lthough its design uses existing reactor technology, the Advanced Neutron Source (ANS) will be better suited to neutron research than any other research reactor in the world. To ensure that this unique Department of Energy (DOE) facility will operate effectively and safely as an intense, steady-state source of neutrons for experiments, research is being conducted at ORNL and other laboratories. The best expertise and facilities throughout the nation are being called upon to work with ORNL researchers to achieve one goal: to make the ANS the best possible reactor for meeting the needs of the scientific, technological, and medical communities.
Doug Selby, head of the ANS Project's research and development program, says that 13 types of program activities address technical issues associated with the ANS, which will be used for neutron scattering research, radioisotope production, and materials irradiation. These activities and some of the issues addressed are described in this article.
"This work provides input to preliminary and final safety analysis reports and support to design teams," says Selby. For the conceptual design, the task group recommended a two-element core, somewhat larger than ORNL's High Flux Isotope Reactor (HFIR) core, with 93% enriched uranium silicide (U2Si3) fuel--that is, fuel containing uranium atoms, 93% of which are fissionable uranium-235.
Can the ANS be operated with fuel of lower enrichment than 93%? This issue raised by DOE and various review groups has been studied by a DOE task force headed by Brookhaven National Laboratory that includes researchers from ORNL, INEL, and Argonne National Laboratory (ANL). This study that has just been completed included reactor physics and some thermal-hydraulic evaluations of various alternative enrichments.
"We looked at several enrichments_50%, 35%, and 20%," Selby continues. "It appears that fuel with an enrichment as low as 50% could be used without much penalty in reactor performance_that is, neutron flux and other parameters. If the enrichment level is dropped below 50%, the performance tends to degrade rapidly. Right now we get a peak thermal neutron flux of approximately 7.4 times 10E19 per square meter per second. At the peak reflector flux location, we can more than likely get within 15 to 20% of that value with enrichment in the 50% range.
"However, for both 50 and 35% enrichment, we must switch from a two-element core to a three-element core. As a result, refueling outages (downtime) will be longer, the problem of what to do with spent fuel elements will be compounded, and fuel costs will be greater. Also, even with the three-element core, at 20% enrichment enough fuel cannot be put in to maintain a reasonable core life at a reasonable power level."
"Our evidence to date is that under ANS conditions, uranium silicide performs even better as a fuel than uranium oxide," Selby says. "At the higher burnup levels of the ANS, the silicide fuel swells less, has better thermal conductivity, and better retains fission products."
Uranium silicide was first tested and used as a high-enrichment fuel in 1987 by Copeland and his team using the Oak Ridge Research Reactor; U3Si2 is now being tested in many countries, including Japan, France, Germany, Indonesia, Denmark, Sweden, and Russia. After the 1987 tests at ORNL, a team led by Colin West, ANS Project director, Copeland, and Selby proposed uranium silicide as a reference fuel for the ANS.
"We are now using the HFIR to test capsules containing uranium silicide fuel specimens," Selby says. "The fuel can be taken up to well above the average fission density we would expect to see in the ANS. Two capsules were irradiated in the reactor, each for one HFIR cycle, and a third capsule will be irradiated later this year. Argonne researchers are currently performing postirradiation examinations of the specimens from the first two capsules. Microstructural evaluation techniques, such as metallography and scanning electron microscopy, are used to judge the retention of fission products, particle swelling, and particle-matrix interactions. This information is used to validate the mechanistic fuel performance model also being developed at ANL."
Selby says that Argonne and ORNL researchers have published some results of their fuel examinations, further reinforcing the belief that uranium silicide is the best fuel choice.
"The Argonne scientists," he says, "found that the interactions between the fuel and aluminum cladding form a shell of aluminide around the fuel particles, enhancing their stability. Results to date indicate that the uranium silicide fuel will allow the ANS to reach its full planned performance."
Remedies proposed for this corrosion problem have included cladding surface treatment, changes in coolant chemistry, and changes in coolant flow and velocity. To test the effects of these changes on fuel plates at different simulated power densities, ORNL built the Corrosion Test Loop Facility, a high-pressure heated water loop that has been operating since January 1988. The facility is designed to measure the corrosion of aluminum and other effects under the heating levels and coolant velocity expected for the ANS fuel.
The operation of this facility at ORNL's Engineering Technology Division is conducted by the ANS Project's Corrosion Tests and Analyses task group led by Dick Pawel of the Metals and Ceramics Division. The facility is more than halfway through its planned program of 64 tests.
Selby says that the corrosion of the aluminum alloy cladding can affect reactor operation in two ways. First, film growth influences system temperatures. Second, internal reactions may limit structural and containment capability.
"The severity of internal reactions was observed to be related to the heat flux," Pawel says. "These reactions occurred only after thicker oxide films had formed and spallation had started." Spallation is a loss of portions of the oxide film as a result of buildup of mechanical stresses in the system.
The HFIR also has aluminum cladding, and experiments performed in the 1960s established oxide growth and constraints under HFIR operating conditions. "In the ANS studies we found that some assumptions made when the HFIR was designed were not applicable in the range of ANS operating conditions," Selby says. "We found that we must consider additional variables that are important to growth of the oxide film. ANS conditions not found in the HFIR, such as higher heat fluxes and coolant velocities, must be taken into consideration when determining oxide growth."
New results from tests at the ANS Corrosion Test Loop Facility indicate that oxide growth on cladding can be controlled to acceptable levels if the heavy-water coolant has the right chemistry.
"The rate of oxide growth on aluminum alloy cladding is sensitive to the acidity level of the coolant," Pawel says. "We found that controlling the water chemistry to keep the pH at or slightly below 5.0 inhibits oxide growth.
"Under some conditions," Pawel continues, "we have linked the lower oxide growth to formation of a thin iron-rich film on the outer surface of the oxide film. In addition to creating a viable data base with our experiments, we are trying to unravel the interactions of water chemistry, film properties, and oxide growth rates."
"We developed a new correlation, or set of equations, that predicts oxide growth on aluminum cladding as a function of heat flux, coolant temperature, and flow rate," Selby says. "This correlation has been published in the literature."
The task group is also studying the remote possibility that the ANS might have a loose reactor part that could, unless appropriate precautions are taken, partially block coolant flow. Occasionally, a component of a pump, experiment, or refueling machine might drop into the heavy-water coolant stream. In the ANS, as in other reactors, strainers will catch loose items at various locations in the loop to reduce the probability of their entering the core. Even so, coolant flow blockage by a loose piece, although an unlikely event, is still a potential concern.
To address this concern, the Core Flow Tests task group constructed a Flow Blockage Test Facility to determine the sizes of blockages that the ANS could take without incurring fuel plate damage. The researchers are using the facility to determine the effects on fuel plate temperature of different sizes and locations of blockages and the distances downstream of the blockages where these effects persist. Data have been obtained on fuel plates' heat transfer capability and coolant velocity with blockages of 10% and 25% on the edge of the channel and 10%, 15%, and 35% in the center of the channel.
"If a blockage occurs, there is some area behind it where the cooling capability is degraded," Selby says. "Farther downstream the coolant flow re-forms and the core is subjected to normal cooling. But between those points, the fuel plate receives less cooling than expected.
"We're trying to measure actual cooling capabilities and fuel plate temperatures downstream of a blockage," Selby says. "The measurements are obtained from a thermosensitive film placed on the plate downstream of the blockage. Different colors in a computer image of the film show the variations in film temperatures, indicating the different degrees of cooling to which the fuel plate is subjected.
"Our goal is to determine how big a blockage we can accommodate in our design of the ANS," Selby adds. "This information will help determine how large the ANS strainer should be."
Passive safety features will be incorporated into the ANS design to make it safer during operation and shutdown. During shutdown the ANS will be designed to use the natural heat from the fuel's decay to drive the heavy-water coolant and continue to remove heat from the core. The Core Flow Tests task group will use a Natural Circulation Test Facility to be built to test this passive design feature for low-flow conditions. "In the tests," Selby says, "we will determine when natural circulation driven by heat in the fuel elements supplies sufficient cooling for the core so that forced cooling is no longer needed following shutdown."
The task group is headed by Brian Worley and consists of his colleagues in EPMD with some support from the I&C Division.
"We obtained measurements of reactor physics parameters for ILL and its critical experiments, built reactor physics models, and compared our predictions with their experimental results," Selby says. "A recently published report documents the work performed at INEL and shows good agreement between our calculations and ILL's experimental data."
The third and final phase of the preoperational reactor physics validation, Selby adds, is to perform a series of reactor physics experiments specific to the ANS geometry. Five sites being considered for those experiments are the Los Alamos, Oak Ridge, and Sandia national naboratories as well as Chalk River Laboratory in Canada and Winfrith Technology Centre laboratory in England.
"We have run four tests using epoxy plates and scaled the results to aluminum fuel plates like those that will be used in the ANS," Yahr says. "We chose epoxy plates because we can test theoretical models using lower pressures and coolant velocities than would be needed to test aluminum plates. We are now preparing to test aluminum plates to show their stability under ANS conditions."
The ANS will provide the highest neutron flux ever, so efforts are being made to ensure that its structural materials will withstand damage from neutron irradiation. The same task group has irradiated two capsules (Hansal-1 and Hansal-2), which contained specimens made of aluminum 6061-T651, the reference design material for many ANS structural components. Tests on specimens from the first capsule, which was irradiated in the HFIR for three reactor cycles, showed little effect of irradiation on fracture toughness except at the highest temperature tested. Tests have not yet started on specimens from the second capsule, which was irradiated in the HFIR for about 2 years (21 reactor cycles) for an accumulated total thermal neutron fluence of 8 1026 per square meter. This fluence simulates the degree of radiation damage that might be expected in some ANS components.
"Information from tests like these," Selby says, "will be used to develop a materials data base, which will be used to complete the final design of the ANS." Data are also being collected so that researchers can build computer models to predict thermal stresses in candidate reactor structural materials under various ANS conditions.
Cold sources at the ILL, for example, allow liquid deuterium to boil. The lighter boiled-off gas then rises to a heat exchanger where it is reliquefied. This heat exchanger is cooled by helium gas refrigerant from a refrigerator system.
"Because of the high heat load in the two ANS cold sources, a system that works with both a liquid and a gas leaves many complex problems to be solved," Selby says. "It would be extremely hard to assess the relative proportions of liquid and gas in the boiling vessel for all possible operating conditions."
ORNL researchers have proposed an all-liquid cold neutron source using deuterium. They say that a more efficient moderator can be produced if the liquid is cooled well below its boiling point and circulated in fully liquid form by a mechanical circulator. An external review group including representatives from BNL and the National Institute of Standards and Technology concurs that this concept offers the lowest technical risk.
The ANS Project's Cold Source Development task group, under Trevor Lucas of the Engineering Technology Division, has developed such a system. Two prototype liquid circulators have been ordered and will be tested and developed by ORNL researchers beginning this year. Deuterium, an isotope of hydrogen, has been chosen as the neutron moderator because of its ability (unlike normal hydrogen) to slow down neutrons without absorbing them.
In the ORNL cold source, liquid deuterium will enter the moderator vessel at 20 K. The 30-kW heat load will raise this temperature to 25 K, which is still several degrees below its boiling point. To prevent boiling, the pressure of the liquid will be controlled. A mechanical circulator will move the liquid around the loop between the cold source and the heat exchanger, where the helium refrigerant recools the deuterium.
Each cold source vessel will be located close to the reactor core to take maximum advantage of the available neutron flux. The slowed, or cold, neutrons will then be guided through cold neutron guide tubes to experiments as far away as 70 to 100 meters without significant losses in the strength of the neutron beam.
In the past 5 years, new devices have been developed by the ANS Project's Instrument and Beam Tube Development task group under Hayter and Ralph Moon, associate director of ORNL's Solid State Division. One device, developed by Herb Mook and Hayter, both of the Solid State Division, received an R&D 100 Award in 1989. Called a transmission polarizer, this device produces polarized beams for neutron scattering research while minimizing the loss of neutron intensity in the polarization process. In other words, it deflects neutrons of one spin state without affecting neutrons of the other spin state. The device was used by researchers to study the magnetic properties of materials such as high-temperature superconductors.
The ORNL polarizer allows efficient use of short-wavelength neutrons (2.5 angstroms) that will be abundantly available in the ANS. It uses a stack of 80 magnetic supermirrors_excellent reflective surfaces for neutrons.
Supermirrors are being developed at ORNL in conjunction with NIST to minimize the losses of cold neutrons in a beam as they are transported from the reactor to the experimental sample in the guide hall. "We are developing new surfaces and new ways to transmit neutrons from the source to the experiment and reduce the neutron losses between both points," Selby says. "If improvements can be made in neutron beam transmission, additional experiments may be moved out of the space-constrained containment area to a larger neutron guide hall area that has lower radiation background."
The key to designing an effective supermirror for a cold neutron-beam guide is to make a surface that continuously reflects back those low-energy neutrons that leak out of the beam.
"Once neutrons enter a beam guide, if the surface has the right shape and composition, they keep reflecting off surfaces with only small losses until they reach the experiment," Selby says. "For any given neutron energy, we try to improve the angle of the guide tube's acceptance--the neutron angle that will allow the neutrons to be reflected down the guide tube rather than transmitted through the tube wall. For a given neutron energy, there is some angle that will reflect a neutron perfectly while neutrons of higher energies pass through the surface. We're trying to improve surfaces to increase the angle so more neutrons can be delivered to experiments."
Neutrons from the reactor can also be used to produce positrons (positively charged electrons) for research. A new conceptual design study for an ANS positron source has been completed. In this concept, a copper-64 target would be irradiated in the intense neutron flux near the reactor core, producing copper-65. During decay, copper-65 emits positrons. Copper-64 was chosen over other metals because it does not transmute to materials that decay into long-lived radioactive wastes. It was selected instead of krypton gas as a positron source to eliminate concerns about the consequences of an accidental release of a radioactive gas. Examples of applications for positrons are measurement of electron momentum densities in metals and alloys and three-dimensional mapping of material defects.
"The energy of neutrons entering the block from all directions would be raised by transfer of energy from the graphite block," Selby explains. "The neutron energy spectrum shifts into the desired range--say, from 0.04 electron volts to the 0.2 to 0.4 electron volt range--and are then transported to experiments via beam tubes."
"We also evaluate high-energy neutron and gamma radiation that is unfortunately transmitted through the biological shield via the beam tube path," Selby says. "We try to determine beam tube orientations that minimize this gamma and high-energy neutron radiation that is considered noise to the experimenters. We found that neutron and gamma contamination of beam tubes is high if they are oriented so that they are 'looking' at the core. If we change the beam tube orientation to look away from the core, fast neutron and gamma flux is reduced. Because the thermal neutron source is nearly isotropic at the beam tube mouth, any beam orientation will work for neutrons in the energy range desired by researchers."
Selby says the "primary biological shield" planned for the ANS is reinforced heavy-duty concrete. Other materials are being looked at to provide shielding for areas where radioactive fuel elements are moved for storage.
The task group is also using computer modeling to predict the neutron and gamma heating of various reactor components under ANS conditions. By determining how much energy is deposited, the researchers can learn how much heat must be removed to keep the components cool. This information will affect the design of many components in the reflector region.
Selby says that one of the main challenges of the design is to obtain the desired control system response time, which requires fast actuation of control components. Through use of computer simulations, the I&C team has developed a series of rod performance requirements and is continually interacting with the engineering design team to evaluate design options.
The control rods are just an example of how modern design techniques are influencing the final outcome. Throughout the design, the ANS Project has taken advantage of the availability of new computer-based modeling tools. One example is the ANS Dynamic Model, which allows designers to simulate with a personal computer the dynamic response of the ANS core and its cooling circuits. This model has been used extensively to test different design options, and it has also been used to set initial requirements for the control and plant protection systems.
The I&C research team in cooperation with the Electric Power Research Institute is developing a reactor protection system using application-specific integrated circuits (ASICs). These circuits are being investigated for potential use in commercial power plants and in the ANS. Selby says that ASICs can be designed to perform as well as computers, and the task team is working to show that they are simpler, cheaper, and more reliable than software-based systems.
A proposed test would examine the seal that separates the primary loop from the neutron reflector area. "We hope to design a device to increase pressure on one side of a prototypic seal to test how much pressure the seal can take before weakening," Selby explains. "We also want to test latching devices for control rods and measure the time it takes to release them."
Many ANS operations will be performed remotely, so ORNL is testing certain remote operations. "Key features of certain components will be modeled with high fidelity and the interfaces with remote operation devices will be examined," Selby says. These tests will provide indications of potential remote handling interface problems as early in the design as possible.
In conclusion, Selby notes that almost half of the research and development on the ANS Project is scheduled to be performed at ORNL because of the wide-ranging reactor expertise and facilities here. However, he adds, "We work very hard to determine the best place to do different pieces of ANS research. We are building a team of experts and using the best facilities from all over the nation to solve problems to ensure safe, reliable operation of the ANS."
[ Next article | Search | Mail | Contents | Review Home Page | ORNL Home Page ]