SL-1
|
|
The SL-1, the Stationary Low-Power Reactor Number One, was a U.S. experimental military nuclear power reactor. It was destroyed in the first nuclear power plant accident in the United States. Part of the Army Nuclear Power Program, during design and build it was called the Argonne Low Power Reactor (ALPR). It was intended to provide electrical power and heat for small, remote military facilities, such as radar sites near the Arctic Circle, such as those in the DEW Line. The design power was 3 megawatts (thermal). Operating power was 200kW electrical, 400kW thermal, for space heating. For testing it was located approximately forty miles west of Idaho Falls, Idaho, in the National Reactor Testing Station.
US_AEC_SL-1.JPG
| Contents |
Accident description
At 9:01 pm on January 3, 1961, after a shutdown of eleven days for the holidays, and during maintenance procedures, the SL-1 went prompt critical. In four milliseconds, the heat generated by the resulting enormous power surge caused water surrounding the core to begin to explosively vaporize. The water vapor caused a pressure wave to strike the top of the reactor vessel. This propelled the control rod upwards, which killed the operator who had been standing on top of the vessel, leaving him pinned to the ceiling. The other two military personnel, who were supervising the maintenance operations, were also killed. The victims were Army specialists John Byrnes and Richard McKinley and Navy Electricians Mate Richard Legg.
Reactor principles and events
In all water moderated reactors, to operate the reactor requires the presence of water to moderate (slow down) the neutrons produced by the nuclear reaction and allow the reaction to proceed. If the reaction becomes too great, several factors limit the reaction.
The first limiting factor is normal heat expansion of the core as its components heat up and spread apart, decreasing the reaction rate. For this design, that could automatically regulate the reaction if the period (rate of growth of power) was about 100ms or greater.
If the core expansion can't react fast enough or with enough reduction in power, the next limiting process is the water turning to steam, quickly and automatically removing the water moderator and shutting down the nuclear reaction. How long it takes for this to happen depends on the design of the reactor. In this design it was about 6ms before steam formation started, in part because of the thickness of the aluminum core cladding - a thinner cladding would have let it respond more quickly by not insulating the core as much.
As a small reactor, the SL-1 was designed with a main central control rod which was able to produce a very large excess reactivity if it was completely removed. The excess reactivity is a measure of how much more capacity there is to accelerate the nuclear reaction than is required to start a controlled nuclear reaction for power generation. The potential for excess reactivity is always required because the fuel becomes less reactive over time. A greater excess reactivity causes a faster increase of the rate of the nuclear reaction. In normal operation, the control rods are withdrawn only enough to cause sufficient reactivity for a sustained nuclear reaction and power generation.
In this accident, the very great reactivity addition produced a period estimated at 3.6ms. That was too fast for the heat from the core to get through the aluminum cladding and boil enough water to fully stop the power growth in all parts of the core.
Instead, the final control method happened in parts of the core: destructive vaporization and consequent conventional explosive expansion of the parts of the reactor core where the greatest amount of power was being produced most quickly. It was estimated that this core heating and vaporization process happened in about 7.5ms, before enough steam had been formed to shut down the reaction, beating the steam shutdown by a few ms.
Events after the power excursion
There were no other people at the reactor site. The ending of the nuclear reaction was caused solely by the design of the reactor and the basic physics of heated water and core elements vaporizing, separating the core elements and removing the moderator.
Heat sensors above the reactor set off an alarm at the central test site security facility. The first response crew, of firemen, arrived nine minutes later and initially noticed nothing unusual, with only a little steam rising from the building, normal for the cold (-20F) night. The control building appeared normal. On approaching the reactor building their radiation detectors jumped sharply, to above their maximum range limit, and they withdrew, unable to know whether they could safely proceed or for how long they could remain.
At 9:17PM a health physicist arrives and he and a fireman, both wearing air tanks and masks with positive pressure in the mask to force out any potential contaminants, approached the reactor building stairs. Their detectors read 25 Roentgens per hour as they started up the stairs and they withdrew.
Some minutes later a health physics response team arrived with Jordan Redectors, radiation meters capable of measuring gamma radiation up to 500 Roentgens per hour, and full-body protective clothing. One plus two firefighters ascended the stairs and from the top could see damage in the reactor room. With the meter showing maximum scale readings they withdrew rather than approaching the reactor more closely.
Around 10:30PM the supervisor for the contractor running the site and health physicist arrived. They entered the reactor building and found two mutilated men: one clearly dead, the other moving slightly. With a one minute and one entry per person limit, a team of five men with stretchers recovered the operator who was still breathing; but he didn't recover consciousness and died of his head injury at about 11PM. Even stripped, his body was so contaminated that it was emitting about 500 Roentgens per hour. They looked for but didn't find the third man. With all potential survivors now recovered, safety of rescuers took precedence and work was slowed to protect them.
On the night of 4 January a team of six volunteers used a plan involving teams of two to recover the second body. Radioactive gold-198 and copper-64 from his gold watch buckle and a screw in a cigarette lighter subsequently proved that the reactor had indeed gone critical.
The third man wasn't discovered for several days because of the wrecked state of the room. On 9 January a team of eight men in relays of two at a time with 65 seconds limit used a net and crane arrangement to recover his body from its pinned to the ceiling position.
The bodies of all three are buried in lead-lined caskets sealed with concrete and placed in metal vaults with a concrete cover. All had major physical injuries, including severed limbs and fragments of fuel assemblies in wounds. Richard Leroy McKinley is buried in section 31 of Arlington National Cemetery.
Cause
One of the required maintenance procedures called for the main control rod to be manually withdrawn approximately three inches in order to attach it to its automated control mechanism, from which it had been disconnected. Post accident calculations estimate that the main control rod was actually withdrawn approximately twenty inches, causing the steam explosion. The three most common theories proposed for this discrepancy are sabotage by one of the operators, inadvertent withdrawal of the main control rod, which was known to be "sticky," or an intentional attempt to "exercise" the sticky rod, to make it travel more smoothly within its sheath. The maintenance logs do not address what the technicians were attempting to do and thus the actual cause of the accident is unlikely to ever be known. There were scratches suggesting that at the time of the accident the operator was attempting to reinsert the control rod and prevent the runaway reaction.
Consequences
The remains of the SL-1 reactor are now buried near the original site.
The accident caused this design to be abandoned and future reactors to be designed so that a single control rod removal wouldn't have the ability to produce the very large excess reactivity which was possible with this design. Today this is known as the "one stuck rod" criteria and requires complete shutdown capability even with the most reactive rod stuck in the fully withdrawn position. The reduced excess reactivity limits the size and speed of the power surge which is possible.
The accident also showed that in a real extreme accident both the vaporizing of the core and the water to steam conversion would shut down the nuclear reaction, demonstrating in a real accident the inherent safety of the water moderated design against the possibility of a nuclear explosion.
In addition to a sudden power surge, a nuclear explosion requires sufficient force to hold the reacting nuclear components together for some time, achieved by surrounding the core with a carefully engineered, symmetrical, inward-facing conventional explosion, an element not present in nuclear reactors. Without that explosive compression to hold the vaporized core components together they just fly apart as they did in this accident, ending the reaction. Net result: a badly damaged reactor core but not a nuclear bomb.
Even without an engineered containment building like those used today, the building contained most of the radioactivity, though Iodine-131 levels on plants during several days of monitoring reached fifty times background levels downwind.
Radiation exposure limits prior to the accident were 100 Roentgens to save a life and 25 to save valuable property. During the response to the accident 22 people received doses of 3 to 27 Roentgens total body exposure and 3 doses above 27 R. In March 1962 the Atomic Energy Commission awarded certificates of heroism to 32 participants in the response.
The documentation and procedures required for operating nuclear reactors expanded substantially, becoming far more formal as procedures which had previously taken two pages expanded to hundreds. Radiation meters were changed to allow higher ranges for emergency response activities.
After a pause for evaluation of procedures the Army continued its use of reactors, operating the Mobile Low-Power Reactor (ML-1), which started full power operation on 28 February 1963, becoming the smallest nuclear power plant on record to do so. That design proved too advanced for the materials available and was eventually abandoned after corrosion problems. While the tests had shown that nuclear power was likely to have lower total costs, the financial pressures of the Vietnam War caused the Army to favor lower initial costs and it abandoned its reactor program in 1965.
See also
- BORAX Experiments, 1953-4, which proved that the transformation of water to steam would safely limit a boiling water reactor power excursion, as happened in this accident.
References
- INEEL Publication "Proving the Principle" (http://www.inl.gov/proving-the-principle/)
- Chapter 15 "The SL-1 Reactor" (http://www.inl.gov/proving-the-principle/chapter_15.pdf) 9.5MB PDF
- Chapter 16 "The Aftermath" (http://www.inl.gov/proving-the-principle/chapter_16.pdf) 4.3MB PDF
- US Department of Energy Idaho Freedom of Information Act request archives (http://www.id.doe.gov/foia/archive.htm), including eight documents on the SL-1 accident.
- "SL-1 Reactor Accident on January 3, 1961, Interim Report" (http://www.id.doe.gov/foia/IDO-19300a.pdf), May 1961. From the above page. 15.5MB PDF.
- SL-1 Accident (http://www.radiationworks.com/sl1reactor.htm)
- Atomic Energy Insights - July 1996 (http://www.ans.neep.wisc.edu/~ans/point_source/AEI/jul96/AEI_Jul96.html)
- Brain Candy #59 - The SL-1 Accident (http://home.neo.rr.com/catbar/brain_candy/brncdy59.htm)
- ALPR (http://www.anlw.anl.gov/anlw_history/reactors/alpr.html) summary page for this design.
- Dr. George Voelz, M.D. (http://www.eh.doe.gov/ohre/roadmap/histories/0454/0454toc.html#Fatal), one of two doctors at the site, oral history of the events.