April 28, 2000

 

 

Bruce Alberts

President

National Academy of Sciences

2101 Constitution Avenue, NW

Washington, DC 20055

 

 

Dear Dr. Alberts:

 

            We are writing to you in reference to a report of the Board on Chemical Sciences and Technology of the National Research Council (NRC) entitled "Electrometallurgical Techniques for DOE [Department of Energy] Spent Fuel Treatment," which was released on April 18.  We are concerned that some important conclusions of this report are misleading and call into question the technical judgment, if not the objectivity, of NRC's Committee on Electrometallurgical Techniques for DOE Spent Fuel Treatment ("the Committee").  Consequently, we strongly recommend that final publication be suspended pending an independent review of the report to determine whether its conclusions are supported by the underlying data. 

 

            This report contains the Committee's final assessment of a demonstration of DOE's electrometallurgical technology (EMT), which is being developed at Argonne National Laboratory (ANL) for the treatment of certain DOE spent fuels.  In response to public concerns regarding the safety, effectiveness and proliferation resistance of EMT, DOE requested in 1994 that the NRC form a committee to evaluate the technology.  In June 1996, DOE commenced a three-year demonstration of EMT to treat a fixed number of spent fuel assemblies (100 "driver" and 25 "blanket" assemblies) from the now-shutdown Experimental Breeder Reactor-II (EBR-II) in Idaho.  The Committee requested that ANL provide a set of criteria according to which the demonstration project would be evaluated.  ANL provided these criteria in May 1997, and NRC approved them.  These criteria were reaffirmed in the Committee's status reports on EMT, up to and including its 1999 report. 

 

            The first criterion for success of the demonstration project is as follows:

 

      "Demonstration that the 100 driver and 25 blanket EBR-II assemblies can be treated in FCF [Fuel Conditioning Facility] within three years, with a throughput rate of 16 kg/month for driver assemblies sustained for a minimum of 3 months and a blanket throughput rate of 150 kg for one month." [1]

     

            The record, as confirmed to us by ANL spokesman Paul Pugmire, shows that the EMT demonstration project failed to meet this fundamental criterion.  In particular, by June 1999, when the demonstration project was scheduled to end, only five blanket assemblies, or 20% of the goal quantity, had been fully treated.  Moreover, at that time ANL had not yet demonstrated that a blanket throughput rate of 150 kg in one month was achievable.  This demonstration did not even begin until late July 1999, well after the project was supposed to have been completed. [2]   Even by September 30, 1999, three months after the demonstration was supposed to have been completed, only 14 blanket assemblies out of 25 had been processed. To date, only 18 blanket assemblies, or 72% of the original goal, have been processed.  Moreover, some subsidiary goals, such as the production of ten ceramic high-level radioactive waste samples, also appear not to have been met in a timely manner.  

 

            In view of ANL's failure to complete the demonstration project on time, the Committee's conclusion that "ANL has met all of the criteria developed for judging the success of its electrometallurgical demonstration project" is puzzling.  How did they arrive at this judgment?

 

            The mystery is solved on page 69 of the Committee's report, when one learns (in a minor footnote reference) that DOE recommended a change in the wording of Criterion 1 from "Demonstration that 125 EBR-II assemblies can be treated ..." to "Demonstration that 100 driver and up to 25 blanket EBR-II assemblies can be treated ..." (emphasis supplied), and the Committee agreed to the change.  An NRC staffer confirmed that this change was made in Spring 1999. 

           

            In other words, as the end of the demonstration project in June 1999 approached, DOE realized that it could not meet the processing target established in Criterion 1, so it altered the criterion instead.  This sleight-of-hand, sadly, is all too typical of the way DOE conducts its business.  It is highly inappropriate for the Committee to acquiesce in such behavior by changing a fundamental evaluation criterion late in the game.  The Committee claims that the DOE demonstration was able to meet the "intent" of the criterion without processing the entire quantity of  blanket fuel.  However, this claim is not justified, amplified or explained in the report.  Moreover, it is contradicted by ANL's previous arguments that processing fewer than 25 blankets would be insufficient.  

           

            The Committee's report is structured in such a way that it is very difficult for the reader to ascertain what actually happened during the demonstration.  First, the revised Criterion 1 is presented early in the document (second page of Chapter 2) without any mention that it was not the original criterion.  Nowhere is it plainly stated that the demonstration project failed to meet one of its initial goals.  Second, the document does not contain a straightforward accounting of the work that was completed by June 1999, the end of the demonstration period, including a detailed description of what problems were encountered in trying to meet the original goal and what, if any, outstanding processing problems remain.

           

            For example, the Committee's final report does not mention one major cause of the delays experienced by the demonstration project that was described in a previous report: [3] namely, three serious incidents within four weeks in the FCF processing cell which resulted in (1) external radiological contamination of 11 personnel (four of whom also received internal exposures), (2) a violation of maintenance procedures that could have led to a radiological release, and (3) failure of a heavy piece of transfer equipment. Together, these incidents led to a two-month loss of processing time.  

           

            Since the FCF is the facility to be employed if DOE decides to use EMT for its entire inventory of sodium-bonded spent nuclear fuel, these incidents, rather than being not worthy of mention in the Committee's final report, are in fact central to a determination of whether such an undertaking should be authorized, given the questionable levels of safety and equipment performance displayed during the FCF demonstration project.

 

            Even a two-month shutdown cannot fully explain why ANL was approximately seven months behind schedule in processing the blanket fuel assemblies. [4]   The Committee's report should be revised to include a thorough analysis of the performance of the Mark-V electrorefiner, which was used to process the blanket fuel, so that the reasons for the additional delays can be understood by the reader.

 

            The issue is not simply that DOE was unable to complete a project according to schedule, which is hardly earth-shattering news.  Rather, the issue is that DOE agreed at the outset (and continually reaffirmed until near the very end) that the demonstration project would be evaluated against a set of fixed criteria, including a well-defined processing goal.  Given the propensity of DOE projects to experience severe delays, cost overruns, safety problems and unmet technical goals, the EMT demonstration project should be viewed as an essential practical test of the ability of the ANL EMT team to complete a project within a fixed time frame, thereby providing confidence in its ability to successfully manage larger projects.  While ANL argues instead that the demonstration was merely an R&D project and that delays could be expected, we would point out that the Committee's Criterion 3 requires "demonstration of an overall dependable and predictable process, considering uptime, repair and maintenance, and operability of linked process steps."  This implies an expectation that ANL should have been able to complete its processing targets even in the event of contingencies.    

 

            For the Committee to effectively dismiss both Criteria 1 and 3 and suggest that they were irrelevant raises serious questions about the technical judgment it has applied to this matter. Further, the Committee's failure to accurately describe the inability of the demonstration project to meet these criteria can only lessen the confidence of the public in the Committee's assessment of the demonstration project in relation to the other success criteria, which are much more technical and subjective in nature.  The Committee should not put itself in the position of being viewed as a mere rubber stamp of DOE activities.  Such a perception could be very damaging to the credibility of the NRC.

 

            There is also a crucial policy question at stake here.  After years of insisting otherwise, DOE has admitted that it is not necessary to use EMT to process sodium-bonded blanket fuels, which make up the vast majority of DOE's inventory of sodium-bonded spent fuel.  Instead, these fuels can be directly disposed of in a geologic repository, after a relatively simple and inexpensive process is used to remove the sodium layer.  The public should be fully informed of any evidence that problems are being experienced in applying EMT to blanket fuels, so that the direct disposal alternative can be fully assessed.

 

            In view of the defects reviewed above, the Committee's assessment of the demonstration project is simply not acceptable.  We strongly urge that the final publication of the report be suspended until an independent review of the Committee's work is conducted.  The final report should contain, at a minimum, a clear statement from the outset that the demonstration project did not achieve its original goal of processing 25 blanket assemblies and demonstrating a 150 kg/month throughput within three years, and that the Committee's assertion of success is based on a revised, weakened criterion which was proposed near the very end of the demonstration.  It also should explain the reasons why the goal was not met, including a full discussion of the safety significance of the incidents at the FCF that led to a two-month shutdown, and of the causes of the additional extensive delays experienced during processing of blankets and ceramic waste forms. 

 

            Because the report is soon to be published, we respectfully request your reply to this letter within seven business days.  Thank you for your prompt attention to this urgent matter.

 

 

                                                                    Sincerely,

                                         

 

                        Edwin S. Lyman, PhD                                         Paul L. Leventhal

                        Scientific Director                                                President

 

 

cc:  Bill Richardson, DOE  


[1] National Research Council, Committee on Electrometallurgical Techniques for DOE Spent Fuel Treatment, Electrometallurgical Techniques for DOE Spent Fuel Treatment:  Status Report on Argonne National Laboratory's R&D Activity Through Spring 1997 (Washington, DC:  National Academy Press, 1997), 10.  Later statements of this criterion grouped the driver and blanket assemblies together.

[2] National Research Council, Committee on Electrometallurgical Techniques for DOE Spent Fuel Treatment, Electrometallurgical Techniques for DOE Spent Fuel Treatment:  Final Report (Washington, D.C.:  National Academy Press, 2000), 109.

[3] National Research Council, Committee on Electrometallurgical Techniques for DOE Spent Fuel Treatment, Status Report on Argonne National Laboratory's R&D Activity as of Fall 1998 (Washington, D.C.:  National Academy Press, 1999), Appendix C.

[4] Blanket processing began in August 1998.  ANL projected after the shutdown that processing of the blanket assemblies would not be complete until late August 1999.  However, only 14 blankets were processed by that time.  If  the Mark-V electrorefiner had operated, say, at an average throughput of 120 kg/month (80% of its maximum throughput), the time required to process the additional 11 blankets (520 kg of uranium) would have been about five months, putting the project over six months behind schedule.