INEEL/EXT-04-02530
Feasibility Study of Supercritical Light
Water Cooled Reactors for Electric Power
Production
Nuclear Energy Research Initiative Project 2001-001
Westinghouse Electric Co. Award Number: DE-FG07-02SF22533
Final Report
12th Quarterly Report
Principal Investigators: Philip MacDonald, Dr. Jacopo Buongiorno,
Dr. James W. Sterbentz, Cliff Davis, and Prof. Robert Witt
Telephone: 208-526-9634
Fax: 208-526-2930
Email: pem@inel.gov
Collaborating Organizations:
University of Michigan
Principal Investigators: Prof. Gary Was, J. McKinley, and S. Teysseyre
Westinghouse Electric Company
Principal Investigators: Dr. Luca Oriani, Dr. Vefa Kucukboyaci, Lawrence Conway,
N. Jonsson, and Dr. Bin Liu
ii
Executive Summary
The supercritical water-cooled reactor (SCWR) is one of the six reactor technologies selected for research
and development under the Generation-IV program. SCWRs are promising advanced nuclear systems
because of their high thermal efficiency (i.e., about 45% vs. about 33% efficiency for current Light Water
Reactors, LWRs) and considerable plant simplification. SCWRs are basically LWRs operating at higher
pressure and temperatures with a direct once-through cycle. Operation above the critical pressure
eliminates coolant boiling, so the coolant remains single-phase throughout the system. Thus the need for
a pressurizer, steam generators, steam separators and dryers is eliminated. The main mission of the
SCWR is generation of low-cost electricity. It is built upon two proven technologies, LWRs, which are
the most commonly deployed power generating reactors in the world, and supercritical fossil-fired
boilers, a large number of which is also in use around the world.
The reference SCWR design for the U.S. program is a direct cycle system operating at 25.0 MPa with
core inlet and outlet temperatures of 280 and 500 °C, respectively. The coolant density decreases from
about 760 kg/m
3
at the core inlet to about 90 kg/m
3
at the core outlet. The inlet flow splits with about
10% of the inlet flow going down the space between the core barrel and the reactor pressure vessel (the
downcomer) and about 90% of the inlet flow going to the plenum at the top of the rector pressure vessel
to then flow downward through the core in special water rods to the inlet plenum. Here it mixes with the
feedwater from the downcomer and flows upward to remove the heat in the fuel channels. This strategy
is employed to provide good moderation at the top of the core. The coolant is heated to about 500 °C and
delivered to the turbine.
The purpose of this NERI project was to assess the reference U.S. Generation IV SCWR design and
explore alternatives to determine feasibility. The project was organized into three tasks.
• Task 1. Fuel-cycle Neutronic Analysis and Reactor Core Design.
• Task 2. Fuel Cladding and Structural Material Corrosion and Stress Corrosion Cracking.
• Task 3. Plant Engineering and Reactor Safety Analysis.
Task 1. Fuel-cycle Neutronic Analysis and Reactor Core Design
Metallic and oxide fertile fuels in a fast-spectrum SCWR were investigated during Year 1 to evaluate the
void and Doppler reactivity coefficients, actinide burn rate, and reactivity swing throughout the
irradiation cycle. These results were reported in the 1
st
Quarterly. A variety of other core arrangements
and moderator types for a thermal-spectrum SCWR were also assessed during the three years of this
project. Detailed results from the solid moderator studies were presented in the 3
rd
Quarterly and 4
th
Quarterly Reports and in two papers by Buongiorno and MacDonald (2003a & 2003b). Results from an
analysis of an alternative thermal-spectrum SCWR design based on vertical power channels, hexagonal
fuel assemblies, and water moderation between the fuel assemblies were reported in the 2
nd
Annual
Report. Also reported in the 2
nd
Annual Report were the results of the steady-state thermal-hydraulic
analyses
for two other thermal spectrum SCWRs, one design with solid moderator rods, and one design
with water filled moderator rods.
This report presents in Chapter 3 the results of a two neutronic evaluations for two different SCWR fuel
assembly designs. The first evaluation is for a 25x25 fuel assembly that used MA956 oxide dispersion
steel for the fuel rod cladding, water rod duct, and assembly duct materials. The second is for a 21x21
fuel assembly that used silicon carbide (SiC) for the fuel rod cladding, water rod duct, and assembly duct
materials.
iii
The 25x25 fuel assembly with MA956 cladding and duct material contains a 6x6 array of square water
rods interspersed uniformly within the UO
2
fuel pin array to increase neutron moderation and assembly
reactivity. The assembly exhibits many desirable neutronic characteristics that include sufficient
reactivity to achieve burnups of at least 31.0 GWD/MTU, a strongly negative Doppler coefficient, and a
negative void worth for both the coolant and water rods. The assembly also exhibits some characteristics
that may complicate the design, such as a wide spread in the required radial enrichments (3.2-12.4 wt%
U-235) to flatten the radial power profile, and relatively lower reactivity than in LWRs because of the
high parasitic neutron absorption of the MA956 cladding and duct material. In addition, the assembly
axial power profile exhibits a strong sensitivity to small changes in axial enrichments that may lead to
power oscillations under normal operation, if not properly controlled with burnable poisons and control
rods. The sensitivity is believed to be primarily due to the interplay between the non-uniform axial water
density profiles that affect neutron moderation and the time-dependent axial burnup of the fissile heavy
metal.
The 21x21 fuel assembly with the duplex SiC/SiC for the fuel pin cladding and fuel assembly duct
material exhibited better neutronic characteristics than the MA956 assembly with steel structures. This
assembly showed a significant increase in core reactivity due to the relatively low parasitic neutron
absorption of the silicon carbide. This low parasitic neutron absorption in turn translates into
significantly higher burnup (41.0 GWD/MTU) when compared to the burnup of the MA956 steel
assembly (31.0 GWD/MTU). In addition, the SiC assembly exhibits a strong negative Doppler
coefficient (–2.5 pcm/°C), and negative void worth for both the coolant and water rods. The assembly
does however require again a relatively wide spread in the fuel rod radial enrichment (3.2-12.4 wt% U-
235) to flatten the radial power profile at beginning-of-life conditions, and in addition would require at
least a three-zone axial enrichment to flatten and center the unrodded axial power profile about the core
midplane at beginning-of-life. As with the MA956 assembly, the fuel assembly axial power profile
appears to exhibit a strong sensitivity to small changes in axial enrichments (and therefore burnup) that
could lead to power oscillations under normal operation, if not properly controlled with burnable poisons
and control rod movement.
Task 2. Fuel Cladding and Structural Material Corrosion and Stress
Corrosion Cracking
The existing data base on the corrosion and stress-corrosion cracking of austenitic stainless steel and
nickel-based alloys in supercritical water is very sparse. Therefore, the focus of this work has been
corrosion and stress corrosion cracking testing of candidate fuel cladding and structural materials. During
Year 1, a high temperature autoclave with carefully controlled chemistry and containing a constant rate
mechanical test device was built and tested at the University of Michigan.
During Years 2 and 3, a variety of austenitic and ferritic-martensitic alloys were tested. The results of
that work are presented in Section 4 of this report and briefly summarized below. The austenitic alloys
were tested in deaerated water (dissolved oxygen of the order of a few ppb) at temperatures between 400
and 550 °C and they all showed varying degrees of susceptibility to intergranular stress-corrosion
cracking (SCC). Susceptibility was determined by examination of the fracture surface, the gage surface,
and by analysis of cross-sections of the tensile bars. All these measurements are required to provide a
complete description of the cracking behavior. Alloy 625 is the most susceptible, displaying the highest
degree of intergranular fracture and some of the deepest cracks along with a very high crack density. The
304L stainless steel is the next most susceptible material, showing the deepest intergranular cracks.
Alloys 690 and 316L are the least susceptible austenitic alloys from all measures considered; crack
density, crack depth and crack length.
iv
The degree of intergranular SCC of the austenitic alloys increases with increasing temperature. As the
temperature increases, the crack density decreases but the crack length and depth increase, resulting in a
net increase in the intergranular cracking severity as measured by the crack length per unit area.
There is very strong temperature dependence to the oxidation behavior of the austenitic alloys. The
oxidation rate, as measured either by the weight gain or oxide thickness, increases faster with increasing
temperature. By 550 °C, the austenitic alloy oxide thickness is approaching 10 µm within a few hundred
hours. The predominant feature among all of the austenitic alloy oxides was the two-layer structure
consisting of an iron-rich outer layer and chromium-rich inner layer. X-ray diffraction has shown that the
outer layer was magnetite, Fe
3
O
4
. The outer oxide on the Alloy 690 was probably NiO.
The ferritic-martensitic alloys do not display any evidence of intergranular SCC as determined by fracture
surface and gage surface analysis. They all display strain softening and ductile rupture. However, the
oxidation rates of the ferritic-martensitic alloys are very high compared to the austenitic alloys. At 500
°C, the ferritic-martensitic oxidation rates are a factor of 10 greater than those for the austenitic alloys at
the same temperature. These alloys also display a two-layer structure, in which the outer layer is
identified as magnetite, Fe
3
O
4
. The O/M ratio of the inner layer is closer to hematite, but the structure of
the inner oxide layer was not verified. The addition of 100 ppb oxygen to the water at 500 °C resulted in
a reduction of the total oxide thickness by about 10% and a slight increase in the O/M ratio. These results
are consistent with the objective of combined water chemistry control.
Task 3. Plant Engineering and Reactor Safety Analysis
SCWR Core Thermal Hydraulic Design Assessment. The Westinghouse Electric Company tasks
included an assessment of the reference core thermal hydraulic design. A complete review of the
Westinghouse SCWR core assessment activities is provided in Westinghouse Report STD-ES-04-45,
while in this report the focus is mostly on the final analyses and the main conclusions of the analyses
effort.
The first step in performing the core thermal-hydraulic assessment was identification of the design limits
that were then used to evaluate the acceptability of the core design. Section 2 of STD-ES-04-45 and the
2
nd
and 3
rd
Quarterly reports for this NERI project (MacDonald et al. 2002a and 2002b) provide the
considerations used in defining the design limits for the SCWR. Once the boundaries of the analysis were
defined, simplified calculations were performed to provide an initial characterization of the design. These
analyses are summarized in Section 3 of STD-ES-04-45. Based on the results of the simplified analysis,
it was concluded that the SCWR, due to its very large enthalpy rise along the core, is sensitive to small
deviations from nominal conditions, especially variations in the flow to power ratio. Thus, even small
effects due to various hot channel factors (coolant flow channel tolerances, operational variations, etc.)
might have a large impact on the peak cladding temperature of some fuel rods. This was considered a
major feasibility issue for the SCWR, and thus it was decided to perform detailed subchannel analysis of
the SCWR core to provide a more in depth assessment of this issue.
The W-VIPRE subchannel analyses code was adapted for the analysis of supercritical water, and new
correlations that are considered adequate for SCWR analyses were implemented in the code. A complete
characterization, including sensitivity studies, of the SCWR with the modified VIPRE core is documented
in Section 5.1.2. Based on these results, sufficient information was available for a preliminary thermal-
hydraulic optimization of the SCWR core design. Temperature profiles for various core geometries were
then analyzed with two different objectives: (1) to identify an optimal geometry that minimizes the
temperature differences between core channels, and (2) to confirm and characterize the sensitivity of the
v
temperature profile to the local flow to power ratios. The need of maintaining uniform conditions at the
exit of the core is dictated by the fact that safety limits need to be verified for the limiting fuel rod, while
the overall plant performance depends on the average core exit conditions. Thus, a uniform temperature
distribution minimizes the “wasted” design margin. These analyses are documented in Section 5.1.3
(Section 5 of STD-ES-04-45).
Results from the optimization studies suggest that it is possible to obtain a better temperature profile
(hence, lower hot channel factors) by employing a more complex assembly configuration. Based on the
results of this study, the design should use a geometric configuration with 10mm outside diameter fuel
rods for the coolant channels facing the water rods and at the assembly periphery, 9.5mm outside
diameter fuel rods for the assembly corners, and 10.2mm outside diameter fuel rods for all other
positions. While this study shows a path to obtain an acceptable thermal hydraulic design, it also
provides the designer with an important design issue: the flow is clearly extremely sensitive to small
variations in the channel flow area. Therefore, rod bowing and even the tolerances in rod dimensions
could be crucial in terms of the temperature peaking. This sensitivity, which can be attributed to large
channel enthalpy rise coupled with a region of low-density coolant and high exit velocities, renders the
design uncertain.
Based on the results of this study, it appears that the reference SCWR design is not feasible. Although
additional design and analysis might allow the recovery of some margin, it is unlikely that a SCWR
assembly and core design can be developed that provides acceptable performance (i.e. low enough hot
channel exit temperature). Therefore, the SCWR core design remains a major feasibility issue for which a
solution has yet to be achieved at this stage of the program.
An Evaluation of an Innovative Safety Concept for the SCWR. Preliminary investigations of the
safety characteristics of a SCWR performed by INEEL and the Westinghouse have resulted in the
development of a novel safety concept for this Generation IV reactor. Previous analyses have shown that
the SCWR can meet transient thermal limits for events initiated by loss of main feedwater only if a large
capacity auxiliary feedwater system is actuated rapidly. However, the required rapid initiation of
auxiliary feedwater was judged to pose significant technical and economic challenges. Consequently,
Westinghouse developed an innovative conceptual design that uses a passive circulation system to
mitigate the effects of loss of main feedwater. This safety concept utilizes two, relatively small,
feedwater tanks that store water for reactor cooling during normal operation and provide sufficient
cooling capacity to mitigate the effects of a loss of main feedwater. Main coolant pumps similar to those
utilized in advanced light water reactors provide the head required to circulate the flow in the reactor.
Although the proposed concept takes advantage of the SCWR once-through, direct cycle concept during
normal operation, it allows the establishment of a recirculation path in the system following containment
isolation, with an isolation condenser that provides long-term decay heat removal.
The safety characteristics of the design were evaluated for loss-of-flow transients using the RELAP5-3D
computer code. The results of these evaluations confirmed the potential of the design. Acceptable short-
term results following loss of flow were obtained by adjusting the coastdown characteristics of the main
coolant pumps. Acceptable long-term decay heat removal following loss of flow was obtained with 300
or more tubes in the isolation condenser.
Preliminary evaluations of loss-of-coolant accidents were also performed. The analysis of an accident
initiated by a large cold leg break showed that significantly lower cladding temperatures were obtained
after the blowdown peak in the proposed design than in a simple, once-through design due to the
recirculation loop and the added coolant inventory provided by the feedwater tanks. The milder evolution
![Figure 144. Temperature distribution [K] for case of Figure 132 using a k-ω turbulence model.](/figures/figure-144-temperature-distribution-k-for-case-of-figure-132-o7am0ny3.webp)
![Figure 145. Temperature distribution [K] for case of Figure 136 using a k-ω turbulence model.](/figures/figure-145-temperature-distribution-k-for-case-of-figure-136-29zei2ky.webp)
