp-up and plant optimization. This means that the RRP now is virtually an AREVA plant (WNN, 2013, 2014b). In October 2013, the RRP completed the vitrification technology development facility, “Rokkasho Vitrification

Laboratory (RVL),” and was ready to start-up operation.

However, new regulations from the Nuclear Regulation Authority (NRA) came into effect in December 2013, including a requirement that facilities have a seismic design basis of 600 Gal (up from 375 Gal). This meant that the RRP plant had to withstand earthquakes five times bigger that it was built for (magnitude 9.0 instead of magnitude 8.3) and oceanic earthquakes 1.5 times bigger (magnitude 7.4 instead of 7.2). Under the new regulations, any building would be required to be anchored to bedrock, an existing water storage tank built before the new regulations must be retrofitted with anchors to secure it to the bedrock and a large amount of piping must be upgraded to higher seismic standards (WNN, 2017a). JNFL filed requests in January 2014 for safety reviews of the RRP. It had hoped to get an operating license by September 2018. But a false report was found in relation to a safety rule violation, which came to light in 2015, leading to a delay in the NRA’s screening. As of 22 December 2017, JNFL announced at its extraordinary press conference that to comply with additional regulatory requirements of the NRA, the schedule for the expected completion of the RRP had been changed from the first half of Japan’s fiscal year 2018 to the first half of FY 2021 and the completion of the MOX Fuel Fabrication Plant (J-MOX) had been changed from the first half of FY 2019 to the first half of FY 2022 (See Figure 3.8) (JNFL, 2017e).

108

Figure 3.8 History of Rokkasho Reprocessing Plant (RRP) Construction and the Expected Date for Start of Operation *

* The chart shows the construction timeline of RRP as indicated on the RRP website. It implies that plant construction started in FY2001. Actually, RRP construction started on 28 April 1993 and the effort to complete the construction has continued for 26 years. Currently, the RRP has reached the stage of “Final Commissioning Test" to meet new regulatory safety standards and hopes to

complete its construction in the first half of FY 2018.

Source: JNFL, 2018b

Japan remains the only non-weapon state that has not abandoned reprocessing. The RRP is large and the operations and maintenance are costly. Data presented by Japan’s AEC in 2012 break down the annual projected costs of reprocessing and associated activities for each year from 2012 through the plant’s eventual decommissioning. Japan’s AEC estimates that the total cost of the Rokkasho reprocessing plant over its life, including capital, operating, and decommissioning costs, along with some allowance for interest during construction, will be $108 billion for reprocessing 32,000 tons of spent fuel.

Meeting the new safety regulations will certainly add to the cost of reprocessed fuel. Moreover, once the plant starts operation, it will be far costlier to decommission than if the plant had never gone into operation. Plants that have operated will have widespread high-level radioactive contamination. Once started, it will be almost impossible to turn back. Thus, the JNFL could be tempted to not start up the RRP and decommission it before it is used. The decommissioning cost would theoretically be the responsibility of the JNFL and FEPC.

Even if the RRP successfully begins its reprocessing operations, the outlook for business does not look positive. The RRP no longer has Monju FBR as its primary source of plutonium, which had been the major reason for the establishment of the RRP in the first place. Also, the growth in the amount of spent fuel on-hand (spent fuel received, as shown in Figure 3.9) has not increased

because so few nuclear power plants are in operation in Japan. Unless this situation changes and the RRP has access to more feed material, it will have little work to do.

109

Figure 3.9 Trends in Japan’s Spent Fuel Received for Reprocessing; Total = 3,390 tU Source: JNFL, 2017f.

By comparison, Japan has relatively less spent fuel in its inventory compared to states such as Canada and the United States (See Table 3.9). Given the number of power reactors that have been in operation (prior to the Fukushima disaster), Japan’s experience is most comparable to France. At the end of 2007, for example, Japan had 54 reactors operating and about 19,000 tons of spent fuel on hand, whereas France had 58 reactors and about 13,500 tons of spent fuel. The reason France has a relatively small inventory of spent fuel is that it reprocesses its own spent fuel. Canada and the United States do not and therefore they have several times more spent fuel in their

inventories—considering the size of their reactor fleets. Their experience with direct disposal of spent fuel, without reprocessing, would indicate that this is a viable strategy for the long term.

Japan’s government has long resisted the economic arguments that have been made against reprocessing, however.

At this point, Japan still has a choice with respect its nuclear policy: it could continue the present course, depending upon expensive reprocessed fuels and accepting the greater risk that is associated with more complex nuclear technology, or it could shift to reliance on enriched uranium and long-term direct disposal of spent fuel.

1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 tU 8 24 96 340 312 524 425 540 265 391 239 93 85 18 13 8 5 4 0

100 200 300 400 500 600

tU

110 Table 3.9 Spent Fuel Management Programs of Selected Countries

Country Began Commercial Operations

Number Operating Reactors

Spent Fuel Inventory

(as of 2007 in Tons Heavy Metal)

Spent Fuel Management Program Commercial Reprocessing Capacity (Tons Per Year)

Canada 1968 18 38,400 Does not process spent nuclear fuel

India 1957 22 Not

transparent about its fissile material stocks

Reprocesses its own spent nuclear fuel as well as spent fuel from other countries

India (PHWR, 4 plants) 360

Japan 1966 54 19,000 Reprocesses spent nuclear fuel;

Historically, Japan has shipped its spent nuclear fuel to France and the UK

Japan (Rokkasho) 800*

Russia 1963 33 17,895 Reprocesses some spent nuclear fuel as well as spent fuel from other countries

Russia, Ozersk (Mayak) 400 France 1964 58 13,500 Reprocesses its own spent nuclear fuel

as well as spent fuel from other countries France, La Hague 1700 South

Korea 1978 23 10,900 Does not reprocess spent nuclear fuel

Germany 1969 9 5,850 Shipped most of its spent nuclear fuel to France and United Kingdom for

reprocessing, until 2005 United

Kingdom

1956 18 5,850 Reprocesses its own spent nuclear fuel

as well as spent fuel from other countries

UK, Sellafield (THORP) 600 UK, Sellafield (Magnox) 1500

Sweden 1972 10 5,400 Does not reprocess spent nuclear fuel

Finland 1972 4 1, 600 Does not reprocess spent nuclear fuel United

States

1977 100 70,000 Does not reprocess spent nuclear fuel

* Now expected to start operation in the first half of FY 2021.

Source: Table is based on authors’ interpretation and analysis of data from the United States General Accounting Office, 2012; the European Nuclear Society, no date; Business Standard, 2011; Nuclear Power Corporation of India, 2015; Albright and Kelleher-Vergantini, 2015);

Nuclear Energy Institute (NEI), 2017).

111