Conferences, Videos & Testimony
Report of December 8, 1997 Workshop
by Kevin L. O'Neill
December 8, 1997
report author: Kevin L. O'Neill
with review and concurrence of David Albright,
Chair of the Source Term Subcommittee
Final draft submitted: February 18, 1998
IntroductionOn December 8, 1997 the Health Advisory Panel's (HAP) Source Term Subcommittee conducted a workshop for concerned stakeholders to review Radiological Assessment Corporation's (RAC's) estimates of plutonium releases from the 1957 fire in Building 71 at the Rocky Flats Plant. This workshop sought to dispel remaining controversies around the fire by providing stakeholders with detailed information about what is known about the 1957 fire and allowing these stakeholders to comment on and challenge RAC's methodology and conclusions. As in the past, the workshop provided an opportunity for RAC's principle investigator, Paul Voillequé, and the public to engage in a more detailed and thorough discussion of the fire than is possible at HAP meetings or other public meetings sponsored by HAP or the Colorado Department of Public Health. This workshop was the fourth in a series that has provided stakeholders with an opportunity to review and comment on RAC's assessments of plutonium releases from the plant. While RAC's estimates of plutonium releases from routine plant operations and from the 1969 fire have been the subject of previous workshops, RAC's analysis of the 1957 fire has been a recurring topic. By giving the public an opportunity to thoroughly understand RAC's assessments and to constructively challenge these assessments, this workshop series has proved to be a useful public education tool. A second purpose of this workshop was to present illustrative findings on uranium releases. These findings, presented by HAP member David Albright, were presented at the request of the Environmental Information Network (EIN), which has asked HAP to consider the health effects of these releases. The purpose of this report is to summarize the presentation and comments from the December 8, 1997 workshop and to recommend further issues for investigation.
Reanalysis of Plutonium Releases During the 1957 FireThe principal presentation during the workshop was given by Paul Voillequé, who described his methodology and offered a "crude estimate" of plutonium released from the fire. Voillequé introduced his presentation by underscoring the importance of the public's input in the effort to estimate plutonium releases from the fire. He noted that his original goal was to make a best estimate (with uncertainty) of the total plutonium release during the fire by using historical plant data, such as drawings and reports, and reported information about the fire. Voillequé noted, however, that his initial estimate was based on incomplete information about the fire, and therefore was highly uncertain. Since this initial estimate was presented to the HAP over a year ago, Voillequé said that he has been engaged in a follow-up analysis to consider a broad range of fire scenarios. He also has developed a release chronology to facilitate RAC's calculation of dose assessments based on meteorological data during the fire. Voillequé noted that this reanalysis also incorporated new information based on the suggestions, questions and comments from the public and from members of the HAP.
Chronology of Events and Building LayoutVoillequé described the known sequence of events during the fire and the layout of Building 71's ventilation systems. He noted that a key assumption in his analysis is that the fire started at about 9:45 PM, or about 20 to 30 minutes before it was discovered at 10:10 PM. He stressed that the actual time that the fire started remained unknown. Voillequé said that the fire probably started when plutonium "skulls" spontaneously ignited in a glovebox line in room 180 of Building 71. The plexiglass gloveboxes containing the plutonium and other combustible materials, such as rags and cutting oils, also caught on fire. When the fire was discovered, Voillequé said, the gloveboxes were already breached. Shortly after the fire was discovered at 10:10 PM, the building's ventilation fans were turned to high speed to clear smoke from the rooms. Initial efforts to extinguish the fire with carbon dioxide failed, so water was used. According to reports, the fire in room 180 was extinguished at 10:38 PM. At 10:39 PM, once the fire in room 180 had been extinguished, fire fighters reported an explosion, which they later described as a "whoosh" rather than a "bang." Voillequé concluded that this explosion should more properly be described as a "rapid deflagration" in the main filter plenum, where firefighters discovered a secondary fire. The force of the explosion or deflagration knocked out the filter plenum's four exhaust fans, but the supply fans likely remained on until 11:10 PM, when the building lost power completely. Voillequé noted that the fire in the main filter plenum was declared "knocked down" at 2:00 AM, approximately 4 hours after it was discovered. He interpreted this to mean that the there were no flames in the plenum, although filters were probably still smoldering. The fire was not declared "out" until 11:20 AM on the following day. Voillequé next described Building 71's ventilation system. He noted that the building was maintained under negative pressure. Air pressure in the gloveboxes was the lower than in the rooms to help prevent airborne plutonium from contaminating workers and equipment in the rooms and hallways of the building. Air from all gloveboxes was exhausted through a separate booster filter system before exhausting through the main filter plenum. Room air exhausted directly through the main filter plenum. The main filter plenum consisted of a single wall of filters bisecting a long room on the second floor of the building. Exhaust ducts from the booster system and from the rooms entered the plenum on the upstream side of the filter bank. On the downstream side, four fans exhausted the plenum to a long tunnel that exited the building and led to the stack.
Sources and Quantity of Plutonium Involved in the FireVoillequé next presented his estimate of the amount of plutonium involved in the fire, including: the amount of plutonium in room 180 that was involved in the fire, plutonium holdup in the booster system main filter plenum, and plutonium holdup in the booster system ducts. According to Voillequé, interview-based reports and material accountability records indicated that 40 to 50 kilograms of plutonium was located in room 180 at the time of the fire. Of this amount, an estimated 9 to 18 kilograms was involved in the fire. This material was in various forms and subject to different degrees of combustion. Voillequé next discussed his estimate of plutonium holdup in the booster system and main filter plenum filters at the time of the fire. Plant records were used to make these estimates; the plant took daily, shift-long air samples from the tunnel from the main filter plenum and from the exhaust ducts from the booster system and from main process rooms. Daily records were later summarized in monthly reports, giving the average and peak concentrations and, in the case of the tunnel from the main filter plenum, the number of samples taken. Voillequé said that the plant also collected room air samples when high exposure operations were being conducted. Plant data indicted that there were four large releases of plutonium to the main filter plenum prior to the fire that constituted the vast bulk of held up material. He estimated the quantity from these four events to be slightly more than 130 grams. Voillequé said that no direct measurements were taken of releases on the upstream side of the booster system filters. Therefore, the plutonium holdup had to be calculated by dividing the measured discharge from the booster system duct into an estimated filter efficiency of 98 to 99 percent for each of the booster system sets. He estimated the loading to the booster system to be 14 to 54 grams. Voillequé then discussed the quantity of plutonium holdup in the booster system ducts. He estimated that no more than 6 grams of plutonium was held up in the ducts at the time of the fire. This estimate was based on a linear extrapolation from data taken in the 1970s when plant production capacity was much higher than it had been in the 1950s. He also noted that since much of the plutonium in the ducts was already oxidized, it was not combustible at the time of the fire, therefore contributing little to the overall release. Voillequé's discussion of the amount of plutonium involved in the fire elicited much discussion from stakeholders. In particular, several stakeholders were concerned that sampling data from the plant was not representative, leading to a higher estimate of the loading to the plenum. Other stakeholders argued that daily releases to the main filter plenum should logically outweigh releases from four events; weren't there any other "upset conditions" that could have contributed additional releases to the filter? Voillequé acknowledged that the single-point sampling was not representative. Under ideal conditions, he said, it would be possible to conduct a study using tracers to determine the behavior of air flowing through the ducts. However, the duct system was replaced after the fire, so a tracer study was not possible. However, Voillequé argued that he had adopted methodology to account for, with uncertainty, all routine releases and "upset conditions." Voillequé explained that the plant began in the 1960s to take three in-line sampling points in the tunnel leading from the main filter plenum. Data from these points could be used to model concentrations in the tunnel and the ducts through Monte Carlo calculations. This methodology, he pointed out, would capture all "upset conditions" and formed the basis for his estimate of routine plant releases under a separate source term investigation for HAP. Some stakeholders were concerned that Voillequé underestimated the quantity of plutonium in the ducts at the time of the fire. These stakeholders argued that the use of glove box filters were not commonly used in the plant until the 1960s, and when they were used, workers frequently poked holes in the filters. Consequently, they argued, the vents should have contained "a large amount" of plutonium that could have been resuspended during the fire. Voillequé disagreed, arguing that his estimate of plutonium in the ducts was probably too high. By using plant employment data as a surrogate to measure capacity, Voillequé showed that production capacity grew dramatically in the 1960s and 1970s. By comparison, plant capacity prior to the fire was very low. By taking a linear extrapolation from data collected in the 1970s, Voillequé arrived at an estimate of duct holdup that was probably in excess of the true value.
Plutonium Releases, Oxidation Conditions and Filter Degradation ScenariosVoillequé then discussed the Airborne Release Fractions (ARFs) that he used to calculate the plutonium released in the fire. These values were derived experimentally and recommended for this study by expert reviews. Voillequé noted that the ARF for metallic plutonium had the greatest variability of all the ARFs for the different forms of plutonium involved in the fire. Therefore different assumptions about how the metallic plutonium oxidized in the fire had the greatest impact on the overall release. Voillequé discussed what was known about the oxidation conditions during the fire. Reports following the fire indicated that piles of plutonium oxide were found in the room, indicating relatively quiet conditions. However, flashes and minor explosions were also reported, indicating more vigorous oxidation. Since there was no empirical basis to determine combustion conditions, Voillequé considered a broad range of oxidation conditions involving equal quantities of metallic plutonium. He also assumed that the fire in room 180 grew more vigorous over time, leading to larger releases over time, until it was extinguished. Voillequé explained the release pathways from room 180 and through the vents. While the fire was contained in the glovebox, airborne plutonium traveled though the glovebox exhaust system to the booster system, and from the booster system to the main filter plenum. Once the glovebox in room 180 was breached, the main pathway was through the room air vent directly to the main filter plenum. Voillequé then described four filter degradation scenarios that he considered in his analysis. Booster system and main filter plenum filters would continue to catch plutonium released during the fire in room 180, albeit at decreasing efficiencies. The more rapidly the filters failed, the larger the quantity of plutonium that would be released directly up the stack. Voillequé also explained the importance of estimating releases over time, which enabled other RAC investigators to estimate public exposures and doses based on changing meteorological data. The final report, Voillequé said, will include plots of releases over time for each of the filter degradation scenarios used in this analysis.
- Under the first scenario, Voillequé assumed that the booster system fails within forty minutes of the fire's ignition, while the main filter plenum remains largely intact (30 percent penetrated) until the deflagration at 10:39 PM. Under this scenario, a total of 80 grams (median value) of plutonium was released to the environment.
- Under the second scenario, both the booster system and main filter plenum fail more quickly. The booster system fails within 30 minutes of the start of the fire and the main filter plenum fails (100 percent penetration) at the time of the deflagration. Under this scenario, an estimated 200 grams (median value) was released to the environment.
- Under the third scenario, the booster system and main filter plenum fail even more rapidly than in the previous scenarios. The main filter plenum fails five minutes before the deflagration. Under this scenario, an estimated 240 grams (median value) was released to the environment.
- Under the fourth scenario, the main filter plenum fails 15 minutes before the deflagration. Voillequé said that this scenario was unrealistic, since this scenario would not allow for a buildup of combustible gasses in the plenum to cause the deflagration. However, this scenario provides a worst-case, upper bound to the estimate. Under this scenario, an estimated 320 grams of plutonium (median value) was released to the environment.