Fwd: Criticality Accident Summary Significantly Updated

["Michael C. Baker" <mcbaker@lanl.gov>]

>From: "Valerie L Putman" <VPUTMAN@inel.gov>
>Date: Wed, 6 Oct 1999 20:36:44 -0600
>Subject: Criticality Accident Summary Significantly Updated (991006)
>
>
>=================================
>ACCIDENT SUMMARY TO DATE (991006 MDT pm)
>=================================
>
>DISCLAIMER:  This summary is an in-progress work.  It includes
>additions and corrections to previous summaries.  Information is based
>on a copies of a somewhat technical Japanese presentation made in
>Vienna, news articles, other's experience with fuel fabrication in the
>U.S., and email messages from technical personnel who are not directly
>involved with the incident, incident response, or incident investigation.
>Faxed presentation copies are poor quality, with much information
>blurred due to size and/or the copy generation.  Available news articles
>are mostly non-technical, often sensationalized, and sometimes
>contradictory, as expected.  In addition, there are some questions
>arising from different nomenclatures and possible translation problems.
>
>
>BACKGROUND EMAIL INFORMATION
>
>For convenience,
>- low enriched uranium (LEU) = less than 10% enriched
>- intermediate enriched uranium (IEU) = 10-60% enriched
>- all times are Tokaimura local and specified in a 0000 to 2400
>format.
>- uncertain and speculative information, and comments, are marked,
>often with brackets ([]) or question marks (?).
>
>To avoid limitation problems inherint in some email software, this
>message has no attachments, is in US ASCII, has a very long bcc list,
>and will be divided into parts (source citations and news URLs will
>eventually be sent in separate messages)
>
>Radiation dose values:
>-  1Sv = 100 rem [I do not know any exposure values in rad and cannot
>tell you neutron spectra for the various exposures or contributions
>from neutrons and gamma rays for most exposures]
>-  20mSv is worker annual dose limit, according to JCO presentation
>-  7Sv is considered lethal, according to interviewed Japanese medical
>personnel [some radiological technicians say 7Gy, not 7Sv]
>-  0.25Sv is usually the point at which some radiation sickness
>symptoms are expected, according to interviewed Japanese medical
>personnel [some radiological technicians say .25Gy, not .25Sv]
>
>=======================================
>
>
>INTRODUCTION
>
>A criticality accident occurred at 1035 on September 30, 1999.  It
>occurred in a conversion test building of a nuclear fuel fabrication
>facility in Tokaimura, about 140km northeast of Tokyo, in the Ibaraki
>Prefecture.  Japanese authorities provisionally rated this accident a
>4 on the International Nuclear Event Scale.  Their basis was radiation
>exposures to two operators and to two off-facility people.  However,
>on October 7th, the Daily Yomiuri reported authorities are considering
>rating the accident a 5 based on accident causes as well as on
>radiation exposures.
>
>
>THE FACILITY AND OPERATIONS
>
>The facility is part of JCO Co.'s Tokai Works site.  This  entire
>facility is unshielded because process material is unirradiated or
>"cleaned" of fission products.  The plant had no criticality accident
>response plans because it was assumed "critical fission chain
>reactions could not occur."  [More probably, "Emergency planning was
>not required because management and authorities assumed a criticality
>accident was not credible if an appropriate criticality safety program
>was implemented."]  Apparently small radiation doses to workers are
>credible because the subject area has gamma-ray monitors.  However,
>there are indications that JCO has no portable neutron monitors,
>although other nearby nuclear facilities might have such instruments.
>
>The nearest site boundary is rather close to the subject building.  An
>apparently public road is within 80m.  A building identified as a
>private residence is no more than 100m from the test building.
>
>This plant usually converts UF6 to UO2 for LEU (5% max.) PWR fuel.
>However, it occassionally converts U3O8 to uranyl nitrate for IEU
>(about 20%) breeder reactor fuel. This time the product was IEU uranyl
>nitrate for JOYU, a fast breeder research reactor.  [Even some
>Japanese presentation materials confuse the oxides and processes,
>which caused considerable confusion here in understanding the subject
>process and accident.]
>
>The last campaign with IEU was about three years ago.  Apparently this
>was either the first campaign, or first campaign with IEU, for two
>workers.  In addition, the third worker, possibly this team's
>supervisor, had only a few months experience with the subject
>operation, with handling IEU, or both.
>
>A manual provides written instructions for all operations.
>Operational limits change when needed to accommodate the LEU and IEU
>enrichments.  Japan adopted the criticality safety double contingency
>principle [but I do not yet know how it is stated or if it is a
>requirement or recommendation].
>
>Among other things, U3O8 powder is converted to uranyl nitrate
>solution in the subject building.  This batch process includes using a
>dissolution vessel and a sedimentation tank.  Reports are unclear as
>to the sedimentation tank's original purpose in this IEU process
>[precipitation to purify materials?, further mixing?, a hold point
>which, in the LEU process, is used to precipitate UO2?].
>
>According to the regulator-approved manual, U3O8 powder must be first
>weighed and then added to a small dissolution vessel in which uranyl
>nitrate is prepared.  The solution is then transferred, via plant
>piping, to a sedimentation tank, apparently where further processing
>is completed and/or the product is temporarily stored.  Batch size for
>the U3O8 dissolution, sedimentation tank, or both is operationally
>limited to 2.4kg uranium if the material is 16-20% enriched.  News
>reports seem to indicate the physical design prevents, or at least
>minimizes, over-batching.  [One INEEL chemical process engineer
>speculates that the 2.4kg limit applies to the dissolution vessel.  He
>believes this limit and the vessel's volume in turn limit
>uranyl-nitrate-solution concentration.  Hence, multiple batches from
>the dissolution vessel can be safely accumulated in the sedimentation
>tank for at least a short while.]
>
>However, two to five years ago the company approved a manual change,
>without regulator concurrence or notification.  This change
>accelerates the overall process.  It allows workers to mix U3O8 powder
>and nitric acid in steel buckets and then use funnels to add uranyl
>nitrate directly to the sedimentation tank.  Apparently the 2.4kg
>limit was then applied to the bucket, the sedimentation tank, or both.
>Some reports indicate the buckets are 50L containers or mop buckets.
>One article quotes an official who said regulators would not have
>approved the change because uranium oxide, nitric acid, and stainless
>steel react to produce a toxic gas. [Chemists tell me open and
>partially open containers, not steel, are the problem. The chemicals
>and their reaction involve readily airborne particles and vapors we
>should not breathe.]
>
>The sedimentation tank is a vertical cylinder with a flat top, dished
>bottom, 50cm inner diameter, 70cm(?) inner height (top to lowest point
>in bottom), 0.3cm stainless steel walls, and water cooling jacket. The
>top has a few penetrations, most notably a centrally located, vertical
>pipe about 25cm in diameter, an angled pipe [charging port?] about
>10cm in diameter, and four vertical smaller diameter pipes
>[charging/sample/reagent ports?].  There is also a drain line in the
>tank bottom and two horizontal lines on the side, a few cm above the
>bottom.  The 55cm outer diameter water jacket surrounds at least the
>bottom and lower 1/3 of the tank.
>
>
>THE ACCIDENT
>
>Before the accident, supervisors and, possibly, managers directed
>personnel to take shortcuts to accelerate processing further.  Workers
>were directed to use the buckets, overbatch, and, possibly, skip other
>steps.  Workers might also have decided to skip more steps than their
>oral directions specified.  Apparently the end result is that U3O8
>powder and nitric acid were added to the tank directly, and then mixed
>in the tank.  These shortcuts apparently reduced a three hour task to
>30 minutes.  Interviewed workers and supervisors said they knew
>nothing about the dangers of overbatching.  Some management personnel
>agreed, indicating worker training included almost nothing on
>criticality accident consequences and did not emphasize criticality
>accident prevention.
>
>The accident involved 18.8% uranium.  On Wednesday, workers poured
>about 9.2kg uranium from four buckets into the sedimentation tank.  On
>Thursday workers added about 6.9kg uranium from three buckets.
>Workers were most likely aware of the total accumulated mass.  They
>were certainly aware of the mass accumulated Thursday.  The resultant
>solution, or reflected slurry, went flash critical [prompt critical?].
>One email indicates the solution was approximately 370g/L uranium
>with, possibly, 1mole/L nitric acid.
>
>Contrary to earlier reports from other sources, the Japanese
>presentation indicates this system oscillated (or pulsed) between
>super- and sub-critical states for more than 17 hours.  [Each pulse
>probably went subcritical as solution was forced into lines connecting
>with the vessel top, and a new pulse was initiated as solution drained
>back into the vessel.  A stirring device in the tank and further U3O8
>dissolution might have contributed to the phenomena.]
>
>Available reports do not indicate the number of pulses, their
>magnitude, or their frequencies.  Measured radiation dose-rate values
>at the nearest site boundary seem fairly steady for hours, indicating
>pulse frequency was probably rapid enough to overwhelm radioactive
>decay evidence.  Fission yields are not yet reported for any pulse or
>for the reaction duration.
>
>Tokyo Electric Power Company rushed some 880lbs of borated material
>[boric acid?] to the JCO plant.  However, responders could not use it
>right away because they had no readily apparent means of remotely
>adding neutron absorber to the tank.
>
>It took about 3 hours on October 1st to drain cooling water from a
>water jacket around the tank. At 02:35 two workers with radiation
>protection coveralls took pictures of a cooling-water-drain-valve
>outside the conversion test building.  The area with the valve was
>highly contaminated [more probably, highly radioactive].  Although
>they stayed in the area only three minutes, one worker received
>11.92mSv gamma and 91.2mSv neutron.  At 03:00 another two workers went
>to a pump yard(?) and confirmed a cooling-water-circulation-pump(s)
>was working well.  At 03:22 another two workers opened the valve.
>However responders could not confirm the water jacket was successfully
>drained.  Finally other workers cut a drain pipe outside, at the
>valve's up-stream side, and injected gas.  At 06:00 responders
>confirmed the water jacket was drained.  The tank was finally
>subcritical.  18 workers total were involved in draining the water
>jacket.
>
>Apparently boron was then added to the system.  Boron might also have
>been sprayed into the room earlier.
>
>Officials declared the reaction terminated and the system safely
>subcritical at 09:20 October 1st.  However, some premature news
>reports indicated the reaction had terminated at various times after
>18:30 the previous night.
>
>Apparantly the building was not physically damaged.  [I do not know if
>the tank was damaged.  It might be intact given its wall thickness and
>all of its penetrations.]
>
>[FYI:  There is a separate, JCO nuclear fuel reprocessing plant in
>Tokaimura.  It apparently was shutdown after a 1997 fire/explosion
>accident.  If I remember correctly, that was a chemical or industrial
>accident but it released radioactive material.  Since the conversion
>test building was not damaged in the criticality accident, I speculate
>BBC video of a building with a hole in its roof is actually file
>footage of this other plant, taken during or after that earlier
>accident.]
>
>The entire plant/facility is now shutdown pending safety reviews and
>safety inspections.  As of October 7th, radiation levels were still
>too high to allow anyone into the immediate area [at least not for
>more than a few minutes].
>
>
>RADIOLOGICAL INFORMATION
>
>The three workers in the room saw a blue flash.  Reportedly they felt
>sick immediately.  [I'd feel sick that quickly just knowing what the
>blue flash means, whether or not physical symptoms were evident that
>quickly.]
>
>Presumably the area's gamma alarms activated, and everybody in the
>area left as quickly as they could.  The radius for this initial
>evacuation is not reported.  Most plant personnel were probably first
>evacuated to the further plant boundaries if not to offsite locations.
>
>News reports indicate some 7000 people were checked for radiological
>exposure.  Significant exposures were apparently limited to the three
>workers in the room, 36 other plant workers, three firemen [ambulance
>crew?], and up to seven residents who were near site boundaries at the
>time.  The three workers in the room respectively received estimated
>17, 10, and 3Sv doses.  [The first two workers were probably next to
>the tank, and the third worker was probably a few feet away.]  The next
>highest doses were to response and clean-up workers, each of whom
>reportedly received less than a 0.1Sv dose.
>
>The LA Times reported the three most severely exposed workers were not
>wearing dosimeter badges; their badges were apparently found some 160
>yards from the tank.  [However, were these badges actually removed
>during initial treatment at the plant, or did they off during the workers'
>evacuation, or were they truly not worn?]
>
>Firemen [ambulance crew?] were exposed when they entered the area
>without appropriate personal protective equipment.  Apparently they
>were not advised of conditions or accident type before they entered.
>
>The highest measured gamma and neutron dose rates reported so far are
>about 4.5 and 0.85mSv/hr, respectively, at the nearby site boundary.
>These reported measurements do not include information from the
>initial pulse or immediately thereafter.
>
>Although the building was not damaged, all fission products were
>released to the atmosphere.  [Room and building filters either failed
>or were not designed to handle fission products.]
>
>[I do not know if solution splashed out of the tank or if any
>radiological contamination was tracked about as people evacuated and
>responders moved in the area.]
>
>
>EMERGENCY RESPONSE OUTSIDE PLANT
>
>Plant personnel completed initial notifications to JCO officials
>within ten minutes.  Some notification information was not clear
>because at least one company official did not understand they
>were dealing with a criticality accident.  Apparently none of these
>officials instructed plant personnel to notify and/or establish
>communications with city or regulatory authorities.
>
>City authorities were notified approximately one hour after the
>initial pulse.  They apparently determined response actions for
>residents on their own, or with a little help from plant personnel.
>
>Residents were notified up to 2.5 hours after the first pulse.  About
>160 people within a 350m radius were evacuated until the afternoon of
>October 1st.  However, after a night in temporary shelters, some
>evacuated residents reportedly returned home to care for pets
>and/or retrieve fresh clothing.
>
>Authorities advised people within 10km to shelter (stay inside with
>doors and windows closed) at least until the evening of October 1st.
>Apparently others stayed inside as well because the city is said to
>have resembled a ghost town.
>
>Authorities also warned people they should not eat produce or drink
>milk from the area until testing was complete.  That ban was lifted by
>October 4th.  [I speculate authorities attempted to scale
>reactor-accident guidance to this accident.]
>
>It now seems responders were notified and activated separately from
>authority notifications.  Notifications to offsite responders might have
>warned offsite authorities.  Initial radiological responders were
>apparently from plant personnel but, as response continued, they
>might have been supplemented by personnel from other plants.  Other
>in-field and city responders were from the civil police, civil firefighters,
>and army.  The army's role is not indicated but their chemical warfare
>unit responded.
>
>
>MEDICAL TREATMENT
>
>Only the three most severely exposed workers exhibited radiation
>sickness symptoms.  They exhibited all classic early symptoms,
>including Na-24 in their vomit.  They were taken by helicopter to be
>treated by radiation experts at the National Institute for Radiological
>Science in Chiba Prefecture.  (One or more of these experts
>assisted Russian doctors treat victims of the Chernobyl accident.)
>
>These patients were placed in placed in sterile rooms, their individual
>symptoms are treated as needed, and all precautions are taken to
>avoid infections.  Doctors say it is most encouraging that all three
>patients have survived this long (October 7th a.m.).
>
>Fifty-four year old Yutaka Yokokawa received an estimated 3Sv (300
>rem).  Unlike the other two workers, he did not lose consciousness and
>walked into the hospital.  Apparently he could answer a few questions
>immediately after the first accident pulse but he was not fully
>coherent.  He recovered sufficiently to answer police questions
>October 2nd.  He is expected to recover fully in a month without a
>bone marrow or blood cell transplant, if he can avoid any serious
>infections.  However, he was listed in serious condition as of October 5th.
>
>Medical personnel indicate the other two workers received lethal
>radiation doses.  However, medical advances might save them.  Doctors
>started preparations for a blood cell transplant soon after these men
>were hospitalized.  Apparently their prognosis is hopeful if they
>survive these transplants by one month.
>
>Thirty-five year old Hisashi Ouchi received an estimated 17Sv (1700
>rem).  One of Ouchi's siblings [news articles first indicated a sister,
>now indicate a brother] is a near-perfect match.  This sibling agreed to
>be a donor and doctors gave the sibling medicines to stimulate cell
>production.
>
>Doctors decided a peripheral blood stem cell transplant is the most
>appropriate procedure.  (Stem cells are immature blood cells that can
>develop into either red or white blood cells, or platelets.)  This
>technique is about 10 years old, but this is its first use on
>radiation victims. It's advantages include no need to obtain bone
>marrow, requires less cell volume, and can be repeated at three-day
>intervals for a while.  Doctors performed the first transplant on
>October 6th, although news articles reported Ouchi showed signs of
>pneumonia just hours before the procedure.
>
>Thirty-nine year old Masato Shinohara received an estimated 10Sv (1000
>rem).  Since none on Shinohara's relatives is a close enough match,
>doctors decided an umbilical cord blood cell transplant is the most
>appropriate procedure.  Shinohara was transferred to Tokyo University
>Hospital and transplant preparations are proceeding.  However, news
>articles have not reported when the transplant will be performed.
>
>
>INVESTIGATION
>
>Accident investigation for response purposes began immediately.
>Whether of not responders had applicable procedures, their
>information-gathering methods apparently are similar to methods
>typically used in the USA.
>
>Further investigation apparently began in earnest the afternoon of
>October 1st.  JCO conducted an in-house investigation.  Police are
>conducting an extensive official investigation.  News articles do not
>indicate if the regulator (Science and Technology Agency) is
>also conducting an accident investigation, either jointly with or
>separate from the police investigation.
>
>So far the police investigation included interviews with Yokokawa on
>October 2nd, a raid of JCO's Ibaraki and Tokyo offices by October 4th,
>and further raids of JCO's Tokyo offices on October 5th.  Police are
>also searching the plant, but they cannot yet enter the accident room.
>Some news articles indicate police actions might be part of a criminal
>investigation (suspicion of JCO negligence and violating nuclear
>safety laws), separate from an accident investigation.
>
>
>MISCELLANEOUS CONSEQUENCES, RESPONSE, and REACTIONS
>
>Tokaimura resident attitudes range from calm to angry over nuclear
>safety issues.  However, most do not seem to be worried about
>long-term health effects from this accident.
>
>JCO established counseling services for area residents and farmers.
>
>JCO and its parent company, Sumitomo Metal Mining Co., promised to pay
>compensation.  JCO will also probably be fined.
>
>The regulator apparently has not [yet] decided to revoke JCO's license,
>although Kyodo News reported otherwise on October 6th.
>
>Stock in Sumitomo Metal Mining Co. dropped.
>
>Two Greenpeace activists measured radiation and/or contamination
>levels around the JCO facility and Tokaimura.
>
>Oversight (Science and Technology Agency and Nuclear Safety
>Committee) inspection effectiveness is questioned.
>
>The regulator ordered inspections for all 51 [power?] reactors.  News
>articles do not mention non-reactor nuclear facilities but they are
>apparently included in this inspection order.
>
>The regulator plans to search offices of 20 nuclear-related
>facilities.
>
>The Japanese government probably will not revise its nuclear policy,
>despite an apparant drop in public confidence.  However, legislators
>plan to write a nuclear emergency law and to revise an existing
>nuclear safety law.
>
>If USA experience is indicitive, the JCO plant's entire safety basis
>will be scrutinized and reanalyzed.
>
>USA and UK personnel are reviewing criticality safety programs and
>safety bases at all non-reactor and, possibly, reactor nuclear
>facilities in light of the accident.  Many reviews began before
>regulating authorities ordered these reviews.
>
>US President Clinton threatened to (or did) veto a bill regarding
>Yucca Mountain and nuclear waste.
>

************************************************************************
The RADSAFE Frequently Asked Questions list, archives and subscription
information can be accessed at http://www.ehs.uiuc.edu/~rad/radsafe.html